Discrimination of Heat Degradation Gasoline Using Different Calculation Method of Regression Analysis

• Published in: BUNSEKI KAGAKU Vol. 55; no. 11; pp. 869 - 874
• Main Authors: Okuyama, Syuji, Satoh, Hiroyasu, Hida, Minemasa, Mitsui, Toshiyuki
• Format: Journal Article
• Language: English; Japanese
• Published: Tokyo The Japan Society for Analytical Chemistry 2006; Japan Science and Technology Agency
• Subjects: criminal research; different calculation method of regression analysis (DCR); gasoline; heat degradation; multiple regression coefficient
• ISSN: 0525-1931
• DOI: 10.2116/bunsekikagaku.55.869

Forty liquid gasoline samples and extracts from forty-seven combustion residue samples were calculated from multiple regression coefficients (R) against two kinds of marketing gasolines as references with different calculation methods of the regression analysis by using 12 components that eluted after xylene. A quantitative evaluation for heat degradation of the samples was performed with the R values. If a sample R value of more than 0.85 was obtained, we could judge whether the sample was a gasoline only from a gas chromatogram. For sample R values from 0.85 to 0.50, additional information was required to judge whether the samples were gasoline. A sample R value of less than 0.5 clearly showed that it was not gasoline. This method was tried to apply a difference discrimination using both one subject sample and one object sample. We could prove that the subject and the object samples were the same kind of gasoline if the calculated R value was 0.999 or more.