Rapid Detection of Pretilachlor in Fish and Water Samples Using a Fluorescence Sensor Based on Dual-Emission Molecularly Imprinted Polymers and Machine Learning Random Forest Algorithm

Chenxi Liu,Jingxin Liao,Yong Zheng,Ying Chen,Hongsheng Liu,Xizhi Shi

DOI: https://doi.org/10.2139/ssrn.4110877

2022-01-01

SSRN Electronic Journal

Abstract:A novel, sensitive and efficient fluorescence sensor based on dual-emission molecularly imprinted polymers (Dual-em-MIPs) was successfully developed using random forest algorithm in machine learning for rapid detection of pretilachlor. The SiO 2 coating on red-emission CdSe/ZnS quantum dots (r-SiO 2 @QDs) as an intermediate light-emitting component is not selective for pretilachlor, and molecularly imprinted layer by coated on blue-emission nitrogen-doped graphene quantum dots (N-GQD S ) is selective for pretilachlor. The fluorescence images of Dual-em-MIPs were acquired. The red, green and blue (RGB) values of the image color were analyzed using machine learning of a random forest algorithm and the classifier was trained by 103 fluorescent images for automatic analysis. Under optimized conditions, an excellent linear relationship of the sensor for pretilachlor were obtained in the range of 0.001-5.0 mg/L (R 2 , 0.9958). In addition, the satisfactory recoveries of Dual-em-MIPs were from 92.2% to 107.6% for the real samples with relative standard deviation (RSD) below 6.5%. The satisfactory recoveries of random forest (RF) model based on fluorescence sensor were 84.2-108.2% with RSD below 6.4%. Overall, the proposed fluorescence sensor based on Dual-em-MIPs and machine learning method were successfully used to determine pretilachlor in environment and aquatic products.