Effect of microstructure and grain boundary chemistry on slow crack growth in silicon carbide at ambient conditions

Effect of microstructure and grain boundary chemistry on slow crack growth in silicon carbide at ambient conditions.

N. Al Nasiri°¹, N. Ni°, E. Saiz°, J. Chevalier^*, F. Giuliani° and L.J. Vandeperre°

°Centre for Advanced Structural Ceramics, Department of Materials, Imperial College London, South Kensington Campus, London SW7 2AZ, United Kingdom

*Université de Lyon, INSA Lyon – Université Claude Bernard Lyon 1, MATEIS, UMR CNRS 5510, 69621 Villeurbanne Cedex, France.

Abstract

Silicon carbide (SiC) is being used increasingly as a room temperature structural material in environments where moisture can’t always be excluded. Unfortunately, there have been almost no reports on slow crack growth (SCG) in SiC at room temperature. To address this gap, SCG in SiC was studied using constant stress rate and double torsion tests in water. SiC based materials were produced with a wide range of grain boundary chemistries and microstructures, which may affect their slow crack growth behaviour. To clarify the role of chemistry and microstructure respectively, solid state (SS) sintering with carbon and boron along with liquid phase (LP) sintering using oxides additives were used to produce materials with fine and coarse grains. The LP-SiC was three times more sensitive to SCG than SS-SiC materials. Moreover, the larger grained material with a higher toughness was less sensitive to SCG than the materials with fine grains.

Keywords: Slow crack growth, SiC, toughness, grain morphology.

1. 
   Introduction
   

Silicon carbide has attracted much attention as a candidate material for high temperature applications. Therefore there is a wide body of research on the mechanical behaviour of SiC at elevated temperature and its interaction with several combustion environments. It is well known that in these environments, the oxidation of SiC causes slow crack growth [1-7].
However, SiC’s properties such as high specific stiffness, low thermal expansion and high thermal conductivity, make it an interesting structural material for many technologies that don’t involve high temperatures. For example, silicon carbide is playing an increasingly more prominent role in the scientific missions of the European Space Agency. A prime example is the 3.5 m diameter silicon carbide main mirror on ESA’s Herschel telescope [8]. Such instruments tend to be manufactured over relatively long periods, and components sometimes need to be stressed for prolonged times on earth in environments where moisture can’t always be excluded. Moreover, SiC is becoming a potential candidate for biomedical applications such as orthopaedic implants [9, 10] and for mechanical seals in hydraulic systems [11], where the components are stressed for extended periods of time in wet environments.
Unfortunately, there have been almost no studies of the SCG in SiC at room temperature. In itself this is not surprising as the covalent bonds of silicon carbide should make it immune to corrosive attack by water [12, 13]. However, all practical silicon carbides are sintered using a diverse range of additives and it is well documented that these additives modify the grain boundary chemistry [14, 15]. As cracks tend to propagate through grain boundaries, their chemistry is likely to have large influence on environmentally assisted failure [16]. In addition to additives, different thermal treatments may lead to a wide range of microstructures that can also affect slow crack growth [13].
The aim of this paper is to quantify the effect of grain boundary chemistry and microstructure on the sensitivity of SiC to SCG. Two typical silicon carbide materials were considered: a solid state sintered material (SS-SiC) with additions of boron and carbon and a liquid phase sintered material (LP-SiC) with additions of alumina and yttria. To determine the effect of microstructure, we studied a fine grain and a larger grain material for each chemistry. The results provide guidance for the selection of materials and the fabrication of silicon carbide parts designed to work under wet environments.

2. 
   Experimental methods
   

2.1 Material processing

For the SS-SiC materials, 3 wt.% of carbon and 0.2 to 0.5 wt% boron (Grade II, H.C. Starck, Germany) was used [17]. Fine grains were achieved by hot pressing α-SiC (UF-25, H.C. Starck, Germany) at 2050 °C for 30 minutes with 0.2 wt.% boron. After several trials, microstructures with elongated grains were obtained by increasing the boron content to 0.5 wt.% and hot pressing β-SiC (BF-17, H.C. Starck, Germany) at 2150 °C for 3 hours. The carbon was added in the form of a phenolic resin (CR-96, Novolak, Crios Resinas, Brazil) with a 50% carbon yield after pyrolysis at 400 ⁰C for 1 hour in an argon atmosphere.

For the LP-SiC materials, a mixture of 6 wt.% of alumina (AKP-30, Sumitomo, Japan) and 4 wt.% yttria (Grade C, H.C. Starck, Germany) was used as sintering additives. The fine equiaxed-grain material was produced by hot pressing α-SiC at 1950 °C for 30 minutes, whereas the larger grains were obtained by hot pressing β-SiC at 2050 °C for 3 hours [18-21].

The SiC powder and additives were wet mixed using methyl ethyl ketone (VWR, London, UK) for 24 hours using 10 mm diameter milling media (Union process, Akron, USA) of either alumina for batches containing oxide additives or silicon nitride for batches with non-oxide additives.

For the non-oxide mixture, the SiC powder and additives were mixed by ball milling using silicon nitride media (Union Process, Akron, USA) and for the oxide- mixture alumina media for the oxide mixture (Union Process, Akron, USA) both in methyl ethyl ketone (VWR, London, UK) for 24 hours. The slurries were dried by rotory evaporation (R-20 BUCHI Rotavapor, Switzerland). After drying, the powders were crushed and sieved through a 100 µm sieve. Hot pressing was conducted in 80 mm diameter graphite dies at heating rate of 10⁰C min^-1 under flowing argon gas in a graphite furnace (FCT, Rauenstein, Germany) under a pressure of 25 MPa.
2.2 Characterisation

The density was determined by Archimedes method with distilled water as the immersion medium according to ASTM standard C8300-00 [22]. The densities were compared to a theoretical density value of 3.28 g/cm³ and 3.21 g/cm³ for the LP-SiC and SS-SiC materials respectively.

The microstructure was examined using scanning electron microscopy (SEM, S-3400N, Hitachi, Japan). The specimens were polished to 1 µm using diamond suspension and chemically etched with boiling Murakami’s solution [23]. The average grain size of 300 grains was calculated using the linear intercept method [24]. For the large grain samples, the average grain length and width was calculated after measuring 300 grains.
The structures and chemistries of the grain boundaries were characterized by Transmission Electron Microscopy (TEM). Cross-sectional TEM foils were prepared by focused ion beam (FIB) milling (Helios NanoLab 600). TEM work was carried out on a FEI Titan 80-300 S/TEM operated at 300 kV. Energy dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS) were performed in STEM mode, and the incident angle α and collection angle β for EELS acquisitions were ~10 and ~14 mrad, respectively.

The Young’s modulus was measured by means of the impulse excitation method [25]. The resistance to fracture was measured in two ways. Firstly, toughness measurements in 3 point bending were performed according to ASTM 1421 using the single notched edge beam (SENB) method [26] with root notch radius of 15 µm. Then R-curve measurements were carried out in situ in the SEM (vacuum level of 1 x 10^-3 Pa) using a constant moment double cantilevered test rig, built in line with the set-up proposed by Sorensen et al. [27]. The silicon carbide specimens were 65 mm by 10 mm by 5 mm. One side of the samples was polished to facilitate the observation of the crack tip. The speed of the stage motor was 0.1 mm min¹. Every time crack propagation was observed, the applied load was recorded and a SEM micrograph was taken.

Slow crack growth (SCG) testing was carried out using both the constant stress rate test in four point bending[28] and the double torsion test [29]. For constant stress rate testing, the specimens measured 4 mm by 3 mm by 40 mm. The tensile face was polished with 1 µm diamond suspension and all the edges were bevelled to eliminate stress concentrations. A 2 kg Vickers indent was placed in the centre of the tensile face with the indent orientated such that 2 of the cracks emanating from the corners created a crack perpendicular to the applied tensile stress. The indentations were made in air with a 10 second holding time, immediately prior to starting the bending test the samples were immersed in distilled water. The fracture stress was measured with an inner span of 10 mm and an outer span of 20 mm at cross head speeds ranging from 0.001 mm min¹ to 1 mm min¹. For every test, the fracture force and time were recorded in order to calculate the fracture stress and stress rate. To determine the exponent n of the SCG law, the following relationship was used [12]:

()

where, is the crack velocity, K_I is the applied stress intensity factor, A and n are material and environment dependent SCG parameters. The slope of log fracture stress versus the log stress rate,, was converted to n using [30]:

()

All fractured samples were checked after the test to confirm that the fracture occurred because of the indent. A total of 5 specimens were tested at each stress rate. For comparison, one set of indented samples were tested in mineral oil with silica desiccant balls with an estimated humidity of less than 1 % to obtain a 100% water free environment at a stress rate of 1 mm min¹.

For double torsion testing, specimens were 40 mm by 20 mm by 2 mm polished with 1 µm diamond suspension. A notch 10 mm long (a₀) with a root radius of 0.3 mm was introduced using a diamond blade. At the end of the notch, two Vickers indentations of 10 kg each were placed in order to obtain a well-aligned sharp crack. A sharp pre-crack length (a_i) of 12-13 mm was achieved by loading the specimen slowly at a cross head speed of 0.05 mm min^-1 followed by fast unloading once the crack propagation was noted by a load drop of 1%. The pre-crack load was recorded and the crack length was measured by means of an optical microscope. The specimen was later loaded at a cross head speed of 0.2 mm min^-1 and once 90-95% of the pre-crack loading value was reached, the cross head position was held fixed to allow load relaxation with time. The crack velocity was calculated from variation of load versus time during relaxation using the following equation [31]:

()

where, is the crack velocity, a_f is the final crack length, P_f is the final crack length load and is the slope of P(t) at a given time. The applied stress intensity was calculated from the load at each time from [32]:

()

where, w_m is the span length, t is the specimen’s thickness, w is the specimen’s width, ν is the Poisson ratio, is a calibration factor for thick specimen determined using elastic theory [33] and m/k is a material dependant correction factor determined using the method explained by Chevalier et al. [29]. The values obtained for SiC are given in Table 2. For each material, two samples were tested in air (45% humidity) and two samples were tested in water.

The applied stress intensity K_I was also calculated according to the approach suggested by Evans [34]. K_I is directly proportional to the applied load:

(5)

where, B a constant that depends on the specimen dimensions and the test device, which can be calculated using the fracture toughness (K_IC) of the material and the force needed to break the double torsion specimen. For each material, one sample was tested in air (45% humidity) at a crosshead speed of 1 mm min^-1. The applied stress intensity was calculated using the constant B and the force values from the load-relaxation data.

3. 
   Results and discussions
   

The microstructures are shown in Figure 1. The LP-Fine consists of homogenous fine equiaxed grains with average grain size of 1.0 ± 0.1 µm. LP-Coarse has elongated grains with a length of 16 ± 1 µm and a width of 5 ± 0.3 µm. This has been achieved using β-SiC powder, which is known to lead to elongated grains due to the enhanced β α transformation at temperatures above 1850⁰C [18, 35-38]. For the SS-Fine, the grains have an average size of 4.0 ± 0.4 µm and are therefore larger than the LP-Fine grains, this is due to the boron addition which enhances grain growth [39]. The grains in the SS-Coarse have a length of 18 ± 1 µm and a width of 5 ± 0.3 µm.
Table 1 shows a summary of the density, Young’s modulus and K_IC for each of the materials. The values of the Young’s moduli are within the range of 421-554 GPa predicted from the single crystal elastic constants [40]. The toughness measured using 3-point bending, agree with typical values reported for these materials. The toughness of LP-SiC with equiaxed grains is typically reported to be in the range of 4.0 to 4.6 MPa m^1/2 [18, 19, 37], while elongated grains can lead to toughness values between 6.0 and 7.5 MPa m^1/2 [19-21, 41, 42]. The toughness of SiC sintered with boron and carbon is consistent with typical literature values of 3 MPa m^1/2 and 3.5 MPa m^1/2 for fine and coarse grained system respectively [20, 21]. This confirms that the observations made here are for materials comparable with other published SiC systems.
The TEM analysis of both LP-Fine and SS-Fine shows evidence of the formation of an amorphous grain boundary film (Figure 2). In the case of LP-Fine, the X-ray energy dispersive spectroscopy (EDS) analysis of this film showed enrichment in aluminium, oxygen and yttrium. However, for the SS-Fine, neither boron or carbon enrichment nor oxide enrichment was observed and oxygen could not be detected in the film. Isolated pockets of a second phase were observed at the triple junctions in LP-Coarse. This phase is amorphous and enriched in aluminium, oxygen and yttrium (Figure 3).

R-curve measurements using double cantilever beam tests were performed for cracks up to 3 mm (Figure 4). The short crack toughness values measured from the bending tests agree quite well with the values for short cracks obtained from the R-curve. The bend test results are somewhat higher even though care was taken to minimize the notch radius to below 50 µm by sharpening with diamond paste. However, the difference is within the range that might be expected. The remarkable rise in the R-curve in LP-Coarse was believed to be due to crack deflection and grain bridging mechanisms as a result of intergranular fracture mode (Figure 5), while these toughening mechanisms were absent in the SS-Coarse due to their transgranular fracture as shown in Figure 5.

The fracture stress of the indented samples measured in water as a function of the stressing rate is plotted in Figure 6. Higher fracture stresses were observed for the LP-SiC materials, but more important is the fact that the fracture stress of the solid state sintered SiC’s hardly varies with loading rate. On the other hand, the fracture stresses of the LP-SiC decrease when the loading rate is reduced.
The exponents, n, and the values of the fracture stresses measured in oil are listed in Table 2. The insensitivity to slow crack growth of the SS-SiC is reflected in the higher value of n comparing to the ones of the LP-SiC. It is also noticeable that the materials with coarse grains exhibit higher n values than the materials with fine grains.
The crack velocity versus the applied stress intensity (-K_I) curves measured using double torsion tests in air and in water are plotted in Figure 7. The calculated values of K_I using Evan’s approach (equation 5) were ~0.5 MPa m^1/2 higher than those calculated using the correction factor (m/k) from Chevalier et al. [29]. This shifts the stress intensity values closer to K_IC. Overall, the differences are small compared to the error of the measurements and therefore the following discussion will be based on the results obtained using the analysis of Chevalier et al. in order to compare the trends for the different materials. The clear shift in the -K_I curve of the LP- SiC when the test environment is changed from air to water indicates that the data should correspond to region I, where the crack growth is highly dependent on the applied stress intensity and testing environment. A shift in -K_I curve between measurements made in air and in water has also been observed in glass [43], alumina [32, 44, 45], yttria stabilized zirconia [31] and some silicon nitrides with a certain amount of glassy phase [46]. It is worth noting that Bhatnagar et al. [47] didn’t observe a shift of the -K_I curve of silicon nitride ceramic sintered with Al₂O₃, Y₂O₃ and MgO additives tested in water versus those tested in dry air. Kruzic et al [48] also did not observe a shift for long cracks in alumina in dry and moist air. However, both investigations were conducted using cyclic loading. It has been shown by Bloyer et al. [49] that crack growth rates under cyclic loading in an aqueous environment are lower than what could be expected from static experiments. This could explain the difference with our observation of a shift under static loading.

It is generally argued that oxide ceramics are prone to slow crack growth because of stress corrosion cracking assisted by water molecules. Indeed, the slow crack growth sensitivity of a given material depends on its affinity to water. Materials presenting ionic bonds are more sensitive to water than covalent bond materials. However, when failure occurs at the grain boundary of covalent materials (as it is the case of the two LP-SiC materials of this study), it is the grain boundary chemistry what dominates the sensitivity to water, hence to slow crack growth. There was no evidence of plateau value, i.e. region II, where the crack velocity becomes independent on the applied stress, as corrosion is controlled by the diffusion rate at which water can reach the tip of the crack. Region II develops approximately at crack velocity above 10^-4 ms^-1 [12], whereas, the measurement of the -K_I of this work was limited to crack growth rates below 10^-4 ms^-1 and this may explain the absence of region II. In contrast, for the SS-SiC there is no difference in the crack velocities measured in air and in water and the very steep slope of the graphs suggest that this may correspond to fast fracture (region III) only (no effect of water molecules on the crack propagation, even for slow crack rates).

The exponents of SCG, n, of all materials were calculated using the same power relationship (equation 1) and have been added to Table 2. The n values obtained from both types of tests are very similar, which confirms the validity of the results. Previously, SiC was not considered susceptible to slow crack growth due to its strong covalent bonding [50] and the very high exponents of the SS-SiC confirm that slow crack growth is negligibly slow below K_IC in these materials. However, for the LP-SiC, the strength does reduce for slower loading and therefore SCG can be observed below K_IC. At n = 60 and 68, the exponent for the LP-Coarse is similar as what has been measured for Si-Y-Al-O-N glasses [51] and the value of the LP-Fine, n = 45 and 44, is comparable to the n = 43 that has been observed for alumina in water [50]. This clearly indicates that the grain boundary chemistry is the dominant factor controlling SCG in these materials. Oxide grain boundary films promote slow crack that is suppressed for materials with oxygen-free interfaces. This is also in agreement with the observation that slow crack growth occurs in silicon nitride (with n = 30) sintered using oxide additives [52].
It is worth mentioning that the exponent n increases with grain size even within the same grain boundary chemistry, which suggests that coarse grained materials have more resistance to slow crack growth than fine grained materials. Ebrahimi et al. [45] also showed that the (-K_I) curve for coarse alumina (12.7 µm) has a higher slope in region I accompanied with a shift to higher K_I compared to fine alumina (1.9 µm). Bridging can shift the (-K_I) curve to higher K_I values, but the shift varies with crack length as it was suggested by Lawn [12]. Therefore, it is suggested here that for short cracks, the elastic bridging is not constant and it varies proportionally with the applied load and the fracture toughness as follows:

(8)

where, K₀ is a constant value, K_IC(a) is a function of the crack length, a. Grain bridging shields the crack tip from the applied stress and stress concentration at the tip, K_tip, is:

(9)

Substituting equation (8) into equation (9) and rearranging the SCG equation (1):

(10)

or

(11)

Following this model, two different situations may occur. For long cracks, which are on the plateau of the R-curve, no change in K_tip is expected and hence no change in slope (n). However, for short cracks on the rising R-curve, the change in will give a change in n.  During the double torsion tests, it is likely that the different microstructures are in different regions of the R-curve leading to changes in the value of K_Itip and therefore variations in the value of n.

Alternatively, it may be that diffusion of water through a tortuous crack is more difficult than through a flat crack leading to higher n values.

5. Conclusions

Silicon carbide materials with different grain sizes and grain boundary chemistries were produced. Due to the formation of grain boundary films containing oxygen, SiC sintered with oxide additives showed a clear evidence of slow crack growth whereas it was virtually non-existing in the solid state sintered materials. The measured slow crack growth exponents, n, for LP-SiC are similar to those reported for glasses and oxide ceramics despite the fact that the amount of additives is limited to 10 wt% and there is a continuous network of grain boundary of oxide nature no continuous oxide phase in the structure. The exponent is not only sensitive to chemistry, but also to microstructure and can be ~ 50% higher for materials with the same chemistry but exhibiting larger grain size.

Acknowledgments

The authors would like to thank the Engineering and Physical Sciences Research Council (EPSRC) of the UK for the funding of this project through grant EP/F033605/1.

References

1. D. J. Rowcliffe and P. A. Huber, Hot Gas Stress Corossion of Silicon Nitride and Silicon Carbide. Proc. Br. Ceram. Soc., 1975. 25: p. 239-259.

2. J. Henshall and D. Rowcliffe, KIc and delayed fracture measurements on hot pressed SiC. Journal of the American Ceramic Society, 1978. 60(1-2): p. 36-41.

3. K. D. Mchenry, E. Richard, and R. E. Tressler, High temperature dynamic fatigue of hot pressed SiC and sintered alpha SiC. The American Ceramic Society, 1980. 59(4): p. 459-461.

4. K. Mchenry and R. Tressler, Subcritical crack growth in silicon carbide. Journal of Materials Science, 1977. 12(6): p. 1272-1278.

5. S. R. Choi, J. A. Salem, and N. N. Nemeth, High temperature slow crack growth of silicon carbide determined by constant stress rate and constant stress testing. Journal of materials Science, 1998. 33: p. 1325-1332.

6. T. E. Easler, R. C. Bradt, and R. E. Tressler, Strength Distribution of SiC ceramics after oxidation and oxidation under load. The American Ceramic Society, 1981: p. 731-734.

7. T. Fett, D. Munz, and K. Keller, Subcritical crack growth in silicon carbide. Journal of Materials Science Letters, 1997. 16(4): p. 324-327.

8. ESA. Small French firm behind Herschel’s giant mirror. 2010; Available from: http://www.esa.int/SPECIALS/Technology/SEMBD6ZNZBG_0.html.

9. J. Will, A. Hoppe, F. Müller, C. Raya, J. Fernández, and P. Greil, Bioactivation of biomorphous silicon carbide bone implants. Acta Biomaterialia, 2010. 6(12): p. 4488-4494.

10. J. Chevalier and L. Gremillard, Ceramics for medical applications: A picture for the next 20 years. Journal of the European Ceramic Society, 2009. 29(7): p. 1245-1255.

11. X ZHAO, Y. LIU, Q. WEN, and Y. WANG Frictional performance of silicon carbide under different lubrication conditions 2013. 1-6 DOI: 10.1007/s40544-014-0039-1.

12. B. Lawn, Fracture of brittle solids. 2n ed. 1993, Cambridge: Cambridge University Press.

13. C. Olagnon, J. Chevalier, and V. Pauchard, Global description of crack propagation in ceramics. Journal of the European Ceramic Society, 2006. 26: p. 3051-2059.

14. G. Choi and S. Horibe, Static fatigue in ceramic materials: Inflences of an intergranular glassy phase and fracture toughness. Journal of Materials Science, 1993. 28(21): p. 5931-5936.

15. Y. Yamauchi, S. Sakai, M. Ito, T. Ohji, W. Kanematsu, and S. Ito, Dynamic fatigue behavior of ceramics at room temperature. The Ceramic Society of Japan, 1988. 96(9): p. 885-889.

16. S. J. Dillon, M. P. Harmer, and J. Luo, Grain bounday complexions in ceramics and metals: An overview. The journal of the Minerals, Metals & Materials Society, 2009. 61(12): p. 38-44.

17. L. Stobierski and A. Gubernat, Sintering of silicon carbide II Effect of boron. Ceramics international, 2003. 29: p. 355-361.

18. N. Padture, In situ toughened silicon carbide. Journal of the American Ceramic Society, 1994. 77(2): p. 519-523.

19. S. Lee and C. Kim, Effects of alpha SiC versus beta SiC powders on microstructure and fracture toughness of SiC sintered with Al2O3-Y2O3 additives. Journal of the American Ceramic Society, 1994. 77(6): p. 1655-1658.

20. S.K. Lee, D.K. Kim, and C.H. Kim, Flaw-tolerance and R-curve behavior of liquid phase sintered silicon carbide with different microstructures. Journal of the American Ceramic Society, 1995. 78(11): p. 65-70.

21. S.K. Lee, Y.C. Kim, and C.H. Kim, Microstructural development and mechanical properties -sintered SiC with plate-like grains using Al2O3-Y2O3 additives. Journal Materials Science, 1994. 29: p. 5321-5326.

22. ASTM C830-00: Standard Test Methods for Apparent Porosity, Liquid Absorption, Apparent Specific Gravity, and Bulk Density of Refractory Shapes byVacuum Pressure. 2011: p. 1-5.

23. W. Lee and W. Rainforth, Ceramic Microstructure. 1994: Kluwer Academic Publishers.

24. M. Mendelson, Average grain size in polycrystalline ceramics. American Society for Testing and Materials, 1969. 52(8): p. 443-446.

25. G. Roebben, B. Bollen, A. Brebels, J. Van Humbeeck, and O. Van der Biest, Impulse excitation apparatus to measure resonant frequencies, elastic moduli and internal friction at room and high temperature. American Institute of Physics, 1997. 68(12): p. 4512-4515.

26. ASTM C1421-10: Standard Test Methods for Determination of Fracture Toughness of Advanced Ceramics at Ambient Temperature. The American Ceramic Society, 2010: p. 1-31.

27. B. Sorensen, A. Horsewell, O. Jorgensen, and A. Kumar, Fracture Resistance measurement method for in situ observation of crack mechanisms. Journal of the American Ceramic Society, 1998. 81(3): p. 661-669.

28. S. M. Wiederhorn, Subcritical crack growth in ceramics, in Fracture Mechanics in ceramics, R.C. Bradt, D.P.H. Hasselman, and F.F. Lange, Editors. 1974, Pelnum Press: New York. p. 613-646.

29. J. Chevalier, M. Saâdaoui, C. Olagnon, and G. Fantozzi, Double Torsion Testing of a 3Y-TZP Ceramic. Ceramics International, 1996. 22(2): p. 171-177.

30. E. Fuller, B. Lawn, and R. Cook, Theory of fatigue for brittle flaws originating from residual stress concentrations. Journal of The American Ceramic Society, 1983. 66(5): p. 314-321.

31. J. Chevalier, C. Olagnon, G. Fantozzi, and B. Cales, Crack propagation behavior of Y-TZP ceramics. Journal of the American Ceramic Society, 1995. 78(7): p. 1889-1894.

32. A. G. Evans, A method for evaluating the time dependent failure characteristic of brittle materials and its application to polycrystalline alumina. Journal Materials Science, 1972. 7: p. 1137-1146.

33. B. J. Pletka, E. R. Fuller, and B. G. Koepka, An Evaluation of double torsion testing-experimental. ASTM STP 678, 1979. 12: p. 29-45.

34. A. Evans, A method for evaluating the time dependent failure characteristic of brittle materials and its application to polycrystalline alumina. Journal of Materials Science, 1972. 7(10): p. 1137-1146.

35. A. H. Heuer, G. A. Fryburg, L.U. Ogbuji, and T. E. Mitchell, The Beta to Alpha transformation in polycrystalline SiC: I, Microstructural aspects. Journal of the American Ceramic Society, 1978. 61(9-10): p. 406-411.

36. C. Greskovich and J. Rosolowiski, Sintering of covalent solids. Journal of the American Ceramic Society, 1976. 59(7-8): p. 336-343.

37. M. Nader, F. Aldinger, and M. Hoffmann, Influence of the alpha/beta SiC phase transformation on microstructural development and mechanical properties of liquid phase sintered silicon carbide. Journal of Materials Science, 1999. 34(6): p. 1197-1204.

38. T.E. Mitchell, L. U. Ogbuji, and A.H. Heuer, Beta to Alpha transformation in polycrystalline SiC: II, Interfacial energetics. Journal of the American Ceramic Society, 1978. 61(9-10): p. 412-413.

39. S. Prochazka, The role of boron and carbon in the sintering of silicon carbide, in Special ceramics 6; Proceedings of the Sixth Symposium, Stoke-on-Trent, Staffs. 1974: England; United Kingdom. p. 171-181.

40. H. A. Bornstein, Elektrische, Piezoelektrische, Pyroelektrische, Piezooptische, Elektrooptische Konstanten und Nichtlineare Dielektrische Suszeptibilitaten, Springer, Editor. 1979: Berlin.

41. D.H. Cho, Y. W. Kim, and W. Kim, Strength and fracture toughness of in situ toughened silicon carbide. Journal Materials Science, 1997. 32: p. 4777-4782.

42. Y.W. Kim, M. Mitomo, and H. Hirotsuru, Grain growth and fracture toughness of fine grained silicon carbide ceramics. Journal of the American Ceramic Society, 1995. 78(11): p. 3145-3148.

43. D. P. Williams and A. G. Evans, A simple method for studying slow crack growth. Journal of testing and evaluation, 1973: p. 264-270.

44. T. Fett and D. Munz, Determination of v-KI Curves by a Modified Evaluation of Lifetime Measurements in Static Bending Tests. Journal of the American Ceramic Society, 1985. 68(8): p. C-213 to C-215.

45. M. E. Ebrahimi, J. Chevalier, and G. Fantozzi, Slow crack growth behavior of alumina ceramics. Journal Materials Science, 1999. 15(1): p. 142-147.

46. W. Hongjie, W. Yonglan, J. Zhihao, and Z. Huijiu, The effect of microstructure on the static fatigue behavior of Si3N4. Journal Materials Science, 1997. 32(21): p. 5775-5778.

47. A. Bhatnagar, M. Hoffmann, and R. Dauskardt, Fracture and subcritical crack growth behavior of Y-Si-Al-O-N glasses and Si3N4 ceramics. Journal of the American Ceramic Society, 2000. 83(3): p. 585-596.

48. J. Kruzic, R. M. Cannon, and R. Ritchie, Effects of Moisture on Grain-Boundary Strength, Fracture, and Fatigue Properties of Alumina. Journal of the American Ceramic Society, 2005. 88(8): p. 2236-2245.

49. D. Bloyer, J. McNaney, R. Cannon, E. Saiz, A. Tomsia, and R. Ritchie, Stress corrosion crack growth of Si-Na-K-Mg-Ca-P-O bioactive glasses in simulated human physiological environment. Biomaterials, 2007. 28: p. 4901-4911.

50. R. F. Pabst and G. Popp, Critical and subcritical crack extension in Al2O3 and SiC as functions of temperature, environment and loading rate. Surface Flaws, Statistics and Microcracking, ed. R.C. Bradt, A.G. Evans, D.P.H. Hasselman, and F.F. Lange. Vol. 5. 1983, New York: Plenum Press.

51. D. De Graaf, H. Hintzen, and G. De With, Subcritical crack growth and power law exponent of Y-Si-Al-O-(-N) glasses in aqueous environment. Journal of Materials Science, 2006. 41(18): p. 6031.

52. G. G. Trantina, Strength and life prediction for hot presses Silicon Nitride. The American Ceramic Society, 1978: p. 377-380.

1^ Corresponding author Dr.Nasrin Al Nasiri. Tel: [phone]

Email address: [e-mail]