ANALYSIS OF GASOLINES BY FT-IR SPECTROSCOPY

Citation
Ge. Fodor et al., ANALYSIS OF GASOLINES BY FT-IR SPECTROSCOPY, Analytical chemistry, 68(1), 1996, pp. 23-30
Citations number
12
Categorie Soggetti
Chemistry Analytical
Journal title
ISSN journal
00032700
Volume
68
Issue
1
Year of publication
1996
Pages
23 - 30
Database
ISI
SICI code
0003-2700(1996)68:1<23:AOGBFS>2.0.ZU;2-G
Abstract
An experimental and computational protocol was established for the sim ultaneous determination of several key gasoline properties from a sing le Fourier transform infrared (FT-IR) spectrum, The study has shown th at midband FT-IR spectroscopy combined with multivariate calibration a nalysis is a versatile, efficient, and accurate technique for the simu ltaneous estimation of key gasoline properties within about 1 min with less than 2 mt of sample, The FT-W-derived values of gasoline propert ies include research and motor octane numbers, aromatic, olefinic, and saturated hydrocarbon content, benzene content, and concentrations of ethanol, methyl tert-butyl ether, and total oxygen, Concentrations of other oxygenated compounds are expected to be equally predictable, Ho wever, since these oxygen-containing species have not been adequately represented among the currently commercially available gasoline sample s, their calibration may only be achieved using laboratory fuel blends , Midrange boiling point data may also be estimated. Fuel properties d etermined by minor concentrations of fuel components, e.g., flash poin t, sulfur content, etc., may not be modeled because the corresponding ET-IR signals are below detection limits of presented experimental pro tocol, The precision of this procedure was shown to be comparable to r eproducibility of the standard laboratory analyses used for direct mea surement of specific fuel properties, with squared correlation coeffic ient (R(2)) ranging from 0.94 to 0.99 between the two sets of measurem ents, This new methodology could increase the corresponding output of the petroleum laboratories by a factor of over 200 to 1 while maintain ing data integrity and minimizing sample requirements, environmental h azards, and cost.