Thesis

Extending the scope of mid-infrared spectrometry for in situ process analysis through ATR immersion probes

Downloadable Content

Download PDF
Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2011
Thesis identifier
  • T12988
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • This study concerned developments and applications of attenuated total reflection (ATR) mid-infrared (MIR) insertion probes featuring novel polycrystalline silver halide fibres. Improvements in probe performance from the original design to the newer designs were observed; although some changes provided manufacturing benefits rather than performance improvements. Detailed evaluation of two 12 mm diameter probes and a 2.7 mm diameter single fibre probe was performed for analysis of mixtures of acetone, ethanol and ethyl acetate. Calibration transfer was attempted for situations when either the ATR probe or MIR spectrometer were changed. Direct transfer introduced severe calibration errors, so two standard calibration methods, direct standardisation (DS) and piecewise direct standardisation (PDS), and a new method spectral space standardisation (SST), were compared. The SST procedure incorporating a scaling factor demonstrated advantages over DS and PDS giving lower errors of prediction and simpler implementation over PDS. Two other application areas were also investigated: identification of counterfeit Scotch whisky and study of a fermentation reaction. The identification of counterfeit Scotch whiskies was based on the determination of ethanol concentration and the spectra of the dried residues of the suspect samples. The latter measurement also proved successful as a procedure to gain a greater understanding of the impact of manufacturing variables on the generation of the colour of whisky. Near-infrared (NIR) and MIR spectrometry were used to study a fermentation reaction; selected spectral regions of the NIR or MIR data can be used to monitor three properties of the fermentation process: optical density, glycerol and ammonium concentrations. PLS models built with NIR data produced better results than those using MIR data for the prediction of optical density and ammonium. However, the results for the predictions of glycerol were comparable; indications of the potential benefits of data fusion were apparent when the MIR and NIR data were combined.
Resource Type
DOI
Date Created
  • 2011
Former identifier
  • 939347

Relations

Items

Thumbnail Title Date Uploaded Visibility Actions
file details PDF of thesis T12988 2021-07-02 Public Download