Abstract
Abstract
Background
Leaves of Rhus punjabensis var. sinica, Rhus potaninii, Caesalpinia spinosa (Tara), and Phyllanthus emblica (Indian gooseberry), and other species are rich in tannic and gallic acids, which are polyphenols widely utilized in the pharmaceutical, food, and chemical industries. Conventional quantification involves tedious solvent extraction and chemical assays, which are time-consuming, costly, and incapable of simultaneous multi-component analysis.
Objective
To enable rapid, non-destructive detection, Fourier transform near-infrared spectroscopy was used to establish quantitative models for synchronous determination of tannic acid, gallic acid, and moisture in polyphenol-rich plant leaves.
Methods
Eighty-five leaf samples from six species were collected; reference values were obtained by chemical methods. Near-infrared spectra of plant leaves were fitted to traditional chemical values. Synergy interval partial least squares (SIPLS) selected characteristic intervals to build a partial least squares (PLS) model; performance was assessed by correlation and error metrics.
Results
Optimal pretreatments were FD+MSC (moisture), MSC (tannic acid), and FD (gallic acid) with number of principal factors 4, 9, and 6. Characteristic bands: moisture 9403.7–6098.1 and 5450.1–4597.7 cm⁻¹; tannic acid 9403.7–7498.2 and 6102–4246.7 cm⁻¹; gallic acid 9403.7–8451 and 6102–5446.3 cm⁻¹. For moisture, tannic acid, and gallic acid, Rc² = 0.9574, 0.9595, 0.9115; RMSECV = 0.433%, 1.47%, 0.362%; RPD = 4.81, 4.96, 3.36; Rp² = 0.9669, 0.9792, 0.9443; RMSEP = 0.274%, 1.25%, 0.241%. Bias was not significant (P > 0.05), and average RSD
Conclusion
The results demonstrate that near-infrared spectroscopy combined with chemometric methods demonstrates high predictive capability and practical application value for the quantitative determination of tannic acid, gallic acid, and moisture in polyphenol-rich plant leaves.
Highlights
A rapid, non-destructive NIR spectroscopy method was developed for simultaneous quantification of tannic acid, gallic acid, and moisture in polyphenol-rich plant leaves. Optimized chemometric modeling (SIPLS–PLSR) achieved high prediction accuracy (Rp² > 0.94, RPD > 3.0), demonstrating strong robustness and suitability for practical applications.