J. She et al. Materials Science and Engineering 4325(2002)19-24 solution(Merck, Germany) for 2 h to achieve alu- 3. Results and discussion ninum-hydroxide [Al(OH)3] gelation. After drying on a hot plate, the specimens were further heated at A350 Fig. 1 shows the change of weight increase and oC for 15 min to decompose Al(OH)3 into Al,O3. The matrix porosity with cyclic infiltrations for porous mul- processes of infiltration, gelation and decomposition lite/mullite composites. As expected, the weight gain repeated ten times. To stabilize amorphous Al, O, increases and the matrix porosity decreases as the num n equilibrium o-structure, a final heat-treatment ber of infiltration cycles increases. After ten infiltra- was given at 1200C for 30 min. The weight of the tions, the matrix porosity decreased from 65 to 52% specimens before infiltration and after heat-treatment due to the introduction of c 12 wt% alumina into the vas measured to calculate the amount of Alo that void space within the composite. Based on a theoretical was incorporated into the mullite matrix. Density and analysis [ll], the matrix porosity, Pm, is related to the porosity were determined by Archimedes' method number of infiltration cycles, N, through the following Flexural strength was measured by a three-point bend distance of 20 Pm=P(1-Y2), cross-head speed of 0.5 mm min. Tests were inter rupted when the load dropped below 5N where Pm is the initial content of pores in the matrix. cross-head displacement exceeded 0.4 mm. Load-dis and Ye is the ceramic yield by volume of the infiltrant placement responses were recorded with a computerized It can be seen in Fig. 1(b) that the fitted curve with data-acquisition system. Fracture energy was evaluated Pm=64.6% and Y=2.28% is in good agreement with from the area under the load-displacement curve in the the experimental data. If an infiltrant with a higher Yc non-elastic region. Polished sections and fracture sur value is used, the infiltration process would be more faces were observed in a scanning electron microscope effective in reducing the porosity (SEM) It has been demonstrated in the earlier work [12] that there is a non-uniform distribution of the residual pores within the infiltrated composites. As shown in Fig. 2, the matrix porosity is lower in the surface region, and tends to increase towards the interior. This is thought to result from two events which may occur during infiltration and gelation. As reported elsewhere [12], the intruded solution is filtered' as it passes through the pore channels due to a good adsorbability of amor- phous alumina that was introduced into the pore space 6 by previous cycles. On the other hand, the gelling process of the intruded solution in the central region is prevented due to the formation of the Al(OH)3 gelation in the surface region. Since the former may produce a decreasing concentration of the intruded solution from 0 the surface inward and the latter may cause an incom- plete gelation of the intruded solution in the interior region, any or both of them could give rise to less alumina and thus more pores in the interior region relative to the surface region. This phenomenon is less 62 noticeable for the composites within four infiltrations, while it becomes substantially more pronounced for five to ten infiltrations. Moreover, it was found that the thickness of the 'dense surface region increased with the infiltrated composites, some closer inspections were further performed on the polished sections. Fig. 3 illus trates a typical microstructure of the composite after 0 4 8 10 ten infiltration cycles, where mullite appears gray and Number of Infiltration Cycles alumina is in bright contrast. Evidently, the alumina phase is always located within the interstices between Fig. 1. Weight increase and matrix porosity as a function of the matrix particles or between matrix and fibers. Since number of infiltration cycles for porous mullite/mullite composites amorphous alumina sinters readily above 800C [13].20 J. She et al. / Materials Science and Engineering A325 (2002) 19–24 solution (Merck, Germany) for 2 h to achieve alu￾minum-hydroxide [Al(OH)3] gelation. After drying on a hot plate, the specimens were further heated at
350 o C for 15 min to decompose Al(OH)3 into Al2O3. The processes of infiltration, gelation and decomposition were repeated ten times. To stabilize amorphous Al2O3 to an equilibrium
-structure, a final heat-treatment was given at 1200 o C for 30 min. The weight of the specimens before infiltration and after heat-treatment was measured to calculate the amount of Al2O3 that was incorporated into the mullite matrix. Density and porosity were determined by Archimedes’ method. Flexural strength was measured by a three-point bend￾ing test with a support distance of 20 mm and a cross-head speed of 0.5 mm min−1 . Tests were inter￾rupted when the load dropped below 5 N or the cross-head displacement exceeded 0.4 mm. Load–dis￾placement responses were recorded with a computerized data-acquisition system. Fracture energy was evaluated from the area under the load–displacement curve in the non-elastic region. Polished sections and fracture sur￾faces were observed in a scanning electron microscope (SEM). 3. Results and discussion Fig. 1 shows the change of weight increase and matrix porosity with cyclic infiltrations for porous mul￾lite/mullite composites. As expected, the weight gain increases and the matrix porosity decreases as the num￾ber of infiltration cycles increases. After ten infiltra￾tions, the matrix porosity decreased from 65 to 52% due to the introduction of
12 wt.% alumina into the void space within the composite. Based on a theoretical analysis [11], the matrix porosity, Pm, is related to the number of infiltration cycles, N, through the following expression: Pm=P m 0 (1−Yc) N, (1) where P m 0 is the initial content of pores in the matrix, and Yc is the ceramic yield by volume of the infiltrant. It can be seen in Fig. 1(b) that the fitted curve with P m 0 =64.6% and Yc=2.28% is in good agreement with the experimental data. If an infiltrant with a higher Yc value is used, the infiltration process would be more effective in reducing the porosity. It has been demonstrated in the earlier work [12] that there is a non-uniform distribution of the residual pores within the infiltrated composites. As shown in Fig. 2, the matrix porosity is lower in the surface region, and tends to increase towards the interior. This is thought to result from two events which may occur during infiltration and gelation. As reported elsewhere [12], the intruded solution is ‘filtered’ as it passes through the pore channels due to a good adsorbability of amor￾phous alumina that was introduced into the pore space by previous cycles. On the other hand, the gelling process of the intruded solution in the central region is prevented due to the formation of the Al(OH)3 gelation in the surface region. Since the former may produce a decreasing concentration of the intruded solution from the surface inward and the latter may cause an incom￾plete gelation of the intruded solution in the interior region, any or both of them could give rise to less alumina and thus more pores in the interior region relative to the surface region. This phenomenon is less noticeable for the composites within four infiltrations, while it becomes substantially more pronounced for five to ten infiltrations. Moreover, it was found that the thickness of the ‘dense’ surface region increased with infiltration cycles. In order to highlight the microstructural features of the infiltrated composites, some closer inspections were further performed on the polished sections. Fig. 3 illus￾trates a typical microstructure of the composite after ten infiltration cycles, where mullite appears gray and alumina is in bright contrast. Evidently, the alumina phase is always located within the interstices between matrix particles or between matrix and fibers. Since amorphous alumina sinters readily above 800 o C [13], Fig. 1. Weight increase and matrix porosity as a function of the number of infiltration cycles for porous mullite/mullite composites