In-line fiber-optic spectroscopy : applications to polymer process monitoring

Author

Srinivas Vedula

Date of Award

5-1997

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Chemical Engineering

Major Professor

Marion G. Hansen

Committee Members

Tsewei Wang, Roberto Benson

Abstract

Near-infrared (NIR) spectroscopy has several advantages over conventional mid-infrared (MIR) techniques, such as remote data collection coupled with rapid data analysis methods, availability of fiber-optics, and lack of sample handling problems. These advantages have led to the widespread use of NIR spectroscopy for in-line chemical process measurements. Earlier research studies have addressed problems related to development of fiber-optics, feasibility of in-line monitoring of processes in harsh process environments, data analysis techniques, new process applications, and real-time process control strategies. In the area of polymer process engineering, NIR spectroscopy is being primarily used for measurements of polymer composition in copolymers, polymer blends during extrusion and injection molding; monitoring reaction kinetics; and estimation of physical properties, such as density, relative viscosity, et cetera,/em>. In this research, the applicability of NIR spectroscopy has been extended to make simultaneous measurements in an in-line extrusion process of rheological properties and composition of flowing polymer melts. A methodology is proposed and demonstrated for estimating polymer melt index and linear viscoelastic properties for a system of ethylene-vinyl acetate random copolymers. This methodology involved developing calibration models, which were subsequently utilized for real-time monitoring. Qualitative analysis is provided for assignation of spectral regions for the rheological properties and the comonomer ratio. The research also addressed the issue of in-line polymer additive monitoring using fiber-optic near-infrared and ultraviolet spectroscopy. Predictive calibration models and exploratory data analysis techniques were used as definitive studies for examining the feasibility of in-line trace-level additive monitoring.