My Paper - 3

Received: 26 December 2023 Revised: 6 March 2024 Accepted: 8 March 2024
DOI: 10.1002/pc.28337
RESEARCH ARTICLE
Experimental estimation and numerical validation of

cohesive zone parameters in hydroxyl functionalized
MWCNTs-reinforced CFRP under pure mode II loading
Pran Jyoti Saikia 1 | Munna Kumar 1 | Avneesh Kumar 1 | Nelson Muthu 2
1
Department of Mechanical Engineering,
Indian Institute of Technology Guwahati, Abstract
Guwahati, India The present study concerns the interface mode II fracture of hydroxyl functiona-
2
Department of Mechanical Engineering, lized multi-walled carbon nanotubes (MWCNTs) reinforced carbon fiber epoxy
Indian Institute of Technology Palakkad,
composites. The well-known cohesive zone model (CZM) is employed for inves-
Palakkad, India
tigating crack propagation. The major difficulty in applying the CZM is the
Correspondence parameter identification. In this work, the CZM's strength and energy parame-
Nelson Muthu, Department of Mechanical
Engineering, Indian Institute of
ters are determined reliably from the Short beam shear (SBS) and End-notched
Technology Palakkad, Palakkad, Kerala flexure (ENF) tests, respectively, for pristine and three MWCNTs variations—
678623, India. 0.1 wt.%, 0.2 wt.%, and 0.3 wt.%. The derived CZM parameters from the tests are
Email: nelsonm@iitpkd.ac.in
imputed to a bilinear cohesive law to simulate the global load vs. displacement
Funding information responses, which correlate well with the experimental findings. It is observed
Science and Engineering Research Board,
that the addition of 0.2 wt.% of MWCNTs increases the cohesive strength and
Grant/Award Number: ECR/2018/001638;
Indian Institute of Technology Guwahati, fracture energy up to 14.14% and 17.78%, respectively, when compared to the
Grant/Award Number: AGS-X 100 kN pristine composites. Finally, the field emission scanning electron microscopy
UTM
(FESEM) analysis is carried out to investigate the fracture mechanisms.
Highlights
• Cohesive properties under mode II are estimated using standard fracture
tests.
• The homogenization technique is used to calculate the effective properties
of CFRP.
• The maximum % error among the experimental and the numerical peak
load is 3.01%.
• A higher density of hackles with an increase of MWCNTs up to 0.2 wt.%.
KEYWORDS
carbon nanotubes, carbon/epoxy composites, cohesive zone modeling, fractography,
homogenization, mode II crack propagation
1 | INTRODUCTION development. Composites exhibit superior strength-weight

ratios and better mechanical properties with higher chemi-
Carbon fiber-reinforced plastic (CFRP) composites are cal and corrosion resistance, providing them a suitable
widely utilized in the automotive, aerospace, marine, and replacement for metals. However, they are susceptible to
structural industries in the contemporary era of industrial delamination due to weak out-of-plane properties that lead
Polymer Composites. 2024;1–15. wileyonlinelibrary.com/journal/pc © 2024 Society of Plastics Engineers. 1

15480569, 0, Downloaded from https://4spepublications.onlinelibrary.wiley.com/doi/10.1002/pc.28337 by PRAN JYOTI SAIKIA - Indian Institute Of Technology Guwahati , Wiley Online Library on [18/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 SAIKIA ET AL.
to catastrophic failure upon external impact. The most The calendaring method has also proven effective for
common method for increasing the resistance to delami- homogenizing large volumes of CNTs and achieving higher
nation in CFRP is to add some toughening materials— production rate.13
such as CNTs, graphene, etc. to epoxy and polyester.1,2 Understanding the failure mechanism of advanced
The mechanical properties of the composite can be sig- materials and structures is essential for determining the
nificantly enhanced by incorporating the nanofillers into component's structural integrity. Delamination is
the polymer matrix.3,4 Carbon nanotubes (CNTs) are the most frequent failure mode in the composite lami-
widely used nanofillers due to their higher aspect ratio nate, which can be effectively predicted using the virtual
and surface area. Furthermore, the addition of CNTs sig- crack closure technique (VCCT) and the CZM. The
nificantly improves not only the mechanical properties but major drawback of the VCCT is that it is only valid for
also enhances the thermal and electrical properties of the LEFM problems. In the majority of cases of modern
polymer composites.5 The influence of single, double, and composites, this assumption (LEFM) is not valid since
MWCNTs on the mechanical properties of epoxy-based large fracture process zones develop. From that perspec-
nanocomposites was investigated by Gojny et al.6 The tive, the CZM can be considered an appealing method
addition of 0.5 wt.% of amino-functionalized double- for simulating delamination in composites. It combines
walled carbon nanotubes resulted in a significant increase the strength of material and fracture mechanics
(43%) in fracture toughness. Borowski et al.7 investigated approach; the former is used to initiate damage, while
the influence of MWCNTs on the mode I (opening mode) the latter regulates the crack growth.14 The failure
interlaminar fracture toughness (IFT) of CFRP composites. behavior of the CZM is defined using a constitutive rela-
Incorporating three different MWCNTs variations— tionship (popularly known as traction separation law—
0.5 wt.%, 1 wt.%, and 1.5 wt.% resulted in improvements of TSL) between the cohesive traction ðσ Þ to the relative
opening mode IFT by 25%, 20%, and 17%, respectively. separation ðδÞ. There are three essential parameters in
Burkov and Eremin8 examined the IFT of CFRP with TSLs; cohesive stiffness ðK Þ, strength ðσ c Þ, and fracture
varying concentrations of single-walled carbon nanotubes energy ðGC Þ.15 While Barenblatt16 and Dugdale17 first
(SWCNTs) from 0.1 wt.% to 0.5 wt.%. Although the pres- presented the idea of the CZM, different shapes of cohe-
ence of CNTs significantly increased (35%) in mode II IFT, sive laws have been proposed for a better representation
a notable decrease was observed in mode I IFT. In one of of the failure phenomenon, i.e., bilinear cohesive law for
the recent works, Shin and Kim9 examined the shear brittle fracture, exponential softening for more ductile
strength and mode II IFT of CNT buckypaper epoxy com- fracture.18,19 Bilinear cohesive law is popularly used for
posites reinforced with carbon fiber. A significant increase delamination in composites due to its simplicity in imple-
of 45.90% in the mode II IFT was observed with the inser- mentation and straightforward physical meaning.20–22
tion of the buckypaper in the middle ply of the laminate The key aspect of the CZM is to obtain its parameters
ahead of the precrack. The shear strength, however, was for the failure of any bonded interface. The CZM parame-
reduced by 4% in contrast to the pristine laminate. Gener- ters can be estimated either by an indirect or direct
ally, the MWCNTs are functionalized with carboxyl, approach. Most direct approaches used J-integral con-
hydroxyl, and amino-functionalized groups that present cepts, where the cohesive law can be estimated using spe-
diverse surface chemistries and reactivity, making them cific fracture characterization tests. A technique that is
applicable in various domains such as materials science, entirely dependent upon the J-integral to estimate the
nanotechnology, and biomedical. Most of the researchers cohesive law for double cantilever sandwich beams sub-
investigated the influence of amino and carboxyl- jected to pure bending moment was investigated by
functionalized MWCNTs and comparatively less emphasis Sørensen.23 Additionally, the effect of the loading rate
on investigating the effect of hydroxyl-functionalized and the adhesive layer thickness on the cohesive law's
MWCNTs on the mode II IFT of CFRPs, carried out in the shape was studied. Digital image correlation (DIC) tech-
existing study. niques combined with J-integral to estimate the cohesive
Although including CNTs in a composite matrix can laws for the bonded interface are also popular methods
improve the mechanical properties, it depends entirely on used in recent times.24,25 However, the processes men-
the effectiveness of the dispersion technique.10 As CNTs tioned above are time-consuming, requiring high-
are hydrophobic, they may tend to agglomerate, leading resolution instruments to calculate the J-integral values
to a potential decrease in the overall properties of the from the actual output.26 In one of the recent studies,
multiscale composites.7 Various dispersion techniques are Arrese et al.27 proposed a novel methodology based on
available, with one commonly used method being ultra- the correlation between the energy release rate (ERR)
sonication. Several researchers have extensively used ultra- and the crack tip opening displacement (CTOD) to deter-
sonication to effectively disperse CNTs within the resin.11,12 mine the cohesive law for unidirectional composite
SAIKIA ET AL. 3
laminates subjected to pure mode II conditions. They nano Technologies, India.31 The epoxy consists of resin
derived the cohesive law without the use of external dis- Epofine 1564 and hardener 3486-2 provided by FINE
placement measuring instruments and specified FINISH, INDIA. The matrix has a combined elastic mod-
cohesive laws. ulus of 3.42 GPa and Poisson's ratio of 0.30.
On the other hand, the indirect approach assumes a
predefined cohesive law, is relatively straightforward,
and involves computational intensive iterative proce- 2.1 | Dispersion of OH-MWCNTs into
dures/optimization strategies to fit the numerical data the epoxy resin
with the experimental findings. In that context, an
inverse approach based on distributed strain data was for- Three different MWCNTs wt.% of 0.1%, 0.2%, and 0.3%
mulated by Sorensen et al.28 Azevedo et al.29 employed are used in the present study. Due to the relatively lesser
an inverse curve fitting approach to examine the mode II density of hydroxyl-functionalized MWCNTs compared
cohesive law characteristics of bonded aluminum joints to other functional groups, it becomes operationally chal-
using Araldite® AV138, Araldite® 2015, and Sikaforce lenging to achieve a uniform dispersion of more than
7752 adhesives. They successfully derived a distinct set of 0.3 wt.%. Initially, the required proportions of MWCNTs
CZM parameters, including the strength and energy are thoroughly mixed with the resin for 5 min. The resin
parameters, for each specimen to effectively capture the mixture is subjected to a SONICS 500-watt, 20 kHz tip
behavior of the joints. However, in the present day, sonicator for 2 hrs. of sonication time. The sonicator's
the indirect approach is becoming less significant as it amplitude is fixed at 40% for the initial 5 min and 20% for
involves a tiresome trial-and-error process, and different the remaining 1 h and 55 min of sonication time, respec-
combinations of CZM parameters may provide identical tively. During the sonication process, the container is
load–displacement responses. placed in an ice bath to lower the temperature of the
Therefore, the novelty of the current work is to evaluate solution. The modified resin mixture is then subjected to
the cohesive zone parameters reliably with a routine uni- a magnetic stirring process for another 1.5 h at 600 RPM
versal testing machine (UTM) experimental setup that and 60 C. The solution is then cooled to its normal tem-
reduces the experimental complexity. Moreover, the varia- perature; estimated proportions of hardener are added to
tions of CZM parameters with MWCNTs content are also produce the nanofiller-modified epoxy resin. The
reported. The composites are made with pristine (neat) and MWCNTs modified epoxy resin was then kept in a vac-
other weight percentages of 0.1%, 0.2%, and 0.3% hydroxyl uum desiccator to eliminate air bubbles within the solu-
functionalized MWCNTs. A bilinear cohesive law is tion mixture.
assumed, and the strength and energy parameters of the
CZM for pure mode II loading are explicitly evaluated from
the SBS and ENF tests, respectively. The benefit of the pre- 2.2 | Fabrication of carbon
sent approach is the reduction in experimental complexity fiber-reinforced epoxy nanocomposites
as it doesn't need a specific crack sliding displacement
(CSD) monitoring device. The obtained CZM parameters Due to the higher aspect ratios, the MWCNTs may segre-
are imputed into a bilinear cohesive law to accurately pre- gate while passing through multiple layers of fibers. There-
dict the delamination behavior of MWCNTs-reinforced fore, a conventional hand layup and the vacuum bagging
CFRP composites under mode II loading. The FESEM process are adopted to fabricate the composite laminate.
analysis of fracture surfaces is also carried out to under- The UD carbon fabric is cut into 20 cm 20 cm dimen-
stand the undergoing failure mechanism. sions and placed on the top of the composite mold. The
MWCNTs modified resin is spread uniformly over the fiber
using a paintbrush and hand roller. The above procedure is
2 | EXPERIMENTAL repeated for all layers of fibers. Weight ratios between the
M E T H O DO L O G Y fiber and matrix are kept constant so that they roughly
equate to a 50% fiber-to-matrix volume fraction. The fiber-
The unidirectional carbon fiber (grade 200 GSM) is sup- to-matrix ratio is based upon ensuring a uniform mixing of
plied by FIBER REGION, INDIA. The fiber diameter is fibers with the epoxy resin and FESEM imaging.32 A Teflon
5–8 μm with the elastic modulus, and the tensile strength sheet is placed in the middle half of the laminate and
equals 230 GPa and 2750 MPa, respectively.30 The elastic extended up to a length of 70 mm to fabricate the three-
modulus of hydroxyl functionalized MWCNTs is point bending (TPB) specimens. A peel ply and green mesh
400 GPa, with the geometric dimensions of length l ¼ 5– are used to cover the entire composite assembly. Finally, a
10 μm and diameter d ¼ 2–15 nm purchased from Ad vacuum bag is used to seal the assembly using sealant tape,
4 SAIKIA ET AL.
F I G U R E 1 Composite fabrication
process (A) Manual mixing of MWCNTs
with the resin (B) Ultra-sonication
(C) Magnetic stirring (D) Composite
fabrication process (E) Fabricated composite
plate.
which is then vacuumed. The vacuum creates a pressure layers. The SBS test is a simple, popularly used
gradient and compresses the stacked fabric against the method for estimating the ILSS of unidirectional
external pressure that helps in increasing the fiber vol- CFRP composites.33 Here, the specimens are sup-
ume fraction. After curing for a day, the composite lami- ported between two cylindrical supports with another
nate is removed from the assembly and postcured at cylinder head connected to the machine actuator to
80 C for a day. With the 0.20 mm thickness of individ- apply load at the center of the specimens. The failure
ual carbon fiber mats, the number of carbon fiber mats load of the specimens is used to calculate the ILSS of
used for the TPB and SBS tests is 16 and 30, respectively. the composites.
The entire fabrication process is shown in Figure 1. The SBS test is performed in compliance with the
standard ASTM D 2344.34 The specimens are cut into a
dimension of 36 12 6 mm3 and subjected to TPB test
2.3 | Mechanical testing in a SHIMADZU 100 kN UTM with 1 mm/min crosshead
speed, as shown in Figure 2A,B.
Mechanical testing is essential to understand the inter- Classical Euler-Bernoulli beam theory is used to eval-
face failure of laminated structures. These tests offer uate the ILSS of the composite laminate along the inter-
valuable design information, allowing us to evaluate the face. For a rectangular specimen geometry, the ILSS can
performance of the composite in specific applications. To be evaluated as
ensure consistency in the results, it is imperative
to adhere to distinct standards (ASTM or ISO) for each Pmax
τs ¼ 0:75 ð1Þ
testing procedure. The SBS and ENF tests are widely bt
employed to assess the interlaminar shear strength (ILSS)
and IFT of the composite, respectively. Three specimens where Pmax is the peak load under the failure response. b
are tested for pristine and each variation of MWCNTs for and t are the width and thickness of the specimen,
all test configurations. respectively. The distance between the support rollers is
four times the specimen thickness.
The main advantage of the SBS technique is its ease
2.3.1 | Estimation of interlaminar shear of use. Making the specimens is not too difficult, and the
strength test procedures require minimal fixture setup. Numerous
studies were conducted to estimate the ILSS of the unidi-
A transverse load on the composite laminate gener- rectional CFRP composite based on SBS tests.35,36 How-
ates shear stress between the individual composite ever, the reliability of this test is questionable as, in some
SAIKIA ET AL. 5
FIGURE 2 SBS test (A) specimen geometry (B) experimental setup.
FIGURE 3 Mode II fracture toughness (A) specimen geometry (B) experimental setup.
cases, the composite laminate may fail by flexure and which will be used for conducting the compliance
crushing at the point of load application or a combination calibration (CC) tests. The estimated value of the
of the two. The specimen is considered to undergo inter- peak force for the CC test is 50% of the maximum
laminar shear failure if the failure load of the specimen possible value of the critical force at that particular
drops by 30% once it reaches the peak load. crack length. Another marking is made 30 mm from
the crack tip, with a support span of 100 mm, which
will be used for the fracture test. The detailed testing
2.3.2 | Estimation of mode II interlaminar procedure for estimating the mode II fracture tough-
fracture toughness (IFT) ness is provided in the standard ASTM D 7905.37,38
Here, the mode II IFT is evaluated as
For the ENF test, the specimens are precisely trimmed
and polished to a dimension of 140 21 3:2 mm3. The 3mPmax a20
GIIC ¼ ð2Þ
specimen comprised 50 mm of delamination length, 2b
with a 30 mm crack length from the roller support, as
shown in Figure 3A. Finally, the tests are performed where m is the compliance calibration coefficient
with 0.50 mm/min crosshead speed, as shown in obtained by plotting compliance against the cube of the
Figure 3B. crack length a0 . b and Pmax are the width of the specimen
White paint is applied on the specimens' edges to and peak load under the fracture test, respectively. a0 is
visualize the delamination tip. Marks are made on the crack length used for the fracture test and has a value
the edges 20 and 40 mm from the Teflon insert's tip, of 30 mm.
6 SAIKIA ET AL.
3 | R ESULTS A ND DISCUSSIONS The average ILSS value of the pristine laminate is

approximately 28.21 MPa, which increases up to 32.20 MPa
Following the experimental procedures, the load– for 0.2 wt.% of MWCNTs modified CFRP. The average ILSS
displacement behavior of pristine and MWCNTs-reinforced
specimens is plotted. The mean interlaminar shear stress
and the IFT is evaluated using the load–displacement
behavior, employing Equations (1) and (2), respectively.
3.1 | Interlaminar shear stress (ILSS)

studies
The load–displacement behavior of all the specimens

subjected to the SBS test is shown in Figure 4A–D.
An interlaminar shear failure of the specimens is
indicated by a gradual increase in load with displace-
ment up to the peak load and then an abrupt
decrease. The average ILSS of pristine and nanofiller-
modified specimens with their standard deviations F I G U R E 5 Variation of average interlaminar shear stress for
are shown in Figure 5. different MWCNTs variations.
F I G U R E 4 Load–displacement response obtained from the SBS test (A) Pristine, and three different MWCNTs variations—(B) 0.1 wt.%,
(C) 0.2 wt.%, (D) 0.3 wt.%.
SAIKIA ET AL. 7
F I G U R E 6 Load–displacement response obtained from mode II fracture tests (A) Pristine and three MWCNTs variations—(B) 0.1 wt.%,
(C) 0.2 wt.%, (D) 0.3 wt.%.
values for 0.1 wt.% of MWCNTs and 0.3 wt.% reinforced ENF test are shown in Figure 6A–D. The load increases
MWCNTs are approximately 31.52 and 30 MPa, respec- linearly for all the specimens, followed by a nonlinear
tively. More MWCNT content enhances the adherence of increase in load with the displacement before it reaches
fibers and matrix and facilitates efficient stress transfer its peak. Once the load reaches its optimum value (peak
between MWCNTs and epoxy. Nonetheless, it has been load), the crack propagates, and as a result, the load
observed that a rise in MWCNT content correspondingly decreases rapidly. For mode II fracture tests, there is no
raises the matrix's viscosity, which leads to poor wettability fiber bridging mechanism. Therefore, the mode II IFT is
and fiber-matrix adhesion. Accordingly, a decrease in inter- assumed to be the crack initiation fracture toughness
laminar shear stress is observed for CFRP reinforced with ðGIIC Þ corresponding to the peak load.
higher wt.% of MWCNTs (0.3%). The fracture toughness values for pristine and three
different variations of MWCNTs (0.1 wt.%, 0.2 wt.%, and
0.3 wt.%) reinforced composites are approximately 2.08,
3.2 | Mode II IFT studies 2.35, 2.45, and 2.15 kJ/m2, as shown in Figure 7. Among
all variations, 0.2 wt.% of MWCNTs reinforced CFRP
The load–displacement behavior for pristine and all wt.% exhibits maximum fracture toughness. The epoxy resin
variations of MWCNTs reinforced CFRP subjected under will become tougher with the inclusion of MWCNTs.
8 SAIKIA ET AL.
FIGURE 8 Bilinear cohesive law.
FIGURE 7 Average fracture toughness of different MWCNTs

δfs δmax δcs
variations. d¼ ð5Þ
δmax δfs δcs
However, for larger values of MWCNTs wt.%, the nano- δmax defines the instantaneous displacement during
particle may try to agglomerate under the strong Van der the loading history. δcs and δfs represent damage initiation
Waals forces.7 This may be the possible reason that and final displacement, respectively. δcs and δfs for a bilin-
decreases the fracture toughness value at a higher wt.% ear cohesive law, as depicted in Figure 8, can be
(0.3%) variation of MWCNTs. obtained as:
τc
δcs ¼ ð6Þ
4 | NUMERICAL VALIDATION K
The CZM technique is used to simulate mode II interla- and

minar failure. A bilinear CZM, as depicted in Figure 8, is
used for validation as it agrees well with the experimental 2GIIC
δfs ¼ ð7Þ
failure load and progressive delamination response.39 τc
The maximum stress criteria mentioned in Equation (3)
is used for damage initiation, which implies that the The ENF specimens are numerically modeled using
damage is initiated when the stress ratio is equal to one. the FE package ABAQUS 2017 ®. The objective of the
The damage initiation criteria can be defined as. numerical modeling is to achieve a deeper understand-
ing of how MWCNTs influence the overall load–

τ displacement response. The strength and energy
max ¼1 ð3Þ
τc parameters of the CZM are obtained experimentally
using the SBS and ENF tests. The numerical model is
where τ represents the instantaneous shear stress in the fitted with the experimentally determined CZM
first direction of shear. τc represents the shear strength in parameters to predict the delamination phenomenon
the respective direction. observed in the composite laminate under pure mode
Once the damage initiation criteria are satisfied, the II loading conditions.
material interface will undergo damage evolution. The cohesive stiffness K is an essential parameter for
The softening phenomena are defined using a linear sca- undamaged cohesive elements to establish stiff connec-
lar damage variable, d. Initially, d has a zero value. Fol- tions with the adjacent layers. A larger value of cohesive
lowing the onset, the damage gradually increases from stiffness causes oscillations of traction in the elements,
zero to one until the component fails. The instantaneous whereas a too low value generates fictitious compliance
value of stress after damage initiation is given as in the numerical model. In this present study, the cohe-
sive stiffness is considered 1 1013 N/m3, which is in the
τ ¼ ð1 dÞτc ð4Þ range of 1012 1015 N/m3, used by most researchers.40
The dimensions of the ENF specimen geometry are
Additionally, the damage variable d is defined as identical to the tested specimens used in the
SAIKIA ET AL. 9
experimental section. Initially, the effective elastic modu-

lus of the MWCNTs reinforced matrix is evaluated. The
elastic constants of a composite laminate reinforced
with fibers and MWCNTs are determined based on the
properties of the fibers and MWCNTs, using the theory
of the rule of mixtures and representative volume ele-
ment (RVE) analysis.
4.1 | Evaluation of the effective

properties of nanocomposites FIGURE 9 Representative volume element (RVE).
Computational modeling is an efficient tool for esti-

mating the effective properties of CNTs reinforced As the RVEs can be assembled to form the complete
composite laminates with intricate morphologies. Var- lamina, the displacement should be continuous at the
ious methods, including the rule of mixture and the boundary of the neighboring RVE. Furthermore, there
strength of materials concept, can be used to evaluate should be an equal and opposite traction distribution at
the overall properties of the composite laminate. the opposing parallel faces of the RVE. The following set
Another effective approach is the homogenization of equations can be used to represent the displacement
technique, which utilizes the FE tool to convert a het- field of the pair of opposite parallel faces.
erogeneous substance into an equivalent continuous

homogeneous medium. As it is physically challenging ujþ jþ
i ¼ ϵik x k þ ui
to estimate all orthotropic properties of the composite
laminate experimentally, a computational approach
uj j
i ¼ ϵik x k þ ui ð10Þ
has been adopted.
Here, a two-step methodology is adopted to evaluate
the equivalent properties. Initially, the equivalent proper- where u is the periodic function from one RVE to
ties of epoxy nanocomposites are computed using the another, k þ and k denote the k th pair of opposite paral-
rule of mixture. If ρr denotes the density ratio between lel boundary surfaces of the RVE, and ϵik is the global
the MWCNTs and epoxy, the volume fraction of strain tensor of the periodic structure. For the opposing
MWCNTs corresponding to a particular weight percent- faces of RVE, the periodic component of the displace-
age can be derived as follows41 ment field u is the same, which results in
1
ρr ujþ j jþ j
i ui ¼ ϵik x k x k ¼ ϵik Δx jk ð11Þ
vcnt ¼ ρr þ 1 ð8Þ
wcnt
Once the displacement field is obtained, the average

where wcnt is the proportion of MWCNTs by weight. The stresses and strains in the RVE can be evaluated. Finally,
effective modulus of epoxy-based nanocomposite E nc is the specific elastic modulus Cij is obtained from the aver-
calculated using the rule of mixtures using the formula age values of stress and strain. The detailed formulation
procedure is already mentioned in some earlier published
E nc ¼ E m vm þ E cnt vcnt ð9Þ works.32,42
A cylindrical fiber that corresponds to a particular
where E m , vm, and E cnt , vcnt represent Young's modulus fiber volume fraction of V f ¼ 0.50 (shown in Figure 9) is
and the volume percentage of MWCNTs and resin, assumed inside a cubical RVE in the present analysis.
correspondingly. The EasyPBC plugin43 is utilized in the ABAQUS to
Next, an RVE is assumed, and the homogenization access the equivalent properties of the composites for a
technique is adopted to compute the equivalent proper- given fiber volume fraction. The equivalent elastic prop-
ties of the nanocomposites with reinforced fibers. erties obtained for pristine and different wt.% of
Because of the recurring patterns within the composites, MWCNTs are mentioned in Table 1. The longitudinal
the RVE shown in Figure 9 is appropriate. The homoge- elastic modulus ðE xx Þ improves marginally with the addi-
nization technique is employed through periodic bound- tion of MWCNTs, aligned with the observation of earlier
ary conditions (PBCs). published literature.44
10 SAIKIA ET AL.
TABLE 1 Material properties for multiscale composites with all variations of MWCNTs.
wt:% MWCNTs Exx ðGPaÞ E yy ¼ E zz ðGPaÞ νxy ¼ νxz νyz Gxy ¼ Gxz ðGPaÞ Gyz ðGPaÞ
Neat 116:42 11:24 0:24 0:26 3:88 2:81
0:1% 116:54 11:95 0:24 0:26 4:13 2:99
0:2% 116:65 12:65 0:24 0:26 4:38 3:17
0:3% 116:77 13:35 0:24 0:26 4:62 3:36
F I G U R E 1 0 (A) FE mesh of the model

geometry. (B) Deformed configuration of the
model geometry.
A 3-D deformable solid part is created for each half of the bonded side of the laminate. The remaining length
the ENF specimens. Although the 2D FEA can be accu- along the bonded interface contains cohesive surface
rately used to predict the delamination in simplified ENF interaction property, allowing crack propagation along
specimen geometry,45 the purpose of using the 3D FEA is the predefined path. The specimen geometry is posi-
to represent the model's geometry and boundary condi- tioned with a span length of 100 mm between two fixed
tions accurately. Moreover, 3D FEA can be used to obtain roller supports while the top roller is subjected to a down-
the distribution of stress and strain throughout the entire ward displacement of 6 mm.
volume of the structure. Upon examining the crack slid- The FE simulations are carried out for pristine and all
ing distance, it has been observed from the simulations wt.% of MWCNTs. The initial and deformed FE model
that the crack front mostly remains straight as the crack for the ENF specimens is shown in Figure 10A,B, respec-
propagates. tively. It is observed that the numerical model's load–
Following the mesh convergence study, 9052 8-noded displacement response closely matches the outcomes of
linear brick elements with a global size of 0.80 mm are the experiments. However, a minimal discrepancy in load
taken into consideration on either half of the beam. To displacement is observed during the crack growth phe-
generate the precrack for initial delamination, no contact nomenon in all cases, as shown in Figure 11A–D. It may
interaction is provided for the initial 50 mm length along be due to the use of the constant value of interface
SAIKIA ET AL. 11
F I G U R E 1 1 Comparison between experimental and numerical results for (A) Pristine and three MWCNTs variations—(B) 0.1 wt.%,
(C) 0.2 wt.%, (D) 0.3 wt.%.
T A B L E 2 Disparity between the peak load calculated not taken into account by classical fracture criteria.
numerically and experimentally. Nonetheless, the FE model provides insight into how
MWCNTs affect CFRP's mode II interlaminar fracture
Experimental Numerical
wt:% of peak load peak load toughness. The difference between the experimental and
P exp P num
MWCNTs (N) P exp (N) ðP num Þ % Error numerical peak load with % error is mentioned in
P exp
Table 2.
Neat 701:10 717:04 3:01%
0:1% 748:04 767:35 2:58%
0:2% 756:46 771:17 1:98% 5 | S T U D Y OF SU R F A C E
0:3% 708:32 726:47 2:57% M O R PH O LO G Y
The specimen's fracture surfaces are studied to compre-

fracture energy ðGIIC Þ in the numerical model, although hend the impact of MWCNTs on the toughening mecha-
the experimental fracture toughness changes with the nism. After conducting the mode II fracture tests, the
delamination growth.7 Additionally, adding MWCNTs interface of the specimens ahead of the precrack are
results in non-linearity in interlaminar fracture tough- chopped into a dimension of 5 mm 5 mm and placed
ness, and the failure mechanism of nanocomposites is under ZEISS Gemini 300 FESEM. The fractured surfaces
12 SAIKIA ET AL.
FIGURE 12 Fracture surfaces under FESEM (A) Pristine and three MWCNTs variations—(B) 0.1 wt.%, (C) 0.2 wt.%, (D) 0.3 wt.%.
of the pristine specimens are relatively smooth and clean A specific surface texture of the shear lip and saw-
compared to the other MWCNTs-reinforced specimens, as tooth hackles is observed in MWCNTs reinforced speci-
shown in Figure 12A. In addition, several resin-rich areas mens. The hackles and the shear lip are prominent on
and broken fibers are observed in the pristine specimen. the fractured surface with increasing contents of
On the other hand, the surface texture of the MWCNTs MWCNTs (prominently for 0.2 wt.% of MWCNTs content
reinforced specimens is relatively rougher. The crack- observed from Figure 12C). Furthermore, nano-fillers in
deflecting mechanism in the presence of MWCNTs is the epoxy resin repelled the matrix cracking by a bridging
mainly responsible for uneven surface morphology.46,47 mechanism, which requires a more considerable amount
SAIKIA ET AL. 13
of energy for crack propagation. The MWCNTs rein- ACKNOWLEDGMENTS

forced specimens, therefore, exhibit a larger fracture The work was supported by the Science and Engineering
toughness compared to pristine composite specimens. Research Board, India, under the scheme ‘Early Career
The highly dense MWCNTs may result in agglomera- Research Award’, No: ECR/2018/001638. The authors
tion of MWCNTs in the reinforced composite speci- also acknowledge the DST-FIST funded facility,
mens.8,44 Agglomeration lowers the stress transfer Shimadzu AGS-X 100 kN UTM, at the Department of
between MWCNTs and matrix and fiber/matrix adhe- Mechanical Engineering, Indian Institute of Technology
sion, respectively.48 This may be the possible reason why Guwahati.
the fracture toughness of 0.3 wt.% of MWCNT specimens
declined compared to the other MWCNTs reinforced C O N F L I C T O F I N T E R E S T S T A TE M E N T
specimens. The authors declare no conflict of interest.
DA TA AVAI LA BI LI TY S T ATE ME NT
6 | C ON C L U S I ON S Raw data supporting the findings of this study will be
made available on request.
In the present study, the CZM parameters are estimated
experimentally for pristine and MWCNTs-reinforced
ORCID
CFRP composite laminate under pure mode II condi-
Nelson Muthu https://orcid.org/0000-0002-2589-3234
tions. The mode II cohesive strength τc and fracture
energy GIIC for all cases are estimated from the SBS and
RE FER EN CES
ENF tests, respectively. The cohesive stiffness is, how-
1. Tanoglu M, Tugrul SA. Investigating the effects of a polyester
ever, estimated iteratively based on the previous litera-
preforming binder on the mechanical and ballistic perfor-
ture. Estimating cohesive stiffness from experimental mance of E-glass fiber reinforced polyester composites. Int J
studies or other direct approaches will reduce the compu- Adhes Adhes. 2003;23(1):1-8. doi:10.1016/S0143-7496(02)
tational effort on interface failure and can be considered 00061-1
part of the future scope. The equivalent orthotropic prop- 2. Zhu J, Imam A, Crane R, Lozano K, Khabashesku V,
erties of the nanofiller-reinforced CFRP are calculated Barrera E. Processing a glass fiber reinforced vinyl ester com-
with the help of the rule of mixtures and homogenization posite with nanotube enhancement of interlaminar shear
strength. Compos Sci Technol. 2007;67(7–8):1509-1517. doi:10.
technique. The obtained material properties and cohesive
1016/j.compscitech.2006.07.018
parameters are directly imputed into the FE model with 3. Liu W, Li L, Zhang S, Yang F, Wang R. Mechanical properties
a bilinear TSL to validate the load–displacement response of carbon nanotube/carbon fiber reinforced thermoplastic poly-
for pristine and all MWCNT proportions. The agreement mer composite. Polym Compos. 2017;38(9):2001-2008. doi:10.
between the numerical results with the experimental 1002/pc.23771
findings validates the suitability of SBS and ENF tests for 4. Sahu R, Ponnusami SA, Weimer C, Harursampath D. Interface
determining the mode II CZM parameters of MWCNTs engineering of carbon fiber composites using CNT: a review.
reinforced CFRP. The highest percentage error in the Polym Compos. 2023;45(1):9-42. doi:10.1002/pc.27772
peak load between the numerical and experimental data 5. Saikia PJ, Muthu N. Extrinsic/intrinsic cohesive zone model-
ling and experiments on interface failure of multi-walled car-
is less than 3.01%.
bon nanotube reinforced adhesively bonded joints under mode
The enhancement of IFT of the CFRP with the accu- II loading. Theor Appl Fract Mech. 2023;126(May):103967. doi:
mulation of relatively less utilized hydroxyl functionalized 10.1016/j.tafmec.2023.103967
MWCNTs is another finding of the current work. In com- 6. Gojny FH, Wichmann MHG, Köpke U, Fiedler B, Schulte K.
parison to the pristine composites, it is seen that the addi- Carbon nanotube-reinforced epoxy-composites: enhanced stiff-
tion of 0.1 wt.% and 0.2 wt.% of MWCNTs enhances the ness and fracture toughness at low nanotube content. Compos
fracture energy to 12.98% and 17.78%, respectively. How- Sci Technol. 2004;64(15):2363-2371. doi:10.1016/j.compscitech.
ever, higher MWCNTs contents, that is, 0.3 wt.% of 2004.04.002
MWCNTs, reduce the effective stress transfer and IFT 7. Borowski E, Soliman E, Kandil U, Taha M. Interlaminar frac-
ture toughness of CFRP laminates incorporating multi-walled
of the composites. The same is confirmed from the investi-
carbon nanotubes. Polymers (Basel). 2015;7(6):1020-1045. doi:
gations of the FESEM, which reveal that as the MWCNTs' 10.3390/polym7061020
content increases until 0.2 wt.%, the interlaminar fracture 8. Burkov M, Eremin A. Evaluation of fracture toughness of
surface shows a higher density of hackles. This higher den- hybrid CNT/CFRP composites. Mech Adv Mater Struct. 2023;
sity suggests an increased resistance to crack propagation. 30(14):2872-2881. doi:10.1080/15376494.2022.2064569
However, the presence of hackles gets reduced with any 9. Shin YC, Kim SM. Enhancement of the interlaminar fracture
further increase in the wt.% of MWCNTs. toughness of a carbon-fiber-reinforced polymer using
14 SAIKIA ET AL.
interleaved carbon nanotube Buckypaper. Appl Sci. 2021; of cohesive laws. Compos Part A Appl Sci Manuf. 2016;90:234-
11(15):6821. doi:10.3390/app11156821 242. doi:10.1016/j.compositesa.2016.07.004
10. Esawi AMK, Salem HG, Hussein HM, Ramadan AR. 25. Dias GF, de Moura MFSF, Chousal JAG, Xavier J. Cohesive
Effect of processing technique on the dispersion of car- laws of composite bonded joints under mode I loading. Com-
bon nanotubes within polypropylene carbon nanotube- pos Struct. 2013;106:646-652. doi:10.1016/j.compstruct.2013.
composites and its effect on their mechanical properties. 07.027
Polym Compos. 2010;31(5):772-780. doi:10.1002/pc.20859 26. Abdel Monsef S, Ortega A, Turon A, Maimí P, Renart J. An
11. Miyagawa H, Mohanty AK, Drzal LT, Misra M. Nanocompo- efficient method to extract a mode I cohesive law for bonded
sites from biobased epoxy and single-wall carbon nanotubes: joints using the double cantilever beam test. Compos Part B
synthesis, and mechanical and thermophysical properties eval- Eng. 2019;178(June):107424. doi:10.1016/j.compositesb.2019.
uation. Nanotechnology. 2005;16(1):118-124. doi:10.1088/0957- 107424
4484/16/1/024 27. Arrese A, Insausti N, Mujika F, Perez-Galmés M, Renart J. A
12. Yu N, Zhang ZH, He SY. Fracture toughness and fatigue life of novel experimental procedure to determine the cohesive law in
MWCNT/epoxy composites. Mater Sci Eng A. 2008;494(1–2): ENF tests. Compos Sci Technol. 2019;170(April 2018):42-50.
380-384. doi:10.1016/j.msea.2008.04.051 doi:10.1016/j.compscitech.2018.11.031
13. Jiménez-Suarez A, Campo M, Sanchez M, Rom on C, Ureña A. 28. Sorensen L, Botsis J, Gmür T, Humbert L. Bridging tractions in
Dispersion of carbon nanofibres in a low viscosity resin by cal- mode I delamination: measurements and simulations. Compos
endering process to manufacture multiscale composites by Sci Technol. 2008;68(12):2350-2358. doi:10.1016/j.compscitech.
VARIM. Compos Part B Eng. 2012;43(8):3104-3113. doi:10. 2007.08.024
1016/j.compositesb.2012.04.030 29. Azevedo JCS, Campilho RDSG, da Silva FJG, Faneco TMS,
14. Fernandes RMRP, Chousal JAG, De Moura MFSF, Xavier J. Lopes RM. Cohesive law estimation of adhesive joints in mode
Determination of cohesive laws of composite bonded joints II condition. Theor Appl Fract Mech. 2015;80:143-154. doi:10.
under mode II loading. Compos Part B Eng. 2013;52:269-274. 1016/j.tafmec.2015.09.007
doi:10.1016/j.compositesb.2013.04.007 30. Hollaway LC. Key issues in the use of fibre reinforced polymer
15. Nafar Dastgerdi J, Lotf Yasouri M, Hosseini TH. Microstruc- (FRP) composites in the rehabilitation and retrofitting of con-
ture-sensitive investigation on the mechanical behavior of crete structures. Serv Life Estimation Ext Civ Eng Struct. 2011;3-
CNT-reinforced composites considering debonding damage 74. doi:10.1533/9780857090928.1.3
based on cohesive finite element method. Mater Today Com- 31. Kalamkarov AL, Georgiades AV, Rokkam SK, Veedu VP,
mun. 2022;31(January):103458. doi:10.1016/j.mtcomm.2022. Ghasemi-Nejhad MN. Analytical and numerical techniques to
103458 predict carbon nanotubes properties. Int J Solids Struct. 2006;
16. Barenblatt GI. The mathematical theory of equilibrium cracks 43(22–23):6832-6854. doi:10.1016/j.ijsolstr.2006.02.009
in brittle fracture. Advances in Applied Mechanics. Vol 7. Else- 32. Rakesh P, More A, Kumar M, Muthu N. Probabilistic fail-
vier; 1962:55-129. doi:10.1016/S0065-2156(08)70121-2 ure prediction in a double composite cantilever beam with
17. Dugdale DS. Yielding of steel sheets containing slits. J Mech Phys single and double source uncertainty. Compos Struct. 2022;
Solids. 1960;8(2):100-104. doi:10.1016/0022-5096(60)90013-2 279(August 2021):114870. doi:10.1016/j.compstruct.2021.
18. Camacho GT, Ortiz M. Computational modelling of impact 114870
damage in brittle materials. Int J Solids Struct. 1996;33(20–22): 33. Kirmse S, Ranabhat B, Hsiao KT. Experimental and analytical
2899-2938. doi:10.1016/0020-7683(95)00255-3 investigation on the interlaminar shear strength of carbon fiber
19. Tvergaard V, Hutchinson JW. The relation between crack composites reinforced with carbon nanofiber z-threads. Mater
growth resistance and fracture process parameters in elastic- Today Commun. 2020;25(January):101512. doi:10.1016/j.
plastic solids. J Mech Phys Solids. 1992;40(6):1377-1397. doi:10. mtcomm.2020.101512
1016/0022-5096(92)90020-3 34. Precision WS. Standard test method for short-beam strength of
20. Wang K, Zhao L, Hong H, Zhang J. A strain-rate-dependent polymer matrix composite materials. Annu B ASTM Stand.
damage model for evaluating the low velocity impact induced 2011;1-8. doi:10.1520/D2344
damage of composite laminates. Compos Struct. 2018;201- 35. Fan Z, Santare MH, Advani SG. Interlaminar shear strength of
(June):995-1003. doi:10.1016/j.compstruct.2018.06.046 glass fiber reinforced epoxy composites enhanced with multi-
21. Zhao L, Wang Y, Zhang J, Gong Y, Hu N, Li N. XFEM-based walled carbon nanotubes. Compos Part A Appl Sci Manuf.
model for simulating zigzag delamination growth in laminated 2008;39(3):540-554. doi:10.1016/j.compositesa.2007.11.013
composites under mode I loading. Compos Struct. 2017;160: 36. Rahman MM, Zainuddin S, Hosur MV, et al. Effect of
1155-1162. doi:10.1016/j.compstruct.2016.11.006 NH2-MWCNTs on crosslink density of epoxy matrix and ILSS
22. Banea MD, Da Silva LFM, Campilho RDSG. Mode i fracture properties of e-glass/epoxy composites. Compos Struct. 2013;95:
toughness of adhesively bonded joints as a function of tempera- 213-221. doi:10.1016/j.compstruct.2012.07.019
ture: experimental and numerical study. Int J Adhes Adhes. 37. ASTM D7905. Standard test method for determination of the
2011;31(5):273-279. doi:10.1016/j.ijadhadh.2010.09.005 mode II interlaminar fracture toughness of unidirectional
23. Sørensen BF. Cohesive law and notch sensitivity of adhesive fiber-reinforced polymer matrix composites. Astm. 2014. doi:10.
joints. Acta Mater. 2002;50(5):1053-1061. doi:10.1016/S1359- 1520/D7905
6454(01)00404-9 38. Prasad V, Sekar K, Varghese S, Joseph MA. Enhancing mode I
24. Sarrado C, Turon A, Costa J, Renart J. An experimental analy- and mode II interlaminar fracture toughness of flax fibre rein-
sis of the fracture behavior of composite bonded joints in terms forced epoxy composites with nano TiO2. Compos Part A Appl
SAIKIA ET AL. 15
Sci Manuf. 2019;124(May):105505. doi:10.1016/j.compositesa. 45. Yang QD, Thouless MD, Ward SM. Elastic–plastic mode-II
2019.105505 fracture of adhesive joints. Int J Solids Struct. 2001;38(18):3251-
39. Katsivalis I, Thomsen OT, Feih S, Achintha M. Development 3262. doi:10.1016/S0020-7683(00)00221-3
of cohesive zone models for the prediction of damage 46. Silva H, Ferreira JAM, Capela C, Richardson MOW. Mixed
and failure of glass/steel adhesive joints. Int J Adhes mode interlayer fracture of glass fiber/nano-enhanced epoxy
Adhes. 2020;97(November 2019):102479. doi:10.1016/j. composites. Compos Part A Appl Sci Manuf. 2014;64:211-222.
ijadhadh.2019.102479 doi:10.1016/j.compositesa.2014.05.011
40. Zhao L, Gong Y, Zhang J, Chen Y, Fei B. Simulation of 47. Wetzel B, Rosso P, Haupert F, Friedrich K. Epoxy
delamination growth in multidirectional laminates under nanocomposites – fracture and toughening mechanisms. Eng
mode I and mixed mode I/II loadings using cohesive ele- Fract Mech. 2006;73(16):2375-2398. doi:10.1016/j.engfracmech.
ments. Compos Struct. 2014;116(1):509-522. doi:10.1016/j. 2006.05.018
compstruct.2014.05.042 48. Gojny F, Wichmann M, Fiedler B, Schulte K. Influence of differ-
41. Tjong SC. Carbon Nanotube Reinforced Composites. Vol 3. ent carbon nanotubes on the mechanical properties of epoxy
Wiley; 2009. doi:10.1002/9783527626991 matrix composites – a comparative study. Compos Sci Technol.
42. Kumar A, Muthu N, Narayanan RG. Equivalent orthotropic 2005;65(15–16):2300-2313. doi:10.1016/j.compscitech.2005.04.021
properties of periodic honeycomb structure: strain-energy
approach and homogenization. Int J Mech Mater Des. 2023;
19(1):137-163. doi:10.1007/s10999-022-09620-x
43. Omairey SL, Dunning PD, Sriramula S. Development of an
ABAQUS plugin tool for periodic RVE homogenisation. How to cite this article: Saikia PJ, Kumar M,
Eng Comput. 2019;35(2):567-577. doi:10.1007/s00366-018- Kumar A, Muthu N. Experimental estimation and
0616-4 numerical validation of cohesive zone parameters
44. El Moumen A, Tarfaoui M, Lafdi K. Computational homoge-
in hydroxyl functionalized MWCNTs-reinforced
nization of mechanical properties for laminate composites
reinforced with thin film made of carbon nanotubes. Appl
CFRP under pure mode II loading. Polym Compos.
Compos Mater. 2018;25(3):569-588. doi:10.1007/s10443-017- 2024;1‐15. doi:10.1002/pc.28337
9636-2

My Paper - 3

Uploaded by

Copyright:

Available Formats

My Paper - 3

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

My Paper - 3

Uploaded by

Copyright:

Available Formats

Received: 26 December 2023 Revised: 6 March 2024 Accepted: 8 March 2024

Experimental estimation and numerical validation of

Pran Jyoti Saikia 1 | Munna Kumar 1 | Avneesh Kumar 1 | Nelson Muthu 2

1 | INTRODUCTION development. Composites exhibit superior strength-weight

Polymer Composites. 2024;1–15. wileyonlinelibrary.com/journal/pc © 2024 Society of Plastics Engineers. 1

FIGURE 2 SBS test (A) specimen geometry (B) experimental setup.

3 | R ESULTS A ND DISCUSSIONS The average ILSS value of the pristine laminate is

3.1 | Interlaminar shear stress (ILSS)

The load–displacement behavior of all the specimens

FIGURE 8 Bilinear cohesive law.

FIGURE 7 Average fracture toughness of different MWCNTs

The CZM technique is used to simulate mode II interla- and

experimental section. Initially, the effective elastic modu-

4.1 | Evaluation of the effective

Computational modeling is an efficient tool for esti-

Once the displacement field is obtained, the average

F I G U R E 1 0 (A) FE mesh of the model

The specimen's fracture surfaces are studied to compre-

of energy for crack propagation. The MWCNTs rein- ACKNOWLEDGMENTS

You might also like