Statistical Classification of Cigarettes Using High Pressure Liquid Chromatograms of Cigarette Filter Tar as Fingerprint

Abstract

In this work, we have showed that a liquid chromatogram can be used as finger print to characterize and classify cigarettes. A liquid chromatographic (high pressure liquid chromatography) procedure was first developed to obtain chromatograms of different cigarettes. The cellulose filter of a smoked cigarette was first dissolved in 5 mL acetonitrile. Then the organics were partitioned into the n-hexane phase. After concentrating the n-hexane phase to 100 mu L, 5 mu L was injected to a high pressure liquid chromatography. The chromatograms were obtained using a C18 reversed phase column. The mobile phase was 80 % acetonitrile/water with 1 mL min(-1). The detector was UV spectrometer working at 254 nm. For the evaluation the chromatograms as fingerprints, five peaks were chosen randomly. The heights of the corresponding peaks were used for the statistical evaluation which was based on the discriminate analysis for classifying problems introduced by Anderson. To achieve this, a Matlab computer program was developed. The proposed method can be performed in a relatively short time.