Infrared spectroscopy chemometric model for determination of phenolic content of plant leaf powder

Abstract

Introduction: The leaf is the part of the plant often used in traditional medicine because it is rich in one class of secondary metabolite compounds, namely phenolic compounds. Objectives: This study aims to establish a chemometric model for determining the phenolic content of plant leaf powders using the infrared spectroscopy (FTIR and NIR) method with a combination of chemometrics. Methods: The dried and powdered plant leaves were scanned using FTIR and NIR spectroscopy. Spectra were used to form calibration models. The calibration models were Partial Least Square (PLS), Principal Component Regression (PCR), and Support Vector Regression (SVR). The selected calibration model was validated using LOOCV and external cross-validation. The best calibration model was PCR, with R2 and RMSEC values of FTIR and NIR of 0.9918885; 0.9752648 and 0.8675906; 1.5150245, respectively. Results: The results of the Paired-Sample T-test analysis of actual samples determined by the selected calibration model compared to the comparison method showed no significant difference.