Component Determination And Geographical Origin Identification Of Black Tea Using Fourier Transform Near Infrared Relfectance Spectroscopy (FT-NIRS)

Black tea is one of the most popular beverages throughout the world. Recently, with the growing consumption of black tea, quality control of black tea becomes increasingly of importance. Traditional analysis method is performed by sensory evaluation and chemical analysis for black tea quality. However, the result assessed by tasters is often less coherent and less impartial due to be infuenced easily by tasters’ subjectivity or physiological factors. The results of chemical analysis can achieve high accuracy. Nevertheless, detection process can’t meet the demand of rapid determination to the main compositions of tea in manufacturing and circulating process owing to complex operation, wasting much time, and high expense in chemical analysis. Besides, most commercially black tea is processed into broken black tea in the international market. Quality and identification of geographical origins are assessed from external morphology for broken black tea with difficulty. Therefore, a rapid and accurate analytical method is essentially required to determine main chemical compositions contents in black tea and discriminate their geographical origins whether in the import or export business.Near-infrared spectroscopy technique is a fast, accurate, easy and non-destructive analysis technique that can be considered as a replacement of time-consuming conventional chemical analysis method applied widely in determination and origin discrimination of agricultural produce. The dissertation attempts the feasibility to use near infrared (NIR) spectroscopy to determine main chemical compositions contents in black tea and discriminate black tea from different geographical origins. The main results of the dissertation involve:1. NIR spectroscopy combined with partial leas square (PLS) was employed for building the quantitative models to predict water extract, total polyphenols, free amino acids and caffeine contents in black tea. In building model, the effects on the spectral region, the spectral preprocessing methods and principal components factors (PCs) to results were discussed. Experimental results showed that in the spectra region between6502cm-1and5446.2cm-1. According the sequence of water extract, total polyphenols, free amino acids and caffeine. Spectral preprocessing of the models were min/max normalization, min/max normalization, SNV+1st derivative and MSC. The optimal numbers of PLS factors were10,10,10,8. The values of the correlation coefficient (R) between the predicted and the reference results for the train set were0.992,0.983,0.961, 0.983. The values of the regression coefficient (a) and RMSECV were0.993,0.967,0.925,0.980and0.547,0.453,0.157,0.102in the train set. The values of R and RMSEP were0.986,0.962,0.951,0.955and0.599,0.581,0.167,0.160in the prediction set. The values of relative prediction deviation (RPD) were8.51,5.14,4.30,3.38. Meanwhile, the predicted results of models were compared with those of standard chemical analysis methods with no significant difference by paired t-test. The results demonstrated that it was feasible to apply NIR spectroscopy to determine rapidly main chemical compositions contents in black tea.2. NIR spectroscopy combined with factorization method was attempted to build the identification model from different geographical origins (viz. Anhui, Hubei, Yunnan, India, Kenya, Sri Lanka, and Burma). The effects on the spectral region, the spectral pretreatment methods and principal components factors (PCs) to results were discussed. Experimental results showed that the optimal number of PLS factors were4, The spectral ranges were4042.2-3749.1cm-1,4701.8-4512.8cm-1,5345.9-4867.7cm-1and6155.9-5781.8cm-1, and spectral pretreatment method was1st derivative+SNV for the best identification model. Identification accuracy rate of the model was up to92.8%for unknown samples from the prediction set. The results indicated that NIR spectroscopy technique with factorization method could be efficiently employed to discriminate black tea from different geographical origins.3. NIR spectroscopy with factorization method was employed to build the identification model from different geographical origins (viz. Anhui, Hubei, Yunnan, India, Kenya, Sri Lanka, and Burma). The model could also be efficiently applied to discriminate between middle-small-leaf varieties origins (e.g. Anhui and Hubei) and broad-leaf varieties origins (e.g. India and Yunnan). A new idea by the quick and precise identification of tea material varieties was offered in this dissertation.