Abstract
The sintering of alumina has been studied with a view to produce high quality experimental data to validate theoretical predictions made by a computational model designed to address two key issues in the sintering of real powders: variations in particle size and the removal of large pores. Having produced porous materials the effect of large pores on the mechanical properties and thermal shock behaviour of a monolith was examined. The initial stage of solid state sintering of alumina has been studied using a high temperature in situ monitoring technique incorporating a method for accounting for sample radiance at high temperature. Data were acquired from compacts composed of nominally coarse and fine powders, and powder blends containing different proportions of these powders. Broad agreement was found between experimental observations and predictions with the model accurately describing some of the sintering characteristics associated with the variation of particle size found in real powders i. e. relative rates of shrinkage. Discrepancies between the model requirements and the characteristics of real powders preclude full agreement. Changes in the size of large pores during the sintering of compacts were determined from measurements of sample density following sintering. The results showed that volume of the large pores decreased by a factor of 0. 85 during the sintering process. These observations of pore shrinkage corroborate model predictions of sintering behaviour. The effect of large pores on the mechanical properties and thermal shock resistance was subsequently examined. It was demonstrated that strength and Young's modulus decreased with increased levels of porosity, with strength exhibiting a higher porosity dependence. Experimental data was fitted using an exponential equation in the form of X = Xoexp(-b Vfp) where b was found to be equal to 3.60 and 5.22 for Young's modulus and strength respectively. The porosity dependence observed was greater than that reported in the literature, and attributed to the local variation in pore concentration due to pore clustering. Measurements of work of fracture indicate that the effective surface energy increases by a factor of 1.4 over the range of porosity studied. Material property behaviour was then used to predict the effect of porosity on the thermal shock resistance parameters. It was shown that the critical temperature for the onset of crack growth should decrease and the resistance to crack propagation and the stability of existing cracks should increase with increases in the level of clustered porosity. Measurements of residual strength following thermal shock showed the presence of clustered porosity to result in increases in the relative retained strength comparable to those predicted.