The atmosphere is one of the most important parameters for microstructure development during sintering or heat-treatment of alloys. In the present study, the effect of atmosphere, gaseous or solid, on the microstructure of $SrTiO_3$ has been investigated. First, the densification of powder compacts in insoluble gas atmosphere has been calculated (chap.2). Based on the calculation, the effect of atmosphere gas on the densification of $SrTiO_3$ has been studied(chap.3). In recent, the interface migration caused by the change of surrounding atmosphere has been found to be common phenomenon in ceramics as well as in metals. Since the grain boundary migration in $SrTiO_3$ can cause considerable change of electric properties, the cause of the phenomenon has been studies by using the powder-packing technique during heat-treatment(chap.4). $CaTiO_3$ and $BaTiO_3$ were used as the packing-powder atmosphere.
In chapter 2, limits to the shrinkage of isolated pores and to the densification of powder compacts have been calculated for various sintering conditions. The impeding effect of entrapped gases diminishes with lower gas pressure in the sintering atmosphere, with higher solid-gas interfacial tension, with smaller initial pore size, and with higher dihedral angle. Over 99.5% relative density can be achieved under conventional conditions when sintering fine powders with a pore size of less than a few micrometers. In contrast, the final density can be much reduced in coarse powder compacts and under high initial atmosphere pressure. Therefore the gas pressure sintering at the beginning is not recommended if the entrapped gases are not fast diffusing. The driving force and the rate of densification of fine powder compacts are not seriously impeded by the entrapped gases of 1 atm except for final densification; the sintering time for reaching the limiting density is about the same as that for full densification without entrapped gases. The increase of attainable density by excess external pressure, similar to sintering+HIP, is considerable even for medium-pressure sintering of an order of 10-100 atm. Higher external pressure, however, has its justification in faster densification; the sintering time can be reduced by a factor of the increase of external pressure. The effect of pore coalescence due to grain growth on attainable density is also calculated for sintering+HIP as well as pressureless sintering. A scaling of the initial pore size by the number of pores in coalescence can give an estimation for the maximum attainable density for any experimental condition.
In chapter 3, $SrTiO_3$ specimens sintered at 1400℃ for 4 h in $H_2$ or $H_2+N_2$ have been annealed at 1400℃ for 4 to 8 h in $O_2$. The relative density of specimens initially sintered in $H_2$ or $H_2+N_2$, was between 83 and 85%. During the annealing(heat-treatment) in $O_2$, density of specimen sintered in $H_2$ increases to 94%, but the density of specimen sintered in $H_2+N_2$ is not changed. The difference in density between the two different specimens increase with heat-treatment time in $O_2$. Higher density during annealing of the specimen sintered in $H_2$ is attributed to the diffusion-out of hydrogen from the specimen. For specimens sintered in H2 and heat-treated in $O_2$, denser region spreads from surface region to center region with the increase of heat-treatment time. The diffusion of hydrogen is therefore thought as the major factor controlling the densification during sintering of $SrTiO_3$ in $H_2$ or $H_2+N_2$.
In chapter 4, grain boundary migration induced by chemical composition change has been investigated. Compacts of $SrTiO_3$ powder have initially been sintered at 1450℃ for 4 h in air. The sintered specimens have been packed in CaO, $CaTiO_3$ or $BaTiO_3$ powder and annealed in air at various temperatures between 1200 to 1400℃ and various times from 20 to 160 h. In most specimens, the grain boundary migrates, forming a new solid solution containing cations of the packing powder. In specimens packed with CaO powder, the migration distance increase linearly with the annealing time. The migration of some grain boundaries results in a corrugated boundary shape. When $CaTiO_3$ is used as packing powder, grain boundary migration as well as grain growth are observed in the same specimen heat-treated at 1400℃. The migrated the grown areas contain Ca ion. The linear dependance of migration distance with time in CaO packed specimen and the continuous change of microstructure in $CaTiO_3$-packed specimen suggest that the driving force for the migration and the microstructure change is the coherency strain energy in solute diffusion zone of shrinking grain in front of the moving boundary.
In order to test conclusively the coherency strain energy model for the grain boundary migration in $SrTiO_3$, sintered $SrTiO_3$ has been packed with various $BaTiO_3/CaTiO_3$ powder mixtures and annealed at 1400℃ for various times. At the surface region of most $SrTiO_3$ specimens, the grain boundary has migrated with solutioning of Ba and Ca. The ratio of Ba and Ca solutioned varies with the depth from surface and the packing-powder composition, and is not equal to the cation ratio composition of packing powder. When the Ba/Ca ratio of packing powder is 1, cessation of migration has been observed. The composition of the area without migration is approximately equal to the composition for the matching of original grain($SrTiO_3$) and the diffusion zone formed by solutioning. The coherency strain energy induced by diffusion of solute atoms is therefore believed to be the major driving force for the grain-boundary migration of $SrTiO_3$ under chemical instability.
