silicon carbide composites by polymer infiltration and pyrolysis process

Composites Science and Technology 72 (2012) 461–466

Contents lists available at SciVerse ScienceDirect

Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech

Fabrication of multi-walled carbon nanotube-reinforced carbon fiber/silicon carbide composites by polymer infiltration and pyrolysis process Hai-zhe Wang ⇑, Xiao-dong Li, Jun Ma, Gong-yi Li, Tian-jiao Hu College of Science, National University of Defense Technology, Changsha 410073, China

a r t i c l e

i n f o

Article history: Received 12 August 2011 Received in revised form 5 December 2011 Accepted 11 December 2011 Available online 17 December 2011 Keywords: A. Carbon nanotubes A. Ceramics A. Ceramic–matrix composites (CMCs) B. Mechanical properties

a b s t r a c t Multi-walled carbon nanotube (MWNT)-reinforced carbon fiber/silicon carbide (Cf/SiC) composites were prepared using a polymer infiltration and pyrolysis (PIP) process. The MWNTs used in this study were modified using a chemical treatment. The MWNTs were found to be well dispersed in the matrix after ultrasonic dispersion, and the mechanical properties of the Cf/SiC composite were significantly improved by the addition of MWNTs. The addition of 1.5 wt.% of MWNTs to the Cf/SiC composite led to a 29.7% increase in the flexural strength, and a 27.9% increase in the fracture toughness. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Cf/SiC composites—which exhibit excellent properties in terms of their low density, high strength, high fracture toughness, and oxidation resistance—are considered as desirable candidate materials for high-tech applications. The automotive, aerospace, and aviation industries are particularly concerned with the application of Cf/SiC composites [1,2]. It is well known that the mechanical properties of Cf/SiC composites are generally determined by three factors: (1) the properties of the fibers, (2) the interface between the fibers and the matrices, and (3) the microstructure of the matrices [3,4]. The addition of reinforcing agents with unique mechanical properties (and consistency with the matrices) is therefore expected to play a positive role in improving the mechanical properties of Cf/SiC composites. In recent work, nanoscale additives were introduced into Cf/SiC composites to make them ‘‘stronger’’; carbon nanotubes (CNTs) were considered as one of the most outstanding candidates [5]. CNTs have attracted much attention because of their excellent mechanical [6–8] and electronic properties [9]. The incorporation of CNTs can improve the mechanical properties of materials, and CNTs have been extensively applied in fabricating polymer/CNT [10–12] and metal/CNT [13–15] composites. However, the number of studies and the achievements in the ceramic/CNT composite field [16–18] have been less significant than those in the

⇑ Corresponding author. Fax: +86 84574786. E-mail address: [email protected] (H.-z. Wang). 0266-3538/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2011.12.009

polymer/CNT and metal/CNT fields. The matrices used in CNT-reinforced ceramic composites in previous studies were primarily Al2O3 [5,13,19,20], and SiC [21]. In the preparation of ceramic/CNT composites, several key issues must be considered [19]. Firstly, the CNTs should be dispersed uniformly in the matrices. Both single- and multi-walled CNTs can easily form twisted aggregate structures, because of the high aspect ratio of CNTs, and the strong Van der Waals forces among the tubes. As a result, it is difficult to disperse CNTs homogeneously in the matrices. The second problem is the interphase between the MWNTs and the matrices. Only an appropriate interface will lead to an enhanced stress transfer capability from the matrices to the CNTs [21]. In this study, MWNT-reinforced Cf/SiC composites were prepared using a PIP [22–24] method, with antimony-substituted polymethysilane (A-PMS) used as a precursor. The mechanical properties and microstructures of these composites were investigated.

2. Experimental procedures The Cf/SiC/MWNT composites were prepared using a PIP process; one cycle of this process had four steps. Firstly, pretreated MWNTs were added to the A-PMS (the MWNT:A-PMS weight ratios were 0%, 0.5% and 1.5% for the different samples), with subsequent agitation and ultrasonic dispersion performed at 800 W for 60 min at room temperature. The chemical modification of the MWNTs was performed as follows: MWNTs with a diameter of 3080 nm and a length of 20100 lm were prepared in our lab by chemical vapor deposition (CVD). These MWNTs were treated

462

H.-z. Wang et al. / Composites Science and Technology 72 (2012) 461–466

(a)

(b)

500nm

200nm Fig. 1. Cross-sectional SEM images of cured A-PMS/1.5 wt.% MWNT composite.

-SiC

(a)

(b)

(a)

4000

3000

2000

1000

(b)

Wavenumbers (cm-1) Fig. 2. FT-IR spectra for (a) A-PMS after ultrasonic treatment, and (b) A-PMS. The strong peaks at 2106-2166 cm1 correspond to the Si–H bonds’ stretching vibration.

(c)

10

20

30

40

50

60

70

80

90

2 (°) Fig. 5. XRD patterns for the ceramics derived from (a) A-PMS, (b) A-PMS/0.5 wt.% MWNTs, and (c) A-PMS/1.5 wt.% MWNTs.

(a) 2.10 (b)

4000

3000

1000

2000

Wavenumbers (cm-1) Fig. 3. FT-IR spectra for (a) A-PMS, (b) cured A-PMS, and (c) cured A-PMS/1.5 wt.% MWNT composite. The strong peaks at 2106–2166 cm1 correspond to the Si-H bonds’ stretching vibration.

Density (g•cm-3)

1.90 (c)

1.70 1.50 1.5wt.%MWNTv/A-PMS 0.5wt.%MWNTv/A-PMS A-PMS

1.30 1.10 0

2

4

6

8

10

Cycles

105

Fig. 6. Density–cycle curves for the composites prepared by PIP technology using A-PMS, A-PMS/0.5 wt.% MWNTs, and A-PMS/1.5 wt.% MWNTs.

95

(b) (a)

85

75

0

20 0

400

600

800

1000

1200

1400

Temperature (°C) Fig. 4. TG curves for (a) cured A-PMS and (b) cured A-PMS/MWNT composite after ultrasonic treatment.

in concentrated HNO3 for 8 h at room temperature; they were then reacted with acryloyl chloride at 66 °C for 12 h. After this chemical modification process, it was found that the length of the MWNTs had decreased to about 510 lm, and the acryloyl chloride had polymerized on the surface of MWNTs and had encapsulated them (this work has been submitted for publication in separate manuscripts). Secondly, 3D carbon fiber preforms (Jilin Carbon, China, braided in Nanjing, China) were infiltrated with the A-PMS/ MWNTs, under vacuum. Afterwards, the samples were cured at 470 °C for 1 h in flowing N2. Thirdly, the cured samples were

H.-z. Wang et al. / Composites Science and Technology 72 (2012) 461–466

infiltrated with A-PMS/MWNTs in an autoclave, under a pressure of 3 MPa, under N2; this was followed by crosslinking at 470 °C for 1 h, under N2. This process was then repeated. Finally, the samples were pyrolyzed at 1300 °C for 1 h; this process was carried out in N2 ambience. All of the above procedures make up one cycle of this process. In this study, 10 cycles were performed. Fourier transform infrared (FT-IR) spectra were obtained in the range of 4000–400 cm1, using a Nexus 670 Fourier transform infrared spectrophotometer. KBr discs were used, which were prepared by compressing a finely ground mixture of about 5 mg of the sample and 300 mg of KBr powder. Thermal analysis (TG, Netzsch STA 449C) was performed by raising the temperature at the rate of 25 °C/min, up to a temperature of 1300 °C, in flowing Ar gas. The structure of the pyrolyzed products was studied using X-ray diffraction (XRD, Siemens D500, Cu Ka). The Archimedes principle was employed to measure the bulk densities of the resulting composites. The flexural strength was measured using a three-point-bending test on specimens of dimensions 4.0 mm  4.5 mm  60 mm, with a 50 mm span and a 0.5 mm/min crosshead speed (Instron-1342). To determine the fracture toughness (KIC), single edge notched beam (SENB) tests were applied on notched specimens of dimensions 4.0 mm  8.0 mm  60 mm (with a notch 0.3 mm in width and 4.0 mm in depth), with a 0.05 mm/min crosshead speed and a

350 300

(a)

Load (N)

250 200 150

Cf/SiC/1.5wt.% MWNTs

100

Cf/SiC/0.5wt.% MWNTs Cf/SiC

50 0.2

0.4

0.6

0.8

1.0

1.2

Flexural strength (GPa)

Displacement (mm) 460

(b)

30 mm span. The reported flexure strength and fracture toughness values were taken as an average, based on five specimens. After the mechanical properties tests were performed, scanning electron microscopy (SEM, JSM-5600LV) was employed to observe the fracture surfaces of the composites. Detailed information on the morphologies of the composites was obtained using transmission electron microscopy (TEM, HF-2000 Hitachi, Japan). Before TEM observation, the samples were milled in an agate mortar and ultrasonically dispersed in ethanol, and then transferred on carboncoated copper grid.

3. Results and discussion After ultrasonic treatment, the dispersion state of the MWNTs in the A-PMS was investigated using SEM. Fig. 1a shows the APMS containing the homogeneously dispersed MWNTs with no entanglement and agglomeration. The higher magnification (Fig. 1b) image shows that some individual MWNTs partly protruded out of the surface of the precursor, and that some pores in the surface resulted from the removal of MWNTs; these findings indicated that the MWNTs were well-dispersed in the A-PMS, and that the MWNTs were well wetted by the precursor. In view of the fact that the A-PMS was synthesized via the reaction of PMS with excess SbCl3 at room temperature and under ultrasonic conditions, the A-PMS structure may have changed during the ultrasonic dispersion of the MWNTs. Fig. 2 shows the FT-IR spectra for A-PMS, and for A-PMS after ultrasonic treatment. The Si–H absorption peak at 2106 cm1 decreased slightly in magnitude after the ultrasonic treatment, while the other peaks remain unchanged; this indicated that the Si–H groups in the A-PMS reacted with the residual SbCl3 under the ultrasonic treatment. To improve the ceramic yields of the A-PMS/MWNT composites, the samples were cured after each infiltration procedure. Fig. 3 shows FT-IR spectra for the A-PMS, cured A-PMS, and cured APMS/MWNT composite. Fig. 3 shows that the peak at 2106 cm1—which was attributed to the Si-H stretching vibration—significantly decreased in magnitude after curing, indicating that some Si–H was consumed in the curing process. The reaction could be described by the following scheme:

420

 SiAH þ HASi ! SiASi  þH2 380 340 300

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.2

1.4

1.6

MWNTs contents (wt.%)

Fracture toughness (MPa•m1/2)

463

22

(c)

21 20 19 18 17

0

0.2

0.4

0.6

0.8

1.0

MWNTs contents (wt.%) Fig. 7. (a) Load-displacement curves for the obtained composites, (b) flexural strength, and (c) fracture toughness of the obtained composites as a function of MWNT content.

Moreover, the Si–H peak for both the cured A-PMS and the cured A-PMS/MWNT composite split into two peaks (2104 and 2158 cm1), corresponding to the formation of Si–C units [25] and the change of the chemical environment. The pyrolysis behavior of the cured A-PMS and A-PMS/1.5 wt.% MWNTs was investigated using TG (see Fig. 4). In Fig. 4, we can see that both the cured A-PMS and the A-PMS/1.5 wt.% MWNT composite exhibited three-stage weight loss, with losses occurring at approximately 400 °C, 800 °C and above 800 °C. In the first stage, almost no weight loss could be observed. That indicated that the samples were converted into semi-organic structures with networks and 3-dimensional structure. During the second stage, when the temperature was above 400 °C, the TG curves slope rapidly downwards. The weight loss was approximately 18%, meaning that the samples underwent a rapid conversion into inorganic structures. The weight loss in the cured A-PMS was higher than in the A-PMS/1.5 wt.% MWNT composite, which could be attributed to the introduction of the MWNTs. In the last stage, almost no weight loss was observed, which indicated that the samples had almost entirely converted into inorganic structures. Moreover, the ceramic yield of the cured A-PMS and the cured A-PMS/1.5 wt.% MWNT composite exceeded 80% at 1300 °C; the ceramic yield of the cured

464

H.-z. Wang et al. / Composites Science and Technology 72 (2012) 461–466

(a)

(b)

300 m

15 m

(c)

(d) MWNTs

300 nm

220 nm

(e)

(f)

SiC matrix

SiC matrix

MWNTs

MWNTs 220 nm

SiC matrix

Crack deflecting

150 nm

Fig. 8. Fracture surface of the obtained Cf/SiC/1.5 wt.% MWNT composites. (a) SEM image of the fracture surface of the composites, (b) magnified SEM image of the fracture surface of the composites, (c) magnified SEM image of the fracture surface of the composites, where MWNTs can be seen in the fracture surfaces, (d) MWNT protruding out of the fracture surface of the composites, with the small change in the MWNT diameter visible, (e) another MWNT protruding out of the fracture surface of the composites, with the small change in the MWNT diameter visible, and (f) the crack deflection caused by the MWNTs can be observed in the composites.

A-PMS was a little lower than that of the A-PMS/1.5 wt.% MWNT composite, which was attributed to the addition of the MWNTs. XRD was employed to determine the structure of the obtained composites. Fig. 5 shows peaks at 2h = 36°, 61°, and 72°, which correspond to the (1 1 1), (2 2 0), and (3 1 1) planes of b-SiC (JCPDS Card No.73-1665), respectively, which indicate that b-SiC was formed via the pyrolysis of cured A-PMS at 1300 °C. Fig. 5b and c show that the intensity of the peak at 2h = 36° was lower than that in Fig. 5a, and that the peaks at 2h = 61° and 72° were too weak to identify; this suggested that a lower quantity of crystalline b-SiC or smaller b-SiC grains were synthesized by the pyrolysis of the cured A-PMS/ MWNTs at 1300 °C. This can be ascribed to the introduction of MWNTs into the composite; the introduced MWNTs usually exist at the boundaries of grains, and their presence would be likely to hinder the further growth of the grains [26]. Moreover, in Fig. 5b and c, a small peak is observed at 26°, which corresponds to the (0 0 2) plane of the graphite in the MWNTs. To determine the efficiency of the PIP process, the bulk densities of the composites were measured after each cycle. Fig. 6 shows the density–cycle curves for the composites. Fig. 6 shows that the density of the composites increased with the number of cycles, but the rate of increase became very slow when the number of cycles exceeded five. The density of both composites reached 1.80 g cm3 after only five cycles, and exceeded 2.0 g cm3 after 10 cycles.

The density of the Cf/SiC/MWNT composites was higher than that of the Cf/SiC composite for the first five cycles, but in the latter five circles the Cf/SiC composite had higher density. As a consequence, it was determined that there were two significant factors affecting the density of the composites in this PIP process; the first is the ceramic yield of the cured precursor, and the second is the addition of MWNTs. Because the ceramic yield of the cured A-PMS/MWNT composite was slightly higher than that of the cured A-PMS, the density of the Cf/SiC/MWNT composites should be higher (when the samples were prepared using identical procedures). However, the sintered density of the composite will decrease because of the enrichment of MWNTs at the grain boundaries, which hinders the achievement of full density in the composites [26]. Since the two factors had opposite effects on the density in the Cf/SiC/MWNT composites, the composites did not reach higher densities during the first set of cycles which were dominated by the first factor. However, the composites were denser during the later cycles, when the second factor played a more important role in the growth of SiC microcrystals. The load/displacement curves and mechanical properties of the composites are presented in Fig. 7. From Fig. 7a, it is clear that all of the composites failed non-catastrophically. Fig. 7b and c shows that the additive effects in the mechanical properties—which were expected to result from the hybridization of Cf/SiC with the

H.-z. Wang et al. / Composites Science and Technology 72 (2012) 461–466

(a)

465

(b) SiC matrix

SiC matrix

MWNTs 200 nm

30 nm

(c)

(d)

MWNTs

Walls of MWNTs

SiC matrix

MWNTs 15 nm

2nm

SiC matrix

Fig. 9. (a) TEM image of the Cf/SiC/MWNT composites, with an MWNT bridged between two grains, (b) high-resolution TEM image of the Cf/SiC/MWNT composite, (c) TEM images of a broken nanotube; part of the MWNT embedded in the SiC matrix and the arrows indicate the minimal change at the location of the break in the MWNT, and (d) TEM image of the interface between the MWNTs and the SiC matrix.

MWNTs—were observed. The flexural strength and fracture toughness of the obtained composites increased with increasing MWNT content. Indeed, when the MWNT content in the Cf/SiC was raised to 1.5 wt.%, the resulting flexural strength and fracture toughness of the Cf/SiC/1.5 wt.% MWNT composite were 423 MPa and 23.35 MPa m1/2, respectively, much higher than those of the Cf/ SiC composite (326 MPa and 18.25 MPa m1/2). The addition of 1.5 wt.% of MWNTs led to a 29.7% increase in the flexural strength, and a 27.9% increase in the fracture toughness. The fracture surfaces of the obtained composites are shown in Fig. 8. A fiber pullout (which accounted for the gliding fracture) is evident in Fig. 8a and b, suggesting that the prevalent toughening mechanism operating in the obtained composites was the debonding and pullout of fibers. Detailed microstructure information is shown in Fig. 8c–f for the obtained composites. Some features should be noted. Firstly, MWNTs were found to be individually dispersed within the SiC matrix, and numerous individual MWNTs were protruding from the fracture surface (Fig. 8c). Secondly, the protruding MWNTs showed that the length of the pulled-out section was typically approximately 1–1.8 lm. The MWNT diameter was drastically attenuated toward the tube tip (Fig. 8d and e), quite similar to the phenomenon observed in MWNTs broken under tensile load (‘sword-in-sheath’ failure) [27]. Thirdly, crack deflecting was observed in the composites, which was caused by the MWNTs (Fig. 8f). TEM studies revealed the detailed features of the obtained composites. From Fig. 9a and b, it is clear that the MWNTs were able to form bridges between the SiC grains, which helped to improve the crack resistance of the composites. Fig. 9c shows that a MWNT was broken, which indicated the transfer of the load between the matrix and the MWNTs embedded in it. From the location of the break in the MWNT (Fig. 9c, indicated by the arrow), we can see that part of the outer shell of the broken MWNT was conserved. Fig. 9d shows that there were no obvious apertures or pores between

the matrix and the MWNTs embedded in it, which indicated a good chemical compatibility between the MWNTs and the SiC matrix. As a result, a 29.7% increase in the fracture toughness and a 27.9% increase in the flexural strength were measured for the Cf/SiC/MWNT composites, with 1.5 wt.% of added MWNTs. Similarly as previously reported by Yamamoto et al [28], the MWNTs in this study broke in the sword-in-sheath fracture mode; the outer shells broke and the inner core was subsequently pulled away, leaving fragments of the outer shells in the matrix. This caused the MWNTs in the composites to have a far higher load carrying capacity; this could lead to the development of composites with higher fracture toughness. Our present study supports Yamamoto’s viewpoints.

4. Conclusion Cf/SiC/MWNTs composites were prepared using a PIP method. It was found that the chemically modified MWNTs were well-dispersed in the precursor after ultrasonic treatment, and were conserved in the obtained ceramic composites after the application of the PIP method. The MWNTs embedded in the composites increased the pullout resistance, bridged the crack gaps, and caused crack deflection in the composites. The MWNTs broke in the sword-in-sheath fracture mode; the outer shells broke and the inner core was then pulled away, leaving fragments of the outer shells in the matrix. These fragments caused the MWNTs in the composites to have a much higher load carrying capacity; as a result, the mechanical properties of the obtained composites were improved. The addition of 1.5 wt.% of MWNTs resulted in 29.7% and 27.9% increases in the flexural strength and the fracture toughness, respectively. Using this method, large-scale Cf/SiC/MWNTs composites can be conveniently prepared; Cf/SiC/MWNTs components with good mechanical properties have been prepared in our lab (images of these components are shown in Supplementary

466

H.-z. Wang et al. / Composites Science and Technology 72 (2012) 461–466

Data, Fig. S1). This study suggests a promising future for the application of MWNTs in reinforcing structural ceramic components, as well as in other materials systems. Acknowledgment The Authors are thankful for the help from Dr. Qingling Fang and Mr. Yang Yong, as well as the financial support from the National Natural Science Foundation of China (Grant No. 51102281) and the Research Fund of NUDT (No. JC10-02-18). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.compscitech.2011.12.009. References [1] Zhou XG, You Y, Zhang CR, Huang BY, Liu XY. Effect of carbon fiber pre-heattreatment on the microstructure and properties of Cf/SiC composites. Mater Sci Eng A 2006;433:104–7. [2] Dong R, Hirata Y, Sueyoshi H, Higo M, Uemura Y. Polymer impregnation and pyrolysis (PIP) method for the preparation of laminated woven fabric/mullite matrix composites with pseudoductility. J Eur Ceram Soc 2004;24:53–64. [3] Desprec JF, Monthioux M. Mechanical properties of C/SiC composites as explained from their interfacial features. J Eur Ceram Soc 1995;15:209–24. [4] Ly HQ, Taylor R, Day RJ. Carbon fibre-reinforced CMCs by PCS infiltration. J Mater Sci 2001;36:4027–35. [5] Ahmad I, Unwin M, Cao H, Chen H, Zhao H, Kennedy A, et al. Multi-walled carbon nanotubes reinforced Al2O3 nanocomposites: mechanical properties and interfacial investigations. Compos Sci Technol 2010;70:1199–206. [6] Treacy MMJ, Ebbesen TW, Gibson JM. Exceptionally high Young’s modulus observed for individual carbon nanotubes. Nature 1996;381:678–80. [7] Barber AH, Andrews R, Schadler LS, Wagner HD. On the tensile strength distribution of multiwalled carbon nanotubes. Appl Phys Lett 2005;87:203106. [8] Demczyk BG, Wang YM, Cumings J, Hetman M, Han W, Zettl A, et al. Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes. Mater Sci Eng A 2002;334:173–8. [9] Mintmire JW, White CT. Electronic and structural properties of carbon nanotubes. Carbon 1995;33:893–902. [10] Lourie O, Wagner HD. Evidence of stress transfer and formation of fracture clusters in carbon nanotube-based composites. Compos Sci Technol 1999;59:975–7.

[11] Schadle LS, Giannaris SC, Ajayan PM. Load transfer in carbon nanotube epoxy composites. Appl Phys Lett 1998;73:3842–4. [12] Ajayan PM, Stephan O, Colliex C, Trauth D. Aligned carbon nanotube arrays formed by cutting a polymer resin–nanotube composite. Science 1994;265:1212. [13] Laha T, Chen Y, Lahiri D, Agarwal A. Tensile properties of carbon nanotube reinforced aluminum nanocomposite fabricated by plasma spray forming. Composites: Part A 2009;40:589–94. [14] Zeng X, Zhou G, Xu Q, Xiong Y, Luo C, Wu Jicai. A new technique for dispersion of carbon nanotube in a metal melt. Mater Sci Eng A 2010;527:5335–40. [15] Esawi AMK, Morsi K, Sayed A, Gawada AA, Borah P. Fabrication and properties of dispersed carbon nanotube aluminum composites. Mater Sci Eng A 2009;508:167–73. [16] Yamamoto G, Omori M, Hashida T, Kimura H. A novel structure for carbon nanotube reinforced alumina composites with improved mechanical properties. Nanotechnology 2008:19. [17] Yamamoto G, Mamoru O, Kenji Y, Toshiyuki H, Koshi A. Structural characterization and frictional properties of carbon nanotube/alumina composites prepared by precursor method. Mater Sci Eng B 2008;148:265–9. [18] Yamamoto G, Omori M, Kenji Y, Toshiyuki H. Mechanical properties and structural characterization of carbon nanotube/alumina composites prepared by precursor method. Diam Relat Mater 2008;17:1554–7. [19] Zhang SC, Fahrenholtz WG, Hilmas GE, Yadlowsky EJ. Pressureless sintering of carbon nanotube–Al2O3 composites. J Eur Ceram Soc 2010;30:1373–80. [20] Inam F, Yan H, Peijs T, Reece MJ. The sintering and grain growth behaviour of ceramic–carbon nanotube nanocomposites. Compos Sci Technol 2010;70:947–52. [21] Yamamoto G, Yokomizo K, Omori M, Sato Y, Jeyadevan B, Motomiya K, et al. Polycarbosilane-derived SiC/single-walled carbon nanotube nanocomposites. Nanotechnology 2007;18:145614. [22] Ziegler G, Richter I, Suttor D. Fiber-reinforced composites with polymerderived matrix: processing, matrix formation and properties. Composites: Part-A 1999;30:411–7. [23] Lorca J, Singh RN. Influence of fiber and interfacial properties on fracture behavior of fiber-reinforced ceramic composites. J Am Ceram Soc 1991;74:2882–90. [24] Brennan JJ. Interfacial Characterization of slurry-cast melt-infiltrated SiC/SiC ceramic-matrix composite. Acta Mater 2000;48:4619–28. [25] Liu L, Li X, Xing X, Zhou C, Hu H. A modified polymethylsilane as the precursor for ceramic matrix composites. J Organomet Chem 2008;693:917–22. [26] Vasiliev AL, Poyato R, Padture NP. Single-wall carbon nanotubes at ceramic grain boundaries. Scripta Mater 2007;56:461–3. [27] Yu MF, Louire O, Dyer MJ, Moloni K, Kelly TF, Ruoff RS. Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load. Science 2000;287(5453):637–40. [28] Yamamoto G, Shirasu K, Hashida T, Takagi T, Suk JW, An J, et al. Nanotube fracture during the failure of carbon nanotube/alumina composites. Carbon 2011;49:3709–16.