New Instrumental and Chemometric Developments form the on-line hyphenation of Liquid Chromatography and Infrared Spectroscopy

Resumen

The coupling of Liquid Chromatography (LC) to infrared (IR) spectrometry is an interesting analytical tool, because high resolution provided by LC is combined with the non-destructive and molecular specific information of IR spectrometry. However, the potential of on-line LC-IR employing a flow cell interface, characterized by its technical simplicity, cannot yet be fully exploited due to the limited sensitivity of IR detection and difficulties in on-line measurements arising from the dominating absorption of most of the commonly used mobile phase components. Accordingly, the objective of this thesis can be divided into three parts: Innovations in the field of isocratic separations (discussed in Chapter 2), new methods for chemometric background correction, especially focusing on gradient LC applications (described in Chapters 3 to 6) and instrumental developments of LC-IR set-ups (Chapter 7). In summary, on-line LC-IR instrumentation as well as new chemometric tools for user-friendly background compensation, both necessary for enabling new applications of this analysis technique, should be improved or developed. Chapter 2 describes latest improvements in the field of isocratic separations with on-line IR detection. A method for the determination of lecithin and soybean oil in dietary supplements using on-line Gel Permeation Chromatography - Attenuated Total Reflectance - Infrared (GPC-ATR-IR) was developed including a simple procedure to select the optimum wavenumber used for the extraction of characteristic elution profiles of each analyte. Another method was developed for the monitoring of polymerized triglycerides in deep-frying oil by on-line GPC-IR spectrometry using the Science Based Calibration (SBC) multivariate approach for the extraction of chromatograms improving both, sensitivity and selectivity of the technique as compared to the use of the 'classical' univariate approach consisting of the monitoring of the change in absorption in a defined spectral interval with time. Additionally, for the determination of glycolic acid in cosmetics a simple on-line LC-IR procedure was developed employing the rapid scan data acquisition mode. Chapter 3 deals with different methods for background correction in on-line LC-IR using reference spectra matrices (RSM). In the course of this, the determination of the critical eluent composition for polyethylene glycols using methanol as organic modifier was discussed employing the absorbance ratio as identification parameter for background correction. Results found for the estimation of the chromatographic critical conditions of PEG correspond to a methanol:water composition of 86.5:13.5% v/v and were in good agreement with previously published results. For the determination of sugars in beverages employing acetonitrile as organic modifier, again the absorbance ratio was used as identification parameter. Fructose, glucose, sucrose and maltose could be separated, identified and quantified in commercial samples with limits of detection between 0.4 and 0.6 mg mL-1. Moreover, a Partial Least Squares (PLS) procedure for automated background correction was introduced. Its applicability for the measurement of the analyte spectra in isocratic and gradient LC has been tested, achieving accurate results in both experimental conditions. Finally, the application of point-to-point matching algorithms for background correction was discussed. The proposed background correction process has been used in on-line LC-IR using acetonitrile (0.08% v/v TFA):water (0.08% v/v TFA) gradients between 35 and 85% acetonitrile for the determination of nitrophenols as model compounds, obtaining accurate results. Chapter 4 deals with the compensation of the background contribution employing column-wise techniques. On the one hand side, polynomial regressions modelling the background absorbance at each wavenumber throughout the gradient run were used. The proposed method has been tested on real reversed-phase on-line LC-IR data sets using a mobile phase composed by acetonitrile and water. This approach reduces the influence of the size of the RSM on the accuracy of the background correction process, thus allowing the use of reduced size RSMs. On the other hand, cubic smoothing splines were used for background correction. This method works without the need of a RSM and it was tested and evaluated by means of simulated as well as real data sets obtained employing different alcoholic organic modifiers (methanol, 2-propanol and ethanol). In Chapter 5, background correction based on factor analysis, multivariate curve resolution and reference spectra matrices was discussed. The contribution of the background absorption to overall LC-IR signals could be compensated using two approaches based on Principal Component Analysis (PCA) and simple-to-use interactive self modeling analysis (SIMPLISMA), obtaining good results. Multivariate curve resolution-alternating least squares (MCR-ALS) was applied to background-corrected data improving peak and spectral resolution and eliminating remaining signal variation due to background absorption and detector drift. The chemometric extraction of 'analyte-specific' chromatograms in on-line LC-IR was discussed in Chapter 6. Obtained results confirm that the SBC method is particularly well suited for recovering an 'analyte-specific' signal from on-line LC-IR chromatograms. The usefulness of this method could be confirmed even when the analyte was injected in the presence of unknown interfering compounds and no deep knowledge of the sample constituents was needed. The last chapter describes instrumental developments in on-line LC-IR. First, on-Line IR detection in gradient capillary liquid chromatography using micromachined nanoliter-flow cells was discussed. Four model compounds were separated and identified using an acetonitrile:H2O gradient with limits of detection in the concentration range of 35-94 ng µL-1, representing an increase in mass sensitivity by a factor of approximately 30 as compared to LC systems employing a 4.6 mm ID column. In spite of using a gradient technique, high quality analyte spectra could be recovered by employing background correction using a reference spectra matrix and the relative absorbance as an identification parameter. The development of LC with on-line dual Quantum Cascade Laser (QC-laser) detection is also considered in this chapter. Compared to the use of state-of-the-art FTIR spectrometers, the developed QC-laser based system provided a significant improvement in both, sensitivity and data acquisition frequency.