CERAMICSINTERNATIONAL

Available online at www.sciencedirect.com

http://dx.doi.org0272-8842/& 20

nCorrespondinE-mail addre

(2016) 1762–1768

Ceramics International 42 www.elsevier.com/locate/ceramint

The effects of stitched density on low-velocity impact damageof cross-woven carbon fiber reinforced silicon carbide composites

Hui Mein, Changkui Yu, Hongrui Xu, Laifei Cheng

Science and Technology on Thermostructural Composite Materials Laboratory, Northwestern Polytechnical University, Xi'an, Shaanxi 710072, PR China

Received 12 July 2015; received in revised form 23 September 2015; accepted 23 September 2015Available online 1 October 2015

Abstract

Two dimensional carbon fiber reinforced silicon carbide composites (2D C/SiCs) subjected to low-velocity impact (LVI) damage wereinvestigated, in order to evaluate the efficiency of stitching as a reinforcing mechanism able to improve the delamination resistance of 2D C/SiCs.The damage microstructures of the specimens at different stitched density (SD) were observed by infrared thermography and industrial computedtomography scanners. While the damage depth of specimens with the SD of 10 mm/needle was greater than that of specimens with SD of 5 or15 mm/needle, the residual tensile strength of the specimens with the SD of 10 mm/needle was the highest. With the decreasing of SD, the realdamage radius of 2D C/SiCs measured by thermography increased whereas the residual tensile strength did not appear the same phenomenon.The 2D C/SiCs with the SD of 5, 10, and 15 mm/needle had good damage resistance after the LVI, with the tensile strength still retaining72.43%, 95.20%, and 91.49%, respectively.& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Ceramic matrix composites; Stitched density; Impact behavior; Low velocity impact; Nondestructive tests

1. Introduction

The carbon fiber reinforced silicon carbide composites (C/SiCs) for its superior properties (such as excellent hightemperature resistance, low density, and high wear resistance)have an increasingly important application prospect in rotaryultrasonic machining and high temperature structural applica-tions especially in the aerospace field [1–3]. The ratios ofstiffness to mass and strength to mass and high temperatureproperties of structural material are of great interest in theaerospace field. The aeroplane composed of C/SiCs, however,may be damaged by hail impact, accidental impacts fromdropped tools during maintenance and servicing, or impactsfrom stones on the tarmac during take-offs and landings. Thesedamages may seriously destroy the structural integrity of theaircraft, so it is necessary to use impact test to simulate theeffect of the foreign object damage (FOD) on the composites.The FOD imparted to a thermal barrier system in a turbine

/10.1016/j.ceramint.2015.09.13715 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

g author. Tel.: þ86 29 88495312; fax: þ86 29 88494620.ss: phdhuimei@yahoo.com (H. Mei).

engine has been researched [4]. The FOD behaviors of three-dimensionally woven silicon carbide SiC/SiC composites atroom temperature and 800 1C were investigated and theembrittled damage characteristics in thermally exposed speci-mens were observed [5].In order to simulate the FOD process veritably, the impact test

has been widely reported [6–11]. The static indentation and impacttests were undertook on four carbon fiber reinforced polymermaterials and then evaluated the influence of fiber type, fabricweave pattern, and resin system [6]. The ceramic matrixcomposites were received relatively limited attention, mainlyfocusing on the high velocity impact (4100 m/s) behavior ofSiC/SiC composites [7]. First the term ‘low-velocity impact’ (LVI)was defined and the major impact-induced damage modes weredescribed. The LVI is considered potentially dangerous mainlybecause the damage may be left undetected. Whereafter, thin threedimensions-woven SiC/SiC specimens subjected to LVI tests atroom temperature were investigated [8,9]. The damage behaviorof 2D C/SiCs after the LVI was experimentally investigated[10,11]. The residual mechanical and thermo-mechanical proper-ties of C/SiC, which was exposed to the LVI, were studied. The

Supporter

Sample

Impactor

IR detector

Fig. 1. Photogragh of weight drop impact testing machine with the config-uration of the sample, impactor, IR detector, and supporter. The inset showingthe supporter and sample.

H. Mei et al. / Ceramics International 42 (2016) 1762–1768 1763

changes in microstructure and thermo-mechanical properties of C/SiC through exposure to multiple experimentally simulated re-entries were investigated.

To evaluate the resistant ability of C/SiCs to FOD, not onlywas the impact process to 2D C/SiCs considered, but also thepreparation and structures of 2D C/SiCs were investigated. Thein-plane dimensions showed a great influence on the mechan-ical behavior of composite structures under the LVI [12].Certainly, it was investigated that the carbon fiber's stackingsequence made a difference to low-velocity impact character-istics and residual tensile strength of carbon fiber compositelaminates [13,14]. Stitching did not increase the load at whichdelamination begins to propagate, but greatly reduced theextent of delamination growth at the end of the impact event[15]. A comprehensive study [16] demonstrated that thestitching was particularly effective in improving thecompression-after-impact strength and Mode I fracture tough-ness of carbon/epoxy laminates, and moderately effective inimproving the Mode II fracture toughness.

However, so far few LVI damage date with the effects ofdifferent stitched densities (SDs) on 2D C/SiCs was reported.In this paper, the effects of SD on the mechanical property andmicrostructure of 2D C/SiCs were discussed. Three non-destructive testing (NDT) methods (such as infrared (IR)thermograph, X-ray computed tomography, and microscopy)were used to detect the 2D C/SiC specimens after the LVI. Itwas suggested that thermography should be used to evaluatethe LVI damage of C/SiC [17,18].

2. Experiments and methods

2.1. Materials

T300™ carbon fiber [01/451] fabric (Toray industries Inc.,Tokyo, Japan) was used to prepare two-dimensional (2-D)preform by laminating layer by layer. The planar laminateswere stitched by T300™ carbon fiber to 5, 10, and 15 mm/needle of specimens, manually. Pyrolytic carbon (PyC) inter-face was deposited on the fibers inside the 2D preform by achemical vapor deposition (CVD) method at around 900 1C,and through a chemical vapor infiltration (CVI) method thensilicon carbide (SiC) matrix was infiltrated into the preform ataround 1000 1C. The detailed CVD and CVI conditions andparameters were described elsewhere [19,20]. In order toinvestigate the effect of the SD on tensile strength, the C/SiC specimens after the LVI were cut from the as-fabricatedcomposite plates into dimensions of 200 mm (length)� 50 mm(width)� 3 mm (thickness) for tension. The specimens withthree stitched densities were denoted as S1 (5 mm/needle), S2(10 mm/needle), and S3 (15 mm/needle). The densities ofcomposite specimens were 2.0 g/cm3 in average and thevolume fraction of carbon fiber approximated to 40%.

2.2. Low velocity impact test

LVI tests were conducted on an automatic drop-weightimpact testing machine (Sans Materials Co., Shenzhen, China).

The machine controlling landing height which is able tocontrol the low-velocity process, as visible in the graph ofFig. 1, has the ability of automatic lifting and holding thehammer system of preventing secondary impact. The impactexperimental machine is made up of sample, impactor, IRdetector, and supporter. The impactor weighted 30 kg is madefrom 45 steel in 16 mm diameter according to ASTM D7136-05 standard [21]. The deformation of impactor material wasnot considered. The damage depths of the specimens weremeasured by a Digital Indicator (Guanglu Number MeasuringInstrument Co., Shenzhen, China) and then the observeddamage radii were calculated by,

ROB ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiR2�ðR�HÞ2

qð1Þ

where ROB is the observed damage radius, R is the impactorradius and H is the damage depth.The monolithic tensile tests of the damaged C/SiC speci-

mens were conducted on an electronic universal testingmachine (Mechanical Engineering Research Institute Co.Ltd., Changchun, China) according to ASTM C1275-00standard [22]. The mechanical tests were implemented in adisplacement controlled mode with a loading rate of 0.5 mm/min, and both ends of the specimens were bonded withAluminum tab in order to prevent the specimen end from gripcrush. The plate specimen number of each condition was fiveto guarantee data efficiency. Microstructures and morphologiesof the C/SiC specimens were observed by scanning electronmicroscopy (SEM, Hitach S-2700, Tokyo, Japan).

2.3. Nondestructive tests

The damages of the specimens after the LVI were examinedby using an infrared (IR) thermograph (EchoThem, TWI Co.Ltd., USA). The instrument consists of the infrared cameras,thermal excitation systems and computer with special software,image acquisition and processing system. The heat application

Fig. 2. Photographs of impact damage to the C/SiC composites at SD¼5 mm/needle showing (a) front side and (b) back side.

Fig. 3. The effect of stitched density (SD) on the damage depth at differentstitched densities of SD¼5, 10, and 15 mm/needle.

H. Mei et al. / Ceramics International 42 (2016) 1762–17681764

was achieved by directing the output of four 2.4 kJ xenon flashlamps contained within a hood assembly which helps to focusthe energy onto the detection surface. The front (impact test)sides of 2D C/SiC plate specimens were shone with the flash oflight momentarily. After the flash, the irradiated surface slowlyconducted heat into the sub-surface. With time, the IR imagesof the surface and subsequently the images of the specimens atdeeper thickness were received when the heat was conducted.

An X-ray computed tomography (CT, BT500, Indintro Co.Ltd., Moscow, Russia) was provided to examine the damagesof the C/SiC specimens after the LVI. Basic principle of theCT inspection method was described in detail according to thepreceding work [23,24]. The CT system allowed 380 kV tubevoltages with a spatial resolution of up to 30 Lp/cm and singleslice thickness of 0.5 mm for detecting changed in density aswell as defect.

3. Results and discussion

3.1. LVI damage and characteristic

As can be seen in Fig. 2, the final fracture of the compositeplate looks like a ‘triadius’ crack in the back side (Fig. 2b) dueto fiber tensile breaking perpendicular to the cracks ([01/451]fiber alignments), and ‘spherical cap’ pit in the front side (Fig.2a) due to SiC matrix compression shear cracking. Thedamage of specimen (S1) in Fig. 2 contains actually the wholeimpact process: (a) the first stage of linear elastic compressionof the SiC matrix; (b) the second stage of pores collapse;(c) the third stage of delamination; and (d) the fourth stage ofthe fiber transferring to bear the load the fiber fracture. Finally,the C/SiC material is seriously damaged and thus loses theability to bear the LVI.

Fig. 3 illustrates the effect of SD on the damage depthobserved by the human eyes. The damage depths were left atthe surface of the substrate after impactor rebounding. In termsof the damage depths of specimens (S1, S2, and S3), the threegroups do not show a great difference, but the damage depthsof S2 is slightly deeper than that of S1 and S3. This can beexplained in the following argumentation. While the higher SDof the 5 mm/needle destroys the reinforcing fiber and offersmore deposition channels for the CVI preparation, it also

results in more deposition of SiC matrix. Consequently, thedamage depth is relatively small. Meanwhile, the lower SD,such as 15 mm/needle, provides the lower deposition channeland lower SiC matrix deposition compared to 5 mm/needle,but it damages slightly the reinforced carbon fiber to produce asmaller damage depth of the material. These findings demon-strate that neither higher nor lower SD makes the damagedepth the biggest. In order to obtain a better impact damageresistance of materials, only from the surface damage analysis,there should be an optimal SD.Of course, only the observation of the surface damage

cannot validly explain the degree of the specimens after theLVI. In order to characterize the extent of the damage better,therefore, the specimens are detected by thermography. Asshown in Fig. 4, the size and shape features of the damage in arelatively lighter color are distributed at the center of the darkerbackground of the specimens. It indicates that the variation ofdamage radii of the specimens (S1, S2, and S3) appear differentcompared to the apparent damage radius ROB. The infraredthermograph results reveal that the damage radii of specimensincrease with a reduction in specimens' SD. The real damageradius RIR is actually much larger than the apparent damageradius ROB, as shown in Table 1. Fig. 4 also shows that thedamage area contains the invisible delamination which occursalways along the interfaces of tensile stress states, resultingaccordingly in much larger real damage areas than theapparent ones.

delaminationdelamination

Fig. 4. Infrared images of stitched samples after the LVI at different stitched densities of SD¼(a) 5 mm/needle, (b) 10 mm/needle, and (c) 15 mm/needle.

Table 1Statistical results of original or residual tensile strength, modulus, and damage radius of the C/SiC specimens with different stitched densities after the LVI.

Specimen Tensile strength (MPa) Modulus (GPa) Strength retaining ratio (%) Damage radius R (mm)

Original Residual Observed IR

S1 124.7973.77 90.3972.66 112.3474.42 72.43 6.4670.49 7.4570.12S2 126.1574.88 120.1074.75 110.8973.28 95.20 6.5170.29 8.2470.17S3 126.99711.65 116.19712.22 79.8070.02 91.49 6.0470.31 8.8570.42

15 mm /needle

10 mm /needle

5 mm /needle

2RIR

2ROB

Delamination

Cone cracking

Fig. 5. Computed tomography slice images of 2D C/SiCs specimens with low velocity impact at the different stitched densities of SD¼5, 10, and 15 mm/needle.

H. Mei et al. / Ceramics International 42 (2016) 1762–1768 1765

The computed tomography slice images of specimens (S1,S2, and S3) are observed in Fig. 5. The damage area decreasedas the SD increased, which is also shown by Fig. 4. It isreasonable to infer that the degree of the delamination of 2D C/SiCs reduced by the high SD is more than that minimized bythe lower SD. As can be observed in Fig. 5, there exist theinvisible damage zones of delamination in gray color at thedeeper thickness of the composites between the light color ofdirectly damaged zone and the dark color of non-damagedzone. Importantly, the delamination occurs actually along withthe interfaces of the tensile stress states in the compositesfollowed by the cone cracks of the tensile stress. It is foundthat the 2D C/SiCs' delamination increases dramatically withthe decrease of the SD. Especially, the internal delamination ofspecimens is more serious as the SD is low. There are twocauses for the phenomenon. The primary one is that the lowSD results in less matrix deposition and poor interlayerproperty. So the delamination appears heavily after the speci-mens are subjected to the LVI. The specimens, on the otherhand, do not exist the Z-direction reinforced phase. The

internal delamination shows a great influence on damagetolerance from the residual property analysis. As a conse-quence, the main injury patterns are found out within the scopeof application of 2 D C/SiCs before the optimal SD isdetermined.

3.2. Residual mechanical properties

As can be seen in Fig. 6, it indicates that the fracturemorphology appears brittle fracture and slightly changed withthe decrease of SD. According to Mei [25] proposed stretchinjury model, Fig. 6 shows that tensile fracture occurs slight‘layer pull’ phenomenon with the lower SD and fracturemorphologies of the specimens would be plain with the higherSD. Shown by the stress–strain curves of specimens in Fig. 8,this finds are mainly due to the fact that the SD of thespecimen is too high and then the fibers are seriously damagedduring the preparation of material, which reduces the tensilestrength of the fiber and leads to plain fracture morphologies.Furthermore, the higher the SD, the more the deposition

50 mm 50 mm 50 mm

Fig. 6. Fractured surface of the stitched specimens under tensile load with different stitched densities of SD¼ (a) 5 mm/needle, (b) 10 mm/needle, and (c) 15 mm/needle.

Axial crack Radial fracture

Debonding

Axial crack

Pull-out

Fiber breaking

Crack

Fibre bundle cracking

Fibre Crack Pull-out

Fig. 7. SEM micrographs of the LVI damaged C/SiC composites showing (a, b) matrix cracking and fiber fracture at SD¼5 mm/needle, (c, d) axial cracking, radialcracking, debonding, and pull-out at SD¼10 mm/needle, (e, f) matrix cracking, pull-out, and fiber bundle cracking at SD¼15 mm/needle.

H. Mei et al. / Ceramics International 42 (2016) 1762–17681766

channel and then the more deposited SiC substrate in the layer,which reduces the intermediate layer shear force. It suggeststhat the optimal SD can predominantly improve the CVIprocess of 2D C/SiCs. The associated damage micrographsderived at the LVI are presented in Fig. 7 to basically show thefiber fracture, axial cracking, slight debonding, less pull-outand fiber bundle cracking with decrease of SD.

The residual tensile strengths of specimens were calculated,which were obtained by a tensile test force and displacementcurve. The decreasing ratio of tensile strength of specimenswas defined, which was damaged relatively to the originalstrength. It can be calculated by the following formula:

r¼ sori�sdamð Þ=sori � 100% ð2Þ

Fig. 8. Stress–strain curves of the specimens after the LVI with differentstitched densities of SD¼5, 10, and 15 mm/needle.

H. Mei et al. / Ceramics International 42 (2016) 1762–1768 1767

where r is the decreasing ratio of tensile strength, sori is theoriginal strength, and sdam is the tensile strength of thedamaged specimen.

Statistical results including original or residual tensilestrength, modulus, and damage radius of the C/SiC specimenswith different SDs after the LVI are listed in Table 1. Table 1demonstrates that the initial tensile strengths of specimens (S1,S2, and S3) do not show tremendous variations, indicating thatthe SD has a slight influence on the initial tensile strengths ofspecimens. However, the SD has a dramatic influence on theresidual tensile strengths of specimens. To some extent, it isdiscovered that the specimens S1, S2, and S3 have the gooddamage resistance, with tensile strength after the LVI stillretaining 72.43%, 95.20%, and 91.49%, respectively. Inaddition, Table 1 also illustrates that the decreasing rate ofthe tensile strength of the specimens S1 is more 19% than thatof the S3. This can be explained by based on the followingargumentation. On the one hand, the tensile strength dependsmainly on the residual fibers to bear the load, while the SD of5 mm/needle has caused more serious damage to the fiberbundles than the SD 10 or 15 mm/needle. On the other hand,the stitching process will bring about stresses around thepinprick of 2D preform and give rise to matrix cracks aroundthe pinprick of specimens. The new matrix cracks of speci-mens initiate and propagate during the LVI process, especiallythe matrix cracks will connect for the short distances amongthe cracks and cracks propagation in S1. As a consequence thestrength retaining ratio of the specimens with higher SD issmaller than that of the specimens with lower SD.

As shown in Fig. 8, all tensile stress–strain curves exhibitnonlinear behavior. Fig. 8 illustrates that the higher the SD, thebigger the elastic modulus. Such nonlinear behavior can bequalitatively understood by the damage and fracture occurringin the composite under increasing tensile stresses, as intraditional cross-woven C/SiC composites [26]. As shown inTable 1, it indicates that the elastic modulus of the specimensS1, S2, and S3 after the LVI is 112.34, 110.89, and 79.80 GPa,respectively. The elastic modulus increases mainly because thedeposition of SiC matrix of specimens is influenced by theincreasing SD which improves the content of the SiC matrix.Because the elastic modulus of SiC matrix is higher than thatof the fiber bundle, the higher SD is, the more the SiC matrix

increase and the bigger elastic modulus of the specimenincreases.Firstly, the damages brought about by the LVI and

processing-induced cracks propagated with increasing tensilestress cause the tensile stress–strain to nonlinear. As the tensilestress increased, more processing-induced cracks propagated,and at the same time, new matrix cracks initiated especially atthe stitching sites, leading to the most nonlinear part. When thepressure was higher than 40 MPa, the number of matrix cracksformed to maximum of the tensile load was mainly borne bycarbon fibers, so the slope of the tensile stress–strain curvestended to be increased. Between 50 MPa and 80 MPa, thetensile load was mainly borne by carbon fibers, so the slope ofthe tensile stress–strain curves tended to be stable. Whatshould be noticed is that, in S1 (5 mm/needle), the ultimatetensile stress is minimum among the specimens before finalfailure of the composite due to fiber breaking occurred at thestress levels above 80 MPa (as shown in Fig. 7a).To further investigate the effect of stitching on tensile

behavior, the fracture surfaces were observed in detail. Fig. 7shows the tensile fracture surfaces of the tested specimens forthe stitched C/SiC composites. It can be seen that there areseveral similar characteristics in the fracture modes for onetypes of C/SiC composites with three SDs, namely, all testedspecimens appear the matrix cracks. However, the differencesamong them are also obvious. Firstly, severe fiber breaking areobserved for specimens S1 (as shown in Fig. 7a), while slightfiber breaking is found in the other two composites (as shownin Fig. 7b–f), indicating that the SD is so large that the stitchescould cause serious damage to fiber. Secondly, slight pull-outsand debondings are observed on the fracture surfaces of the S2and S3 whereas raduial fractures are discovered on the fracturesurfaces of the S2. Therefore, the decreasing ratio of the tensilestrength of specimens with the moderated SD is the least thanthat of specimens with low or high SD. It indicates that SD hasa significant impact on the residual tensile strength and fracturemorphologies of C/SiCs.

4. Conclusions

The results showed that the SD had a significant effect onthe CVI process, residual mechanical properties and fracturemorphologies of C/SiCs. The real damage radius of 2D C/SiCsmeasured by thermography reduced with the increase of SD,however, the residual tensile strength of specimens with theSD of 10 mm/needle was highest. The real damage radius RIR

containing the invisible delamination was actually much largerthan the apparent damage radius ROB. Compared to the surfacedamage of the same specimens, the real damage of specimenswas serious. The 2D C/SiCs with the SD of 5, 10, and 15 mm/needle had the good damage resistance, with tensile strengthstill retaining 72.43%, 95.20%, and 91.49% after the LVI,respectively. In addition, the decreasing ratio of the tensilestrength of specimens with the high SD is more 19% than thatof specimens with low SD. Consequently, while the higher SD

H. Mei et al. / Ceramics International 42 (2016) 1762–17681768

could effectively reduce the delamination of 2D C/SiCs duringthe impact process, it had caused more serious damage to thefiber bundles and a lot of micro-cracks during 2D preform'preparation, resulting in the reduction of residual tensilestrength of specimens.

Acknowledgments

This work has been financially supported by NationalNatural Science Foundation of China (51272210 and50902112), Program for New Century Excellent Talents inUniversity (NCET-13-0474), Foreign Talents Introduction andAcademic Exchange Program of China (B08040), and North-western Polytechnical University (NPU) Foundation for Fun-damental Research (NPU-FFR-JC201135).

References

[1] K. Ding, Y.C. Fu, H.H. Su, Y. Chen, X.Z. Yu, G.Z. Ding, Experimentalstudies on drilling tool load and machining quality of C/SiC compositesin rotary ultrasonic machining, J. Mater. Process. Technol. 214 (12)(2014) 2900–2907.

[2] L.T. Zhang, L.F. Cheng, Discussion on strategies of sustainable devel-opment of continuous fibre reinforced ceramic matrix composites, ActaMater. Compos. Sin. 24 (2) (2007) 1–6 in Chinese.

[3] M.S. Liu, Y.L. Li, F. Xu, et al., Dynamic compressive mechanicalproperties and a new constitutive model of 2D C/SiC composites, Mater.Sci. Eng. A 489 (2008) 120–126.

[4] X. Chen, R. Wang, N. Yao, A.G. Evans, J.W. Hutchinson, R.W. Bruce,Foreign object damage in a thermal barrier system: mechanisms andsimulations, Mater. Sci. Eng. A352 (2003) 221–231.

[5] S.M. Wiederhorn, M.K. Ferber, Silicon nitride for gas turbines, Curr.Opin. Solid State Mater. Sci. 5 (2001) 311–316.

[6] M.S. Found, J.R. Lamb, Damage assessment of impacted thin CFRPpanels, Composites A 35 (2004) 1039–1047.

[7] K. Ogi, T. Okabe, M. Takahashi, S. Yashiro, A. Yoshimura,T. Ogasawara, Experimental characterization of high-velocity impactdamage behavior in a three-dimensional woven SiC/SiC composite,Composites A 41 (4) (2010) 489–498.

[8] M.O.W. Richardson, M.J. Wisheart, Review of low-velocity impactproperties of composite materials, Composites A 27A (1996) 1123–1131.

[9] V. Herb, E. Martin, G. Couégnat, Damage analysis of thin 3D-wovenSiC/SiC composite under low velocity impact loading, Composites A 43(2012) 247–253.

[10] X.Y. Tong, L.J. Yao, Z.S. Li, Q.Y. Cheng, Damage behavior of 2D C/SiC composites under low velocity impact, J. Inorg. Mater. 3 (2010/25)311–314.

[11] R. Knoche, A. Dröse, Structural integrity of low-velocity impacted C/SiCceramic matrix composites, in: Proceedings of the 12th EuropeanConference on Spacecraft Structures, Materials and EnvironmentalTesting, European Space Agency, (Special Publication) ESA SP, March20, 2012–March 23, 2012.

[12] Z. Aslan, R. Karakuzu, B. Okutan, The response of laminated compositeplates under low-velocity impact loading, Compos. Struct. 59 (2003)119–127.

[13] S.X. Wang, L.Z. Wu, L. Ma, Low-velocity impact and residual tensilestrength analysis to carbon fibre composite laminates, Mater. Des. 31(2010) 118–125.

[14] F. Mili, B. Necib, Impact behavior of cross-ply laminated compositeplates under low Velocities, Compos. Struct. 51 (2001) 237–244.

[15] B.V. Sankar, H.S. Zhu, The effect of stitching on the low-velocity impactresponse of delaminated composite beams, Compos. Sci. Technol. 60(2000) 2681–2691.

[16] S.K. Sharma, B.V. Sankar, Effects of Through the Thickness Stitching onImpact and Inter Laminar Fracture Properties of Textile Graphite/epoxyLaminates, National Aeronautics and Space Administration, Washington(DC), 1995.

[17] L.J. Yao, Z.Y. Feng, Q.Y. Cheng, Low velocity impact damageevaluation of 2D C/SiC composite material, Adv. Mater. Res. 79–82(2009) 1835–1838.

[18] S.S. Li, X.Y. Tong, L.J. Yao, B. Li, Validity of several NDT methods onlow velocity impact damage evaluation of 2-dimensional plain woven C/SiC composites, Adv. Mater. Res. 581–582 (2012) 764–767.

[19] H. Mei, L.F. Cheng, Comparison of the mechanical hysteresis of carbon/ceramic-matrix composites with different fibre preforms, Carbon 47(2009) 1034–1042.

[20] H. Mei, L.F. Cheng, Q.Q. Ke, L.T. Zhang, High-temperature tensileproperties and oxidation behavior of carbon fibre reinforced siliconcarbide bolts in a simulated re-entry environment, Carbon 48 (11) (2010)3007–3013.

[21] ASTM International, ASTM D7136-05, 2005.[22] ASTM C1275-00, Standard Test Method for Monotonic Tensile Behavior

of Continuous Fiber-reinforced Advanced Ceramics with Solid Rectan-gular Cross-section Test Specimens at Ambient Temperature.

[23] H. Mei, Current development in non-destructive testing and damageevaluation for ceramic matrix composites, Adv. Appl. Ceram. 108 (2)(2009) 84–91.

[24] H. Mei, Y.D. Xu, L.F. Cheng, L.T. Zhang, Nondestructive evaluation andmechanical characterization of a defect-embedded ceramic matrix com-posite laminate, Int. J. Appl. Ceram. Technol. 4 (4) (2007) 378–386.

[25] H. Mei, L.D. Zhang, H.R. Xu, L.F. Cheng, Damage mechanism of acarbon fibre ceramic composite during the step-loading indentation andits effect on the mechanical properties, Composites B 56 (2014) 142–148.

[26] M. Wang, C. Laird, Characterization of microstructure and tensilebehavior of a cross-woven C/SiC composite, Acta Mater. 44 (4) (1996)1371–1387.