Optofluidic identification of single microorganisms using fiber‐optical‐tweezer‐based Raman spectroscopy with artificial neural network

Rapid and accurate detection of microorganisms is critical to clinical diagnosis. As Raman spectroscopy promises label‐free and culture‐free detection of biomedical objects, it holds the potential to rapidly identify microorganisms in a single step. To stabilize the microorganism for spectrum collection and to increase the accuracy of real‐time identification, we propose an optofluidic method for single microorganism detection in microfluidics using optical‐tweezing‐based Raman spectroscopy with artificial neural network. A fiber optical tweezer was incorporated into a microfluidic channel to generate optical forces that trap different species of microorganisms at the tip of the tweezer and their Raman spectra were simultaneously collected. An artificial neural network was designed and employed to classify the Raman spectra of the microorganisms, and the identification accuracy reached 94.93%. This study provides a promising strategy for rapid and accurate diagnosis of microbial infection on a lab‐on‐a‐chip platform. An optofluidic method is developed for identifying single microorganisms in microfluidic channels using fiber‐optical‐tweezer (FOT)‐based Raman spectroscopy with artificial neural network (ANN). The individual moving microorganisms are trapped by the FOT incorporated into the microfluidic channel, followed by the collection of their Raman spectra. With the ANN to analyze the spectral features of each of the 15 microorganism species, the average identification accuracy reached 94.93%, which facilitates the rapid and accurate diagnosis of microbial infection on a lab‐on‐a‐chip platform.