Use of multivariate curve resolution to monitor an esterification reaction by near-infrared spectroscopy

Multivariate curve resolution-alternating least squares (MCR-ALS) methodology was applied to near-infrared (MR) spectra with view to estimating the reaction profile for the esterification of a mixture of caprylic and capric acids with glycerol. The reaction was conducted in excess glycerol, so the major products were the monoglyceride and diglyceride; the triglyceride was obtained in considerably lower proportions. Esterification processes were performed in a laboratory-scale reactor from which samples were withdrawn to record MR spectra. Some samples were also analyzed by gas chromatography and acid-base titration in order to determine the composition of the reaction mixture. Concentration and spectral profiles were obtained by using the MCR-ALS algorithm. Subsequently, concentration values and the pure spectra of reagents and the triglyceride were used to refine the models. The spectra obtained were processed with MCR-ALS in new esterification batches to obtain their concentration profiles. Copyright © Taylor & Francis, Inc.