Absorption of CO2 and subsequent viscosity reduction of an acrylonitrile copolymer

Acrylonitrile (AN) copolymers (AN content greater than about 85 mol%) are traditionally solution processed to avoid a cyclization and crosslinking reaction that takes place at temperatures where melt processing would be feasible. It is well known that carbon dioxide (CO2) reduces the glass transition temperature (T-g) and consequently the viscosity of many glassy and some semi-crystalline thermoplastics. However, the ability of CO2 to act as a processing aid and permit processing of thermally unstable polymers at temperatures below the onset of thermal degradation has not been explored. This study concentrates on the ability to plasticize an AN copolymer with CO2, which may ultimately permit melt processing at reduced temperatures. To facilitate viscosity measurements and maximize the CO2 absorption, a relatively thermally stable, commercially produced AN copolymer containing 65 mol% AN was investigated in this research. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) indicated that CO2 significantly absorbs into and reduces the T-g of the AN copolymer. Pressurized capillary rheometry indicated that the magnitude of the viscosity reduction is dependent on the amount of absorbed CO2, which correlates directly to the T-g reduction of the plasticized material. Up to a 60% viscosity reduction was obtained over the range of shear rates tested for the plasticized copolymer containing up to 6.7 wt% CO2 (31 degreesC T-g reduction), corresponding to as much as a 30 degreesC equivalent reduction in processing temperature. A Williams-Landel-Ferry (WLF) analysis was used to estimate the viscosity reduction based on the T. reduction (and corresponding amount of absorbed CO2) in the plasticized AN copolymer, and the predicted viscosity reduction based on using the universal constants was 34-85% higher than measured, depending on the amount of absorbed CO2. Van Krevelen's empirical solubility relationships were used to calculate the expected absorbance levels of CO2, and found to be highly dependent on the choice of constants within the statistical ranges of error of the Van Krevelen relationships. (C) 2004 Elsevier Ltd. All rights reserved.