Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Composite Structures 147 (2016) 143–154

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

FRP–brick masonry bond degradation under hygrothermal conditions


Hamid Maljaee ⇑, Bahman Ghiassi, Paulo B. Lourenço, Daniel V. Oliveira
ISISE, University of Minho, Department of Civil Engineering, Azurém, 4800-058 Guimarães, Portugal

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

Article history: Externally bonded reinforcement (EBR) of masonry structures with Fiber Reinforced Polymers (FRPs) has
Received 20 November 2015 received extensive attention during the last years. Despite the vast literature on mechanics and short-
Revised 10 March 2016 term performance, the durability and long-term performance of these systems still remain insufficiently
Accepted 24 March 2016
studied. Structures are subjected to harsh environments such as coupled temperature and moisture vari-
Available online 25 March 2016
ations (hygrothermal conditions) during their service life. A clear understanding on the performance of
the strengthening system under these conditions is critical at the design stage.
Keywords:
This paper presents an experimental investigation on the effect of long-term hygrothermal conditions
FRP
Masonry
on FRP-strengthened masonry components. The focus is on the interfacial bond behavior and materials
Bond mechanical properties. The effect of mechanical surface treatment on the bond durability is also
Durability investigated.
Environmental degradation Ó 2016 Elsevier Ltd. All rights reserved.
Mechanical testing

1. Introduction The combined effect of moisture and temperature, also called as


hygrothermal conditions, is thus the main focus of this study.
Externally bonded reinforcement (EBR) of masonry structures Moisture, after diffusion in epoxy resin and FRP-substrate inter-
with Fiber Reinforced Polymers (FRPs) is found to be a promising face, can lead to plasticization, degradation of mechanical proper-
technique. Several studies have focused on structural behavior, ties and loss of integrity at the bond level through hydrolytic
design parameters and failure modes. The effectiveness of EBR sys- breakdown of the interface between matrix and FRP [17,18]. Expo-
tems is known to be highly dependent on the bond performance sure to elevated temperatures, near or above the glass transition
between the strengthening material and the substrate [1–4]. The temperature, leads to significant reduction of mechanical proper-
bond performance can be seriously affected under environmental ties. While short-term exposure can be reversible, long-term expo-
conditions, which leads to premature debonding or change of sure to high temperatures may induce irreversible changes in their
expected failure modes. In addition to the short-term performance, properties [19]. On the other hand, the freeze–thaw cycling can
understanding the long-term behavior and active degradation lead to degradation of mechanical properties and of the bond per-
mechanisms under different environmental conditions is thus crit- formance [20,21]. Temperature cycles below glass transition tem-
ical for service life predictions [5]. perature (Tg) can also affect the bond performance due to
The dearth of available information on durability and long-term thermal fatigue [18]. Internal stresses due to thermal incompatibil-
performance of externally bonded FRP systems applied to masonry ity (difference of thermal expansion coefficient in materials) may
walls demands a broader investigation. While several investiga- affect the mechanical properties, structural integrity and conse-
tions can be found in the literature focused on durability of FRP- quently exacerbate the bond degradation [22,23]. The moisture
concrete systems [6–12], the available information on FRP- induced degradation can be accelerated when coupled with tem-
strengthened masonry elements is still rare [13–16]. While most perature [6] as the moisture absorption rate increases with incre-
of the available studies focus on the effect of an isolated degrading ment of temperature [17,24]. On the other hand, moisture can
agent on the bond behavior in FRP-strengthened systems, less affect thermal properties of materials and thus change the degra-
attention has been given to the coupled effects (as happens in real dation rates.
condition) even in case of FRP-strengthened concrete elements. Ghiassi et al. [15] investigated the effect of 225 cycles of
hygrothermal exposure (temperature cycles ranging from +10 °C
to +50 °C and 90% R.H.) on the bond and material properties of
⇑ Corresponding author.
GFRP-strengthened extruded solid clay bricks. Besides a significant
E-mail addresses: h.maljaee.civil@gmail.com (H. Maljaee), bahmanghiassi@civil.
uminho.pt (B. Ghiassi), pbl@civil.uminho.pt (P.B. Lourenço), danvco@civil.uminho. reduction of bond strength, epoxy and GFRP mechanical properties,
pt (D.V. Oliveira). a clear degradation trend was not observed and the need for exten-

http://dx.doi.org/10.1016/j.compstruct.2016.03.037
0263-8223/Ó 2016 Elsevier Ltd. All rights reserved.
144 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

sion of the exposure period was reported. Extensive FRP delamina- 40 150 10
tion was also reported in the specimens after exposure. The
Brick
observed delamination was reported to be due to thermal incom-

50
patibility between GFRP and masonry substrate. The failure mode GFRP sheet
also changed from cohesive to adhesive failure.
In a recent study, the effect of hygrothermal conditions (con- 450
stant temperature at +40 °C and 90% R.H.) on FRP-strengthened
natural stones was reported [13]. The authors reported significant Aluminum plate Unbonded area Bonded area
degradation of stones’ mechanical properties after 25 weeks of (a)
exposure. The failure mode was also changed from cohesive to
adhesive failure. It was also reported that specimens prepared with 200
Leccese stones showed higher degradation of bond in comparison 180
to the specimens prepared with Neapolitan tuff. This difference Brick
was attributed to the differences in porosity of the substrates.

100
GFRP sheet

70
This paper extends the previous research of Ghiassi et al. [15] to
longer exposure periods. GFRP-strengthened bricks are exposed to
temperature cycles ranging from +10 °C to +50 °C and relative
humidity of 90% for 960 cycles (each cycle lasts for 6 h). The influ- (b)
ence of mechanical surface treatment on the durability of bond, not
addressed before, is also assessed here. The effect of accelerated Fig. 1. Geometry of the specimens prepared for: (a) single-lap shear tests; (b) pull-
off tests (dimensions are in mm).
hygrothermal exposure on the material properties and bond
behavior are deeply investigated. The bond degradation is investi-
gated through performing pull-off and single-lap shear debonding surfaces (denoted as ORG-specimens). In the second group, the
tests. The mechanical tests are combined with thermal analysis of bricks’ surfaces were grinded for about 5 mm with a mechanical
primer, epoxy (bulk adhesive) and epoxy taken from interface to saw before application of the GFRP (denoted as GR-specimens).
deeply understand the degradation mechanisms. A decay model In both specimen types, the GFRP sheets were applied over a
is also proposed for each property to be used for long-term predic- 150  50 mm2 area, leaving a 40 mm unbonded length at the
tions in hygrothermal conditions. Finally the design parameters loaded end as shown in Fig. 1a.
presented in CNR-DT 200 are calibrated based on the experimental The pull-off test specimens were prepared without application
results. It should be noted that, throughout this study, the word of any mechanical surface treatment. The GFRP sheets were
‘‘delamination” refers to the separation of FRP sheet from the applied over a 180  70 mm2 area in these specimens as shown
masonry substrate due to the exposure before the test execution, in Fig. 1b.
and ‘‘debonding” describes the bond failure phenomenon in FRP- A total number of 35 specimens (of each material) were also
strengthened specimens after bond mechanical tests. prepared, according to relevant codes, for materials’ mechanical
characterization during the exposure. The specimens included
40 mm cubic bricks, dog-bone shape epoxy resin and primer, and
2. Experimental program
GFRP coupons with the geometrical details shown in Fig. 2.
The experimental program consisted of exposing FRP-
2.2. Material properties
strengthened bricks and material samples to hygrothermal expo-
sure conditions in a climatic chamber. The specimens were taken
Mechanical properties of materials were experimentally charac-
from the climatic chamber after different periods of exposure to
terized according to relevant test standards. The tests included
investigate possible changes in their mechanical performance.
compressive tests on brick specimens, tensile tests on dog-bone
The main focus was on the changes of FRP–masonry interfacial
shape epoxy resin and primer specimens, and tensile tests on GFRP
bond and materials mechanical properties.
coupons. Five specimens were tested for each material. Table 1
presents the mean value and coefficient of variation (CoV) of the
2.1. Materials and specimens’ preparation experimental results. The obtained results were also used as

The specimens consisted of extruded solid clay bricks with


dimensions of 200  100  50 mm3 strengthened with Glass Fiber
Reinforced Polymer (GFRP) following the wet layup procedure. A 190
40
unidirectional E-glass fiber and a two-part epoxy resin were used 40
10

t=4
20

to prepare the composite material. The specimens were prepared


40

with specific geometrical details for single-lap shear bond tests 90


(a total of 70 specimens) and pull-off tests (a total of 28 speci-
(a) (b)
mens), as presented in Fig. 1. The bricks were initially cleaned
and dried in an oven at 100 °C for 24 h. After drying, a primer layer
200
was applied on the bricks’ surfaces. A layer of epoxy resin was then
100
applied followed by impregnation of the glass fibers with epoxy
resin and then application to the bricks’ surfaces. A slight pressure 35 35
15

was finally applied on the GFRP surface with a roller to remove any
50 50
air voids at the interface.
Two groups of specimens were prepared for single-lap shear
(c)
bond tests to investigate the effect of surface mechanical prepara-
tion on the bond durability. The first group was prepared without Fig. 2. Geometry of the specimens prepared for material characterization: (a) brick
application of any mechanical surface treatment on the bricks’ cubes; (b) epoxy resin and primer; (c) GFRP coupon (dimensions are in mm).
H. Maljaee et al. / Composite Structures 147 (2016) 143–154 145

Table 1
Material properties (CoV is given inside brackets).

Compressive strength (MPa) Tensile strength (MPa) Elastic modulus (GPa) Peak strain (%) Tg* (°C)
Masonry brick 15.38 – – – –
(8%)
Epoxy resin – 58.84 2.86 2.83 Tg1: 54.2
(4%) (8%) Tg2: 66.1
Primer – 42.04 2.85 1.98 Tg1: 56.3
(18%) (10%) Tg2: 61.7
GFRP coupons – 1185.14 58.84 2 –
(8%) (5%)
*
Tg1 and Tg2 are obtained from the first and second heating scans, respectively.

reference for investigating the effect of exposure conditions on similar elastic modulus due to the linear behavior of GFRPs until
material properties. tensile rupture.
Compressive tests were performed on brick specimens accord- Differential Scanning Calorimetry (DSC) tests were also per-
ing to ASTM C67 [25] and EN 772-1 [26]. The specimens were dried formed on epoxy and primer specimens (after 100 days of curing
in an oven for 24 h at 100 °C before performing the tests. The tests and just before starting the hygrothermal exposure) in order to
were performed using a Lloyds testing machine under force control quantify the glass transition temperature, Tg, and its changes dur-
at a rate of 150 N/min, in flatwise direction. In each test, a pair of ing the environmental exposure. Three samples from each material
friction-reducing Teflon sheets with a layer of oil in between was (10–14 mg) were placed in the measuring pans to be tested in each
placed between the specimen and the compression plates to exposure period. The pans were then heated between 20 °C and
reduce the friction and ensure a uniform distribution of stresses 200 °C at a constant heating rate of 10 °C/min under nitrogen
at the center [27], see Fig. 3a. The compressive strength of brick atmosphere. Two heating scans were performed on epoxy resin
was obtained as the maximum experimental compressive force specimens (taken from dog-bone shaped specimens) by cooling
per unit area. the samples with same rate after the initial heating and then heat-
Tensile tests on epoxy resin and primer were performed accord- ing again for a second time. The first heating scan represents the
ing to standard ASTM D638 [28]. The specimens were cured in lab- thermal history of the sample including aging and processing
oratory conditions (T = 20 °C and R.H = 60%) for 100 days before [32,33]. In the second heat scan, the thermal history is eliminated
performing the tests. The tests were conducted using a Lloyds test- and thus the inherent properties of materials is captured [34]. Tg is
ing machine at a displacement rate of 2 mm/min, see Fig. 3b. The determined as a midpoint temperature which is the point on the
specimens were instrumented with a clip gauge at the middle to thermal curve corresponding to 1/2 of the heat flow difference
monitor the deformation during the test. The elastic modulus between the extrapolated onset and extrapolated end [35]. The
was obtained as the initial slope of the experimental stress–strain Tg obtained from the first and second heating scans was 54.2 °C
curve. and 66.1 °C for reference epoxy resin samples, and was 56.3 °C
Tensile tests on GFRP coupons were performed following the and 61.7 °C for reference primer samples, see Table 1.
ASTM D3039 [29] and ISO 527-1 [30]. The specimens were cured DSC tests were also performed on epoxy samples taken from the
in laboratory conditions (T = 20 °C and R.H = 60%) for 100 days interfacial zone between FRP and brick after performing the
before performing the tests. The tests were conducted with an debonding tests. The aim was to investigate the differences in
Instron testing machine at a displacement rate of 2 mm/min, see the curing condition and degradation between bulk epoxy and
Fig. 3c. Tensile stresses are determined based on the equivalent interfacial epoxy. These tests were performed with only one heat-
thickness concept according to ACI 440.3R [31]. This is a usual pro- ing scan. The Tg of the reference specimens taken from the inter-
cedure when the composite material is prepared following the wet face was, 55.3 °C, very close to the tensile test specimens.
layup procedure [15,21]. The elastic modulus was obtained from
the stress–strain curves as the slope of the chord between 20% 2.3. Bond characterization tests
and 60% of the maximum tensile stress (according to the code pro-
posal). The initial slope of the stress–strain curves also provides a Bond characterization tests included pull-off tests (only on
ORG-specimens) and single-lap shear debonding tests (on both
ORG- and GR-specimens).
For performing the pull-off tests, two 50 mm diameter partial
cores were initially drilled on each specimen with an approximate
depth of 5 mm inside the brick, see Fig. 4a. Aluminum disks were
then glued on the GFRP surface using a high-strength epoxy paste
adhesive. The specimens were firmly clamed to a steel frame, as
shown in Fig. 4a, to avoid any movements during the tests. A
closed-loop servo-controlled testing machine was used to apply
tensile load to the disks with a displacement rate of 0.2 mm/min.
The load was monotonically applied until debonding of the disk
from the specimens.
For single-lap shear bond tests, the specimens were placed on a
rigid supporting steel frame and firmly clamped to it in order to
avoid any misalignment during load application, see Fig. 4b. A
closed-loop servo-controlled testing machine was again used for
Fig. 3. Schematic of test setups for material characterization: (a) compressive test
on bricks; (b) tensile test on epoxy resin and primer; (c) tensile test on GFRP
monotonically application of tensile displacements at a rate of
coupons. 0.3 mm/min. The relative slip between GFRP sheet and masonry
146 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

Fig. 4. Schematic of test setups for bond characterization: (a) pull-off test; (b) single-lap shear test.

brick was monitored by means of four LVDTs placed on different due to the complexity of controlling the relative humidity at low
locations along the bonded length. Two LVDTs were mounted at temperatures.
the loaded end, one at the middle of the bonded length, and Two ORG-specimens (from single-lap shear specimens), two
another at the free end, as shown in Fig. 4b. bricks, two epoxy resin and two primer specimens were instru-
Four pull-off tests (on ORG-specimens) and five single-lap shear mented with strain gauges, according to the details shown in
tests (on ORG- and GR-specimens) were performed on specimens Fig. 6, to monitor the deformation of the specimens during the
before starting the hygrothermal exposure tests (untreated exposure period. The strain gauges were carefully sealed with pro-
condition). tective layers to avoid impregnation of relative humidity and thus
affecting the obtained results. These specimens were kept inside
2.4. Hygrothermal exposure the climatic chamber until the end of exposure period and the
strains were monitored and recorded.
Specimens were exposed to hygrothermal conditions in a cli- Five specimens of each kind were taken from the climatic cham-
matic chamber after 100 days of curing in laboratory conditions. ber every month (around 120 cycles and always when the temper-
The exposure included 6-h temperature cycles ranging from ature was stabilized at 10 °C) to perform post-exposure tests. The
+10 °C to +50 °C with constant relative humidity of 90%, see specimens were weighed and visually inspected, and stored in a
Fig. 5. The relative humidity inside the climatic chamber drops to controlled environment room (20 °C temperature and 60% R.H.)
60% when the temperature reaches +10 °C and then goes back to for 7 days before testing. The post-exposure tests included
90% after a short period. This is a usual situation and it happens materials’ mechanical characterization, DSC tests and bond

160 100
Relative Humidity, RH [%]

140
80
Temperature [° C]

120
100 60
RH (real condition)
80
RH (planned condition)
60 Temperature 40

40
20
20
0 0
0 6 12 18 24
Time [hrs]

Fig. 5. Hygrothermal exposure tests condition.


H. Maljaee et al. / Composite Structures 147 (2016) 143–154 147

Fiber direction
sg.3 sg.2 sg.1 Brick
L
T

GFRP L
coupons sg.5 sg.4

T Epoxy resin and primer

Fiber direction

Fig. 6. Schematic position of strain gauges on the specimens.

characterization tests. The DSC tests were performed on primer properly during the tests, the changes of Coefficient of Thermal
and epoxy resin samples taken from tensile test specimens and Expansion (CTE), a, can be obtained from the variation of strain
on epoxy resin samples taken from debonded FRP sheets after in each cycle (De = emax  emin) as De = aDT. Looking at the figures,
the shear tests. The test methods and procedures were as two different phenomena are recognized; change of CTE and
explained in Sections 2.2 and 2.3. change of the mean strain ((emax + emin)/2). The latter can be due
to micro-cracking, delamination or swelling of the specimens due
3. Results and discussion to moisture absorption.
Thermal expansion of the brick seems to be similar in both lon-
3.1. Strain measurements gitudinal and transverse directions with a slight difference in CTE,
being around 1.5  105/°C and 1.36  105/°C, respectively. A
The variation of specimens’ surface strains along the exposure gradual increase in the mean strain can be observed in both curves,
period are presented in Fig. 7. Assuming that strain gauges worked which can be due to swelling of the brick with moisture
absorption.

Fig. 7. Strain changes in different specimens with exposure cycles.


148 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

For the epoxy resin, the mean strain increases in the first 120 then stabilize until the end of exposure. This change is due to pro-
cycles and thereafter has negligible changes until the end of expo- gression of FRP delamination until the complete separation of FRP
sure period. This increment is coincident with maximum moisture from the brick surface has occurred at around 240 cycles. After this
absorption rate, see Fig. 11, and can be due to moisture swelling point CTE is 7.98  105/°C, which is very close to the one from
[22,36]. The CTE of epoxy resin seems to be constant along the tests epoxy (8.28  105/°C) showing that FRP delamination has
and is equal to 8.28  105/°C. A similar change of mean strain can occurred. Both mean strain and De are smaller in sg.4 as this strain
be observed in primer, which again can be attributed to moisture gauge was located far from the loaded end. De increases only after
swelling, see also moisture absorption curves in Fig. 12. In con- 480 cycles showing that FRP delamination or interfacial micro-
trary, the CTE of primer decreases around 40% along the exposure cracking has occurred at this location. However, as the CTE is much
period changing from a0 = 8.96  105/°C to a480 = 5.52  105/°C. less than that of epoxy (is 2.72  105/°C at cycle 480) only micro-
The strain measurements on GFRP specimens show larger strain cracking has probably occurred at the interface and FRP is not com-
variations and thus larger CTE in transverse direction pletely debonded. Moisture induced swelling can also be responsi-
(a = 9.28  105/°C) in comparison to longitudinal direction ble for a percentage of the observed changes in the mean variation
(a = 1.15  105/°C). The CTE and mean strain variation in trans- in strain gauges bonded to the ORG-specimens.
verse direction are in the same order of epoxy resin, as expected.
In FRPs prepared with unidirectional fibers, the properties are gov- 3.2. Thermal properties
erned by fiber properties in fiber direction (longitudinal direction
in this study) and by epoxy in transverse direction. This can be The variation of Tg obtained from DSC tests is presented in
the reason why the effect of epoxy moisture swelling is not Fig. 8. The DSC tests on bulk epoxy (taken from tensile tests spec-
observed in the mean strain in longitudinal direction. imens) included two heating scans. Both reference epoxy samples
In FRP-strengthened specimens, the strain variation captured (taken from the interface and from tensile specimens) showed a
by sg.1–2 does not show any specific changes during the exposure similar Tg (around 55 °C for the 1-stage DSC tests), demonstrating
besides a slight increment of mean strain. However, the slope of a similar degree of cure at both conditions. The Tg of the bulk adhe-
the mean strain in sg.3, placed at the loaded end, increases at sive in both heating scans increased about 17% during the first 240
around 400 cycles, which can be due to FRP delamination from cycles of exposure and reached a plateau thereafter until the end of
the substrate. De is similar to the one observed in GFRP in longitu- the tests (Fig. 8). However, the Tg of epoxy at the interface is
dinal direction, although the mean strain is slightly larger in increased at a lower rate leading to a total 7% increase at the
bonded specimens. end. The Tg of primer, only measured at the beginning and at the
De is much larger in sg.4 and sg.5 as the behavior is governed by end of exposure period, showed a 20% drop from 56 °C to 45 °C.
epoxy in transverse direction. The transverse CTE recorded by sg.4 The observed changes in the Tg are the result of several counter-
and sg.5 is 1.16  105/°C and 1.97  105/°C at the beginning of acting and interlinked mechanisms, which affect the network
exposure, both far from that of epoxy due to the FRP–brick bond. structure of epoxy resin. Moisture absorption causes plasticization
De and mean strain increase until around 240 cycles in sg.5 and in epoxy resin and thus reduction of its Tg. On the other hand, the
epoxy is exposed to temperatures above its curing temperature
(23 °C) during the hygrothermal cycles (when the temperature
90 rises to 50 °C), which leads to further cross-linking of epoxy and
thus increment of its Tg [37,38]. The curing time, by itself, also
75
leads to improvement of the cross-link density and Tg increment.
60 The effect of moisture absorption on Tg can be simulated as
Tg[° C]

[39]:
45
st T gw ¼ ðAx2 þ Bx þ CÞT gd ð1Þ
30 Epoxy resin (1 heat scan)
nd

15
Epoxy resin (2 heat scan) where Tgw is the Tg of wet epoxy, Tgd is the Tg of dry epoxy, x is the
Epoxy at bond interface
moisture content in the epoxy in the percent weight, and A, B, and C
Primer
0 are constant values obtained by curve fitting of the experimental
0 240 480 720 960
Cycles results. These parameters are calibrated with water immersion tests
performed on the same epoxy resin in [40] and are obtained as
Fig. 8. Variation of Tg with exposure cycles. A = 0.0047, B = 0.056 and C = 1. Eq. (1) is then used for estimating

1.4 120
Normalized value [Tg/Tg0]

96
48
24

1.2 100
0
0
0

Cy
Cy
Cy
0

1.0
cl
cl
Cy

cl

80
es
e
e

s
cl

s
Tg[° C]

se

0.8
2
60
0.6 Tgw/Tgd=0.0047ω −0.056ω+1
40
0.4
20 Hygrothermal condition
0.2 Experimental data
Simulated data Laboratory condition
0
0.0
0.0 0.4 0.8 1.2 1.6 2.0 0 2 4 6 8 10 12
Curing time [months]
Moisture content [%]
(a) (b)
Fig. 9. (a) Tg depression with moisture absorption; (b) changes of Tg in specimens exposed to hygrothermal and laboratory conditions.
H. Maljaee et al. / Composite Structures 147 (2016) 143–154 149

the effect of moisture absorption on Tg during hygrothermal expo- governing mechanism (with the largest contribution) for incre-
sure tests, see Fig. 9a. According to the analytical formula, a total ment of Tg.
reduction of 10% in Tg is expected at the end of exposure period
(corresponding to 1.9% moisture absorption). The difference 3.3. Material properties
between the experimental results and the analytical simulation,
the shaded area in the graph, represents the counteractive effect The changes in mechanical properties and moisture content of
of physical aging (further cross-linking) of epoxy resin due to both the specimens with exposure cycles are presented in this section.
curing time and exposure to temperatures above the curing The moisture content is measured: (i) immediately upon removal
temperature. of the specimens from the climatic chamber; and (ii) at the test
The effect of epoxy curing time on Tg is investigated by perform- moment (after seven days of storage in controlled environmental
ing DSC tests on epoxy samples cured and kept in laboratory con- conditions).
ditions for 2, 6 and 10 months. The results, presented in Fig. 9b in The variation of compressive strength and moisture content of
comparison to hygrothermal exposure results, show a slow but bricks along the exposure is presented in Fig. 10. The changes in
continuous increment of Tg with time (starts from 53 °C and the bricks’ compressive strength can be considered as negligible.
reaches 61 °C at the end of exposure period). Tg increases more The coefficient of variation of the results is at an acceptable range
rapidly in hygrothermal conditions in the first two months, but of 6–14%. The moisture absorption level (0.25% water mass after
then it stabilizes at 61 °C until the end of the tests. The comparison 960 cycles) is negligible in comparison to 12.5% saturation level
of Fig. 9a and b shows that physical aging of epoxy resin due to reported in [40] for the same bricks. The results show suitable
exposure to temperatures higher than curing temperature is the durability of the bricks used in this study under the considered
hygrothermal conditions.
The changes in mechanical properties and moisture content of
20 0.5 epoxy resin with the number of exposure cycles (n) are presented
Moisture content, MC [%]
Compressive strength [MPa]

in Fig. 11. The coefficient of variation (CoV) of the results is at an


16 0.4 acceptable range of 3–14% for elastic modulus and tensile strength
[41], but larger variations were observed for ultimate strain. The
12 0.3 largest reduction in tensile strength and elastic modulus occurred
in the first 120 cycles, when also the largest moisture absorption
8 0.2
was occurred. The rate of degradation in tensile strength is signif-
icantly reduced thereafter. Plasticization of epoxy resin, micro-
4 Compressive strength 0.1
MC after exposure cracking due to swelling or thermal cycles can be the responsible
0 MC at test moment 0.0 mechanisms for the observed degradation in mechanical proper-
0 240 480 720 960
ties. The elastic modulus has increased after the initial decrement
Cycles and then it has decreased again until the end of exposure period.
These fluctuations in the elastic modulus can be due to the coun-
Fig. 10. Changes of bricks’ compressive strength with exposure. teractive effect of moisture plasticization and physical aging. The

70 3.5 1.2
Normalized strength [Ft /Ft0]

4 months of exposure
Moisture content, MC [%]

fte=44.5+14.48exp(-0.004n)
Tensile strength [MPa]

60 3.0 1.0
50 2.5 0.8
40 2.0
0.6
30 1.5 8 months of exposure
0.4
20 1.0
Tensile strength 0.2
10
MC after exposure 0.5 Water immersion
0 MC at test moment 0.0 0.0 Hygrothermal exposure
0 240 480 720 960 0.0 0.5 1.0 1.5 2.0 2.5
Cycles Moisture content, MC [%]

(a) (b)
3.5 3.5 4.5 3.6
Ete=2.29+0.57exp(-0.006n) εe=3.7-exp(-0.003n)
Moisture content, MC [%]

Moisture content, MC [%]

4.0 3.2
Elastic modulus [GPa]

3.0 3.0
Ultimate strain [%]

3.5 2.8
2.5 2.5
3.0 2.4
2.0 2.0 2.5 2.0
1.5 1.5 2.0 1.6
Stress

1.5 1.2
1.0 1.0
Elastic modulus 1.0 960 cycles 0.8
0.5 MC after exposure 0.5 0.5 360 cycles 0.4
0.0 MC at test moment 0.0 0.0 Reference Strain 0.0
0 240 480 720 960 0 240 480 720 960
Cycles Cycles
(c) (d)
Fig. 11. Changes of epoxy resin properties with exposure cycles (a) tensile strength; (b) comparison with water immersion tests; (c) elastic modulus; (d) ultimate strain.
150 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

total degradation in tensile strength and elastic modulus is 26% of extreme plasticization is another reason for the observed exten-
and 18% at 960 cycles, respectively, corresponding to 1.96% water sive degradation. In contrary to the epoxy resin, it seems that the
absorption. An exponential degradation model seems suitable for tensile strength has not reach a plateau and the degradation trend
simulating the changes of tensile strength and elastic modulus, even after the extensive 960 cycles of exposure is not clear. A
see Fig. 11a and c. chemical degradation analysis seems therefore necessary and is
The moisture absorption curve shows that seven days storage proposed for future investigations for interpretation of the
before performing the tests has led to only a slight loss of the observed behavior. A significant decrement of elastic modulus
absorbed water. It also seems that epoxy is near saturation at the (around 57% at the end of the tests) and increment of ultimate
end of exposure period. A comparison with water immersion tests strain (around 88% during the first 120 cycles and reaching 168%
performed on the same epoxy resin in [40], see Fig. 11b, shows that at the end of exposure) can be observed in Fig. 12c and d). An expo-
a similar degradation level is occurred in tensile strength at low nential degradation model is used for each property and is pre-
moisture absorption levels (until 1.6%). However, the degradation sented on the graphs. In case of tensile strength, no conclusion
is larger in hygrothermal conditions at higher moisture absorption could be made for the degradation trend. However, following the
levels in contrary to the larger Tg obtained from DSC tests. This can consistency of fitting curves adopted for other material properties,
be attributed to micro cracking of epoxy resin due to thermal a similar exponential model assuming that the degradation only
cycles. starts after saturation is applied, see Fig. 12a. Again, the stress–
The changes in epoxy ultimate strain are presented in Fig. 11d strain curves show a change of behavior from linear elastic to non-
together with the typical stress–strain curves along the exposure. linear with exposure time.
The ultimate strain increases with exposure time until 240 cycles The changes in mechanical properties and moisture content of
and thereafter it reaches a plateau following an inverse exponen- GFRP coupons with the number of exposure cycles (n) are pre-
tial trend. The stress–strain curves show that the behavior of epoxy sented in Fig. 13. The CoVs of experimental results were in the
resin has changed from linear elastic (with a small nonlinearity range of 4–21%, which is acceptable for wet lay-up manufactured
near failure) to nonlinear (with a small linear region in low stress specimens [14]. The water content curve shows that the maximum
levels) along the exposure time. moisture absorption of GFRP is around 0.9% in this exposure condi-
The variation of mechanical properties and moisture content of tion. Again, most of the degradation occurs in the first 240 cycles
primer are depicted in Fig. 12. The tensile strength of primer pre- and thereafter the tensile strength reaches a plateau. A 21%
sented no changes during the first 240 cycles, although 2.4% mois- decrease in tensile strength is observed after 960 cycles, corre-
ture absorption was achieved, see Fig. 12a. After this point, in sponding to 0.9% water absorption. Fig. 13b shows that the degra-
which the primer seems to be saturated in this environment, the dation starts earlier than water immersion tests performed in [40],
degradation starts and continues until a total drop of 42% at the which can be due to the presence of thermal cycles in the present
end of the tests. The primer reached a similar level of moisture study. The changes in the elastic modulus were negligible during
absorption during three months of water immersion with no sig- the exposure, see Fig. 13c, while the ultimate strain showed a sim-
nificant degradation in tensile strength in [40], see Fig. 12b. The ilar degradation trend as the tensile strength, see also Fig. 13d.
effect of temperature cycles and exposure to high temperatures Mechanical performance of GFRP coupons is affected by mechani-
on the degradation of primer is clear. The large drop of Tg as a sign cal performance of epoxy, fibers and epoxy/fiber interface [42]. The

50 5 1.2
Normalized strength [Ft /Ft0]

3 months of exposure
Moisture content, MC [%]

fte=13+37exp(-0.001n)
Tensile strength [MPa]

1.0
40 4
0.8
30 3
0.6 8 months of exposure
20 2
0.4

10 Tensile strength 1 0.2 Water immersion


MC after exposure Hygrothermal exposure
0 MC at test moment 0 0.0
0 240 480 720 960 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Cycles Moisture content, MC [%]

(a) (b)
3.5 7 6 6
εp= 4.44-2.34exp(-0.005n)
Moisture content, MC [%]

Moisture content, MC [%]


Elastic modulus [GPa]

3.0 6 5 5
Ultimate strain [%]

Etp=1.5+1.42exp(-0.0065n)
2.5 5
4 4
2.0 4
3 3
1.5 3
Stress

2 480 cycles 2
1.0 2
960 cylces
Elastic modulus 1 1
0.5 1
MC after exposure
0.0 MC at test moment 0 0 Reference Strain 0
0 240 480 720 960 0 240 480 720 960
Cycles Cycles
(c) (d)
Fig. 12. Changes of primer properties with exposure cycles (a) tensile strength; (b) comparison with water immersion tests; (c) elastic modulus; (d) ultimate strain.
H. Maljaee et al. / Composite Structures 147 (2016) 143–154 151

1400 2.1 1.2

Normalized strength [Ft /Ft0]


ftf=956.97+229.9exp(-0.015n) 2 months of exposure

Moisture content, MC [%]


1200 1.8 1.0

Tensile strength [MPa]


1000 1.5 0.8
800 1.2
0.6
600 0.9 8 months of exposure
0.4
400 0.6
Tensile strength 0.2
200 Water immersion
MC after exposure 0.3
0.0 Hygrothermal exposure
0 MC at test moment 0.0
0 240 480 720 960 0.0 0.5 1.0 1.5 2.0
Cycles Moisture content, MC [%]

(a) (b)
80 1.6 2.4 1.8
εf=1.6+0.41exp(-0.01n)

Moisture content, MC [%]


Moisture content, MC [%]

Strain at peak load [%]


70 1.4
Elastic modulus [GPa]

2.0 1.5
60 1.2
1.6 1.2
50 1.0
40 0.8 1.2 0.9

Stress
Reference
30 0.6
0.8 480 cylces 0.6
20 0.4
Elastic modulus 120 cylces
10 0.4 0.3
MC after exposure 0.2
0 MC at test moment 0.0 0.0 Strain 0.0

0 240 480 720 960 0 240 480 720 960


Cycles Cycles
(c) (d)
Fig. 13. Changes of GFRP properties with exposure cycles (a) tensile strength; (b) comparison with water immersion tests; (c) elastic modulus; (d) ultimate strain.

glass fibers show reasonable resistance against thermal and mois-


ture exposure [43]. The degradation of GFRP coupons (including
tensile strength and elastic modulus) is thus governed by degrada- Delaminated area
tion of epoxy resin and the fiber/matrix interface. Fiber/matrix
interface is susceptible to moisture ingress and transverse cracking
due to thermal incompatibility between epoxy and fibers, both of
Delamination process

which can cause degradation in the performance of GFRPs. Again,


it seems that an exponential degradation model is suitable for sim-
ulating the tensile strength and ultimate strain.
Delaminated area [%]

1.0
F
0.8
3.4. Bond properties 0.6
0.4
3.4.1. Visual inspection 0.2
The specimens were visually inspected after every 60 cycles of 0.0
exposure (15 days). Due to the relative transparency of the epoxy 0 120 240 360 480 600
Cycles
resin, any possible delamination at the interface level was visible
0 60 120 240 480 600
and could be easily detected, as also reported in [15,21].
Cycles
A progressive FRP delamination was observed in ORG-
specimens, prepared for both pull-off and shear debonding tests, Fig. 14. Progressive delamination in ORG-specimens.
prior to testing. A schematic diagram of the typical delamination
progress in ORG-specimens is presented in Fig. 14. The delamina-
tion generally initiated from the boundaries and propagated from the exposure, leading to the reduction of pull-off strength up to
loaded end to the free end. Although the debonded area varied around 100% in some locations. The failure mode was initially
from one to another specimen in each exposure period, its progres- cohesive with fracture inside the brick. Then it changed to
sion pattern was similar in most of the specimens. While the cohesive-adhesive failure after 240 cycles and to adhesive failure
delamination was significant in ORG-specimens, no delamination after 480 cycles, see Fig. 15.
was observed in GR-specimens showing the effect of surface
preparation on improving the bond quality and durability. 3.4.3. Shear debonding tests
The changes in the debonding force and delaminated area in
3.4.2. Pull-off tests ORG- and GR-specimens are presented in Fig. 16. In contrary to
The variation of pull-off bond strength with exposure cycles is GR-specimens, that did not have any delamination, extensive
presented in Fig. 15. The pull-off tests were only performed on delamination was observed in ORG-specimens (about 50% of the
ORG-specimens. A severe degradation in pull-off strength can be bonded area after 960 cycles). The debonding force in ORG-
observed. The bonded length in the specimens decreased during specimens showed a sudden decrease (about 50%) after 120 cycles.
152 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

1.2 of chemical and mechanical bond mechanisms [44]. The integrity


Zone II of epoxy-substrate bond is believed to be primarily dependent on
Pull-off strength [MPa] 1.0
Zone III the mechanical interlocking. Post-curing of the epoxy resin leads
0.8 to improvement of both chemical and mechanical bonding and
consequently the bond strength [44].
0.6 The typical failure modes of the specimens before and after
exposure are also presented in Fig. 16. An adhesive failure mode
0.4
combined with detachment of a thin layer of brick and tiny cracks
0.2 Zone I on the brick surface was observed in the reference ORG-specimens.
However, the detachment occurred only at the FRP/brick interface
0.0 (adhesive failure) after 120 cycles and remained the same until the
0 240 480 720 960 end of exposure. A similar change was also observed in [15] for the
Cycles specimens exposed to thermal cycles. In contrast, the failure mode
in GR-specimens was cohesive with the fracture inside a thin layer
of the brick in all the exposure period. This is another evidence of
Zone I Zone II Zone III
the improved quality of the bond in GR-specimens. It should also
Cohesive-
Cohesive Adhesive be noted that shear cracks are observable on the surface of the
adhesive
GR-specimens exposed to 960 cycles. These cracks, which are not
present in the failure surface of the reference specimens, are the
evidence of improved mechanical interlocking in these specimens
due to the post-curing of the epoxy resins.
The fracture-based approach is known as an appropriate
approach at the interpretation of debonding problems [7]. There-
fore, in order to evaluate the bond degradation, the degradation
Fig. 15. Variation of pull-off strength and failure mode with exposure cycles. in the interfacial fracture energy is assessed. CNR DT 200 [43] pro-
poses the following formula to correlate the debonding force to the
fracture energy is shown:
Thereafter, the degradation continued with a lower rate and
reached 72% at the end of exposure period, see Fig. 16b. A similar P2max
degradation trend was also reported by Ghiassi et al. [15,21] for Gf ¼ 2
ð2Þ
bf  ð2Ef  tf Þ
225 exposure cycles. In contrary, a slight increase was observed
in the debonding force in GR-specimens, see Fig. 16c. The results where Gf is the fracture energy, Pmax, bf, tf and Ef are the debonding
show clearly the significant effect of surface grinding on improving force, FRP width and thickness, and the elastic modulus of FRP,
the bond performance and durability. The observed increment of respectively. It should be noted that this equation is applicable if
the debonding force can be attributed to the variability of the spec- the bond length is greater than effective bond length [43]. Having
imens, curing time, and post-curing of epoxy adhesive at the inter- the variation of debonding force and FRP’s elastic modulus, the vari-
face level. As discussed in Section 3.3, a 7% increase was observed ation of fracture energy with hygrothermal cycles can be obtained
in Tg of epoxy at the interface being the evidence of further cross- in both sets of specimens. In ORG-specimens, FRP width varies
linking and bond improvement throughout the exposure period. due to delamination, which should be taken into consideration.
The bond performance between epoxy and brick is the resultant The effective bond length is reported to be around 30 mm in [45].

14 70 14 70
Zone II
Zone I

12 60 12 60
Debonding force [kN]

Debonding force [kN]


Debonded area [%]

Debonded area [%]

10 50 10 50
8 40 8 40
6 30 6 30
4 20 4 20
2 10 2 10
0 0 0 0
0 240 480 720 960 0 240 480 720 960
Cycles Cycles

Zone I Zone II 0 cycles 960 cycles


Adhesive Adhesive Cohesive Cohesive

(a) (b)
Fig. 16. Variation of debonding force and failure modes in (a) ORG-specimens; (b) GR-specimens.
H. Maljaee et al. / Composite Structures 147 (2016) 143–154 153

1.0 0.20
Gf=0.8-0.15exp(-0.0024n)

Fracture energy [N/mm]


0.8
0.15

0.6
ORG 0.10

KG
0.4 GR
0.05 KG=0.099+0.00003n
0.2

0.0 Gf=0.13+0.3exp(-0.011n) 0.00


0 240 480 720 960 0 240 480 720 960
Cycles Cycles
(a) (b)
Fig. 17. Variation of: (a) fracture energy; (b) KG in GR-specimens.

Assuming that the effective bond length in ORG-specimens is con- stant relative humidity (90%) in a climatic chamber. The main focus
stant during the tests, the remaining bonded length is more than was on the changes of the bond performance at FRP–brick interface
the effective bond length until the end of exposure and thus Eq. and the material properties. The use of strain gauges for monitor-
(2) can be used. The variation of fracture energy in both sets of spec- ing the environmental induced strain on the specimens during the
imens is presented in Fig. 17a. An exponential model is also pre- exposure helped in understanding the governing mechanisms in
sented for simulating the changes of fracture energy for each materials behavior and changes in the thermal expansion as well
group. The fracture energy in ORG-specimens decreased signifi- as detection of FRP delamination.
cantly up to 68% during 240 cycles. Besides fluctuations, no further The effect of hygrothermal exposure on the compressive
reduction was observed in fracture energy after this point. In con- strength of bricks, used in this study, was negligible. The behavior
trary, an increase of 15% is obtained for GR-specimens. As discussed of epoxy resin and primer changed from linear elastic (with small
before, no degradation was observed in the elastic modulus of GFRP nonlinearity near failure) to nonlinear showing reduction of both
and compressive strength of brick. Thus this slight increment of strength and stiffness with increment of exposure cycles. The
fracture energy can be attributed to the physical aging such as observed degradation was due to the concurrent effect of moisture
post-curing and additional cross-linking in epoxy resin due to the absorption (which leads to plasticization) and post-curing due to
exposure to high temperatures [46]. Increment of joint strength exposure to temperatures above the curing temperature. The DSC
due to physical aging after exposure to thermal cycles was also tests showed an increase of Tg in epoxy resin and a large drop of
reported in [47]. Tg in primer. The degradation in GFRP sheets was limited to the
The fracture energy can be also attributed to the mechanical tensile strength and no specific changes were observed in the elas-
properties of masonry substrate as [43]: tic modulus. The degradation in GFRP was probably due to both
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi plasticization of epoxy resin and fiber/epoxy interface damage.
kb  kG
Gf ¼  f bm  f tbm ð3Þ The degradation trend in epoxy resin and GFRP was observed to
FC
be in direct relation with moisture absorption level following the
where fbm and fbtm are the average compressive and tensile strength exponential trend.
of masonry, kb accounts for the width effects (it can be obtained The bond characterization tests showed that grinding the
from CNR-DT200 [43]), kG is a corrective factor proposed to be equal bricks’ surfaces before application of GFRP significantly improves
to 0.031 mm for FRP systems applied to bricks following the wet the bond durability. While specimens with mechanical surface
layup procedures, and FC is a design confidence factor, which can treatment did not show any degradation in the bond performance,
be assumed equal to 1 for the purpose of this study. The tensile extensive FRP delamination and degradation in bond strength was
strength of the brick can be considered to be equal to 0.1fbm. Thus observed in specimens prepared without any surface treatment.
Eq. (3) can be rewritten as: Increment of fracture energy in GR-specimens was correlated to
pffiffiffiffiffiffiffi the physical aging (post-curing) of epoxy resin at the interface
Gf ¼ kb  kG  0:1f bm ð4Þ
level. Post-curing of epoxy resin led to improvement of bond
It must be noted that this formula can only be applied to GR- between epoxy and brick which was evident in the failure surface
specimens due to the main assumption behind this equation of of the specimens after debonding tests. Predictive models for sim-
having a cohesive failure inside the brick. Having the variation of ulating the material properties, bond fracture energy and correc-
masonry compressive strength from the experimental results and tive design factors for hygrothermal environments were also
fracture energy from Eq. (2) and assuming that kb is constant dur- discussed and proposed.
ing the tests (kb = 1.29), the variation of parameter kG can be plot-
ted versus hygrothermal cycles using Eq. (4), see Fig. 17b. It can be Acknowledgements
observed that kG can have considerable changes after exposure to
environmental conditions, which is contrary to the current The second author acknowledges the financial support of the
assumption of being constant made in design codes. Ministério da Ciência, Tecnologia e Ensino Superior, FCT, Portugal,
under the grant SFRH/BPD/92614/2013.
4. Conclusions
References
The effect of hygrothermal conditions on durability of FRP-
[1] Aiello MA, Sciolti SM. Bond analysis of masonry structures strengthened with
strengthened brick masonry was investigated by performing accel- CFRP sheets. Constr Build Mater 2006;20(1–2):90–100.
erated aging tests. The tests included exposing the specimens to [2] Ghiassi B, Verstrynge E, Lourenço PB, Oliveira DV. Characterization of
concurrent temperature cycles (between 10 °C and 50 °C) and con- debonding in FRP-strengthened masonry using the acoustic emission
technique. Eng Struct 2014;66:24–34.
154 H. Maljaee et al. / Composite Structures 147 (2016) 143–154

[3] Oliveira DV, Basilio I, Lourenço PB. Experimental bond behavior of FRP sheets [25] ASTM C67. Standard test methods for sampling and testing brick and
glued on brick masonry. J Compos Constr 2011;15(1):32–41. structural clay tile, vol. 04, 2014.
[4] Ghiassi B, Xavier J, Oliveira DV, Kwiecien A, Lourenço PB, Zajac B. Evaluation of [26] EN 772-1. Methods of test for masonry units. determination of compressive
the bond performance in FRP–brick components re-bonded after initial strength, 2011.
delamination. Compos Struct 2015;123:271–81. [27] Lourenço PB, Fernandes FM, Castro F. Handmade clay bricks: chemical,
[5] Broughton WR, Mera RD. Review of life prediction methodology and adhesive physical and mechanical properties. Int J Archit Herit August 2013;2009
joint design and analysis software, 1997. (4):38–58.
[6] Silva MAG, Biscaia H. Degradation of bond between FRP and RC beams. Compos [28] ASTM D638-03. Standard test method for tensile properties of plastics, 2014.
Struct 2008;85(2):164–74. [29] ASTM D3039/D 3039M. Standard test method for tensile properties of polymer
[7] Tuakta C, Büyüköztürk O. Deterioration of FRP/concrete bond system under matrix composite materials, 2014.
variable moisture conditions quantified by fracture mechanics. Compos Part B [30] ISO TC 71/SC 6 N. Non-traditional reinforcing materials for concrete
Eng 2011;42(2):145–54. structures-test methods-part 1: FRP bars and grids, 2015.
[8] Amidi S, Wang J. Subcritical debonding of FRP-to-concrete bonded interface [31] ACI 440.3R. Guide test methods for fibre-reinforced polymers (FRPs) for
under synergistic effect of load, moisture, and temperature. Mech Mater reinforcing or strengthening concrete structures, 2012.
2016;92:80–93. [32] Sharma KS, Mudhoo A, editors. A handbook of applied biopolymer technology:
[9] Sen R. Developments in the durability of FRP-concrete bond. Constr Build synthesis, degradation and application. Royal Society of Chemistry (RSC);
Mater 2015;78:112–25. 2011.
[10] Benzarti K, Freddi F, Frémond M. A damage model to predict the durability of [33] Bengoechea C, Arrachid A, Guerrero A, Hill SE, Mitchell JR. Relationship
bonded assemblies. Part I: debonding behavior of FRP strengthened concrete between the glass transition temperature and the melt flow behavior for
structures. Constr Build Mater 2011;25(2):547–55. gluten, casein and soya. J Cereal Sci 2007;45(3):275–84.
[11] Lau D, Büyüköztürk O. Fracture characterization of concrete/epoxy interface [34] Soenen H, Besamusca J, Poulikakos LD, Planche J-P, Das PK, Kringos N.
affected by moisture. Mech Mater 2010;42(12):1031–42. Differential scanning calorimetry applied to bitumen: results of the RILEM
[12] Won J-P, Yoon Y-N, Hong B-T, Choi T-J, Lee S-J. Durability characteristics of NBM TG1 round robin test. In: Kringos N, Birgisson B, Frost D, Wang L, editors.
nano-GFRP composite reinforcing bars for concrete structures in moist and Proc. int. RILEM symp., Stockholm. Springer; June 2013. p. 311–23. 2013.
alkaline environments. Compos Struct 2012;94(3):1236–42. [35] ASTM E1356-08. Standard test method for assignment of the glass transition
[13] Sciolti MS, Aiello MA, Frigione M. Effect of thermo-hygrometric exposure on temperatures by differential scanning calorimetry, 2014.
frp, natural stone and their adhesive interface. Compos Part B Eng [36] Robert RJ, Kural MH, Macke GB. Thermal expansion properties of composite
2015;80:162–76. materials, 1981.
[14] Ghiassi B, Marcari G, Oliveira DV, Lourenço PB. Water degrading effects on the [37] Zhou J, Lucas JP. Hygrothermal effects of epoxy resin. Part II: variations of glass
bond behavior in FRP-strengthened masonry. Compos Part B Eng transition temperature. Polymer (Guildf) 1999;40(20):5513–22.
2013;54:11–9. [38] Wondraczek K, Adams J, Fuhrmann J. Effect of thermal degradation on glass
[15] Ghiassi B, Lourenço PB, Oliveira DV. Accelerated hygrothermal aging of bond in transition temperature of PMMA. Macromol Chem Phys 2004;205
FRP–masonry systems. J Compos Constr 2015;19(3):04014051. (14):1858–62.
[16] Sciolti MS, Aiello MA, Frigione M. Influence of water on bond behavior [39] Chamis CC, Murthy PLN. Simplified adhesively procedures for designing
between CFRP sheet and natural calcareous stones. Compos Part B Eng bonded composite joints. J Reinf Plast Compos 1991;10:29–41.
2012;43(8):3239–50. [40] Maljaee H, Ghiassi B, Lourenço PB, Oliveira DV. Moisture-induced degradation
[17] Maxwell AS, Broughton WR, Dean G, Sims GD. Review of accelerated ageing of interfacial bond in FRP-strengthened masonry. Compos Part B Eng
methods and lifetime prediction techniques for polymeric materials, NPL Rep 2016;87:47–58.
DEPC MPR 016, 2005. [41] Haldar A, Mahadevan S. Probability, reliability, and statistical methods in
[18] Karbhari VM, Chin JW, Hunston D, Benmokrane B, Juska T, Morgan R, et al. engineering design. John Wiley; 2000.
Durability gap analysis for fiber-reinforced polymer composites in civil [42] Schutte CL. Environmental durability of glass-fiber composites. Mater Sci Eng
infrastructure. J Compos Constr 2003;7(3):238–47. R 1994;13(7):265–322.
[19] Hollaway LC. A review of the present and future utilisation of FRP composites [43] CNR-DT 200 R1. Guide for the design and construction of externally bonded
in the civil infrastructure with reference to their important in-service FRP systems for strengthening existing structures, 2015.
properties. Constr Build Mater 2010;24(12):2419–45. [44] Blackburn BP, Tatar J, Douglas EP, Hamilton HR. Effects of hygrothermal
[20] Cromwell JR, Harries KA, Shahrooz BM. Environmental durability of externally conditioning on epoxy adhesives used in FRP composites. Constr Build Mater
bonded FRP materials intended for repair of concrete structures. Constr Build 2015;96:679–89.
Mater 2011;25(5):2528–39. [45] Ghiassi B, Xavier J, Oliveira DV, Lourenço PB. Application of digital image
[21] Ghiassi B, Oliveira DV, Lourenco PB. Hygrothermal durability of bond in FRP- correlation in investigating the bond between FRP and masonry. Compos
strengthened masonry. Mater Struct 2014;47(12):2039–50. Struct 2013;106:340–9.
[22] Vaddadi P, Nakamura T, Singh RP. Transient hygrothermal stresses in fiber [46] Benzarti K, Chataigner S, Quiertant M, Marty C, Aubagnac C. Accelerated ageing
reinforced composites: a heterogeneous characterization approach. Compos behaviour of the adhesive bond between concrete specimens and CFRP
Part A Appl Sci Manuf 2003;34(8):719–30. overlays. Constr Build Mater 2011;25(2):523–38.
[23] Ray BC. Temperature effect during humid ageing on interfaces of glass and [47] Hu P, Han X, Da Silva LFM, Li WD. Strength prediction of adhesively bonded
carbon fibers reinforced epoxy composites. J Colloid Interface Sci 2006;298 joints under cyclic thermal loading using a cohesive zone model. Int J Adhes
(1):111–7. Adhes 2013;41:6–15.
[24] Bao L-R, Yee AF. Effect of temperature on moisture absorption in a
bismaleimide resin and its carbon fiber composites. Polymer (Guildf)
2002;43(14):3987–97.

You might also like