International Journal of Adhesion & Adhesives 68 (2016) 298–304

Contents lists available at ScienceDirect

International Journal of Adhesion & Adhesives

journal homepage: www.elsevier.com/locate/ijadhadh

The influence of high-temperature sulfuric acid solution ageing

on the properties of laminated vinyl-ester joints
M. Lindgren a,n, M. Wallin b, M. Kakkonen c, O. Saarela b, J. Vuorinen c
a
Outotec Research Center, P.O. Box 69, 28101 Pori, Finland
b
Department of Mechanical Engineering, Aalto University, P.O. Box 14300, FI-00076 Aalto, Finland
c
Department of Materials Science, Tampere University of Technology, P.O. Box 589, 33101 Tampere, Finland

art ic l e i nf o a b s t r a c t

Article history: A carbon fiber powder doped corrosion layer is used as an inner layer in large composite tanks to
Accepted 13 April 2016 improve their chemical and wear resistance. In joints fabricated on site, this layer is embedded into the
Available online 21 April 2016 structure. This study evaluates the lap shear strength of specimens, simulating a laminated joint in
Keywords: between the corrosion layer and the inner joint laminate. Lap-shear tests were carried out for as-
Vinyl ester fabricated and aged laminates at room temperature and at elevated temperature. Ageing was carried out
Aging for half a year in a 95 °C pressurized sulfuric acid solution. The tests showed that, after ageing, the room
Sulfuric acid temperature shear strength remained unaltered but high-temperature shear strength was lowered.
Lap-shear When the temperature increased, the failure location shifted from the interface between the doped layer
and the joint laminate to the doped layer. Thermal analysis and microscopy were employed to clarify the
reasons for the observed behavior.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction polymers physically, i.e. plasticization, and/or chemically, i.e.

hydrolysis. Water is chemically very reactive because it has a polar
Fiberglass reinforced plastic composites (FRP) are often used to molecule H–OH, which dissociates readily giving a proton and a
construct tanks, piping, scrubbers, launders and other components hydroxyl ion. A polymeric network contains numerous reactive
for use in corrosive environments, e.g. in the mining and metal- sites and is held together by hydrogen bonds and other secondary
lurgical industry [1]. FRP typically offers superior and cost effective valence forces between the adjacent polymer chains. For example,
corrosion resistance relative to other construction materials. It through hydrogen bonding, water molecules bind to epoxy resins,
comprises three different distinct phases (fibers, matrix, and fiber– forming type I and type II bound water, depending on the bond
matrix interfacial layers). These phases are affected in different strength [8]. Rupture of the interchain bonds allows the polymer
ways when water or some aggressive chemical enters the material network to expand and, as a result, bound water induces swelling
from the environment. [9]. In addition to bound water, there is some free volume in the
The glass fibers that give structural stiffness and strength for polymer (including voids). Obviously, water can occupy these free
FRP laminates are practically insensitive to moisture but various volumes and it is then labeled free water. Compared to polyester,
acids have an adverse effect on the strength properties of normal E vinyl ester is more resistant to degradation since it has fewer ester
glass fibers [2,3]. Additionally, E glass is prone to stress-corrosion groups, which are weak points for hydrolysis [13]. Consequently,
cracking [4–6]. ECR glass has been shown to degrade to a far lesser vinyl ester is favored in corrosive conditions.
extent under acidic conditions [5–7]. Consequently, ECR glass The stability of the fiber–matrix interfacial layer governs the
fibers are applied in corrosive environments either throughout the mechanical properties of a fiber-reinforced composite, especially
structure or within the resin-rich corrosion layer, protecting the under hot and wet conditions [10]. Glass fibers are highly hygro-
underlying structural layer. scopic. Therefore, water accumulates at unprotected glass fiber
The polymeric matrix is sensitive to moisture, in addition to surfaces over periods of time causing a loss of adhesion. A study of
potentially aggressive chemicals, because water interacts with glass-reinforced isophthalic polyester by the nuclear magnetic
resonance technique has demonstrated the tendency of water to
n
concentrate at the fiber–matrix interface; the concentration of
Corresponding author. Tel.: þ 358 40 829 7293; fax: þ 358 20 329 3201
E-mail addresses: mari.lindgren@outotec.com (M. Lindgren),
water there was found to be twice that of the bulk of the matrix
markus.wallin@aalto.fi (M. Wallin), markus.kakkonen@tut.fi (M. Kakkonen), [11]. Strong chemical interaction between the fibers and the
olli.saarela@aalto.fi (O. Saarela), jyrki.vuorinen@tut.fi (J. Vuorinen). matrix, e.g. through silane coating [11,12], prevents moisture

http://dx.doi.org/10.1016/j.ijadhadh.2016.04.011
0143-7496/& 2016 Elsevier Ltd. All rights reserved.
 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304 299

penetration at the interface promoting the retention of mechanical tensile elongation at yield 5%–6%. The first part of the laminate
properties. consisted of a 3 mm thick corrosion barrier layer that was laid onto
Moisture penetrates a laminate mainly through direct diffusion the mold surface, and a 7 mm thick structural layer that was added
of water molecules into the matrix. The other mechanisms that on top of the corrosion layer. Together the layers represented the
may become active are capillary flow along the fiber–matrix inner portion of a tank bottom laminate that is typically joined on
interface and transport through microcracks. At low temperatures, site. The corrosion layer consisted of one surface mat of ECR glass
moisture absorption is often found to follow Fickian-type diffusion and eight layers of Advantex chopped strand mat (Owens Corning
[15]. At higher temperatures, the complementary diffusion paths, M723A 300 g/m2). The resin of the corrosion layer was doped with
through micropores for example, gain importance and more carbon fiber powder (CFM-PYR-H1; pyrolysis recycled HTS-type
complex behavior has been observed [16–18]. A mechanically
carbon fibers, medium fiber length 200 mm, concentration 5% by
damaged laminate has been found to absorb water more quickly
weight,). As a result of the doping, the corrosion layer was black in
and the saturation moisture content has been higher compared to
color. The laminate, later referred to as the base laminate, was
its undamaged counterpart [19].
cured for 4 h at 80 °C and 1 h at 100 °C.
The exposure of a laminate to a hot and wet environment or to
After curing, the doped surface of the base laminate was
certain chemical environments leads to a noticeable deterioration
sandblasted and the joint laminate was prepared on top of the
of mechanical properties, such as tensile strength [21–23], elastic
modulus [22], and interlaminar shear strength [12,24], influencing surface in two phases. First, an infused laminate consisting of two
the integrity of the structures. In addition, changes in physical mat layers (powder bound chopped strand mat 450 g/m2), four
properties that manifest themselves as, for instance, a lowering of fabric/mat layers (Ahlstrom 5520/M100-125-40, 1374 g/m2), and
the glass transition temperature have typically been reported [20]. one mat layer (450 g/m2) was added. Secondly, a hand laminate
The samples, however, have often comprised E glass fibers and no consisting of similar reinforcements was applied. The curing and
corrosion barrier has been used. Both features differ from the post-curing of the joint laminate was similar to the curing of the
laminates aimed for aggressive chemical environments. Conse- base laminate, i.e. 4 h at 80 °C and 1 h at 100 °C.
quently, the reported deterioration rates may be overestimated The glass fractions of the laminates were measured with a
when applied to laminates designed for corrosion service. residual ash determination test where a small sample (weight
In large tanks in the mining and metallurgical industry, joints approximately 1 g) was cut from the laminate and its temperature
between substructures, such as the cylinder shell and the bottom, was raised gradually to 600 °C in an oven. After all the resin had
are partly made in the workshop and partly on site. The joint burned away the remaining ash was weighed and compared to the
between the bottom and the cylinder shell can be critical [25]. In original weight. The glass volume content of the base laminate was
this type of joint, the primary load-carrying member is normally determined to be 27 71 wt% in the corrosion layer and 42 71 wt%
the outer joint laminate and the main function of the inner joint in the structural layer. The glass volume content of the joint
laminate is to seal the joint. Nevertheless, the inner joint inevi- laminate was on average 467 1 wt%, although some differences
tably bears a certain amount of the load but its long-term prop- between the hand laminate and infused laminate were found. The
erties after ageing are not known. In this experimental study, a carbon fiber doped layer contained the largest pores. In fact, all the
laminate representing an inner joint added on top of a carbon fiber visually observable pores were seen in that layer. The joint lami-
powder doped corrosion layer was manufactured and aged in a
nate adjacent to the carbon doped layer contained numerous
sulfuric acid solution and its properties were characterized.
smaller pores, as illustrated in Fig. 1.
Half of the test laminate was aged in four separate pieces. Before
ageing, the edges of the laminates were sealed with vinyl ester resin.
2. Materials and methods
The laminates were weighed before and after immersion. The lami-
nates were aged for half a year completely immersed in a 50 g/l
Several similar test laminate samples were prepared and half of
them were aged for half a year in a high-temperature sulfuric acid sulfuric acid solution (about 5% H2SO4), containing 0.5 g/l ferric sul-
solution. Lap shear tests were carried out for the as-fabricated and fate. Ferric sulfate was added to act as a corrosion inhibitor for the
aged laminate samples at room temperature and at elevated stainless steel parts of the set-up. The nominal temperature of the
temperature. Thermal analysis was also conducted for the as- solution was 95 °C and the nominal pressure 15 bar. The temperature
fabricated and aged samples and selected fracture surfaces were and pressure were monitored continuously. The measured tem-
characterized with a scanning electron microscope. Table 1 sum- perature varied between 93 and 95 °C and the measured pressure
marizes the experimental investigations. between 13 and 15 bar. The H2SO4 concentration was measured by
titration before and after the experiment and was found to remain
2.1. Test laminate unchanged. The iron concentration was measured by inductively
couple plasma (ICP) and was found to increase somewhat due to
The test laminate was prepared from Derakane 441-400 vinyl corrosion of the stainless steel parts.
ester resin and ECR glass fibers using a glass plate as a mold.
According to the datasheet of the resin [26] glass transition tem-
perature Tg of the resin is 125 °C, tensile strength 90 MPa and

Table 1
Summary of the experimental investigations.

Laminate Room tempera- Elevated tem- Thermal Fracture sur-

ture (RT) lap perature lap analysis face
shear tests shear tests examination

As-fabricated X 105 °C X RT, 105 °C

Aged X 80 °C, 95 °C and X RT, 80 °C, 95 °C
110 °C and 110 °C Fig. 1. Optical micrograph of the test laminate around the carbon fiber powder
doped (black) layer.
300 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304

2.2. Thermal analysis

Samples from various locations in the through-thickness

direction of as-fabricated and aged laminates were prepared for
dynamic thermo-mechanical analysis (DMTA) and differential
scanning calorimetry (DSC). Two to three samples as a function of
their relative position in the thickness direction of the laminate
from each layer were measured. DSC equipment was Netzsch DSC
204 F1 and the sample weight was 12–17 mg. Two consecutive
DSC measurements were carried out with a heating rate of 10 K/
min. The experience from the multiple measurements of this type
of samples has given a repeatability of 1–2 °C.
DMTA studies were done with Perkin Elmer Diamond DMA
using the single cantilever loading mode. The loading amplitude
was 40 μm, loading frequency 1 Hz, and the heating rate was 3 K/
min. The sample size was about 4x2x50 mm3 and samples were
taken as a function of their relative position in the thickness
direction of the laminate. The samples were cut using a wheel
grinder and cooling fluid to avoid the heating of the sample.

2.3. Mechanical testing

The applied lap shear test method was modified from ASTM
D5868 [27]. The test specimen was of “thick adherend” type. The
specimen width was 25 mm.
After post-curing, the thickness of the test laminate varied
between 23 and 26 mm and was slightly bent. In addition, the
base laminate was thinner than the joint laminate. Consequently,
the bond line being studied was not in the middle of the laminate.
For these reasons it was decided to make the specimens thinner by
milling them to a thickness of 12 mm. Grooves for the lap joint
were machined using a table saw with a 0.9 mm thick diamond
blade. An overlap length of 12 mm was used.
The grooves extending to the interface between the carbon
fiber powder doped corrosion layer and the joint laminate are
presented in Fig. 2a. An image in Fig. 2b illustrates that the lap
shear joint was not in the middle of the grips in testing due to
other objectives of the tests.
Testing was carried out using a Dartec 100 kN load frame with
an MTS FlexTest 40 controller. The test speed was 0.5 mm/min.
The specimens tested at elevated temperature were first heated in
an oven to the test temperature. The tests were conducted in an
environmental chamber. The temperature of the specimens during Fig. 2. a) Location of the grooves with respect to the studied joint in between the
the tests was monitored with thermocouples and with a hand- joint laminate and carbon fiber powder doped corrosion layer of the base laminate;
held thermometer. The elevated test temperatures were 105 °C for b) Sample attached to the grips: distances of the lap shear joint center from the test
machine grips 25 mm and 50 mm.
the as-fabricated specimens and 80, 95, and 110 °C for the aged
specimens.
reinforcements [28]. The use of different reinforcements and the
After testing, the fracture mode of each specimen was observed
applied pressure during immersion could explain why higher mass
visually and photographed. One sample from each test set was
gain was measured in the sulfuric acid solution than in
taken for further analysis. The fracture surfaces were characterized
with a scanning electron microscope (Jeol 6490LV). distilled water.
The glass transition temperature values (Tg) of the test samples
measured with DSC after post-curing and after ageing are illu-
3. Results strated in Fig. 3. After post-curing, the laminate had an average Tg
of 104 °C. The Tg was a couple of degrees lower in the carbon fiber
3.1. Characterization of the aged laminates doped layer. Ageing significantly increased the Tg value as the aged
laminates had an average Tg of 115 °C. After ageing, no systematic
The mass of the laminates increased by 1.470.03% during differences between the Tg values in difference layers were noted.
immersion in the high-temperature pressurized sulfuric acid In the as-fabricated laminate, the storage modulus of the car-
solution. This is more than the saturation moisture that Yu et al. bon fiber powder doped layer was slightly smaller than that of the
[18] measured for vinyl ester–carbon fiber composites (1.0%) in other laminate layers in the whole test temperature range. After
distilled water at 95 °C. Overall, higher water activity enhances ageing, the difference between the layers disappeared. Fig. 4 pre-
water absorption and, consequently, higher mass gain would be sents the storage modulus of the carbon fiber doped layer as a
expected in distilled water than in sulfuric acid solution. However, function of test temperature for the as-fabricated and aged sam-
it has been observed that the use of glass fiber reinforcements ples. Ageing increased the storage modulus and the increase was
facilitates water absorption compared to the use of carbon fiber slightly higher in the upper part of the test temperature range.
 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304 301

Fig. 5. Maximum values of tanδ in as-fabricated and aged samples.

Fig. 3. Tg values of the first heating of as-fabricated and aged laminates in various
layers measured with DSC. The x-coordinate describes the relative positions of the
measurement points. maximum value of tanδ. That was ascribed to the deterioration of the
matrix and the interfaces [29].

3.2. Lap shear test results

Table 2 summarizes the shear strength values obtained in the

lap shear tests. Fig. 6 displays the obtained values as a function of
temperature. When the test temperature increased, the shear
stress decreased, and the rate of reduction increased as the test
temperature approached the Tg value of the material.
According to Fig. 6, the shear strength of as-fabricated lami-
nates at room temperature was about 14 MPa. The obtained values
were somewhat lower than the typical values that could be
expected, 20 MPa. The reasons for this could be the asymmetry
effects in the testing and the porosity in the samples.
After ageing, the shear strength at room temperature was at the
same level. The measured difference of 2% between the as-fabricated
and aged samples is well within the experimental scatter. The mea-
sured shear strength values are relatively high. For comparison, e.g. for
Fig. 4. Storage modulus of samples taken from two depths of the carbon fiber
powder doped layer in the as-fabricated and aged samples. the interlaminar shear strength of laminates reinforced with chopped
strand mats, the standard EN 13121 [27] stipulates a value of 7 MPa,
All the layers in the laminate showed an increase in Tg after which is half of the value measured here.
ageing but that was associated with an increase in the storage On the other hand, in the high-temperature tests, the measured
modulus only in the carbon fiber doped layer. The increase was shear strength values of as-fabricated samples were lower and the
more noticeable at low temperatures below Tg. In the other layers, ageing effect was seen. The ageing lowered the shear strength
the storage modulus remained unaltered. This suggests that the value by about 30% at 95 °C. Regardless of the reduction, the shear
strength was still at a relatively good level when compared to the
mechanisms behind the Tg increase may be different in the carbon
interlaminar shear strength value referenced above.
fiber powder doped layer. The increase in Tg may be a result of the
anti-plasticization effect i.e. the loss of low molar mass solvents or
3.3. Fracture location
of enhanced cross-linking.
In the anti-plasticization effect, the Tg value increases without a
The test temperature influenced not only the shear strength,
simultaneous increase in the storage modulus. This seemed be the
but also the failure location. When the as-fabricated samples were
case in the laminate layers without doping. Leaching of the
tested at room temperature, the fracture occurred in every test
monomer residues could be responsible for this effect. In contrast,
specimen at the interface of the carbon fiber powder doped layer
in the carbon fiber doped layer, the increase in the Tg value was
and the joint laminate. In contrast, in the high-temperature tests,
associated with the increase in the storage modulus, indicating
the fracture took place within the carbon fiber doped layer (Fig. 7).
extended cross-linking. When the aged samples were tested at room temperature, the
The tanδ of all measured samples showed only one peak, indicat- fracture occurred predominantly at the interface of the carbon
ing that no separation into differently behaving regions had occurred. doped layer and the joint laminate (Fig. 7). When the test tem-
The maximum values of tanδ are given in Fig. 5 for the as-fabricated perature increased, the failure location shifted more and more to
and aged samples. It can be seen that in the undoped layers the the doped layer. The results indicate that, at high-temperature, the
maximum of the tanδ was lower after ageing. In the carbon fiber interlaminar shear strength of the doped layer was less than the
powder doped layer, however, it remained unaltered. Boinard et al. adhesion strength between the doped layer and the joint laminate.
[14] observed a similar behavior in vinyl ester laminates aged in dis- In other words, the carbon fiber powder doped layer itself was the
tilled water. They associated the increase in Tg and decrease in tanδ weakest at high-temperature.
peak with the anti-plasticization effect. In contrast, glass/polyester When the as-fabricated laminate was tested at room tem-
laminates immersed in seawater exhibited a significant increase in the perature, the fracture surface exhibited regions of brittle matrix
302 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304

Table 2
Summary of the lap shear test results for the as-fabricated and aged samples.

As-fabricated Aged

RT T ¼ 105 °C RT T ¼ 80 °C T ¼ 95 °C T ¼ 110 °C

Average shear strength [MPa] 14 11 14 13 11 5.9

Standard deviation [MPa] 1.9 0.5 1.1 1.0 0.8 0.7
Coefficient of variation [%] 13 4.7 7.7 7.8 6.9 12
Failure mode Shearing Shearing Shearing Shearing Shearing Shearing

Fig. 6. Measured average shear strength values as a function of test temperature

for the as-fabricated and aged samples (standard deviations of the measurements
indicated by the bars).

Fig. 7. Failure locations in the aged laminates as a function of test temperature.

Fig. 8. Scanning electron image of the fracture surface of as-fabricated sample

fracture (Fig. 8a) and regions of debonded fibers (Fig. 8b). The tested at room temperature showing typical fracture surface features: a) brittle
appearance of these regions was similar throughout the fracture fracture of the matrix and b) debonding of fibers.
surface. A brittle fracture of the matrix was observed between the
debonded fibers. Small pores were seen in the fracture surface. Fracture surfaces of the aged laminates tested at elevated
After testing at room temperature, the fracture surface of the temperature exhibited similar features to the surfaces of the as-
aged samples resembled that of the as-fabricated samples after fabricated laminates tested at elevated temperature. Smaller areas
room temperature testing. However, it seemed that large bundles of brittle fracture were present and lots of debonded individual
of fibers had debonded from the matrix rather than individual fibers were seen in Fig. 12.
fibers (Fig. 9). A brittle fracture of the matrix had occurred
between the debonded areas.
After testing of the as-fabricated samples at elevated tem- 4. Discussion
perature, the fracture surfaces contained a significant amount of a
small dust-like substance, presumably representing the carbon In large FRP tanks serving in a high-temperature environment
fiber powder doped matrix (Fig. 10). The regions where the matrix where aggressive chemicals and wear-inducing solids are present,
had failed in a brittle manner were visible but their size seemed to a carbon fiber powder doped corrosion layer may be used to
be smaller than after room temperature testing. Fiber debonding improve the wear resistance and to slow down diffusion through
was the dominant failure mechanism: the fiber bundles were the corrosion layer. When substructures of the tank, e.g. the
detached and broken and single fibers were randomly oriented, as cylinder and bottom, are joined on site, the inner joint laminate is
presented in Fig. 11. laid onto the surface of this layer. In the room temperature tests of
 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304 303

Fig. 9. Fracture surface of the aged sample after testing at room temperature Fig. 12. Typical fracture surface features of aged samples tested at 95 °C; small
showing fiber bundles detached from the matrix. areas of brittle fracture present and lots of debonded individual fibers visible.

In high-temperature testing, the performance of the joint

naturally deteriorated as the Tg temperature of the material was
approached. This behavior was observed for both as-fabricated
and aged samples. Furthermore, the failure location shifted more
and more from the interface to the carbon fiber powder doped
layer itself with increasing test temperature. In other words, in
high-temperature tests, the interlaminar shear strength of the
doped layer was lower than its adhesion strength to the adjoining
joint laminate. There are several possible reasons for this behavior.
First, it may be ascribed to the lower glass content of the layer. At
higher temperature, when the matrix softens, the load bearing by
the fibers becomes more and more important and the lower glass
content translates into lower strength. Secondly, the thermal
mismatch stresses between the matrix, the glass fibers, and the
carbon fiber particles may weaken the interfaces and contribute to
the lower shear strength. The scanning electron images of the
Fig. 10. Fracture surface of the as-fabricated sample tested at 105 °C showing dust- failure surface indicated that the strength of the interfaces played
like substance presumably representing the carbon fiber powder doped matrix. a major role as, after high-temperature testing, detached bundles
of fibers were seen rather than the debonding of individual fibers
and brittle fracture of the matrix. The high porosity of the doped
layer in the form of large pores may also affect the behavior.
Ageing seemed to influence the doped layer more than the rest
of the laminate. In every part of the laminate, the Tg value of the
material increased because of the post-curing during ageing.
However, only in the doped layer was this associated with an
increase in the storage modulus, both at low and high tempera-
tures, and a slight increase in tanδ values albeit within the
experimental scatter. The room temperature shear strength was
not influenced by ageing as it was governed by the adhesion
strength between the doped layer and the laminate. In contrast,
high-temperature shear strength was determined by the proper-
ties of the doped layer and was lowered by ageing.
The most probable cause for this behavior is the presence of
additional interfaces between the carbon fiber particles and the
matrix. The presence of large pores and additional interfaces may
be conducive to more severe ageing of the doped layer. The
Fig. 11. Fracture surface of the as-fabricated sample after testing at 105 °C; fiber interfaces are known to be susceptible to moisture and other
debonding was the dominant failure mechanism: the fiber bundles were detached chemicals weakening the bonding. The deterioration of the inter-
and broken and single fibers were randomly oriented. faces inside the doped layer was, however, so small that the effect
was detected only at high temperature, i.e. when the thermal
this study, the interface between such a corrosion layer and the mismatch stresses contributed significantly to the local stress
joint laminate showed high adhesion strength. Furthermore, the state, leading to lower shear strength.
room temperature adhesion strength was not adversely influenced The vinyl ester laminate itself seemed to be unaffected by
by half-year ageing in a high-temperature pressurized sulfuric acid ageing. The tanδ values obtained from the DMTA analysis, which
solution. The adhesion strength remained unaltered and the fail- are reported to increase with material degradation, were even
ure location in the interface of the carbon fiber powder doped lower for the aged laminate. This is in line with the known
layer and the joint laminate remained the same. excellent chemical resistance of the resin.
304 M. Lindgren et al. / International Journal of Adhesion & Adhesives 68 (2016) 298–304

5. Conclusions [8] Zhou J, Lucas JP. Hygrothermal effects of epoxy resin. Part I the nature of water
in epoxy. Polymer 1999;40:5505–12.
[9] Xiao GZ, Shanahan MER. Swelling of DGEBA/DDA epoxy resin during hygro-
Based on the experimental findings, the following conclusions thermal ageing. Polymer 1998;39(14):3253–60.
can be made: [10] DiBenetto AT. Tailoring of interfaces in glass fiber reinforced polymer com-
posites: a review. Mater Sci Eng 2001;A302:74–82.
[11] Kotsikos G, Gibson AG, Mawella J. Assessment of moisture absortion in marine
– The test temperature had a significant influence on the shear GRP laminats with aid of nuclear magnetic resonance imaging. Plast Rubbers
strength of the joint between a carbon powder doped corrosion Compos 2007;36(9):413–8.
layer and a structural laminate added onto the surface of [12] Pavlidou S, Papaspyrides CS. The effect of hygrothermal history on water
sorption and interlaminar shear strength of glass/polyester composites with
the layer. different interfacial strength. Compos Part A: Appl Sci Manuf 2003;34:1117–24.
– At room temperature, the interface between the carbon powder [13] Olmos D, López-Morón R, Gonzales-Benito J. The nature of the glass fibre
doped corrosion layer and the structural laminate failed adhe- surface and its effect in the water absorption of glass fibre/epoxy compsites.
The use of fluorescence to obtain information at the interface. Compos Sci
sively in the main but, as the test temperature increased, the
Technol 2006;66:2758–68.
cohesive fracture within the doped layer became predominant. [14] Boinard E, Pethrick RA, Dalzel-Job J, Macfarlane CJ. Influence of resin chemistry
– Similar temperature dependence of the failure location was on water uptake and environmental ageing in glass fibre reinforced compo-
sites – polyester and vinyl ester resin laminates. J Mater Sci 2000;35:1931–7.
observed for the as-fabricated samples and for samples aged for
[15] Eslami S, Honarbakhsh-Raouf A, Eslami S. Effects of moisture absortion on
half a year at 95 °C in the pressurized sulfuric acid solution. It degradation of E-glass fiber reinforced vinyl ester composite pipes and mod-
was also observed that the shear strength of the doped layer at eling of transient moisture diffusion using finite element analysis. Corros Sci
high temperature decreased due to ageing. Microscopically, the 2015;90:168–75.
[16] Berketis K, Tzetzis D. Long-term water immersion ageing characteristics of
fracture surfaces of the as-fabricated laminates contained more GFRP composites. J Mater Sci 2009;44:3578–88.
individual fibers with random orientation, whereas in aged [17] Verghese KNE, Hayes MD, Garcia K, Carrier C, Wood J, Riffle JR, et al. Influence
laminates, fibers remained in bundles that were debonded from of matrix chemistry on the short term hydrothermal aging of vinyl ester
matrix and composites under both isothermal and thermal spiking conditions.
the matrix. J Compos Mater 1999;33:1918.
– Ageing seemed to have a more severe effect on the doped layer [18] Yu Y, Li P, Sui G, Yang X, Liu H. Effects of hygrothermal aging on the thermal-
than on the vinyl ester laminate. mechanical properties of vinyl ester resin and its pultruded carbon fiber
composites. Polym Compos 2009:1458–64.
– Regardless of ageing, the shear strength of the studied joint [19] Suri C, Perreux D. The effects of mechanical damage in a glass fibre/epoxy
remained at a high level compared to the requirement provided composite on the absorption rate. Compos Eng 1995;5(4):415–24.
by the relevant standard. [20] Mezghani K. Long term environmental effects on physical properties of
vinylester composite pipes. Polym Test 2012;31:76–82.
[21] Banna MH, Shirokoff J, Molgaard J. Effects of two aqueous acid solutions on
glass reinforced vinyl ester resin tubes. Can Metall Q 2012;51(1):91–100.
[22] Alawsi G, Aldajah S, Rahmaan SA. Impact of humidity on the durability of
References
E-glass/polymer composites. Mater Des 2009;30:2506–12.
[23] Kim H-Y, Park Y-H, You Y-J, Moon C-K. Short-term durability test for GFRP rods
[1] Graham J, Johnson T, Kelley D. Optimum material selection for industrial under various environmental conditions. Compos Struct 2008;83:37–47.
hydrometallurgical equipment. In: Proceedings of the 51st conference of [24] Ray BC. Temperature effect during humid ageing on interfaces of glass and
metallurgists on corrosion: recent advances & sustainability. Niagara Falls, ON, carbon fibers reinforced epoxy composites. J Colloids Interface Sci
Canada; 2012. p. 93–104. 2006;298:111–7.
[2] Das B, Tucker BD, Watson JC. Acid corrosion analysis of fibre glass. J Mater Sci [25] Lindgren M, Bergman G, Kakkonen M, Lehtonen M, Jokinen J, Wallin M, et al.
1991;26:6606–12. Failure analysis of a leaching reactor made of glass-fiber reinforced plastic.
[3] Chandler HD, Jones RL. Strength loss in “E” glass fibres treated in string Eng Fail Anal 2016;60:117–36.
solutions of mineral acids. J Mater Sci 1984;19:3849–54. [26] Derakane 441-400 epoxy vinyl ester resin. Technical datasheet. Ashland per-
[4] Caddock BD, Evans KE, Hull D. Stress-corrosion failure envelopes for E-glass formance materials, 2015.
fibre bundles. J Mater Sci 1990;25:2498–502. [27] Standard ASTM D5868-01(2014) Standard test method for lap shear adhesion
[5] Lewis G, Bedder SW, Reid I. Stress corrosion cracking of glass fibres in acidic for fiber reinforced plastic (FRP) bonding.
environments. J Mater Sci Lett 1984;3:728–32. [28] Standard EN 13121-3:2008. GRP tanks and vessels for use above ground.
[6] Sapalidis SN, Hogg PJ, Youd SJ. High temperature acidic stress corrosion of Design and workmanship.
glass fibre composites. J Mater Sci 1997;32:309–16. [29] Gu H. Dynamic mechanical analysis of the seawater treated glass/polyester
[7] Aveston J, Sillwood JM. Long-term strength of glass-reinforced plastics in composites. Mater Des 2009;30:2774–7.
dilute sulfuric acid. J Mater Sci 1982;17:3491–8.