正在加载图片...
M. Verdenelli et al /Journal of the European Ceramic Society 23(2003)1207-1213 acetonate Al(acac)3 at 194C. Decomposition of the temperature(Fig. 5). The results were very similar for precursors occurred at about 210C both samples. They showed a swell which was due to the However, looking closely to the results one can see formation and evolving of the gas phase during the hat the dsc curve for sample 1 showed a broad peak decomposition of the organics around 250-300 oC. at 250 oC, which was spread over a larger range of Then, a contraction was observed for the two samples temperature than sample 2. SDTA results were con-(10-16% linear for, respectively, samples 1 and 2). The sistent with this observation, with an endothermic signal density of the residues was compatible with a very high at 230C, corresponding to the inflexion point of the porous volume, which was consistent with the SEM loss of mass observed in tga characterization. Above 800C, mainly closed macro- An exothermic reaction (AH=300 Jg)was pores were evidenced, with almost no contribution to observed in the range 300-500 oC for 1 and 300-400oc the surface area for 2 corresponding to an oxidation of the latest organic residues. One can also notice on the dsc or SDTA 3.3. Characterization of the coatings curves that the broad exothermic peak following the endothermic phenomena was shifted towards the The coatings elaborated on the fibres in our condi- higher temperature for sample 1 (up to 500C) than tions were usually adherent and crack-free. The thick ple 2. nesses were in the range 0.5-l um depending on the The surface areas were estimated by the Bet method dipping parameters. Cracks were observed when rapid and the results are reported in Fig 4. The surface area thermal annealing(300-500C) was applied. Therefore was relatively high for both systems, even at high tem- drying at 120 oC between the layers and thermal peratures(800 oC). Some differences can be noticed annealing with a rate of 5C/min were performed on between the two samples. The sample 1 (mullite) the samples to prevent cracking. The SEM character showed that the surface area increased between 400 and isations are presented on the 3A1203-2SiO2 composition 600C with a maximum of 510 m-g, while the max-(1)but similar results were observed with the oxides imum of 540 m-g was reached at 400C for the mixture(2). Figs 6 and 7 show typical scanning electron sample 2(Al2O3/mullite). The explanation could be that micrograph for the Al-Si-O coatings. The surface was ne organics were completely removed form the pores very homogeneous and smooth, without apparent defect earlier for sample 2(400C)than for sample 1(600C).(Fig 6a). For a fibre annealed at 500oC in the air, the For both systems the surface area remained high at distribution of the mesopores at the surface looked very 800C(310-370 m'g)and even at 1000C for the regular, and the size of the mesopores(50 nm)was oxides mixture 2(170 m2g). Above 1000C and nearly monodisperse(Fig. 6b). The observation of the 1200C for, respectively, 1 and 2, the surface de ecrease cross-section of a fibre annealed at 1200 oC for 1 h in drastically due to the crystallization of the respective the air showed that the oxide particles size was 50 nr oxides as shown by powder X-ray diffraction and by the(Fig. 7a, b). The pore distribution evidenced on the sur exothermic drift starting at 650C in the SDTA. The face of this fibre was much less homogeneous with sizes surface area decreased less rapidly for 2 because of the of about 100-200 nm(Fig. 7c) highest crystallization temperature of alumina. The thicknesses of the coatings were estimated either The thermo-mechanical analysis(TMA)was used to directly from scanning electron micrographs on the fibres investigate the mechanical behavior as a function of the cross-sections or using laser interferometer. As previously mentioned, the dipping time was one of the most influ ential parameters on the thickness of the films. Usual thickness of the film can be correlated to the withe a300 80o 1000 1200 Termperature〔C Fig. 4. Surface area(BET) for both samples I(A)and 2(.)depend- ing on the temperature Fig. 5. Thermomechanical analysis for samples I and 2.acetonate Al(acac)3 at 194 C. Decomposition of the precursors occurred at about 210 C. However,looking closely to the results one can see that the DSC curve for sample 1 showed a broad peak at 250 C,which was spread over a larger range of temperature than sample 2. SDTA results were consistent with this observation,with an endothermic signal at 230 C,corresponding to the inflexion point of the loss of mass observed in TGA. An exothermic reaction (H=300 J g1 ) was observed in the range 300–500 C for 1 and 300–400 C for 2 corresponding to an oxidation of the latest organic residues. One can also notice on the DSC or SDTA curves that the broad exothermic peak following the endothermic phenomena was shifted towards the higher temperature for sample 1 (up to 500 C) than for sample 2. The surface areas were estimated by the BET method and the results are reported in Fig. 4. The surface area was relatively high for both systems,even at high temperatures (800 C). Some differences can be noticed between the two samples. The sample 1 (mullite) showed that the surface area increased between 400 and 600 C with a maximum of 510 m2 g1 ,while the maximum of 540 m2 g1 was reached at 400 C for the sample 2 (Al2O3/mullite). The explanation could be that the organics were completely removed form the pores earlier for sample 2 (400 C) than for sample 1 (600 C). For both systems the surface area remained high at 800 C (310–370 m2 g1 ) and even at 1000 C for the oxides mixture 2 (170 m2 g1 ). Above 1000 C and 1200 C for,respectively, 1 and 2,the surface decreased drastically due to the crystallization of the respective oxides as shown by powder X-ray diffraction and by the exothermic drift starting at 650 C in the SDTA. The surface area decreased less rapidly for 2 because of the highest crystallization temperature of alumina. The thermo-mechanical analysis (TMA) was used to investigate the mechanical behavior as a function of the temperature (Fig. 5). The results were very similar for both samples. They showed a swell which was due to the formation and evolving of the gas phase during the decomposition of the organics around 250–300 C. Then,a contraction was observed for the two samples (10–16% linear for,respectively,samples 1 and 2). The density of the residues was compatible with a very high porous volume,which was consistent with the SEM characterization. Above 800 C,mainly closed macropores were evidenced,with almost no contribution to the surface area. 3.3. Characterization of the coatings The coatings elaborated on the fibres in our conditions were usually adherent and crack-free. The thicknesses were in the range 0.5–1 mm depending on the dipping parameters. Cracks were observed when rapid thermal annealing (300–500 C) was applied. Therefore drying at 120 C between the layers and thermal annealing with a rate of 5 C/min were performed on the samples to prevent cracking. The SEM characterisations are presented on the 3Al2O3–2SiO2 composition (1) but similar results were observed with the oxides mixture (2). Figs. 6 and 7 show typical scanning electron micrograph for the Al–Si–O coatings. The surface was very homogeneous and smooth,without apparent defect (Fig. 6a). For a fibre annealed at 500 C in the air,the distribution of the mesopores at the surface looked very regular,and the size of the mesopores (50 nm) was nearly monodisperse (Fig. 6b). The observation of the cross-section of a fibre annealed at 1200 C for 1 h in the air showed that the oxide particles size was 50 nm (Fig. 7a,b). The pore distribution evidenced on the surface of this fibre was much less homogeneous with sizes of about 100–200 nm (Fig. 7c). The thicknesses of the coatings were estimated either directly from scanning electron micrographs on the fibres cross-sections or using laser interferometer. As previously mentioned,6 the dipping time was one of the most influential parameters on the thickness of the films. Usually the thickness l of the film can be correlated to the withdrawal Fig. 4. Surface area (BET) for both samples 1 (~) and 2 () depending on the temperature. Fig. 5. Thermomechanical analysis for samples 1 and 2. 1210 M. Verdenelli et al. / Journal of the European Ceramic Society 23 (2003) 1207–1213