Chemometric methods for the analysis of process spectra using minimal reference data

Pedge, Nicholas Ian

Abstract

To construct a spectroscopic multivariate calibration model, a set of representative mixture spectra (independent variables) and the corresponding reference values for the property of interest (dependent variables) must be obtained. For a dynamic system such as a batch or semi-batch chemical reaction, creating such a data set may be very difficult or extremely time consuming. It may not be possible to create synthetic mixtures because reaction between the various reactants may occur. If the reaction proceeds via a reactive intermediate or affords a reactive product, isolated reference standards of those species may not be available. Reactions in industry are often heterogeneous and highly concentrated; sampling the batch throughout the course of the reaction for off-line analysis can be problematic and therefore introduce significant error into measured reference values.

An alternative approach that combined Self-Modelling Curve Resolution (SMCR) methods and Partial Least Squares (PLS) to construct a quantitative model using only minimal reference data was implemented. The objective was to construct a quantitative calibration model to allow real-time in-situ UV/ATR measurements to be used to determine the end-point of a chlorination reaction. Difficult reaction sampling conditions and the absence of isolated reference standards for the product and reactive intermediate required the method to be developed using only a few key reference measurements.

Citation

Pedge, N. I. (2008). Chemometric methods for the analysis of process spectra using minimal reference data. (Thesis). University of Hull. Retrieved from https://hull-repository.worktribe.com/output/4208971

• Thesis Type: Thesis
• Deposit Date: Aug 15, 2011
• Publicly Available Date: Feb 22, 2023
• Keywords: Chemistry
• Public URL: https://hull-repository.worktribe.com/output/4208971
• Additional Information: Department of Chemistry, The University of Hull
• Award Date: Oct 1, 2008