Composites Part B 129 (2017) 1e10

Contents lists available at ScienceDirect

Composites Part B
journal homepage: www.elsevier.com/locate/compositesb

Effect of carbon nanotube and functionalized liquid rubber on

mechanical and electrical properties of epoxy adhesives for aircraft
structures
L. Vertuccio a, *, L. Guadagno a, **, G. Spinelli b, S. Russo c, G. Iannuzzo c
a
Department of Industrial Engineering, University of Salerno, Via Giovanni Paolo II, Fisciano (SA), Italy
b
Department of Information and Electrical Engineering and Applied Mathematics, University of Salerno, Via Giovanni Paolo II, Fisciano (SA), Italy
c
Leonardo SpA, Aircraft Division, Engineering Airframe, Viale dell'Aeronautica, Pomigliano D'Arco (NA), 80038, Italy

a r t i c l e i n f o a b s t r a c t

Article history: New electrical conductive adhesives based on Multi Wall Carbon Nanotubes (MWCNTs) and function-
Received 4 May 2017 alized liquid rubber have been designed and characterized. The elastomeric domains play a very relevant
Received in revised form role in enhancing flexibility and mechanical performance of the adhesive formulation. Lap shear adhe-
5 July 2017
sion tests have shown enhancements in the stress up to 69% for the sample containing 25 phr of elas-
Accepted 13 July 2017
Available online 17 July 2017
tomeric phase in the matrix. The inclusion of MWCNTs in the toughened adhesive can be advantageously
employed for further enhancing adhesive properties simultaneously imparting electrical conductivity,
which results of 11 orders of magnitude higher than the unfilled formulation.
Keywords:
Carbon-carbon composites (CCCs)
© 2017 Elsevier Ltd. All rights reserved.
Adhesion
Electrical properties
Mechanical testing

1. Introduction moment resisting bolted beam-column connections in GFRP

The adhesively bonded joints, in which an adhesive is placed
between the surfaces of adherents, are increasingly being pursued
as alternatives to mechanical joints in load-bearing structures such
as aerospace, automotive or other engineering applications [1e5].
For instance, in the civil engineering field, the feasibility of adhe-
sively bonded connections in FRP frame structures is explored as an
alternative to bolted connections [1]. The adhesive connections
based on an epoxy resin available on the market seem to have
higher strength compared to the steel bolted connections [1]. From
a theoretical point of view, bonded connections can be superior to
bolted connections because the holes made in structural members
with bolted connections cause stress concentrations and increase
the risk of moisture penetration in the members [1]. In addition, as
discussed by the authors of Ref. [1], as it concerns the current
* Corresponding author.
** Corresponding author.
E-mail addresses: lvertuccio@unisa.it (L. Vertuccio), lguadagno@unisa.it
(L. Guadagno), gspinelli@unisa.it (G. Spinelli), salvatore.russo@leonardocompany.
com (S. Russo), generoso.iannuzzo@leonardocompany.com (G. Iannuzzo).
structures a limit of 20% of the flexural capacity of the connected
members has been found. Bonded lap joint connections are able to
partially overcome some of these drawbacks because the absence
of hole allows a uniform distribution of the stresses over the
bonded surfaces [1]. Furthermore, adhesive formulations, in the
transport field, can provide many other advantages over conven-
tional mechanical fasteners such as cheaper manufacturing cost,
improved damage tolerance [6] and lower weight making the
structures able to contrast the rapid depletion of fossil and the
aviation environmental impact. In the electronic industry, electri-
cally conductive adhesives (ECAs) are also employed; conventional
ECAs require 25e30% by volume (80% by weight) of conductive
filler [6]. Due to the high filler loading, most commonly silver
particles, only a small portion of the shear strength is retained with
respect to the unfilled adhesive and existing ECAs are not suitable
for aerospace structural applications. Although structural adhesives
are already common on aircrafts, riveting is still the only current
solution for electrical bonding [7]. A significant progress in the field
of structural joints is the replacement of traditional rivets with
structural epoxy conductive adhesives. The possibility to use ad-
hesives for high structural materials (aeronautical, naval or

http://dx.doi.org/10.1016/j.compositesb.2017.07.021
1359-8368/© 2017 Elsevier Ltd. All rights reserved.
2 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
infrastructural etc.) can simplify the design process of composite
structures by reducing the thickness of structural components with
consequent reduction of the material weight. Furthermore,
together with a cheaper joining process, the use of adhesive
bonded joints entails other advantages among which a) the
corrosion prevention: adhesive joints are not as sensible to crevice
or deposition corrosion as solder and mechanical joints; b) fatigue
resistance: a properly designed adhesively bonded joint eliminates
or reduces the stress concentration at fasteners and increases the
fatigue life of a given joint; c) preservation of the structural integ-
rity of structural components such as carbon fibers reinforced
composites (CFRCs), and d) reduction of additional holes in the
resins impregnating carbon fabric/epoxy laminates and rivet
weight. The development of electrical conductive adhesives with
good structural bonding capability is required in order to address
the mentioned challenges and also to impart functional properties.
The development of conductive epoxy adhesive to be used in the
aeronautic field to join parts of primary structure is a current need
with a view to optimizing efficiency of joints while preserving the
conductivity of lightweight materials able to provide also in the
joints good lightening protection. The enhancement in different
properties of epoxy-based materials and or/adhesives depend on
numerous parameters, such as the chemical nature of nanofiller,
adhesive and adherents, the applied surface treatment, the phase
composition of the adhesive formulation or the tested properties
[7e13]. Many researchers focused their research toward the
development of new multifunctional adhesives for structural ap-
plications, based on appropriate epoxy resin nano-modified with
carbon nanostructured forms, able to hinder the insulator proper-
ties of resin if employed beyond their Electrical Percolation
Threshold (EPT) [14e17]. Carbon Nanotubes (CNTs) are known as
materials with very remarkable electronic, thermal, optical, me-
chanical, spectroscopic and chemical properties, which have been
attributed to the bonding structure of the CNTs [18,19]. The epoxy
resins are reactive monomers, which are commonly cured with
amine to form thermosetting polymers. If the epoxy is cured with
an aromatic amine of sufficient functionality, the result is a highly
cross-linked network with relatively high stiffness, glass transition
temperature (Tg) and chemical resistance. Unfortunately, epoxy
resins are brittle materials that have fracture energies of about two
orders of magnitude lower than modern thermoplastics and other
employed materials [20]. In order to overcome this criticality, the
epoxy adhesives are usually modified for improving their fracture
toughness. Currently, studies dealing with the tailoring of the resin
formulation to enhance the adhesive performance are in the fore-
front of scientific research. The target of these studies is focused on
maximizing the dissipated energy through either the plastic
deformation of the matrix, e.g. the inclusion of elastomers which
are able to increase the resin toughness [21], or through a suitable
control of the crack initiation and propagation process e.g. ceramic
modified polymers that inhibit interlaminar crack propagation. The
most common strategy for toughening epoxy adhesives consist in
the use of liquid rubbers or preformed rubber particles mixed or
dispersed in the epoxy formulation. In the case of liquid rubbers,
the rubber is initially dissolved into the epoxy resin [22]; one of the
major drawback of this preparation method is an undesirable phase
separation during the curing cycles of the adhesive formulation.
Extensive studies have shown that the phase separation process is a
result of the decrease in configurational entropy due to the increase
in molecular weight as the epoxy cures [23e25]. This changes the
free energy of mixing leading to a decrease in the solubility of the
rubber that provides the driving force for phase separation. Thus,
the functionality of the matrix monomers, which control the
development of the network and the cross-link density of the
epoxy matrix, has an effect on the phase precipitation process. The
particle size and the concentration of the precipitated rubber also
depend on the curing process and the interaction between the
rubber and the epoxy precursors. With butadiene-based rubbers,
the solubility may be increased by forming a copolymer with the
more polar acrylonitrile monomer. Another way to modify the
liquid rubber and hence control its interaction with the matrix is by
functionalising the chain ends with carboxyl, amine or epoxide
groups that may couple with the reacting matrix [24,26]. The focus
of this study is to propose a new electrical conductive adhesive
with the aim to expand the use of adhesive bonding of structural
parts to many engineering applications. This goal has been ach-
ieved modifying a recently developed [16,27,28] system by using a
liquid rubber in order to obtain high performance in the lap shear
strength (LSS) with adherents in carbon fiber reinforced compos-
ites. The correlation between the performance in the LSS values and
the structure of the polymer matrix has been studied by means of
mechanical and microscopy analysis.
2. Experimental
2.1. Materials
The composite mixture consists of tetraglycidylmethylenedia-
niline (TGMDA) and the epoxy reactive monomer 1e4 butane-
dioldiglycidyl ether (BDE) that acts as a reactive diluent. The 4,4-
diaminodiphenyl sulfone (DDS) has been employed as curing agent.
These components were supplied by Sigma Aldrich. The elastomer
employed is a carboxyl-terminated butadiene acrylonitrile copol-
ymer (CTBAC) from Hycar-Reactive Liquid Polymers, with
Mn ¼ 3600 containing terminal carboxy groups (COOH content of
0.67  103 equiv/g of CTBAC and 18 w/w% of CN). The triphenyl-
phosphine is employed as catalyst for the reaction between the
epoxy and the carboxyl groups. Multi Wall Carbon Nanotubes
(MWCNTs) (3100 Grade purchased from Nanocyl S.A) are
embedded in the formulation to impart electrical conductivity. In
order to obtain miscibility between epoxy precursor and CTBAC,
two precursor-elastomer (TGMDA-CTBAC) masterbatch are pre-
pared. The first is prepared by addition of 12.5 phr of CTBAC and 10
phr of triphenylphophine in the precursor resin. This mixture is
heated at 170  C, using an oil bath and mechanical agitation pro-
moted by an electric stirrer, for 24 h. The second system is prepared
by addition of 25 phr of CTBAC and 40 phr of triphenylphophine in
the precursor resin. This mixture is heated at 200  C, using an oil
bath and mechanical agitation promoted by an electric stirrer, for
14 h. The reaction related to the formation of the TGMDA-CTBAC
pre-polymer has been studied by FTIR analysis and the results are
described in the section “IR characterization” of the Supplementary
Information. The epoxy mixture is obtained by mixing the TGMDA-
CTBAC system with BDE monomer at a concentration TGMDA/BDE
of 80%:20% by wt. An amount of curing agent of 55 phr with respect
to precursor (TGMDA) is employed to activate the curing reactions.
Epoxy blend and DDS are mixed at 120  C and the MWCNTs are
added and incorporated in the matrix by using an ultrasonication
for 20 min (Hielscher model UP200S-24 kHz high power ultrasonic
probe) in order to obtain a homogeneous dispersion. All the mix-
tures are cured by a two-stage curing cycles: a first isothermal stage
at the lower temperature of 125  C for 1 h followed by a second
isothermal stage at the higher temperatures of 200  C for 3 h.
Table 1 summarises the chemical composition of the developed
adhesive systems, where the acronym TC refers to the formulation
composed of TGMDA (T) and CTBAC (C), whereas EPC refers to the
formulation composed of TGMDA, BDE, CTBAC and DDS. The
subscript x in TCx and EPCx refers to the phr of CTBAC. The adhesive
compositions and the preparation procedures are described in
details in the Supplementary Information.
 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
Table 1
Composition in parts per hundred parts of precursor (phr) for each component in the
final materials where the acronym TC refers to the formulation composed of TGMDA
(T) and the elastomeric phase CTBAC (C), whereas EPC refers to the formulation
composed of TGMDA, BDE, CTBAC and the hardener DDS. The subscript x in TCx and
EPCx refers to the phr of CTBAC.
Sample acronym TGMDA-CTBAC TGMDA-CTBAC-BDE-DDS (phr)
(phr)
TC0 100/0
TC12.5 100/12.5
TC25 100/25
EPC0 100/0/25/55
EPC12.5 100/12.5/25/55
EPC25 100/25/25/55
2.2. Methods
Adhesive tests were carried out using Dual Column testing
system (Instron mod. 5967). The specimens were placed in the
testing machine and the loading by setting a crosshead speed of
12.7 mm min1 (0.5 inch min1) as standard requirements. Five
specimens were tested for each adhesive composition. The adher-
ents are obtained from laminate of fibre reinforced composites
supplied by Leonardo-Finmeccanica. The adhesive strength is
determined by single lap shear test according to ASTM D5868.
Adhesion method test and Preparation of Test Joints are described
in Supplementary Information. Tensile properties are evaluated
with the crosshead speed of 1 mm/min. The obtained values were
taken from an average of five specimens each with a dimension
80  10  3 mm3. Dynamic Mechanical Analysis (DMA) of the
samples were performed with a dynamic mechanical thermo-
analyzer (Tritec 2000 DMA -Triton Technology). Solid samples
with dimensions 2  10  35 mm3 were tested by applying a var-
iable flexural deformation in three points bending mode. The
displacement amplitude was set to 0.03 mm, whereas the mea-
surements were performed at the frequency of 1 Hz. The range of
temperature was from 90  C to 315  C at the scanning rate of 3  C/
min. Micrographs of the formulated nanocomposites were ob-
tained using SEM (mod. LEO 1525, Carl Zeiss SMT AG, Oberkochen,
Germany). All samples were placed on a carbon tab previously
stuck to an aluminium stub (Agar Scientific, Stansted, UK) and were
covered with a 250 Å-thick gold film using a sputter coater (Agar
mod. 108 A). Nanofilled sample sections were cut from solid sam-
ples by a sledge microtome and then the section surfaces were
suitably treated before SEM investigation, as already described in a
previous paper [27]. The electrical characterization of the nano-
composites, focused on the DC volume conductivity measurements,
was performed on disk shaped specimens of 2 mm of thickness and
50 mm of diameter according to methods described in reference
[29].
3. Results and discussion
3.1. Mechanical properties (CTBAC effect)
The adhesive strength, determined by single lap shear test (LSS),
has been evaluated for adhesive compositions characterized by
different concentrations of elastomer CTBAC: a) 0 phr of CTBAC
(sample EPC0); b) 12.5 phr of CTBAC (sample EPC12.5) and c) 25.0
phr of CTBAC (sample EPC25.0).
Fig. 1 shows an enhancement in the LSS values with increasing
CTBAC amount. In particular, enhancements of 27% and 69% are
detected in stress of the samples EPC12.5 and EPC25 respectively
with respect to the sample EPC0. The better mechanical perfor-
mance related to this last parameter is generally indicative of an
3
Fig. 1. Lap shear strength of epoxy resin containing 0.0 (EPC0), 12.5 (EPC12.5) and 25.0
(EPC25) phr of CTBAC.
increase in the toughness of the formulation as discussed hereafter
in this section. A higher toughness in different rubber-modified
epoxy formulations was also found by other authors and was
ascribed a two-phase microstructure consisting of relatively small
rubbery particles well dispersed in the epoxy matrix [30]. The in-
fluence of the CTBAC concentration on the mechanical properties of
the resulting materials has been also investigated analysing their
tensile behaviour. Fig. 2a shows the stressestrain curves of the
three analysed samples EPC0, EPC12.5 and EPC25.0. It is evident that
the inclusion in the resin of CTBAC phase is able to improve the
toughness of the adhesive formulations. In fact, CTBAC-modified
systems show larger deformations and higher strains at-break
with respect to the neat epoxy resin EPC0 which is characterized
by a more brittle fracture and a different range of non-linear
viscoelastic behaviour. In Fig. 2b, c and 2d, tensile strength, Young
modulus and elongation-at-break of the formulated materials are
plotted as function of the CTBAC concentration. Fig. 2b highlights
that the higher value in the tensile strength at-break is obtained for
the sample EPC12.5 for which an increase of about 17.1% with respect
to the pure epoxy (i.e. 38.1 MPa and 44.6 MPa, respectively) is
detected. A such increase could be due to interactions between
rubber particles (CTBAC) and epoxy matrix. A similar result was
also observed by the authors of reference [31] for epoxy systems
modified with functionalized polybutadiene. Data shown in Fig. 2b
also highlight that the highest concentration of CTBAC (i.e. 25 phr),
determines a reduction in the value of tensile strength This
behaviour is most likely due to a lowering in cross-linking density
of the epoxy network determined by a significant amount of rubber
phase domains. Of course, the effect of rubber amount on the
tensile behaviour of the resin is not a trivial issue to understand
because different structural, chemical and morphological parame-
ters are involved: chemical interactions at the interface between
different phases, forms, dimensions, possible interpenetrations
between domains, etc.,. Huang and Kinloch, for similar systems,
explained a such decrease in the tensile strength with the presence
of micro voids [32], whereas Shukla and Srivastava with the stress
concentration effect of the rubber particles [31]. However, a such
decrease is surely expected for samples containing high concen-
trations of interface domains which are responsible to lower the
cross-linking density. Concerning the Young modulus of the ana-
lysed samples, Fig. 2c highlights a continuous decrease with
increasing the CTBAC concentration. In fact, ECP25 system exhibits
4 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
Fig. 2. Mechanical properties of EPC0, EPC12.5 and EPC25 samples: a) tensile stress-strain curves; b), tensile strength at break; c) Young modulus (E) and d) elongation at break.
the lowest modulus with a decrease of about 25% compared to the
sample EPC0. This behaviour can be due to lower modulus of
elastomer phase compared to the epoxy phase. This behaviour has
been also found for other systems containing a modified elasto-
meric phase [33]. The elongation-at-break of the modified epoxies
continuously increases with the increasing concentration of the
rubber phase in the samples (see Fig. 2d) going from 4.9% for the
EPC0 sample to 8.4% for EPC25 sample. This behaviour is very
promising because it is indicative of a good adhesion between the
rubber and epoxy phase and hence of the effectiveness of the re-
action carried out to compatibilize the rubber phase (see
Supplementary Information). The storage modulus, E’ (MPa), and
the loss factor, tand, of the analysed formulation are shown in
Fig. 3aeb, where in particular, tand, expressed as the ratio of loss
modulus to storage modulus, is a measure of the energy lost,
expressed in terms of the recoverable energy, and it represents
mechanical damping in a viscoelastic system. A comparison
Fig. 3. DMA curves of EPC0, EPC12.5 and EPC25 samples: a) loss factor Tand and b)
storage modulus.
between the different systems in terms of storage modulus evi-
dences that all samples show high values of the storage modulus,
higher than 2000 MPa up to 100  C, highlighting that the adhesive
formulation can be advantageously applied in the range of service
temperature of structural materials (i.e. aeronautical, shipbuilding
industries, wind turbine blades, as electronic device, etc.). The
curve related to tand of the sample EPC0 (red curve) exhibits two
distinct peaks centred at about 200  C and 265  C respectively. The
presence of two peak in the dynamic-mechanical spectra of unfilled
epoxy resins was already found in literature [28]; it was correlated
to fractions with different crosslinking density characterized by
two different values of Tg. The amount of these different phases,
related to the intensity of the peaks in the tand curves depend on
the curing degree and the stoichiometric ratio of oxirane rings/
hardener. The introduction of CTBAC in the epoxy matrix de-
termines significant changes in the profile of tand curves. In
particular, the glass transition temperature appears as a single
phase shifted at lower temperature with respect to both the Tg
values of 200  C and 265  C. Furthermore, the peak is widened and
more intense than each peak observable for the sample EPC0. This is
a clear indication that the two phases with different crosslinking
density converge in a single phase, which presents relaxation
mechanisms in a wider range of temperature. The rubber phase is
well distributed in the matrix and affect the dynamic mechanical
properties in the same way in each region contributing to lower the
crosslinking density everywhere. The plasticiser promotes the
onset of the a transition and a higher volume of non-agglomerated
plasticiser causes a lower Tg and an increased adhesive perfor-
mance (see Fig. 1). A lower glass transition temperature results in
lower modulus values (see Fig. 3b), in particular for the mixture
with 25 phr of rubber. These results confirm that the introduction
of CTBAC has a plasticizer action of the matrix. A such plasticizer
action has already found for other thermosetting formulations
containing a rubber phase. It has been hypostasized that the plas-
ticizer action enhances the adhesive properties for different rea-
sons: a) the crack front pinning results when a crack front meets a
number of toughened particles propagating around them. This
creates a deformation in the crack which in turn requires an higher
level of energy in order to maintain its progress, improving the
 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
adhesion of the material [34]; b) this mechanism can also cause
particles de-bonding in severe cases that helps to blunt the crack
tip thus increasing the fracture toughness of the material; c) the
propagation of crack through the rubber particles causes small
cavitations that are a major mechanism of the increase in the
fracture energy of the adhesives [35]; d) the stress is further
dissipated through the growth of such voids until a coalescence
occurs [36]. The crack bridging can occur when the particles are
stretched from their initial sites and the progress of the crack is
inhibited. In this work, it is found that several mechanical proper-
ties can be optimized by suitably choosing the CTBAC amount and
its functionalization whose choice is strictly related to the nature of
the epoxy matrix. It is worth noting that even if the values of the
elastic modulus, the tensile strength and the glass transition tem-
perature decrease, the formulation based on the tetrafunctional
precursor (TGMDA) is characterized by high values in these pa-
rameters and the values achieved after the decrease are within the
range required for many industrial application with the advantage
to obtain a strong improvement in the value of the lap shear
strength. The adopted strategy has allowed to reduce the crosslink
density, to increase the chain flexibility and hence enhance the
ability of the adhesive to inhibit the crack propagation. The best
result in terms of adhesive performance has been found for the
sample EPC25. This last sample has been chosen as the basis
formulation to impart electrical conductivity to the adhesive.
Electrical behaviour, dynamic-mechanical properties and
morphological features are discussed in the next sections.
3.2. Electrical properties
The electrical conductivity is the most intensively investigated
property of composites once the percolation threshold (i.e. EPT) has
been achieved [37,38]. The conduction in composites reinforced
with carbon nanotubes has been explained by considering that,
when the CNTs concentration (i.e x) overcomes a threshold value
(i.e. xc,), conductive paths are formed in the composite which
converts its electrical behaviour from an insulator to a conductor.
The percolation theory describes the dependence of the conduc-
tivity s on the filler concentration by a scaling law of the form
s ¼ s0(x-xc)t where s0 is the intrinsic electrical conductivity of the
filler and t is an exponent depending on the system dimensionality
[39e43]. In particular, correlations between material characteris-
tics (polymeric matrices, CNT type, synthesis method, treatment,
Fig. 4. Electrical conductivity of EPC25 sample versus MWCNT weight percentage;
Inset: Electrical conductivity of EPCx at loading rate of 0.5% by weight of MWCNTs vs.
amount of CTBAC (0, 12.5 and 25.0 phr of elastomeric phase).
5
etc.) and parameters describing the percolation law are discussed in
reference [42]. Fig. 4 shows the DC volume conductivity at room
temperature of the composite EPC25 as function of the filler
amount. The EPT is reached for a value of MWCNT percentage lower
than 0.5 wt%. In fact, for this MWCNT concentration, the sample is
characterized by a value in the electrical conductivity of 6.5 $ 106
which is several orders of magnitude higher than the few pS/m of
the unfilled sample. A value of about 2.6 $ 101 S/m is achieved at
the highest filler loading (i.e. 1.5 wt%). However, the conductivity
does not reach the values obtained for dense films of MWCNTs
(z103 S/m) [44] due to the effect of nanotube/nanotube contact
resistance and to the role of tunnelling resistance, affected by the
thickness of resin wrapping the conductive filler, in determining
the overall electrical properties of the resulting composites [45].
The inset of Fig. 4 shows the influence of the amount of CTBAC on
the electrical conductivity for samples filled with 0.5 wt% of
MWCNTs. A reduction of 5 orders of magnitude is obtained for the
composite with 25 phr of CTBAC compared to EPC0 sample. The
morphological analysis (see section 3.4) provides crucial evidence
for the interpretation of this behaviour. In fact, no MWCNTs have
been found in the rubber phase, hence the rubber domains inter-
rupt percolating paths within the sample. As a result, there is a
change of the electrical resistance associated with the tunneling
between conducting clusters forming the percolating paths inside
the resin, as will be evident in the section on the morphological
analysis. In any case, an electrical conductivity comparable with
that obtained in absence of elastomer (i.e. EPC0 sample), can be
obtained only by increasing the filler concentration of one per-
centage point (i.e. at 1.5 wt%). It is worth noting that, together with
an enhancement in the electrical conductivity, the inclusion of
MWCNTs in the adhesive formulations might be of benefit also to
impart them other functional properties of relevant technological
interest [46, 47].
3.3. Mechanical properties (carbon nanotubes effect)
Data shown in Fig. 5 allow understanding the effect of the filler
concentration on the adhesive and tensile properties of EPC25
sample with increasing MWCNT percentage. In particular, Fig. 5a
shows the LSS of the nanofilled formulations; whereas Fig. 5bed
refer to parameters related to the tensile tests: Young modulus,
Elongation and Tensile strength at break. Fig. 5a shows an
improvement in the adhesive properties with increasing the
amount of MWCNTs. The maximum value of 14.76 Mpa is found for
the sample loaded with 1.0 wt% of MWCNT. LSS value increases of
13% for the sample containing 1% by weight of MWCNT, with
respect to the value of the sample without nanofiller. It is worth
noting that the values in LSS parameter are higher than those ob-
tained in previous papers related to conductive adhesives used to
assemble reinforced laminates adherents which have been tested
according to ASTM 5868 standard [48e51]. It is worth noting that,
while many reports in literature deal with the electrical conduc-
tivity and mechanical properties of bulk CNT-polymer composites,
only few reports on the performance tests based on end-use ap-
plications, such as adhesive joints, are available [49e52]. Such a
study is relevant because the application performance can differ
from that expected on the base of the bulk properties of the sample.
In fact, it has already found that the joint-strength is not necessarily
well predicted by bulk mechanical properties [49,50]. However, the
first studies reported in literature on the influence of CNTs on the
adhesive properties have already shown improvements in lap-
shear strength of z10e40% [50e52]. Improved fracture tough-
ness in wedge tests for aluminumealuminum joints with
SWCNTeadhesive composites has also been demonstrated [49].
Furthermore, concerning the joint electrical performance, which is
6 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10

Fig. 5. Mechanical properties of EPC25 sample versus MWCNT weight percentage: a) LSS, b)Young modulus (E), c) Elongation at break and d) Tensile strength at break.

another relevant technological parameter, few studies have dealt

with the evaluation of this parameter. In particular, electrical con-
ductivities of 0.1 S m1 (with 1.0 CNTs wt%), 1.0 S m1(with 2.0 CNTs
wt%), and 100 S m1 (with 12.0 CNTs wt%) have been reported with
relatively large amounts of MWCNTs in epoxy resins, but with
decreased lap shear strength compared to the unfilled epoxy [52].
In Fig. 5b,c,d, the changes in the mechanical properties of the for-
mulations here described are shown. In particular, a decrease in
mechanical properties (Young modulus, tensile strength at break,
elongation at break) is observed with increasing MWCNT per-
centage. A similar behavior was obtained from Meguid et al. [52]
but only for high concentrations of nanoparticles. Generally, the
dispersion of small percentages of CNTs has proven to improve
several mechanical properties of the nanofilled formulations
[38,53]. The effect of MWCNTs on the mechanical parameters here
shown seems to be opposed to the behavior described in previous
paper [28,54,55] where the inclusion of MWCNTs in epoxy for-
mulations allows to obtain higher storage modulus compared to
the unfilled resin. It is therefore possible, and indeed likely, that this Fig. 6. DMA curves of the EPC25 sample unfilled and at loading rate of 0.5%, 1.0% and
different trend is strictly related to the presence of the elastomeric 1.5% by weight of MWCNTs: a) loss factor Tand and b) storage modulus.
phase which interrupts the nanotube network, determining not
only a reduction in the electrical conductivity (see inset in Fig. 4),
but also a slight decrease in the mechanical parameters. In fact, the
and 90  C. For higher values of MWCNT percentage, the decrease is
segregation of MWCNTs from the elastomeric phase affects in
even more pronounced. Furthermore, the presence of the elasto-
opposite way mechanical properties and adhesive properties. An
meric phase causes a decrease of the storage modulus in the tem-
explanation of such a phenomenon is clear in dynamic mechanical
perature range from 90  C to 0  C (see inset in Fig. 7), where a
curves of the analyzed formulations. Fig. 6 shows the behavior of
clear transition is observed in the tan d curve. The inclusion of
tan d (Fig. 6a) and of the storage modulus (Fig. 6b) vs. the tem-
MWCNTs affects the intensity and the maximum value of this last
perature of the formulation EPC25 at different MWCNT percentage.
transition, which shifts to a lower temperature. The presence of this
MWCNTs embedded in the matrix do not cause large differences in
phase has already been discussed in previous papers. Barcia et al.
the glass transition temperature, which mainly remains at the value
[56] assert that the peak at low temperature appears in the same
of 140  C, except for the system containing 1.5% by weight of
glass transition region relative to the polybutadiene phase. They
MWCNTs characterized by the lower value of 132  C. The storage
assume that, in all the modified networks there is a phase sepa-
modulus, as the Young modulus, decreases with increasing
ration due to the presence of rubber particles in the modified epoxy
MWCNT percentage. Specifically, for the system with 1.0 wt% of
resin which are not fully dissolved into the epoxy matrix. Wise et al.
MWCNTs which is characterized by the highest value in LSS (very
[57]. found that the magnitude of the peak at low temperature
similar to the sample loaded with 1.5 wt% of MWCNTs), the
increases when the rubber level is raised. An explanation can be
decrease is around 0.9 GPa in the temperature range between 0  C
 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10 7

Fig. 8. SEM images of the fracture surface of system with the elastomer at 25 phr
Fig. 7. Magnification of loss factor Tan d and storage modulus (Inset) of EPC25 sample (EPC25).
unfilled and at loading rate of 0.5%, 1.0% and 1.5% by weight of MWCNTs in the tem-
perature range between 90  C and 50  C.

diameters of about 15 mm. The presence of segregate rubbers do-

mains can be responsible of the behaviour discussed in the previous
that the lowering of the CTBAC rubber Tg may be due to composi-
section: effect of plasticization (see presence of the phases
tional heterogeneity of acrylonitrile in the copolymer chains, which
between 90  C and 30  C), the decrease in some of the mechanical
causes compositional fractionation of the acrylonitrileebutadiene
properties with respect to the unmodified resin etc. Barcia el al [56],
copolymer during the phase precipitation process. This could cause
in a previous paper, observed a heterogeneous morphology resul-
the nonpolar, butadiene-rich component of the CTBAC copolymer
ted in opaque samples. The holes observed in the micrograph were
to be preferentially expelled from the more polar matrix thus
ascribed to the rubber particles that were withdrawn from the
leading to a reduction in the Tg of the dispersed phase because the
surface after the treatment with solvent, indicating no chemical
butadiene-rich CTBAC has a lower Tg. In our case, a more evident
bond between the blend components. In the case of the sample
presence of the low-temperature phase is determined by the
here described, the spherical rubber domains are not extracted
introduction of carbon nanotubes. The reason for this, is due to,
with the etching treatment; this is a clear evidence of the effec-
most likely, the occurrence that MWCNTs hinder the crosslinking
tiveness of the adopted strategy to enhance the interfacial adhesion
reaction between the chain segments composed by the epoxy
between the epoxy precursor and the elastomeric phase (see
precursor linked to the elastomer, creating a separate phase (see
Supplementary Information). Fig. 9 shows the system with 25 phr
transition between 0  C and 30  C which partially overlaps on the
of CTBAC/epoxy with a 1.0% by weight of MWCNTs. The image re-
left with the transition between 90  C and 0  C) formed by chain
veals two separate phases (domains): the elastomeric domains and
segments of precursor/elastomer. Finally, although the carbon
the epoxy matrix where MWCNTs are embedded. Probably this
nanotubes cause a reduction in the mechanical properties, reducing
particular feature, composed of two macro phases, a conductive
the cohesive forces of the adhesive, they increase the adhesion in
and the other insulating, determines the presence of Tg at low
the joint. The adhesion enhancement may be attributed to the
temperatures found in Fig. 7. In addition, this phase not only leads
CNTs/epoxy interface which can be responsible to activate mech-
to a decrease of the mechanical properties but also leads to a not
anisms at the nanoscale level such as crack bifurcation and arrest,
good distribution of the carbon nanotubes, interrupting the
delaying thus the global shear failure as already explained by other
authors that found similar results [58]. Interestingly enough, the
choice of sample EPC25 at a loading rate of 1.0 wt% of MWCNTs
allows to simultaneously optimize the performance in electrical,
mechanical and adhesive properties.

3.4. Morphological investigation

In order to analyze the morphological arrangement of the

elastomeric phase and nanofiller dispersion, morphological anal-
ysis by SEM investigation was carried out on etched samples.
Etching procedures have already been successful applied to remove
the resin surrounding MWCNTs, leaving them bare and hence
allowing a better investigation on their dispersion in the matrix.
Fig. 8 shows SEM images of the fracture surface of system with the
elastomer at 25 phr (EPC25). The fracture surfaces of the CTBAC-
modified epoxy matrix show a two-phase morphology with a
scattered rubber phase (CTBAC spherical domains) and a rigid
continuous phase (epoxy matrix). The size of rubber particles in
blend sample containing 25 phr of CTBAC, is in the range of few Fig. 9. SEM images of the fracture surface of system 25 phr CTBAC/epoxy with 1.0% by
micron, except in the case of few domains which can reach weight of the MWCNTs.
8 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
percolation network (see inset of Fig. 4) as discussed before. An
increase of the amount of the carbon nanotubes allows to increase
the electrical conductivity of the system (see Fig. 4), and if on the
one hand the presence of the rubber phase reduces some of the
mechanical properties (see Fig. 5), on the other hand it increases
the toughness of the system determining a better adhesion. The
optimized adhesive formulation EPC25 at loading rate of 1.0 wt% of
MWCNTs has been employed to manufacture joints composed of
aeronautic adherents (CFRCs). The preparation of Test Joints is
described in Supplementary Information. The fracture surfaces of
the tested adhesive joints have been examined using optical mi-
croscopy and scanning electron microscopy (SEM) (see Figs. 10 and
11). Fig. 10 shows the joint surface tested with EPC25 adhesive at a
loading rate of 1.0 wt% of MWCNTs. The adhesive unevenly covers
the surface of the joints. In fact, in some zones (highlighted with a
white ellipse in the optical images at lower magnification (on the
left) a failure in the laminate (delamination) between resin and
Carbon Fibers (CFs) (in the adherent) is observed. The effect is
clearly observed in SEM image of Fig. 11a, where the surface of an
almost bare CF is observed. This is a very interesting result because
it is indicative of an adhesion between the adhesive and the
adherent much more effective than the adhesion forces that bond
epoxy matrix and carbon fiber in the adherents. The macro-
morphology shown in Fig. 10 in the ellipses highlights that the
surface of the joints shows the channel of bare CFs being the sur-
rounding resin stripped together the adhesive part. Hence the
developed adhesive forms cohesive forces with the resin impreg-
nating the CFs stronger than the forces between CFs and resin of the
Fig. 10. Optical images of fracture surfaces of the tested adhesive joints.
adherent and as a consequence the delamination occurs before the
failure along the interface adhesive/adherent. The failure mode is
mixed adhesive/delamination in the joints bonded with nano-
reinforced epoxy adhesive. In fact, in other zone, the joint fails in
the adhesive between the adherents as it can also be observed in
the SEM image of Fig. 11 b, where the spherical domains of the
rubber phase are clearly observable. On the contrary, the joints
assembled using unfilled epoxy adhesive (see Fig. 10 e sample
EPC25) show a different behaviour. In fact, in this case the joint fails
almost completely in the area between the adhesive and the sur-
face of the adherent. As consequence of this failure mode, the
feature of the carbon fabric on the adherent surface is clearly visible
and no delamination phenomena occur.
4. Conclusions
In this study, a conductive adhesive for application on adherents
in carbon fiber reinforced plastics has been designed and charac-
terized. The analysis of the mechanical properties has shown that
the presence of elastomeric domains allows to obtain a matrix with
greater flexibility able to strong increase the adhesive performance.
MWCNTs embedded in the epoxy adhesive are able to further in-
crease the LSS of the joints, reinforcing the strong effect caused by
rubber domains dispersed in the formulation. The fracture of the
joints bonded with the unfilled formulation EPC25 occurs almost
completely along the interface between adhesive and adherent and
no delamination effects or damages are observed in the adherent
parts. In opposite to this behavior, the failure observed on the
 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
Fig. 11. SEM images of the joint surface tested with EPC25 adhesive at loading rate of
1% by weight of MWCNTs.
surfaces of joints assembled with the adhesive loaded with
MWCNTs, results in a detachment between the resin (of the
adherent) and the carbon fibers (belonging to the carbon fabric).
This result highlights that the cohesive forces determined between
adherent resin and adhesive resin are stronger than the adhesive
forces along the interface between adhesive/adherent. Concerning
the effect of MWCNTs on the adhesive performance, a very similar
result has been obtained embedded in the adhesive formulation
(without rubber phase) carbon nanofibers [15]. It is worth noting
that in the case of the formulation EPC25 the presence of the rubber
phase strong increases the adhesive performance and very low
concentrations of MWCNTs have been used to impart electrical
conductivity to the epoxy adhesive.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.compositesb.2017.07.021.
References
[1] Ascione F, Lamberti M, Razaqpur AG, Spadea S. Strength and stiffness of
adhesively bonded GFRP beam-column moment resisting connections. Com-
pos Struct 2017;160:1248e57.
[2] Ozdemir O, Oztoprak N. An investigation into the effects of fabric re-
inforcements in the bonding surface on failure response and transverse
impact behavior of adhesively bonded dissimilar joints. Compos Part B
2017;126:72e80.
[3] Akpinar IA, Gültekin K, Akpinar S, Akbulut H, Ozel A. Research on strength of
nanocomposite adhesively bonded composite joints. Compos Part B 2017;126:
143e52.
[4] Ribeiro FMF, Campilho RDSG, Carbas RJC, da Silva LFM. Strength and damage
growth in composite bonded joints with defects. Compos Part B 2016;100:
9
91e100.
[5] Ascione F. The influence of adhesion defects on the collapse of FRP adhesive
joints. Compos Part B.
[6] Bishopp J. Adhesives for aerospace structures. In: S E, editor. Handbook of
adhesives and surface preparation: William Andrew; 2011. p. 301e44.
[7] Jakubinek MB, Ashrafi B, Zhang Y, Martinez-Rubi Y, Kingston CT, Johnston A,
et al. Single-walled carbon nanotub-epoxy composites for structural and
conductive aerospace adhesives. Compos Part B 2015;69:87e93.
[8] Esposito Corcione C, Freuli F, Frigione M. Epoxy/expanded graphite stacks
nanocomposites for cold-cured adhesives. J Adhes Sci Technol 2017;31(7):
713e25.
[9] Esposito Corcione C, Frigione M. Cure kinetics and physical characterization of
epoxy/modified boehmite nanocomposites. J Adhes Sci Technol 2017;31(6):
645e62.
[10] Gude M, Prolongo S, Go
mez-del Río T, Urena A. Mode-I adhesive fracture
energy of carbon fibre composite joints with nanoreinforced epoxy adhesives.
Int J Adhes Adhes 2011;31(7):695e703.
[11] Lafdi K, Fox W, Matzek M, Yildiz E. Effect of carbon nanofiber-matrix adhesion
on polymeric nanocomposite properties: part II. J Nanomater 2008;2008:5.
[12] Prolongo S, Gude M, Sanchez J, Uren ~ a A. Nanoreinforced epoxy adhesives for
aerospace industry. J Adhes 2009;85(4e5):180e99.
[13] Guadagno L, Naddeo C, Raimondo M, Barra G, Vertuccio L, Sorrentino A, et al.
Development of self-healing multifunctional materials. Compos Part B
2017;128:30e8.
[14] Billias MG, Borders ME. Electrically conductive structural adhesive. Google Pat
31 Jan 1984. U.S. Patent No. 4,428,867.
[15] Gkikas G, Sioulas D, Lekatou A, Barkoula NM, Paipetis AS. Enhanced bonded
aircraft repair using nano-modified adhesives. Mater Des 2012;41:394e402.
[16] Vietri U, Guadagno L, Raimondo M, Vertuccio L, Lafdi K. Nanofilled epoxy
adhesive for structural aeronautic materials. Compos Part B 2014;61:73e83.
[17] Jakubinek MB, Ashrafi B, Zhang Y, Martinez-Rubi Y, Kingston CT, Johnston A,
et al. Single-walled carbon nanotube-epoxy composites for structural and
conductive aerospace adhesives. Compos Part B 2015;69:87e93.
[18] Reich S, Thomsen C, Maultzsch J. Carbon manotubes: basic concepts and
physical properties. John Wiley & Sons; 2008.
[19] Guadagno L, Vietri U, Raimondo M, Vertuccio L, Barra G, De Vivo B, et al.
Correlation between electrical conductivity and manufacturing processes of
nanofilled carbon fiber reinforced composites. Compos Part B 2015;80:7e14.
[20] Rezaifard AH, Hodd KA, Barton JM. Toughening epoxy resin with poly (methyl
methacrylate)-grafted natural rubber. Washington, DC (United States):
American Chemical Society; 1993.
[21] Pearson R, Yee A. Toughening mechanisms in elastomer-modified epoxies.
J Mater Sci 1989;24(7):2571e80.
[22] Sultan JN, McGarry FJ. Effect of rubber particle size on deformation mecha-
nisms in glassy epoxy. Poly Eng Sci 1973;13(1):29e34.
[23] Williams R, Borrajo J, Adabbo H, Rojas A. A model for phase separation during
a thermoset polymerization. Rubber-Modified Thermoset Resins 1983:
195e213.
[24] Verchere D, Sautereau H, Pascault J, Moschiar S, Riccardi C, Williams R. Rub-
ber-modified epoxies. I. influence of carboxy-terminated butadiene-acrylo-
nitrile random copolymers (CTBN) on the polymerization and phase
separation processes. J Appl Polym Sci 1990;41(3e4):467e85.
[25] Moschiar S, Riccardi C, Williams R, Verchere D, Sautereau H, Pascault J. Rub-
ber-modified epoxies. III. analysis of experimental trends through a phase
separation model. J Appl Polym Sci 1991;42(3):717e35.
[26] Chen D, Pascault JP, Sautereau H. Rubber-modified epoxies. I. influence of
presence of a low level of rubber on the polymerization. Polym Int
1993;32(4):361e7.
[27] Vertuccio L, Russo S, Raimondo M, Lafdi K, Guadagno L. Influence of carbon
nanofillers on the curing kinetics of epoxy-amine resin. RSC Adv 2015;5(110):
90437e50.
[28] Guadagno L, Naddeo C, Raimondo M, Barra G, Vertuccio L, Russo S, et al. In-
fluence of carbon nanoparticles/epoxy matrix interaction on mechanical,
electrical and transport properties of structural advanced materials. Nano-
technology 2017;28(9):094001.
[29] Guadagno L, Raimondo M, Vertuccio L, Mauro M, Guerra G, Lafdi K, et al.
Optimization of graphene-based materials outperforming host epoxy
matrices. RSC Adv 2015;5(46):36969e78.
[30] VrD Ramos, da Costa HM, Soares VL, Nascimento RS. Hybrid composites of
epoxy resin modified with carboxyl terminated butadiene acrilonitrile
copolymer and fly ash microspheres. Polym Test 2005;24(2):219e26.
[31] Shukla SK, Srivastava D. Blends of modified epoxy resin and carboxyl-
terminated polybutadiene. I J Appl Polym Sci 2006;100(3):1802e8.
[32] Huang Y, Kinloch A. The toughness of epoxy polymers containing microvoids.
Polymer 1992;33(6):1330e2.
[33] Saadati P, Baharvand H, Rahimi A, Morshedian J. Effect of modified liquid
rubber on increasing toughness of epoxy resins. Iran Polym J 2005;14(7):
637e46.
[34] Baldan A. Adhesively-bonded joints and repairs in metallic alloys, polymers
and composite materials: adhesives, adhesion theories and surface pretreat-
ment. J Mater Sci 2004;39(1):1e49.
[35] Dean G, Crocker L, Read B, Wright L. Prediction of deformation and failure of
rubber-toughened adhesive joints. Int J Adhes Adhes 2004;24(4):295e306.
[36] Chikhi N, Fellahi S, Bakar M. Modification of epoxy resin using reactive liquid
(ATBN) rubber. Eur Polym J 2002;38(2):251e64.
10 L. Vertuccio et al. / Composites Part B 129 (2017) 1e10
[37] Kim YJ, Shin TS, Do Choi H, Kwon JH, Chung Y-C, Yoon HG. Electrical con-
ductivity of chemically modified multiwalled carbon nanotube/epoxy com-
posites. Carbon 2005;43(1):23e30.
[38] Song YS, Youn JR. Influence of dispersion states of carbon nanotubes on
physical properties of epoxy nanocomposites. Carbon 2005;43(7):1378e85.
[39] De Vivo B, Guadagno L, Lamberti P, Raimondo M, Spinelli G, Tucci V, et al.
Electrical properties of multi-walled carbon nanotube/tetrafunctional epoxy-
amine composites. In: D'Amore Alberto, Grassia Luigi, Acierno Domenico,
editors. AIP conference proceedings, vol 1459. AIP; 2012. No. 1.
[40] De Vivo B, Lamberti P, Spinelli G, Tucci V. A morphological and structural
approach to evaluate the electromagnetic performances of composites based
on random networks of carbon nanotubes. J Appl Phys 2014;115(15):154311.
[41] Guadagno L, Naddeo C, Vittoria V, Sorrentino A, Vertuccio L, Raimondo M,
et al. Cure behavior and physical properties of epoxy resin-filled with mul-
tiwalled carbon nanotubes. J Nanosci Nanotechno 2010;10(4):2686e93.
[42] Bauhofer W, Kovacs JZ. A review and analysis of electrical percolation in
carbon nanotube polymer composites. Compos Sci Technol 2009;69(10):
1486e98.
[43] Kymakis E, Amaratunga GA. Electrical properties of single-wall carbon
nanotube-polymer composite films. J Appl Phys 2006;99(8):084302.
[44] Guadagno L, De Vivo B, Di Bartolomeo A, Lamberti P, Sorrentino A, Tucci V,
et al. Effect of functionalization on the thermo-mechanical and electrical
behavior of multi-wall carbon nanotube/epoxy composites. Carbon
2011;49(6):1919e30.
[45] Li C, Thostenson ET, Chou T-W. Dominant role of tunneling resistance in the
electrical conductivity of carbon nanotube-based composites. Appl Phys Lett
2007;91(22):223114.
[46] Vertuccio L, Guadagno L, Spinelli G, Lamberti P, Tucci V, Russo S. Piezoresistive
properties of resin reinforced with carbon nanotubes for health-monitoring of
aircraft primary structures. Compos Part B 2016;107:192e202.
[47] Vertuccio L, Vittoria V, Guadagno L, De Santis F. Strain and damage monitoring
in carbon-nanotube-based composite under cyclic strain. Compos Part A
2015;7:9e16.
[48] Rosca ID, Hoa SV. Method for reducing contact resistivity of carbon nanotube-
containing epoxy adhesives for aerospace applications. Compos Sci Technol
2011;71(2):95e100.
[49] Ashrafi B, Johnston A, Jakubinek M, Martinez-Rubi Y, Simard B, Hubert P. Ef-
fect of single-walled carbon nanotubes on processing and property
enhancement of epoxy adhesives and laminates. In: International SAMPE
technical conference; 2012.
[50] Zhang J, Luo R, Yang C. A multi-wall carbon nanotube-reinforced high-tem-
perature resistant adhesive for bonding carbon/carbon composites. Carbon
2012;50(13):4922e5.
[51] Liu Z, Bae DH. A study on joining technology of aluminum alloy sheet using
nano-adhesives. Int J Mod Phys B 2011;25(31):4265e8.
[52] Meguid S, Sun Y. On the tensile and shear strength of nano-reinforced com-
posite interfaces. Mater Des 2004;25(4):289e96.
[53] Schadler L, Giannaris S, Ajayan P. Load transfer in carbon nanotube epoxy
composites. Appl Phys Lett 1998;73(26):3842e4.
[54] Guadagno L, Vertuccio L, Sorrentino A, Raimondo M, Naddeo C, Vittoria V,
et al. Mechanical and barrier properties of epoxy resin filled with multi-
walled carbon nanotubes. Carbon 2009;47(10):2419e30.
[55] Guadagno L, Raimondo M, Vittoria V, Vertuccio L, Naddeo C, Russo S, et al.
Development of epoxy mixtures for application in aeronautics and aerospace.
RSC Adv 2014;4(30):15474e88.
[56] Barcia FL, Amaral TP, Soares BG. Synthesis and properties of epoxy resin
modified with epoxy-terminated liquid polybutadiene. Polymer 2003;44(19):
5811e9.
[57] Wise C, Cook W, Goodwin A. CTBN rubber phase precipitation in model epoxy
resins. Polymer 2000;41(12):4625e33.
[58] Gkikas G, Sioulas D, Lekatou A, Barkoula N, Paipetis A. Enhanced bonded
aircraft repair using nano-modified adhesives. Mater Des 2012;41:394e402.