The use of chemometric techniques in process analytical method development and operation

Abstract

Process analytical chemistry is one area in which chemometrics can have a considerable practical impact. In many cases, a good multivariate calibration method is necessary, but not sufficient, for successful development and implementation of a process analytical method. This statement is particularly true in the case of spectral process analyzers, which can produce large volumes of very complex and non-selective data. The implementation of a process spectroscopy method can be thought of as a series of discrete procedures, each of which can be greatly aided through the use of chemometric methodologies: (i) gross outlier removal; (ii) data filtering/pre-linearization; (iii) sample selection; (iv) variable selection; (v) calibration; (vi) on-line method fault detection; and (vii) calibration updating/maintenance. The purpose of this paper is to both present recent chemometric approaches to each of these steps, and, in light of the experimental results reported, to discuss possibilities for new or improved methodologies. Recent approaches to most of these steps will be presented using data obtained from a process spectrometer installed on a polymer production unit.