Microstructural variations in SiC ceramics sintered with AlN-Y 2 O 3. Keywords: Silicon carbide, Liquid phase sintering, Microstructure

Transcription

1 Microstructural variations in SiC ceramics sintered with AlN-Y 2 O 3 K.Strecker 1 ; C.Santos 2 ; M. J. Bondioli 2 ; M.-J.Hoffmann 3 1 Universidade Federal de São João del-rei, Departamento de Mecânica, Brazil 2 Faculdade de Engenharia Química de Lorena, Departamento de Materiais, Brazil 3 University of Karlsruhe, Institute for Ceramics in Mechanical Engineering, Germany Keywords: Silicon carbide, Liquid phase sintering, Microstructure Abstract: In this work, SiC ceramics were prepared by liquid phase sintering using 10 wt.% of AlN and Y 2 O 3 mixtures in the proportion of 4:1 and 2:3 as additives Sintering was done in a graphite resistance heated furnace at 2080 o C under 0.2 MPa N 2 atmosphere for 1h. Part of the samples was further heat-treated at 2000 o C for 4h to allow grain growth. The microstructures of the sintered samples were analyzed using polished and plasma etched surfaces. Furthermore, relative densities, weight loss during sintering and SiC-polytype distributions are reported. Introduction Silicon carbide is a compound of relatively low density, high hardness, elevated thermal stability and good thermal conductivity, resulting in good thermal shock resistance. Because of these properties, SiC materials are widely used as abrasives and refractories. The conventional densification route of SiC via solidstate sintering using small quantities of B or Al and C or compounds thereof [1] is widely applied today to produce components on a commercial base. In order to increase the diffusion coefficient, sintering temperatures in the range of o C are applied. An innovative approach to solid-state sintering of SiC has been initiated in the 1980s by Omori and Takei [2], who sintered SiC to high densities via liquid-phase sintering using Al 2 O 3 and Y 2 O 3 as additives. Since then a growing interest for liquid-phase sintered SiC ceramics is noted, because this type of material offers the opportunity of an increased fracture toughness by microstructural control [3 6]. In general, the microstructural variation is achieved making use of the β- to α-sic phase transition. The formation of platelet shaped a-sic grains is thereby influenced by the a-sic content of the starting powders, the additive system, besides the sinter parameters temperature, time and atmosphere. Fracture toughness values of 7 8 Mpam 1/2 have been reported [7, 8], signifying an almost 100% increase when compared to solid-state sintered SiC ceramics. Rixecker and co-workers [9-11] investigated the pressureless sintering of SiC with oxynitride additives, using AlN Y2O3 powder mixtures. They report that using AlN as additive and a nitrogen sintering atmosphere the weight loss during sintering can be effectively minimized, because the decomposition of AlN is avoided by the nitrogen atmosphere. On the other hand, it is known that a nitrogen atmosphere reduces the sintering rates and also hinders the b to a-sic transformation [12, 13].

2 In this work, the influence of microstructural variation of LPS-SiC ceramics on the fracture toughness of liquid-phase sintered silicon carbide ceramics using AlN Y 2 O 3 powder mixtures as additives has been investigated. Experimental procedure For the powder mixtures, β-sic, H. C. Starck, B10, α-sic, Norton AS, FCP-15, AlN, H. C. Starck, Grade C, and Y 2 O 3, H. C. Starck, Grade C were used, as shown in Table 1. Table 1: Composition of powder batches. Sample designation β-sic [wt-%] α-sic [wt-%] AlN [wt-%] Y 2 O 3 [wt-%] A A The powder batches were obtained by attrition milling during 4h at 1000rpm in isopropanol, using Si 3 N 4 balls as milling media, drying and sieving. Green bodies of approximately 40 x 60 x 6 mm3 were obtained by uniaxial pre-pressing and cold isotatic pressing under 400 MPa. The sintering of the samples was done in a furnace with graphite heating elements, Astro Industries (Santa Barbara, USA, CA).. The samples were placed in a graphite crucible using powder beds of identical composition as the samples. The applied sinter programmes consisted in heating up under vacuum up to 1000 o C with a heating rate of 20 o Cmin, then injecting 0,1MPa of nitrogen, and further heating up to 1600 o C with an isothermal holding time of 30min. At the end of this stage the gas pressure was increased to 0,2MPa. Finally, the samples were heated up under a heating rate of 10 o C/min until the maximum sintering temperature of 2080 o C has been reached. Isothermal sintering of the samples was done for 1h. Part of the samples was further heat-treated at 2000 o C for 4h. The cooling rate was 10 o C/min, until the inertia of the furnace prevailed. The theoretical densities of the various sample compositions have been calculated by the rule of mixtures. The bulk density of the sintered samples has been determined by the Archimedes method in distilled water, according Equation 1: md ρ b = ρ H 2O (1) md mi where ρ b represents the bulk density, m d the mass of the dry sample, m i the mass of the sample immersed in water and ρ H2O the density of water. The mass variation of the samples during sintering was determined by the difference of the sample mass prior and after sintering. The phase composition of the sintered samples was determined by X-ray diffraction (Kristalloflex D 500, Siemens, Germany), using monochromatic CuK_ radiation. The samples were investigated between the diffraction angles of 20 to 80o, using a step width of 0,02o and a time of 5s per step. The phases were identified by the calculation of the interplanar spacings d, according Bragg s law, Equation 2, and comparison with the ASTM index: n λ = 2d sen Θ (2)

3 where λ is the wavelength of the radiation, d the interplanar spacing and Θ the diffraction angle. The quantitative analysis of the SiC polytypes 15R, 4H, 6H and 3C has been performed by a programme based on the work of Ruska and Gauckler [14]. For the calculation of the polytypes the intensities of the peaks at 2Θ 33.7, 34.2, 34.9, 35.7, 38.2 and 41.5 o are determined and computed by the programme. Recently, Ortiz et al. [15] compared different methods for the quantitative calculation of the polytype composition by X-ray diffraction. They determined a mean error of about 5 % for the method used in this work. The microstructure of the sintered or hot-pressed SiC samples was investigated by scanning electron microscopy, SEM (Stereoscan 440, LEO, Germany), of polished and etched surfaces. First, the samples were cut with a diamond saw and embedded in resin. After grinding and polishing the samples, the microstructure was developed by etching in a commercial plasma etching apparatus (RF Plasma Barrel Etcher PT 7150, Bio Rad Laboratoric GmbH, Germany). The etchant gas was a mixture of CF 4 and O 2 ; etching was conducted at 100W for up to 120min. The CF 4 gas dissociates under the electromagnetic field forming F radicals, which attack the SiC under formation of SiF 4 and C. The oxygen reacts with the C, forming CO and CO 2. The intergranular phase is not attacked by this method. After etching, the specimens were coated with gold to avoid charging. Results and Discussion The results of the green density, bulk density and relative density after sintering, as well as the weight changes during sintering are listed in Table 2. Table 2: Theoretical and relative density, as well as mass variation of the sintered and heattreated samples. Sample Sinter Conditions Th. density [g/cm 3 ] A 2080 o C, 1h A 2080 o C, 1h o C, 4h A o C, 1h A o C, 1h o C, 4h Rel. density [%] ± ± ± ± 0.28 m [%] From the results presented in Table 2 the following observations can be made: samples with high AlN content, composition A80, revealed very low mass losses during sintering of only 1.2%. This mass loss increased during the subsequent heat-treatment to 2.2%, which still is considered to be quite low. On the other hand, samples of composition A, containing less AlN in the additive system, showed significantly increased weight loss

4 of about 4.7%, which increased to almost 6% during the heat-treatment. This different behavior is attributed to the fact that in the case of the AlN rich composition volatilization of the additives and formation of volatile compounds due to reactions between the additives and SiC are significantly reduced, in aggreement with literature data [9, 11]. The results of the SiC polytypes found by quantitative X-ray analysis of the sintered and heat-treated samples are shown in Table 3. In the starting powder mixture, 4.5% α-sic (4H) have been detected, higher than the intentionally added 1% α-sic as seeds, which is attributed to some α-sic impurities in the β-sic powder, as well as to the precision of the used analyzing method. During sintering the β- to α-sic phase transformation occurs. In the case of samples with lower AlN content, samples A, the predominantly formed polytype is 4H, while in the AlN rich samples, A80, the predominantly formed polytypt is 6H. During the subsequent heat-treatment the α-sic content increases further for both compositions studied. While for samples A the predominant polytype continues to be the 4H modification, but in the case of samples A80 a large part of the 6H modification transforms into 4H. Table 3: Quantitative SiC-polytype composition of initial powder mixtures, sintered and heat-treated samples. Sample Sinter Conditions SiC - polytype 3C 15R 6H 4H As-prepared A A A80 A o C, 1h 2080 o C, 1h o C, 4h 2080 o C, 1h 2080 o C, 1h o C, 4h The microstructures of the sintered and heat-treated samples are shown in Figure 1. Samples of composition A exhibit equiaxial SiC grains after sintering. During the subsequent heat-treatment grain growth is observed, whithout changing the morphology of the grains. Increased porosity of the heat-treated samples is also evident, consistent with the results of the relative density presented in Table 2. On the other hand samples of composition A80 showed elongated SiC grains of high aspect ratio, but also of much smaller grain size as in samples A. During heat-treatment grain growth also occurred, but to a smaller extent when compared to samples of composition A. Furthermore, almost no porosity is observed in samples of composition A80, in aggreement with the relative density shown in Table 2. The difference in the morphology of the SiC grains between samples A and A80 is attributed to the different AlN content in the additive system used. This observation is in aggreement with other works reported in the literature [11], investigating the microstructural development in regard of the additive system used.

5 (a) (b) (c) (d) Figure1: Microstructures of sample A (a) after sintering and (b) after heat-treatment and of sample A80 (c) after sintering and (d) after heat-treatment. Conclusions Significant differences in the liquid-phase sintering of SiC ceramics have been verified in this work, depending on the composition of the additive system AlN-Y 2 O 3 investigated. Higher AlN content led to higher densification and lower weight losses during sintering, because the decomposition of AlN is effectively inhibited by the nitrogen sinter atmosphere and because AlN does not react with SiC forming volatile compounds. Furthermore it has been found that a higher AlN content reduces grains growth during sintering, but favors the formation of elongated, platelet shaped SiC grains. A microstructure composed of high aspect ratio grains may result in increased fracture toughness and strength. Therefore it is of great importance to control the microstructure.

6 References [1] S. Prochazka. The role of Boron and Carbon in the sintering of Silicon Carbide. In P. Popper, editor, Special Ceramics 6, pages , Stoke-on-Trent, UK, The British Ceramic Research Association. [2] M. Omori and H. Takei. J. Am. Ceram. Soc., 65 (1982), p. C 92. [3] N. P. Padture, J. Am. Ceram. Soc., 77 (1994), p [4] Y.-W. Kim, M. Mitomo, and H. Hirotsuru, J. Am. Ceram. Soc., 78 (1995), p [5] Y.-W. Kim, M. Mitomo, and H. Hirotsuru. J. Am. Ceram. Soc., 80 (1997), p. 99. [6] J. K. Lee, H. H. Kang, Y. J. Kim, E. G. Lee, and H. Kim. Key Eng. Mat., (1999), p [7] K. Y. Chia and S. K. Lau, Ceram. Eng. Sci. Proc., 12 (1991), p [8] N. P. Padture and B. R. Lawn, J. Am. Ceram. Soc., 77 (1994), p [9] G. Rixecker, K. Biswas, I. Wiedmann, and F. Aldinger, J. Cer. Processing Res., 1 (2000), p. 12. [10] G. Rixecker, I. Wiedmann, A. Rosinus, and F. Aldinger., J. Eur. Cer. Soc., 21 (2001), p [11] K. Biswas, G. Rixecker, I. Wiedmann, M. Schweizer, G. S. Upadhyaya, and F. Aldinger, Mat. Chem. Phy., 67 (2001), p [12] T. Grande, H. Sommerset, E. Hagen, K. Wiik, and M.-A. Einarsrud J. Am. Ceram. Soc., 80 (1997), p [13] T. Nagano, K. Kaneko, G.-D. Zhan, and M. Mitomo, J. Am. Ceram. Soc., 83 (2000), p [14] J. Ruska and L. J. Gauckler, J. Mat. Sci., 14 (1979), p [15] A. L. Ortiz, F. Sanchez-Bajo, N. P. Padture, F. L. Cumbrera, and F. Guiberteau, J. Eur. Cer. Soc., 21 (2001), p