[Zurück]

G. Quintas, A. Domínguez-Vidal, B. Lendl:
"MULTIVARIATE CURVE RESOLUTION IN ON-LINE GRADIENT HPLC-FOURIER TRANSFORM INFRARED SPECTROSCOPY";
Poster: 11as JORNADAS DE ANÁLISIS INSTRUMENTAL, Barcelona; 15.11.2005 - 17.11.2005.

Mid-infrared spectroscopy provides detailed molecular specific information on the chemical composition of almost any sample under investigation. While this feature is of advantage in many analytical application it makes direct on-line FTIR spectroscopic detection in HPLC difficult due to strong solvent absorptions of the liquid phase. Whereas solvent absorption can be handled in case of isocratic HPLC by recording an appropriate background spectrum advanced techniques of data analysis are required in case of gradient HPLC-FTIR to cope with the changing background during a chromatographic run.

In this contribution we present a novel strategy to correct for changes in solvent composition during gradient HPLC-FTIR. With the help of multivariate curve resolution alternating least squares (MCR-ALS) the changes in background spectra can be efficiently modelled. Using this advanced data analysis technique the elution profiles of the analytes as well as their FTIR spectra can be extracted from the recorded data set. The principle of this novel analytical technique will be shown on the example of the determination of four phenolic acids (gallic acid, 3,4-dihydroxybenzoic acid, p-hydroxybenzoic acid and vanilic acid) by using a Methanol:H2O (1% v/v acetic acid) gradient. For on-line recording of FTIR spectra a 25 µm flow-cell made of CaF2 windows has been used. The calculated spectral matches of the recovered spectra were very satisfactory reassembling the standard spectra up to 98 %. Furthermore quantitative information could also be obtained from the calculated elution profiles covering a typical concentration range from 0.5 to 2 mg ml-1.

http://publik.tuwien.ac.at/files/pub-tch_5265.pdf