The present work describes a methodology for estimation of monomer concentration during homopolymerization reactions by Raman spectroscopy. The estimation is done using linear models based on two different approaches: a univariate approach and a multivariate approach (with principal component regression, PCR, or partial least squares regression, PLS). The linear models are fitted with data from spectra collected from synthetic samples, i.e., samples prepared by dispersing a known concentration of monomer in polymer emulsions. Homopolymerizations of butyl acrylate and of vinyl acetate were monitored by collecting samples from the reactor, and results show that the methodology is efficient for the model fitting and that Raman spectroscopy is a promising technique for on-line monitoring of the emulsion polymerization process.
Polymerization process monitoring; FT-Raman spectroscopy; Multivariate linear models; Partial least squares regression
