Phenomena during thermal removal of binders

The research presented herein has focused on debinding of an ethylene copolymer from a SiC based molded ceramic green body. Examination of the binder burnout process was carried out by breaking down the process into two distinct regions: those events which occur before any weight loss begins, and those events occurring during binder removal. Below the temperature of observed binder loss (175sp°C), both reversible and irreversible displacement was observed to occur. The displacement was accounted for by relaxation of molding stresses, thermal expansion of the system, and melting of the semicrystalline copolymer occurring during heating. Upon further heating the binder undergoes a two stage thermal degradation process. In the first stage, acetic acid is the only degradation product formed, as determined by GC/MS analysis. In this stage, component shrinkage persisted and it was found that one unit volume of shrinkage corresponded with one unit volume of binder removed, indicating that no porosity developed. The escaping acetic acid effluents must diffuse through liquid polymer filled porous regions to escape. The gas pressure of the acetic acid species produced in the first stage of the thermal degradation may exceed the ambient pressure promoting bubble formation. Controlling the heating rate of the specimen maintains the gas pressure below the bubbling threshold and minimizes the degradation time. Experiments have determined the kinetics of the reaction in the presence of the high surface area (10-15msp2/g) ceramic powder and then verified that acetic acid was diffusing through the polymer phase to the specimen surface where evaporation is taking place. The sorption method measured the diffusivity and activity of acetic acid within the filled ceramic system within a TGA. These data were incorporated into a Fickian type model which included the rate of generation of the diffusing species. The modeling process involved prediction of the bloating temperature as a function of the sample size for a given heating rate. The predicted results and experimental results show good correlation. The model was used to optimize the heating schedule which minimized the binder removal time.