Chinese Materials Research Society

Progress in Natural Science: Materials International

Progress in Natural Science: Materials International 2013;23(2):157–163

1002-0071 & 2013 Chhttp://dx.doi.org/10.10

nCorresponding authE-mail address: g

Peer review under r

www.elsevier.com/locate/pnsmiwww.sciencedirect.com

ORIGINAL RESEARCH

Microstructure and flexural properties of carbon/carboncomposite with in-situ grown carbon nanotube assecondary reinforcement

Hai Zhang, Lingjun Guon, Qiang Song, Qiangang Fu, Hejun Li, Kezhi Li

State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an 710072, China

Received 15 October 2012; accepted 20 January 2013Available online 6 April 2013

KEYWORDS

Carbon nanotube;C/C composite;Chemical vaporinfiltration;Microstructure;Flexural property

inese Materials Res16/j.pnsc.2013.03.00

or. Tel./fax: þ86 2uolingjun@nwpu.ed

esponsibility of Chin

Abstract Carbon nanotubes (CNTs) were in-situ grown in carbon felts using ferric chloride as catalystand natural gas as carbon precursor via thermal gradient chemical vapor infiltration (TGCVI).Subsequently, the carbon felts were densified to obtain CNT reinforced carbon/carbon (C/C) compositesin the same furnace. Effects of CNTs on the microstructure and flexural property of C/C composites wereinvestigated by polarized light microscopy, Raman spectroscopy, scanning electron microscopy anduniversal mechanical testing machine. The results of PLM observation and Raman analysis showed thatCNTs have two-sided effects on the microstructure of pyrocarbon: the pyrocarbons in the region withoutCNTs show medium texture; while, in the region full of CNTs, the microstructure was low-textured oreven isotropic though the TGCVD conditions would lead to the deposition of pure low texturepyrocarbons. Analysis based on stress–strain curves demonstrated that the flexural strength increasedfirst and then decreased with the CNT content increasing. When the CNT content was 5.23 wt%, theflexural strength was maximum and had a nearly 35% improvement compared with pure C/C composite.Besides, after adding CNTs, the flexural modulus of the composites decreased and the ductility increasedobviously, indicating CNTs can toughen C/C composites.

& 2013 Chinese Materials Research Society. Production and hosting by Elsevier B.V. All rights reserved.

earch Society. Production and hostin1

9 8849 4197.u.cn (L. Guo).

ese Materials Research Society.

1. Introduction

Carbon/carbon (C/C) composites have been widely used inaviation and aerospace industries due to their good thermal andmechanical properties such as high thermal conductivity, lowdensity, high specific strength, good chemical and mechanicalablation resistance [1]. The performance of C/C composites mainlydepends on the type of pyrocarbon matrix, spatial structure of

g by Elsevier B.V. All rights reserved.

www.elsevier.com/locate/pnsmidx.doi.org/10.1016/j.pnsc.2013.03.001dx.doi.org/10.1016/j.pnsc.2013.03.001dx.doi.org/10.1016/j.pnsc.2013.03.001mailto:guolingjun@nwpu.edu.cndx.doi.org/10.1016/j.pnsc.2013.03.001

Felt Φ70mm Specimens for observation

of pristine CNT

Center of

edge of specimen

H. Zhang et al.158

primary reinforcement and interfacial strength between carbonfiber and pyrocarbon matrix. Too weak interfacial strength andbrittle pyrocarbon can lead to bad performance [2].

Carbon nanotubes (CNTs) have attracted many researchers’attention for their excellent mechanical properties and uniquerolled graphitic layers since it was discovered by Sumio Iijima.Abundant theoretical and experimental studies demonstrated thattheir Young's modulus can be up to 1 TPa and tensile strength canapproach to 60 GPa [3,4], which makes CNTs one of the stiffestmaterials available. With such promising properties, CNTs arenatural candidates for reinforcement of advanced structural com-posites. For example, CNTs have been used for addictives in metal[5], polymer [6], cement [7] and ceramic [8] matrix composites.Hung et al. [9] conducted CNTs grown on carbon fibers andemployed the CNT-attached fibers to make composites with multi-scaled reinforcement for modulate composites’ intrinsical proper-ties. Their results showed that increasing of interfacial area andpulling out of nanotubes can improve the tensile strength of as-prepared composites. Li [10] et al. prepared CNT reinforced C/Ccomposites by film boiling CVI method with ferrocene and tolueneas catalyst and precursor, respectively. The test revealed that CNTsembedded in the matrix of C/C composites can be the bridgebetween fiber and carbon matrix and offer both intra-laminar andinter-laminar reinforcement, thus improving the delaminationresistance and the flexural performance of C/C composites.

Up to now, the main motivation of doped nano-scaled CNTs isto alleviate the existing limitations associated with the matrixdominated material and to improve the performances of thecomposites. In general, there still are a few reports on thedevelopment of CNTs as reinforcement for brittle C/C composites.In our study, C/C composites with multi-scaled reinforcements(i.e. micro-scaled carbon fiber and nano-scaled carbon nanotube)were fabricated by densifying two-dimensional carbon fiber feltsdoped with CNTs in-situ grown in them via thermal gradientchemical vapor infiltration (TGCVI). The microstructure andflexural mechanical property of these composites were alsoinvestigated further.

specimen

Fig. 1 Schematic diagram of sampling for observation of pristineCNT (a), device for three points flexural test (b).

2. Experimental

2.1. Preparation of CNT-C/C composites

Two-dimensional (2D) carbon fiber felts (carbon fiber: T300; size:Φ 70 mm� 10 mm; 0.4 g/cm3) were used as the starting materials.They were immersed into FeCl3 � 6H2O aqueous solution withdifferent concentrations (0.5 wt%, 1 wt%, 1.5 wt%, 2 wt% and 4 wt%) at room temperature for 10 h and dried in air at 80 1C. Afterwards,they were introduced into a TGCVI reactor for CNT growth.

In our work, the CNT growing temperature was set to be1000 1C and growing time was 5 h. Before reaching the desiredtemperature, hydrogen and Ar (0.1 m3/h and 0.08 m3/h, respec-tively) were fed into the TGCVI reactor for protection andreduction of catalyst. Upon reaching 1000 1C, the hydrogen andAr were turned off and Natural gas was fed into the furnace with aflow rate of 0.6 m3/h as the carbon source of CNT growth. After5 h growth, the felts were taken out from reactor and weighedusing an analytical balance. The mass gain was ascribed to theformation of CNT, content of which are 1.01 wt%, 5.23 wt%,6.64 wt%, 8.41 wt%, 12.93 wt%, respectively. And then, theywere densified by TGCVI to obtain CNT reinforced C/C (CNT-C/C) composites in the same furnace at 1000 1C with natural gas

as carbon source under ambient pressure. These composites weremarked by S1, S2, S3, S4 and S5, respectively. For comparison,pure C/C composite marked by S0 was also prepared under thesame conditions. Finally, all the samples were treated at 2100 1Cfor 2 h under Ar atmosphere.

2.2. Characterizations of CNTs and the composites

The morphology of CNTs and the fracture surfaces of all thecomposites were observed using Quanta-600FEG field emissionscanning electron microscopy (SEM) and Tecnai VEGA3 microscopy.The sampling for SEM observation was shown in Fig. 1(a). In order toinvestigate the microstructure of CNTs, the carbon felts with in situgrown CNTs was cut down and put into alcohol solution for ultrasonicvibration. A copper grid was immersed into alcohol to capture thesuspended CNTs, and then, the copper grid was dried for TEMobservation (TEM), which was carried out on a Tecnai F30G2 filedemission transmission electron microscopy. Specimens, with dimen-sions approximately 30 mm� 10 mm� 10 mm, were cut from thedensified composites. These specimens were pre-polished using SiCabrasive paper (600, 1000, 1500, 2000 grit successively). The finalpolishing was completed with 0.5 μm grain size diamond polishingpaste. Finishing the preparation of the specimens, microstructure ofpyrocarbon was studied with polarized light microscopy (PLM, LeicaDMLP optical microscope) and Raman spectroscopy (Renishaw inVIA, with 514.5 nm excitation) at polished cross-section.

Microstructure and flexural properties of carbon/carbon composite with in-situ grown carbon nanotube as secondary reinforcement 159

Three points flexural test was carried out with a universaltesting machine (INSTRON 8872) to evaluate the effects of CNTson the mechanical properties of C/C composites at room tempera-ture. Testing specimens were machined into rectangular bars of55 mm� 10 mm� 4 mm from the densified composites along theaxial direction of carbon fiber bundles. Before tests, the bulkdensities of the samples were measured. The tests were carriedwith the constant loading rate of 0.5 mm/min and a span of40 mm. At least six specimens were tested for each composite.Fig. 1(b) is the device of flexural test in our experiment. The stressand strain values were recorded automatically. A ductility factor(FD) was used to determine the ductility of the composites. FD canbe calculated from the following equation [11].

FD ¼ 1−ðEsecant=EoriginÞ ¼ 1−ðεlin=εtÞ ð1Þwhere Esecant is secant modulus (the slope of the line from theorigin to the stress at failure in the stress–strain curve), Eorigin iselastic modulus (the slope of the linear part of the stress–straincurve), εlin is the strain in the linear part of the stress–strain curveat failure and the εt means the strain at failure.

3. Results and discussion

3.1. Morphology and microstructure of CNTs

SEM observations of pristine CNTs in the center and edge ofspecimen are presented in Fig. 2(a and b), respectively. It can be

500 nm

Fig. 2 (a) SEM micrograph of as obtained CNTs in the edge of specim(c) TEM micrograph of CNTs in the carbon felts. (d) HRTEM micrograp

seen that the straight CNTs with the diameter of 100–150 nm inFig. 2(a) are distributed disorderly in the space between carbonfibers. The surface of carbon fiber is smooth (Fig. 2(a)) and noCNT is grown on it. The length of most CNTs is over 5 μm andeven up to 15 μm (Fig. 2(a)). It can be seen that the amounts ofCNTs in the center of the specimen are almost similar comparedwith that on the edge. However, there is still a little differencebetween the center and edge in CNT's morphology. The inset inFig. 2(b) reveals that besides some coarsen and straight CNTs,there are lots of thinner and shorter CNTs intertwining with thebigger sized CNTs. TEM images (Fig. 2(c)) show the innerdiameter and the wall thickness of CNTs are 30 nm and 40 nm,respectively. According to Fig. 2(d), the CNTs are multi-walledand the crystal lattice fringes are somewhat bent, implying thepartial crystallization of CNTs.

3.2. Pyrocarbon microstructure of the composites

Microstructures of pyrocarbon matrix of C/C composites withCNTs and pure C/C were investigated by PLM and Ramanspectroscopy. According to PLM observation of the polishedspecimens in Fig. 3, it is evident that pyrocarbons in pure C/Ccomposite (S0) are low-textured due to the low optical activitiesand the unclear extinction cross around carbon fibers (Fig. 3(a)).However, CNT doped composites (Fig. 3(b and c)) have differentoptical activities in the reflection of polarized light compared withpure C/C composite (Fig. 3(a)). The thickness of pyrocarbon is

500 nm

5 nm

en S2. (b) SEM micrograph of CNTs in the center of specimen S2.h of carbon nanotube wall.

H. Zhang et al.160

about 10 mm and the boundary of pyrocarbons grown arounddifferent fibers is extremely distinct. As for CNT-C/C composites,the pyrocarbons have higher optical activities, indicating a highertexture than that of pure C/C composite. While, in the region fullof CNTs (labeled by white rectangular frame in Fig. 3(c)), low-textured pyrocarbon or isotropic pyrocarbon is induced to form.The texture diversity of pyrocarbon is more apparent in Fig. 3(cand d). It can be seen that more close to the fiber, denser carbonnanotubes are and lower the optical activities of pyrocarbon is. Theformation of low-textured pyrocarbon should be attributed tothe CNTs agglomeration and the depositing randomness ofpyrocarbon around them [12]. As shown in Fig. 2(b), thenanometer scaled interval among CNTs and twisted CNTs willprovide large surface area in a small space, i.e. lager surface area/volume ratio, which can accelerate the deposition of pyrocarbonon the CNTs as active nucleation sites [13]. The nano-scaled poreswould constrict the size of pyrocarbon grain around CNTs. Thepyrocarbon grain originated from adjacent CNT would contacteach other soon in densification process. This deposition ofpyrocarbon is largely dependent on the randomly oriented CNTsas the substrate of deposition. As a result, the pyrocarbon growingfrom CNTs would align in different angles controlled by thepristine disordered three dimensional CNTs network. Therefore,the randomness of pyrocarbon in orientation, especially in theCNTs agglomeration region, would exhibit characterization of ISOlayer or lower texture in polarized light. Nevertheless, pyrocarbon(shown by white arrows in Fig. 3(c)) deposited around theindividual CNT can grow larger (the thickness is about severalmicrometers) and shows some characteristics of medium-texturedpyrocarbon. The π–π conjugation effect of CNT curved graphiticlayers, which attracts similar structured polyaromatic molecules to

20 �m

20 �m

Carbon fiber

Individual CNT Agglomeration

zone of CNT

Fig. 3 PLM images of composites. (a) Microstructure of co

arrange parallel each other to form regular pyrocarbon perhaps haswider influence on microstructure, which is considered to inducethe formation of medium-textured pyrocarbon. For all the C/Ccomposites doped by CNTs, the microstructure of pyrocarbon isnot homogeneous at all. Namely, medium-textured and ISOpyrocarbons coexist in the matrix. In order to further study themicrostructure of CNTs doped composites. Raman characteriza-tion of pyrocarbon in S0, S3 and the pristine CNTs was conducted.As shown in Fig. 4(b), the characteristic peak (D band) at1350 cm−1 and shoulder peak at 1620 cm−1 (D' band) correspondto presence of disorder and distortion in carbon atomic layers.The second prominent band is G band related to high-frequencyE2g first order mode and the band at 2700 cm

−1 is the overtone ofD band, called 2D band. The D and G band position of CNTsare several cm−1 lower than composites. This downshift is due tothe less carbon atomic layers of CNTs. The ratio of IG to ID(IG means the intensity of G band) depends on the perfect crystalplanar domain size and the degree of order of pyrocarbon [14].The IG/ID value for pyrocarbon in pure C/C is the lowest,corresponding to the lowest crystal ordered degree [15].The second lowest IG/ID value corresponding to pyrocarbonoriginated from fiber (labeled with point 1). Meanwhile, the IG/ID value of pyrocarbon grown from individual CNT (labeledwith point 2) is a little larger compared with that at point 1, whichis normally due to the improvement of crystal ordered degree andfewer defects in microstructure. According to the highest ratiovalue presented in the Raman spectrum of pristine CNTs, there-fore, it is considered that the CNTs have the highest texture. As forpyrocarbon of CNTs agglomeration region (labeled with point 3),the second larger IG/ID ratio value may virtually be attributedto the presence of CNTs within the laser spot. In our work, these

CNT

fiber

20 �m

20 �m

Medium

texture ISO pyrocarbon fiber

mposite S0. (b and c) Composite S2. (d) Composite S3.

600 900 1200 1500 1800 2100 2400 2700 300-200

0200400600800

100012001400160018002000220024002600280030003200

2D band

G band

pure C/C

pristine CNTs

point 2

point 3

inte

nsity

/a.u

.

Raman shift /cm-1

point 1

D band

D' band

Point 1

Point 2

Point 3

20 �m

Fig. 4 (a) Microstructures of composite S3 under normal light (left) and polarized light (right). Point 1, 2 and 3 show the positions of Ramantesting (b) Raman spectra of pure C/C, composite S3 at different points and pristine CNTs.

Table 1 Three point flexural strength and modulus for all composites.

Composites S0 S1 S2 S3 S4 S5

Density (g cm−3) 1.75 1.62 1.57 1.60 1.57 1.59Flexural strength (MPa) 8673 9772 116715 10878 6576 50.978Flexural modulus (GPa) 16.7772.1 12.2472.2 12.9773.5 13.4373.8 8.871.2 10.2870.8FD o0.005 0.23370.059 0.48870.073 0.45370.098 0.41170.055 0.42470.92

0.000 0.005 0.010 0.015 0.020 0.025 0.030

0

20

40

60

80

100

120

140

stre

ss (M

Pa)

strain

S0

S3

S1

S4

S5

S2

Fig. 5 Stress–strain curves for composites S0, S1, S2, S3, S4 and S5.

Microstructure and flexural properties of carbon/carbon composite with in-situ grown carbon nanotube as secondary reinforcement 161

findings indicate that besides the CVI deposition parameters,CNTs and its distribution can also affect the microstructureof pyrocarbon. Concretely, CNTs can induce the formation ofhigher- or lower-textured pyrocarbon, which is similar to other’swork [16].

3.3. Flexural properties and fracture behaviors of C/Ccomposites

The results of three points bending test are listed in Table 1. Itpresents the average flexural strength and modulus for the sixcomposites. The average flexural strengths of composites S0, S1,S2, S3, S4, and S5 are 80 MPa, 97 MPa, 116 MPa, 108 MPa,65 MPa and 50.9 MPa, respectively. The flexural strengthincreases first and then decreases with the CNT content increasing.When the CNT content is 5.23 wt%, the flexural strength is

maximum and has a nearly 35% improvement compared with pureC/C (S0). However, when the CNT content increases to 8.41 wt%or more, the flexural strength decreases from the maximum value.On the hand, this may be the result of the degradation of carbonfiber tensile strength caused by the dissolution of more iron catalystinto fibers and more formation of low texture pyrocarbon incomposites with higher concentration catalyst [17]. On the otherhand, another reason for the reducing of flexural strength might bethe aggregation of the CNTs, which leads to the sealing of nano-scaled pores and poor infiltration of the precursor and deposition ofpyrocarbon. The defects in the aggregation region of CNTs, forinstance, pores exists in matrix (Fig. 6g) and make the continuouspyrocarbon matrix separated, causing lower the contribution ofmatrix to the flexural strength. Flexural modulus for all hybridcomposites S1–S5 are less than that of composite S0 completely,which is different from variation of flexural strength. This variationof flexural modulus is similar with that in literature [18]. Thisreducing of flexural modulus can be attributed to the ISOpyrocarbon formed in matrix under the influence of agglomeratedCNT. As seen in Fig. 6(g and h), the coarsened CNTs and theconcave pits are obvious, and most coarsened CNTs in theagglomeration zone are more liable to detach from the matrixrather than the expected fulling out of CNTs or cracking. As asequence, the strength and modulus of CNT cannot be fullyutilized. However, with CNT content increasing, the CNTsaggregation becomes more serious in composites S4 and S5,which deteriorate the ultimate properties due to the defects in theregion as discussed above.

The stress–strain curves for composites are shown in Fig. 5. Asseen from the curves, it is noticeable that the composites S1, S2,S3 have higher fracture strengths and larger strains than compositeS0. For CNT-doped composites (S1, S2, S3, S4, S5), after themaximum stress, the curves drop slowly and have a larger ductilityfactor FD, exhibiting a pseudo plastic fracture mode, whereas the

Debonding

of fiber

Pulling out

of fiber Debonding

of CNTs

Bridge

Pulling out of CNT

Holes left by pulled out CNTs

Agglomeration zone of

CNTs with low texture Pulling out of

CNT

Pores

Fracture of coarsened

CNTs

Pyrocarbon

Pulling out of CNT

Fracture of

coarsened CNTs

Concave pits left by the coarsened

CNTs separation from matrix

Fig. 6 (a) SEM images of the fracture surfaces of composite S0. (b) Magnified images of fracture surfaces of composite S0. ((c)–(f)) SEM imagesof the fracture surface of composite S2: (c) Debonding and pulling out of fibers occurs in the surface. (d) Bridge and debonding of CNTs.(e) Fracture surface of pyrocarbon deposited in the agglomeration zone of CNTs. (f) Fracture surface of pyrocarbon deposited around individualCNT. ((g) and (h)) Fracture morphology of composite S4.

H. Zhang et al.162

Microstructure and flexural properties of carbon/carbon composite with in-situ grown carbon nanotube as secondary reinforcement 163

pure C/C composite S0 shows a brittle fracture behavior due tosharp decline of stress and small fracture strain indicatingintroduction of CNTs is contributed to the improvement of flexuralductility. In the platform of curves, it is thought that the partialfibers have cracked and the rest fibers and CNTs still bear the load,thus, they absorb the energy during being pulled out resulting information of platform. The curves for hybrid composites S1–S5display obvious platform, meaning better ductility.

To investigate the reinforcing mechanism of CNTs, the fracturesurfaces of composite S0, S2 and S4 are observed by SEM (Fig. 6).For pure C/C S0, the fracture surface is flat (Fig. 6(a)) and only afew fibers are pulled out from carbon matrix (Fig. 6(b)) whichconfirms the strong fiber/matrix interface bonding and brittlefracture mode. As seen from Fig. 6(c), the extraction anddebonding of carbon fibers shown by white arrows can be seenin composite S2, besides, CNT pull-out from matrix is also found,both of which can toughen C/C composites. It is evident that thedoping of CNTs into matrix can tremendously increase theinterface area between reinforcement and matrix. Therefore, whenthe stress is loaded on the specimens, fracture is more difficult tooccur for CNT-doped composites by virtue of additional surfaceenergy, pull-out work and friction work because of the newlyformed surface area [19]. The CNTs inserted in pyrocarbon shownin Fig. 6(d) link adjacent pyrocarbon layers together as a bridge,which can avoid the formation of concentric crack around carbonfiber to a certain extent. Hence, it is beneficial to the loadtransmitting and sufficient utilization of carbon fiber. And whenthe CNTs were pulled out, an overview of many holes left inmatrix is clearly shown in Fig. 6(e). The multilayer carbon matrix(Fig. 3(c)) has been formed under the influence of interlaced CNTsshown in Fig. 6(e and f). During the fracture, multi-layered carbonmatrix can lead cracks to spread along multiple paths and results inan uneven fracture surface. Therefore, the crackle would havemuch greater interaction with the pyrocarbon matrix duringpropagating, which further improves mechanical interlockingbetween reinforcements and matrix and toughness [20,21]. Fromthe discussed, it will be significant on improvement of themechanical properties of CNTs doped composite to control theproper content and distribution of pristine CNTs in felts.

4. Conclusions

C/C composites doped by CNTs (as secondary reinforcements)were prepared by TGCVI. According the observations of PLM andthe Raman analysis, CNTs in-situ grown in carbon felts had two-side effects on the microstructure of pyrocarbon. ISO pyrocarbonwas obtained in the aggregation region of CNTs; while, medium-textured pyrocarbon was induced in the sparse region of CNTs andaround individual CNT. As contrast, the pyrocarbon of pure C/Cwas low-textured. Mechanical tests and fracture surface analysisshow that the suitable CNTs content can improve the flexuralstrength and ductility, while, the modulus of prepared multi-scaledC/C composites with CNTs generally decreased compared withpure C/C composite.

Acknowledgments

This work has been supported by the programme of introducingtalents of discipline to universities (B08040), key grant project ofChinese ministry of education (313047) and national naturalscience foundations of China (51275417 and 51221001).

References

[1] T. Windhorst, G. Blount, Carbon–carbon composite: a summary ofrecent developments and application, Materials and Design 18 (1)(1997) 11–15.

[2] A. Oberlin, Pyrocarbons, Carbon 40 (1) (2002) 7–18.[3] J.P. Lu, Elastic properties of carbon nanotubes and nanoropes,

Physical Review Letters 79 (7) (1997) 1297–1300.[4] M. Sammalkorpi, A. Krasheninnikov, A. Kuronen, Mechanical

properties of carbon nanotubes with vacancies and related defects,Physical Review B: Condensed Matter 70 (24) (2004) 245416–245418.

[5] J. Stein, B. Lenczowsk, N. Frety, Mechanical reinforcement of a high-performance aluminium alloy AA5083 with homogeneously dispersedmulti-walled carbon nanotubes, Carbon 50 (6) (2012) 2264–2272.

[6] S.S. Wicks, R.G. de Villoria, B.L. Wardle, Interlaminar and intrala-minar reinforcement of composite laminates with aligned carbonnanotubes, Composites Science and Technology 70 (1) (2010) 20–28.

[7] B.M. Tyson, R. Al-Rub, A. Yazdanbakhsh, et al., Carbon nanotubesand carbon nanofibers for enhancing the mechanical properties ofnanocomposite cemetitious materials, Journal of Materials in CivilEngineering 23 (7) (2011) 1028–1035.

[8] A. Peigney, C. Laurent, E. Flahaut, A. Rousset, Carbon nanotubes innovel ceramic matrix nanocomposites, Ceramics International 26 (6)(2000) 677–683.

[9] K.H. Hung, W.S. Kuo, Processing and tensile characterization ofcomposites composed of carbon nanotube-grown carbon fibers,Composites Part A: Applied Science and Manufacturing 40 (8)(2009) 1299–1304.

[10] X.T. Li, K.Z. Li, H.J. Li, Microstructures and mechanical propertiesof carbon/carbon composites reinforced with carbon nanofibers/nanotubes produced in situ, Carbon 45 (8) (2007) 1662–1668.

[11] B. Reznik, M. Guellali, D. Gerthsen, Microstructure and mechanicalproperties of carbon–carbon composites with multilayered pyrocarbonmatrix, Materials Letters 52 (1-2) (2002) 14–19.

[12] K.Z. Li, Q. Song, Influence of carbon nanotube content on micro-structures and mechanical properties of carbon/carbon composite,Chinese Journal of Inorganic Chemistry 27 (5) (2011) 1001–1008.

[13] J. Antes, Z. Hu, W. Zhang, Chemistry and kinetics of chemical vapordeposition of pyrocarbon VII: Confirmation of the influence of thesubstrate surface area/reactor volume ratio, Carbon 37 (1999) 2031–2039.

[14] W.Z. Li, H. Zhang, C.Y. Wang, Raman characterization of alignedcarbon nanotubes produced by thermal decomposition of hydrocarbonvapor, Applied Physics Letters 70 (20) (1997) 2684–2686.

[15] E.L. Honorato, P.J. Meadows, Fluidized bed chemical vapor deposi-tion of pyrolytic carbon-I. Effect of deposition conditions onmicrostructure, Carbon 47 (2009) 396–410.

[16] Q.M. Gong, Z. Li, Fabrication and structure: a study of alignedcarbon nanotube/carbon nanocomposites, Solid State Communica-tions 131 (6) (2004) 399–404.

[17] Q. Hui, A. Bismarck, E.S. Greenhalgh, Carbon nanotube graftedcarbon fibers: a study of wetting and fibers fragmentation, CompositesPart A: Applied Science and Manufacturing 41 (9) (2010) 1107–1114.

[18] K.Z. Li, H.L. Deng, Microstructure and mechanical properties ofcarbon/carbon composites doped with LaCl3, Materials Science andEngineering: A 529 (25) (2011) 177–183.

[19] X.W. Wu, R.Y. Luo, Deposition mechanism and microstructure ofpyrocarbon prepared by chemical vapor infiltration with kerosene asprecursor, Carbon 47 (6) (2009) 1429–1435.

[20] Q. Song, K.Z. Li, H.L. Li, Grafting straight carbon nanotubes radiallyonto carbon fibers and their effect on the mechanical properties ofcarbon/carbon composites, Carbon 50 (10) (2012) 3949–3952.

[21] H.L. Li, H.J. Li, J.H. Lu, Effects of a pre-deposited pyrocarbon layeron the microstructure and mechanical properties of carbon/carboncomposites, Materials Science and Engineering: A 556 (2012) 295–301.

Microstructure and flexural properties of carbon/carbon composite with in-situ grown carbon nanotube as secondary...IntroductionExperimentalPreparation of CNT-C/C compositesCharacterizations of CNTs and the composites

Results and discussionMorphology and microstructure of CNTsPyrocarbon microstructure of the compositesFlexural properties and fracture behaviors of C/C composites

ConclusionsAcknowledgmentsReferences