Green tea, while popular for its flavor and health benefits, deteriorates rapidly due to polyphenol oxidation and degradation during storage. Traditional methods like sensory evaluation and chemical analysis (GC-MS, HPLC, LC-MS) are time-consuming and subjective. Spectral analysis techniques, such as hyperspectral imaging and near-infrared spectroscopy, have shown promise but haven't identified specific indicator compounds linked to storage time. Surface-enhanced Raman spectroscopy (SERS), a rapid, objective, and sensitive technique, offers a potential alternative for monitoring green tea quality changes. This research aimed to investigate the sensory and chemical changes in green tea over time, acquire Raman signals using SERS with silver nanoparticles, develop an accurate predictive model for storage time, and identify characteristic Raman peaks and indicator compounds.

Existing literature highlights the challenges of assessing green tea quality changes during storage. Sensory evaluation, while efficient, is subjective and depends on expert panelists. Chemical analysis methods, though accurate, are often time-consuming and require specialized expertise. Previous studies have explored spectral analysis techniques like hyperspectral imaging and near-infrared spectroscopy to evaluate green tea storage quality, but these approaches haven't pinpointed specific compounds that correlate with storage time. This study leverages SERS, a technique previously shown to be successful in measuring catechin concentrations in tea and detecting quality changes in other beverages (wine), to address these limitations.

The study used Tieguanyin green tea samples (TPHK) collected annually from 2015 to 2020. Sensory evaluation, following Chinese national standards, assessed aroma, taste, color, and overall acceptability on a 9-point hedonic scale. Chemical analysis using UPLC determined the concentrations of various polyphenols (EGCG, ECG, GCG, EGC, C, EC, GC, GA, caffeine, quercetin, kaempferol, and myricetin). Silver nanoparticles were synthesized using a microwave-assisted method and characterized via SEM and EDS. Green tea infusions were mixed with silver nanoparticles, and SERS spectra were acquired (300–1800 cm⁻¹) using a Horiba Jobin Yvon Lab RAM HR Evolution Raman spectrometer. Spectral data were preprocessed using Savitzky-Golay smoothing and standard normal variate (SNV) transformation. PCA was employed to reduce dimensionality, followed by PCA-LDA and PCA-SVM model building for storage time prediction. Correlation analysis linked SERS peaks with polyphenol concentrations, and a standard curve was developed to quantify myricetin using SERS.

Sensory evaluation revealed a significant decrease in overall acceptability with increased storage time. Chemical analysis showed decreases in most polyphenols (EGCG, ECG, GCG, EGC, C, EC, GC, GA) and increases in others (C, M) with storage. SERS spectral analysis revealed changes in peak intensities at 731.1, 1320.5, 1237.9, 1589, and 1631.3 cm⁻¹. PCA showed distinct clustering of samples based on storage year. The PCA-SVM model, using six principal components, exhibited high accuracy (98.61% for the calibration set and 97.22% for the prediction set) in predicting storage time. Correlation analysis strongly linked the Raman peak at 730 cm⁻¹ to myricetin concentration, which exhibited a linear positive correlation with storage time (R² = 0.9346). SERS-based myricetin quantification showed good agreement with UPLC results (maximum error of 5.81%).

The high accuracy of the PCA-SVM model in predicting green tea storage time using SERS demonstrates the technique's potential as a rapid and reliable quality assessment method. The strong correlation between the Raman peak at 730 cm⁻¹ and myricetin concentration, coupled with myricetin's known role in green tea degradation, establishes myricetin as a valuable indicator compound. This method offers a significant improvement over traditional sensory and chemical methods due to its speed, objectivity, and simplicity. The findings contribute to a more efficient quality control system for green tea, potentially leading to better quality assurance and reduced waste.

This research successfully developed a rapid and accurate method to predict green tea storage time using SERS combined with a PCA-SVM model. The identification of myricetin as a key indicator compound, based on its unique Raman peak at 730 cm⁻¹, provides a valuable tool for quality control. Future studies could explore the application of portable Raman devices for on-site green tea quality assessment and further investigate the underlying mechanisms of myricetin's role in green tea degradation.

The study's limitations include the use of a single green tea variety and the limited range of storage conditions. The model's generalizability to other green tea varieties or diverse storage environments needs further validation. While the SERS method shows promise, the development of a portable device for on-site analysis would enhance its practical application.