Diamond & Related Materials 29 (2012) 84–88
Contents lists available at SciVerse ScienceDirect
Diamond & Related Materials journal homepage: www.elsevier.com/locate/diamond
Sintering behaviors of fine-grained cBN–10 wt.% Al3.21Si0.47 system under high pressure Lili Zhang, Zili Kou ⁎, Chao Xu, Kaixue Wang, Chengliang Liu, Bo Hui, Duanwei He Institute of Atomic and Molecular Physics, Sichuan University, Chengdu 610065, China
a r t i c l e
i n f o
Article history: Received 19 April 2012 Received in revised form 10 August 2012 Accepted 14 August 2012 Available online 20 August 2012 Keywords: PcBN Fine-grained Sintering behavior High pressure and high temperature
a b s t r a c t Sintering behaviors of fine-grained cBN–10 wt.% Al3.21Si0.47 system under high pressure of 5.0 GPa and temperatures up to 1600 °C were investigated. By analyzing the phase components of sintered samples through x-ray diffraction (XRD), and detecting the in-situ electrical resistance signal during high pressure and high temperature (HPHT) sintering, we suggested the reactions between cBN and Al3.21Si0.47 happened at about 600 °C. Between 600 °C and 900 °C, AlN, AlB2 and Si formed, while above 1000 °C, the content of AlN, AlB12 and Si phases increased with AlB2 disappearing. Scanning electron microscope (SEM), Vickers hardness tests as well as cutting performance tests showed that the well-sintered samples in cBN–10 wt.% Al3.21Si0.47 with homogeneous microstructure and best mechanical performance could be obtained at the sintering P–T condition of 5.0 GPa, 1400 °C for 10–15 min. Crown Copyright © 2012 Published by Elsevier B.V. All rights reserved.
1. Introduction Cubic boron nitride (cBN) has superior properties, such as high hardness, excellent thermal conductivity and high thermal stability. In addition, it does not react with ferrous materials as a result of its high chemical inertness. Because of those unique properties, cBN is widely used for cutting hardened steel, cast iron, ferrous powder metal and heat resisting alloy [1–3]. The cBN single crystal is difficult to grow up and it is extremely brittle due to anisotropy, while the properties of polycrystalline cubic boron nitride (PcBN) are relatively homogeneous and the cost is low, showing excellent mechanical properties with high hardness and wear resistance suitable for cutting tools [4–6]. As a result, PcBN is widely applied in industrial production. The influence of the grain size of the cBN on the performance of PcBN tool is very important [6–8]. Amanda Mckie shows that the hardness of the PcBN materials increases with decreasing cBN grain size [9]. Kevin Chou and Chris J. Evans indicate that the greater wear resistance and lower wear rates are acquired for PcBN tools with the smaller grain size [10]. According to Hall–Petch relationship [11], the material yield stress and hardness are inversely proportional to the square root of the grain size. So within a certain range, the smaller the grain size, the greater the yield stress and hardness of the material. Grain refinement can improve the material toughness and ductility, which make up the deficiency of the material with high hardness and poor toughness. What's more, the smaller the grain size of cBN, the better is the surface quality of the sintered bodies as well as the impact resistance ⁎ Corresponding author. Tel./fax: +86 28 8540 1306. E-mail address:
[email protected] (Z. Kou).
and wear resistance, which are conducive to machine the high surface quality of the workpiece [8,10]. So, the cBN grain size of 1–2 μm is used as the starting material in this study. Fine-grained sintered bodies of cBN can be produced at high pressure and high temperature (HPHT) by the direct transformation from the hexagonal boron nitride (hBN) or sintering of fine-grained cBN grits with or without additives. The former method involves a complicated transformation reaction from hBN to cBN. Moreover, this method requires higher pressure (>7 GPa) and higher temperature (>2000 °C) which make the cost increase for the industrial application [12,13]. The latter method without additives requires the relatively high sintering conditions [14]. Hence in this investigation, sintering of the finegrained cBN grits with additives is adopted. Conventional binders used for cBN sintering have mainly been metals of the groups IV, V and VI of the periodic table or their compounds. Metallic elements like Ni, Co, Ti, Al and materials which have the catalytic activity for hBN to cBN conversion such as AlN and Si are used as well [15–18]. Adding the appropriate additives to the cBN grits decreases the sintering conditions, including the pressure and temperature [19]. In this study, Al3.21Si0.47 is used as the binder in the sintering process, since it has the low melting point and can react with cBN under a wide temperature range, making cBN particles that are bonded by the reaction production to be denser and the wear performance of the sintered bodies improved. However, little information about the fine-grained cBN with Al3.21Si0.47 has been reported at present yet. Thus the aim of this study is to extend our current understanding on the sintering behaviors of fine grained cBN with Al3.21Si0.47 alloy as well as the properties of the sintered products. This paper studied the reaction mechanisms and related the chemical as well as mechanical properties of the cBN composites. In order to illustrate the reaction mechanisms, x-ray diffraction (XRD)
0925-9635/$ – see front matter. Crown Copyright © 2012 Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.diamond.2012.08.001
L. Zhang et al. / Diamond & Related Materials 29 (2012) 84–88
analysis, combining in-situ electrical resistance measurements at different P–T conditions was conducted. Scanning electron microscope (SEM) with energy dispersive spectrometer (EDS) analysis, Vickers hardness and cutting performance tests was also applied to study the sintered samples which were prepared under various conditions. 2. Experimental procedures 2.1. Sample preparation The cBN powder (grain size: 1–2 μm, Zhengzhouzhongnanjiete Co., Ltd, China) and Al3.21Si0.47 powder (grain size: 80 nm, purity: 99.9%, Shanghai Chao Wei Nanotechnology Co., Ltd, China) as the starting materials were mixed according to the mass ratio of 9:1. Powders were manually mixed in ethyl alcohol using an agate mortar and pestled for 4 h. In order to remove the ethanol and vapors adsorbed on the powder surface, the mixtures were treated in vacuum of 3.0 × 10 −3 Pa and temperature of 500 °C for 1 h. Afterwards, about 1 g mixture powders were put into zirconium capsule (11 mm in diameter, 5 mm in thickness) and sintered at 5.0 GPa, high temperature at 500 °C–1600 °C for different heat treatment times of 5–20 min, using DS6 × 8MN cubic press [20]. The sample assembly for the HPHT sintering experiment can be seen elsewhere (see Fig. 1 in [17]). The cell temperature was directly measured with Pt6%Rh–Pt30%Rh thermocouples, and pressure was calibrated with a method of aluminum melting-point [21]. In our experiments, samples were first compressed to a required pressure and then heated to the desired temperature with a heating rate of about 200 °C/min. Because the heating rate was relatively great, there was a certain gradient of temperature in the sample chamber in our experiments. However, the temperature difference between the center part of the sample and the edge part was estimated to be less than 5 °C. After keeping for a desired treatment time, the samples were quenched to room temperature with cooling rate of about 70 °C/min and then decompressed to ambient pressure with 12 min. The well shaped specimens were ground into a wafer of about 9.5 mm in diameter and 4 mm in thickness. Then the samples were polished to a smooth mirror surface by a polishing machine with 10 μm and 1.5 μm diamond pastes.
In order to determine the mechanical properties of the sintered samples, Vickers hardness and cutting performance tests were designed to measure. Vickers hardness of the polished samples was tested by a Vickers hardness tester (FV-700B, Future-Tech, Japan) with 29.4 N of applied load and 15 s dwelling time. Moreover, the sintered bodies were polished to a cutting-tool shape with 0.2 mm × 20° chamfer for cutting tests so as to evaluate the cutting performance. The tests were conducted on a numerically controlled lathe (SK50P/750, Baoji, China). Tool holders were CRSNR2525M09. The tool wear was gauged by a stereomicroscope (XTL-3400, Shanghai, China). The hardened steel with the hardness at about 62 HRC was used as the working piece. The testing parameters were as follows: 120 m/min of cutting speed, depth of cut of 0.15 mm, feed of 0.15 mm/rev. No lubricant or coolant was used during turning. 3. Results and discussion 3.1. Investigate sintering mechanisms by XRD and in-situ electrical resistance measurement According to the XRD patterns of the samples sintered at 5.0 GPa, and different temperatures, with heating times of 15 min in Fig. 1, the sintered bodies are composed of different phases at different temperatures. At 5.0 GPa, 500 °C for 15 min, the major phases detected by XRD were cBN and Al3.21Si0.47 [Fig. 1(b)]. When temperature increased to 600 °C, phase components changed to be cBN, AlN, AlB2, Si and Al3.21Si0.47. When the temperature was set from 700 °C to 900 °C, the major phases of the samples were still cBN, AlN, AlB2, Si and Al3.21Si0.47, but the amount of Al3.21Si0.47 decreased and AlN, AlB2 and Si increased [Fig. 1(c)–(f)]. When we continue increasing the temperature from 1000 °C to 1600 °C, AlB2 and Al3.21Si0.47 disappeared, and cBN, AlN and Si were the final phases [Fig. 1(g)–(i)]. We deduce the possible reaction mechanisms between Al3.21Si0.47 and cBN according to XRD data. At 500 °C, the major components of the samples detected by the XRD were cBN and Al3.21Si0.47 which revealed Al3.21Si0.47 and cBN did not react at such P–T condition. Until temperature increased up to 600 °C, reactions happened and AlN, AlB2 and Si formed. We speculate the possible chemical reaction at 600 °C and temperatures up to 900 °C may proceed as follows:
2.2. Sample characterization The phase composition of the sintered samples was investigated by the XRD analysis (DX-2500, Dandong, China) and in-situ electrical resistance measurements at different P–T conditions were conducted to monitor the HPHT sintering behavior. SEM (S-4800, Hitachi, Japan) equipped with EDS (IE250, Oxford, England) was carried out to analyze the microstructure. In in-situ electrical resistance measurements, the sample assembly we used can be seen elsewhere [see Fig. 5 in [19]. A constant current of 10 mA was provided by a constant current source in the measurement circuit. If the electrical resistance of the sample changed, then the voltage drop of the sample would also change. Thus by detecting the voltage signal of the sample, we could get the in-situ electrical resistance changes of our samples during HPHT sintering. The cBN–10 wt.% Al3.21Si0.47 mixture was used as the starting materials. Because of the non-conduction of the starting materials at room pressure and temperature, a constant resistance of 10 Ω was connected in parallel with the sample to show the tendency of the sample resistance changes obviously. Heating power with a constant increasing rate was controlled by a programmed power controller. The temperature was measured by a Pt6%Rh– Pt30%Rh thermocouple which was put into the sample zone. Thus the temperature and voltage signal across the sample could be monitored simultaneously and separately recorded by a multi-channel data recorder, and we could get the relations between resistance changes of the sample and temperature.
85
2BNðsÞ þ
3 3 Al3:21 Si0:47 ðlÞ ¼ 2AlNðsÞ þ AlB2 ðsÞ þ 0:47 SiðsÞ 3:21 3:21 ð1Þ
Fig. 1. X-ray diffraction patterns of starting materials (a) and the sintered samples from the cBN–10 wt.% Al3.21Si0.47 mixture at: (b) 500 °C; (c) 600 °C; (d) 700 °C; (e) 800 °C; (f) 900 °C; (g) 1000 °C; (h) 1400 °C and (i) 1600 °C under 5.0 GPa for15 min.
86
L. Zhang et al. / Diamond & Related Materials 29 (2012) 84–88
We reason that when temperature was 600 °C–900 °C, Al3.21Si0.47 alloy melt first and filled the interstices between cBN grains, and reactions between cBN and Al3.21Si0.47 occurred next. From the reaction equation started above, AlN and AlB2 existed simultaneously with 2:1 mole ratio, which consisted of the XRD data. As shown in Fig. 1, the strongest peaks of AlN from samples prepared at 600 °C–900 °C are much higher than that of AlB2 and Si. And because of low temperature and limited maintaining time of heating, Al3.21Si0.47 was not consumed completely, and remained in the quenched samples. At higher temperature above 1000 °C, the reaction between Al3.21Si0.47 and cBN conducted very quickly, and in the quenched samples, only cBN, AlN and Si left. It is likely that these phases form by reaction: 2BNðsÞ þ
3 3 Al Si ðlÞ ¼ 2AlNðsÞ þ AlB2 ðsÞ þ 0:47 SiðsÞ ð1Þ 3:21 3:21 0:47 3:21
6AlB2 ðsÞ ¼ 5AlðlÞ þ AlB12 ðsÞ
ð2Þ
13AlðlÞ þ 12BNðsÞ ¼ 12AlNðsÞ þ AlB12 ðsÞ
ð3Þ
We believe that the system may undergo Eq. (1) when temperature is below 1000 °C. In the meantime, according to the phase diagram of AlB2 and the papers reported before [22,23], as the temperature is higher than 1000 °C, AlB2 may no longer be stable, and it decomposes into Al and AlB12, as we describe in reaction Eq. (2). As reported, at high temperature above 1000 °C, Al can easily react with cBN, and AlN and AlB12 will be produced [24]. As temperature continues to increase, the high temperatures might make the reaction between Al3.21Si0.47 and cBN toward AlN, AlB12 and Si thermodynamically preferred. In the meantime, it obviously improved the reaction rate. In our experiment, when temperature was higher than 1000 °C, both AlB2 and Al3.21Si0.47 disappeared, and cBN, AlN, and Si were the only phases in the XRD data. Thus actually we can describe the reaction of our system at 5.0 GPa and temperatures above 1000 °C as follows: 12BNðsÞ þ
13 13 Al3:21 Si0:47 ðlÞ ¼ 12AlNðsÞ þ AlB12 ðsÞ þ 0:47 SiðsÞ 3:21 3:21 ð4Þ
In Eq. (4), it was clear that when 12 mol of BN was consumed in the reaction, 12 mol of AlN would be formed, while only 1 mol of AlB12. Thus in the XRD experiment, we can clearly detect the existence of AlN, but because of the very small quantity, detecting AlB12 would be difficult, and the same results have been reported before [25]. We also treated the pure Al3.21Si0.47 from 600 °C to 1600 °C with heat treatment times of 15 min under 5.0 GPa. The XRD results of quenched samples are shown in the Fig. 2. As demonstrated, except a little
Fig. 2. X-ray diffraction patterns of the sintered samples from the pure Al3.21Si0.47 at: (a) 600 °C; (b) 1000 °C; (c) 1600 °C under 5.0 GPa for 15 min.
oxidization of Al3.21Si0.47, there is no change in phase components of the samples at different temperatures. Because of high temperature treatment, Al3.21Si0.47 would be easy to oxidize and with only a small amount of Al2O3 being detected. According to the chemical formula of Al3.21Si0.47, the mole ratio of Al and Si is larger than 6:1, thus in the treated Al3.21Si0.47 samples, Si or the oxidation of Si would be difficult to be detected by XRD. In our sintering experiment, Al2O3 would be more difficult to detect by XRD, because Al3.21Si0.47 was only 10 wt.% of the mixtures. Hence, we believe that when the reaction between Al3.21Si0.47 and cBN took place, Al3.21Si0.47 did not decompose. On the contrary, because of Al3.21Si0.47's chemical activity, it directly reacted with cBN at a low temperature. Fig. 3 shows the variation of the resistance with temperature under 5.0 GPa. Two abrupt changes in resistance are shown in Fig. 3, at about 200 °C (point A) and 1000 °C (point B), respectively. At about 200 °C, the non-conductive sample changed to be conductive, because the conductivity of the sample was improved at temperature above 200 °C. According to XRD results, no reactions of cBN with Al3.21Si0.47 occurred at such conditions. Meanwhile, cBN doesn't change. Thus we guess that at about 200 °C under 5.0 GPa, the sample may have a temperature induced transformation in which the conductivity of the sample would be improved, and this may be most likely because of the transformation of Al3.21Si0.47. In order to illustrate the speculations, in-situ electrical resistance measurements were done with pure Al3.21Si0.47 alloy as the starting material. A series of experiments at the same condition confirmed the results shown in Fig. 4 were reproducible. In Fig. 4, electric resistance of pure Al3.21Si0.47 diminished quickly at about 200 °C, which also occurred when the cBN–10 wt.% Al3.21Si0.47 mixture was heated at about 200 °C. Thus we believe that the first abrupt change of resistance was due to the temperature induced transformation of the Al3.21Si0.47 alloy. But it is difficult to determine the specific reason of the transformation at present. Further studies about the transformation of the Al3.21Si0.47 alloy will be done in the days ahead. At about 1000 °C, the second abrupt change happened, and the sample changed from conductive to non-conductive as shown in Fig. 3. According to XRD results, cBN began to react with Al3.21Si0.47 at about 600 °C, and the reaction rate greatly increased when temperatures increased above 1000 °C with reaction product of AlN, AlB12 and Si. Because of the rapid production of a large amount of AlN, AlB12 and Si above 1000 °C, the sample became electrically insulated, and the resistance increased rapidly as shown in Fig. 3. 3.2. Microstructure and mechanical property tests 3.2.1. Microstructure characterization Fig. 5 (1) shows the back scanning electron microscopy (BSEM) of the polished surface of the samples sintered at 1400 °C for 15 min
Fig. 3. Plot of resistance versus temperatures with cBN–10 wt.% Al3.21Si0.47 mixture under 5.0 GPa.
L. Zhang et al. / Diamond & Related Materials 29 (2012) 84–88
87
[Fig. 5(b)] which should correspond to the reaction products of AlN, AlB12 and Si. And the white areas are the rich zones of the elements of silicon, aluminum, and oxygen [Fig. 5(c)]. We think such areas contain Si and Al2O3. The formation of Al2O3 was due to the oxidation of Al3.21Si0.47 at high temperature. Owing to the very little amount of Al2O3, XRD did not detect it, which was in accordance with our previous discussion. From Fig. 5 (2), we can see that the distribution of cBN in the samples is relatively homogeneous. So we think the samples were well-sintered.
Fig. 4. Plot of resistance versus temperatures with pure Al3.21Si0.47 under 5.0 GPa.
under 5.0 GPa. Combining XRD and BSEM with EDS, we believe the dark areas are the rich zones of elements of boron and nitrogen [Fig. 5(a)] corresponding to the existence of cBN. The gray areas are rich in the elements of aluminum, nitrogen, boron and silicon
3.2.2. Vickers hardness test Fig. 6 shows the relationship of the samples' Vickers hardness against both sintering temperature and heating durations. Five Vickers indentations were made on each PcBN sample, and the final hardness value was taken according to their average. As can be seen from Fig. 6, at a fixed temperature, the hardness of the sintered bodies increased first and achieved the highest hardness at around 36 GPa when the heat treatment times grew to about 15 min, and finally decreased with continuing heat treatment times. With a fixed heating duration of 15 min, the hardness of samples increased from 33 GPa to about 36 GPa when the temperature increased from 1300 °C to 1500 °C. In the meantime, we could find that below 1400 °C, temperature might have greater influence on the hardness of the samples
Fig. 5. Microstructure of the samples sintered at 1400 °C for 15 min under 5.0 GPa.
88
L. Zhang et al. / Diamond & Related Materials 29 (2012) 84–88
Fig. 6. Vickers hardness of cBN–10 wt.% Al3.21Si0.47 samples versus sintering temperatures and heating durations under 5.0 GPa.
than that when temperatures were from 1400 °C to 1500 °C, and the hardness of the samples was nearly the same when temperatures were from 1400 °C to 1500 °C. We guess that this result is attributed to inhomogeneous and incompact microstructure of the samples sintered at lower temperature such as 1300 °C, which causes the lower hardness. Thus according to hardness tests, we find that with the starting material of cBN–10 wt.% Al3.21Si0.47 at 5.0 GPa, to sinter a sample with high hardness, temperatures of above 1400 °C and heating duration of about 15 min are needed. 3.2.3. Cutting test Cutting tests of the sintered samples were conducted to investigate the tool life. In our experiment, tool wear (error of the tool wear measurement b1/1000 mm) was measured as the width of the wear land on the flank surface (Vb), and the relationship between tool wear and cutting length as shown in Fig. 7 could be used to evaluate the mechanical performance of the sintered samples. For each sintered sample, five passes were conducted, with wear land of each pass being measured. As shown in Fig. 7, the sample sintered at 1400 °C for 10 min under 5.0 GPa had the best cutting performance. Only at the fourth pass, the sample sintered at 1200 °C for 10 min shows the better wear result than the sample sintered at 1400 °C for 10 min. At other passes, the product sintered at 1400 °C for 10 min shows the better wear results. In general, the sample sintered at 1400 °C for 10 min shows the best wear results. The width of the wear land on the flank face was only 0.28 mm at 2100 m cutting length. Thus we suggest that to achieve the best mechanical performance, cBN–10 wt.% Al3.21Si0.47 system needs a favorable sintering P–T condition of 5.0 GPa, 1400 °C for around 10 min. 4. Conclusion In summary, the reactions between cBN and Al3.21Si0.47 happened at about 600 °C under 5.0 GPa. Between 600 °C and 900 °C, AlN, AlB2 and Si formed. Above 1000 °C, AlB2 was not stable. It decomposed and produced AlB12. At the same time, the content of AlN, AlB12 and Si phases increased. The results of SEM, Vickers hardness tests as well as cutting performance tests showed that the well-sintered samples in cBN–10 wt.% Al3.21Si0.47 with homogeneous microstructure and
Fig. 7. Evolution of flank wear of the samples as functions of cutting length. Sintering conditions: (a) 1400 °C 10 min; (b) 1200 °C 10 min; (c) 1400 °C 15 min; (d) 1200 °C 15 min; (e) 1300 °C 15 min under 5.0 GPa.
best mechanical performance could be obtained at about 1400 °C in processing times of 10–15 min under 5.0 GPa. Acknowledgements This work was funded by the sponsor-id="gs1">Sichuan University and partially supported by the National Natural Science Foundation of China (Grant No. 51072123). References [1] A. Braghini Jr., R.T. Coelho, Int. J. Adv. Manuf. Technol. 17 (2001) 244–257. [2] K.S. Neo, M. Rahman, X.P. Lia, H.H. Khoo, M. Sawa, Y. Maeda, J. Mater. Process. Technol. 140 (2003) 326–331. [3] Y. Sahin, J. Mater. Process. Technol. 209 (2009) 3478–3489. [4] T. Taniguchi, S. Yamaoka, Rev. High Press. Sci. Technol. 7 (1998) 980–982. [5] T. Taniguchi, S. Yamaoka, J. Cryst Growth 222 (2001) 549–557. [6] T.K. Harris, E.J. Brookes, C.J. Taylor, Int. J. Refract. Met. Hard Mater. 22 (2004) 105–110. [7] Z. Lv, J. Feng, F. Lin, X. Xu, Mater Sci. Forum 532 (533) (2006) 41–44. [8] K. Fujisaki, H. Yokota, N. Furushiro, Y. Yamagata, T. Taniguchi, R. Himeno, A. Makinouchi, T. Higuchi, J. Mater. Process. Technol. 209 (2009) 5646–5652. [9] A. McKie, J. Winzer, I. Sigalas, M. Herrmann, L. Weiler, J. Rodel, N. Can, Ceram. Int. 37 (2011) 1–8. [10] Y. Kevin Chou, C.J. Evans, Wear 212 (1997) 59–65. [11] A. Lasalmonie, J.L. Strudel, J. Mater. Sci. 21 (1986) 1837–1852. [12] H. Sumiya, S. Uesaka, S. Satoh, J. Mater. Sci. 35 (2000) 1181–1186. [13] H. Sumiya, Adv. Sci. Technol. 45 (2006) 885–892. [14] T. Taniguchi, M. Akaishi, S. Yama, J. Mater. Res. 14 (1999) 162–169. [15] E. Benko, P. Klimczyk, S. Mackiewicz, T.L. Barr, E. Piskorska, Diam. Relat. Mater. 13 (2004) 521–525. [16] B.K. Agarwala, B.P. Singh, S.K. Singhal, J. Mater. Sci. 21 (1986) 1765–1768. [17] R. Lv, J. Liu, Y.J. Li, S.C. Li, Z.L. Kou, D.W. He, Diam. Relat. Mater. 17 (2008) 2062–2066. [18] D.W. He, M. Akaishi, T. Tanaka, Diam. Relat. Mater. 10 (2001) 1465–1469. [19] Y.J. Li, S.C. Li, R. Lv, J.Q. Qin, J. Zhang, J.H. Wang, F.L. Wang, Z.L. Kou, D.W. He, J. Mater. Res. 23 (2008) 2366–2372. [20] L.M. Fang, D.W. He, C. Chen, L.Y. Ding, X.J. Luo, High Press. Res. 27 (2007) 367–374. [21] S.M. Wang, D.W. He, W.D. Wang, L. Lei, High Press. Res. 29 (2009) 806–814. [22] X.M. Wang, J. Alloys Compd. 403 (2005) 283–287. [23] C. Rizzoli, P.S. Salamakha, O.L. Sologub, G. Bocelli, J. Alloys Compd. 43 (2002) 135–141. [24] H.S.L. Sithebe, D. McLachlan, I. Sigalas, M. Herrmann, Ceram. Int. 34 (2008) 1367–1371. [25] Xiao-Zheng Rong, T. Yano, J. Mater. Sci. 39 (2004) 4705–4710.