The Total Status Of Taste Characteristic And The Relationship With Chemical Components In Flue-cured Tobacco Leaves Of Main Tobacco-growing Areas In China

Taste characteristic is an important quality characteristic index to evaluate the sensory qualityof tobacco leaves; it mainly includes the aftertaste, irritation and dryness. The purpose of theresearch is study the general characteristics of flue-cured tobacco in China, and to find out thechemical quality indicators that have close relationship with the taste characteristic. So the paperhave great significances to guide the cigarette industrial enterprise to use the tobacco materialsrationally. This study selected153samples of flue-cured tobacco leaves from maintobacco-growing areas of China (including9samples from northern region,26samples fromsoutheast region,15samples from Huanghuai tobacco planting area,77samples from Southwesttobacco planting area and26samples from Yangtze river tobacco planting area) and27samplesfrom abroad included14Brazil’s samples and13Zimbabwe’s samples as the materials. At thesame time, by using a variety of statistical analysis methods, the paper researched the relationshipsbetween the taste characteristic and the content of routine chemical components, neutral aromasubstances, polyphenols and organic acids respectively, and established statistical models as well.In the end, discriminant analysis of taste characteristic was applied based on chemical components.The main results are as follows:1. The basic characteristics of taste characteristic of flue-cured tobacco in main tobaccogrowing areas were analyzed. The scores of taste characteristic in flue-cured tobacco leaves of ourcountry ranged from8.50to14.00, with the mean of11.11and coefficient variation (CV) was9.75%. The scores of75.82%of the samples were between10.00~12.00,10.46%samples’ scoreswere less than10.00, and only13.72%of the total samples had higher taste characteristic scores,which were more than12. In general, the overall performance of sensory pleasurableness of ourcountry tobacco was medium.2. Difference analysis on taste characteristic of flue-cured tobacco in main tobacco growingareas was conducted. The results showed that the scores of taste characteristic were significantlydifferent among different parts, different flavor stypes, different tobacco growing regions anddifferent industrial use.(1) Among different parts, the score of taste characteristic showed in theorder of middle leaf> lower leaf> upper leaf, and the scores of taste characteristic of middleleaves and lower leaves significantly higher than the upper leaves;(2) The average score of taste characteristics of light flavor type tobacco was highest, was11.42, followed by the untypicalflavor, the full flavor tobacco with lowest score. The scores of light flavor type tobacco leaveswere significantly higher than that of untypical flavor and the full flavor, while no difference ontaste characteristics scores between untypical flavor and full flavor.(3) Southwest tobaccogrowing area, as a representative of light flavor type tobacco leaves, the expression of tastecharacteristic, was better than the Southeast region, significantly better than the Yangtze rivertobacco planting area, Northern and Huanghuai regions.(4) Compared with the tobacco leavesabroad, the performance of taste characteristic of tobacco leaves in Zimbabwe was best, and hadno difference between the tobacco leaves that from Brazil and Southwest tobacco growing areas,but it displayed better significantly than that in the Southeast, Northern, Yangtze River andHuanghuai tobacco planting areas.(5) Among industrial uses, the score of taste characteristic as tothe main ingredient material showed the following trend that up-market>high price> low price,while as for the assistant materials the score of taste characteristic showed in the order of highprice>medium price>low price.3. The relationship between the taste characteristic and routine chemical components wasstudied, the results indicated that:(1) The taste characteristics had extremely negative correlationwith total plant alkaloid, total nitrogen, protein, volatile bases, and had positive correlation withtotal sugar, reducing sugar content, pH and nitrogen/nicotine, sugar/nicotine.(2) There existedsignificant or extremely significant regression relationships between taste characteristics and totalplant alkaloid, total sugar, reducing sugar, total nitrogen, protein, volatile bases andnitrogen/nicotine, sugar/nicotine. Under the condition of linear case, fitting degree of regressionequations of total plant alkaloid, reducing sugar and taste characteristics were the best, theregression relationship between taste characteristics and other chemical components were the bestunder the condition of nonlinear.(3) The contents of most chemical components were significantdifferent among different taste characteristics score clusters except for potassium, chlorine andpotassium/chlorine. The tobacco leaves that belonged to high taste characteristics score clustershad higher total sugar, reducing sugar and pH, while the contents of total plant alkaloid, totalnitrogen, protein, volatile bases in the leaves that belonged to low taste characteristics scoreclusters were all the lowest. 4. the relationship between the taste characteristic and neutral aroma substances was studied.The results indicated that:(1) As to the single neutral aroma component, the contents of2-acetylpyrrole, benzaldehyde, phenethyl alcohol,6-methyl-5-hepten-2-one, geranylacetone,megastigmatrienone and solanone were different significantly among different clusters. In additionto the products of browing reaction, the contents of products of phenylalanine, pruducts degradedby carotene, products degraded by cumbrance and the total contents of aroma components were allthe highest in the group with high taste characteristic scores, while the group with low tastecharacteristic scores was opposite.(2) Among different clusters, comprehensive factor scoresranked in the order of high cluster>medium cluster>low cluster. In the high and medium tastegroups, most of the samples had higher comprehensive factor scores, while the factor scores of themajority (of samples) were lower in the low taste group. Factor analysis based on neutral aromasubstances to evaluate the quality of tobacco among different taste characteristics was feasible.(3)Taken together, phenethyl alcohol,6-methyl-5-hepten-2-one, geranylacetone, megastigmatrienoneand solanone may be the important neutral aroma components that had prominent contributions tothe formation of good taste characteristic. The content of neutral aroma components could be usedto provide a reference for evaluating the taste characteristic to a certain extent.5. The relationships between the taste characteristic and polyphenols, organic acids wereanalyzed. The results showed that:(1) Gray relational grade analysis of taste characteristic andpolyphenol substances showed that the grey related degree between chlorogenic acid and tastecharacteristic was of the maximum, followed by rutin, and scopoletin had the least influence oncomfortableness of taste. Grey superior analysis of the taste characteristic and polyphenolsubstances showed that chlorogenic acid was the quasi-optimal factor of affecting tastecharacteristic of tobacco leaves, followed by rutin, the effect of scopoletin on taste characteristicwas least.It can be indicated that chlorogenic acid and rutin had positive effects on the score of tastecharacteristic, which was based on BP artificial neural network of taste characteristic andpolyphenols. Overall, taste characteristics score was increased with the increase of the contents ofchlorogenic acid and rutin. While scopoletin had a negative effect on taste characteristic, and withthe increase of the content of scopoletin, taste characteristics score decreased. (2) Through the grey relational grade analysis of taste characteristic and organic acids，thegrey relational order was：linoleic acid＞stearic acid＞palmitic acid＞succinic acid＞myristicacid＞oleic acid＞lauric acid＞malonic acid＞oxalic acid＞citric acid＞malic acid. It could beindicated that stearic acid, linoleic acid, palmitic acid, succinic acid and lauric acid were themain organic acids advantage index effecting taste characteristic of flue-cured tobacco by usinggrey superior analysis of taste properties. Taken together, the organic acids greatly affecting tastecharacteristic of flue-cured tobacco were mainly stearic acid, linoleic acid, oleic acid, palmiticacid, lauric acid belonging to higher aliphatic acid and higher aliphatic acid belonging to multipleorganic acid.It could be indicated that oxalic acid, malonic acid, succinic acid, malic acid and citric acidhave different effects on taste characteristic of flue-cured tobacco, which was based on BPartificial neural network of taste properties and multiple organic acid. The score of tastecharacteristic was tending towards stability at first, then increased and dramatically reduced in theend with the increase of oxalic acid, which was a complicated phenomenon. While malonic acid,succinic acid and citric acid had a negative influence on taste characteristic. The score of tastecharacteristic increased at first, and then decreased gradually with the increase of malic acid.It could be seen from the curve fitting of artificial neural network that senior fatty acidsexcept myristic acid had significant effect on taste properties. The taste characteristic declined atthe beginning, then slowly increased with the increase of lauric acid. While palmitic acid, linoleicacid and oleic acid had positive effect on taste characteristic, especially for palmitic acid, linoleicacid and oleic acid, the score of taste characteristic showed increasing trend with increase of them,when they reached a certain degree, it had little effect on taste characteristic. It had a positiveeffect on taste characteristic when the contents of stearic acid was lower than0.65mg/g, while ithad negative effect on taste characteristic when the contents of stearic acid was higher than0.65mg/g, which trend to be a positive effect as a whole.In a word, organic acids, especially senior fatty acids had significant effect on comfort degreeof taste characteristic of flue-cured tobacco.(3) It could be seen from the curve fitting of BP artificial neural network and the regressionanalysis result of monadic polynomial, bivariate polynomial and polynomial regression that polyphenol in tobacco and most of organic acid had respective large or small effect on tasteproperties of tobacco, and most of them presented high-dimensional nonlinear relation. Using BPartificial neural network could get a better study and matching of the data comparing withpolynomial regression model, which could preferably reflect the characteristics of the data,therefore, the result was more accurate and reliable.6. Taking routine chemical composition in tobacco as variables, using general discriminantanalysis and stepwise discriminant analysis techniques to establish mathematical model toauthenticate taste characteristic, it turned out that:(1) Using general discriminant analysis toestablish three Fisher-discriminate functions, the rates of accuracy, of returned classification of thethe original sample using auto-controlled trial and cross-validation methodology, reached to75.1%and71.2%.(2) Using stepwise discriminant analysis technique to eliminate four variables,only total nitrogen, total sugar, volatile acid, starch, starch, pH value and total vegetable alkaloidthese seven variables were reserved, and three discriminate functions were established. The ratesof accuracy, of returned classification using auto-controlled trial and cross-validation methodology,reached to77.8%and73.2%.(3) It showed that the accuracy of discrimination reached higherthan70%, so discriminant functions were verified well in discrimination results.