In-line monitoring of batch emulsion polymerization by fiber-optic Raman spectroscopy

Raman spectroscopy has been recognized as an excellent technique for monitoring polymerization reaction systems. It provides direct information of the changes in the chemical bonds as the monomers polymerize. The use of optical fibers makes it possible to conduct the measurements in remote. However, major limitations in the availability of adequate probes have prevented the implementation of the technique in industrial processes. In this research, the interaction of emulsion systems with several probe designs was evaluated. Multi-fiber probe designs were found to suffer from strong fiber interferences that mask the emulsions' Raman scattering signals. However, the major problems associated with multi-fiber probes were eliminated by the use of filter-based probes. First, the filter rejects the intereferences produced from the interaction of the excitation source with the excitation fiber. Second, it blocks the elastic scattering (Rayleigh and Mie) from the sample from entering the collection fiber and thus masks the Raman signal from the sample. Experiments were conducted to monitor emulsion reactions using a Raman system coupled to a filter-based fiber-optic probe. An emulsion reaction of styrene-butyl acrylate copolymerization was studied. The total concentration of monomers was found to correlate with the Raman peak at 1,631 cm^-1. The styrene monomer concentration was found to correlate with the Raman peak at 1,412 cm^-1 attributed solely to the styrene monomer. A methyl methacrylate emulsion polymerization system was also monitored with a Raman system coupled to a filter-based fiber-optic probe. Two methyl methacrylate reactions were run under similar conditions and both were monitored with Raman spectroscopy. The monomer concentration was determined using two different regions of the spectra. A model for the concentration of the first region using the peak at 1,644 cm^-1 was built using a univariate technique. The second region between 805-855 cm^-1 contained two peaks; one at 840 cm^-1 attributed to the monomer, and the second at 820 cm^-1 attributed to the polymer. A model for the concentration of the second region was built using a multivariate technique. These two regions gave similar monomer profiles for each reaction. The concentration models built using the data of one reaction were capable of predicting the concentration profile of the second reaction. This study showed the feasibility of using Raman spectroscopy coupled with a filter-based probe to monitor in real time the progress of emulsion polymerization batch reactions.