Joining of silicon carbide by a heat-resistant phosphate adhesive

Abstract

In this paper, SiC was bonded with a phosphate inorganic adhesive prepared from H₃PO₄ and Al(OH)₃ as the matrix and Si and B₄C as the inorganic fillers. The properties of the obtained adhesive were analyzed by XRD, TG-DSC, IR, shear strength tests and SEM. The results showed that the phosphate adhesive exhibited outstanding heat-resistant properties and excellent bonding strength. The addition of B₄C could greatly improve the adhesive's mechanical properties due to the formation of borosilicate glass and the volume compensation came from B₄C's oxidation. The shear strength of bonded specimens could be up to about 10 MPa after heat treatment at 1300 °C, attributed to the generation of mullite and borosilicate as well as the reasonable compensation for the adhesive's volume.

---

1. Introduction

As a construction material, silicon carbide (SiC) is well known as a useful material for applications in harsh environments. Due to its excellent properties such as good thermal and environmental stability, high hardness and good wear resistance, low density and coefficient of thermal expansion (CTE), high thermal conductivity and excellent oxidative stability,^1,2 SiC has been widely used in aerospace, semiconductor, automotive and nuclear industries, etc.^3,4 In many of these practical applications, components with large sizes and complicate shapes are often required. However, due to the high sintering temperature and cost of post processing as well as the poor machinability of silicon carbide, it is difficult to fabricate SiC samples with large sizes and complex shapes directly.⁵ In this case, the joining of relatively small parts together is an alternative route.

Up to now, various joining techniques have been developed, such as brazing,^6,7 diffusion bonding,^8,9 reaction bonding,^10,11 glass or glass ceramic joining,^12,13 and combustion synthesis joining.^14,15 All these techniques require high temperature treatment, in some cases, high pressure, thus reducing the process feasibility.¹⁶ A relative low-temperature and pressure-free joining method can facilitate the fabrication and repair of large structures onsite under a wide variety of field assembly conditions. Thereby, the adhesive bonding is the most promising method because it can join SiC at low temperature without pressure. Besides, the adhesive bonding method is simpler and cheaper.

Nowadays, the application of SiC–SiC joints at high temperature conditions is urgently needed. Of course, the commonly-used organic adhesives for bonding SiC can resist high temperature through adding inorganic fillers.^17,18 Nevertheless, these adhesives have certain limitations affecting their facilities, such as high curing temperature and inert atmosphere protection.¹⁹ So, it is significant to find a low-temperature-curing adhesive providing appropriate bonding property after being treated at high temperature in the air. The phosphate adhesive possesses properties like high temperature resistance, low temperature curing, small curing shrinkage, low-cost production, simple preparation technology and short production cycle.²⁰ If the phosphate inorganic adhesive can join SiC effectively and be applied well at high temperature, it will facilitate the fabrication and repair of SiC components. However, it is known that bonding SiC composites by using heat-resistant phosphate adhesive has rarely been reported. In this article, aluminum phosphate was chosen as the matrix to prepare the room-temperature-curing heat-resistant inorganic adhesive. In order to overcome its volume shrinkage, B₄C and Si are added into the adhesive as volume regulators. Its bonding mechanism and heat-resistant property were studied by using X-ray Diffraction (XRD), Thermogravimetry-Differential Scanning Calorimetry (TG-DSC), Infrared Radiation (IR) and Scanning Electron Microscope (SEM). Meanwhile, the bonding strength was evaluated by measuring the apparent shear strength of bonded specimens.

2. Experimental procedure

SiC sintered substrates with sizes of 40 × 10 × 6 mm and 18 × 10 × 6 mm were supplied by Ankai Seal Co., Ltd, China. The density and the room-temperature bending strength of SiC specimens in this study are 3.05 ± 0.01 g cm⁻³ and over 350 MPa, respectively. In order to avoid the influence of surface roughness on the bonding strength, the bonding surfaces of substrates were polished sufficiently by diamond wheel at first. The average surface roughness (Ra) of SiC was measured to be 0.145 ± 0.07 μm. Then they were ultrasonically cleaned in anhydrous ethanol and dried at 80 °C for 4 h.

Phosphate adhesive consisted of synthetic adhesive matrix and inorganic fillers. Firstly, 85 wt% H₃PO₄ (Kewei Company of Tianjin University, Tianjin, China) was diluted with water to 60 wt% and then the dilute phosphoric acid was heated to 85 °C followed by adding Al(OH)₃ (analytically pure; Kewei Company of Tianjin university, Tianjin, China) based on the molar ratio of 1 : 0.85 (Al : P). The mixture was stirred until the synthetic matrix solution reached suitable viscosity. Finally, the phosphate adhesive was prepared by uniformly dispersing the inorganic fillers of B₄C (d = 6.5–10 μm; Mudanjiang Chenxi Boron Carbide Company, Mudanjiang, China) or Si (d = 0.5 μm; Guangzhou Tuoyi Trade Company, Guangzhou, China) into the matrix solution, based on the mass radio of 5 : 1 (solution: fillers). In order to obtain the optimum dosage of fillers in the adhesive, several adhesives consisting of different concentration of B₄C (0, 5%, 8.75%, 12.5%) are prepared.

The adhesive was applied on bonding surfaces in uniform thickness and the effective bonding area was 20 × 10 mm². Besides, in order to avoid the torsion action in the process of shear strength test, SiC substrates were superposed and bonded to a configuration as shown in Fig. 1 and then cured at room temperature (RT). Its RT-curing property was based on the polymerization reaction between matrix materials.²¹ Bonded specimens were heat-treated at different temperatures ranging from 300 °C to 1500 °C in the air. Each bonded sample was calcined at a given temperature for 1 h. Their shear strength was tested at RT by the Electronic Universal Testing Machine (CSS-44001; Changchun, China) at a cross-head speed of 0.2 mm min⁻¹. It should be noted that shear strength was obtained as the average of five specimens.

Fig. 1 Sample configuration and procedure of shear strength test.

Besides, with the temperature increasing, the influence of volume regulators on the volume change of adhesives was analyzed by comparing the volume of adhesives before and after heat-treatment. Regular cube pieces with the size of 5 × 5 × 5 mm were cut from the curing adhesives. The obtained pieces were also heat-treated at given temperatures mentioned above for 1 h. The accurate size of each piece before and after each temperature heat-treatment was measured by using the micrometer.

In addition, the crystalline phases in the adhesive were indentified by using XRD (Cu Kα radiation D/Max-2500; Rigaku, Akishima, Japan). The chemical bonding was tested by using IR (WQF-510; Mitech Instrument, Beijing, China). The thermostability of the adhesive was tested by TG-DSC (STA 449C; Netzsch Gerätebau, Bavaria, Germany). The field emission scanning electron microscope (S-4800; Hitachi, Tokyo, Japan) was also employed to investigate the cross-section morphology of the bonding area.

3. Results and discussion

3.1. Selection of the optimal adding amount of B₄C

Fig. 2 showed shear strength of specimens changed with the concentration variation of B₄C after being heat-treated at 900 °C and 1300 °C, wherein all data were the average values. As Fig. 2 illustrated, shear strength increased with the elevating concentration of B₄C at the beginning zones of the two curves, therefore, it could be inferred that the addition of B₄C could enhance the adhesive's bonding strength availably. As the concentration of B₄C continued to increase, the shear strength corresponding to two temperatures reduced. At 1300 °C, the shear strength reached the maximum value of 9.7 MPa when the concentration was 5 wt%, while it reduced apparently to 7.9 MPa when the concentration increased to 8.75 wt%. At 900 °C, adhesives performed well at both 5 wt% and 8.75 wt% of B₄C and the shear strength was 5.2 MPa and 5.3 MPa, respectively. Therefore, 5 wt% was considered as the ideal concentration of B₄C and the adhesive with 5 wt% of B₄C was referred as AP + Si + B₄C. In addition, the adhesive consisting of AP and Si (referred as AP + Si) was also prepared as the blank control group and the influence of B₄C would be shown in the following part.

Fig. 2 Dependence of shear strength upon concentration of B₄C.

3.2. The shear strength of specimens at different temperatures

Fig. 3 showed shear strength results of specimens bonded by two kinds of adhesives after being treated at temperatures from RT to 1500 °C, wherein all data were the average values. It was obvious that the addition of B₄C can enhance the adhesive's bonding strength evidently, especially at the heat-treatment temperature range of above 500 °C. Two adhesives' bonding strength had the similar tendency at the temperature range from 900 °C to 1500 °C. For AP + Si + B₄C, as Fig. 3 indicated, the curve consisted of five zones: firstly, from RT to 500 °C, the shear strength of bonded specimens without heat-treatment was 6.7 MPa and then decreased gradually with the temperature increasing and the value was reduced to 5.2 MPa at 500 °C. Secondly, from 500 °C to 700 °C, the shear strength increased to 5.7 MPa at 700 °C. Thirdly, from 700 °C to 900 °C, the strength decreased again to 5.2 MPa at 900 °C. Fourthly, from 900 °C to 1300 °C, the shear strength was improved drastically with the increasing temperature and reached up to about 10 MPa at 1300 °C. Lastly, from 1300 °C to 1500 °C, the value decreased again but still remained above 5.7 MPa. Compared with AP + Si + B₄C, the bonding strength of AP + Si without heat-treatment was close to (still less than) that of AP + Si + B₄C under the same condition, however, the bonding strength of AP + Si was obviously lower than that of AP + Si + B₄C at the temperature range from 900 °C to 1500 °C. The shear strength of samples bonded by AP + Si was only about 3.5 MPa at the treatment temperature of 900 °C and 1500 °C. The bonding mechanism of AP + Si + B₄C and the modification effect of B₄C were discussed in the following analysis.

Fig. 3 The shear strength of bonded specimens after being treated at different temperatures.

3.3. The compositional variation of the adhesive in the whole heat-treatment process

The compositional variation of adhesive with the temperature increasing was closely tied with the change trend of bonding strength. In order to analyze the bonding mechanism of AP + Si + B₄C, its phase change and thermostability with the temperature increasing were analyzed by using XRD, IR and TG-DSC.

3.3.1 XRD analysis of the adhesive at different temperatures. Fig. 4 summarized the XRD spectra of AP + Si + B₄C after heat-treatment at different temperatures from RT to 1500 °C. At RT, the adhesive was composed of amorphous phase, un-reacted Al(OH)₃, Si and B₄C. After being treated at 500 °C, the amorphous phase decomposed and the mixture of crystalline Al(HP₂O₇)·2.5H₂O, Al(PO₃)₃ and hexagonal AlPO₄ was formed. Besides, Al(OH)₃ disappeared and dehydrated into Al₂O₃. It was inferred that Al₂O₃ generated in the process was amorphous as there was no Al₂O₃ peak in the XRD spectrum. Meanwhile, the peak intensity of B₄C decreased at 500 °C, indicating that B₄C had been oxidized into B₂O₃. After thermal treatment at 900 °C, the orthorhombic AlPO₄ appeared, accompanying with the disappearance of Al(HP₂O₇)·2.5H₂O as well as the reduction of hexagonal AlPO₄. In addition, Si began to be oxidized into SiO₂ as the indentified peaks of Si weakened obviously. With the temperature increasing, much more Si was oxidized and the peak of Si almost disappeared at 1100 °C. The indentified phases of the adhesive after treatment at 1100 °C were mullite and orthorhombic AlPO₄. Mullite formed from the reaction of SiO₂ and Al₂O₃ while orthorhombic AlPO₄ came from the transition of hexagonal AlPO₄. According to the spectrum at 1100 °C, there was also an obvious amorphous phase indentified from the pattern and the peaks of B₄C and Si disappeared, implying that it was one of borosilicate glasses. The content of mullite and orthorhombic AlPO₄ increased as the temperature continued to grow and achieved the highest at 1300 °C. When the temperature was up to 1500 °C, the content of glass apparently decreased, which was ascribed to the volatilization of B₂O₃.

Fig. 4 XRD spectra of AP + Si + B₄C after being treated at different temperatures from RT to 1500 °C.

Fig. 5 illustrated the IR spectra of AP + Si + B₄C after heat-treatment from RT to 1500 °C. Contrasted with the spectrum at RT, it was found that the absorption peak intensity at 1090 cm⁻¹ and 500 cm⁻¹ v(–PO) as well as 700 cm⁻¹ v(–AlO) increased apparently after the heat-treatment at 500 °C and 700 °C, corresponding to the decomposition of amorphous phosphates and Al(OH)₃. The absorption peak height at 3434 cm⁻¹ v(–OH) became weak with the heat-treatment temperature increasing and almost disappeared at 900 °C, demonstrating that Al(HP₂O₇)·2.5H₂O remained below 900 °C. The peaks caused by B–O and Si–O began appeared at 500 °C and 700 °C, respectively, which was consistent with the XRD analysis. In addition, when the temperature was above 900 °C, the decrease of peak intensity which was caused by Al–O suggested that Al₂O₃ participated in the formation of glass.

Fig. 5 IR spectra of AP + Si + B₄C after being treated at different temperatures from RT to 1500 °C.

Fig. 6 showed TG-DSC spectra of AP + Si + B₄C after heat-treatment from RT to 1500 °C. It could be clearly concluded that the main stage of mass loss was induced by the decomposition and dehydration of phosphates and Al(OH)₃ as there were three endothermic peaks on the DSC curve before 500 °C. A strong exothermic peak at about 750 °C indicated the oxidation of Si. Meanwhile, the mass of adhesive continued to increase from 750 °C to 1300 °C, which also indicated the oxidation of Si. When the temperature was above 1300 °C, the secondary mass loss on the TG curve as well as the decline of DSC curve was induced by the decomposition of glass and the volatilization of B₂O₃.

Fig. 6 The TG-DSC curves of the AP + Si + B₄C adhesive.

In summary, as the processing temperature increased, the phenomena appeared in the adhesive successively: the decomposition and dehydration of phosphates and Al(OH)₃, the transformation of aluminum phosphates, the oxidation of B₄C and Si, the formation of mullite and borosilicate as well as the regression of the borosilicate. These phase changes with the temperature increasing would definitely influence the adhesive's bonding effect.

3.4. The reaction between the adhesive and SiC substrates at the bonding interface

Through the XRD test on the polished surface of SiC substrates, the residual Si was detected in the SiC. The existence of Si on the surface of SiC substrate would inevitably impact the bonded specimens' strength due to the residual Si would react with the adhesive. Thereby, the reaction between the adhesive and SiC substrate at the bonding interface after heat-treatment was also analyzed by XRD. In order to exactly indentify the phase at the interface, the remaining adhesive on the SiC substrate was scraped completely after shear test. The XRD spectra of the bonding interface after being treated at RT, 900, 1100 and 1300 °C were summarized in Fig. 7 wherein the peak intensity of SiC was subtracted by 50 percent. According to the spectra, no reaction occurred at the interface when the temperature was not above 900 °C. The residual Si on the surface of substrates began to be oxidized into SiO₂ when the temperature was above 700 °C according to the above analysis and then SiO₂ reacted with Al₂O₃ to form mullite. Note that mullite generated at the interface at 1100 °C and could partly diffuse into SiC substrates, thus enhancing the strength of SiC–SiC joints.²² With the temperature increasing, much more mullite generated at the interface as the peak intensity of mullite at 1300 °C became stronger. The chemical reactions at the interface were listed as follows:

Si (s) + O₂ (g) = SiO₂ (s)

2SiO₂ (s) + 3Al₂O₃ (s) = 3 Al₂O₃·2SiO₂ (s)

Fig. 7 XRD spectra of the bonding interface at different temperatures (RT, 900, 1100 and 1300 °C).

3.5. The bonding mechanism of the adhesive

3.5.1 Effect of the concentration of B₄C. The optimum efficiency of adding volume regulators should be able to compensate the volume shrinkage of the adhesive appropriately without causing extra expansion. The excessive expansion could bring about internal stress at the bonding area and decreased the bonding strength. It was well known that the oxidation of B₄C and Si produced different degrees of volume expansion,²³ a proper mass radio between B₄C and Si could reach the desirable expectation. Fig. 8 showed the volume change of adhesives consisting of different concentrations of B₄C with the temperature increasing. When there was only Si filler being included in the adhesive, its complete oxidization could not make up for the entire volume shrinkage. As seen from Fig. 8, the relative volume of the adhesive including no B₄C filler was only 95% after silicon's complete oxidation. However, when the concentration of B₄C was greater than 5 wt%, there was much more extra volume expansion occurring in adhesive. The extra volume expansion of adhesives could be up to 8% and 15%, respectively, corresponding to 8.5 wt% and 12.5 wt% B₄C filler after treatment at 1300 °C. In contrast, only when the concentration was 5 wt%, the adhesive could be compensated reasonably with the relative volume being restored to 100% at 900 °C and no more than 2% extra volume expansion was produced at 1300 °C, revealing that 5 wt% was the optimum concentration of B₄C in the adhesive. This consequence conformed to the shear strength test results. In addition, the material composition of adhesives consisting different content of B₄C after being treated at 1300 °C was analyzed by XRD, as seen from Fig. 9. With the concentration of B₄C increasing, the content of glass in the adhesive decreased gradually through comparing the amorphous phase peak area in a range of 2 theta between 18 degree and 20 degree. It was known that B₂O₃'s high vapor pressure leaded to its volatilization when the temperature was above 1000 °C and the production of borosilicate glass could improve the stability of B₂O₃ phase.^24,25 Since the addition of fillers was constant, the increase of the proportion of B₄C leaded to the decrease of the content of Si. Therefore, when the concentration of B₄C was 12.5 wt%, there was a smaller amount of Si participating in the formation of glass and most B₂O₃ would volatilize. In addition, comparing with the volume expansion caused by the oxidization of B₄C, according to Fig. 8, the volume shrinkage induced by the volatilization of B₂O₃ was very small, implying that there would be a lot of pores produced in the adhesive, thus reducing its bonding strength.

Fig. 8 The volume change of adhesives consisting of different concentration of B₄C with the temperature increasing.

Fig. 9 XRD spectra of adhesives consisting of different concentration of B₄C after being treated at 1300 °C.

3.5.2 Effect of the temperature. Combining the phases change in adhesive and the reaction at the bonding interface which were mentioned in the foregoing parts, the bonding strength variation of AP + Si + B₄C with increasing temperature was concluded in this part. As was know, amorphous or crystalline condensed phosphates had the appropriate bonding property, the bonding effect was extremely relied on phosphates' physical adhesion at temperature range RT to 500 °C since no reaction occurred at the bonding interface.^20,26 The strength of the physical adhesion was mainly depended on the surface roughness of bonding surface. As has been noted, in order to tightly investigate the bonding effect of phosphate adhesive, the SiC substrates had been polished and the bonding surface was very smooth with the Ra of 0.145 ± 0.07 μm. Therefore, the bonding strength of AP + Si + B₄C in this temperature range was not relatively high, but it still had a useable strength performance in the practical application.²⁷ It could be seen from the Fig. 11A that the joining between the adhesive and SiC substrates at RT showed the perfect uniform continuity and no cracks appeared at the interface. At this temperature range, the bonding strength decreased with the temperature increasing, which was due to the dehydration and decomposition of phosphates that leaded to mass loss of the adhesive and the generation of pores in the body. Compared Fig. 10A with Fig. 10B, the adhesive at RT was more compact than that after being treated at 500 °C. Moreover, the mass loss of the adhesive also leaded to the discontinuity between adhesive and substrate and the production of pores at the interface, as seen from Fig. 11B. From 500 °C to 700 °C, the increase of bonding strength was attributed to the generation of B₂O₃. On the one hand, the oxidization of B₄C could compensate certain volume;²⁸ On the other hand, since B₂O₃ melted at 450 °C, the melted B₂O₃ with good fluidness could impregnate into pores and cracks at the bonding layer and B₂O₃ had a good adhesion to SiC.²⁹ Although much more B₂O₃ generated when the temperature reached to 900 °C, the bonding strength decreased again. This phenomenon was mainly abstracted to the completely decomposition of Al(HP₂O₇)·2.5H₂O which was the another main binding phase of the adhesive. And it could also be clearly seen from the curve of AP + Si in Fig. 3 that there was a sharp drop from 700 °C to 900 °C. As Fig. 10C and 11C showed, the density of adhesive increased at 900 °C, but the contact between adhesive and substrate was still not good.

Fig. 10 The SEM images of the internal morphology of AP + Si + B₄C after being treated at RT (A), 500 °C (B), 900 °C (C) and 1300 °C (D).

Fig. 11 The SEM cross-section images of bonded specimens after being treated at RT (A), 500 °C (B), 900 °C (C) and 1300 °C (D).

As the temperature continued to grow, more and more B₄C and Si were oxidized, compensating the volume reasonably. At 1100 °C, mullite that a kind of refractory material generated in the adhesive, enhancing the adhesive's heat-resistance; Moreover, the residual Si of SiC substrates was oxidized and then reacted with Al₂O₃ of adhesive to generate the mullite at the interface, leading to the partial penetration of adhesive in the SiC substrates, thus enhancing the bonding strength. Besides, borosilicate glass formed in the adhesive and it had a good wetting and spreading ability on SiC surface,^30,31 contributing to the increase of the adhesive's strength. When the heat-treatment temperature was up to 1300 °C, the content of mullite and glass increased so that the bonding strength reached the maximum value. Meanwhile, the structure of the adhesive was very compact, no pores and cracks could be seen form Fig. 10D and the glass state was obvious. It could also be observed from Fig. 11D that glass wetted the SiC substrates well and the continuity of interface layer was agreeable. Nevertheless, there were still some large pores in the bonding area, which was caused by the volatilization of B₂O₃.

In conclusion, the enhancement of bonding strength from 900 °C to 1300 °C was attributed to the formation of glass and mullite in the adhesive and the generation of mullite at the interface as well as the volume compensation of adhesive. However, after heat-treatment at 1500 °C, the bonding strength of the adhesive decreased again, which was ascribed to the huge amounts volatilization of B₂O₃. As Fig. 6 indicated, B₂O₃ volatilized seriously above 1300 °C, leading to the decrease of glass.

3.5.3 Effect of the addition of B₄C. It was suggested from Fig. 3 that the addition of B₄C could effectively enhance the bonding strength of the adhesive. The improvement effect of B₄C included three ways: Firstly, as a kind of particle-reinforced phase,³² B₄C could directly improve the strength of the adhesive to some extent. Secondly, B₄C could be oxidized into B₂O₃ at the temperature above 500 °C accompanied with 2.5 times volume expansion, filling pores and compensating the volume shrinkage of adhesive significantly.²⁸ As seen from Fig. 10C and 12A, after heat-treatment at 900 °C AP + Si + B₄C was more compact than AP + Si under the same temperature, and there was much more boron glass phase in AP + Si + B₄C while AP + Si contained a large amount of tiny grains. Thirdly, the addition of B₄C leaded to the formation of borosilicate glass which had very good adhesion effect for SiC.^30,31 Fig. 13 illustrated the XRD spectra of AP + Si at 1100 °C, 1300 °C and 1500 °C. Compared Fig. 4 with Fig. 13, there was no glass phase generated in AP + Si as no amorphous glass phase peak appeared on spectrum and the main phases were AlPO₄ and mullite at this high temperature range. Although the oxidization of Si brought in some volume compensation, the morphology of AP + Si was more grainy and looser after treated at 1300 °C as seen from Fig. 12B.

Fig. 12 The SEM images of the internal morphology of AP + Si after being treated at 900 °C (A) and 1300 °C (B).

Fig. 13 The XRD spectra of AP + Si after being treated at different temperatures (1100, 1300 and 1500 °C).

4. Conclusion

In this study, a heat-resistant adhesive was prepared from H₃PO₄ and Al(OH)₃ as the matrix and Si and B₄C as the inorganic fillers. The addition of B₄C could greatly improve the adhesive's mechanical properties due to the formation of borosilicate glass and the volume compensation came from B₄C's oxidation.

The bonding mechanism of adhesive included physical binding and chemical reaction binding. When the heat-treatment temperature was not above 900 °C, the bonding effect was mainly based on the physical binding which came from the phosphates' own cohesiveness as well as the good wettability and adhesion of B₂O₃ with SiC, the shear strength in this temperature range was about 5.5 MPa. When the temperature was above 900 °C, chemical reaction binding occurred due to the reaction between the adhesive and the residual Si on the SiC substrates, thus enhancing the bonding strength. In addition, the generation of mullite and borosilicate as well as the reasonable compensation of adhesive's volume made the bonding strength reach the maximum of 10 MPa at 1300 °C. The adhesive was potentially applicable to the manufacture of complex components and more importantly it was usable for the repair onsite.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Project no. 51172156 and Project NO. 51272171).

References

1. X. H. Zhang, B. S. Xu, C. Q. Hong, J. C. Han, F. X. Qin, W. B. Han, H. M. Cheng, C. Liu and R. J. He, RSC Adv., 2014, 4, 6591.
2. W. B Tian, H. Kita, H. Hyuga and N. Kondo, J. Eur. Ceram. Soc., 2012, 32, 149.
3. H. L. Wang, X. G. Zhou, J. S. Yu, Y. B.Cao and W. B. Liu, Mater. Lett., 2010, 64, 1691.
4. J. J. Cao, W. J. MoberlyChan, L. C. De Jonghe, C. J. Gilbert and R. O. Ritchie, J. Am. Ceram. Soc., 1996, 79, 461.
5. M. Ferraris, M. Salvo, V. Casalegno, M. Avalle and A. Ventrella, Int. J. Appl. Ceram. Technol., 2012, 9, 795.
6. M. C. Halbiga, B. P. Coddington, R. Asthana and M. Singh, Ceram. Int., 2013, 39, 4151.
7. X. Y. Li, J. Q. Zhang, P. X. Yang and M. Z. An, RSC Adv., 2014, 4, 5853.
8. H. Charles, J. Henager and R. J. Kurtz, J. Nucl. Mater., 2011, 417, 375.
9. H. Y. Dong, S. J. Li, Y. Y. Teng and W. Ma, Mater. Sci. Eng., B, 2011, 176, 60.
10. W. B. Tian, H. Kita, H. Hyuga, N. Kondo and T. Nagaoka, J. Eur. Ceram. Soc., 2010, 30, 3203.
11. S. k. Tang and H. Zhao, RSC Adv., 2014, 4, 11251.
12. M. Herrmann, W. Lippmann and A. Hurtado, J. Nucl. Mater., 2013, 443, 458.
13. T. J. Perham, L. C. De Jonghe and W. J. MoberlyChan, J. Am. Ceram. Soc., 1999, 82, 297.
14. R. Rosa, P. Veronesi, S. Han, V. Casalegno, M. Salvo, E. Colombini, C. Leonelli and M. Ferraris, J. Eur. Ceram. Soc., 2013, 33, 1707.
15. E. Colombini, R. Rosa, P. Veronesi, M. Cavallini, G. Poli and C. Leonelli, J. Mater. Eng. Perform., 2012, 21, 25.
16. Y. Katoh, A. Kohyama, T. Nozawa and M. Sato, J. Nucl. Mater., 2004, 329–333, 587.
17. X. Z. Wang, J. Wang and H. Wang, J. Eur. Ceram. Soc., 2012, 32, 3415.
18. M. Salvo, S. Rizzo, V. Casalegno and M. Ferraris, Int. J. Appl. Ceram. Technol., 2012, 9, 778.
19. M. C. Wang, J. C. Liu, H. Y. Du, F. Hou, A. R. Guo, S. Liu and X. Dong, Ceram. Int., 2013 DOI:10.1016/j.ceramint.2014.03.123.
20. Z. X. Liu, R. N. Sun, Z. P. Mao and P. C. Wang, Surf. Coat. Technol., 2012, 206, 3517.
21. R. H. Hao, J. C. Liu, X. Dong and A. R. Guo, Int. J. Appl. Ceram. Technol., 2013, 10, 978.
22. P. Lemoine, M. Ferraris, M. Salvo and M. Montorsi, Ceram. Trans., 1995, 57, 459.
23. C. J. Engberg and E. H. Zehms, J. Am. Ceram. Soc., 1959, 42, 300.
24. F. J. Buchanan and J. A. Little, Carbon, 1995, 33, 491.
25. T. Piquero, H. Vincent, C. Vincent and J. Bouix, Carbon, 1995, 33, 455.
26. H. Hirai, T. Masui, N. Imanaka and G. Y. Adachi, J. Alloys Compd., 2004, 374, 84.
27. C. Wang, Y. D. Huang and B. Wang, Int. J. Adhes. Adhes., 2006, 26, 206.
28. J. E. Sheehan, Carbon, 1989, 27, 709.
29. Z. H. Luo, D. L. Jiang, J. X. Zhang, Q. L. Lin, Z. M. Chen and Z. R. Huang, Int. J. Appl. Ceram. Technol., 2012, 9, 742.
30. C. Isola, M. Salvo, M. Ferraris and M. A. Montorsi, J. Eur. Ceram. Soc., 1998, 18, 1017.
31. F. T. Lan, K. Z. Li, H. J. Li, Q. G. Fu and X. Q. Lin, Mater. Lett., 2008, 62, 2347.
32. M. J. Chao, X. Niu, B. Yuan, E. J. Liang and D. S. Wang, Surf. Coat. Technol., 2006, 201, 1102.

---