Introduction Obtaining rheological data, such as viscosity vs shear rate, on highly ®lled viscoplastic ¯uids such as propel- lants with low shear rate viscosities >1 ´ 10 5 Pa.s) is usually met with signi®cant diculty when using conventional rheometers. In torque driven devices such as cone-and-plate and parallel-plate rheometers it is usually dicult to obtain rheological data at shear rates greater than 0.1±1.0 s )1 because of slip or fracture of the material at the edges. Furthermore, because of the long relaxation times, residual stresses associated with loading the sample can signi®cantly aect steady shear and dynamic oscillatory measurements. Although it is sometimes possible to reach higher shear rates e.g., 10±20 s )1 ) using capillary rheometers, slip at the Rheol Acta 2002) 41: 245±256 Ó Springer-Verlag 2002 ORIGINAL CONTRIBUTION Tung W. Chan Donald G. Baird An evaluation of a squeeze ¯ow rheometer for the rheological characterization of a ®lled polymer with a yield stress Received: 2 April 2001 Accepted: 3 October 2001 T. W. Chan &) Alliant Ammunition and Powder Company, Radford Army Ammunition Plant, Radford, VA 24141±0100 E-mail: tung_chan@atk.com D. G. Baird Department of Chemical Engineering Virginia Polytechnic Institute and State University, Blacksburg, VA 24061±0211 Abstract This paper is concerned with the evaluation of a squeeze ¯ow rheometer for determining the vis- cosity of a highly viscous ®lled viscoplastic ¯uid which is represen- tative of the behavior of some pro- pellants. Ordinarily materials of this nature are dicult to characterize using conventional capillary and rotary rheometers because of melt and edge fracture. A model system was chosen because its viscosity could still be measured up to rea- sonable shear rates of about 1 s )1 in a parallel disk rheometer. The pow- er-law, Bingham, and Herschel- Bulkley empiricisms of the general- ized Newtonian ¯uid were ®tted to the steady shear viscosity data. The lubrication approximation was then used to predict the squeezing force as a function of gap height and the results compared to experimentally determined values. The best results were obtained for the Herschel- Bulkley model and in this case the lubrication approximation predic- tions were on the average some 20% lower than the experimental values between gap to radius ratios of 0.15 and 0.35. A ®nite element simulation of squeezing ¯ow was carried out to assess the limitations of the lubrica- tion approximation. The numerical solution was found to give a good description of the squeezing force for gap to radius ratios between 0.18 and 0.35. It was estimated that the error involved in the lubrication approximation was about 25% at a gap to radius ratio of 0.3 and decreased to below 2% at a gap to radius ratio of 0.05 for the material studied. Experiments were carried out to assess the eect of slip on the squeezing ¯ow measurements using both roughened and lubricated plates. It was found that no signif- icant slip occurred at squeezing speeds up to about 2.4 mm/s. Lubrication was found to reduce the squeezing force by an order of magnitude. Key words Squeezing ¯ow á Yield stress á Herschel-Bulkley ¯uid á Lubrication approximation á Numerical simulation