The Analysis Of Taste In Cigarette Mainstream Smoke And Establishment Of Forecasting Model

Sensory qualities of cigarette smoke mainly depend on the interaction of flavor &taste components in cigarette smoke;the contribution of flavor & taste components will play an important role to keep the style of Chinese cigarette,thus determination of components in cigarette smoke has become an important subject when balancing the qualities and health of cigarette smoke.Correlations were built up based on content of components and sensory results, the most contributive component would be picked to make a theoretical reference for the cigarettes'sensory control.Traditional sensory analysis is conducted by trained sensory panels, whose results are easily influenced by environment and other factors.As a new sensory technique, the electronic tongue can accomplished the test based on the responsive value of sensors without any interference.To assist the traditional sensory evaluation, the results of Electronic tongue were used to predict the sensory qualities of cigarette based on Supportive vector machine (SVM) model.The main contents of the research were as follows:1. Methnol,dichloromethane and cyclohexane were chosen to examine the extraction effect of flavor components in cigarette smoke,the results showed that dichloromethane had a better capability to extract components compared with other two solvents.Through the orthogonal test and single factor experiment were conducted to optimize the parameters for solvent extraction, finally, soaking time 20min, solvent volum 20mL, extraction power 600W, extraction time 20min were selected as the optimum parameters for this research.2. Volatile organic acids, alkaloids and phenolics were determined by GC/MS combined with selected ion scan; nonvolatile organic acids were determined by high performance liquid chromatography; water-soluble sugars and Cl- were analysed by Ion chromatography; Inductively coupled plasma was used to determine K+ and Na+ in cigarette smoke, through the recovery and accuracy test, all the methods met the determination requirements.3. After a series of conditioning, calibration and diagnoses procedures, the Electronic tongue methods were adopted for determining cigarette mainstream smoke in the artificial saliva; according to the needs of this research, the sensory evaluation standards were set, which contained sourness, sweetness, bitterness, saltiness, astringency, hoarseness, burning and wet mouthing.4. Based on the data of components in Chapter 3 and the sensory evaluation results of 21 samples, correlation analysis was carried out to check the most contributive components for each sensory standard;supportive vector machine (SVM) test was used to build a predictive model, the Electronic tongue was proved to be a good prediction tool for the sensory evaluations except for bitterness,astringency and wet mouthing in cigarette mainstream smoke.