Electrochemical and chemometric authentication of Panax notoginseng from different geographical origins using graphene-modified screen-printed electrodes

Abstract

An effective, rapid, and low-cost platform for herbal authentication was developed by integrating nanomaterial-modified sensor technology with advanced data analysis techniques. In this study, 120 Panax notoginseng samples (40 each from Yunnan, Sichuan, and Guangxi provinces) were collected and analyzed. A graphene-modified screen-printed electrode was fabricated by depositing 10 μL of graphene ink and allowing a 2-hour drying period, which yielded a uniform film with an average flake size of 200 nm. Differential pulse voltammetry was employed to capture the redox behaviors of herbal extracts in 0.1 M phosphate-buffered saline (pH 7.0) and 0.1 M acetate-buffered solution (pH 4.5), resulting in distinct oxidation peaks at 0.47 V, 0.50 V, and 0.45 V for the three sample groups. The acquired signals were normalized and subjected to first-order derivative processing to enhance peak resolution and minimize baseline drift. Subsequent principal component analysis revealed that the first two components accounted for 72 % and 13 % of the total variance, respectively, cumulatively explaining 85 % of the data variability. Furthermore, machine learning algorithms were applied to classify the samples based on their processed electrochemical fingerprints, with the non-linear support vector machine model achieving an accuracy of 92.0 %, precision of 92.5 %, recall of 91.5 %, and an area under the receiver operating characteristic curve of 0.95, misclassifying only 4 out of 108 test instances. These results indicate that the proposed platform can reliably distinguish subtle differences in the chemical profiles of herbal extracts, offering a promising approach for objective quality control and authentication in herbal medicine applications. These innovative findings underscore the potential for widespread implementation of this technique in commercial and research settings.