hBN composites

Author’s Accepted Manuscript Effect of rare earth oxides on the microstructure and properties of mullite/hBN composites Han Jin, Zhongqi Shi, Xiaodan Li, Yongfeng Li, Hongyan Xia, Zhuo Xu, Guanjun Qiao www.elsevier.com/locate/ceri

PII: DOI: Reference:

S0272-8842(16)32191-5 http://dx.doi.org/10.1016/j.ceramint.2016.11.179 CERI14267

To appear in: Ceramics International Received date: 18 August 2016 Revised date: 24 November 2016 Accepted date: 25 November 2016 Cite this article as: Han Jin, Zhongqi Shi, Xiaodan Li, Yongfeng Li, Hongyan Xia, Zhuo Xu and Guanjun Qiao, Effect of rare earth oxides on the microstructure and properties of mullite/hBN composites, Ceramics International, http://dx.doi.org/10.1016/j.ceramint.2016.11.179 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effect of rare earth oxides on the microstructure and properties of mullite/hBN composites Han Jina, Zhongqi Shia*, Xiaodan Lia, Yongfeng Lib,c*, Hongyan Xiaa, Zhuo Xub, Guanjun Qiaoa,d a

State Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong University,

Xi’an 710049, China b

Electronic Materials Research Laboratory, Xi’an Jiaotong University, Xi’an 710049,

China c

School of Electronic Engineering, Xi’an University of Posts and Telecommunications,

Xi’an 710121, China d

School of Materials Science and Engineering, Jiangsu University, Zhenjiang 212013,

China [email protected] [email protected] *

Corresponding author: Tel.: +86 29 82667942; fax: +86 29 82663453;

Abstract Mullite/hBN composites were fabricated with different rare earth oxides additives (ReO: Er2O3, CeO2, La2O3, Lu2O3) by pressureless sintering at 1600oC for 4 h. The impacts of ReO on the phase composition, microstructure, mechanical, dielectric and tribological properties of the composites were investigated. XRD results showed that all the ReO additives were beneficial to the formation of mullite phase. With the decrease in the ionic radius of the ReO, the mullite grains of the composite were refined while their mechanical properties were increased. The sample sintered with Lu2O3 showed the 1

smallest grain size and the most excellent mechanical properties, e.g., its relative density, flexure strength, fracture toughness and hardness reached 93.7%, 222 MPa, 3.2 MPa·m1/2 and 6.02 GPa, respectively. Due to the different porosity and phase composition of the composites, the sample sintered with La2O3 had the lowest dielectric constant while the sample sintered with Er2O3 exhibited the lowest dielectric loss. Attributing to the highest hardness and fracture toughness, the sample sintered with Lu2O3 showed the best tribological properties. Keywords: Mullite; Hexagonal boron nitride; Rare earth oxide; Properties

1. Introduction Both mullite and hexagonal boron nitride (hBN) ceramics are the materials of great importance to vary engineering fields. As for mullite, the only stable binary compound in the Al2O3-SiO2 system, is an important engineering ceramic material due to its good creep resistance, low thermal expansion coefficient, excellent chemical and thermal stability [1-3]. hBN also has many superior properties, such as high temperature resistance, high thermal conductivity, chemical inertness, good machinability and dielectric properties [4-8]. Therefore, it gives the chance to combine the two materials together, so as to fabricate a novel composite with excellent properties, such as good thermal shock resistance, ablation resistance and low dielectric constant [9, 10], which might be used for radome or some high-temperature applications. As known, due to the specifical laminated structure, hBN tends to grow along the layer direction rather than the c-axis in the sintering process, and thus the broad plate hBN crystals are formed with stacking and supporting mutually, which deteriorates the sinterability of the materials seriously. Therefore, in order to destroy this card-house structure, hot pressing

2

(HP) is generally adopted as an appropriate method for improvement in the densification of the hBN composites. Actually, Zhang et al. has fabricated the mullite/hBN composites successfully via reactive HP method [11]. However, the high requirements on equipment and the only simple shape of the products limit their wide application [12-14]. Liquid phase sintering has been reported as another efficient method in majority of the ceramics’ densification, through the flowing of liquid phase in sintering process, which is beneficial to the atoms migration [15]. Hence, the pressureless liquid phase sintering would be the more promising way to fabricate mullite/hBN ceramics. To date, rare earth oxides (ReO) have been found to be effective liquid sources in the ceramics sintering process. These ReO additives could not only react with ceramics and result in the reduction of sintering temperature, but also helpful in strengthening the grain boundaries. Many efforts have been devoted to study the influence of ReO (La2O3, Y2O3, CeO2, etc.) on different kinds of pressureless-sintered hBN composites, such as SiC/hBN, Si3N4/hBN, SiAlON/hBN, Al2O3/hBN, AlN/hBN, etc [16-21]. However, for the mullite/hBN composites, there is little research reported on. In our previous work, we found that the sinterability of hBN ceramics could be improved by a pre-oxidation treatment to the hBN particles. Due to the fact that a BNO amorphous layer was coated, hBN particle could form a transition layer with the ceramic matrices in sintering process which is beneficial to the densification [20, 22]. Hence, in this paper, mullite/hBN composites containing 10 wt.% pre-oxidized hBN powders with different ReO additives (Er2O3, CeO2, La2O3, Lu2O3) were fabricated by pressureless sintering at 1600

o

C, and the phase composition, microstructure,

mechanical, dielectric and tribological properties of the composites were investigated,

3

respectively.

2. Experimental The raw materials used in this experiment were Al2O3 (>99% in purity, 5 μm), SiO2 (>99% in purity, 15 μm), hBN (>99% in purity, 10μm) and ReO (Er2O3, CeO2, La2O3, Lu2O3. >99% in purity, 1 μm). At the beginning, the hBN powders were pre-oxidized in 600 oC with flowing air for 168 h [22]. Fig. 1 shows the SEM images and the XPS spectra of the raw hBN (r-BN) and the oxidized hBN (o-BN), respectively. As can be seen, there is no significant change in the particle size of hBN before and after pre-oxidation, and many debris contained with lots of oxygen is formed on the surface of o-BN. HRTEM image of o-BN is shown in Fig. 2, which demonstrated a core-shell structure with h-BN crystal as the core and BNO amorphous layer of 5~10 nm as the shell. Then, different kinds of powders with the composition of 61 wt.% Al2O3, 24 wt.% SiO2, 10 wt.% o-BN and 5 wt.% of different ReO were mixed with ethanol using Al2O3 balls for 24 h, respectively. The mixed slurry was dried at 80 oC and then sieved to 200 mesh. Followed by pre-compact into rectangular green compacts (30 mm × 30 mm × 6 mm), the samples were isostatically pressed at 100 MPa for higher green density. Finally, the green compacts were pressureless sintered at 1600 oC for 4h in Ar atmosphere. Apparent porosity and bulk density of the sintered samples were measured by Archimedes method in distilled water. Flexure strength was determined by a three point bending test (sample size: 3 mm × 4 mm × 25 mm, span: 16 mm, loading speed: 0.5 mm·min-1). Fracture toughness was determined by the single edge notched beam (SENB) method (sample size: 3 mm × 4 mm × 25 mm, span: 16 mm, gap width: 0.25

4

mm, gap depth: 2 mm, loading speed: 0.05 mm·min-1). Phase composition was analyzed by X-ray diffraction (XRD, X’Pert Pro) with Cu Kα radiation at a scanning rate of 8o min-1. Microstructure of the sintered samples in the fracture surface was observed by a scanning electron microscopy (SEM, FEI Quanta600). Dielectric property was tested by an impedance instrument with an effective frequency of 1 MHz (Hioki 3532). Friction and wear tests were carried out on a pin disc two-body friction and wear tester (ML 10) with a 240-grit SiC abrasive disk. The loading was 10 N. The worn surface was observed by a color 3D laser scanning microscope (KEYENCE, VK9700).

3. Results and discussion 3.1. Phase composition Fig. 3 displays the XRD patterns of the mullite/hBN composites with different compositions sintered at 1600 oC. From the patterns sintered without additives showed in Fig. 3a, it can be seen that in spite of both mullite and hBN phases are detected, there is still -Al2O3 phase residual in the composite, indicating the temperature for the full transformation of mullite sintered with SiO2 and Al2O3 powders could be above 1600 oC. It is in agreement with literatures reported that the synthesis temperature of mullite is about 1700 oC in the case of without sintering aids [1, 3, 11]. Fig. 3b-e display the XRD patterns of mullite/hBN composites added with Lu2O3 (MBLu), Er2O3 (MBEr), CeO2 (MBCe), La2O3 (MBLa), respectively. Obviously, ascribed to the ReO aids, the peaks of -Al2O3 are disappeared, and all the composites exhibit a better crystallinity in the mullite phase. This phenomenon can be attributed to the liquid phase sintering with the low melting point rare earth compounds, in which the mass transportation of the materials would be accelerated and ultimately results in lowering the transformation

5

temperature of mullite [15]. But unfortunately, the peaks corresponding to the rare earth compounds have not been detected in all the patterns due to their little contents. 3.2. Microstructure After polishing and etching by HF acid, the BSE micrographs of MBLa, MBCe, MBEr and MBLu were obtained, as shown in Fig. 4. It can be seen that, for all the samples, the glass phases distribute homogenously at the grain boundaries of mullite, and the mullite grains are well-crystallized and show a plate-like morphology with similar aspect ratio (about 2~3). However, the difference is that the grain size in MBLa is the biggest with the width over 10 μm (Fig. 4a) while that in MBLu is the smallest with the width of about 3~5 μm (Fig. 4d). This might be ascribed to the different growth velocity of the mullite grains in different liquid phases. It is well known that the formation of mullite can be divided into two steps: (1) the mutual diffusion between Al3+ and Si4+ in the low temperature, and (2) the solution–reprecipitation process by means of fluid flow and mass transfer in the high temperature [23, 24]. Hence, the viscosity of the liquid phase inevitably plays a vital role in the high temperature sintering of mullite. Compared to the low viscosity condition which prompts the diffusion and dissolution of chemical elements, the mullite grain growth would be inhibited by the high viscosity in liquid phase because of the high diffusion resistance as well as the strong restraint of dissolution. It has been reported that cationic field strength of the rare earth elements is increased with the decrease of cationic radius, which can increase the viscosity of ReO relevant liquid phases at grain boundary in the sintering process [25-27]. Therefore, the ReO with a smaller ionic radius can lead to a higher viscosity in liquid phase [28, 29]. Since the ionic radius of La3+, Ce4+, Er3+ and Lu3+ are 0.1061, 0.0924, 0.0881 and 0.0848 nm respectively, it can be suggested that MBLu

6

would present the highest viscosity in the sintering process among the four samples, and in that case, the smallest mullite grains should be obtained as a result. Further analyses were performed by EDS, which reveals that the rare earth elements are all uniformly distributed at the grain boundaries. The results can be attributed to that ReO could hardly form solid solution with the ceramic matrix due to their large cation radius compared with Al3+ and Si4+. In order to investigate the morphology and distribution of hBN in the composites, fracture surfaces of the samples are observed and shown in Fig. 5. From Fig. 5a-d, it is surprising that the hBN particles are all refined and distributed homogeneously in the composites with different ReO. Meanwhile, since the finest grains are presented in MBLu (Fig. 5d), the high magnification images of hBN particles are illustrated in Fig. 5e and f. It can be seen that hBN particles distribute both at the grain boundaries and in the grains the mullite matrix, with the width of 0.5~1 μm and the thickness of 50~100 nm, which is much smaller than the original o-BN particles. These refining phenomena of hBN might be concerned with the pre-oxidation treatment [20]. In our precious study, we found an oxygen-rich amorphous layer with the thickness of 5 nm formed on the surface of o-BN particle, which could improve its wettability with ceramic matrices significantly in the subsequent sintering process. However, its concrete refining mechanism is still unclear and needs to further investigate. 3.3. Relative density and mechanical properties Table 1 shows the relative density and mechanical properties of mullite/hBN composites sintered with different ReO. As can been seen, MBLu behaves the most excellent comprehensive performances. Its relative density, flexure strength, fracture toughness and hardness reach 93.7%, 222 MPa, 3.2 MPa·m1/2 and 6.02 GPa,

7

respectively. In fact, whether from the XRD patterns or relative density of the composites mentioned above, the four aids are all confirmed to be beneficial in promoting the sintering of mullite. But the mechanical properties of the samples vary greatly, especially when La2O3 was added, whose flexure strength and fracture toughness are only one-third of the sample added with Lu2O3. As mentioned, the liquid phase with different viscosity would lead to various mullite grain sizes. Generally, the grain growth rate of ceramics in the sintering process is controlled by the migration rate of grain boundary, and the viscosity of the liquid phase decreases with the cation radius of ReO. So it indicates that the ReO with a bigger cation radius might cause a higher grain boundary migration rate of the matrix. However, for the porous composites, pores would be hardly eliminated when the migration rate of grain boundary is faster than the pore migration rate. Thus, MBLa presents a lowest density while MBLu shows a highest one. It has been reported that the mechanical properties of ceramics vary linearly with the cation radius of ReO [25, 30]. Meanwhile, with the decrease of cation radius, the flexure strength and hardness increase while the fracture toughness decreases [25, 30]. However, our results are dissimilar. We found that not only the flexure strength and hardness, but also the fracture toughness increases with the ionic radius deceasing. Actually, the results are accordance with the analysis of the microstructures. As known, when the aspect ratio of grains has little changes in composites, the lower the porosity and the finer the grain size are, the more excellent mechanical properties will be. So MBLu exhibits the best comprehensive properties including flexure strength, fracture toughness and hardness. As a comparison, the mechanical properties of MBLa deteriorate drastically due to the abnormal grain growth and higher porosity.

8

Crack paths of the samples obtained by Vickers indentation method are showed in Fig. 6 to explain the toughening mechanism of the composites. For MBLa, the crack propagates along a quite straight direction with an extremely smooth flaw sections. However, with decreasing cation radius of ReO, the grains are refined and the grain boundaries are increased, which induce the emergence and promotion of crack deflection, especially for MBLu. The driving force of crack propagation is weakened dramatically when the deflection occurred, which indicates a large energy consumption in that process. Furthermore, due to more grain boundaries formed by the refined grains, the submicron hBN would have more tendency to distribute at the grain boundaries of the matrix, which promotes the crack bridging efficiently. As a result, the fracture toughness of the composites is remarkably improved as well. 3.4. Dielectric properties Fig. 7 shows the dielectric constant (  ) and dielectric loss ( tan  ) of mullite/hBN composites sintered with different ReO (Frequency: 1 MHz). It can be seen that the lowest dielectric constant and dielectric loss are obtained in MBLa and MBEr, with the value of 5.63 and 5.82 × 10-3, respectively. According to literatures [31-33],  and

tan  could be influenced not only by the porosity, but also the different phase composition. The relations can be calculated as follows: ln   (1  p)  ln  0

(1)

ln   X1  ln 1  X 2  ln  2

(2)

tan   tan  0  A  Pn

(3)

ln tan   X1  ln tan 1  X 2  ln tan  2

(4)

where  , tan  and  0 , tan  0 are the dielectric constant and dielectric loss of the

9

fabricated composites and fully densified materials,  1 , tan 1 and  2 tan  2 are the dielectric constant and dielectric loss of the different phases, p is the porosity, X1 and X2 are the volume fractions of the different phases, A and n are the constant. Hence, from the equations (1-4) and Fig. 7, it could be deduced that, compared with phase composition, porosity has a far greater influence on the dielectric constant, and the dielectric constant increases apparently with the porosity decrease. So with the highest porosity among the samples, the MBLa exhibits the lowest dielectric constant. However, the results are contrary to the dielectric loss, while the phase composition plays a more vital role. The higher dielectric loss of second phase added, the higher dielectric loss of the composites has [34, 35]. Therefore, since the lowest dielectric loss of the Er2O3 has, the lowest dielectric loss of MBEr presents [35]. 3.5. Tribological properties Fig. 8 presents the friction coefficient of the different mullite/hBN composites sliding against SiC. At the beginning of the tests, ascribe to the removal of surface contaminants by rubbing, the friction coefficient of all the samples appears abnormal fluctuation [36]. As the value stabilized, MBLu shows the lowest friction coefficient of 0.43, while MBEr, MBCe and MBLa exhibit the moderate coefficient of 0.57, 0.70 and 0.74, respectively. For the mullite/hBN composites with same composition, the MBLu sample with finer grains (Fig.3d) is liable to develop a relative smooth surface film in friction process and therefore reduce the friction coefficient of the composite, which is consistent with Gahr’s conclusion [37]. Fig. 9 shows the wear rate of the different mullite/hBN composites sliding against SiC. MBLu exhibits the lowest wear rate of 3.41 × 10-3 mm3/Nm, while MBCe shows the highest wear rate of 6.25 × 10-3 mm3/Nm. It can also be seen that the wear rate of

10

MBLa is slightly lower than MBCe, with the value of 5.92 × 10-3 mm3/Nm. From Table 1, compared with MBCe, MBLa possesses the lowest properties in almost all aspects except hardness. Therefore, the tribological results indicate that the wear rate should mainly relate to the hardness of the composites, that is, a harder material might result in a lower wear rate [36]. The morphologies and the 3D profile of the worn surface of the four composites sliding against SiC are shown in Fig. 10. Comparing with MBEr and MBLu, rougher surfaces can be observed in MBLa and MBCe with a lot of potholes on the worn surfaces, suggesting that they suffered a serious fracture and stripping during the friction process. Generally, a rougher surface may suffer a deeper cutting in friction process, which results in a higher friction coefficient. Therefore, the friction coefficient of MBLa and MBCe is higher than the others correspondingly. In addition, different from the small potholes in MBLa, the potholes dimension in MBCe is relatively larger and even reaches 100 μm. Actually, the heavy spall on the matrix surfaces can be ascribed to the synergistic effect of hardness and fracture toughness. Due to the weak interfacial strength of ceramics, brittle fracture is liable to occurred, and the dropped wear debris would also play a cutting role during the wear process. Therefore, MBCe with the lowest hardness is more affected, leading to a larger worn surface. On the other hand, a lower fracture toughness means a faster crack propagation velocity, which would accelerate wear both in longitudinal and transverse of the worn surface, in that case, amount of deep potholes are formed. As a contrast, MBEr shows a moderate worn surface, while due to the highest hardness and fracture toughness, MBLu presents the smoothest worn surface, with little shallow potholes be detected.

11

4. Conclusion In summary, mullite/hBN composites were successfully fabricated with different ReO additives (Er2O3, CeO2, La2O3, Lu2O3) by pressureless sintering at 1600oC. XRD analyses show that all the ReO additives are beneficial to the formation of mullite phase. SEM images indicate that the rare earth elements mainly exist at the boundaries of mullite grains. With the decrease of cationic radius of the ReO, the mullite grains are refined. The sample sintered with Lu2O3 exhibits the best mechanical performance thank to its lowest porosity and finest grain size, and its relative density, flexure strength, fracture toughness and hardness could reach 93.7%, 222 MPa, 3.2 MPa•m1/2 and 6.02 GPa, respectively. Owing to different porosity and composition, the lowest dielectric constant and dielectric loss are obtained in the composites sintered with La2O3 and Er2O3, with the value of 5.63 and 5.82 × 10-3, respectively. The tribological properties of the composites are relevant to hardness and fracture toughness, the sample sintered with Lu2O3 shows the best tribological property sliding against SiC, with the friction coefficient and wear rate of 0.43 and 3.41 × 10-3 mm3/Nm.

Acknowledgement This research was financially supported by the National Natural Science Foundation of China (No. 51302208), the Program for Young Excellent Talents in Shaanxi Province (2013KJXX-50), Open fund project of State Key Laboratory for Mechanical Behavior of Materials (20131306), the China Postdoctoral Science Foundation (2013M542342) and Natural Science Foundation of Education Department of Shaanxi Provincial Government (12JK0450).

12

References [1] S. Kanzaki, H. Tabata, T. Kumazawa, S. Ohta, Sintering and mechanical-properties of stoichiometric mullite, J. Am. Ceram. Soc. 68 (1985) C6-C7. [2] H. Schneider, J. Schreuer, B. Hildmann, Structure and properties of mullite - A review, J. Eur. Ceram. Soc. 28 (2008) 329-344. [3] I.A. Aksay, D.M. Dabbs, M. Sarikaya, Mullite for structural, electronic, and optical applications, J. Am. Ceram. Soc. 74 (1991) 2343-2358. [4] T. Kusunose, T. Sekino, Y.H. Choa, K. Niihara, Fabrication and microstructure of silicon nitride/boron nitride nanocomposites, J. Am. Ceram. Soc. 85 (2002) 2678-2688. [5] G.J. Zhang, J.F. Yang, M. Ando, T. Ohji, S. Kanzaki, Reactive synthesis of alumina-boron nitride composites, Acta Mater. 52 (2004) 1823-1835. [6] M. Du, J.Q. Bi, W.L. Wang, X.L. Sun, N.N. Long, Influence of sintering temperature on microstructure and properties of SiO2 ceramic incorporated with boron nitride nanotubes, Mater. Sci. Eng. A 543 (2012) 271-276. [7] Z.Q. Shi, J.P. Wang, G.J. Qiao, Z.H. Jin, Effects of weak boundary phases (WBP) on the microstructure and mechanical properties of pressureless sintered Al2O3/h-BN machinable composites, Mater. Sci. Eng. A 492 (2008) 29-34. [8] Z. Shi, J. Wang, G. Qiao, J. Yang, Z. Jin, Machinability, deformation, and cracks behavior of pressureless-sintered Al2O3/h-BN composites: role of weak boundary phases, J. Mater. Sci. 44 (2009) 1580-1587. [9] D. Lewis, R.P. Ingel, W.J. McDonough, R.W. Rice, Microstructure and thermomechanical

properties

in

alumina

and

mullite-boronnitride

particulate

ceramic-ceramic composites, Ceram. Eng. Sci. Proc. 2 (1981) 712-718. [10] A.K. Basu, A. Mitra, S. Ghatak, Synthesis of a powder precursor in the form of 13

hydroxyhydrogel for reaction sintering of BN–mullite composite, Ceram. Int. 32 (2006) 213-219. [11] G.J. Zhang, J.F. Yang, M. Ando, T. Ohji, S. Kanzaki, Mullite-boron nitride composite with high strength and low elasticity, J. Am. Ceram. Soc. 87 (2004) 296-298. [12] M. Du, J.Q. Bi, W.L. Wang, X.L. Sun, N.N. Long, Y.J. Bai, Fabrication and mechanical properties of SiO2–Al2O3–BNNPs and SiO2–Al2O3–BNNTs composites, Mater. Sci. Eng. A 530 (2011) 669-674. [13] T. Jiang, Z. Jin, J. Yang, G. Qiao, Investigation on the preparation and machinability of the B4C/BN nanocomposites by hot-pressing process, J. Mater. Process. Technol. 209 (2009) 561-571. [14] L. Gao, X.H. Jin, J.G. Li, Y.G. Li, J. Sun, BN/Si3N4 nanocomposite with high strength and good machinability, Mater. Sci. Eng. A 415 (2006) 145-148. [15] H. Liang, X. Yao, J. Zhang, X. Liu, Z. Huang, The effect of rare earth oxides on the pressureless liquid phase sintering of α-SiC, J. Eur. Ceram. Soc. 34 (2014) 2865-2874. [16] G.J. Zhang, H.F. Yang, M. Ando, T. Ohji, Reaction synthesis of aluminum nitride-boron nitride composites based on the nitridation of aluminum boride, J. Am. Ceram. Soc. 85 (2002) 2938-2944. [17] T. Kusunose, T. Sekino, Y. Ando, Synthesis of SiC/BN nanocomposite powders by carbothermal reduction and nitridation of borosilicate glass, and the properties of their sintered composites, Nanotechnology 19 (2008) 275603. [18] T. Kusunose, R.J. Sung, T. Sekino, S. Sakaguchi, K. Niihara, High-temperature properties of a silicon nitride/boron nitride nanocomposite, J. Mater. Res. 19 (2011) 1432-1438. [19] Y. Zhao, Y. Zhang, H. Gong, X. Wang, H. Sun, Effects of Y2O3–MgO nanopowders

14

content on mechanical and dielectric properties of porous BN/Si3N4 composites, Ceram. Int. 41 (2015) 3618-3623. [20] Y.F. Li, P. Liu, X.D. Wang, H.Y. Jin, J.F. Yang, G.J. Qiao, Interface structure and formation mechanism of BN/intergranular amorphous phase in pressureless sintered Si3N4/BN composites, Scr. Mater. 63 (2010) 185-188. [21] Y. Li, H. Yu, Z. Shi, H. Jin, G. Qiao, Z. Jin, Synthesis of β-SiAlON/h-BN nanocomposite by a precursor infiltration and pyrolysis (PIP) route, Mater. Lett. 139 (2015) 303-306. [22] H. Jin, Y. Li, X. Li, Z. Shi, H. Xia, Z. Xu, G. Qiao, Functionalization of hexagonal boron nitride in large scale by a low-temperature oxidation route, Mater. Lett. 175 (2016) 244-247. [23] M. Schmucker, W. Albers, H. Schneider, Mullite Formation by Reaction Sintering of Quartz and Alpha-Al2o3 - a Tem Study, J. Eur. Ceram. Soc. 14 (1994) 511-515. [24] B. Saruhan, W. Albers, H. Schneider, W.A. Kaysser, Reaction and sintering mechanisms of mullite in the systems cristobalite/α-Al2O3 and amorphous SiO2/α-Al2O3, J. Eur. Ceram. Soc. 16 (1996) 1075-1081. [25] Y. Zhou, K. Hirao, M. Toriyama, Y. Yamauchi, S. Kanzaki, Effects of intergranular phase chemistry on the microstructure and mechanical properties of silicon carbide ceramics densified with rare-earth oxide and alumina additions, J. Am. Ceram. Soc. 84 (2001) 1642-1644. [26] R. Ramesh, E. Nestor, M.J. Pomeroy, S. Hampshire, Formation of Ln-Si-Al-O-N glasses and their properties, J. Eur. Ceram. Soc. 17 (1997) 1933-1939. [27] F. Lofaj, R. Satet, M.J. Hoffmann, A.R. de Arellano López, Thermal expansion and glass transition temperature of the rare-earth doped oxynitride glasses, J. Eur. Ceram.

15

Soc. 24 (2004) 3377-3385. [28] P.F. Becher, M.K. Ferber, Temperature-dependent viscosity of SiREAl-based glasses as a function of N:O and RE:Al ratios (RE = La, Gd, Y, and Lu), J. Am. Ceram. Soc. 87 (2004) 1274-1279. [29] J. Yang, J.F. Yang, S.Y. Shan, J.Q. Gao, T. Ohji, Effect of sintering additives on microstructure and mechanical properties of porous silicon nitride ceramics, J. Am. Ceram. Soc. 89 (2006) 3843-3845. [30] J. Marchi, D.S. Morais, J. Schneider, J.C. Bressiani, A.H.A. Bressiani, Characterization of rare earth aluminosilicate glasses, J. Non-Cryst. Solids 351 (2005) 863-868. [31] Y.F. Xia, Y.P. Zeng, D.L. Jiang, Dielectric and mechanical properties of porous Si3N4 ceramics prepared via low temperature sintering, Ceram. Int. 35 (2009) 1699-1703. [32] X. Li, L. Zhang, X. Yin, Z. Yu, Mechanical and dielectric properties of porous Si3N4–SiC(BN) ceramic, J. Alloy. Compd. 490 (2010) L40-L43. [33] S.J. Penn, N.M. Alford, A. Templeton, X.R. Wang, M.S. Xu, M. Reece, K. Schrapel, Effect of porosity and grain size on the microwave dielectric properties of sintered alumina, J. Am. Ceram. Soc. 80 (1997) 1885-1888. [34] D. Xue, K. Betzler, H. Hesse, Dielectric constants of binary rare-earth compounds, J. Phys.: Condens. Matter 12 (2000) 3113-3118. [35] S.Y. Cho, K. In-Tae, H. Kug Sun, Microwave dielectric properties and applications of rare-earth aluminates, J. Mater. Res. 14 (1999) 114-119. [36] Y. Zhu, A. Zhou, Y. Ji, J. Jia, L. Wang, B. Wu, Q. Zan, Tribological properties of Ti3SiC2 coupled with different counterfaces, Ceram. Int. 41 (2015) 6950-6955.

16

[37] K.H.Z. Gahr, W. Bundschuh, B. Zimmerlin, Effect of grain-size on friction and sliding wear of oxide ceramics, Wear 162 (1993) 269-279.

Fig. 1 SEM images and XPS spectra of (a) r-BN and (b) o-BN Fig. 2 HRTEM image of o-BN Fig. 3 XRD patterns of (a) mullite/hBN composites (b) MBLu; (c) MBEr; (d) MBCe; (e) MBLa Fig. 4 BSE micrographs and EDS analysis of (a) MBLa; (b) MBCe; (c) MBEr; (d) MBLu Fig. 5 Fracture surfaces of (a) MBLa; (b) MBCe; (c) MBEr; (d-f) MBLu with different magnification Fig. 6 SEM micrographs of crack paths by Vickers indentation of (a) MBLa; (b) MBCe; (c) MBEr; (d) MBLu Fig. 7 Dielectric constant and dielectric loss of mullite/hBN composites sintered with different aids Fig. 8 Friction coefficient of mullite/hBN composites sintered with different aids sliding against SiC Fig. 9 Wear rate of mullite/hBN composites sintered with different aids sliding against SiC Fig. 10 Morphologies and 3D profile of the worn surface of (a, e) MBLa; (b, f) MBCe; (c, g) MBEr; (d, h) MBLu

17

Table 1 Relative density and mechanical properties of mullite/hBN composites sintered with different ReO aids

Aids

Apparent Relative Flexure strength Fracture toughness porosity (%) density (%) (MPa) (MPa·m1/2)

Hardness (GPa)

La2O3

14.2

85.5

75±11

0.9±0.1

4.33±0.3

CeO2

11.1

88.6

130±30

1.7±0.3

3.28±0.4

Er2O3

6.9

89.6

195±25

2.8±0.2

4.35±0.1

Lu2O3

4.6

93.7

222±19

3.2±0.4

6.02±0.2

18

19

20

21

22