Journal of Science: Advanced Materials and Devices (2019) 299e309 Contents lists available at ScienceDirect Journal of Science: Advanced Materials and Devices journal homepage: www.elsevier.com/locate/jsamd Original Article Fracture analysis and mechanical properties of three phased glass/ epoxy laminates reinforced with multiwalled carbon nanotubes Rohit Pratyush Behera a, *, Prashant Rawat a, b, K.K Singh a, Sung Kyu Ha c, Anand Gaurav a, Santosh K Tiwari c a b c Department of Mechanical Engineering, Indian Institute of Technology (ISM), Dhanbad, India College of Civil Engineering, Hunan University, Changsha, China Department of Mechanical Engineering, Hanyang University, Seoul, South Korea a r t i c l e i n f o a b s t r a c t Article history: Received December 2018 Received in revised form March 2019 Accepted 13 March 2019 Available online 28 March 2019 Herein, we report the use of Multi Wall Carbon Nano Tubes (MWCNTs) as nano-compatibilizers based on their astonishing mechanical properties and ease of processing To fabricate laminate samples, pure MWCNTs were homogeneously dispersed in the fiber-reinforced plastic (FRP) composite with 0, 0.5, and 1.5 wt % loading The laminates were prepared with eight plies (4.0 ± 0.1 mm thickness) using the hand layup technique assisted by the compression moulding method It was found that the tensile, compressive and inter-laminar shear strength (ILSS) increase by 103.81%, 139.78% and 36.06%, respectively corresponding to wt % loading of MWCNTs as compared to neat GFRP specimen However, a rapid decrease in strength beyond wt % loading of MWCNTs has been noted Interestingly, the maximum of the tensile strength was higher than that of the compressive strength, and the maximum of the tensile modulus was larger than that of the compressive modulus in the case of wt % loading of MWCNTs It was observed that after a certain loading, the mechanical properties of such laminates can only reach the best value with an optimum loading of MWCNTs In addition, the micromechanical failure modes and effect of MWCNTs loading on internal morphologies of the composites were also intensively explored with the help of Field Emission Scanning Electron Microscopic (FESEM) analysis © 2019 The Authors Publishing services by Elsevier B.V on behalf of Vietnam National University, Hanoi This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/) Keywords: Mechanical properties MWCNTs Fiber-reinforced plastic Tensile strength Compressive strength ILSS properties Failure modes Introduction In the past few decades, nanocomposites have become interesting for different applications, particularly to improve the mechanical properties of polymeric materials using nanoparticles (as nanofillers) These nanofillers may be graphene [1], chopped carbon fibers [2], nanoclay, CNTs and their derivatives [3] Depending on the nature, these nanomaterials have capabilities to influence the mechanical, thermal and electrical properties of the produced nanocomposites Several recent investigations have proved the applicability of CNTs based composites in many fields, especially for lightweight and high-performance structures in aerospace industry and as coating materials in the maritime and chemical industries [4] The astonishing properties of CNTs in mechanical (axial Young modulus of 1e5 TPa [5], stiffness, * Corresponding author E-mail address: rohit.pratyush@gmail.com (R.P Behera) Peer review under responsibility of Vietnam National University, Hanoi strength [6], flexibility [7], fracture toughness [8]) and physical (high thermal conductivity, electrical conductivity [3,9], semiconducting behavior [10]) aspects, have been continuously studied and published The large surface area of nanotubes can act as an interface and bridging agent for uniform load transfer, but an extreme agglomeration of CNTs is caused due to strong attractive forces between CNTs Such aggregation always demises the properties of nanocomposites mainly in the case of polymer composites [11] The specific surface area of the CNTs is dependent on the number of sidewalls and the diameter of the tubes themselves, therefore single-wall CNTs (SWCNTs) have the largest surface area as compared to the double-wall CNTs (DWCNTs) and multiwall CNTs (MWCNTs) Their dispersibility, however, is quite low and the load transfer is difficult as already discussed elsewhere [12] In this line, we have used the MWCNTs because of their ease of bulk production, low cost per unit and high thermal, chemical stability Such CNTs based nano-composites have been under intensive study using different matrix materials, including metals https://doi.org/10.1016/j.jsamd.2019.03.003 2468-2179/© 2019 The Authors Publishing services by Elsevier B.V on behalf of Vietnam National University, Hanoi This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/) 300 R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 [13,14], ceramics [15,16], polymers [17] and so on Moreover, CNTs are being used as a secondary reinforcement in FRP composites to enhance their properties beyond the natural limit, even though there are certain challenges for developing an efficient and considerably tough three-phase glass/epoxy/ MWCNT nanocomposites These challenges are (i) a standardized and homogeneous distribution of CNTs in the two-phase glass/ epoxy composites, (ii) the proper interfacial bonding between the filler and the matrices which affect the uniform load transfer from the matrix to the reinforcement, and (iii) the tedious functionalization of CNTs To minimize these three-shortcoming, numerous efforts have been made and each procedure has its own advantages and disadvantages [15,16] Furthermore, three foremost mechanisms have been taken to explain the load transfer from the matrix to the filler and vice-versa, including: (i) mechanical interlocking, (ii) covalent and non-covalent bonding between the matrix and the CNTs, (iii) Van der Waals interaction between CNTs/glass and epoxy composites In the case of MWCNTs, the first one is very problematic owing to the smooth surface of CNTs and the second one cannot be guaranteed due to the low friction and the only reason left is addressed for the load transfer [18] To explore these issues, Liu et al [19] studied the tensile modulus and the yield strength by dispersing wt % of MWCNTs in a nylon-6 matrix and found an increase in values around 214% and 162%, respectively They proved that all threemechanisms mentioned above are applicable for MWCNTs to explain the outstanding mechanical properties of MWCNT-Nylol6 composites Grimmer et al.[20] examined and found that the incorporation of small volume fractions of MWCNTs to the glassfiber composites greatly diminishes the cyclic delamination crack propagation rates Schadler et al.[18] studied the load transfer in MWCNT-epoxy composites and found that there is a large scatter in the compression modulus, they also reported exceptional trend in tensile modulus and the maximum value of compression modulus is larger than that of the tensile modulus Allaoui et al.[21] considered the mechanical and electrical properties of MWCNT/epoxy composites and found that the Young's modulus is doubled, and the yield strength is quadrupled at and wt % loading of MWCNTs, respectively In this work, we tried to modify the brittle nature of the bidirectionally woven glass fiber composite by adding MWCNTs as a secondary reinforcement and as a nano-compatibilizers The tensile, compressive and inter-laminar behaviors are investigated; the modulus of tension and compression forces are compared; the fracture analysis is explored; and the morphological properties are discussed considering different loadings of MWCNTs Nevertheless, this paper proposed an optimum loading percentage to exploit the mechanical properties of the GFRP composites using MWCNTs, which is the novelty of the paper presented in the Fig 1(a), showing the characteristics of -COOH functionalized carbon nanotubes [24] Modified MWCNTs show intense D and G bands at around 1348 and 1580 cmÀ1, respectively, which disclose the prominent structural disorder caused by the incorporation of oxygen functional moieties on the carbon skeleton of the nanotubes [24,25] The presence of a broad 2D (2690 cmÀ1) peak in the spectrum is also a signature of the edge disorder in the nanotubes due to the incorporation of the eCOOH functional groups [24,25] To evaluate the nature of the functional groups on MWCNTs, FTIR analysis was carried out and the spectrum is presented in Fig 1(b) [26,27] The as-prepared, modified MWCNTs show a clear absorption for eOH, C-O, and -C¼O [26,27] This confirms the successful incorporation of the -COOH groups on the surface of the carbon nanotubes which is in good consistency with the previous investigations [26,27] Fabrication of MWCNTs based nanocomposites Eight layered quasi-isotropic symmetrical GFRP laminates were prepared with different loadings of the pure samples, i.e 0.5, and 1.5 wt % of MWCNTs Mixing of the surface modified MWCNTs in epoxy resins (Bisphenol-A) was completed using a probeultrasonicator (Fig 2(a)) This process is one of the most effective paths for the homogeneous dispersion of MWCNTs in polymer matrices as discussed elsewhere [11] The ultrasonication may cause heat generation in the solution and consequently aggregation of the MWCNTs Therefore to avoid this phenomenon, the beaker was fully covered in an ice blanket [11] Once mixing of the MWCNTs and epoxy was completed, the hardener (K-6) was mixed in 10:1 (epoxy: hardener) ratio followed by 15 minutes continuous ultrasonication Herein, the glass fiber used was bi-directionally [(0/90) and (ỵ45/-45)] woven (600GSM) as provided by M.S Industries, Kolkata, India The stacking sequence of the proposed design for the laminates is shown in Fig 2(b), i.e (00/ỵ900), (ỵ45 / 45 ), (ỵ45 /45 ), (0 /90 )//(0 /ỵ90 ), (ỵ45 /45 ), (ỵ45 / 45 ), (0 /90 ) To engineer the three-phased composite laminates, the first layer was placed over a flat glass surface and the resin solution was applied manually using a soft brush The same procedure was also applied for all other layers as per proposed symmetrical design To remove extra resins from the edges of the wet laminates, an iron roller was rolled after placing one layer over the other manually Authors have adopted the hand layup technique for the preparation of wet laminates and curing of the laminates was carried out with the help of a press-molding machine under 40 KN pressure at room temperature for 24 hours (Fig 2(c)) Experimental testing method 4.1 Tensile test Synthesis and characterization The arc-discharge method for the bulk production of high quality MWCNTs developed by Iijima et al [22] is well known and herein, we have adopted the same procedure for the MWCNTs synthesis and their surface modification as reported by Singh et al.[23] To confirm the functionalization of the as-prepared MWCNTs, Raman and FTIR spectroscopic techniques were used and important features are noted below Raman spectroscopy is one of the best methods to analyze carbon nanomaterials Raman spectra of the as-prepared modified MWCNTs was recorded at an excitation wavelength of 532 nm [24,25] The Raman spectrum of the modified MWCNTs is The prepared laminates with different loadings of MWCNTs (total 20 samples, for each wt %) were cut out in the dimensions of 125 mm  15 mm  mm using a diamond cutter as per ASTM D3039 requirement The aluminum tabs of dimensions 25 mm  15 mm  mm were properly cleaned with propyl alcohol and pasted over the sample laminates as shown in Fig 3(a) to avoid alteration during the testing Each specimen was then gripped in the prepared fixture of the Universal Testing Machine (UTM) which prevented any lateral movements and forces were measured using S-type load cells as shown in Fig 3(c) Five such tests were conducted for each loading amount of MWCNTs at a strain rate of 0.5 mm/min R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 301 Fig (a) Raman spectra of surface modified MWCNTs and (b) FTIR of the as-synthesized and modified MWCNTs revealing the presence of the -COOH and -OH functional groups Fig (a) Set up of Probe Ultrasonication process; (b) Schematic representation of stacking sequence; (c) Set up of press molding machine Fig (a) Specimen sample for the tensile test (ASTM D3039); (b) Specimen sample for compression test (ASTM D3410); and (c) Test fixture in UTM machine 4.2 Compression test To investigate the compressive strength, seven different specimens (each wt % of dimension 100 mm  10 mm  mm) were prepared as per ASTM D3410 requirement and a diamond cutter was used for the sample preparation The properly cleaned aluminum tabs of dimensions of 45 mm  10 mm  mm were pasted over the specimen (with gauge length of 10 mm) as shown in Fig 3(b) to avoid alteration This test was performed under the displacement control with a strain rate of 0.5 mm/min at room temperature The load was measured using the S-type load cell attached to the specimen as discussed in the previous section It is notable that the specimens used were short enough to prevent any buckling and clasping during the compression measurement Further, any effects caused due to the stress concentration at the grips were considered insignificantly because 302 R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 the repeated failures were witnessed at the gage section of all the loaded GFRPs 4.3 Inter-laminar shear strength (ILSS) The inter-laminar shear strength tests were conducted based on the short beam shear strength test (SBS) For this test, seven specimens of dimensions 24 mm  mm  mm were prepared as per obligation of ASTM D2344, shown schematicaly in Fig 4(a) The gage length of the specimen was 20 mm The tests were performed at a strain rate of 0.5 mm/min as shown in Fig 4(b) For the ILSS measurement, the gauge length of the specimens were kept very small to minimize the effect of bending during the failure of laminates under tension and compression Thus, the main failure mechanism was dominated by the pure shear phenomenon The standard equation used for the calculation of inter-laminar shear strength is noted below: F à ¼ 0:75  P bÂh (1) where, F* ¼ ILSS or Short-beam strength (ILSS) (MPa) P ¼ Max load observed during test (N) b ¼ Specimen width (mm) h ¼ Specimen thickness (mm) The cut specimens’ samples are shown in Fig 5(a)e(c) for the tensile, compressive and ILSS test, respectively The UTM (as shown in Figs 3(c) and 4(b)) used for this work was fully computerized and the machine can be operated at loading rates varying from 0.01 to 10 mm/min It had maximum load carrying capacity of 50 KN Henceforth, using the UTM, the maximum values of load and stress for different samples under tension, compression and ILSS were noted and the maximum load carrying capacity was compared for the same wt.% samples in tension and compression for the systematic analysis The maximum tensile and compressive modulus for different wt % were also compared with neat samples Results and discussion 5.1 Tensile properties Most of the important mechanical properties of materials, such as yield strength, elasticity, ultimate tensile strength and ductility can be obtained by the tensile analysis [21] For the present investigation, five samples were tested and the obtained values are presented in Table for better understanding The graph shown in Fig 6a represents the average values of the tensile stress and it can be observed that the maximum stress values for the GFRP laminate composites are 122.70 MPa, 144.02 MPa, 250.08 MPa and 163.78 MPa for neat, 0.5, and 1.5 wt % of MWCNTs loading, respectively The Young's modulus of elasticity (ET) was determined using the stressestrain data based on the tensile tests whose values were 2956.62, 3453.71, 6315.15 and 3937.02 MPa for the neat, 0.5, and 1.5 wt % of MWCNTs/glass/epoxy sample laminates, respectively and all are presented in Table for the sake of simplicity The maximum value for the modulus (ETMax.) is 6315.15 MPa at wt % of MWCNTs loading and the average Young's modulus (ETavg.) for the tensile specimen laminates is found to be 4165.625 MPa It can be observed that there is a continuous increase in the stress value as compared to neat GFRP specimen up to wt % MWCNTs incorporated laminates as shown in Fig This enhancement can be accounted to the delay in the crack generation and propagation owing to the nucleation and bridging effect of MWCNTs [18] The role of MWCNTs to reinforce the mechanical properties is schematically presented in Fig Thus, the optimum loading of MWCNTs attributes a strong cross-linkage and enshrouding between the interfaces of resin which ultimately causes a delay in the crack propagation [11] Moreover, with the increase in the loading percentage of MWCNTs, the tensile strength of the FRP laminate increased as fracture behavior shifted from brittle to ductile like hackle for the homogenously dispersed MWCNTs [11] Thus, the strong nanoparticle covalent bonding between the matrixereinforcement interfaces is also responsible for the improved strength of the MWCNTs filled laminate specimens [18] Further, from the tensile analysis, it can be observed that with the increase of MWCNTs loading, there is a fall in stress value by 34.5% from wt % to 1.5 wt % Such a drastic reduction in the value of stress may be attributed to the aggregation of MWCNTs (as shown in Fig 9d), which takes place when MWCNTs loading increases beyond wt % [11] In the present situation, the aggregation of MWCNTs is mainly due to the selective distribution and the high degree of entanglement of MWCNTs with matrices [11,18] Therefore, the optimum stress value is 250.08 MPa at wt % MWCNTs incorporated GFRP and the average Young's modulus (ETAvg.) for the tensile test specimen is 4165.625 MPa 5.1.1 Failure mechanism and fracture analysis The failure mechanism is one of the most important aspects for the composites and polymeric materials [18] There are so many reasons for the failure in the case of FRPs and similar materials In this particular work, it is assumed that the failure originates from a region where there is the maximum stress concentration or an inherent defect is generated in the composite specimens during the testing as explained by Y Iwahori et al.[28] However, the homogeneous mixing of nanofillers in case of the three-phase composites is a serious concern and cannot be explained through the logic developed by Y Iwahori et al.[28], because the mixing of nanofillers Fig (a) Specimen sample for SBS test (ASTM D2344) and (b) Specimen fixture setup for SBS test R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 303 Fig Specimen test samples for (a) Tensile (b) Compressive and (c) SBS tests Table Statistics of the tensile tests performed for the hybrid MWCNT/glass/epoxy laminate samples Stress in MPa Specimen Number MWCNT (wt.%) Average Tensile strength (MPa) Standard deviation (SD) Standard deviation (%) Average Young's modulus (ET) (MPa) 0.5 1.5 120.57 145.67 247.63 165.67 118.96 145.98 245.97 161.23 125.45 144.65 251.67 160.78 124.23 142.66 253.53 164.45 124.29 141.14 251.16 166.77 122.70 144.02 250.08 163.78 2.782 2.057 3.102 2.67 2.27 1.43 1.24 1.63 2956.62 3453.71 6315.15 3937.02 Fig (a) Stress vs Strain graph for the tensile test of GFRP composite sample laminate and (b) Statistics of the tests performed according to ASTM D3039 with reference to Table into three-phase composites leads to the formation of clusters, cracks and voids owing to the non-uniform stress concentration (as shown in Fig (b)) which can also be the reason for the crack generation and propagation that ultimately leads to failure in the composites [28] The mechanism explained in Fig shows the fracture and cracking of FRP composites during the tensile test which is an indication of adopted mechanism for the failure [11] From the FESEM (Fig (d)) micrographs, it is clear that cracking at the interfaces of matrices occurrs at the point of maximum stress, and then the crack propagates through the fibers and apprears on the external surface of the specimen [11] In other words, the debonding between the matrix and the nanofiber/glass is responsible for the fracture which leads to the pull-out phenomenon as shown in Fig 8(a),(b) It is notable that the stress near the tip of the crack causes the matrix-fiber delamination before the actual bond breakdown and finally the crack reaches to the interface between the two laminates of the composite Fig 7(a) and the FESEM image in Fig 9(a) show the mechanism of delay in the crack propagation caused by the strong cross-linkage of MWCNTs between the two laminate composites [18] 5.2 Compression properties The compressive strength is often used to state applicability of composite materials as it denotes the ability of a material to withstand load tending to reduce size To further examine the mechanical properties of the fabricated composites, the compressive behavior was investigated as per ASTM D3410 and the stress vs strain (Fig 10(a)) curve of the same along with statistical data is presented in Table The compressive results of the studied samples are in good agreement with the tensile properties as mentioned in previous section The maximum compressive stress 304 R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 Fig Schematic diagram of (a) MWCNTs distributed between glass fiber layers which cause an increase in the tensile strength and (b) unmodified GFRP Fig (a) mechanism of tensile failure of neat GFRP specimen; (b) nanofiber pull-out and delay in crack propagation caused due to crosslinking of MWCNTs (c) image of the failure mechanism of specimen under tensile load; and (d) FESEM image of debonding followed by fiber pull-out (sC) values for the specimens are found to be 10.38, 13.44, 24.89 and 21.32 MPa for the neat, 0.5, and 1.5 wt % of MWCNTs loading in GFRP composites, respectively Similarly, the compressive modulus (EC) of the studied specimens were obtained (1069.56, 1383.71, 4194.77 and 2195.31 MPa for the neat, 0.5, and 1.5 wt % of MWCNTs/epoxy, respectively) from the corresponding graphs and the data is presented in Table for detailed information From the stress vs strain graphs (Fig 10(a)), MWCNTs loading up to wt % is accounted to the uniform load transfer from the matrix to the reinforcement which is further strengthened owing to the intrinsic properties of MWCNTs and their homogenous distribution at the interfaces of the composite Due to the aggregation of MWCNTs at R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 305 Fig FESEM images of (a) MWCNTs interlocking; (b) cured sample laminate at wt.% MWCNTs loading containing voids; (c) MWCNTs dispersion; and (d) agglomeration at 1.5 wt.% MWCNTs in GFRP Fig 10 (a) The Stress vs Strain graphs for the compressive test of GFRP composite laminate samples (b) Statistics of the tests performed according to ASTM D3410 with reference to Table Table The statistics of compressive tests performed for hybrid MWCNT/glass/epoxy laminate samples Stress in MPa Specimen Number MWCNT (wt.%) Average Compressive strength (MPa) Standard deviation (SD) Standard deviation (%) Average compression modulus (EC) (MPa) 0.5 1.5 10.56 14.56 24.23 22.56 10.65 14.89 25.17 22.21 9.87 13.65 25.36 20.97 11.23 12.63 23.67 21.25 10.02 13.23 25.87 20.56 10.88 12.02 25.61 20.23 9.43 13.10 24.32 21.46 10.38 13.44 24.89 21.32 0.626 1.018 0.820 0.842 6.03 7.57 3.29 3.95 1069.56 1383.71 4194.77 2195.31 higher loading, the cross-linkage (see Fig 9(a)) at the interfaces get decreased and therefore abrupt drops in strength can be seen above wt % of loading of nanofillers (Fig 9(d)) In summary, the maximum value of the compressive modulus (ECmax.) is found as 4194.77 MPa for the wt % of MWCNTs loading and the average value of the compressive modulus (ECAvg.) for the specimen laminates is found as 2210.837 MPa 5.2.1 Failure mechanism and fracture analysis Composites and nanocomposites possess different phases with different elastic properties Hence, there is a high possibility for the formation of microcracks across various regions of interfaces in the fabricated composites under the applied load [18] A typical fracture of fibers is shown in Fig 11(a),(b) for the specimen under compressive loading The compressive strength of the laminates is controlled by various mechanisms [29] First, it is seen that most of the laminate in the test undergo micro-buckling Buckling is the only area in the field of structural mechanics where failure is not related to the material strength The collapsing of the material due to buckling does not deal with the yield of the material [30] Generally, there are two types of micro-buckling modes of failure according to Rosen's model [31] (i): out of phase buckling of fibers or extension mode, (ii) in-phase buckling of fibers or shear mode as shown in Fig 11(e),(f) Their analysis was based on the assumption of the 2D behavior of composites and on an idealistic approach, so the predicted results in their test were significantly higher than 306 R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 Fig 11 (a) FESEM image of fractured fibers and (b) cracking due to compressive loading; (c) FESEM image of the kink band formation in the test sample specimen; (d) FESEM image failure mechanism of compressive test specimen (Kink-band formation and Splitting); (e) Extension mode or out of phase buckling; (f) Shear mode or in-phase buckling experimentally observed ones In their model, the authors predicted the value of the shear mode failure to be lower than that of the extension mode of failures which was also in agreement with the experimentally observed results Further, the failure modes are highly dependent on the fiber volume fraction (Vf) For Vf < 30%, the extension mode is dominant as with application of the sudden load, the adjacent fibers tend to deform sinusoidally with the deformation patterns being 180 out of phase, whereas for Vf > 30%, the fibers adjacent to each other deform transversely in phase indicating the shear mode failure [30] In our experiment, since the prepared composite is high in fiber content with fiber volume fraction (Vf ~ 55%), therefore, in the coherence to the Rosen's model, it is feasible to assume the micro-buckling to be in-phase as shown in Fig 11 Also, researchers like Tadjbaksh and Wang in their model took into consideration the inter-ply micro-buckling of the crossply laminates [32] They modelled the laminate as a single ply inhomogeneous continuum In this analysis, they observed an increase in resistance of buckling due to adjacent plies Similarly, in the present study, most of the samples were resisting buckling owing to the use of cross-ply laminates [33] Secondly, there is kinking, which is also perceived to be the most common failure mechanism Kinking in FRPs can be said to be the consequence of the combination of plastic micro-buckling and the low strain rate subjected to the fibers A typical kinking failure mechanism is shown in the Fig 11c These initiative mechanisms like kinking and micro-buckling at micro-structural level lead to the global instability of the composite materials [30] Lee and Waas [33] in their investigation on the compressive strength of unidirectional GFRP laminates found that the kink-bands were formed at different fiber-volume fractions of the tested GFRP samples Also, for higher fiber-volume fractions, it remained one of the major modes of failure Similarly, in this study, it was observed that the kink bands were formed But, as the fiber volume fraction of the prepared GFRP was high (Vf ~ 55%), no significant stress drop was observed during testing of the loaded GFRP laminates Thus, it can be concluded that the chances of formation of kink-bands decreases at high fiber volume fractions of loaded bi-directionally woven GFRP laminates A typical kink band formation is shown in Fig 11(c) of the loaded GFRP samples Thirdly, we address the splitting or delamination as another failure mechanism for the rupture of the composites [33] This kind of failure modes is supposed to occur when a pre-existing flaw inside the specimen starts growing under compressive loading conditions In multi-directional composites, where the fibers are continuous, cracks with an opening under compressive loading are easily liable to delamination in a macroscopic level and fiber-matrix debonding in a microscopic level The micro-cracks that grow under compressive loading tend to form transverse cracks These micro-cracks with a gradual increase in the compressive load grow and their domain of circumferential size increases, and they form interfacial de-bonds [33] Debonding is the transition mechanism followed by the micro-cracking which eventually leads to transverse cracking [34] Although, it is assumed that transverse crack originates from the defects or voids present in the matrices owing to improper layering and selective distribution of MWCNTs (see Fig 9(b)) However, the consequence of the same may enhance the cracking and therefore the fiber-matrix interfacial weakening in the case of an uniaxial transverse loading [35] Such crack initiation and propagation cannot afford applied loads and result in fracture via the kink band formation A typical debonding or shear failure mechanism is shown in Fig 12(a),(b) followed as by a transverse cracking in the compression test samples As described in previous sections, there are three major mechanisms of failure for the compressive laminate samples, but it is observed that no particular failure is alone responsible for the breakdown of the test samples The failure is governed majorly by a combination of kink-band formation and splitting But there can be other failure modes like splitting or shear failure alone However, this is the rare case, i.e., one among seven specimen samples So, for high fiber volume fraction bi-directional GFRP samples we designate the failure mode by kink-band formation followed by splitting or shear failure as shown in Fig 11(d) 5.3 Inter-laminar shear strength (ILSS) properties The tests were performed for the short beam shear (SBS) strength and the observed strength values are summarized in Table Also, the statistics of the obtained force values are presented in Fig 13(b) The force vs deflection graph for the test samples (see Fig 13(a)) represents the average values of force for the samples put to test in the Hounsfield UTM R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 307 Fig 12 (a) debonding mechanism leading to transverse crack and (b) hinderance to the transverse crack caused by loading MWCNTs Table The statistics of the SBS test performed for hybrid MWCNT/glass/epoxy laminate samples Force in N Specimen Number Average Force (N) Standard deviation (SD) Standard deviation (%) Inter-laminar shear strength (ILSS) (MPa) 469.23 532.48 637.55 503.12 468.12 536.65 634.23 501.67 472.27 540.45 632.12 491.73 460.57 521.27 621.98 487.33 457.32 526.34 620.45 495.51 460.73 525.25 628.34 493.34 470.12 517.55 658.64 510.64 465.48 528.57 633.33 497.62 5.80 16.42 12.78 7.95 1.25 3.11 2.02 1.60 10.909 12.388 14.843 11.663 MWCNT (wt.%) 0.5 1.5 Fig 13 (a) Force vs displacement graph of GFRP laminate samples obtained from SBS test; and (b) statistics of the tests performed according to ASTM D2344 with reference to Table When the shear load in the transverse direction experienced by a composite laminate exceeds the inter-laminar shear strength (ILSS), a failure will occur between the fiber layers (which are the reinforcing materials in FRPs) and its was known as delamination failure [36] In order to measure the ILSS, a pattern of the pure shear stress should be generated which will make the composite undergo an interlaminar shear failure This failure will occur between the glass woven plies, which act as the reinforcing material In the case of a load applied perpendicular to the fiber layers, an additional failure is possible and, therefore, an accurate ILSS estimation is not possible In this testing, we used the SBS test, where the failure was caused by both the shear failure and the failure due to bending (caused by both tension and compression) The failure is not due to shear alone, so there can be anomaly in the results obtained [36] To avoid the bending effects of failure, we reduced the gage length, so that the moment is reduced which ultimately diminishes the failure due to the bending effects [36] In this case, the gage length as per ASTM D2344 is 20 mm, so it can be assumed that the pure shear failure will dominate the bending failure at a greater scale, therefore, we can neglect the effects of bending From the graph of Fig 13(a) and from the calculations using equation (1), the Inter-laminar Shear Strength values are 10.909 MPa, 12.388 MPa, 14.843 MPa and 11.663 MPa for MWCNT loading values of 0, 0.5, and 1.5 wt %, respectively, which are noted down in Table The reason for the increase in ILSS of the fiber reinforced composites up to wt % can be addressed to the MWCNTs/epoxy suspension between the fiber layers, which create a strong cross-linking bond between the layers as shown in Figs 9(a) and 14(b) Although, the CNTs in this case are not preferentially oriented, they are randomly dispersed, it is seen that 308 R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 Fig 14 (a) The crack propagation mechanism in th neat GFRP laminate; (b) Hinderance in the crack propagation due to MWCNTs aligned along the thickness direction; (c) FESEM image of failure mechanism of short beam shear strength (SBSS) test sample specimen (delamination) some of the MWCNTs tend to orient along the thickness direction which transfers the load from the matrix to the MWCNTs Fan et al.[36] used oxidized multi-wall nanotubes (OMWCNTs) with the vacuum assisted resin transfer molding method (VARTM) and the newly developed injection and double vacuum assisted resin transfer molding method (IDVARTM) to preferentially orient the MWCNTs along the thickness direction, which ultimately leads to increased ILSS due to the effective and uniform load transfer from the matrix to the CNTs The decrease in ILSS at 1.5 wt % of MWCNTs loading from 14.843 to 11.663 MPa indicates that the CNTs after wt % loading tend to re-agglomerate (Fig 9(d)) Thus, it can be assumed that the wt.% loading is an optimum value of reinforcement, and beyond the wt.% loading, owing to the saturation effect and the attractive Van der Waal forces among the nanoparticles, the agglomeration effect tends to dominate and results in major aggregation, indicating that the ILSS is decreased by 21.42% from to 1.5 wt % of MWCNTs loading 5.3.1 Failure mechanisms and fracture analysis The failure mechanism of the ILSS test samples for the SBS test is relatively easy Since, the load is applied on an overhanging test sample with a very short gage length, the crack propagation takes somewhat close to the axial direction [37] Since bending also exists in the failure mechanism, the crack propagation is not along the pure axial direction Therefore, there are two mechanisms which can be accounted for the failure: (i) the shear failure and (ii) the bending failure The shear failure is a result of the delamination caused due to the matrix cracking followed by breaking of the bond between the fiber and the matrix which eventually leads to the delamination Since the matrix is considered to be the weakest material in the composite structure, so it is assumed that the crack initiation takes place from the defects or voids present in the matrix (Fig 9b), and it is expected to be eventually present after the curing process [35] This becomes the point of the maximum stress concentration, so a load up to its maximum bearing capability along the axial direction is considered the maximum ILSS Further application of the load after this point leads to the crack propagation from the defect which causes matrix breaking and the load is transferred from the matrix to the secondary reinforcement, i.e the MWCNTs as shown in Fig 14b The matrix breaking and the crack propagation along the near axial direction leads to the delamination of the fiber and the matrix, which upon further spreading conquers a lot of circumferential area under its domain ultimately leading to the shear failure [37] In this testing process, the most dominant failure mode is the inter-laminar shear failure A typical failure mechanism of the SBSS test is shown in Fig 14(c) Conclusion In summary, surface modified MWCNTs were synthesized and unvaryingly dispersed in the glass fiber-reinforced plastic composites with 0, 0.5, and 1.5 wt % loading The mechanical properties of specimens were inspected mainly in the context of tensile, compressive and ILSS test In the present study, it has been established that an optimum tensile stress value of 250.08 MPa was at wt % of MWCNTs loading along with Young's modulus (ETmax) of 6315.15 MPa and the composites failed owing to the debonding followed by the fiber pullout The optimum compressive stress for the composite in the case of wt % of MWCNTs loading was found of ~24.89 MPa corresponding to 4194.77 MPa compressive modulus Interestingly, the maximum strength value for the tension was found greater than that for the compression which implies the role of the nanotubes as nano-compatibilizers and as reinforcement on the properties of the fabricated composites For the studied samples, the tensile modulus was greater than the compressive one indicating that there was less elastic deformation of the material under tension as compared to compression Moreover, owing to the proper dispersion of MWCNTs in the fabricated composites, the increment in the tensile stress and compressive stress were 103.81% and 139.78%, respectively Also, for the wt.% of MWCNTs loading, the optimum ILSS was found to be 14.843 MPa with an increment of 36.06% as compared to the neat sample Thus, the optimum MWCNTs loaded GFRP can be used in many structural R.P Behera et al / Journal of Science: Advanced Materials and Devices (2019) 299e309 applications where a high stability against tension, compression and inter-laminar strength is needed at the same time Conflicts of interest Authors have no conflict of interest References [1] M Saranya, R Ramachandran, F Wang, Graphene-zinc oxide (G-ZnO) nanocomposite for electrochemical supercapacitor applications, J Sci Adv Mater Dev (2016) 454e460, https://doi.org/10.1016/J.JSAMD.2016.10.001 [2] L Zhang, A Aboagye, A Kelkar, C Lai, H Fong, A review: carbon nanofibers from electrospun polyacrylonitrile and their applications, J Mater Sci 49 (2014) 463e480, https://doi.org/10.1007/s10853-013-7705-y [3] M Farbod, A Zilaie, I Kazeminezhad, Carbon nanotubes length optimization for preparation of improved transparent and conducting thin film substrates, J Sci Adv Mater Dev (2017) 99e104, https://doi.org/10.1016/ J.JSAMD.2017.02.005 [4] C.P Sandhya, B John, C Gouri, Sn/Al2O3/C/CNT composite prepared by wet milling as anode material for lithium-ion cells, J Sci Adv Mater Dev (2017) 210e214, https://doi.org/10.1016/J.JSAMD.2017.04.003 [5] A Guevara-Morales, A Taylor, Mechanical and dielectric properties of epoxyeclay nanocomposites, J Mater Sci 49 (2014) 1574e1584, https:// doi.org/10.1007/s10853-013-7840-5 [6] E Thostenson, T Chou, On the elastic properties of carbon nanotube-based composites: modelling and characterization, J Phys D - Appl Phys 36 (2003) 573e582, https://doi.org/10.1088/0022-3727/36/5/323 [7] J Despres, E Daguerre, K Lafdi, Flexibility of graphene layers in carbon nanotubes, Carbon 33 (1995) 87e92 lu, K Schulte, Tensile mechanical behavior and fracture [8] T Seyhan A, M Tanog toughness of MWCNT and DWCNT modified vinyl-ester/polyester hybrid nanocomposites produced by 3-roll milling, Mater Sci Eng A 523 (2009) 85e92, https://doi.org/10.1016/j.msea.2009.05.035 [9] K Harris, A Elias, H Chung, Flexible electronics under strain: a review of mechanical characterization and durability enhancement strategies, J Mater Sci 51 (2016) 2771e2805, https://doi.org/10.1007/s10853-015-9643-3 [10] T Ebbesen, H Lezec, H Hiura, J Benett, H Ghaemi, T Thio, Electrical conductivity of individual carbon nanotubes, Nature 382 (1996) 54e56, https:// doi.org/10.1038/382054a0 [11] R.P Behera, P Rawat, K.K Singh, Tensile behavior of three phased glass/epoxy laminate embedded with MWCNTs: an experimental approach, Mater Today Proc (2018) 8176e8183, https://doi.org/10.1016/j.matpr.2017.11.506 [12] A Peigney, C Laurent, E Flahaut, R Bacsa, A Rousset, Specific surface area of carbon nanotubes and bundles of carbon nanotubes, Carbon 39 (2001) 507e514, https://doi.org/10.1016/S0008-6223(00)00155-X [13] T Kuzumaki, K Miyazawa, H Ichinose, K Ito, Processing of carbon nanotube reinforced aluminium composite, J Mater Res 13 (2011) 2445e2449, https:// doi.org/10.1557/JMR.1998.0340 [14] N.X Dinh, T.Q Huy, L Van Vu, L.T Tam, A.-T Le, Multiwalled carbon nanotubes/silver nanocomposite as effective SERS platform for detection of methylene blue dye in water, J Sci Adv Mater Dev (2016) 84e89, https:// doi.org/10.1016/J.JSAMD.2016.04.007 [15] A Peigney, C Laurent, E Flahaut, A Rousset, Carbon nanotubes in novel ceramic matrix nanocomposites, Ceram Int 26 (2000) 677e683, https:// doi.org/10.1016/S0272-8842(00)00004-3 [16] E Flahaut, A Peigney, C Laurent, C Marliere, F Chastel, A Rousset, Carbon nanotube-metal-oxide nanocomposites, Chem Eng J 302 (2016) 344e367, https://doi.org/10.1016/j.cej.2016.05.038 [17] M.H Suhail, O.G Abdullah, G.A Kadhim, Hydrogen sulfide sensors based on PANI/f-SWCNT polymer nanocomposite thin films prepared by electrochemical polymerization, J Sci Adv Mater Dev (1) (2019) 143e149, https://doi.org/10.1016/J.JSAMD.2018.11.006 309 [18] L Schadler, S Giannaris, P Ajayan, Load transfer in carbon nanotube epoxy composites, Appl Phys Lett 73 (1998) 3842e3844, https://doi.org/10.1063/ 1.122911 [19] T Liu, I Phang, L Shen, S Chow, W Zhang, Morphology and mechanical properties of multiwalled carbon nanotubes reinforced nylon-6 composites, Macromolecules 37 (2004) 7214e7222, https://doi.org/10.1021/ma049132t [20] C Grimmer, C Dharan, Enhancement of delamination fatigue resistance in carbon nanotube reinforced glass fiber/polymer composites, Compos Sci Technol 70 (2010) 901e908, https://doi.org/10.1016/j.compscitech.2010 02.001 [21] A Allaoui, S Bai, H Cheng, J Bai, Mechanical and electrical properties of a MWNT/epoxy composite, Compos Sci Technol 62 (2002) 1993e1998, https://doi.org/10.1016/S0266-3538(02)00129-X [22] S Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56e58, https://doi.org/10.1038/354056a0 [23] K Singh, S Chaudhary, R Venugopal, A Gaurav, Bulk synthesis of multiwalled carbon nanotubes by AC arc discharge method, Proc Inst Mech Eng Part N J Nanomater Nanoeng Nanosyst 231 (2017) 141e151, https://doi.org/ 10.1177/2397791417712836 [24] I.D Rosca, F Watari, M Uo, T Akasaka, Oxidation of multiwalled carbon nanotubes by nitric acid, Carbon 43 (2005) 3124e3131, https://doi.org/ 10.1016/j.carbon.2005.06.019 [25] Q Li, Q Xue, L Hao, X Gao, Q Zheng, Large dielectric constant of the chemically functionalized carbon nanotube/polymer composites, Compos Sci Technol 68 (2008) 2290e2296, https://doi.org/10.1016/j.matlet.2008.06.047 [26] C.M Chang, Y.L Liu, Functionalization of multi-walled carbon nanotubes with non-reactive polymers through an ozone-mediated process for the preparation of a wide range of high performance polymer/carbon nanotube composites, Carbon 48 (2010) 1289e1297, https://doi.org/10.1016/ j.carbon.2009.12.002 [27] F Liang, A.K Sadana, A Peera, J Chattopadhyay, Z Gu, R.H Hauge, W.E Billups, A convenient route to functionalized carbon nanotubes, Nano Lett (2004) 1257e1260, https://doi.org/10.1021/nl049428c [28] Y Iwahori, S Ishiwata, T Sumizawa, T Ishikawa, Mechanical properties improvements in two-phase and three-phase composites using carbon nanofiber dispersed resin, Compos Part A Appl Sci Manuf 36 (2005) 1430e1439, https://doi.org/10.1016/j.compositesa.2004.11.017 [29] A Jumahat, C Soutis, F.R Jones, A Hodzic, Fracture mechanisms and failure analysis of carbon fibre/toughened epoxy composites subjected to compressive loading, Compos Struct 92 (2010) 295e305, https://doi.org/10.1016/ J.COMPSTRUCT.2009.08.010 [30] J Bai, S.L Phoenix, Compressive failure model for fiber composites by kink band initiation from obliquely aligned, shear-dislocated fiber breaks, Int J Solids Struct 42 (2005) 2089e2128, https://doi.org/10.1016/ j.ijsolstr.2004.08.011 [31] B Rosen, Mechanics of Composite Strengthening: Fiber composite materials, Am Soc Met (1965) 37e75 [32] I Tadjbakhsh, Y Wang, Fiber buckling in three-dimensional periodic-array composites, Int J Solids Struct 29 (1992) 3169e3183 https://doi.org/10 1016/0020-7683(92)90034-Q [33] S Lee, A.M Waas, Compressive response and failure of fiber reinforced unidirectional composites, Int J Fract 100 (1999) 275e306, https://doi.org/ 10.1023/A:1018779307931 [34] R Harrison, M Bader, Damage development in CFRP laminates under monotonic and cyclic stressing, Fibre Sci Technol 18 (3) (April 1983) 163e180, https:// doi.org/10.1016/0015-0568(83)90039-8, 18 (1983) 163e180 [35] L Asp, L Berglund, R Talreja, Prediction of matrix initiated transverse failure in polymer composites, Compos Sci Technol 56 (1996) 1089e1097, https:// doi.org/10.1016/0266-3538(96)00074-7 [36] Z Fan, M Santare, S Advani, Interlaminar shear strength of glass fiber reinforced epoxy composites enhanced with multi-walled carbon nanotubes, Compos Part A Appl Sci Manuf 39 (2008) 540e554, https://doi.org/10.1016/ j.compositesa.2007.11.013 [37] J Zhu, A Imam, R Crane, K Lozano, V.N Khabashesku, E.V Barrera, Processing a glass fiber reinforced vinyl ester composite with nanotube enhancement of interlaminar shear strength, Compos Sci Technol 67 (2007) 1509e1517, https://doi.org/10.1016/J.COMPSCITECH.2006.07.018 ... area of carbon nanotubes and bundles of carbon nanotubes, Carbon 39 (2001) 507e514, https://doi.org/10.1016/S0008-6223(00)00155-X [13] T Kuzumaki, K Miyazawa, H Ichinose, K Ito, Processing of carbon. .. Lafdi, Flexibility of graphene layers in carbon nanotubes, Carbon 33 (1995) 87e92 lu, K Schulte, Tensile mechanical behavior and fracture [8] T Seyhan A, M Tanog toughness of MWCNT and DWCNT modified... wt % of MWCNTs loading and the average value of the compressive modulus (ECAvg.) for the specimen laminates is found as 2210.837 MPa 5.2.1 Failure mechanism and fracture analysis Composites and