Research And Application Of The Fingerprint Method Based On Derivatization Gas Chromatography

The quality stability of cigarette, which is a basic and fast consumable, not only plays the key role in tobacco manufacturing, but has a close relation with the health of numerous consumers of tobacco products as well. In this paper, cigarette product, raw material tobacco and total particular material (TPM) were analyzed based on gas chromatography. Different sorts of tobacco samples were distinguished in order to demonstrate the effect of developed data analysis method related to cigarette quality control.Due to the complex constitution of tobacco sample matrix, the concept of fingerprint chromatogram (FC) was introduced herein. FCs were achieved by capillary gas chromatography (GC) with its high resolution, and represent the whole quality of the tobacco sample.The experimental part of this paper was to analyze and had a research on saccharides and nonvolatile compounds in TPM of mainstream smoke. The principle is to enhance the volatility of these compounds by trimethylsilylization (TMS) and render it possible to determine nonvolatile compounds by GC. GC measurement was adopted to GC coupled with flame ionization detection (FID) and mass spectroscopic detection (MSD). The procedure of analysis is simple and rapid so that the method would be universally applied in measurement of industry manufacturing.Data analysis methods, which have been established and developed to software, were used to render a further analysis and research on GC fingerprint chromatograms. The shifts of retention times in different chromatograms were firstly adjusted. Secondly, chemometric methods such as principal component analysis (PCA) and genetic algorithm (GA) were used to recognize the patterns between FCs of different sorts of tobacco samples. Related parameters were set to describe the difference between the patterns. For the FCs which contain several decades of compounds, GA was applied to select the best features to distinguish and optimize classification.In this paper, a large number of tobacco GC fingerprint chromatograms were tested and investigated. The results demonstrated that the experimental methods adopted hereby could represent tobacco samples quality. Data analysis methods efficiently and correctly recognized data to different patterns. The software which has been developed is convenient and easy to operate. Comprehensively speaking, the demand of quality monitoring and control in tobacco product manufacturing could be basically fulfilled when experimental, data analysis, and software are concerned.