Abstract
Parameters of liquor color difference, taste related constituents and aroma related volatiles in 23 Chinese green tea (Camellia sinensis) samples were analyzed by color difference meter, high performance liquid and gas chromatographs respectively. Seven variables were extracted by principal component analysis from the data sets corresponding to the liquor color, taste and aroma and linear regression of total score of sensory preference (TSSP) upon the 7 extracted variables produced a significant relationship: TSSP = 105.82–0.56 (ΔE)−0.15 (total catechins) + 0.32 (geraniol) + 0.53 (GCG) + 0.44 (linalool oxide I) −0.54 (ascorbic acid) + 0.24 (n-valeraldehyde) (p = 0.003, standard error of the estimation = 2.97). The regressive relationship simplifies the process of chemical and instrumental assessment of sensory preference of green tea and could be used for green tea quality assessment.
INTRODUCTION
Green tea is an unfermented tea and its liquor characteristics are quite different from the fermented black tea. During black tea processing, fresh tea leaf (Camellia sinensis) is rolled and/or cut before drying so that tea polyphenols in tea leaf are contacted with the polyphenols oxidase and then oxidized in the consequent fermentation process, during which red pigments thearubigins and yellow pigments theaflavins are formed. However, fresh leaves are fixed by heat in a heated drum or in a steamer during green tea process so that the polyphenols oxidase is inactivated and polyphenols are intact. Therefore, green tea has green liquor while black tea has red liquor.
During the assessment of tea quality, tea taster observes the appearance of both dry and infused tea leaves. He is primarily concerned with the liquor color, aroma and taste of the brewed tea liquor. The sensory assessment of tea quality is absolutely experiential. Tea researchers have made many attempts to explain tea quality and its quality attributes chemically and physically so as to develop an equipment to replace the experiential sensory assessment. Progresses have been made in this area, especially in black tea. Roberts and Smith[Citation1] showed that theaflavin was an important compound in determining back tea quality. There was a close linear regressive relationship between theaflavin concentration and broker's evaluation of black teas.[Citation2,Citation3] Because black tea pigments are oxidation products of tea catechins, regression analysis of tasters' preferences for black teas against fresh green leaf chemical components showed positive correlations to epicatechin gallate (ECG) and epigallocatechin gallate (EGCG).[Citation4] Our previous studies showed that theaflavin made greater contribution to the brightness of black tea liquor than theaflavin gallates because the pyrogallol groups on theaflavin gallates were involved in tea creaming, which resulted in dull color of the black tea liquour.[Citation5,Citation6] Concentrations of caffeine, nitrogen, amino acids, gallocatechin (GC), epigallocatechin (EGC), catechin (C), epicatechin (EC), epicatechin gallate (ECG), catechin gallate (CG), total catechins, theaflavin, theaflavin-3′-gallate and liquor color difference indicators of ΔL, Δa, Δb and ΔE were significantly correlated to total quality score of black teas given by tea tasters.[Citation7] Capillary electrophoresis, electronic tongue and lipid membrane taste sensor have been applied to black tea quality estimation.[Citation8–11] Quality of black tea and pu-erh tea was estimated by chemical compositions and liquor colors[Citation7,Citation12] Little information about the physico-chemical assessment of green tea sensory preference has been available.
The purpose of the present article was to investigate the relationship of chemical compositions and liquor color difference indicators of green tea to its sensory preference. These findings could offer useful information on developing techniques and equipment for assessing green tea sensory preference.
MATERIALS AND METHODS
Materials
Twenty three samples of green tea were collected from various tea estates in China (). The reference compounds of 8 tea catechins for HPLC and of 18 volatiles for gas chromatography were provided by Dr. Tu from Department of Tea Science of Zhejiang University, China. The other chemical reagents used were of HPLC grade (Shild Biometric Technical Co. Ltd., Tianjing, China). Equipments for the chemical analysis were Shimadzu Model SCL-2010A HPLC and Shimadzu Model GC-14B gas chromatograph (Shimadzu Corporation, Kyoto, Japan) and that for tea liquor color difference analysis was an automatic color difference meter (Model TC-PIIG, Beijing Optical Instrument Factory, Beijing, China).
Table 1 Total scores of sensory preference of the tested green tea samples (mean ± SDFootnote a )
Tea Sensory Preference Test
The tea samples were examined and scored independently by 6 tea tasting panelists from Department of Tea Science of Zhejiang University, China. The grading system was based on a maximum score of 100, of which 10% was awarded for the dry tea appearance, 30% for the aroma, 10% for the liquor color, 30% for the taste and 20% for the infused leaf. The grading system is commonly used to assess green tea quality in China. Two hundred g of dry tea was placed in a 20 cm × 20 cm wood tray and the dry tea appearance was scored according to color, size, tenderness, and evenness of the tea twists. Three g of tea sample was infused in a tasting cup with 150 mL freshly boiled water for 5 min. The tea liquor was then poured into a 200 mL tea tasting bowl. Scores of aroma, liquor color and taste were given by comparison with 3 reference samples kept in the Lab. The infused leaves were transferred into a 20 cm × 25 cm porcelain tray containing cold water and scored according to their color, size, tenderness, and evenness.
HPLC Analysis of Ascorbic Acid, Caffeine, and Tea Catechins
HPLC analysis of ascorbic acid, caffeine and tea catechins including GC, EGC, C, EC, EGCG, GCG, ECG and CG was carried out by the method described in our previous paper.[Citation7] The tea liquor was prepared as method described in above section of tea quality preference and filtered through a “Double-ring” No.102 filter paper (Xinhua Paper Industry Co. Ltd, Hangzhou, China) and 0.22 μm Milipore filter before injected into HPLC (Model SLC-2010A,Shimadzu Co. Ltd., Kyoto, Japan). The HPLC conditions were as follows:
Analysis of Tea Polyphenols, Nitrogen, and Amino Acids
Concentration of tea polyphenols, amino acids and nitrogen in the above tea liquors were determined by the method described in our previous paper [Citation7]. During determination of amino acids, 1 mL tea liquor, 0.5 mL buffer (0.067 M Na2HPO4 and 0.67 M KH2PO4) and 0.5 mL reagent solution containing 20 g L−1 of ninhydrin and 0.8 mg L−1 SnCl2·2H2O were put in a screw-capped tubes and heated in a boiling water bath for 15 min. After heating, the tubes were immediately cooled in cold water bath. Then, the reacted solution was diluted to 25mL in a 25 mL volumetric flask with distilled water and thoroughly mixed. The absorbance (570 nm) of the reaction mixture was measured with a HP 8453 UV-VIS spectrophotometer (Hewlett-Packard Company, Palo Alto, USA). Glutamic acid was used as reference and the results were presented as glutamic acid equivalent concentrations. Nitrogen concentrations were determined by Kjeldalhl's method.
Concentrations of tea polyphenols were determined by a spectrophotometric method. One mL of tea liquor was transferred into a 25-mL volumetric flask to react with 5 mL dyeing solution (containing 3.6 × 10−3 M FeSO4 and 3.5 × 10−3 M potassium sodium tartrate, KNaC4H4O6), 4 mL distilled water and 15 mL buffer (0.067 M Na2HPO4 and 0.067 M KH2PO4). Absorbance at 540 nm of the reaction solution was determined in a 1 cm light-path cell by a HP 8453 UV-VIS spectrophotometer (Hewlett-Packard Company, Palo Alto, USA). Absorbance at 540 nm of a control reaction solution (containing 5 ml distilled water, 5 ml dyeing solution and 15 ml buffer) was determined as earlier. Gallic acid was used as reference and the results were presented as gallic acid equivalent concentrations.
Tea Liquor Color Difference Analysis
According to the three dimensional color coordinate system, color difference indicators of ΔL, Δa and Δb represent the lightness-darkness, redness-greenness, yellowness-blueness respectively and ΔE represents total hue difference between the tested sample and the control sample. The color difference indicators of the above tea liquors were determined on a TC-PIIG automatic color difference meter (Beijing Optical Instrument Factory, Beijing, China). To diminish the errors arising from different determination conditions such as different equipments and temperatures, distilled water was used as control and the indicators of ΔL, Δa, Δb, and ΔE between the samples and water were read and printed out directly by the TC-PIIG automatic color difference meter.[Citation7]
Gas Chromatography Analysis of Volatile Constituents
Volatile constituents from the green tea samples were extracted by a successive distillation extraction (SDE) apparatus and analyzed by gas chromatograph according to our previous method.[Citation12] Fifteen grams of the tea sample and 350 ml freshly boiled distilled water were placed in one of the flasks of the SDE apparatus, which was held in the boiling water bath. The sample was extracted for 1 h, during which the volatiles were evaporated and absorbed in the 30 ml of ethyl ether in another flask held in a 50°C water bath. The ethyl ether phase was then transferred into a 50 ml glass tube and dehydrated with 5 g of Na2SO4 overnight. The dehydrated ethyl ether phase was concentrated to about 0.2 ml under reduced pressure at 42°C. The concentrate was used for gas chromatograph (GC) analysis. A Shimadzu GC-14B gas chromatograph (Shimadzu Co. Ltd., Kyoto, Japan) equipped with HP-FFAP fused capillary column (30 m × 0.22 mm id) was used for the GC analysis and the GC conditions were as follows: the injector was held at a constant 200°C and the detector at 250°C during the analysis, oven temperature was maintained at 50°C for 5 min, followed by a linear programming from 50 C to 210°C, at a rate of 3°C min−1. The carrier gas helium was at 100 kPa.
Data Analysis and Statistics
All the tests in the present paper were carried out in duplicate except for the sensory preference which was assessed by six tasters and mean values of the tests were presented. Spearman's linear correlation coefficient, principal component analysis (PCA) and linear regression were calculated by software of SPSS 10.0 for Windows (SPSS Inc. Chicago, Illinois, USA).
RESULTS AND DISCUSSION
Sensory Preference of Various Green Tea Samples
The total score of sensory preference (TSSP) of the 23 tested green tea samples ranged from 73.9 to 90.0, with a mean 81.4 (). The sample No.15 from Jiangsu had the highest TSSP and the sample No.19 from Guangxi ranked the next. The sample No.21 from Yunnan was the lowest. According to the tasters' comments, the tested teas were normally processed and stored and there was no obvious deterioration caused by processing, such as burnt odor and sour taste. They were normal grades of commercial Chinese green teas marketed.
Taste-Related Constituents of Green Tea Liquors
Polyphenols, caffeine, amino acids, tea catechins, ascorbic acid, and nitrogen were major constituents affecting green tea taste and their concentrations varied from sample to sample (). Concentration of polyphenols ranged from 118.4 mg g−1 of liquor (Sample No. 21) to 226.5 mg g−1 (Sample No. 19), with an average 169.1 mg g−1. The mean concentration of total catechins was 118.9 mg g−1, ranging from 93.31 mg g−1 (Sample No. 1) to 165.7 mg g−1 (Sample No. 19). EGCG and GCG were the abundant components of the detected catechins. The average concentrations of caffeine, nitrogen and amino acids were 47.16 mg g−1, 47.0 mg g−1 and 45.2 mg g−1, respectively.
Table 2 Chemical composition of taste in the tea samples (mean ± SDFootnote a , mg g−1)
Color Differences of Green Tea Liquors
The ΔL of the green tea liquors ranged from −17.22 (Sample No. 21) to −7.64 (Sample No.14) (), which showed that the green tea liquors were darker than the background control (distilled water) because they contained a series of tea components. The value of Δa ranged from −3.35 (Sample No. 12) to −0.08 (Sample No. 21) except for Sample No. 11 which had a plus value (0.22). The value of Δb was ranged from 14.09 (Sample No. 7) to 30.44 (Sample No. 21), respectively. The Δa is a parameter of redness (when it is a plus value) and greenness (when a minus value). The Δb is a parameter of yellowness (when plus value) and blueness (when minus value). The above results suggested that the green tea liquors were light green and yellow in color. Because distilled water was used as control in the present experiment, ΔE represented a total hue difference between the tea liquor and the distilled water. The ΔE values ranged from 16.22 (Sample No. 7) to 34.98 (Sample No. 21), showing that the green tea liquors had deeper hue than water. However, sample No. 21 with the highest value of ΔE had the lowest TSSP. It suggests that liquor of low grade of green tea might have a deeper hue because of oxidation of catechins.
Table 3 Infusion color difference indicators of various tea samples (mean ± SDFootnote a )
Volatile Constituents of Green Tea Samples
The detected volatiles varied greatly between the tested samples (). The mean concentration of the detected individual volatils ranged from 0.12 μg g−1 (2-methyl heptenone) to 5.74 μg g−1 (Linalool oxide I). The mean concentrations of linalool oxide I and linalool amounted for one third of the total detected volatiles. The n-valeraldehyde was detected exclusively in samples No. 13 and No. 22 and n-caproaldehyde was detected only in samples No. 18 and No. 19. The two volatiles were abundant compounds in fresh tea leaf,[Citation13] and they might be lost by heating during fixing and drying processes of green tea manufacture.
Table 4 Concentrations of volatiles in various green tea samples (μg g−1)Footnote∗
Principal Component Analysis (PCA) and Regression Analysis
The price of green tea depends on the TSSP consisting of dry tea and infused leaf appearance, liquor color, taste and aroma. The tea tasters assess the appearance of dry tea and infused leaf according to their tenderness, size, evenness, and color. The color of the dry tea and infused leaf depends on the oxidation degree of tea catechins and the degradation of chlorophyll which are closely related with the liquor color. Therefore, liquor color, and constituents related to taste and aroma might be principally correlated to the TSSP of green tea.
Food preference denends on its color, texture, taste and aroma which are affected by many chemical and physical factors.[Citation14–17] It is difficult to analyze all the related factors when chemical and instrumental methods are used to assess the food preference. The advantages of principal component similarity are the detectabulity of a capacity to classify continuum, easy identification of causes for grouping and a potential of discovering a new group.[Citation18] Principal component analysis (PCA) can extract a few important components which accounting for most of the total variance of the group data and simplify the analysis process.
PCA was used to extract variables which were strongly correlated to important components from the 3 groups of the tested parameters, i.e., liquor color difference parameters, taste related constituents and aroma related volatiles. The result of PCA on liquor color difference parameters showed that variable ΔE was the most strongly correlated to component 1 with coefficient = 0.978 and ΔL to the component 2 (coefficient = 0.907) (). The variance of component 1 and component 2 amounted for 92.5% of total variance of the four liquor color parameters ().
Table 5 Component matrix obtained by PCA on infusion color parametersFootnote a
Table 6 Variance explained for infusion color difference variables
PCA on the taste related variables showed that total catechins was the most strongly correlated to component 1 (coefficient = 0.973) in the data set. According to the value of component coefficient, variables which were the most strongly correlated to component 2, component 3 and component 4 were GCG (coefficient = 0.791), ascorbic acid (0.601) and EGC (0.838), respectively (). Variance analysis showed that the cumulative variance of components 1 to 4 accounted for 87.4% of the total variance of the data set with 14 taste related parameters ( and ).
Table 7 Component matrix obtained by PCA on taste related parametersFootnote a
Table 8 Variance explained for the taste related variables
Similarly, PCA on the volatile parameters showed that variables which were the most correlated to components 1 to 4 in the data set of volatiles were geraniol (coefficient = 0.992), linalool oxide I (0.935), n-valeraldehyde (0.809) and n-caproaldehyde (0.898), respectively (). The variance of the components 1 to 4 accounted for 95.4% of total variance of data set ().
Table 9 Component matrix obtained by PCA on aroma related volatile parametersFootnote a
Table 10 Variance explained for the volatile related variables
Three dimensional PCA plot with the first principal components extracted from the data sets of liquor color, taste related components and volatiles showed that the 23 samples were divided into 4 groups (). Considering sample 7 had the highest concentration of geraniol which was the highest correlation to the first component, two dimensional plot was drawn when the first principal component of volatiles data was omitted (). It showed a same trend as and the sample 7 was divided into an independent group because its ΔE and total catechins were at low levels. Sensory preference scores of samples in group C were at low levels, ranging from 73.9 (sample 21) to 78.6 (sample 3) because they had low levels of total catechins and geraniol ( and , ). Sensory preference scores of samples 8 and 20 in group A were below the mean score level of the 23 samples because their geraniol was low though they had high level of total catechins and low level of ΔE. The results suggest that sensory reference depends on the factors of color, taste and volatiles.
Figure 1 PCA plot by the first principal components of color, taste and volatiles data sets.
Figure 2 PCA plot by the first principles of color and taste data sets.
Considering there were 14 variables in the group of taste related parameters and 19 in the group of volatiles, more indicators from the two groups should be selected as independent variables for regression analysis. Regression of TSSP upon the ΔE and the earlier 3 variables corresponding to components 1 to 3 of taste and aroma related data sets produced the following model:
The significance of the above mathematic model was that the 7 extracted variables used for estimating green tea sensory preference included most of variance of the 3 sets of tested variables and the time-consuming analysis was simplified. PCA begins by finding a linear combination of variables (component) that accounts for as much variation in the original variables as possible. It then finds another variable that accounts for as much of the remaining variation as possible but is uncorrelated with the previous component, continuing in this way until there are as many components as original variables. Though the extracted variable might not be the one that was the most correlated to the predicted dependant (TSSP in this case) among the tested individual variables, it accounted for most of the variance of the tested variables. Spearman's linear correlation analysis showed that the variable which was the most strongly correlated to the first component of corresponding group of the data sets was significantly correlated to many other indicators among the group while the second and third principal variables were relatively independent.[Citation12,Citation18] In this case for example, concentration of total catechins was significantly correlated to 7 other variables () and geraniol was significantly correlated to 13 other volatiles () in the corresponding group. Furthermore, calculation result showed that ΔE was significantly correlated to ΔL (Spearman's linear correlation coefficient r = −0.828, ρ < 0.01) and Δb (r = 0.982, ρ < 0.01). Therefore, ΔE, total catechins and geraniol are important for chemical estimation of green tea sensory preference not only because of the correlation of themselves to TSSP, but also because of their strong correlations to the other individual variables in the related data groups. These may explain why the regressions of TSSP on partial variables extracted by PCA gave statistically significant assessments of the green tea sensory preference. The regressions simplified the chemical and instrumental estimation process of green tea sensory preference and will be interesting for developing equipment for green tea sensory preference assessment.
Table 11 Linear correlation coefficients of total catechins, GCG, and ascorbic acid to the other taste related constituents
Table 12 Linear correlation coefficients of geraniol, linalool oxide I and n-valeraldehyde to the other volatiles
CONCLUSION
Sensory preference of green tea was dependent on its color, taste and aroma. Principal component analysis (PCA) could extract a few important components correlating to chemical and instrumental parameters by which chemical and/or instrumental analysis was simplified. ΔE, total catechins, geraniol, GCG, linalool oxide I, ascorbic acid, n-valeraldehyde were extracted from liquor color, taste and aroma related data groups. The regression of total score of sensory preference (TSSP) of the green tea upon the 7 extracted variables was statistically significant. The extracted parameters could be used to predict the TSSP of green tea not only because of their correlation to TSSP, but also because of their strong correlations to the other individual variables in the related data groups and had most of the variance of the tested variables in corresponding data groups.
ACKNOWLEDGMENTS
The authors would like to thank The National Natural Science Foundation of China for sponsoring the work (Project No. 30571192). We also appreciate Dr. Y Tu for offering the reference compounds for the HPLC and gas chromatography analysis and the tea tasting panel from the Department of tea Science of Zhejiang University for organizing sensory evaluation of the tested green tea samples.
Related Research Data
REFERENCES
- Roberts , E.A.H. and Smith , R.F. 1963 . The Phenolic Substances of Manufactured Tea. IX. The Spectrophotometric Evaluation of Tea Liquors . Journal of the Science of Food and Agriculture , 14 : 689 – 700 .
- Hilton , P.J. and Ellis , R.T. 1972 . Estimation of the Market Value of Central African Tea by Theaflavin Analysis . Journal of the Science of Food and Agriculture , 23 : 227 – 232 .
- Cloughley , J.B. 1980 . The Effect of Fermentation Temperature on the Quality Parameters and Price Evaluation of Central African Black Teas . Journal of the Science of Food and Agriculture , 31 : 911 – 919 .
- Obanda , M. , Owuor , P.O. and Taylor , S.J. 1997 . Flavanol Composition and Caffeine Content of Green Leaf as Quality Potential Indicators of Kenyan Black Teas . Journal of the Science of Food and Agriculture , 72 : 209 – 215 .
- Liang , Y.R. and Xu , Y.R. 2001 . Effect of pH on Cream Particle Formation and Extraction Yield of Black Tea . Food Chemistry , 74 : 155 – 160 .
- Liang , Y.R. and Xu , Y.R. 2003 . Effect of Extraction Temperature on Cream and Extractability of Black Tea (Camellia sinensis (L.) O. Kuntze) . International Journal of Food Science and Technology , 38 : 37 – 45 .
- Liang , Y.R. , Lu , J.L. , Zhang , L.Y. , Wu , S. and Wu , Y. 2003 . Estimation of Black Tea Quality by Analysis of Chemical Composition and Colour Difference of Tea Infusions . Food Chemistry , 80 : 283 – 290 .
- Legin , A. , Rudnitskaya , A. , Vlasov , Y. , Natale , C.D. , Davide , F. and D'Amico , A. 1997 . Tasting of Beverages Using an Electronic Tongue . Sensors and Actuators B , 44 : 291 – 296 .
- Horie , H. and Kohata , K. 1998 . Application of Capillary Electrophoresis to Tea Quality Estimation . Journal of Chromatography A , 802 : 219 – 223 .
- Ivarsson , P. , Homin , S. , Hojer , N.E. , Krantz-Rulcker , C. and Winquist , F. 2001 . Discrimination of Tea by Means of a Voltammetric Electronic Tongue and Different Applied Waveforme . Sensors and Actuators B , 76 : 449 – 454 .
- Ivarsson , P. , Kikkawa , Y. , Winquist , F. , Krantz-Rulcker , C. , Hojer , N.E. , Hayashi , K. , Toko , K. and Lundstrom , I. 2001 . Comparison of a Voltammetric Electronic Tongue and a Lipid Membrane Taste Sensor . Analytica Chimica Acta , 449 : 59 – 68 .
- Liang , Y.R. , Zhang , L.Y. and Lu , J.L. 2005 . A Study on Chemical Estimation of Pu-erh Tea Quality . Journal of the Science of Food and Agriculture , 85 : 381 – 390 .
- Wan , X.C. 2003 . Tea Biochemistry , 119 – 131 . Beijing, , China : Agricultural Publication House of China .
- Wilkie , K. , Wootton , M. and Paton , J.E. 2004 . Sensory Testing of Australian Fragrant, Imported Fragrant, and Non-fragrant Rice Aroma . International Journal of Food Properties , 7 : 27 – 36 .
- Al-Hooti , S.N. , Sidhu , J.S. , Al-Saqer , J.M. , Al-Amiri , H.A. , Al-Othman , A. , Mansour , I.B. and Johari , M. 2004 . Developing Functional Foods Using Red Palm Olein: Objective Color and Instrumental Texture . International Journal of Food Properties , 7 : 15 – 25 .
- Bajaj , S. , Urooj , A. and Prabhasankar , P. 2006 . Effect of Incorporation of Mint on Texture, Colour and Sensory Parameters of Biscuits . International Journal of Food Properties , 9 : 691 – 700 .
- Ayo , J.A. and Nkama , I. 2004 . Effect of Acha (Digitaria exilis) Grain Flour on the Physico-chemical and Sensory Properties of Bread . International Journal of Food Properties , 7 : 561 – 569 .
- Nakai , S. , Dou , J.L. and Richards , J.F. 2000 . New Multivariate Strategy for Panel Evaluation Using Principal Component Similarity . International Journal of Food Properties , 3 : 149 – 164 .