Content uploaded by Dimitris Baltzis
Author content
All content in this area was uploaded by Dimitris Baltzis on Oct 05, 2017
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=yprc20
Download by: [89.210.164.63] Date: 08 April 2016, At: 23:20
Plastics, Rubber and Composites
Macromolecular Engineering
ISSN: 1465-8011 (Print) 1743-2898 (Online) Journal homepage: http://www.tandfonline.com/loi/yprc20
Stainless steel coupled with carbon nanotube-
modified epoxy and carbon fibre composites:
Electrochemical and mechanical study
D. Baltzis, S. Orfanidis, A. Lekatou & A. S. Paipetis
To cite this article: D. Baltzis, S. Orfanidis, A. Lekatou & A. S. Paipetis (2016) Stainless steel
coupled with carbon nanotube-modified epoxy and carbon fibre composites: Electrochemical
and mechanical study, Plastics, Rubber and Composites, 45:3, 95-105
To link to this article: http://dx.doi.org/10.1080/14658011.2016.1144339
Published online: 08 Apr 2016.
Submit your article to this journal
View related articles
View Crossmark data
Stainless steel coupled with carbon nanotube-
modified epoxy and carbon fibre composites:
Electrochemical and mechanical study
†
D. Baltzis, S. Orfanidis, A. Lekatou and A. S. Paipetis∗
The aim of this work is the study of the electrochemical and mechanical behaviour of stainless steel
(SS304) adhesively bonded with carbon nanotube (CNT)-reinforced epoxies to either SS304 or
carbon-reinforced composites substrates. For metal to metal (MtM) joints, the shear strength of
nano-reinforced adhesives was studied using single lap shear specimen geometries. The lap
shear strength was improved by almost 50% and the highest shear strength appeared for 0.6%
CNT weight content in the adhesive. The metal to composite joint performed altogether better
compared to the MtM joint, although the CNT inclusion had an adverse effect on the lap shear
strength attributed to the physical property change of the epoxy. Although the incorporation of
CNTs was found to increase the galvanic effect, it also enhanced corrosion protection, as the
modified adhesives exhibited increased resistance to uniform corrosion and localised corrosion
and prevented the electrolyte from reaching the substrate.
Keywords: Stainless steel, Epoxy, Carbon nanotubes, Electrochemical behaviour, Lap shear strength
Introduction
Stainless steel alloys are among the most commonly used
steel alloys for naval and marine applications. A member
of the steel alloys is stainless steel 304 (SS304) and it is
used in applications such as offshore and underwater
pipelines and storage tanks for liquid natural gas
(LNG) in tanker ships due to its excellent corrosion resist-
ance properties. In such applications, the most frequent
types of damage are stress corrosion and erosion because
of the hostile environment (seawater) and operational
conditions like freeze–thaw cycles. Hence, repair oper-
ations are frequently required in such structures. Tra-
ditional steel repairs, like riveted or welded steel
patches, are heavyweight, time-consuming and incorpor-
ate tedious welding works which restrain their use in pipe-
lines located both underground and underwater. In
addition welding repairs may induce additional stresses
on the weld area further enhancing stress corrosion.
1
Moreover, for applications where flammable gases or
liquids (LNG or gasoline) are present, welding is not an
option due to industrial safety reasons; in order to ensure
safety during the repair, the flammable materials must be
absent (e.g. emptying the storage tank or the pipeline)
increasing thus the cost of the repair.
2
Furthermore, due
to the stress concentration induced by the thermal
gradient during the welding, the weld may potentially
act as a cathode forming a galvanic corrosion cell.
Fibre-reinforced composites (FRCs) may provide an
effective repair alternative for such applications, when
employed as adhesively bonded patches. Composite
patch repair is advantageous and potentially more cost-
effective as it is safer compared to welding due to the
absence of flame and potentially faster particularly if
the cost of operational halt time is taken into account.
3
The performance of the bond between fibre-reinforced
polymers (FRPs) and steel surface is primarily governed
by the adhesive strength of the joint. In the cases of
bonded joints, stresses along the entire area of the compo-
site/steel interface are uniformly transferred and stress
concentrations are minimised.
4
However, concerns may
arise when FRCs are selected as the choice for repair.
Differences in the mechanical and physical properties of
steel and FRCs create difficulties in ensuring a strong
bond and reducing the possibility of delamination, par-
ticularly in the presence of temperature fluctuations.
5–7
Another detrimental factor affecting the joint repair is
the steel surface morphology. Primer coating, grit blasting
and chemical etching are some of the studied surface
modification techniques proposed in the literature.
8
In
addition to surface morphology, the thickness of the
adhesive layer is well known to affect the bond repair,
as too thick or too thin adhesive layers may deteriorate
the bond performance.
9,10
Another parameter that needs attention when steel is
adhesively bonded with FRPs or other metallic alloys is
the performance of the structure in terms of galvanic cor-
rosion. Adhesive bonding can induce galvanic corrosion
Materials Science and Engineering Department, University of Ioannina,
Ioannina 45110, Greece
∗Corresponding author, email paipetis@cc.uoi.gr
†
Fourth International Conference of Engineering Against Failure (ICEAF
IV), 24–26 June 2015, Skiathos, Greece.
© 2016 Institute of Materials, Minerals and Mining
Published by Taylor & Francis on behalf of the Institute
Received 11 November 2015; accepted 16 January 2016
DOI 10.1080/14658011.2016.1144339 Plastics, Rubber and Composites 2016 VOL 45 NO 395
Downloaded by [89.210.164.63] at 23:20 08 April 2016
on the metallic surface due to differences in the corrosion
potential of the various constituent materials that are in
contact. A widely applied technique to avoid such
phenomena is the sealing of the bonded structures with
coatings that prevent the bond area to come in contact
with the environment and eliminate galvanic current.
11
In order to enhance the performance of the steel to FRC
bond, nanofillers like multiwall carbon nanotubes (CNTs)
may be incorporated in the adhesive layer in order to
improve the shear strength of the interface and reduce
the probability of delamination and failure of the joint.
12
An additional advantage for CNT-modified polymers is
that they offergreater protection in corrosive environments
by reducing the water uptake in the polymer mass resulting
in better performance in terms of durability,
13,14
or even
improving the cross-linking of the epoxy.
14
However, the
galvanic corrosion is again an issue for the nano-modified
adhesives since in some cases the new conductive phase, i.e.
the CNTs, may accelerate galvanic corrosion in some met-
allic systems like aluminium.
15,16
In order to fully exploit the enhancement capabilities of
the CNT reinforcement, various parameters have to be
addressed. Such parameters include the selected type of
carbon nanotubes (single wall, multiwall, functionalised
or not, etc.), epoxy system,
17–19
the weight content of
the CNTs in the epoxy
20
and the employed dispersion
methodology. When it comes to dispersion methodology
various techniques like sonication,
21,22
high-speed shear
mixing
23,24
and three-roll calender mixing
25
are
employed.
Previous studies on the mechanical and environmental
performance of Al alloy (for aircraft applications) adhe-
sively bonded with CNT-modified epoxy systems demon-
strated the dependence of the response of the system on
the type of adhesive, hardener, process parameters and
CNT content.
12,13,15
In view of the above, this research focuses on the per-
formance of stainless steel adhesively bonded to neat
epoxy, CNT-modified epoxy and carbon fibre (CF)-
reinforced epoxy, with the objective to explore the poten-
tial of using CNT-modified adhesives to repairs in off-
shore applications. Shear strength and galvanic
corrosion measurements were conducted in order to
study the interfacial performance of the above polymer-
based adhesives to steel substrates in terms of mechanical
and electrochemical response. Due to the high position of
carbon in the galvanic series of metallic materials in sea-
water, the electrochemical study was performed in aqu-
eous NaCl, in order to assess if the repair is
electrochemically active or not when coupled with the
steel structure.
26
The epoxy was reinforced at the nano-
scale with CNTs in order to evaluate and quantify the
potential to improve adhesion and tailor the electroche-
mical response, as has already been done in previous
studies.
15,27
Experimental section
Materials
The SS304 steel alloy was selected for this study as a
substrate material as it is one of the most widely used
steel alloys in naval applications for pipelines, storage
tanks, structural parts of ships, etc. The LY 5052
epoxy resin/Aradur 5052 polyamine hardener system
provided by Huntsman was used both as adhesive and
as a matrix material for the CF composite due to its
low viscosity and curing cycle. The low viscosity ensures
good impregnation of the fibre reinforcement while the
versatile curing cycle enables the system to be applicable
in various operational conditions without the need for
special equipment. Based on the manufacturer data-
sheet, the selected curing cycle was curing for 24 hours
at 25°C and post-curing for 4 hours at 100°C in order
to achieve the optimum properties for the epoxy. The
Graphistrength multiwall CNTs provided by Arkema
(hereafter denoted as CNTs) were employed. Typical
diameters ranged from 10 to 15 nm while length ranged
from 1 to 10 µm. These CNTs came in the form of
agglomerates with average dimensions of 400 µm with
a range from 50 to 900 µm.
The fibre reinforcement was a 140 g m
−2
unidirectional
(UD) CF provided by R&G Composites. The CF-
reinforced polymers (CFRPs) were manufactured with
the aforementioned epoxy resin as the matrix material
with vacuum assisted hand lay-up, as it is an easily appli-
cable technique for on-site repairs with no special equip-
ment requirements and the final material possesses very
satisfactory properties.
Specimen preparation for electrochemical
testing
Τhe austenitic stainless steel 304 (SS304) was used as a
substrate for the electrochemical measurements. Neat
and nano-modified resins were employed in five different
configurations, as shown in Table 1. Rectangular cou-
pons were cut from steel sheets to be used as bare
metal reference specimens and as substrates to the coat-
ings presented in Table 1. The bare metal surfaces were
polished using emery paper (grit from 400 to 1000) and
rinsed with distilled water and acetone. On the back sur-
face of each coupon, a copper wire was attached using
aluminium conductive tape. The assembly was insulated
using vacuum sealant tape and Teflon to mask the per-
iphery and the back of the specimens, so as to ensure
that: (i) only a surface area of ∼1cm
2
would be exposed
to the electrolyte, (ii) the cutting edges and the periphery
of the specimens, being susceptible to stress corrosion
cracking because of cold working (cutting) due residual
stresses and dislocations, would not affect the corrosion
process.
Table 1 Nomenclature of specimens and coatings used in the lap shear and electrochemical tests
Epoxy resin
configuration
Neat
epoxy
resin
Epoxy resin +
0.4% w/w CNTs
Epoxy resin +
0.5% w/w CNTs
Epoxy resin +
0.6% w/w CNTs
CF-reinforced
epoxy
CF-reinforced epoxy
+ 0.6% w/w CNTs
Specimen
nomenclature
MtM neat MtM doped
0.4%
MtM doped
0.5%
MtM doped
0.6%
MtC neat MtC doped
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
96 Plastics, Rubber and Composites 2016 VOL 45 NO 3
Downloaded by [89.210.164.63] at 23:20 08 April 2016
Specimen preparation for mechanical testing
For all single lap-joint shear tests, SS304 and CFRPs were
used in two different configurations: one employing only
metallic substrates bonded together (metal to metal con-
figuration, MtM, according to ASTM D 1002-01) and a
second one was CFRP strips were bonded with SS304
strips (metal to composite, MtC, according to D 5868-
01) in a co-cure protocol. The dimensions of the final
specimens can be seen in Fig. 1. The bond area was 2.5
cm in length. In order to ensure the aforementioned
bond length, Teflon tape was wrapped around the speci-
mens at the predefined length. The vacuum assisted
hand lay-up method was used for the CFRP composite
manufacturing and the steel strips were co-cured with
the CFRPs. To achieve similar stiffness as the steel sub-
strate, 15 layers of UD CF fabric were used for the
CFRP laminate. Before bonding, the stainless steel sur-
face was sand blasted.
Electrochemical testing
All electrochemical measurements were performed in a
standard three-electrode electrochemical cell based on
ASTM G5,
28
with Ag/AgCl (3.5 M KCl, E
AgCl
=E
SHE
−200 mV) as the reference electrode, a platinum gauze
as the auxiliary electrode (AE) and the specimen as the
working electrode (WE). The GILL AC was provided
by ACM Instruments. Galvanostat/potentiostat was
employed. Potentiodynamic polarisation tests were car-
ried out at a scanning rate of 10 mV min
−1
. The rest or
steady state potential (E
rest
) was determined after 2 h of
immersion in 3.5% NaCl, at 25°C (open circuit state). Fol-
lowing the determination of the E
rest
, potentiodynamic
polarisation curves were recorded. Compensation for IR
drop was automatically implemented on the potentiostat
by measuring the solution resistance between the WE
and the reference electrode with an AC signal and then
adjusting the output voltage (using Ohm’s law) with a pro-
portional integral derivative control algorithm.
Reverse polarisation was adopted in order to study the
susceptibility of the systems to localised degradation. The
main concept of this technique is that localised
degradation would occur if the current density of the
anodic portion of the return scan is higher than the cur-
rent density of the forward scan for the same anodic
potential. This type of hysteresis is labelled as ‘negative
hysteresis’.
29
In order to study the galvanic effect between the metal
structure and the repair patch, the galvanic current vs.
time was continuously recorded by electrically connecting
the couple constituents to a zero resistance ammeter (Gill
AC by ACM Instruments of current range 10 pA to 500
mA). The galvanic effect of the following couples was
evaluated: SS304–SS304 (as a baseline), SS304–MtM
neat, SS304–MtM doped 0.6% and SS304–MtC neat.
SS304 coupon was connected to the working electrode 1
(WE1) input of the galvanostat and the coated steel to
the working electrode 2 (WE2). Should the galvanic cur-
rent of the couple receive positive values, then WE1 is,
by default, anodic to WE2. The electrolyte provided a
means for ion migration, whereby metallic ions moved
from the anode to the cathode. A positive current density
denoted the movement of the metallic ions from the
anode to the cathode through the electrolyte. The anode
would corrode more quickly than if it was on its own in
the electrolyte and, at the same time, the cathode cor-
rosion would be retarded. Both of the aforementioned
measurements were conducted in a 3.5% w/w NaCl aqu-
eous solution.
Mechanical testing
For the mechanical testing, the single lap-joint shear test-
ing was employed based on ASTM D 1002-01, D 5868-01
for the MtM and MtC geometry, respectively.
30,31
Two
groups of specimens (MtM and MtC) were manufactured
in six different configurations as indicated in Table 2.
The shear strength tests were conducted on an Instron
5967 universal testing machine with extension rate set to
1.3 mm s
−1
. Neat resin and doped resins with 0.4, 0.5
and 0.6% w/w CNTs were used as adhesives for the
MtM specimen configuration. As far as the MtC con-
figuration is concerned, two composite plates were manu-
factured and co-cured with the metallic strips. One plate
incorporated neat resin and the second one incorporated
doped resin in 0.6% w/w CNTs loading as the matrix.
CNTs–epoxy dispersion process
The CNTs dispersion in the epoxy resin was conducted in
a laboratory dissolver apparatus, (Dispermat AE dissol-
ver) equipped with a bead milling module (Torus Mill).
Seventy grams of 0.1 mm diameter beads was used. The
dispersion lasted for 45 minutes at 2000 rpm for each
different CNT loading. All doped resins were extracted
from a master batch doped resin with 2-hour dispersion
at 2000 rpm. Both the Dissolver and the Torus Mill mod-
ule were equipped with double walled containers and a
Grant pump and cooling bath was connected in order
to control the temperature at 25 ± 1°C.
Results and discussion
Single lap-joint shear test
As mentioned before, two different single lap-joint speci-
men configurations were used. In Fig. 2 typical indicative
load vs. extension diagrams are depicted from the respect-
ive tests. As can be seen, both of the studied
1 Specimen dimensions for the single lap-joint specimen
(top) and manufactured specimens from the MtC configur-
ation (bottom)
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
Plastics, Rubber and Composites 2016 VOL 45 NO 397
Downloaded by [89.210.164.63] at 23:20 08 April 2016
configurations exhibit curves indicative of brittle
materials, as was expected. Figure 3 depicts a cumulative
bar chart with the average shear strength for the two
groups, derived from data of five test specimens for each
configuration and CNTs loadings. Ιn the case of MtM
configuration, a significant increase in shear strength is
exhibited after the addition of the CNTs in the host
matrix. The shear strength for the doped resin specimens
is improved by 35, 43 and 47% for 0.4, 0.5 and 0.6% CNT
loading, respectively, when compared to the neat resins.
All the respective average and maximum shear stress
values are shown in Table 2. The results indicate that
the shear strength is CNT loading dependent since for
increasing CNT loading the shear strength improves mar-
ginally. Improvements in shear strength have been
reported in the literature for nano-enhanced epoxy
adhesives,
12
albeit, to the authors’knowledge, not at the
order of 50%. It is generally accepted that CNTs act
both as reinforcement and as tougheners (via crack
deflecting and bridging) in the epoxy matrix resulting in
improved properties, in our case shear strength. The high-
est shear strength was achieved for 0.6% CNT loading
however the average values for the doped resins are
close and within standard deviation values. Based on the
aforementioned results, the 0.6% w/w loading was used
for the doped CFRP lamina for the MtC specimen
manufacture.
For the MtC configuration two CFRP plates were man-
ufactured, one with neat resin (MtC neat) and one with
0.6% doped resin (MtC doped) as the matrix material.
As can be seen in Fig. 3 and Table 2, the presence of
the CNTs in the matrix resulted to the reduction of the
shear strength of the adhesive bond. This may be attribu-
ted to poor wettability of the modified resin on the steel
surface, in conjunction with the increased viscosity caused
by the incorporation of CNTs. Indeed, this behaviour can
clearly be seen in Fig. 4, where stereoscopic micrographs
of the fractured lap shear faces are depicted. The figure is
covering, in both cases at approximately the same scale,
slightly more than two UD tows, with the transverse
stitching appearing in the foreground for one of them.
In Fig. 4a, the neat composite face of a lap shear coupon
is depicted after failure; the matrix is clearly seen to infil-
trate the fibres, and almost uniformly covering the steel/
composite interface. In contrast, the doped composite
face (Fig. 4b) exhibits areas where the matrix has not
been able to infiltrate the fibres and fill the metal/compo-
site interface, particularly in the areas around the stitching
of the UD fabric. Thus, it is clearly seen that due to the
physical property changes invoked by the inclusion of
the CNTs, i.e. viscosity and/or surface tension, there is a
considerable reduction to the ‘effective’adhesion area
which is responsible for the deteriorated properties of
the doped system. In view of these observations, it may
be concluded that in order to fully exploit the potential
provided by the inclusion of CNTs, provisions must be
taken to improve the interface between the substrate
and the composite, either by including a doped resin
layer prior to the lay-up of the wet fabrics or by altering
the manufacturing conditions (higher vacuum and/or
higher temperature) if possible. In all cases, and as is
clearly seen in Fig. 3, MtC specimens (both neat and
doped) exhibit considerably higher values when compared
to the MtM ones (neat and doped). Hence from a repair
point of view, it is concluded that a CFRP patch repair
can lead to greater performance when compared to a
steel bonded repair.
Galvanic effect
Figure 5 presents the galvanic current vs. immersion time
plots for the couples: SS304–SS304, SS304–neat resin,
Table 2 Results from single lap-joint testing
Specimen MtM neat MtM 0.4% MtM 0.5% MtM 0.6% MtC neat MtC doped
Mean shear strength/MPa 4.68 ± 0.23 6.32 ± 0.32 6.71 ± 0.34 6.87 ± 0.34 10.51 ± 0.53 8.74 ± 0.44
Highest value/MPa 4.77 7.17 7.88 7.24 15.23 10.26
2 Typical load vs. displacement curves for the two systems. (MtM on the left and MtC on the right)
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
98 Plastics, Rubber and Composites 2016 VOL 45 NO 3
Downloaded by [89.210.164.63] at 23:20 08 April 2016
SS304–0.6 wt-% CNT doped resin and SS304–CF-
reinforced epoxy. As expected, the SS304–SS304 couple
exhibits current values close to nil. All electrode couples
exhibit positive current density values indicating that
the SS304 substrate is less noble when compared to the
various coatings. The relative nobility of the potential
repair patches implies that in the case that the electrolyte
and its reactive ions reach the metallic substrate, its cor-
rosion would be accelerated, as the substrate is anodic
to the patch. This set-up is susceptible to detachment of
the patch from its substrate in the case of electrolyte
access to the substrate.
32
The CNT doped epoxy presents the highest galvanic
effect with the substrate, whilst the epoxy reinforced by
CF presents the lowest galvanic effect. This finding is
unexpected, since the CNTs are expected to increase the
conductivity of the polymer
13,16,27
and, hence, decrease
the galvanic effect between a neat polymer and a (highly
conductive) metal. However, this notable galvanic effect
is compatible with the fact that the conductivity of com-
posites with identical filler concentration seems to vary,
with some exceptions, by one or two orders of magnitude
for identical matrices and by 10 or more orders of magni-
tude for different concentrations.
27
A major reason for
this is the existence of polymer tunnelling barriers
between CNTs in the form of sheathing layers around
the nanotubes, which result in the decrease in the overall
composite conductivity.
The current fluctuations (in the form of spikes) in the (i
vs. t) curve of couple SS304–MtM doped 0.6% in Fig. 5
can be attributed to instantaneous metastable pitting
phenomena in the anodic electrode (SS304), namely: pit
formation on SS304, deposition of instable corrosion pro-
ducts, dissolution of these products, reformation of pits,
re-passivation and so on. The maximum observed in the
(ivs. t) curve of couple SS304–MtM doped 0.6% at
about 400 min of immersion suggests a peaking of active
corrosion activity in SS304. Immersion for longer time
relieves this activity by re-passivation as the decreasing
trend of current density with time (sustained for the rest
of the immersion time) shows.
The lowest galvanic effect is recorded for the SS304–
CF-reinforced epoxy couple and it is attributed to the
high conductivity of the CFs.
Potentiodynamic polarisation
Overview
Figure 6 depicts the polarisation behaviour of all systems
immersed in 3.5% NaCl, at room temperature. Table 3
lists the critical electrochemical values of potential. As
can be seen in Fig. 6, the polarisation curves of the
0.6% CNT-modified epoxy and SS304 present the highest
current density differences, in compatibility with the gal-
vanic effect data.
Table 3 and Fig. 6 show that the CF-reinforced epoxy
presents the noblest corrosion potential most likely due
to the noble nature of the graphite reinforcement.
Before discussing the electrochemical response of the
patches, one has to bear in mind that water uptake in
polymer matrix composites (PMCs) follows three
routes:
33
(i) direct penetration into the bulk polymer, (ii)
water adsorption at the reinforcement/matrix interfaces
and (iii) water transport through various defects in the
matrix (pores, microcracks and, in general, areas of
poor cross-link density, etc.). Water diffuses into polymers
to a different extent depending on a number of molecular
and microstructural aspects, such as: (i) the polarity of the
molecular structure, (ii) the degree of cross-linking, (iii)
the degree of crystallinity in the case of a thermoplastic
and (iv) the presence of residuals in the material.
34
Also,
it has long been known that when only small amounts
of water diffuse into the polymer, conduction may occur
by a process of ‘activated diffusion’, whereby fixed iono-
genic groups attached to the polymer backbone (e.g.
amide/amine groups in an epoxy) act as sites for ion
3 Average shear strength for the two single lap joint speci-
men configurations (MtM and MtC)
4 Optical stereoscopic images of the fracture surfaces of the MtC coupons –composite face. aNeat composite and bdoped
composite
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
Plastics, Rubber and Composites 2016 VOL 45 NO 399
Downloaded by [89.210.164.63] at 23:20 08 April 2016
transfer through the membrane.
35
Exposure of PMCs to
aqueous chlorides leads to an increase in the effective
water diffusivity mostly owing to favouring water trans-
port through the bulk of the polymer.
35
The conductive behaviour of the neat epoxy (MtM
neat), seen in Fig. 6, is attributed to an increase in the
inherent low conductivity of the epoxy by the following
processes:
36
(i) water sorption by the free volume of the
polymer, (ii) hydrogen bonding of water molecules into
hydrophilic sites and (iii) localised water sorption in
areas of poor cross-link density due to solidification and
curing. In epoxy systems, regions of poor cross-link den-
sity can form networks (more or less extended), even if
they occupy very low volume fractions. At this point it
should be noted that the epoxy system used in the present
work shows different electrochemical values than the
epoxy systems employed in previous efforts,
12,13
suggesting that the electrochemical behaviour of polymers
is material and preparation method dependent. Indeed, in
the earlier work by Brown and Coomber
37
it was shown
that open circuit potentials of CFRP–metal couples can
vary significantly between resin systems.
The very low conductivity of the 0.6% CNT doped
epoxy is compatible with the galvanic effect
measurements and it can be explained, as aforemen-
tioned, by the existence of polymer tunnelling barriers
between CNTs in the form of sheathing layers around
the nanotubes. Two complementary reasons for the
observed very low corrosion kinetics of the 0.6% CNT
doped epoxy coating are: the hydrophobic nature of
CNTs
38
and the uniform distribution of CNT into the
polymer matrix. The latter may limit transport of electro-
lyte through microcracks or other forms of microdamage,
such as pores or small channels already present in the neat
polymer or generated by water attack, since CNTs can act
as crack deflectors. Zhang et al.
39
have shown that CNTs
may suppress fatigue crack growth in a 0.5 wt-% CNT-
reinforced epoxy by pull-out of nanotube fibres that
bridge the crack interface. Xia et al.
40
noted three tough-
ening mechanisms in CNT-reinforced Al
2
O
3
coatings:
crack bridging by CNTs, crack deflection at the CNT/
matrix interface and CNT pull-out on the fracture
surfaces.
The CF-reinforced polymer exhibits an intermediate
kinetic behaviour; CFs may also limit water transport
by the free polymer volume by acting as crack deflectors.
Anodic polarisation
The most common cathodic reaction in PMCs in water is
the reduction of oxygen:
37
O2+2H2O+4e−4OH−(1)
However, there is an uncertainty in identifying the
occurring anodic reactions. Organic polymer coatings,
formed on metallic substrate, show ideal behaviour as
dielectrics. During exposure to a corrosive agent, electro-
lyte activates conductive ‘tracks’in organic coatings that
may lead the electrolyte to the coating/metal interface;
there, anodic dissolution of metal will take place.
41
Hence, in the case of a steel substrate, a likely anodic reac-
tion would be the oxidation of Fe in discontinuities of the
passive film. However, Fig. 6 shows that the polymer coat-
ings exhibit anodic polarisation behaviour that is greatly
different than that of SS304, in terms of anodic polaris-
ation curve shape, overpotential and current density
values. Hence, it is indicated that the electrolyte inter-
action with the coated substrate is much different than
the electrolyte interaction with the bare steel.
A possible interaction can be postulated taking into
account the galvanic effect between the patches and
SS304, as firstly described in the technical literature by
Faudree.
42
Electrochemical reactions will initiate by the
anodic dissolution of iron in the defects of the passive
film.
2Fe 2Fe2++4e−(2)
According to Faudree,
42
accumulation of OH
−
ions
produced by the reduction of oxygen at the graphite
fibre/polyamide interface will lead to an increase in the
pH in these confined areas. There, a form of ‘crevice cor-
rosion’of the polymer will take place leading to the for-
mation of amide salts. However, epoxies are considered
resistant to this type of attack, unless oxygen reduction
has occurred in areas of low cross-link density. Further-
more, there is always the possibility of production of per-
hydroxyl ions (HOO
−
) by oxygen reduction:
43
O2+2H2O+2e−2HOO−+2OH−(3)
5 Galvanic current vs. time for the couples: SS304–SS304,
SS304–MtM neat, SS304–MtM doped 0.6% and SS304–
MtC neat; 3.5 wt-% NaCl, 25°C
6 Potentiodynamic anodic polarisation curves of bare stain-
less steel (SS304) and stainless steel coated by neat
epoxy (MtM neat), CF-reinforced epoxy (MtC neat) and
0.6% CNT doped epoxy (MtM doped 0.6%), in 3.5 wt-%
NaCl at 25°C
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
100 Plastics, Rubber and Composites 2016 VOL 45 NO 3
Downloaded by [89.210.164.63] at 23:20 08 April 2016
HOO−+H2O+2e−3OH−(4)
HOO
−
ions, in the presence of an electrophilic centre, can
react very rapidly with the polymer, before they decom-
pose by reaction (4).
Another likely anodic reaction in the three types of
patches could be carbon dissolution at areas of low
cross-link density:
C+2H2OCO2+4H++4e−(5)
The electrochemical potential of reaction (5) is E
0
=
0.007 V vs. Ag/AgCl in acid media, but it is believed to
be slow in that potential range. Therefore, carbon oxi-
dation should normally occur at high overpotentials,
unless a catalytic process lowers the activation energy.
Nevertheless, a carbon with a relatively high degree of dis-
order, as carbon in poor cross-linked polymer areas,
would be prone to oxidisation more rapidly than graphi-
tised/ordered carbon.
44
In the case of existence of electroactive surface func-
tional groups in the polymer, another likely anodic reac-
tion could be their oxidation.
45
As pH increases at the cathode, due to the liberation of
OH
−
by reaction (1),CO
2
produced from reaction (5) as
well as CO
2
in the aerated NaCl solution will undergo
hydrolysis:
46
H2O+CO2H++HCO−
3(6)
HCO−
3H++CO2−
3(7)
As pH increases at the cathode, CO2−
3production is ther-
modynamically favoured.
46
It is well known that osmosis occurs when the resin acts
as a semi-permeable membrane allowing only water to
diffuse. Therefore, if the resin contains water-soluble sub-
stances (such as carbonates, NaOH from reaction of Na
2+
with OH
−
or water-soluble impurities), thermodynamics
will drive water further into the resin. Osmotic pressure
(π) is a colligative property directly proportional to the
molar concentration of the soluble substance:
34
p
=RTc (8)
where Ris the gas constant, Tis the absolute temperature
and cis the molar concentration of the solubilised parts.
Therefore, water-soluble substance segregation in
microsurfaces such as, microdefects or filler/matrix
interfaces, will drive further epoxy degradation by
osmosis.
The anodic polarisation curves of the neat epoxy coat-
ing and the 0.6 CNT-modified coating present stabilis-
ation current regimes at the high anodic potentials,
which are sustained through the whole range of the
measured potentials. The anodic polarisation curve of
the neat epoxy reinforced by CF exhibits a ‘passive’
regime starting at −54 mV vs. Ag/AgCl and sustained
for about 870 mV. These regimes can be explained by
the saturation of the polymer
35
with water, the deposition
of polymer reductive degradation products that limit dif-
fusion of O
2
or even the deposition of unstable polymer
oxidation products.
47
In the case of the CNT- and CF-
reinforced polymers, these current stabilisation regimes
start at E
cp
potentials that are much closer to the cor-
rosion potential in comparison with the neat polymer
coating (see (E
cp
−E
corr
) values in Table 3), because: (i)
they present a polymer free volume that is smaller than
that of the neat polymer and (ii) the reinforcements may
limit water transport through defects of the polymer
structure, as previously mentioned.
Figure 6 and Table 3 also show that the forward anodic
curves of the neat epoxy and the CF-reinforced epoxy
present breakaway potentials (i.e. flattened gradients sus-
tained for one to two orders of current density magni-
tude, similar to those observed in the polarisation
curves of metallic materials susceptible to localised
forms of corrosion). Thus, these flattened gradients
may be explained by localised degradation of low resist-
ance conduction pathways associated with the localised
presence of deficient cross-link density regions.
13
In the
case of CF-reinforced epoxy, an additional reason for
the localised degradation can be the localised existence
of weak matrix/CF interfaces that induce preferential
water adsorption at the fibre/matrix interfaces.
15
(In
PMCs, water has the tendency to segregate at the
reinforcement/matrix interface,
48
where dissolution of
carbon may take place; further water ingress due to
osmotic effect will intensify localised degradation, as
aforementioned.)
The 0.6% CNT doped epoxy does not present deviation
from a uniform corrosion behaviour, indicating a uniform
CNT distribution that has densified localised poor cross-
link density regions.
Reverse polarisation
Figure 7 shows the cyclic polarisation behaviour of all sys-
tems in aqueous 3.5% NaCl. The negative hysteresis in
Fig. 7a(namely higher ‘reverse’currents than the ‘for-
ward’ones at the same potential in the anodic scan)
suggests that bare SS304 is susceptible to pitting in aqu-
eous NaCl. Pitting in stainless steels is associated with
local discontinuities in the passive film. Two of the most
common local discontinuities are sulphide (Mn,Fe)S
inclusions and small regions of low Cr
2
O
3
in a wide
matrix of high Cr
2
O
3
content.
49
Regarding the first type
of discontinuities, the low pH and high sulphide ion con-
centration resulting from the dissolution of the sulphide
Table 3 Electrochemical values of the materials immersed in 3.5% NaCl
Material
E
corr
/mV,
Ag/AgCl
E
b
/mV,
Ag/AgCl
E
cp
/mV,
Ag/AgCl E
b
−E
cp
/mV E
cp
−E
corr
/mV
SS304 −362 −85 (E
b1
) 350(E
b2
)−56 406 306
MtM neat −656 −47 −7 609
MtM doped 0.6% −497 −280 217
MtC neat −170 797 −54 967 116
E
corr
, equilibrium potential; E
b
, breakaway or breakdown potential; E
cp
, potential at the start of
current limiting domain.
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
Plastics, Rubber and Composites 2016 VOL 45 NO 3101
Downloaded by [89.210.164.63] at 23:20 08 April 2016
inclusions produce a local environment that prevents
establishing passivity in the pit.
50
Regarding the second
type of discontinuities, reductive dissolution of the
Fe
2
O
3
component of the Cr
2
O
3
-poor passive film
occurs:
49
Fe2O3+6H++2e−2Fe2++3H2O (9)
Reaction (9) leads to a decrease in the film thickness and
localised breakdown of passivity.
The neat epoxy and the CF-reinforced epoxy coatings
present negative hysteresis indicative of localised degra-
dation phenomena in compatibility with the aforemen-
tioned breakaway state upon forward polarisation. The
negative hysteresis in the polarisation curves of the neat
epoxy patch and the CF-reinforced patch suggests that
surfaces during reverse scanning are more conductive
than the ones during forward scanning at the same poten-
tial. The comparatively high conductivity of the involved
surfaces during reverse polarisation is attributed to the
fact that they contain more water and water-soluble sub-
stances than the respective ones during forward polaris-
ation. The much larger area of the negative hysteresis
loop in the curve of the CFR epoxy, as compared to
that of the neat epoxy, can be explained by the additional
water uptake by the CF/matrix interfaces. There, the
osmosis driven increased concentration of water-soluble
products has led to surface conductivities that are con-
siderably higher than the respective ones during forward
polarisation.
The only system presenting a positive hysteresis is the
CNT doped coating due to the absence of localised degra-
dation phenomena. The slightly lower surface conduc-
tivity of the 0.6% CNT doped patch during reverse
polarisation in comparison with the surface conductivity
during forward polarisation indicates a reversibility in
the water ingress processes. Thus, it is suggested that the
0.6% CNT doped patch has not suffered irreversible
damage (e.g. damage due to hydrolysis of chemical
bonds or interface rupture due to osmotic pressure).
Indeed, Fig. 8 confirms the evidence of cyclic polarisation
for the CNT-modified epoxy adhesive: the stainless steel
substrate does not present any signs of corrosion after cyc-
lic polarisation in aqueous NaCl. The absence of any cor-
rosion products from the surface of the steel substrate is
manifested in Fig. 9. The line scan across the imaged
cross-section reveals a uniform distribution of oxygen
along the line, confirming the absence of any corrosion
depositions on the steel substrate. Should the electrolyte
reached the surface of the substrate, oxygen concentration
7 Cyclic polarisation behaviour of abare steel (SS304), bsteel coated by neat epoxy (MtM neat), csteel coated by 0.6% CNT
doped epoxy (MtM doped 0.6%) and dsteel coated by CF-reinforced epoxy (MtC neat), in 3.5 wt-% NaCl at 25°C
8 SEM cross-sectional micrograph of stainless steel sub-
strate and 0.6% CNT-modified epoxy patch after cyclic
polarisation in 3.5% NaCl, 25°C. The surface of the sub-
strate is free of corrosion signs
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
102 Plastics, Rubber and Composites 2016 VOL 45 NO 3
Downloaded by [89.210.164.63] at 23:20 08 April 2016
at the surface would be higher than that in the bulk of
steel. Surface concentration of Fe would also be different.
(The abrupt decrease in the concentration of Fe in the line
scan of the steel corresponds to the large crack formed
upon cross-sectioning and grinding.)
Therefore, although anodic polarisation may have been
initiated by dissolution of Fe
2+
(reaction (2)), carbon dis-
solution (reaction (5)) and/or oxidation of polymer sur-
face functional groups seem most likely as anodic
reactions at least for the continuation of oxidation,
depending on the pH and the anodic overpotential.
Cathodic polarisation
As expected, the stainless steel presents the highest cathodic
current densiti es (Fig.6), namely the highest fraction of con-
ductive surfaces large enough to support cathodic reactions.
The 0.6% CNT-modified epoxy adhesive presents the lowest
cathodic currents, namely the lowest fraction of conductive
surfaces large enough to support cathodic reactions.
At the polymer cathode, oxygen will be reduced accord-
ing to reaction (1). Alternatively, reductive degradation of
the polymer may occur as a result of electron transfer to
functional groups. The process is enhanced by electroac-
tive agents, such as solvent hydrolysis products.
51
The cathodic polarisation curve of SS304 differs from
the cathodic curve of the epoxy-based coatings in that it
presents a deflection denoting the change in the governing
cathodic reaction. In more detail, the cathodic branch has
started with reduction of oxygen ending in a short concen-
tration polarisation regime; at lower potentials, the deflec-
tion in the curve shows that the reduction of hydrogen
became operative.
Conclusions
The present work focused on the evaluation of the
adhesion properties of nano-reinforced epoxies and their
composites bonded on typical stainless steels for offshore
applications. Both the mechanical and electrochemical
behaviour of the studied systems was evaluated. The
study was performed in order to assess the potential appli-
cation of FRP bonded repair in naval structures.
The assessment of the adhesion efficiency was per-
formed for two configurations, i.e. MtM bonding with
neat and nano-modified epoxy employed as an adhesive
and MtC bonding where the composite was co-cured
with the steel substrate in either plain (neat epoxy matrix)
or doped (CNT-reinforced matrix) form. Vacuum assisted
hand lay-up was employed as the manufacturing method-
ology, as it is the most suitable candidate method for on-
site applications. The overall assessment of the nano-
reinforcement indicated that the inclusion of CNTs may
lead to beneficial results, both for electrochemical and
mechanical properties.
The nano-modification of the epoxy in the case of the
MtM bonding indicated an impressive improvement of
almost 50% in single lap shear strength. The improvement
was attributed to the well-known stiffening and toughen-
ing effects of the CNTs, resulting cumulatively from the
CNT elastic properties, as well as from crack bifurcation
and arrest, crack reflection and bridging. As should be
noted, although CNTs have been reported to improve
lap shear strength in similar systems (e.g. aluminium to
aluminium bonds) an improvement of almost 50% is
reported for the first time to the authors’knowledge.
The best performance was achieved for the highest CNT
content examined (i.e. 0.6% w/w CNTs) but, within exper-
imental error, it was achieved already at 0.5% w/w.
The MtC bonds performed significantly better to MtM
bonds, exhibiting more than double the lap shear strength
in the case of the neat systems, clearly showing the potential
of FRP repairs. However, the ‘CNT-modified MtC bond’
did not performed as good as the ‘neat MtC bond’. This
was attributed to the physical property changes of the
matrix due to the inclusion of the CNTs which should be
accounted for in orderto fully exploit and properly quantify
the potential of the matrix nano-modification.
As far as the electrochemical behaviour is concerned, the
incorporation of CNTs in the epoxy increased the galvanic
effect between epoxy and SS304. In all cases, the SS304 sub-
strate appears less noble compared to the various coatings/
composites. The CNT doped epoxy is exhibiting the highest
galvanic effect compared to the other galvanic couples
attributed to the existence of polymer tunnelling barriers
around CNTs. The lowest galvanic effect is recorded for
couple SS304-CF reinforced epoxy and it is attributed to
the high conductivity of the CFs. Additionally, both neat
9 SEM cross-sectional micrograph of stainless steel sub-
strate and 0.6% CNT-modified epoxy patch after cyclic
polarisation in 3.5% NaCl, 25°C. The surface of the sub-
strate is free of corrosion signs. The line scan across
the imaged cross-section reveals a uniform distribution
of oxygen along the line
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
Plastics, Rubber and Composites 2016 VOL 45 NO 3103
Downloaded by [89.210.164.63] at 23:20 08 April 2016
epoxy and CF-reinforced epoxy systems present negative
hysteresis during the reverse polarisation semi-cycle indica-
tive of localised degradation. In its turn, this is indicative of
higher conductivity and is attributed to water uptake and
osmosis driven localised degradation. In the case of the
CF-reinforced epoxy system, localised degradation may
be promoted due to additional water uptake and osmotic
effects at the CF/matrix interface. Interestingly enough,
this is not the casefor the CNT doped resin, where the mar-
ginally lower surface conductivity during reverse polaris-
ation in comparison with the surface conductivity during
forward polarisation directly implies the absence of loca-
lised degradation phenomena, suggesting that the patch
has not sufferedirreversible damage. The latter is confirmed
by SEM/EDX examination of the 0.6% CNT/epoxy coated
SS304 specimen.
Recapitulating, the CNT modification may result to an
impressive increase in the bond strength of steel to steel
joints. The FRP to metal bond performs considerably bet-
ter compared to the MtM bond, but caution should be
taken in order to exploit the CNT matrix modification,
as the bond strength is governed by the changes invoked
in the physical properties of the matrix. In this case,
further study is required in order to assess if the effect
of CNT inclusion is beneficial. Finally, the inclusion of
CNTs was found to substantially increase the galvanic
effect, but at the same time to suppress localised degra-
dation phenomena leading to increased protection of
the steel substrate.
ORCID
A. S. Paipetis http://orcid.org/0000-0001-9668-9719
References
1. M. Shamsuddoha, M. M. Islam, T. Aravinthan, A. Manalo and K.-
T. Lau: ‘Effectiveness of using fibre-reinforced polymer composites
for underwater steel pipeline repairs’,Compos. Struct.,2013,100,
40–54.
2. J.-S. Lee, Y. S. Ryu, N.-I. Kim, Y.-K. Kim and M.-H. Kim: ‘Stud
welding for fixation of cryogenic insulation of membrane tanks in
LNG ship building’,Trans. Nonferr. Metals Soc. China,2009,19,
271–275.
3. X.-L. Zhao and L. Zhang: ‘State-of-the-art review on FRP strength-
ened steel structures’,Eng. Struct.,2007,29, (8), 1808–1823.
4. M. Gholami, A. R. M. Sam, J. M. Yatim and M. M. Tahir: ‘A review
on steel/CFRP strengthening systems focusing environmental per-
formance’,Constr. Build. Mater.,2013,47, 301–310.
5. Ç. Özes and N. Neser: ‘Experimental study on steel to FRP bonded
lap joints in marine applications’,Adv. Mater. Sci. Eng.,2015,2015,
1–6.
6. C. Wu, X. Zhao, W. Hui Duan and R. Al-Mahaidi: ‘Bond charac-
teristics between ultra high modulus CFRP laminates and steel’,
Thin Wall. Struct.,2012,51, 147–157.
7. I. Yoshitake, H. Tsuda, J. Itose and N. Hisabe: ‘Effect of discrepancy
in thermal expansion coefficients of CFRP and steel under cold
temperature’,Constr. Build. Mater.,2014,59,17–24.
8. A. Baldan: ‘Review adhesively bonded joints and repairs in metallic
alloys, polymers and composite materials: adhesives, adhesion the-
ories and surface treatment’,J. Mater. Sci.,2004,39,1–49.
9. E. Jarryand R. A. Shenoi: ‘Performance of butt strap joints for marine
applications’,Int. J. Adhes. Adhes.,2006,26, (3), 162–176.
10. T. Yu, D. Fernando, J. G. Teng and X. L. Zhao: ‘Experimental study
on CFRP-to-steel bonded interfaces’,Compos. Part B-Eng.,2012,
43, (5), 2279–2289.
11. D. Schnerch, M. Dawood, S. Rizkalla, E. Sumner and K. Stanford:
‘Bond behavior of CFRP strengthened steel structures’,Adv. Struct.
Eng.,2006,9, (6), 805–817.
12. G. Gkikas, D. Sioulas, A. Lekatou, N. M. Barkoula and A. S.
Paipetis: ‘Enhanced bonded aircraft repair using nano-modified
adhesives’,Mater. Des.,2012,41, 394–402.
13. G. Gkikas, A. Lekatou, D. Sioulas and A. S. Paipetis: ‘Effect of car-
bon nanotube enhanced adhesives on degradation of bonded joints
in corrosive environments’,Plast. Rubber Compos.,2014,43, (10),
322–329.
14. G. Gkikas, D. D. Douka, N. M. Barkoula and A. S. Paipetis: ‘Nano-
enhanced composite materials under thermal shock and environ-
mental degradation: a durability study’,Compos. Part B-Eng.,
2015,70, 206–214.
15. G. Gkikas, A. Paipetis, A. Lekatou, N. M. Barkoula, D. Sioulas, B.
Canflanca and S. Florez: ‘Corrosion and environmental degradation
of bonded composite repair’,Int. J. Struct. Integr.,2013,4, (1), 67–
77.
16. R. Ireland, L. Arronche and V. La Saponara: ‘Electrochemical
investigation of galvanic corrosion between aluminum 7075 and
glass fiber/epoxy composites modified with carbon nanotubes’,
Compos. Part B-Eng.,2012,43, (2), 183–194.
17. A. Godara, L. Mezzo, F. Luizi, A. Warrier, S. V. Lomov, A. W. van
Vuure, L. Gorbatikh, P. Moldenaers and I. Verpoest: ‘Influence of
carbon nanotube reinforcement on the processing and the mechan-
ical behaviour of carbon fiber/epoxy composites’,Carbon,2009,47,
(12), 2914–2923.
18. M. T. Kim, K. Y. Rhee, J. H. Lee, D. Hui and A. K. T. Lau: Property
enhancement of a carbon fiber/epoxy composite by using carbon
nanotubes’,Compos. Part B-Eng.,2011,42, (5), 1257–1261.
19. P.-C. Ma, S.-Y. Mo, B.-Z. Tang and J.-K. Kim: ‘Dispersion, inter-
facial interaction and re-agglomeration of functionalized carbon
nanotubes in epoxy composites’,Carbon,2010,48, (6), 1824–1834.
20. N. W. Khun, B. C. R. Troconis and G. S. Frankel: ‘Effects of carbon
nanotube content on adhesion strength and wear and corrosion
resistance of epoxy composite coatings on AA2024-T3’,Prog. Org.
Coat.,2014,77, (1), 72–80.
21. G. Gkikas, N. M. Barkoula and A. S. Paipetis: ‘Effect of dispersion
conditions on the thermo-mechanical and toughness properties of
multi walled carbon nanotubes-reinforced epoxy’,Compos. Part B-
Eng.,2012,43, (6), 2697–2705.
22. A. Montazeri and M. Chitsazzadeh: ‘Effect of sonication par-
ameters on the mechanical properties of multi-walled carbon nano-
tube/epoxy composites’,Mater. Des.,2014,56, 500–508.
23. A. K. Chakraborty, T. Plyhm, M. Barbezat, A. Necola and G. P.
Terrasi: ‘Carbon nanotube (CNT)–epoxy nanocomposites: a sys-
tematic investigation of CNT dispersion’,J. Nanopart. Res.,2011,
13, (12), 6493–6506.
24. G. C. Papanicolaou, K. P.Papaefthymiou, A. F. Koutsomitopoulou,
D. V. Portan and S. P. Zaoutsos: ‘Effect of dispersion of MWCNTs
on the static and dynamic mechanical behavior of epoxy matrix
nanocomposites’,J. Mater. Sci.,2011,47, (1), 350–359.
25. J. Z. Kovacs, R. E. Mandjarov, T. Blisnjuk, K. Prehn, M. Sussiek, J.
Muller, K. Schulte and W. Bauhofer: ‘On the influence of
nanotube properties, processing conditions and shear forces on the
electrical conductivity of carbon nanotube epoxy composites’,
Nanotechnology,2009,20, (15), 155703.
26. Z. F. Yin, M. L. Yan, Z. Q. Bai, W. Z. Zhao and W. J. Zhou:
‘Galvanic corrosion associated with SM 80SS steel and Ni-based
alloy G3 couples in NaCl solution’,Electrochim. Acta,2008,
53, (22), 6285–6292.
27. W. Bauhofer and J. Z. Kovacs: ‘A review and analysis of electrical
percolation in carbon nanotube polymer composites’,Compos.
Sci. Technol.,2009,69, (10), 1486–1498.
28. ‘Standard reference test method for making potentiostatic and
potentiodynamic anodic polarization measurements’, G5, ASTM,
1999.
29. A. Lekatou, D. Sioulas, A. Karantzalis and D. Grimanelis: ‘A com-
parative study on the microstructure and surface property evalu-
ation of coatings produced from nanostructured and conventional
WC–Co powders HVOF-sprayed on Al7075’,Surf. Coat. Technol.,
2015,276, 539–556.
30. ‘Standard test method for apparent shear strength of single-lap-joint
adhesively bonded metal specimens by tension loading (metal-to
metal)’, D1002, ASTM, 2001.
31. ‘Standard test method for lap shear adhesion for fiber reinforced
plastic (FRP) bonding’, D5868, ASTM, 2001.
32. J. M. Guilemany, S. Dosta and J. R. Miguel: ‘The enhancement of
the properties of WC-Co HVOF coatings through the use of nanos-
tructured and microstructured feedstock powders’,Surf. Coat.
Technol.,2006,201,(3–4), 1180–1190.
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
104 Plastics, Rubber and Composites 2016 VOL 45 NO 3
Downloaded by [89.210.164.63] at 23:20 08 April 2016
33. G. Marom: ‘The role of water transport in composite materials’in
‘Polymer permeability’, (ed. J. Comyn), 341–374; 1985, Dordrecht,
Springer Netherlands.
34. F. Jones: ‘The effects of aggressive environments on long-term
behaviour’,in‘Fatigue in composites’, (ed. B. Harris), 117–146;
2003, Cambridge, Woodhead Publishing Limited.
35. A. Lekatou, S. E. Faidi, D. Ghidaoui, S. B. Lyon and R. C.
Newman: ‘Effect of water and its activity on transport properties
of glass/epoxy particulate composites’,Compos. Part A,1997,
28A, 223–236.
36. A. Lekatou, S. E. Faidi, S. B. Lyon and R. C. Newman: ‘Elasticity
and fracture in particulate composites with strong and degraded
interfaces’,J. Mater. Res.,1996,11, (5), 1293–1304.
37. A. Brown and D. E. Coomber: ‘Behaviour of couples of aluminium
and plastics reinforced with carbon fibre in aqueous salt solutions’,
Br. Corros. J.,1972,7, (5), 232–235.
38. J. H. Walker, R. Jaffe, T. Halicioglu and P. Koumoutsakos: ‘Carbon
nanotubes in water: structural characteristics and energetics’,
J. Phys. Chem. B,2001,105, (41), 9980–9987.
39. W. Zhang, R. C. Picu and N. Koratkar: ‘Suppression of fatigue
crack growth in carbon nanotube composites’,Appl. Phys. Lett.,
2007,91, (19), 193109.
40. Z. Xia, L. Riester, W. A. Curtin, H. Li, B. W. Sheldon, J. Liang, B.
Chang and J. M. Xu: ‘Direct observation of toughening mechanisms
in carbon nanotube ceramic matrix composites’,Acta Mater.,2004,
52, (4), 931–944.
41. Z. Z. Lazarevic, V. B. Miškovic-Stankovic, Z. Kacarevic-Popovic
and D. M. Drazic: ‘The study of corrosion stability of organic
epoxy protective coatings on aluminium and modified aluminium
surfaces’,J. Brazil. Chem. Soc.,2005,16, (1), 98–102.
42. M. C. Faudree: ‘Relationship of graphite/polyimide composites to
galvanic processes’,SAMPE J.,1991,2, 1288–1301.
43. E. Yeager, S. Gupta and J. A. Molla: ‘The electrochemical properties of
graphite and carbon’,Proc.Electrochem.Soc.,1983,84–85,123–157.
44. D. Wei, Y. Liu, L. Cao, H. Zhang, L. Huang, G. Yu, H. Kajiura and
Y. Li: ‘Selective electrochemical etching of single-walled carbon
nanotubes’,Adv. Funct. Mater.,2009,19, (22), 3618–3624.
45. E. Frackowiak and F. Beguin: ‘Carbon materials for the electroche-
mical storage of energy in capacitors’,Carbon,2001,39, 937–950.
46. W. H. Hart, C. H. Culberson and W. S. Smith: ‘Calcareous deposits
on metal surfaces in seawater—a critical review’,Corrosion,1984,
40, (11), 609–618.
47. N. C. Billingham: ‘Degradation and stabilization of polymers’,
Mater. Sci. Technol.,2013.doi:10.1002/9783527603978.mst0412.
48. D. Hull and T. Clyne: ‘An introduction to composite materials’;
1996, Cambridge, Cambridge University Press.
49. K. Sugimoto: ‘Corrosion protection function and breakdown mech-
anisms of passive films on stainless steels’,in‘Characterization of
corrosion products on steels’, (eds. Y. Waseda and Shigeru
Suzuki), Vol. 17; 2006, Berlin, Springer.
50. E. E. Stansbury and R. A. Buchanan: ‘Fundamentals of electroche-
mical corrosion’;2000, Novelty, OH, ASM International.
51. A. A. Pud, S. P. Rogalsky and G. S. Shapoval: ‘Reactions at the
lower potential limit in aprotic medium at a platinum cathode
revisited: their role in indirect electrochemical reductive degradation
of polymers’,J. Electroanal. Chem.,2000,480,(1–2), 1–8.
Baltzis et al. SS coupled with CNT-modified epoxy and CF composites
Plastics, Rubber and Composites 2016 VOL 45 NO 3105
Downloaded by [89.210.164.63] at 23:20 08 April 2016
