Liquid phase sintering of silicon carbide with AlN/Y2O3, Al2O3/Y2O3 and SiO2/Y2O3 additions

Abstract

In this work, the influence of the additive system on the liquid phase sintering of silicon carbide has been investigated. The additives employed were mixtures of AlN/Y2O3, Al2O3/Y2O3 and SiO2/Y2O3. The total additive content was fixed at 20 vol.-%, maintaining the Y2O3 content in each additive system at 35 vol.-%. Cold isostatically pressed samples were sintered at 1900, 2000 and 2100 °C under Ar atmosphere during 30 min. The most promising results have been obtained by samples with AlN/Y2O3 additions sintered at 2000 °C, exhibiting the smallest weight loss of about 6% and the highest flexural strengths of about 433 MPa. Samples with Al2O3/Y2O3 and SiO2/Y2O3 additions exhibited high weight loss, because of reactions of Al2O3 and SiO2 with the SiC matrix, forming gaseous species such as Al2O, SiO and CO, resulting in depletion of the liquid phase, and, consequently, in inferior final densities and mechanical properties. Concerning the SiO2/Y2O3 additive system, the reactions seem to be completed already at temperatures below 1900 °C, turning this additive mixture unsuitable. The microstructural analysis indicated only the presence of the <FONT FACE="Symbol">b</FONT>-SiC phase for all samples; no phase transformation of the <FONT FACE="Symbol">b</FONT>-SiC into <FONT FACE="Symbol">a</FONT>-SiC has been observed.

silicon carbide; liquid phase sintering; microstructure

Liquid Phase Sintering of Silicon Carbide with AlN/Y₂O₃, Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃ Additions

Kurt Strecker^a, Sebastião Ribeiro^a, Daniela Camargo^a, Rui Silvab^*,

Joaquim Vieira^b, Filipe Oliveira^b

^aFaculdade de Engenharia Química de Lorena, Departamento de Engenharia de Materiais, C.P. 116, 12600-000 Lorena - SP, Brazil

^b Universidade de Aveiro, Departamento de Cerâmica e Vidro, 3810 Aveiro, Portugal

Received: January 20, 1999 Revised: August 3, 1999

In this work, the influence of the additive system on the liquid phase sintering of silicon carbide has been investigated. The additives employed were mixtures of AlN/Y₂O₃, Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃. The total additive content was fixed at 20 vol.-%, maintaining the Y₂O₃ content in each additive system at 35 vol.-%. Cold isostatically pressed samples were sintered at 1900, 2000 and 2100 °C under Ar atmosphere during 30 min. The most promising results have been obtained by samples with AlN/Y₂O₃ additions sintered at 2000 °C, exhibiting the smallest weight loss of about 6% and the highest flexural strengths of about 433 MPa. Samples with Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃ additions exhibited high weight loss, because of reactions of Al₂O₃ and SiO₂ with the SiC matrix, forming gaseous species such as Al₂O, SiO and CO, resulting in depletion of the liquid phase, and, consequently, in inferior final densities and mechanical properties. Concerning the SiO₂/Y₂O₃ additive system, the reactions seem to be completed already at temperatures below 1900 °C, turning this additive mixture unsuitable. The microstructural analysis indicated only the presence of the b-SiC phase for all samples; no phase transformation of the b-SiC into a-SiC has been observed.

Keywords: silicon carbide, liquid phase sintering, microstructure

1. Introduction

Silicon carbide, SiC, is a covalent compound of low density, high hardness, high thermal stability – decomposition occurs at 2300 °C – and good thermal conductivity, resulting in a good thermal shock resistance. Furthermore, SiC is relatively stable in aggressive environment. Because of these properties SiC ceramics is widely used in refractories and heat exchangers. Other applications include abrasives and heating elements. Only its relatively low fracture toughness is a limiting factor for a wider range of applications as a structural ceramic.

Due to its covalent nature of bonding, the production of dense, monolithic SiC ceramics without sintering aids is impossible. The most common additives used are small quantities of C and B or Al and their compounds, such as B₄C^1-4. In such way it is possible to densify SiC at temperatures between 2100 and 2300 °C up to 99%. The materials obtained – denominated SSiC (sintered silicon carbide) – consist basically of a-SiC grains without an intergranular phase and due to the strong interfacial bonding of the grains a transcristalline fracture mode prevails, resulting in a low fracture toughness of about 3 - 4 MPa m^1/2.

In the last decade were considered the liquid phase sintering of SiC, using oxide mixtures as additives^5-9. These materials, called LPS-SiC (liquid phase sintered silicon carbide), can be densified at temperatures below 2100 °C. Besides a lower sintering temperature, the LPS-SiC materials exhibit another interesting aspect as compared to SSiC, which is that the fracture toughness can be increased by the control of the phase transformation of b-SiC into a-SiC, leading to elongated a-SiC grains formation and resulting in situ reinforcement, similar to Si₃N₄ ceramics¹⁰. The mechanisms contributing to the increase of the fracture toughness are crack deflection and crack bridging¹¹, caused by the microstructural modifications, as well as internal stresses created by the different thermal expansion coefficients of the SiC matrix and the secondary intergranular phase¹².

Even though the majority of the published works report the use of Al₂O₃/Y₂O₃ mixtures as sintering additives, other compounds can be also used as sintering aids, as shown by K. Negita¹³. However, weight loss due to the evaporation of the additives remains a major problem in liquid phase sintering of SiC ceramics.

The principal objective of the present work is to compare the influence of the additive systems AlN/Y₂O₃, Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃ on the liquid phase sintering of SiC ceramics, the microstructure and the mechanical properties.

2. Experimental

The starting powders used in this study were b-SiC B20, Y₂O₃ Grade Fine and AlN all by H. C. Starck and SiO₂ of Merck. Three powder mixtures with three different combinations of sintering additives, AlN/Y₂O₃, Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃, were prepared by attrition milling in isopropilic alcohol. The additive content was fixed at 20 vol.-% of the total mixture and the Y₂O₃ content in the additive systems was kept constant at 35 vol.-%. The overall compositions of the different powder mixtures as well as their denominations are presented in Table 1.

After homogenization, the powder mixtures were dried in a rotoevaporator and deagglomerated by sieving. Particle size distribution was measured by sedigraphy (SYMPATEC HELOS). Samples of approximately 4 x 4 x 50 mm³ were pressed first uniaxially under 20 MPa and then isostatically under 300 MPa. Green density of the pressed samples were determined.

The samples were sintered in a furnace with a graphite heating element under Ar atmosphere, using grafite crucibles and SiC powder as powder bed. The heating rate employed was 10 K/min up to the maximum sintering temperatures of 1900, 2000 and 2100 °C, with a holding time of 30 min. The cooling rate has been the same until the inertia of the furnace prevailed.

The sintered samples were characterized for final density determined by the Archimedes method, weight loss during sintering and linear shrinkage. Prior to the mechanical and microstructural characterization, the samples were grinded and polished with diamond paste. Phase analysis was accomplished by X-ray diffraction, using Cu_Ka-radiation. Young's modulus was determined by the acoustic resonance method, and the Vickers hardness was determined under a load of 9.81 N. Strength was evaluated by the 3-point flexural strength measurements.

Microstructural characterization by scanning electron microscopy (SEM) was conducted on fracture surfaces and on polished surfaces using backscattered electrons in order to obtain the necessary contrast between the SiC grains and the intergranular phase. Relative contents of Y, Al and Si of samples sintered at 2100 °C were determined by EDS analysis.

3. Results and Discussion

Figure 1 shows the particle size distributions of the three different powder mixtures prepared by attrition milling. The average particle size of all three powders is about 0.75 mm. The distribution curves are very similar, indicating the efficiency of attrition milling and the homogeneity of the three powder mixtures prepared.

The green, theoretical and relative densities of the compacted samples prior to intering are resumed in Table 2. The calculations of the theoretical density were based on the rule of mixtures.

The results of the weight loss (Dm/m₀) during sintering, linear shrinkage (Dl/l₀), and final relative densities (r_rel) of the sintered samples are presented graphically in Figs. 2 - 4. As it can be seen, the samples with Al₂O₃/Y₂O₃ and SiO₂/Y₂O₃ additions suffer extremely high weight loss of up to 19.7 and 15.5%, respectively at the sintering temperature of 2100 °C. This may be explained by the interaction between silicon carbide and the additives resulting in gaseous species formation, such as SiO, Al₂O, CO and CO₂^14,15. In the case of SiO₂/Y₂O₃ additions the reactions seem to be completed below 1900 °C, as can be concluded by the almost constant weight loss at all sintering temperatures (Fig. 2). In the case of Al₂O₃ additions the reactions proceed up to 2100 °C, causing a constant increase of weight loss. The higher weight losses observed for the AY samples in relation to the SY samples are due to the higher initial Al₂O₃ weight content in comparison to the initial SiO₂ content (Table 1). On the other hand, samples with the AlN/Y₂O₂ additive system exhibited only minor weight loss, thus maintaining a higher liquid phase content during sintering. These findings are confirmed by EDS analysis of polished surfaces, as shown in Table 3, where the normalized relative contents of Si, Al and Y in the starting powder mixtures (calculated) and after sintering at 2100 °C are listed.

Although the O and C contents could not be analyzed by this method and although these analysis do not represent an overall quantitative analysis of the samples, the trends described above are confirmed, i.e. the Si and Al contents diminish due to the reactions between SiO₂ and Al₂O₃ with SiC, leading to an increase of the Y content. Higher loss of Al is observed for Al₂O₃/Y₂O₃ system in comparison with the AlN/Y₂O₃ one, which is in agreement with the weight loss presented in Fig. 2. Si loss was highest for the SY samples.

As it can be seen from Figs. 3 and 4, the highest densification was achieved for the samples with AlN/Y₂O₃ additives, sintered at 2000 and 2100 °C, although they exhibited a smaller linear shrinkage as the samples with Al₂O₃/Y₂O₃ additives. For the samples sintered at 1900 °C, the highest density was achieved by samples with Al₂O₃/Y₂O₃ additives. This behavior can be explained by the higher refractoriness of the AlN as compared to Al₂O₃, which therefore needs higher sintering temperatures for complete densification. In the case of the SY samples, the linear shrinkage as well as the final relative densities are significantly lower as compared with other samples due to the loss of SiO₂ additive at temperatures below 1900 °C.

Considering all these observations, it can be concluded that the AlN/Y₂O₃ additive system is the most adequate for the liquid phase sintering of SiC ceramics. The weight losses might be reduced even further by the use of an appropriate powder bed. A sintering temperature of 2000 °C is sufficient to obtain high densification, while higher temperatures cause lower final densities due to increasing weight loss caused by evaporation of the additives. Furthermore, the final densities might be improved by longer sintering time.

The results of the Youngs modulus E, Vickers hardness HV1 and flexural strength measurements are presented in Figs. 5 and 6, respectively. Due to the high porosity of the samples with SiO₂/Y₂O₃ additions no meaningful hardness measurements were possible.

As can be observed in Fig. 5, the samples with AlN/Y₂O₃ additions reach their highest flexural strengths of about 430 MPa after sintering at 2000 °C. The other two materials had lower flexural strength, due to their lower final densities. The Young's modulus values plottet vs. sintering temperature in Fig. 6 exhibit similar trends to the ones of the final density of the sintered samples, demonstrating the direct influence of porosity on the elastic behavior.

The phase analysis of the sintered samples at 2100 °C by X ray diffraction revealed only the b-SiC phase in all cases. Trace amounts of the Al₅Y₃O₁₂ phase appeared in the case of Al₂O₃ and AlN additions. No secondary cristalline phase was identified in the case of SiO₂ additions. Figure 7 shows a X-ray diffractogram of a sample with Al₂O₃/Y₂O₃ additions sintered at 2100 °C. No phase transformation of b-SiC into a-SiC has been observed, probably due to the high purity in terms of the initial a-SiC phase content of the SiC starting powder and due to the short sintering time of only 30 min.

The microstructures of polished and fracture surfaces of the samples sintered at 2100 °C are shown in Fig. 8. In the case of AlN/Y₂O₃ and Al₂O₃/Y₂O₃ quite similar microstructures are observed. Both exhibit a fine grained microstructure of equiaxial SiC grains with grain size smaller than 2 mm. No exaggerated grain growth or typically elongated a-SiC grains have been observed, being in agreement with the results of the X ray analysis. In comparison, the samples with SiO₂/Y₂O₃ additions exhibit significant grain growth in relation to the other two compositions, with grain sizes between 3 to 5 mm. Again, no phase transformation of b-SiC into a-SiC has been observed in this case. The fracture surfaces are quite smooth, indicating a predominantly transcristalline fracture mode, leading to low fracture toughness. No definitive conclusion can be drawn in the case of the SiO₂/Y₂O₃ doped samples, because of their very low densities.

4. Conclusions

Considering the results presented, it is concluded that the additive system based on the AlN/Y₂O₃ mixture is the most adequate for the liquid phase sintering of SiC. Due to the higher refractoriness of AlN, in relation to Al₂O₃, samples with AlN addition reach the highest densities at higher sintering temperatures. The weight losses of these samples might be reduced using powder beds with identical compositions, resulting in higher densities and hence in improvement of mechanical properties. The SiO₂/Y₂O₃ system is inadequate for the liquid phase sintering of SiC ceramics, due to the SiO₂ loss at low temperatures.

The fact that no b-SiC to a-SiC phase transformation has been observed, may be explained by the short sintering time of only 30 min and by the absence of a-SiC grains in the starting powder, which hinders the nucleation of the a-SiC phase. To induce the phase transition of b-SiC into a-SiC a prolonged heat treatment or the use of a-SiC seeds seems to be necessary.

In the case of the SiO₂/Y₂O₃ additive system, an intensive grain growth has been observed as compared to the two other additive systems investigated. The high porosity of this samples indicates that the grain growth probably occurs via an evaporation – condensation mechanism.

Considering the final relative density of about 95% of the samples with AlN/Y₂O₃ additions and their flexural strengths of about 430 MPa, the high potencial of LPS-SiC as a structural ceramic material was demonstrated in this work.

Acknowledgement

The authors would like to thank the financial support by CAPES-ICCTI under grant no. 019/97 and by FAPESP under grant no. 95/06378-5.

• 1. Prochazka, S. Proceedings of the Conference on Ceramics for High Performance Applications, Hyanuis, MA, 1973. Burke, J.J.; Gorum, A.E.; Katz, R.M., eds., Brook Hill Publishing Co., p. 7-13, 1975.
• 2. Prochazka, S. Special Ceramics 6, British Ceramic Res. Assoc., Stoke-on-Trent, p. 171-181, 1975.
• 3. Johnson, C.A., Prochazka, S. Ceramic Microstructures `76, edited by Fulrath and Park, p. 366-378, 1977.
• 4. Murata, Y.; Smoak, R.H. Proceedings of the International Symposium of Factors in Densification and Sintering of Oxide and Non-oxide Ceramics, October 3-6, Hakone, Japăo, p. 382-399, 1978.
• 5. Cutler, R.A.; Jackson, T.B. Ceramic Materials and Components for Engines. Proceedings of the Third International Symposium Tennery,V.J., ed., The American Ceramic Society, Westerville, OH, p. 309-318, 1989.
• 6. Cordery, L.; Niesz, D.E.; Shanefield, D. J. Sintering of Advanced Ceramics, v. 7, Handwerker, C.A.; Blendell, J.E.; Kaysser, W.A., eds., The American Ceramic Society, Westerville, OH, p. 618-636, 1990.
• 7. Kim, D.H., Kim, C.H. J. Am. Ceram. Soc., v. 73, n. 5, p. 1431-1434, 1990.
• 8. Chia, K.Y., Lau, S.K. Ceram. Eng. Sci. Proc., v. 12, n. 9-10, p. 1845-1861, 1991.
• 9. Kim, Y.W. et al. J. Am. Ceram. Soc., v. 78, n. 11, p. 3145-3148, 1995.
• 10. Tani, E. et al. J. Am. Ceram. Bull., v. 65, n. 9, p. 1311-1315, 1986.
• 11. Lee, S.K. et al. J. Am. Ceram. Soc., v. 78, n. 1, p. 65-70, 1995.
• 12. Padture, N.P. J. Am. Ceram. Soc., v. 77, n. 2, p. 519-523, 1994.
• 13. Negita, K. J. Am. Ceram. Soc., v. 69, p. C-308 - C-310, 1986.
• 14. Jacobson, N.S. J. Am. Ceram. Soc., v. 75, p. 1603-1611, 1992.
• 15. Hue, F.; Yorand, Y.; Dubois, J.; Fantozzi, G. J. Europ. Ceram. Soc., v. 17, p. 557-563, 1997.

Publication Dates

• Publication in this collection
  28 Jan 2000
• Date of issue
  Oct 1999

History

• Reviewed
  03 Aug 1999
• Received
  20 Jan 1999