Screening Brazilian commercial gasoline quality by hydrogen nuclear magnetic resonance spectroscopic fingerprintings and pattern-recognition multivariate chemometric analysis

Abstract

The identification of gasoline adulteration by organic solvents is not an easy task, because compounds that constitute the solvents are already in gasoline composition. In this work, the combination of Hydrogen Nuclear Magnetic Resonance ((1)H NMR) spectroscopic fingerprintings with pattern-recognition multivariate Soft Independent Modeling of Class Analogy (SIMCA) chemometric analysis provides an original and alternative approach to screening Brazilian commercial gasoline quality in a Monitoring Program for Quality Control of Automotive Fuels. SIMCA was performed on spectroscopic fingerprints to classify the quality of representative commercial gasoline samples selected by Hierarchical Cluster Analysis (HCA) and collected over a 6-month period from different gas stations in the São Paulo state, Brazil. Following optimized the (1)H NMR-SIMCA algorithm, it was possible to correctly classify 92.0% of commercial gasoline samples, which is considered acceptable. The chemometric method is recommended for routine applications in Quality-Control Monitoring Programs, since its measurements are fast and can be easily automated. Also, police laboratories could employ this method for rapid screening analysis to discourage adulteration practices. (C) 2010 Elsevier B.V. All rights reserved.