Materials and Design 143 (2018) 81–92

Contents lists available at ScienceDirect

Materials and Design

j ourna l homepage: www.e lsev ie r .com/ locate /matdes

Enhancing mode-I and mode-II fracture toughness of epoxy and carbonfibre reinforced epoxy composites using multi-walled carbon nanotubes

Dong Quana, Josu Labarga Urdánizb, Alojz Ivankovića,*a School of Mechanical and Materials Engineering, University College Dublin, Irelandb ETSI Caminos, Canales y Puertos, Universidad Politécnica de Madrid, Madrid, Spain

H I G H L I G H T S

• MWCNTs were used to enhanceepoxy and carbon fibre composites.

• Mode-I and Mode-II fracturebehaviour was studied.

• The addition of MWCNTs moderatelyincreased Mode-I fracture toughness.

• The addition of MWCNTs significantlyincreased Mode-II fracture toughness.

G R A P H I C A L A B S T R A C T

A R T I C L E I N F O

Article history:Received 17 December 2017Received in revised form 25 January 2018Accepted 25 January 2018Available online 31 January 2018

Keywords:Multi-walled carbon nanotubesEpoxyCarbon fibre reinforced epoxy compositesFracture toughness in mode-I and mode-II.

A B S T R A C T

Multi-walled carbon nanotubes (MWCNTs) were added to an epoxy resin in an effort to improve the fracturetoughness of bulk epoxy and also when used as matrix for carbon fibre reinforced epoxy composites (CFRPs).The incorporation of MWCNTs to bulk epoxy and CFRPs moderately increased the mode-I fracture energy,and significantly increased the mode-II fracture energy, i.e. the average mode-II fracture energy of CFRPsincreased from 2026 J/m2 to 3406 J/m2 due to the addition of 0.5 wt% MWCNTs, and further to 5491 J/m2 dueto the addition of 1 wt% MWCNTs. The superior toughening performance of MWCNTs in mode-II fractureis attributed to two reasons: 1) increased MWCNT breaking and crack deflection mechanisms under shearload; and 2) large fracture process zone accompanied with extensive hackle markings and micro-cracksahead of the mode-II crack tip of CFRPs, which resulted in significant number of MWCNTs contributing totoughening mechanisms.

© 2018 Elsevier Ltd. All rights reserved.

1. Introduction

The application of carbon fibre reinforced plastics (CFRPs) hasexpanded extensively in a wide range of industries, including auto-motive, aerospace and wind energy. Epoxies are used most widely

* Corresponding author.E-mail address: alojz.ivankovic@ucd.ie (A. Ivanković).

as the matrices for CFRPs due to their favourable engineering prop-erties, such as high modulus, high strength, low creep and excel-lent thermal stability. However, the highly cross-linked structureof epoxies results in inherently low fracture toughness and hencepoor resistance to fracture. As a consequence, CFRPs possess rel-ative low interlaminar fracture toughness. Blending second phasemodifiers, such as silica particles [1,2], rubber particles [3,4], car-bon nanotubes [5–7] and graphene [8,9], to epoxies was reported tobe a prevalent method of improving the fracture toughness. Further

https://doi.org/10.1016/j.matdes.2018.01.0510264-1275/© 2018 Elsevier Ltd. All rights reserved.

https://doi.org/10.1016/j.matdes.2018.01.051http://www.ScienceDirect.com/http://www.elsevier.com/locate/matdeshttp://crossmark.crossref.org/dialog/?doi=10.1016/j.matdes.2018.01.051&domain=pdfmailto:alojz.ivankovic@ucd.iehttps://doi.org/10.1016/j.matdes.2018.01.051

82 D. Quan et al. / Materials and Design 143 (2018) 81–92

developments of incorporating hybrid rubber-silica nanoparticlesto epoxy have been made in [10,11] with considerable success inenhancing the fracture toughness.

The superior mechanical properties of carbon nanotubes (CNTs)make them attractive candidates as toughening agents of epoxyresins and fibre reinforced epoxy composites. Tang et al. [12,13]reported that blending 1 wt% MWCNTs into an epoxy increased themode-I fracture energy by 56%, while a concurrent improvement inYoung’s modulus was also observed. Cha et al. [14] presented thatthe addition of 1 wt% poly(4-aminostyrene) functionalized CNTs toan epoxy increased the Young’s modulus and tensile strength from2.76 GPa and 61.51 MPa to 3.89 GPa and 82.57 MPa, respectively. Inanother study, Cha et al. [15] employed noncovalently functional-ized carbon nanotubes to enhance the epoxy nanocomposites. Themost significant improvements of Young’s modulus, ultimate tensilestrength and fracture toughness were reported to be 64%, 22% and95–100%, respectively. Saboori et al. [16] measured mode-I, mode-IIIand mixed-mode fracture toughness of MWCNT/epoxy nanocom-posites. It was reported that the mode-I fracture energy increasedsteadily by 27 wt% as the MWCNT content increased to 0.5 wt%,and then decreased slightly for 1 wt% MWCNTs. More pronouncedimprovement was achieved for mode-III and mix-mode fracturetoughness, with no down-ward trend observed as the MWCNTincreased to 1 wt%. It is noteworthy that adding third phase mod-ifiers, such as nanoclay [17,18], to CNT modified epoxies demon-strates a promising method to further enhance the mechanical andfracture properties. In general, the addition of a small amount ofMWCNTs could moderately increase the mode-I fracture toughness.However, limited work [5] has been performed to study the effectof CNTs on mode-II fracture behaviour of epoxies and the fracturemechanisms are not yet fully understood.

CNTs were normally introduced into fibre reinforced epoxy com-posites (FRPs) either by adding CNT/fibre interleaves between thelaminates [19–21], or by grafting/growing CNTs directly on the car-bon fibres [22–24]. Xu et al. [19] employed continuous carbon nan-otube film as interleave to enhance CFRPs. The flexural strength andinterlaminar shear strength were increased by 16% and 25%, respec-tively, for adding 0.22 wt% CNT film. Zheng et al. [20] fabricated sand-wiched carbon nanotube/polysulfone nanofiber (CNTs/PSF) papersas interleaves to improve the interlaminar fracture toughness ofCFRPs. It was reported that adding 10% CNT/PSF interleaves toCFRPs increased the mode-I and mode-II fracture toughness by 53%and 34%, respectively. Additionally, the flexural strength and flex-ural modulus were improved by 27% and 29%, respectively. Zhouet al. [21] evaluated the use of hierarchical carbon nanotube-shortcarbon fibre (CNT-SCF) as interleaves on the interlaminar fracturetoughness of CFRPs. Increases of 125% and 98% of the fracture tough-ness, compared to the control material, were achieved by adding1.0 and 2.0 mg/cm2 CNT-SCF interleaves. Davis and Whelan [22]managed to deposit fluorine functionalized CNTs on the mid-planeof fibre reinforced epoxy composites. It is found that depositing0.5 wt% CNTs increased the mode-II fracture propagation energyfrom 1906 J/m2 of the control to 2419 J/m2. In another work [23],CNTs were grown in-situ on carbon fibres using a flame synthesismethod. The interfacial shear strength of the CFRPs was increased byapprox. 70% after the CNTs were grown for only 3 min. Based on theliterature review, it is clear that the integration of a small amountof CNTs in the mid-plane of FRPs could significantly increase thefracture toughness. However, limited work employed CNT modifiedepoxies as matrices of CFRPs to study the fracture behaviour of suchcomposite laminates. Also, there is a lack of studies on the mode-Iand mode-II fracture mechanisms of CFRPs based on CNT modifiedepoxy.

In the present work, MWCNTs were blended in an epoxy resinto enhance the fracture toughness. Such MWCNT modified epoxieswere then used as matrices to manufacture CFRPs. The effects of

MWCNTs on the mechanical and fracture properties of bulk epoxyand CFRPs were investigated.

2. Experimental

2.1. Materials

The epoxy (CYCOM 890 RTM) is a commercially available, one-part liquid resin system supplied by Cytec Solvay Group. This resinsystem possesses high viscosity at room temperature. However, itcould achieve a low viscosity of 250 cps and stay below 350 cps for24 h at 80 ◦C. The MWCNTs were obtained in powder form fromGraphene Supermarket, USA. They have an average outer diameterof 50–85 nm and a length of 10–15 lm. They appear in an entan-gled cotton-like form. These MWCNTs were not functionalized andwere used as-received. The carbon fibres are biaxial non-crimp fabric(Toray T700Sc 50C), provided by Saertex GmbH, Germany.

2.2. Preparation of MWCNT modified epoxies

High shear mixing process was used to disperse the MWCNTs inthe epoxy. This process has been reported to be able to achieve goodMWCNT dispersion [7,12,25]. The MWCNT/epoxy mixture was firstlypre-mixed using an IKA RW20 digital mixer operating at 2000 rpmfor 2 h at 80 ◦C. The low viscosity of the mixture at 80 ◦C allowed aneasier processing. Then, a shear mixer (Silverson L4RT) was used forhigh-shear mixing at 3000 rpm initially at 80 ◦C with the tempera-ture slowly reduced to 30 ◦C over 2 h. The mixture was subsequentlyshear mixed for another 2 h at approx. 30 ◦C. In this process, a vis-cous system at relative low temperature could generate sufficientshear force to effectively break up MWCNT agglomerates. Followingthe mixing, the mixture was degassed in a vacuum oven at 80 ◦C. Theconcentrations of MWCNTs were 0.5 wt% and 1 wt%.

A portion of the mixture was then cast into aluminium mouldsto manufacture the specimens for evaluating the mechanical andfracture properties of the bulk epoxy. The cure cycle consists of a 2-hour ramp from room temperature to 180 ◦C followed by a 2-hourhold, or dwell, at 180 ◦C. After the curing schedule, the samples wereallowed to cool down naturally to room temperature in the oven. Theremaining mixture was used as matrices to manufacture the CFRPs.

2.3. Fabrication of CFRPs

Due to the high aspect ratio (up to several thousands) of theMWCNTs, resin transfer moulding process is not suitable for manu-facturing the MWCNT enhanced CFRPs, i.e. the MWCNTs in the epoxyresin are not able to flow through multiple layers of fibres uniformlyand hence results in poor overall nanotube dispersion [26]. In thecurrent work, to tackle this problem, the carbon fibre fabrics wereimpregnated by the MWCNT modified epoxy manually. Detailed pro-cedure for manufacturing the CFRP specimens is described in thefollowing and schematically shown in Fig. 1.

(a, b) The epoxy resin and one ply of carbon fibre fabric with aweight ratio of 3/2 were placed into a preheated oven at80 ◦C for approx. 30 min. The low viscosity of the epoxyat 80 ◦C and high temperature of the carbon fibre fabricallowed for easier processing in the next step.

(c, d) The epoxy resin was poured onto the carbon fibre fabric andthen spread evenly on the fabric using a plastic scraper.

(e, f) The carbon fibre fabric was subsequently placed into thepreheated oven at 80 ◦C for additional 30 min. This allowedthe resin to impregnate the carbon fibre fabric.

(g) Eight plies of carbon fibre fabrics with fibre direction ofeither 0 ◦/90 ◦ or ±45 ◦ were prepared by repeating the pro-cedure (a–f). They were then laid-up in an order as shown

D. Quan et al. / Materials and Design 143 (2018) 81–92 83

80 C

Epoxy Carbon fibreOven

80 C

Oven

a b c d

ef

0 /90

± 45

± 45

0 /90

PTFE film

0 /90

± 45

± 45

0 /90

g

h

i

jk

Bottom Plate Release PlyApplied VacuumSealant Tape

PrepregBagging FilmBreathing Fabric

Bottom Plate Release PlyApplied VacuumSealant Tape

Prepreg Bagging film

Compressed Air at 80 psi

Applied 30 Tonne of Force (from hydraulic press)

Top Lid

Breathing Fabric

Load blocks

PTFE insert as initial crack

Fig. 1. Schematic of the fabrication of CFRP samples.

in Fig. 1 (g). A strip of PTFE release film with a thickness of12.5 lm was inserted between the fourth and fifth plies andserved as the initial crack.

(h) A debulking process (under vacuum for 45 min) was thenapplied to the prepreg layup to remove air pockets and toconsolidate the layup. An illustration of the layup proce-dure is shown in Fig. 1 (h).

(i) The prepreg was then cured inside an in-house pressclave,and the setup is shown in Fig. 1 (i). In the curing process,an internal pressure of 80 psi (approx. 5.5 bar) was appliedin the chamber from a compressed air supply line. A vac-uum was also applied to the base plate throughout thecure cycle. The cure cycle consists of a 2-hour ramp fromroom temperature to 180 ◦C followed by a 2-hour hold, ordwell, at 180 ◦C. After the curing schedule, the pressclavewas allowed to cool down naturally to 80 ◦C for approx. 4 hwhile keeping the layup under full pressure and vacuum.

(j, k) After the pressclave cooling down to room temperature,the composite panel was taken out and cut into requireddimensions for subsequent tests. Load blocks were thenadhesively attached to the end of each specimen wherethe crack initiator was located. The sides of each specimenwere painted with a thin layer of water-based correctionfluid (Tipp-Ex). Once the fluid had dried, vertical lines weredrawn on the side of the specimen for indicating cracklength.

2.4. Test procedure

Uniaxial tensile test was conducted to measure the Young’s mod-ulus and tensile strength of the bulk polymer according to BS ISO

527 Standard [27]. Dumbbell specimens with 25 mm gauge lengthand 5 mm × 4 mm cross-sectional area were machined from a curedplate. The tests were conducted at a loading rate of 0.5 mm/min atroom temperature (normally 20±1 ◦C). At least five replicate testswere conducted for each material.

Single edge notch three-point bend (3PB) test, see Fig. 2 (a), wasemployed to measure the Mode-I fracture toughness of the bulkepoxy according to ASTM D5045-99 standard [28]. The sharp pre-crack was introduced by tapping a liquid nitrogen-chilled razor bladeinto the bottom of the v-shape notch. The tests were conducted atroom temperature with a constant displacement rate of 1 mm/min.At least six replicate tests were conducted for each material.

Asymmetric four-point bend (A4PB) test [5], see Fig. 2 (b), wasused for determining the mode-II fracture toughness of the bulkepoxy. The tests were conducted at room temperature with a con-stant displacement rate of 1 mm/min. At least six replicate tests wereconducted for each material. A static equilibrium analysis of the A4PBconfiguration reveals that the shear force Q and the bending momentM at the crack plane can be written in terms of the load P as:

Q =P(L1 − L2)

L1 + L2and M = cQ (1)

when the crack tip is directly underneath the load, i.e. c = 0, thebending moment vanishes and the sample is in pure mode-II loading.The mode-II fracture toughness, KIIC, can be determined as:

KIIC =Q

BW1/2f(

aW

)(2)

84 D. Quan et al. / Materials and Design 143 (2018) 81–92

48mm

a 6mm

=6mm

P

W=12mm

B=6mm

(a) 3PB test

L1=24mm

W=12mma

B=6mm

=6mm

P

L2=12.5mm

12.5mm 24mm

c

(b) A4PB test

a0=45 mm

b=25 mm

h=6 mm

L=150 mm

P

P

(c) DCB test

a0=65 mm

b=25 mm

h=6 mm

LF=117.5 mm

P

L=170 mm

(d) ELS test

Fig. 2. Illustrations of 3PB, A4PB, DCB and ELS tests. The red lines in (a) and (b) indicate the sharp precrack. The green lines in (c) and (d) indicate the crack starter.

where f(a/W) is the geometry function expressed as [5]:

f(

aW

)= 9.763

(a

W

)4− 15.036

(a

W

)3+ 8.667

(a

W

)2

+1.695(

aW

)− 0.037

(a

W

)≤ 0.7 (3)

The fracture energy, GIIC, was calculated using the relation:

GIIC =K2IIC

E

(1 − m2

)(4)

where E is the Young’s modulus and m is the Poisson’s ratio of theepoxy.

Three-point bend flexure test was used to determine the flex-ural modulus and flexure strength of the CFRPs according tothe standard ISO: 14125:1998 [29]. The sample dimension was80 mm × 20 mm × 5.5 mm. The tests were conducted at room tem-perature with a constant displacement rate of 2 mm/min. Five sam-ples were tested for each material.

Mode-I double cantilever beam (DCB) test and mode-II endloaded split (ELS) test were carried out to measure the crackpropagation energy of the CFRPs according to the standards ISO:15024:2001 [30] and ISO: 15114:2014 [31], respectively. The testconfigurations are shown in Fig. 2 (c) and (d). The tests were con-ducted at room temperature with a constant displacement rate of2 mm/min. Five tests were repeated for each material. It should benoted that a 5 mm long sharp precrack was generated by loading thesamples under opening load.

In order to examine the length of the MWCNTs after the shearmixing process, a small amount of uncured epoxy/MWCNT mixturewas added to acetone to dissolve the epoxy matrix. The solution wasthen placed in a low energy ultrasonic bath for 1 h to disperse theMWCNTs. After that, a drop of solution containing remaining MWC-NTs was placed on a piece of tin foil. Scanning electron microscopeequipped with a field emission gun (SEM, FEI Quanta 3D) was usedto image the MWCNTs on the tin foil. The length of the MWCNTs on

the SEM images was measured using a Java-based image process-ing and analysis program called ImageJ. About 400 measurementswere performed. The dispersion of MWCNTs in the cured epoxywas studied using transmission optical microscope (TOM, NikonE80i(Orina)). The samples were ground and fine polished to thinsections of approximately 40 lm thickness, according to the tech-nique described in [32]. The fracture surfaces of the 3PB specimens,A4PB specimens, DCB specimens and ELS specimens were studiedusing SEM. The samples were gold sputter coated at a current of30 mA for 15 s to get a gold layer of approximately 5 nm.

3. Results and discussion

3.1. Length of the MWCNTs

Fig. 3 (a) presents a typical SEM micrograph of the remainingMWCNTs after removing the epoxy matrix, and the measured length-distribution of the MWCNTs is shown in Fig. 3 (b). The length ofthe MWCNTs was measured by dividing the curved MWCNTs into anumber of segments, as shown in the inset of Fig. 3 (b). It is found thatthe MWCNTs have been severely damaged from their initial lengthof 10–15 lm to an average length of 2.2 lm. This was also observedin literature [33], and is typical for brittle MWCNT fibres after shearmixing with polymers [33,34].

3.2. Mechanical properties

3.2.1. Bulk epoxyThe Young’s modulus and tensile strength of the bulk epoxies are

summarised in Table 1. A Young’s modulus of 3.20 GPa was mea-sured for the control. The addition of 0.5 wt% MWCNTs resulted in amoderate increase of the Young’s modulus to 3.42 GPa. This indicateseffective load transfer between the MWCNTs and the epoxy at theelastic deformation stage of the tensile test [35]. The Young’s modu-lus did not increase further for blending more MWCNTs (1 wt%) intothe epoxy. This is in agreement with the literature [7,35,36], and wasattributed to the increasing amount of MWCNT agglomerates. An

D. Quan et al. / Materials and Design 143 (2018) 81–92 85

5 µm

(a)

0

6

12

18

24

Fre

quency (

%)

Length of MWCNTs (µm)

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 7.00

(b)

Fig. 3. (a) SEM micrograph of MWCNTs collected after removing epoxy matrix, and (b) the length-distribution of MWCNTs and the inset is a schematic showing the measurementof a curved MWCNT.

indication of the dispersion characteristics of the MWCNTs is givenin Fig. 4. It presents TOM images of the MWCNTs dispersed in thecured epoxy. It is found that relatively good MWCNT dispersion wasachieved for the epoxy with 0.5 wt% MWCNTs, while considerableMWCNT agglomeration can be observed for the epoxy with 1 wt%MWCNTs.

The tensile strength of the control was measured to be 72.4 MPa,see Table 1. The incorporation of MWCNTs to the epoxy had negligi-ble effects on the tensile strength. While this behaviour is in agree-ment with the observation in the literature [7,36], others [14,15]presented a different trend, i.e. blending a small amount of MWC-NTs to epoxies considerably increased the tensile strength. This isdue to the varying interfacial adhesion between MWCNTs and epox-ies in different systems, i.e. effective load transfer between epoxymatrix and MWCNTs is required to active reinforcement potentialof MWCNTs. In the current work, the interfacial adhesion is suffi-cient to provide effective load transfer at the elastic deformationstage under loading, and hence an increase in the Young’s modu-lus was observed for 0.5%MWCNTs. However, the retention of thetensile strength when adding MWCNTs to epoxy indicates that theinterfacial adhesion is not strong enough to maintain effective loadtransfer at the failure stage of the tensile test. Surface modification ofMWCNTs demonstrates a promising way to achieve good CNT/epoxyinterfacial adhesion [14,37,38], and hence, to achieve effective rein-forcement in both tensile modulus and strength. It is important tonotify that the mechanical properties of CNT modified epoxies alsodepend on the dispersion of the CNTs.

3.2.2. CFRPsThe flexural properties of the CFRPs are given in Table 2. The

flexural modulus and flexural strength of the control were mea-sured to be 48.6 GPa and 505.7 MPa, respectively. Interestingly, theaddition of a small amount of MWCNTs enhanced the mechanicalproperties of CFRPs more efficiently than that of the bulk epoxy.The flexural modulus and flexural strength of the CFRPs increased to54.6 GPa and 520.1 MPa, respectively, due to the addition of 0.5 wt%MWCNTs, and further to 57.6 GPa and 543.0 MPa, respectively, dueto the addition of 1 wt% MWCNTs. Similar trend was also reported

Table 1Mechanical properties of the epoxy.

Sample Young’s modulus (GPa) Tensile strength (MPa)

Control 3.20 ± 0.07 72.4 ± 1.60.5 wt% MWCNTs 3.42 ± 0.11 71.2 ± 1.51 wt% MWCNTs 3.37 ± 0.06 72.1 ± 2.1

by Zheng et al. [20], and the favourable effect on the flexural prop-erties of CFRPs can be explained in terms of an interaction betweenMWCNTs and carbon fibres under bending load.

3.3. Fracture properties

3.3.1. Bulk epoxyFig. 5 shows the measured mode-I and mode-II fracture energy

of the bulk epoxies. The mode-I fracture energy, GIC, of the con-trol was measured to be 118.5 J/m2. The value of GIC was increasedto 136.5 J/m2 due to the addition of 0.5 wt% MWCNTs, and furtherto 161.1 J/m2 due to the addition of 1 wt% MWCNTs. A more pro-nounced enhancement was achieved for the mode-II fracture energy,i.e. GIIC increased from 177.2 J/m2 of the control to 283.6 J/m2 for0.5 wt% MWCNTs, and further to 411.8 J/m2 for 1 wt% MWCNTs.Extensive research, such as [7,35,39,40], has reported similar trendon the mode-I fracture toughness of MWCNT reinforced polymers,i.e. the addition of a small amount of MWCNTs moderately increasedthe mode-I fracture toughness. However, to our best knowledge, verylimited study has been performed on the mode-II fracture of MWCNTmodified polymers, to date.

3.3.2. Fractographic studies: bulk epoxyTypical SEM images of the fracture surfaces of the 3PB specimens

are shown in Fig. 6. The dashed line indicates the front of the precrackand the arrow indicates the crack growth direction. A very smoothfracture surface is observed for the control, as shown in Fig. 6 (a). Thefracture surface of MWCNT modified epoxy appears slightly rougherthan the control (Fig. 6 (b)). Fig. 6 (c) and (d) compare the fracturesurfaces of the control and the 1 wt% MWCNT modified epoxy athigher magnification. While a mirror-like surface is observed for thecontrol, a large number of MWCNT segments along with many holes(with approx. the same diameter as the MWCNTs) and scratchinglines can be seen on the fracture surfaces of the 1 wt% MWCNT mod-ified epoxy. It is worthy noticing that both long and short MWCNTsegments are observed on the fracture surface of the 1 wt% MWCNTmodified epoxy, see Fig. 6 (d). Similar structures have been previ-ously reported in the literature [7,12,13,35] and were identified tobe formed by two different failure mechanisms: immediate fractureof the MWCNTs once the crack front passes it (short segments); andMWCNT pull-out (long segments). Hence, these MWCNT segmentsare the evidence of nanotube breaking and bridging (nanotube pull-out) mechanisms of the MWCNTs, and the scratching lines wereaccompanied with the crack deflection mechanism of the MWC-NTs [7,35]. All these mechanisms contributed to the increase of themode-I fracture energy.

86 D. Quan et al. / Materials and Design 143 (2018) 81–92

200 µm

(a) 0.5 wt.% MWCNTs

200 µm

(b) 1 wt.% MWCNTs

Fig. 4. Dispersion of the MWCNTs in cured epoxy.

Fig. 7 presents typical SEM images of the fracture surfaces of theA4PB specimens. A large number of river-lines are observed on thefracture surfaces of both the control and the 1 wt% MWCNT modi-fied epoxy (Fig. 7 (a) and (b)). The river-lines are more dense on thefracture surface of the 1 wt% MWCNT modified epoxy. This mightbe attributed to the crack deflection of the MWCNTs. Fig. 7 (c) and(d) presents the zoom-in of the positions between the river-lines inFig. 7 (a) and (b). A smooth surface is observed for the control, asexpected. The fracture surface of the 1 wt% MWCNT modified epoxyhas a much rougher appearance, identified with numerous MWCNTsegments and micro-crack marks. Hence, the toughening mecha-nisms of MWCNTs in mode-II fracture are determined to be crackbridging, nanotube breaking and crack deflection. These are similarto the detected toughening mechanisms in mode-I fracture.

Although similar toughening mechanisms are observed in mode-I and mode-II fracture, the toughness enhancement of epoxy foradding MWCNTs was much more pronounced in mode-II than inmode-I (see Fig. 5). For the MWCNT segments on the fracture sur-faces, the short ones were generated by the breaking of the MWCNTsas the crack front passes it, while the long ones were formed by thepull-out of MWCNTs [7,35]. When comparing Figs. 6 (d) and 7(d),it is clear that only a small proportion of short MWCNT segmentsare observed on the mode-I fracture surface (all the others are long),while the majority of the MWCNT segments on the mode-II fracturesurface are very short. Moreover, a large number of holes generatedby MWCNT pull-out are observed on Fig. 6 (d), while only limited ofthem present on Fig. 6 (d). Hence, the majority of the MWCNTs werepulled out (crack bridging mechanisms) during the mode-I fractureprocess, while most of the MWCNTs fractured (nanotube breakingmechanism) during the mode-II fracture process. This might be dueto the presence of higher stress in shear loading of mode-II frac-ture [41]. Given the superior mechanical properties of MWCNTs, it isnot surprising that the nanotube breaking mechanism is more effec-tive in preventing the crack growth and deflecting the crack paththan the crack bridging mechanism. Hence, the addition of MWCNTsis more effective for mode-II toughness enhancement.

3.3.3. CFRPsR-curves of the mode-I DCB tests and the mode-II ELS tests of all

the CFRPs are shown in Fig. 8. No significant R-curve behaviour isobserved for the DCB tests of the control and the MWCNT modified

Table 2Flexural properties of the CFRPs.

Sample Flexural modulus (GPa) Flexural strength (MPa)

Control 48.6 ± 3.4 506 ± 160.5 wt% MWCNTs 54.6 ± 1.3 520 ± 201 wt% MWCNTs 57.6 ± 1.5 543 ± 4

CFRPs, i.e. the mode-I fracture energy remains essentially constantalong the crack length. The average mode-I fracture energy, GIC, ofthe CFRPs is given in Table 3. The value of GIC was measured to be427 J/m2 for the control. It was increased to 462 J/m2 due to the addi-tion of 0.5 wt% MWCNTs, and further to 537 J/m2 due to the additionof 1 wt% MWCNTs.

As shown in Fig. 8 (b), ‘rising’ R-curve behaviour is observedfor the ELS tests of all the CFRPs, i.e. the fracture energy increasedsteadily with the crack length. This can be caused by an increasingdamage zone ahead of the crack tip [42]. The average value, GAvgIIC ,minimum value, GMinIIC , and maximum value, G

MaxIIC , of the R-curves are

presented in Table 3. It is found that the incorporation of MWCNTsdramatically enhanced the mode-II fracture toughness of the CFRPs.GAvgIIC increased from 2026 J/m

2 of the control to 3406 J/m2 of the0.5% MWCNTs modified CFRP, and further to 5491 J/m2 of the 1 wt%MWCNTs modified CFRP.

3.3.4. Fractographic studies: CFRPsFig. 9 presents typical micrographs of the fracture surfaces of

the DCB specimens. A number of broken fibres are observed on thefracture surfaces of both the control and the 1 wt% MWCNT mod-ified CFRPs, as shown in Fig. 9 (a) and (b). All the other fibres arewell attached to the epoxy matrix. Fig. 9 (c) and (d) presents micro-graphs of the broken fibres and the surrounding matrix under highermagnification. It is clear that some epoxy matrix is still attachedto the fibres. All these observations demonstrate good interfacialadhesion between the matrix and the fibres, and the main toughen-ing mechanisms of the carbon fibres in CFRPs are fibre peeling andfibre breaking. Another observation from Fig. 9 (a) and (b) is thatthere exist numerous flake-like fracture features in the interstitial

0

100

200

300

400

500

Control 0.5wt.% MWCNTs 1wt.% MWCNTs

Fra

ctu

re E

nerg

y (

J/m

2) GIIC

GIC

Fig. 5. Fracture energy of the bulk epoxies.

D. Quan et al. / Materials and Design 143 (2018) 81–92 87

Precrack

100 µm

(a) Control

Precrack

100 µm

(b) 1 wt.% MWCNTs

5 µm

(c) Control

5 µm

Long MWCNT segments

Short MWCNT segments

(d) 1 wt.% MWCNT

Fig. 6. Typical SEM micrographs of the fracture surfaces of 3PB specimens for the control and the 1 wt% MWCNTs modified epoxy. The dashed line indicates the tip of the precrack.The arrow indicates the crack growth direction.

matrix between the carbon fibres for both the control and the 1 wt%MWCNT modified CFRPs. However, it is more extensive for the 1 wt%MWCNT modified composites, which resulted in a rougher fracturesurface. The micrographs of the interstitial matrix between the fibresare shown in Fig. 9 (e) and (f). Apparently, the fracture surface ofthe 1 wt% MWCNT modified CFRPs appears rougher than that of thecontrol, and a large number of MWCNT segments (generated by nan-otube breaking and pull-out) are observed on the fracture surface of1 wt% MWCNT modified CFRP composite.

Fig. 10 shows typical micrographs of the fracture surfaces of themode-II ELS specimens as a representative of the unmodified andMWCNT modified CFRPs. One can see that all the fibres remained onthe fracture surface of the bottom part of the ELS specimens, leav-ing corresponding groves on the top part. In the rest of this work, thefracture surface of the bottom part of the ELS specimens is referredto as the male surface, and the fracture surface of the top part isreferred to as the female surface. On the male surfaces, there is noevidence of fibre breaking, and all the fibres are well attached tothe matrix. Hence, the main toughening mechanism of the carbonfibres in mode-II fracture is fibre peeling off in shearing mode. It isalso found that the female surface appears much rougher than themale surface for having extensive hackle markings in the interstitialmatrix between the carbon fibres. The hackle markings were createdunder the shearing deformation in the fracture process and they arealigned in a direction of approx. 45 ◦ to the fracture surface.

Fig. 11 presents SEM images of the male and female fracture sur-faces of the ELS specimens. By comparing Fig. 11 (a) and (b) to (c) and(d), it is found that the incorporation of MWCNTs into CFRPs signifi-cantly increased the roughness of both the male and female fracturesurfaces. Images of the hackle markings with higher magnification

are presented in Fig. 11 (e) and (f). Apparently, the surfaces of thehackle markings are very smooth for the control. This is typical forbrittle epoxies. Contrarily, there exists extensive micro cracks in thehackle markings for the 1 wt% MWCNT modified CFRPs. These microcracks were created by the crack deflection mechanism due to thepresence of MWCNTs. Moreover, significant number of MWCNT seg-ments are observed on the fracture surfaces of the 1 wt% MWCNTmodified CFRPs, see Fig. 11 (f).

Based on the observations, the toughening mechanisms of theMWCNTs in CFRPs are suggested to be crack bridging, nanotubebreaking and crack deflection in both mode-I and mode-II fracture.This is identical to the toughening mechanisms observed in the bulkepoxy. However, the toughness enhancement of the CFRPs due to theaddition of MWCNTs was significantly more pronounced in mode-IIthan in mode-I, i.e. the incorporation of 1 wt% MWCNTs increased themode-II fracture energy by 171%, but only increased the mode-I frac-ture energy by 26%. There are two reasons for the different toughen-ing performance of MWCNTs in mode-I and mode-II fracture. Firstly,the MWCNTs are more effective to improve the fracture toughnessof the epoxy matrix under shear loading than under opening load-ing condition, as discussed in Section 3.3.2. Secondly, the length ofthe fracture process zone for mode-II fracture is much longer thanthat for mode-I fracture. Xie et al. [43] modelled the length of thefracture process zone of CFRPs in different fracture mode using acohesive zone model. It was reported that the length of the mode-IIfracture process zone is about 6 times of the mode-I fracture pro-cess zone. It is worth noticing that Xie et al. [43] also calculated thelength of the fracture process zone using various analytical models,and all of them indicated a much longer fracture process zone inmode-II fracture than in mode-I fracture. Fig. 12 schematically shows

88 D. Quan et al. / Materials and Design 143 (2018) 81–92

100 µm

(a) Control

100 µm

(b) 1 wt.% MWCNTs

5 µm

(c) Control

5 µm

(d) 1 wt.% MWCNTs

Fig. 7. Typical SEM micrographs of the fracture surfaces of A4PB specimens for the control and the 1 wt% MWCNTs modified epoxy. The red arrow indicates the crack growthdirection.

0

200

400

600

800

40 60 80 100 120

Fra

ctu

re E

nerg

y (

J/m

2)

Crack Length (mm)

Control

0.5wt.% MWCNT

1wt.% MWCNT

(a) Mode-I DCB test

0

2500

5000

7500

10000

65 70 75 80 85

Fra

ctu

re E

nerg

y (

J/m

2)

Crack Length (mm)

Control

0.5wt.% MWCNT

1wt.% MWCNT

(b) Mode-II ELS test

Fig. 8. R-curves for the fracture tests of CFRPs.

the fracture process zone ahead of the crack tip in the CFRPs undermode-I and mode-II fracture condition. The longer fracture processzone accompanied with extensive hackle markings and micro-crackscould include significantly more MWCNTs to introduce the tough-ening mechanisms. Hence, the incorporation of MWCNTs into CFRPsdramatically increased the mode-II fracture toughness, but onlyresulted in moderate increase in the mode-I fracture toughness.

Previous attempts to improve the interlaminar fracture tough-ness of CFRPs using CNTs have shown a variety of success. Zhenget al. [20] et al. used carbon nanotubes/polysulfone nanofiber(CNTs/PSF) paper as interlayer to enhance the interlaminar fracture

toughness of CFRPs and achieved 53% and 34% increases of themode-I and mode-II fracture toughness, respectively. Zhu et al. [44]reported 52–95 % improvement in the mode-I fracture toughness and

Table 3Fracture energy of the CFRP composite laminates.

Sample GIC (J/m2) GAvgIIC (J/m

2) GMinIIC (J/m2) GMaxIIC (J/m

2)

Control 428 ± 27 2026 ± 180 903 ± 312 3106 ± 2930.5 wt% MWCNTs 462 ± 25 3406 ± 334 2231 ± 528 4333 ± 2651 wt% MWCNTs 537 ± 34 5491 ± 373 4213 ± 260 6315 ± 446

D. Quan et al. / Materials and Design 143 (2018) 81–92 89

100 µm

Broken fibres

(a) Control

100 µm

Broken fibres

(b) 1 wt.%MWCNT

5 µm

Carbon Fibre

(c) Control

5 µm

Carbon Fibre

(d) 1 wt.% MWCNTs

5 µm

(e) Control

5 µm

(f) 1 wt.% MWCNTs

Fig. 9. Typical SEM micrographs of the fracture surfaces of the mode-I DCB specimens of the control and the 1 wt% MWCNTs modified CFRPs.

74-109% improvement in the mode-II fracture toughness of CFRPsdue to the incorporation of interlayer made of glycidyloxypropyl-trimethoxysilane and carbon nanotubes. Carbon fibre/epoxy inter-leaves were used to toughen CFRPs in [45]. It was found that themode-I and mode-II fracture energy had been improved by 26% and

47%, respectively. In current work, the mode-II fracture energy ofCFRPs was increased by 170% for adding 1% MWCNTs. This is morepronounced than the other studies. This might be due to the differ-ent methods of incorporating MWCNTs into the CFRPs. This resultedin different composite structures, as shown in Fig. 13. The structure

P

200 µm200 µm

Bottom-male surface Top-female surface

Fig. 10. Representative fracture surfaces of the mode-II ELS tests for the CFRPs. The arrows indicate the crack growth direction.

90 D. Quan et al. / Materials and Design 143 (2018) 81–92

30 µm

(a) Control-Male

30 µm

(b) Control-Female

30 µm

(c) 1 wt.% MWCNTs-Male

30 µm

(d) 1 wt.% MWCNTs-Female

5 µm

(e) Control-Female

5 µm

Micro cracks

(f) 1 wt.% MWCNT-Female

Fig. 11. Typical SEM micrographs of the fracture surfaces of mode-II ELS specimens for the control and the 1 wt% MWCNTs modified epoxy.

of the CFRPs in this work might result in more efficient interac-tion/load transfer between the MWCNTs and the carbon fibres, andsubsequently benefit the improvement of mechanical and fractureproperties.

4. Conclusions

The present work studies the effect of adding MWCNTs to anepoxy resin on the mechanical and fracture properties of bulk epoxyand when used as the matrix of CFRPs, with an emphasis on thefracture toughness and toughening mechanisms. A number of con-clusions can be drawn from the current work.

The addition of 0.5 wt% MWCNTs moderately increased theYoung’s modulus of the bulk epoxy. No further improvement in theYoung’s modulus is observed when adding 1 wt% MWCNTs due tothe increasing MWCNT agglomeration. The incorporation of MWC-NTs shows negligible effects on the tensile strength. The adhesion

between the MWCNTs and the epoxy is sufficient to create effec-tive load transfer at the elastic deformation stage (resulting in theincrease of the Young’s modulus), but insufficient to maintain effec-tive load transfer at the failure stage in the tensile test (resulting inthe retention of the tensile strength).

The flexural modulus and flexural strength of the CFRPs weresteadily increased from 48.6 GPa and 506 MPa of the control to57.6 GPa and 543 MPa of the 1 wt% MWCNT modified CFRPs, respec-tively. This demonstrates effective load transfer between the epoxy,the MWCNTs and the carbon fibres under bending load.

Blending MWCNTs into bulk epoxy slightly increased the mode-I fracture energy but significantly increased the mode-II fractureenergy. The toughening mechanisms of MWCNTs appeared to becrack bridging (nanotube pull-out), nanotube breaking and crackdeflection in both mode-I and mode-II fracture. However, moreintensive nanotube breaking mechanism accompanied with higherdensity of crack deflection were detected in the mode-II fracture.This resulted in more pronounced toughness enhancement.

Crack

fracture process zone

(a) Mode-I

Crack

hackle markings micro-cracks fracture process zone

(b) Mode-II

Fig. 12. Schematics of the form of the fracture process zone in CFRPs.

D. Quan et al. / Materials and Design 143 (2018) 81–92 91

Crack

MWCNTs

Carbon

fibres

(a) MWCNTs blended in matrix

CrackMWCNT

interlayer

(b) MWCNTs incorporated as interlayer

Fig. 13. Schematics of the form of MWCNT reinforced CFRPs in (a) this work and (b) the literature [20,44,45].

Similarly to the bulk epoxy, the fracture energy of CFRPsincreased slightly in mode-I fracture but dramatically in mode-IIfracture. The outstanding toughening performance of MWCNTs inmode-II fracture was attributed to two factors. Firstly, intensive nan-otube breaking and crack deflection mechanisms took place undershear load. Secondly, the large mode-II fracture process zone (accom-panied with extensive hackle markings and micro-cracks) of CFRPsresulted in considerable number of MWCNTs contributing to tough-ening mechanisms.

Acknowledgments

The authors gratefully acknowledge the financial support fromthe Irish Composites Centre. The carbon fibres and epoxy matrixwere supplied by Bombardier Aerospace (Belfast).

References

[1] D. Carolan, A. Ivankovic, A.J. Kinloch, S. Sprenger, A.C. Taylor, Toughening ofepoxy-based hybrid nanocomposites, Polymer 97 (2016) 179–190. https://doi.org/10.1016/j.polymer.2016.05.007.

[2] S. Rahmanian, A. Suraya, B. Roshanravan, R. Othman, A. Nasser, R. Zahari, E.Zainudin, The influence of multiscale fillers on the rheological and mechanicalproperties of carbon-nanotube silica-reinforced epoxy composite, Mater. Des.88 (2015) 227–235. https://doi.org/10.1016/j.matdes.2015.08.149.

[3] D. Quan, A. Ivankovic, Effect of core-shell rubber (CSR) nano-particles onmechanical properties and fracture toughness of an epoxy polymer, Polymer66 (2015) 16–28. https://doi.org/10.1016/j.polymer.2015.04.002.

[4] D. Quan, N. Murphy, A. Ivankovic, Fracture behaviour of epoxy adhesive jointsmodified with core-shell rubber nanoparticles, Eng. Fract. Mech. 182 (2017)566–576. https://doi.org/10.1016/j.engfracmech.2017.05.032.

[5] M.R. Ayatollahi, S. Shadlou, M.M. Shokrieh, Fracture toughness ofepoxy/multi-walled carbon nanotube nano-composites under bending andshear loading conditions, Mater. Des. 32 (4) (2011) 2115–2124. https://doi.org/10.1016/j.matdes.2010.11.034.

[6] T. Subhani, M. Latif, I. Ahmad, S.A. Rakha, N. Ali, A.A. Khurram, Mechanicalperformance of epoxy matrix hybrid nanocomposites containing carbon nan-otubes and nanodiamonds, Mater. Des. 87 (2015) 436–444. https://doi.org/10.1016/j.matdes.2015.08.059.

[7] D. Quan, D. Carolan, C. Rouge, N. Murphy, A. Ivankovic, Carbon nanotubes andcore-shell rubber nanoparticles modified structural epoxy adhesives, J. Mater.Sci. 52 (8) (2017) 4493–4508. https://doi.org/10.1007/s10853-016-0695-9.

[8] B. Ahmadi-Moghadam, M. Sharafimasooleh, S. Shadlou, F. Taheri, Effectof functionalization of graphene nanoplatelets on the mechanical response ofgraphene/epoxy composites, Mater. Des. 66, Part A (2015) 142–149. https://doi.org/10.1016/j.matdes.2014.10.047.

[9] P. Li, Y. Zheng, M. Li, T. Shi, D. Li, A. Zhang, Enhanced toughness andglass transition temperature of epoxy nanocomposites filled with solvent-freeliquid-like nanocrystal-functionalized graphene oxide, Mater. Des. 89 (Supple-ment C) (2016) 653–659. https://doi.org/10.1016/j.matdes.2015.09.155.

[10] D. Quan, R.A. Pearson, A. Ivankovic, Interaction of toughening mechanismsin ternary nanocomposites, Polym. Compos. (2017) in press. https://doi.org/10.1002/pc.24368.

[11] D. Carolan, A. Ivankovic, A.J. Kinloch, S. Sprenger, A.C. Taylor, Tough-ened carbon fibre-reinforced polymer composites with nanoparticle-modifiedepoxy matrices, J. Mater. Sci. 52 (3) (2017) 1767–1788. https://doi.org/10.1007/s10853-016-0468-5.

[12] L.-C. Tang, Y.-J. Wan, K. Peng, Y.-B. Pei, L.-B. Wu, L.-M. Chen, L.-J. Shu,J.-X. Jiang, G.-Q. Lai, Fracture toughness and electrical conductivity of epoxy

composites filled with carbon nanotubes and spherical particles, Compos. A:Appl. Sci. Manuf. 45 (Supplement C) (2013) 95–101. https://doi.org/10.1016/j.compositesa.2012.09.012.

[13] L.-C. Tang, H. Zhang, X.-P. Wu, Z. Zhang, A novel failure analysis of multi-walled carbon nanotubes in epoxy matrix, Polymer 52 (9) (2011) 2070–2074.https://doi.org/10.1016/j.polymer.2011.03.002.

[14] J. Cha, S. Jin, J.H. Shim, C.S. Park, H.J. Ryu, S.H. Hong, Functionalization ofcarbon nanotubes for fabrication of CNT/epoxy nanocomposites, Mater. Des. 95(2016) 1–8. https://doi.org/10.1016/j.matdes.2016.01.077.

[15] J. Cha, G.H. Jun, J.K. Park, J.C. Kim, H.J. Ryu, S.H. Hong, Improvement of mod-ulus, strength and fracture toughness of CNT/epoxy nanocomposites throughthe functionalization of carbon nanotubes, Compos. Part B Eng. 129 (2017)169–179. https://doi.org/10.1016/j.compositesb.2017.07.070.

[16] B. Saboori, M.R. Ayatollahi, Experimental fracture study of MWCNT/epoxynanocomposites under the combined out-of-plane shear and tensile load-ing, Polym. Test. 59 (2017) 193–202. https://doi.org/10.1016/j.polymertesting.2017.01.028.

[17] M.N. Salimi, M.T. Merajin, M.K.B. Givi, Enhanced mechanical proper-ties of multifunctional multiscale glass/carbon/epoxy composite reinforcedwith carbon nanotubes and simultaneous carbon nanotubes/nanoclays,J. Compos. Mater. 51 (6) (2017) 745–758. https://doi.org/10.1177/0021998316655201.

[18] M. Hosur, T.H. Mahdi, M.E. Islam, S. Jeelani, Mechanical and viscoelasticproperties of epoxy nanocomposites reinforced with carbon nanotubes, nan-oclay, and binary nanoparticles, J. Reinf. Plast. Compos. 36 (9) (2017) 667–684.https://doi.org/10.1177/0731684417691365.

[19] H. Xu, X. Tong, Y. Zhang, Q. Li, W. Lu, Mechanical and electrical propertiesof laminated composites containing continuous carbon nanotube film inter-leaves, Compos. Sci. Technol. 127 (2016) 113–118. https://doi.org/10.1016/j.compscitech.2016.02.032.

[20] N. Zheng, Y. Huang, H.-Y. Liu, J. Gao, Y.-W. Mai, Improvement of inter-laminar fracture toughness in carbon fiber/epoxy composites with carbonnanotubes/polysulfone interleaves, Compos. Sci. Technol. 140 (2017) 8–15.https://doi.org/10.1016/j.compscitech.2016.12.017.

[21] H. Zhou, X. Du, H.-Y. Liu, H. Zhou, Y. Zhang, Y.-W. Mai, Delamination tough-ening of carbon fiber/epoxy laminates by hierarchical carbon nanotube-shortcarbon fiber interleaves, Compos. Sci. Technol. 140 (2017) 46–53. https://doi.org/10.1016/j.compscitech.2016.12.018.

[22] D.C. Davis, B.D. Whelan, An experimental study of interlaminar shear frac-ture toughness of a nanotube reinforced composite, Compos. Part B Eng. 42 (1)(2011) 105–116. https://doi.org/10.1016/j.compositesb.2010.06.001.

[23] X. Du, F. Xu, H.-Y. Liu, Y. Miao, W.-G. Guo, Y.-W. Mai, Improving theelectrical conductivity and interface properties of carbon fiber/epoxy compos-ites by low temperature flame growth of carbon nanotubes, RSC Adv. 6 (2016)48896–48904. https://doi.org/10.1039/C6RA09839H.

[24] M.S. Chaudhry, A. Czekanski, Z.H. Zhu, Characterization of carbon nanotubeenhanced interlaminar fracture toughness of woven carbon fiber reinforcedpolymer composites, Int. J. Mech. Sci. 131-132 (2017) 480–489. https://doi.org/10.1016/j.ijmecsci.2017.06.016.

[25] W. Zhao, M. Fang, F. Wu, H. Wu, L. Wang, G. Chen, Preparation of graphene byexfoliation of graphite using wet ball milling, J. Mater. Chem. 20 (2010) 5817–5819. https://doi.org/10.1039/C0JM01354D.

[26] A.T. Seyhan, M. Tanoglu, K. Schulte, Mode I and mode II fracture toughness ofE-glass non-crimp fabric/carbon nanotube (CNT) modified polymer based com-posites, Eng. Fract. Mech. 75 (18) (2008) 5151–5162. https://doi.org/10.1016/j.engfracmech.2008.08.003.

[27] BS ISO 527-2:1996, Plastics - Determinatin of Tensile Properties - Part 2: TestConditions for Moulding and Extrusion Plastics, British Standard Institute, BSI.1996.

[28] ASTM D5045-99, Standard Test Method for Plane-Strain Fracture Toughnessand Strain Energy Release Rate of Plastic Materials, 1999.

[29] ISO 14125:1998, Fibre-reinforced Plastic Composites Determination of FlexuralProperties, 1998.

[30] ISO 15024:2001, Fibre-reinforced Plastic Composites Determination of Mode IInterlaminar Fracture Toughness, GIC for Unidirectionally Reinforced Materials,2001.

https://doi.org/10.1016/j.polymer.2016.05.007https://doi.org/10.1016/j.polymer.2016.05.007https://doi.org/10.1016/j.matdes.2015.08.149https://doi.org/10.1016/j.polymer.2015.04.002https://doi.org/10.1016/j.engfracmech.2017.05.032https://doi.org/10.1016/j.matdes.2010.11.034https://doi.org/10.1016/j.matdes.2010.11.034https://doi.org/10.1016/j.matdes.2015.08.059https://doi.org/10.1016/j.matdes.2015.08.059https://doi.org/10.1007/s10853-016-0695-9https://doi.org/10.1016/j.matdes.2014.10.047https://doi.org/10.1016/j.matdes.2014.10.047https://doi.org/10.1016/j.matdes.2015.09.155https://doi.org/10.1002/pc.24368https://doi.org/10.1002/pc.24368https://doi.org/10.1007/s10853-016-0468-5https://doi.org/10.1007/s10853-016-0468-5https://doi.org/10.1016/j.compositesa.2012.09.012https://doi.org/10.1016/j.compositesa.2012.09.012https://doi.org/10.1016/j.polymer.2011.03.002https://doi.org/10.1016/j.matdes.2016.01.077https://doi.org/10.1016/j.compositesb.2017.07.070https://doi.org/10.1016/j.polymertesting.2017.01.028https://doi.org/10.1016/j.polymertesting.2017.01.028https://doi.org/10.1177/0021998316655201https://doi.org/10.1177/0021998316655201https://doi.org/10.1177/0731684417691365https://doi.org/10.1016/j.compscitech.2016.02.032https://doi.org/10.1016/j.compscitech.2016.02.032https://doi.org/10.1016/j.compscitech.2016.12.017https://doi.org/10.1016/j.compscitech.2016.12.018https://doi.org/10.1016/j.compscitech.2016.12.018https://doi.org/10.1016/j.compositesb.2010.06.001https://doi.org/10.1039/C6RA09839Hhttps://doi.org/10.1016/j.ijmecsci.2017.06.016https://doi.org/10.1016/j.ijmecsci.2017.06.016https://doi.org/10.1039/C0JM01354Dhttps://doi.org/10.1016/j.engfracmech.2008.08.003https://doi.org/10.1016/j.engfracmech.2008.08.003http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0135http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0140http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0145http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0150

92 D. Quan et al. / Materials and Design 143 (2018) 81–92

[31] ISO 15114:2014, Fibre-reinforced Plastic Composites Determination of ModeII Fracture Resistance for Unidirectionally Reinforced Materials Using the Cal-ibrated End-loaded Split (C-ELS) Test and an Effective Crack Length Approach,2014.

[32] A.S. Holik, R.P. Kambour, D.G. Fink, S.Y. Hobbs, Grinding and polishingtechniques for thin sectioning of polymeric materials for transmission lightmicroscopy, Microstruct. Sci. 7 (1979) 357–367.

[33] S.-Y. Fu, Z.-K. Chen, S. Hong, C.C. Han, The reduction of carbon nanotube (CNT)length during the manufacture of CNT/polymer composites and a method tosimultaneously determine the resulting CNT and interfacial strengths, Carbon47 (14) (2009) 3192–3200. https://doi.org/10.1016/j.carbon.2009.07.028.

[34] G. Sui, W. Yu, Q. Zhang, F. Chen, Q. Fu, Achieving excellent dispersionand electrical conductivity of olefin block copolymer/MWCNTs composites effi-ciently via high-shear processing, Polymer 123 (2017) 65–72. https://doi.org/10.1016/j.polymer.2017.07.007.

[35] F.H. Gojny, M.H.G. Wichmann, U.K. pke, B. Fiedler, K. Schulte, Car-bon nanotube-reinforced epoxy-composites: enhanced stiffness and fracturetoughness at low nanotube content, Compos. Sci. Technol. 64 (15) (2004) 2363–2371. developments in carbon nanotube and nanofibre reinforced polymers.https://doi.org/10.1016/j.compscitech.2004.04.002.

[36] F.H. Gojny, M.H. Wichmann, B. Fiedler, K. Schulte, Influence of differentcarbon nanotubes on the mechanical properties of epoxy matrix composites -a comparative study, Compos. Sci. Technol. 65 (15-16) (2005) 2300–2313.https://doi.org/10.1016/j.compscitech.2005.04.021.

[37] T. Tsafack, J.M. Alred, K.E. Wise, B. Jensen, E. Siochi, B.I. Yakobson, Exploringthe interface between single-walled carbon nanotubes and epoxy resin, Carbon105 (2016) 600–606. https://doi.org/10.1016/j.carbon.2016.04.066.

[38] S.A. Sydlik, J.-H. Lee, J.J. Walish, E.L. Thomas, T.M. Swager, Epoxy functional-ized multi-walled carbon nanotubes for improved adhesives, Carbon 59 (2013)109–120. https://doi.org/10.1016/j.carbon.2013.02.061.

[39] P.J.F. Harris, Carbon nanotube composites, Int. Mater. Rev. 49 (1) (2004) 31–43.https://doi.org/10.1179/095066004225010505.

[40] J.N. Coleman, U. Khan, W.J. Blau, Y.K. Gun’ko, Small but strong: a review ofthe mechanical properties of carbon nanotube polymer composites, Carbon 44(9) (2006) 1624–1652. https://doi.org/10.1016/j.carbon.2006.02.038.

[41] D. Bhattacharjee, J.F. Knott, Effect of mixed mode I and II loading on the frac-ture surface of polymethyl methacrylate (PMMA), Int. J. Fract. 72 (4) (1995)359–381. https://doi.org/10.1007/BF00040373.

[42] B.R.K. Blackman, A.J. Kinloch, M. Paraschi, The determination of the modeII adhesive fracture resistance, GIIC, of structural adhesive joints: an effectivecrack length approach, Eng. Fract. Mech. 72 (6) (2005) 877–897. https://doi.org/10.1016/j.engfracmech.2004.08.007.

[43] J. Xie, A.M. Waas, M. Rassaian, Estimating the process zone length of fracturetests used in characterizing composites, Int. J. Solids Struct. 100 (2016) 111–126. https://doi.org/10.1016/j.ijsolstr.2016.07.018.

[44] Y. Zhu, C.E. Bakis, J.H. Adair, Effects of carbon nanofiller functionalizationand distribution on interlaminar fracture toughness of multi-scale reinforcedpolymer composites, Carbon 50 (3) (2012) 1316–1331. https://doi.org/10.1016/j.carbon.2011.11.001.

[45] M. Arai, Y. Noro, K. ichi Sugimoto, M. Endo, Mode I and mode II interlaminarfracture toughness of CFRP laminates toughened by carbon nanofiber inter-layer, Compos. Sci. Technol. 68 (2) (2008) 516–525. https://doi.org/10.1016/j.compscitech.2007.06.007.

http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0155http://refhub.elsevier.com/S0264-1275(18)30065-0/rf0160https://doi.org/10.1016/j.carbon.2009.07.028https://doi.org/10.1016/j.polymer.2017.07.007https://doi.org/10.1016/j.polymer.2017.07.007https://doi.org/10.1016/j.compscitech.2004.04.002https://doi.org/10.1016/j.compscitech.2005.04.021https://doi.org/10.1016/j.carbon.2016.04.066https://doi.org/10.1016/j.carbon.2013.02.061https://doi.org/10.1179/095066004225010505https://doi.org/10.1016/j.carbon.2006.02.038https://doi.org/10.1007/BF00040373https://doi.org/10.1016/j.engfracmech.2004.08.007https://doi.org/10.1016/j.engfracmech.2004.08.007https://doi.org/10.1016/j.ijsolstr.2016.07.018https://doi.org/10.1016/j.carbon.2011.11.001https://doi.org/10.1016/j.carbon.2011.11.001https://doi.org/10.1016/j.compscitech.2007.06.007https://doi.org/10.1016/j.compscitech.2007.06.007

Enhancing mode-I and mode-II fracture toughness of epoxy and carbon fibre reinforced epoxy composites using multi-walled carbon nanotubes1. Introduction2. Experimental2.1. Materials2.2. Preparation of MWCNT modified epoxies2.3. Fabrication of CFRPs2.4. Test procedure

3. Results and discussion3.1. Length of the MWCNTs3.2. Mechanical properties3.2.1. Bulk epoxy3.2.2. CFRPs

3.3. Fracture properties3.3.1. Bulk epoxy3.3.2. Fractographic studies: bulk epoxy3.3.3. CFRPs3.3.4. Fractographic studies: CFRPs

4. ConclusionsAcknowledgmentsReferences