Monitoring the esterification reactions of carboxylic acids with alcohols using near-infrared spectrocopy and multivariate calibration methods

Abstract

Simultaneous determination of mixtures of alcohols, acids, esters, and water using near-infrared spectroscopy (NIR) and four different multivariate calibration methods were realized. The four multivariate calibration methods were Genetic Inverse Least Squares (GILS), Genetic Regression (GR), Principle Component Regression (PCR), and Partial Least Squares (PLS). Four different esterification reactions were investigated. These are methyl acetate, ethyl acetate, propyl acetate, and butyl acetate.The sample set contains 40 ternary mixtures of these esterification processes. Duplicate measurements were done for each sample and 80 NIR absorbance spectra that cover the range from 4000 to 10000 cm.1 were obtained. Of these 80 spectra, 50 were used as calibration set, 30 were reserved for the prediction purposes. Several calibration methods were built for each component of each esterification reactions. Standard error of calibration (SEC) and standard error of prediction (SEP) were calculated for each calibration model and comparison of these four multivariate calibration methods was done.