Content uploaded by Carlo Naddeo
Author content
All content in this area was uploaded by Carlo Naddeo on Mar 21, 2018
Content may be subject to copyright.
Water effects on the bond strength of concrete/concrete
adhesive joints
M. Frigione
a,*
, M.A. Aiello
a
, C. Naddeo
b
a
Department of Innovation Engineering, University of Lecce, 73100 Lecce, Italy
b
Department of Chemical and Food Engineering, University of Salerno, 84084 Fisciano (SA), Italy
Received 24 March 2004; received in revised form 22 October 2004; accepted 30 June 2005
Available online 19 September 2005
Abstract
The paper discusses the experimental work by the authors investigating bond strength of epoxy adhesives and their efficiency
when joining to concrete elements; the epoxies studied were those currently used in the construction industry. Flexural tests were
undertaken to determine the mechanical properties of the exposed and the control specimens of three different epoxy adhesives.
In addition, the water resistance of concrete/concrete epoxy joints was investigated by comparing bond strength with those of con-
trol samples; the maximum period of immersion was one month. A reduction in the glass transition temperature and the stiffness at
short immersion time was found for all the adhesives employed, with a subsequent slight increase for prolonged immersion, while the
effects on the strengths resulted almost proportional to their initial values. The effect of water on the adhesion of the joints was found
to be significant, especially at longer immersion times; the bond strength of concrete–adhesive specimens reduced by 30% after one
month of immersion in water.
2005 Elsevier Ltd. All rights reserved.
Keywords: Epoxy adhesives; Concrete/concrete joints; Durability in water
1. Introduction
In recent years, fiber reinforced composites (FRP),
based on polymeric thermosetting resins, have demon-
strated to be an attractive alternative for rehabilitation
or renewal of civil infrastructures, providing significant
advantages to the restoration applications not often
attainable with conventional materials. Widespread uti-
lization of FRPs in construction has, however, been hin-
dered by the lack of long term durability and
performance data on which to base design calculations,
especially when it is realized that FRP composites used
in infrastructure applications are intended to have a ser-
vice life in excess of 50 years.
Durability of a structure can be described as the abil-
ity of the system to maintain designed performance
strength over time under harsh and changing environ-
mental conditions; these durability considerations are
generally more important than the materialÕs pristine
condition. The adverse conditions that may affect dura-
bility of FRPs during their lifetime can be hypothesized
to be: repeated loading, aqueous environment (i.e. high
atmospheric humidity, seawater, rain water, acid rain),
changes in temperatures, exposure to freeze–thaw cycles,
deteriorating chemicals and alkaline environment in the
proximity of Portland cement concrete. Any material is
subjected to microstructural and morphological trans-
formations during its service life, leading to property
changes due to physical and chemical aging. Thus, dura-
bility of a polymeric reinforced/restored structure deals
with the assessment of the initial or design strength of
the repaired structure that may have been lost due to
0950-0618/$ - see front matter 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2005.06.015
*
Corresponding author. Fax: +39 0832 297 215.
E-mail address: mariaenrica.frigione@unile.it (M. Frigione).
Construction and Building Materials 20 (2006) 957–970
Construction
and Building
MATERIALS
www.elsevier.com/locate/conbuildmat
the harmful physical–chemical attacks during its service
life.
The presence of moisture in the composite, in partic-
ular, can initiate undesirable structural changes within
the matrix, in the fiber reinforcement, at their interface
and at matrix/concrete interface. In any case, the result
is the reduction of the durability of the FRP reinforce-
ment. Hence, problems pertinent to the role of perme-
ability of polymeric matrices for composites and
adhesives in conjunction with the concrete adherent
are of prime consideration as the limiting factors of their
performance in service.
The presence of humidity is probably the most harm-
ful environment that can commonly be encountered by
epoxies used as adhesives for civil engineering applica-
tions. The sorption of water can greatly influence the
physical properties of this thermosetting polymer and
its composites. Water may enter a resin either by diffu-
sion or by capillary action through cracks and crazes.
Once inside, the water may alter the properties of the
polymer either in a reversible manner, for example by
plasticization, or in an irreversible manner, for example
by hydrolyzation, cracking or crazing. In the case of
epoxy resins, water molecules bind with resins through
hydrogen bonding. In this way, water is able to disrupt
the interchain Van der Waals forces inside the network
producing an increase of segmental mobility [1].Asa
consequence, the absorption of limited amounts of
water can be regarded as beneficial in terms of both im-
proved toughness, static fatigue resistance and plastic
deformation of the cured resin. On the other hand, an
excessive presence of water is generally considered
harmful leading to a reduction in modulus and strength
with a consequent marked unsuitable decrease of load-
bearing capacity through plasticization effects [2,3].It
is well known that the good properties of epoxy resins
usually undergo a considerable decay after a long period
of immersion in water [4]. Finally, it is not be easy to re-
move the sorbed water completely.
Most of the modern adhesives are not easily hydro-
lyzed, showing a good chemical resistance to water.
However, physical interaction in the form of plasticiza-
tion is a universal consequence of absorption of water.
Plasticization is always accompanied by the lowering
of the T
g
value of the cured resin. This result is particu-
larly worrying for cold-curing epoxies whose typical
glass transition values, when dry, lie in the range
40–55 C, i.e. not much higher than the possible service
temperatures. Water absorption, therefore, will gener-
ally produce a deterioration in the already poor high
temperature load-bearing capacity of epoxy adhesives
cured at room temperature. Hence, the need arises to se-
lect adhesives whose T
g
values do not drop substantially
with water sorption or to assure controlled ambient con-
ditions, when possible. However, relatively short term
exposure to water lead to more or less reversible
plasticization, and an almost complete recovery of the
original T
g
value when water is removed [5].
Finally, the presence of water can be particularly dan-
gerous when the adhesive is used to join two dissimilar
adherends. Water is a highly polar molecule that is per-
meating most polymers, and it is practically impossible
to prevent water from migrating to the interface where
a high-energy surface adherend is present. Although
water plasticizes polymers, it is in the interface regions
where water is believed to reduce the strength. Mays
and Hutchinson [2] reported that water is a harmful fac-
tor for epoxy adhesion joints also for its ability to cause
displacement of adherents by penetrating the interface
of the joint. Moreover, the displacement is even aug-
mented by pre-existing microcracks or debonded areas
at the interface, which originate from poor wetting by
the adhesives [6]. In the case of FRP composites having
an epoxy as matrix, fiber/matrix debonding is among
the major reasons for strength decay in samples aged
in distilled water [7,8]. In a different study, it was found
that the presence of sufficient water at an unsized glass/
epoxy interface causes sudden and catastrophic delami-
nation [9]. The presence of water or moisture accelerates
creep phenomena through plasticization of epoxy matrix
in FRP composites [7]. It reduces fracture energy and
decreases creep-rupture time.
Referring to the effect of the presence of water on the
performances of concrete, a relevant amount of water
can reduce mechanical properties of concrete at the cur-
ing stage [10]. Moreover, once the concrete is hardened,
the presence of water can represent a harmful agent only
for the steel reinforcement, accelerating its electrochem-
ical corrosion [11].
Actual data on durability of cold-curing epoxy adhe-
sives joint to concrete elements related to presence of
water is sparse, not always well documented and, when
available, not easily accessible to the designer. Few re-
search works published in the last years on this subject
indicated a noticeable decrease in the bond strength
(50%) after prolonged immersion in water [12].
Water may easily penetrate through a permeable
adherend like the concrete, which possesses from 10%
to 40% of volumetric fraction of voids and capillary
pores [13], and it can diffuse or be transmitted along
the interfaces through capillary action. After having ac-
cessed the joint, water may cause deterioration of the
bond by altering mechanical properties and adhesive
displacement at the interface [2].
The objectives of this research were to characterize
the degradation behavior of epoxy adhesives in isolation
and when bonded with concrete elements if exposed to
water, to explain the mechanism involved and, finally,
to determine the suitability for utilizing epoxy resins as
adhesives to bond concrete elements in such aggressive
conditions. To this aim, three different epoxy adhesives,
in isolation or in junction with concrete elements, were
958 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
immersed in distilled water for different time spans.
After the scheduled time, each sample was mechanically
tested in order to evaluate the effect of exposure to water
on the mechanical resistance of each adhesive and on its
adhesion strength with concrete.
2. Materials
2.1. Concrete
Different concrete mixes (furnished by FICES
S.p.a., Lecce) were used, with a target compressive
strength of 25 MPa (f
ck
25) and 50 MPa (f
ck
50),
respectively, where f
ck
indicates the characteristic com-
pressive strength of concrete. Details of compositions
of each mix are reported in Table 1. The mechanical
properties (compression and tension strengths) of the
different concretes were evaluated by means of
standard tests, i.e. UNI 6132-72 and UNI 6135-72,
respectively. Their average compressive and tension
strengths and the corresponding standard deviations
are reported in Table 2.
2.2. Adhesives
Different commercial cold-curing epoxy adhesives,
supplied by SIKA Italia S.p.A. and MAC S.p.A., were
selected in this study and they are indicated as S50,
M16 and M20.
S50 is a bisphenolic epoxy resin having a low molec-
ular weight (MW < 700) and a low viscosity (viscosity =
290 MPa s at 20 C). It is used in the restoration of con-
crete to fill and repair cracks of small width and to join
concrete to concrete and also to different materials.
M16 is a bisphenolic epoxy resin with the addition
of 66% of an inorganic filler. It is mainly used to
bond concrete and steel structures and to fix steel
reinforcements within damaged concrete elements.
The filler is largely composed of quartz. Fillers are
commonly added to structural adhesives to improve
their mechanical properties, reduce costs and, possibly,
sensitivity to moisture. Silicates and silica are added to
formulations as either hydrophobic, usually non rein-
forcing, particles or as hydrophilic reinforcing filler
particles [5].
M20 is a bisphenolic epoxy resin with the addition of
49% of an inorganic filler, i.e. calcium oxide. The viscos-
ity of M20 is lower than that of M16 adhesive and M20
is mainly used to bond fresh concrete to hardened con-
crete and to bond concrete and steel structures.
3. Experimental investigation
3.1. Characterization of adhesives
Thermal and mechanical properties of epoxy adhe-
sives were investigated analyzing samples of S50, M16
and M20 previously cured for 20 days at room temper-
ature.
Two differential scanning calorimeters (DSC) were
used to perform the thermal analysis, i.e. a thermoana-
lyzer Mettler – TA 4000 equipped with DSC 30 cell
and a thermoanalyzer Perkin–Elmer DSC-7. All the
thermal scans were carried out between 50 and
250 C with a heating rate of 10 C/min, under nitrogen
atmosphere. The glass transition temperature (T
g
)of
each adhesive was calculated as the mean value of four
experiments.
Flexural characteristics (Young modulus, E; yield
strength, r
y
; and strain at break, e
b
) were measured
using an Instron tensile testing machine (Series 4300),
fitted with a three-point bending fixture at a cross-head
speed of 2 mm/min, following the standard ASTM D
790-92 [14]. The dimensions of the specimens were
80 ·10 ·4 mm and the span to thickness ratio was set
at 16:1. Five samples were tested to determine the
repeatability of the results.
3.2. Tests of water absorption on the adhesives
Tests of water absorption were performed on the
epoxy adhesives cured at ambient temperature for
20 days, following the standard ASTM D 570-81 [15].
Before the test, the samples were subjected to a condi-
tioning procedure, reported in the code, as follows: the
samples were dried in an oven for 24 h at 50 C and then
cooled in a desiccator. Thermal and mechanical tests
were performed on conditioned samples in order to
Table 1
Details of concretes composition
Concrete type Sand (kg/m
3
) Gravel (kg/m
3
) Cement (kg/m
3
) Water (kg/m
3
) Filler (kg/m
3
) Additive (%)
f
ck
25 1009 1747 250 217 330 1.24
f
ck
50 930 1610 360 214 480 0.6
Table 2
Mechanical properties of the concrete
Concrete type f
c
(MPa) sfc(MPa) f
ct
(MPa) sfct (MPa)
f
ck
25 30.20 0.75 2.52 0.07
f
ck
50 66.23 1.59 4.46 0.90
f
c
= mean compressive strength; sfc¼standard deviation of fc;f
ct
=
mean tensile strength; sfct ¼standard deviation of fct.
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 959
evaluate the effects of this treatment on the properties of
the three adhesives.
The percentage of water absorbed after immersion of
24 h and the percentage of water absorbed by samples
substantially saturated, both normalized to the effective
resin content of each adhesive, were determined gravi-
metrically. Epoxy adhesives are prone to water absorp-
tion, because they possess polar sites that attract water
molecules. Water molecules are typically hydrogen
bonded to hydrophilic groups of the cured resin, mainly
hydroxyl and amine groups [3]. Both the amount of
absorbed water and the rate of absorption depend on
formulation variables, such as the epoxy resin and cur-
ing agent types employed, together with environmental
variables, such as temperature and relative humidity,
as well as curing conditions. A wide range of equilibrium
water concentration values and diffusion coefficients
have been quoted in literature for an equally wide range
of formulations and absorption conditions [16–18].In
particular, equilibrium concentrations from 0.25% to
10% by weight have been estimated.
The glass transition temperature and the flexural
characteristics of samples immersed for 14 and 28 days
in distilled water and substantially saturated were calcu-
lated with the same procedure used for dry samples. Be-
fore any test, the samples were left for 2 days in air at
ambient temperature. Each measure was performed on
five samples and the results averaged.
3.3. Adhesion tests
The strength of the bond between each epoxy adhe-
sive and the different concretes was studied in accor-
dance with ASTM C 882-91 [19].
Each adhesive was used to bond together two equal
sections (76.2 mm ·152.4 mm) of concrete cut at a 30
angle from vertical of a concrete cylinder (see Fig. 1).
Before the application of the adhesive, any concrete sur-
face was carefully dried and cleaned. After 20 days,
which was considered the time required to reach the
complete setting of the resin, adhesion tests were per-
formed using a compression testing machine. The bond
strength (r
b
) of the composite cylinder was determined
as the ratio between the load carried by the specimen
at failure (F
u
) and the effective area of the bonded sur-
face (A
b
). Three different thicknesses of each adhesive
(0.5, 2, 5 mm) were employed to study their possible
influence on the bond strength. Each measure was per-
formed at least on three samples and the results aver-
aged.
3.4. Tests after immersion in water on concrete–adhesive–
concrete samples
The physical effects of water exposure on the bond
developed between any adhesive and concrete was, final-
ly, studied. The samples of concrete bonded with differ-
ent adhesives were immersed in distilled water
maintained at a temperature of 23 ± 1 C for different
periods of time: 2, 7, 14 and 28 days. After the different
immersion periods, the samples were left for 2 days in
air at ambient temperature and they were subjected to
compression tests. A total of three specimens for each
test condition were examined. Due to the lack of stan-
dards on this kind of test, the authors chose the
described test procedure trying to simulate some real
service conditions and following the indications of other
researchers [12].
4. Results and discussion
4.1. Properties of adhesives
The main physical (thermal and mechanical) proper-
ties of the cross-linked (cured) resins S50, M16 and M20
are reported in Table 3. It is confirmed that the epoxy
based adhesives cured at ambient temperature possess
a relatively low T
g
, never exceeding 60 C.
Referring to the effect of fillers on the mechanical
characteristics of M16 and M20 adhesives, higher stiff-
ness values for both resins were registered with respect
to that found for the unfilled one. On the other hand,
the inclusion of fillers in the epoxy adhesives did not
show a definite influence on their maximum strength.
4.2. Water absorption properties of adhesives
The water absorption test on adhesive S50 was per-
formed in a previous study [20] and the results are sum-
marized in Table 4. After a 24 h immersion the cured
samples gained 0.62% in weight. The total water ab-
sorbed by the samples substantially saturated was about
1.56% in weight and was reached after 19 weeks of
immersion in water.
Table 4 reports water absorption results also for M16
and M20 adhesives. The percentage of water absorbed
by the epoxy adhesives containing inorganic fillers, nor-
malized to the effective resin content, after one day
76.2 mm
mm 2.241
152.4mm
mm 2.01
30˚
Cutting surface
Fig. 1. Concrete/concrete adhesive joint specimen.
960 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
immersion was in both cases lower than that calculated
for the unfilled resin, i.e. S50. Referring to the water
content values found for substantially saturated sam-
ples, M20 adhesive showed a water content at saturation
conditions similar to that found for S50 resin. For M16
adhesive, on the other hand, a much lower water content
at saturation, i.e. 0.42% wt, was found, possibly due to
the hydrophobic nature of its filler. The time required
for these adhesives to reach saturation conditions, final-
ly, was substantially higher (38 weeks) than that
observed for the unfilled S50 adhesive.
In Tables 5–7 the results of the thermal and mechan-
ical tests performed on immersed samples S50, M16 and
M20, respectively, after different time spans, are
reported. For comparison purposes, the properties cal-
culated on un-conditioned samples as well as on samples
conditioned before the immersion (i.e. 24 h at 50 C) are
reported, in order to assess the influence of the ther-
mal treatment on the final properties of the three adhe-
sives.
The conditioning procedure performed at 50 Con
samples before the immersion in water can be regarded
as a thermal treatment that can influence the properties
of the adhesives. The effects of a thermal treatment on
an epoxy resin depend on the initial structure and the
thermal properties of the resin. In particular, the heating
of a cross-linked epoxy for prolonged time at a temper-
ature close or higher than the T
g
of the resin can cause
one or both of the following: (a) the erasing of physical
aging and (b) the post-curing of the resin (in addition to
the removal of the water eventually contained in the
samples). These effects have important influence on the
Table 3
Thermal and flexural mechanical properties of epoxy adhesives S50,
M16 and M20
Adhesive T
g
(C) E(GPa) r
y
(MPa) e
b
(mm/mm)
S50 46 ± 2 0.830 ± 0.110 27.1 ± 4.5 0.130 ± 0.020
M16 58 ± 2 5.795 ± 0.805 21.2 ± 1.5 0.005 ± 0.001
M20 51 ± 2 4.487 ± 0.487 51.0 ± 6.1 0.011 ± 0.001
T
g
= glass transition temperature; E= Young flexural modulus;
r
y
= yield flexural strength; e
b
= strain at break.
Table 4
Water absorption characteristics of cured adhesives S50, M16 and
M20
Adhesive % Water
(24 h)
% Water
(saturation)
Saturation time
(weeks)
S50 0.62 ± 0.07 1.56 ± 0.17 19
M16 0.25 ± 0.02 0.42 ± 0.09 38
M20 0.08 ± 0.00 1.37 ± 0.11 38
% Water (24 h) = percentage of water absorbed after immersion of
24 h, normalized to the effective resin content; % water (satura-
tion) = percentage of water absorbed by samples substantially satu-
rated, normalized to the effective resin content; saturation time = time
for water saturation; T
g
= glass transition temperature, measured on
samples substantially saturated.
Table 5
Thermal and flexural mechanical properties of epoxy adhesive S50 as a function of immersion time in distilled water
Adhesive S50 T
g
(C) E(GPa) s
E
(GPa) DE(%) r
y
(MPa) sry(MPa) Dr
y
(%) e
b
(mm/mm) seb102(mm/mm)
Un-conditioned 46 ± 2 0.830 0.17 27.1 7.91 0.130 2.0
Conditioned 46 ± 0 0.800 0.07 – 24.8 1.55 – 0.080 4.0
14 days immersion 38 ± 0 0.615 0.14 23 25.3 0.41 +2 0.047 0.55
28 days immersion 41 ± 0 0.969 0.08 +21 28.1 1.61 +13 0.037 0.41
Saturation (19 weeks) 43 ± 2 0.663 0.14 17 22.5 2.60 9 0.064 0.77
Table 6
Thermal and flexural mechanical properties of epoxy adhesive M16 as a function of immersion time in distilled water
Adhesive M16 T
g
(C) E (GPa) s
E
(GPa) DE(%) r
y
(MPa) sry(MPa) Dr
y
(%) e
b
(mm/mm) seb102(mm/mm)
Un-conditioned 58 ± 2 5.795 0.66 21.2 1.30 0.005 0.06
Conditioned 73 ± 4 5.000 1.73 – 25.0 3.20 – 0.009 0.20
14 days immersion 56 ± 1 2.980 0.32 40 24.7 1.87 1 0.008 0.13
28 days immersion 57 ± 2 3.348 0.46 33 23.8 1.24 5 0.010 0.08
Saturation (38 weeks) 61 ± 0 3.208 0.25 36 22.0 1.97 12 0.012 0.09
Table 7
Thermal and flexural mechanical properties of epoxy adhesive M20 as a function of immersion time in distilled water
Adhesive M20 T
g
(C) E(GPa) s
E
(GPa) DE(%) r
y
(MPa) sryDr
y
(%) e
b
(mm/mm) seb102(mm/mm)
Un-conditioned 51 ± 2 4.487 0.40 51.0 5.18 0.011 0.11
Conditioned 53 ± 2 5.395 0.73 – 55.0 4.14 – 0.014 0.06
14 days immersion 47 ± 0 3.930 0.12 27 55.8 2.42 +1 0.014 0.08
28 days immersion 49 ± 1 4.275 0.31 21 46.0 3.30 17 0.011 0.10
Saturation (38 weeks) 53 ± 0 4.244 0.63 21 45.1 4.98 18 0.011 0.17
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 961
properties of the cured products and must be taken into
account.
Physical aging is a universal phenomenon that oc-
curs in all the amorphous polymers below the glass
transition temperature (T
g
) and that leads to a reduc-
tion in the polymerÕs free volume over time, i.e. in a
‘‘densification’’ [21]. The reduction in free volume re-
duces the polymer mobility and increases the relaxation
time. Structural relaxation in the glassy state is a very
slow process, while it is a quicker process at and above
the glass transition temperature [22,23]. The effects of
physical aging are generally manifest in a reduction
of creep compliance, in a stiffening, in a reduction of
ultimate elongation, in an increase in yield strength
[22,24–28]. However, physical aging is a thermorevers-
ible phenomenon that can be erased by heating the
polymer above its glass transition temperature [21,22].
As mentioned, a conditioning temperature of 50 Cis
very close to or higher than the T
g
of the adhesive res-
ins employed in civil engineering applications. As a
consequence, the heating of the aged adhesives at tem-
peratures higher than their T
g
will cause their de-aging
and the reestablishment of their initial properties. The
erasing of physical aging does not affect the T
g
of a
cured epoxy resin. Eventual variations in T
g
, that can
be observed in conditioned samples, can be attributed
to a post-curing process which takes place on samples
which are not fully cured.
It has been observed that the thermal treatment at
50 C produces an increase in T
g
of cold-curing epoxy
adhesives; a higher value is obtained by increasing the
time of heating [29]. Although after curing times (four
months) are well above those suggested by suppliers,
i.e. 15 days, the T
g
of the resin reaches a constant value,
the resin system may not be fully cross-linked. The cur-
ing (ambient) temperature, i.e. around 23 C, is about
30 C lower than the final T
g
of system and any further
cross-linking reaction may be slowed by kinetic
restraints [30]. Hence, if the resin is heated at a temper-
ature higher than the ambient temperature, i.e. 50 C,
but still lower than its T
g
, a post-curing process takes
place. In this condition, the cross-linking reactions start
again and the T
g
increases by increasing the post-curing
time. The amount of post-curing depends on the initial
T
g
of the system when compared with the conditioning
temperature.
The thermal treatments used to condition the adhe-
sive samples before the immersion in water can, there-
fore, produce different effects on mechanical properties
in relation to the different extents of proceeding of
de-aging (erasure of physical aging) and/or post-curing
processes.
When analyzing a fully cross-linked adhesive, pos-
sessing a T
g
lower than the conditioning temperature,
the thermal treatment performed on this adhesive will
erase the physical aging, while it does not produce any
post-cure. This is eventually the case of S50 adhesive,
which T
g
does not change as a consequence of the con-
ditioning procedure. Though it is not completely cross-
linked, in fact, the completion of curing reactions begins
at higher temperatures, i.e. above 90–100 C. Since the
major effects of physical aging on mechanical properties
of a thermosetting resin are: (a) the stiffening of the
glassy material and (b) the reduction of ultimate elonga-
tion and the increase in yield strength, the remove of
physical aging should produce reductions in modulus
and maximum strength and an increased ultimate strain.
Thus, small reductions in flexural modulus and maxi-
mum stress are found for S50 adhesive, as the result of
de-aging process. However, the noticeable reduction in
ultimate strain found for the same resin does not match
with the expectations.
Referring to M20 adhesive, the conditioning proce-
dure removed most of the effects of physical aging, while
the post-cure process again did not take place, since the
T
g
increases by only 2 C. Only the slight increase of
ultimate strain, as a consequence of the de-aging proce-
dure, was in line with the expectations. A less clear situ-
ation is the behavior of this adhesive concerning its
stiffness and strength, both of which increase slightly.
When the adhesive sample possesses a higher T
g
than
the conditioning temperature but is not fully cured, as in
the case of M16 adhesive, the de-aging procedure does
not take place but the thermal treatment, on the other
hand, is able to partly post-cure the adhesive, with the
result that the T
g
is increased by 15 C. Consequently,
the maximum strength is increased after the thermal
treatment. However, a decrease in flexural modulus is
also observed as a consequence of the conditioning pro-
cedure, this effect being due to post-curing, as reported
by other researchers [30]. An increase in rupture strain
was also observed.
It must be emphasized, however, that the flexural test
employed does not allow an accurate definition of the
stiffness and ultimate strength of the thermosetting
materials, since it is based on the hypothesis of elastic-
linear behavior of the samples up to the collapse. A ther-
mosetting resin, in fact, will present such a behavior only
within the first stage of loading. Therefore, a further
analysis of the mechanical properties of the adhesives
by means of tensile tests, accurately measuring the
deformation by electrical resistance strain-gauges during
the test, has been considered. The aim of the authors
was to report the preliminary results in order to qualita-
tively compare the properties of the materials after dif-
ferent exposure conditions, even if the measured
properties should not be considered as reference values
without any further confirmation.
With respect to the analysis of the results of thermal
characterization of the epoxy adhesives immersed for a
prolonged time in water, the comparison must be per-
formed with the values obtained for each adhesive on
962 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
conditioned specimens. In all cases, at short immersion
time an initial decrease in T
g
was measured; this was
probably due to plasticization effects. Following this, a
new increase in glass transition at longer immersion time
was noted. A maximum decrease in T
g
was measured for
the adhesive with the highest initial T
g
, i.e. M16
(DT
g
=17 C). In a saturation condition, moreover,
the adhesives M20 and S50 almost recover the initial
T
g
value, while M16 adhesive showed a decrease in T
g
by about 12 C.
The results obtained in this study seem to confirm
those of several authors. Referring to the effect of an
immersion in water on the thermal properties of epoxy
adhesives, relatively short time of exposure lead to more
or less reversible plasticization, producing a lowering of
the T
g
[31]. The decrease of T
g
as a consequence of
immersion in water is a physical change that can par-
tially be reversed upon drying. The glass transition tem-
perature is a very important parameter of epoxy resin
and epoxy matrix composites because the T
g
establishes
the service environment for the materialÕs usage. Usu-
ally, when the material is exposed to a hygrothermal
environment the T
g
decreases and, as a consequence,
the service temperature of the material changes. This
modification in T
g
reflects the degree of resin plasticiza-
tion and water/resin interactions occurring in the mate-
rial. As already pointed out, this effect is of particular
concern for cold-curing epoxies whose typical glass tran-
sition temperatures, when dry, are not much higher than
the possible service temperature. Hence, the need arises
to select adhesives whose T
g
Õs do not drop substantially
with water sorption or to assure controlled ambient con-
ditions, when possible.
On the other hand, the increase in T
g
after a longer
immersion time is most likely due to additional cross-link-
ing during exposure to water. Additional cross-linking
can take place, as the epoxy samples would not be fully
cured at room temperature and immersion in water can
cause plasticization of the resin with a consequent reduc-
tion in T
g
of the cured wet resin [32]. The lowering in T
g
upon moisture ingress allows the polymer chains to
become mobile; this allows a limited displacement of
polymeric segments promoting post-curing. To confirm
this hypothesis, it was reported that higher values of T
g
resulted for longer immersion time and higher exposure
temperature [33].
The greatest depression in T
g
as consequence of
immersion in water seems to be related to: (a) the initial
T
g
values, (b) the higher the initial T
g
value and (c) the
greater the reduction in T
g
. Moreover, the initial T
g
val-
ues influence also: (a) the new increase in T
g
after the
first immersion period, (b) the higher the initial T
g
values and (c) the greater the increase in T
g
Õs after the
maximum reduction.
As reported from other studies, the effect of an immer-
sion in water for prolonged time on the mechanical
properties of mild-cured bisphenolic epoxy adhesives
(T
g
=76C) can be summarized in an initial increase
in ultimate tensile strength (up to 21% after a three-
monthÕs immersion), followed at longer immersion times
(i.e. five months) by a decrease to values similar to that of
the unaged polymer [7]. The initial increase in the ulti-
mate tensile strength was again explained in terms of
an increase in cross-link density [34]. Later reductions
in strength, finally, were the result of degradation due
to the presence of water. From the same study, the
YoungÕs modulus of the aged epoxy resulted marginally
lower than that of the control samples, increasing the
reduction in modulus by increasing the immersion time
(after a five months immersion the reduction approached
18%). The elongation at break, moreover, tended to
increase initially (up to 76% after a three-monthÕs immer-
sion), but, at later times, the material became brittle
(with a final increase of about 30% with respect to
unaged samples after a five-monthÕs immersion). It has
been reported that a reduction in the failure strain can
be regarded as a clear and sensitive indicator of polymer
degradation [35].
All the adhesives analyzed in the present study
showed values of flexural modulus that decrease at the
initial stages of immersion and then slightly increase at
longer immersion times, reaching a constant value after
about one month. The effects of post-curing during
immersion could lead to a higher modulus in the epoxy
systems. The differences are very small for the resins S50
and M20; and, in fact, their values of flexural modulus
at saturation are reduced by about 17% and 21%,
respectively, with respect to the initial value. In the case
of M16, the reduction in flexural modulus in saturation
conditions is more marked, i.e. about 36% with respect
to the initial value.
The influence of water at short exposition times on
the strength of S50 adhesive leads to a slight increase
of this property, while on M16 adhesive it is rather insig-
nificant. At longer exposition times, on the other hand,
the reductions of maximum stress for both resins are
around 10% of the initial value of the pristine condition
of samples. These results are in accordance with the
mentioned literature. For M20 adhesive, on the other
hand, the immersion in water causes a limited decrease
of this property. In particular, after one month of
immersion in water the reduction in maximum stress is
by about 17% and it retains this value up to saturation
condition.
The effect of water on ultimate strain is almost always
a decrease of deformability. For M20 adhesive, contain-
ing 49% wt of filler, the already low value of strain at
break is reduced by about 22% after one month of
immersion. For the unfilled adhesive, i.e. S50, the reduc-
tion in deformability is even more severe. An immersion
period of one month reduces the ultimate strain of S50
by about 54%. At saturation, however, this reduction
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 963
is partly recovered, i.e. becoming 20% of initial value.
The only exception is M16 (filler content = 66% wt).
For this adhesive, in fact, the initial low value of strain
is improved by about 33% when saturation is reached. It
must be emphasized, however, that M16 presented a dif-
ferent behavior in terms of increase in T
g
and water
absorbed at saturation.
In conclusion, the effect of permanence in water on T
g
of the cured adhesives is a reduction in T
g
at short
immersion time, by about 3–12 C; this can be explained
by the plasticization phenomena. Then, this property
increases slightly for a prolonged immersion. It is also
confirmed that the system with a higher initial T
g
shows
a greater reduction in T
g
at the first stages of immersion
and subsequently a greater new increase in T
g
at longer
immersion times. The effect of water on stiffness of these
three adhesives seems to be very similar to that of water
on their T
g
Õs. It was noted that the effect of water on the
strength of epoxy adhesives is proportional to the initial
value of this property.
4.3. Adhesion tests
The main results of adhesion tests previously deter-
mined are given in Table 8.
The epoxy adhesives, which are used in concrete-
to-concrete bonding, often possess mechanical strengths
greater than those of Portland cement concrete [12].In
such cases, the fracture takes place within the concrete,
when its tensile strength is achieved. As a confirmation,
all the samples prepared with f
ck
25 concrete, i.e. the
concrete with a low resistance comparable to those of
M16 and M20 adhesives, showed a collapse typical of
the entire concrete specimens under compression load
irrespective of the adhesive employed, i.e. vertical cracks
within the whole samples. In these cases, the kind of the
adhesive as well as the thickness of the adhesive layer do
not influence the bond strength of the joint. For both
M16 and M20 resins, in fact, values of bond strength
around 15–16 MPa were generally measured.
Considering the analysis of the joined samples made
with f
ck
50, and since this concrete possesses a strength
appreciably higher than those of each adhesive, the
bond strength, as well as the kind of failure observed
in the specimens, is mainly influenced by the strength
of the specific adhesive. A higher bond strength was
achieved by using the more resistant adhesive (M20),
even when compared to specimens produced with M16
adhesive, that possesses similar modulus but different
strength value with respect to M20. When adhesives,
having similar resistance but different stiffness (i.e.
M16 and S50) were used to bond specimens of f
ck
50,
lower bond strength values were recorded with the resin
possessing the lowest modulus (i.e. S50).
For each of the adhesives employed, by increasing the
thickness of the adhesive layer a lower bond strength
was recorded. This is explained by a higher deformation
in the adhesive joint, resulting in an early failure.
As expected from a concrete with a high strength, the
failure mechanism, generally, is of a mixed type, with
simultaneously crushing within the concrete and inside
the adhesive resin and the interface debonding. Failure
at interface was more frequently observed in samples
prepared with M20 resin, i.e. by using a more resistant
and stiff adhesive. On the other hand, when M16 and
S50 resins were used, fracture inside the adhesive layer
was often recorded, because these resins have strength
values almost halve with respect to M20.
4.4. Adhesion tests after immersion in water
Referring to the samples made with the f
ck
25 con-
crete, reported in Tables 9 and 10, the bond strength
reduced by increasing the time of permanence in water
Table 8
Results of adhesion tests performed on joints obtained with different concretes and epoxy adhesives
System Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%)
f
ck
25 + M16 0.5 120.00 19.00 16 14.64 2.44 17
2.0 138.52 22.25 16 16.85 2.64 16
5.0 124.45 27.70 22 15.28 3.46 23
f
ck
25 + M20 2.0 131.03 21.30 16 16.01 2.52 16
5.0 129.79 6.75 5 15.94 0.93 6
f
ck
50 + S50 0.5 164.96 1.61 1 19.56 0.12 1
2.0 152.20 0.97 1 17.62 0.03 1
5.0 146.32 17.98 12 16.51 2.00 12
f
ck
50 + M16 0.5 215.28 34.02 16 25.98 4.29 16
2.0 205.81 18.37 9 24.53 1.99 8
5.0 190.82 17.12 9 22.83 2.10 9
f
ck
50 + M20 2.0 274.22 74.16 27 32.82 8.56 26
5.0 243.88 48.36 20 29.23 5.80 20
964 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
and it is only in some extent influenced by the behavior
of the adhesive resin when immersed in water.
As can be observed in Table 9, the specimens bonded
with M16 resin at short immersion times retain the ini-
tial values of bond strength, while, at longer immersion
times, the decay of the bond strength reaches values
around 10–20%. The results, however, were very scat-
tered and the influence of the adhesive thickness on
the bond strength degradation is not very clear.
As can be seen in Table 10, the specimens prepared
with M20 resin showed a decrease in bond strength at
longer exposure times to water (i.e. when exceeding 2
weeks of immersion), increasing the reduction in bond
strength with immersion time. However, at shorter
immersion times, an increase in the bond strength was,
actually, recorded. Moreover, a more relevant degrada-
tion of bond properties is observed by using the highest
resin thickness (i.e. 5 mm). The adhesive joint reflects the
behavior of the resin M20 when immersed in water. In
fact, after one month of immersion in water, a decrease
of its strength by about 17% is observed.
An initial slight increase in joint strength with time
was found for various heat-cured epoxy adhesive
bonded metal joints exposed to wet environment [36–
38], attributed to the relief of shrinkage stresses in the
adhesive due to the presence of water or moisture [39].
At longer exposure time, however, the average shear
strength was found to decrease with time. The visual
inspection of failure surfaces, moreover, revealed that
the failure mode becomes increasingly interfacial as the
exposure time was increased.
The kind of collapse observed in specimens immersed
in water depends on several parameters, i.e. the period
of immersion, the kind of concrete and resin employed
and, in a few cases, the thickness of the adhesive layer.
Under dry conditions, failure of structural joints nor-
mally occurs by cohesive failure of the adhesive layer
or within the concrete, depending on the resistance of
the single components. Prolonged exposure to a wet
environment, however, shifts the failure mode to adhe-
sive failure through the polymer–substrate interface
[6,40]. This trend is favored by increasing exposure time.
Table 9
Results of adhesion tests performed on joints obtained with f
ck
25 concrete and M16 epoxy adhesive after different immersion time in distilled water
Days of immersion Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%) Dr
b
(%)
0 0.5 120.00 19.00 16 14.64 2.44 17 –
2.0 138.52 22.25 16 16.85 2.64 16 –
5.0 124.45 27.70 22 15.28 3.46 23 –
2 0.5 132.61 21.28 16 16.28 2.50 15 +11
2.0 127.38 19.52 15 15.60 2.35 15 7
5.0 125.49 4.04 3 15.37 0.42 3 0
7 0.5 117.81 2.90 2 14.39 0.32 2 2
2.0 122.03 7.44 6 14.87 0.98 7 12
5.0 116.42 3.34 3 14.21 0.44 3 7
14 0.5 89.48 5.19 6 10.87 0.67 6 26
2.0 123.35 11.44 9 15.15 1.35 9 10
5.0 114.48 8.05 7 14.02 0.99 7 8
28 0.5 107.24 14.20 13 13.02 1.80 14 11
2.0 124.06 1.33 1 15.01 0.12 1 11
5.0 142.72 7.96 6 17.32 1.00 6 +13
Table 10
Results of adhesion tests performed on joints obtained with f
ck
25 concrete and M20 epoxy adhesive after different immersion time in distilled water
Days of immersion Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%) Dr
b
(%)
0 2.0 131.03 21.30 16 16.01 2.52 16 –
5.0 129.79 6.75 5 15.94 0.93 6 –
2 2.0 158.03 15.53 10 19.37 1.92 10 +21
5.0 142.82 3.07 2 17.48 0.61 3 +10
7 2.0 144.20 16.36 11 17.69 2.03 11 +10
5.0 157.66 20.40 13 19.37 2.54 13 +21
14 2.0 130.16 22.68 17 15.88 2.87 18 1
5.0 111.24 7.64 7 13.60 0.92 7 15
28 2.0 127.94 6.96 5 15.63 0.81 5 2
5.0 92.70 25.40 27 11.36 3.09 27 29
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 965
The loss of joint strength due to water is, therefore, be-
lieved to be due to degradation of the interface rather
than weakening of the bulk adhesive when low resistant
concrete is employed.
Referring to the kind of failure observed in the sam-
ples made with f
ck
25 concrete, the influence of the per-
manence in water was noticed only after 7 days
treatment. In this case, the analysis of the tested speci-
mens evidenced the presence of slip at the interface in
addition to the vertical cracks within the concrete, as
in the case of the control specimens. An explanation
for the more frequent occurrence of slip at the interface
could be the weakened adhesion between concrete and
adhesive due to the presence of water at the interface.
It must be emphasized, however, that such a phenome-
non prevailed in samples prepared with M20 adhesive,
in agreement with the greater water content absorbed
by this adhesive at long immersion times.
After 29 days of immersion in water, interface deb-
onding was the dominant mechanism of fracture which
occurred in samples prepared with M20 adhesive, partic-
ularly when the highest adhesive thickness was em-
ployed. In this case, the longer the period of exposure
the more appreciable diffusion of water took place
toward the concrete/resin interface and, in addition, a
significant decay of the mechanical properties of the
adhesive; both phenomena growing with the amount
of resin used in the specimens. After the same period
of immersion in water, the samples made with M16
adhesive were affected to a lesser extent by the presence
of water. They showed slip at the interface in several
cases, even though the decisive collapse was always
caused by fracture inside both the resin and the
concrete.
With f
ck
50 concrete a greater influence of the resis-
tance of the adhesives on the resistance of the whole sys-
tem was expected, since the resistance of the concrete in
this case is appreciably higher than that of each adhe-
sive. The results of bond strength tests performed on
specimens made with f
ck
50 and S50, M16 and M20
adhesives, reported in Tables 11–13,andinFigs. 2–4,
respectively, confirmed this assumption. The presence
of water, in fact, influences the bond strength of the
specimens to a larger extent with respect to the samples
produced with f
ck
25 concrete, especially employing the
adhesives with a higher water uptake at saturation (i.e.
S50 and M20). It has been reported in the literature that
a critical water concentration exists below which water-
induced damage of the joint may occur to a minor ex-
tent. For any epoxy system, it is estimated to be 1.35%
wt [40]. Any loss in the joint strength by the absorbed
water can be restored upon re-drying if the equilibrium
moisture uptake is below the critical water concentra-
tion [6]. The low amount of water uptake at saturation
of M16 adhesive, therefore, could explain the limited ef-
fect of water on the bond strength of specimens joined
with this resin.
In addition, the bond strength at which failure occurs
generally falls progressively with time of exposure to
water. There is an indication for S50 and M20 adhesives
that the strength values decay to a minimum level after
14 days of exposure to water, although there is some
scatter in the data.
Analyzing in detail the results for S50 adhesive, re-
ported in Table 11, it is observed that after a 2 days
immersion similar results to control specimens were
found, but by increasing the exposition time to 7 days,
a general decrease of bond strength around 20% was
registered. As mentioned, after two weeks of immersion
in water, the decrease in bond strength reached an
almost constant value of about 35%, confirmed also
for 28 days of exposition to water. All the results seemed
Table 11
Results of adhesion tests performed on joints obtained with f
ck
50 concrete and S50 epoxy adhesive after different immersion time in distilled water
Days of immersion Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%) Dr
b
(%)
0 0.5 164.96 1.61 1 19.56 0.12 1 –
2.0 152.20 0.97 1 17.62 0.03 0 –
5.0 146.32 17.98 12 16.51 2.00 12 –
2 0.5 153.34 0.53 1 18.02 0.02 1 8
2.0 169.20 7.51 4 19.68 1.01 5 +12
5.0 135.82 22.78 17 15.65 2.30 15 5
7 0.5 138.33 3.06 2 15.78 0.33 2 19
2.0 118.59 5.12 4 14.04 0.62 4 20
5.0 – –
14 0.5 105.59 11.38 11 12.52 1.30 10 36
2.0 96.09 11.22 12 11.49 1.34 12 35
5.0 94.83 7.1 7 11.15 0.75 7 32
28 0.5 104.38 15.65 15 12.41 1.76 14 36
2.0 92.37 15.87 17 10.81 1.90 18 39
5.0 92.65 9.71 10 10.73 1.26 12 35
966 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
to show that the adhesive thickness has the same influ-
ence on the bond strength values as that registered
on dry samples (i.e. it slightly decreases passing from
the lowest thickness to the highest). This could again
be explained by the mechanical properties of the con-
crete which are higher than those of the adhesive resin.
Table 12
Results of adhesion tests performed on joints obtained with f
ck
50 concrete and M16 epoxy adhesive after different immersion time in distilled water
Days of immersion Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%) Dr
b
(%)
0 0.5 215.28 34.02 16 25.98 4.29 16 –
2.0 205.81 18.37 9 24.53 1.99 8 –
5.0 190.82 17.12 9 22.83 2.10 9 –
2 0.5 153.77 46.80 30 18.87 5.74 30 27
2.0 228.02 27.26 12 27.83 3.10 11 +13
5.0 175.85 27.45 16 21.67 3.44 16 5
7 0.5 156.26 22.75 14 19.21 2.93 15 26
2.0 199.30 20.41 10 24.38 2.71 11 1
5.0 206.07 18.40 9 25.26 2.17 6 +11
14 0.5 135.68 27.36 20 16.68 3.42 20 36
2.0 213.65 30.25 14 25.93 3.44 13 +6
5.0 195.13 6.21 3 24.12 0.90 4 +6
28 0.5 162.13 36.47 22 19.23 4.15 22 26
2.0 197.20 35.85 18 23.64 4.04 17 4
5.0 196.12 43.70 22 23.63 5.03 21 +4
Table 13
Results of adhesion tests performed on joints obtained with f
ck
50 concrete and M20 epoxy adhesive after different immersion time in distilled water
Days of immersion Adhesive thickness
(mm)
F
u
(kN) sFu(kN) COV (%) r
b
(MPa) srb(MPa) COV (%) Dr
b
(%)
0 2.0 274.22 74.16 27 32.82 8.56 26 –
5.0 243.88 48.36 20 29.23 5.80 20 –
2 2.0 257.04 36.04 14 30.79 3.98 13 6
5.0 210.80 50.04 24 25.15 5.70 23 14
7 2.0 227.83 6.16 3 27.97 0.81 3 15
5.0 238.43 20.39 8 28.42 2.48 9 3
14 2.0 197.82 72.38 37 23.63 8.84 37 28
5.0 190.78 64.84 34 22.75 7.57 33 22
28 2.0 166.81 20.22 12 20.30 2.54 13 38
5.0 196.81 8.03 4 24.02 0.97 4 18
10
12
14
16
18
20
22
0102030
Immersion time (days)
Bsdnotretgn(hMPa)
t=5.0
t=2.0
t=0.5
Fig. 2. Bond strength vs. immersion time for joints obtained with f
ck
50 concrete and S50 epoxy adhesive (t= thickness of the adhesive
layer).
14
16
18
20
22
24
26
28
30
0102030
Immersion time (days)
Bond streng htM(Pa)
t=0.5
t=2.0
t=5.0
Fig. 3. Bond strength vs. immersion time for joints obtained with f
ck
50 concrete and M16 epoxy adhesive (t= thickness of the adhesive
layer).
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 967
Therefore, by increasing the thickness of the adhesive
layer, the influence of the strength of the resin becomes
more relevant.
In Table 12 the results of the tests of the adhesion
strength performed on specimens of f
ck
50 bonded with
M16, are presented. As already stated, the effect of
immersion time on the resistance of the joint is rather
negligible. This result is in good accordance with the
trend of the strength of the pure resin when immersed
in water, since it shows only a slight decrease even after
28 days of immersion. On the other hand, the influence
of the adhesive thickness on the bond strength is signif-
icant, irrespective of the immersion time. When using
the lowest thickness, a higher degradation of the bond
strength was constantly registered, around 30%. On
the other hand, with higher thicknesses, the bond
strength values remain roughly unchanged even after
one month of exposition to water. These results appear
quite different to those observed for the same resin
bonded with f
ck
25. However, as can be seen in Table
12, the results of the tests performed on the specimens,
using the higher resistance concrete, were even more
scattered than those obtained with f
ck
25 concrete, par-
ticularly in the case of specimens prepared with the low-
est thickness, as confirmed by the higher values of the
covariance.
As already observed, a continuous decrease in bond
strength by increasing the immersion time up to 14
days was registered for the specimens prepared with
M20 resin, as can be seen in Table 13. Also in this
case, the effect of the presence of water on the mechan-
ical strength of the adhesive reflects more severely on
the bond strength of the joint than on the specimens
produced with M20 adhesive and the less resistant con-
crete (Table 10). This confirms the critical role of the
adhesive when using high strength concretes. The
thickness of the adhesive layer has a certain effect on
the degradation of the resistance of the joint, increasing
this degradation by using the lowest thickness. It must
be emphasized, however, that also in this case very
high values of covariance for the results were regis-
tered.
The presence of water influences in a similar way the
fracture behavior of the specimens made with f
ck
50 con-
crete and S50 adhesive. In fact, after 14 days of perma-
nence in water a noticeable decrease in bond strength
was registered and it remained constant with increasing
immersion time, irrespective of the thickness of the
adhesive layer. Comparing these results with those
found for the neat resin immersed for prolonged period
of time, it seems that the ultimate strain of the adhesive
is the most influential parameter for the strength of the
joint, that reduces even for short immersion times, as
both the modulus and the maximum strength remained
almost constant.
Considering the samples prepared with M16 adhe-
sive, they frequently showed slip at the interface after
2 and 7 days of immersion. This tendency was even
accentuated at longer exposition times. After 14 and
29 days of immersion in water, the collapse was charac-
terized in most cases by interface debonding.
Analyzing the samples produced with M20 and S50
adhesives, the kind of collapse was always of mixed
type, involving either the concrete, the resin and, possi-
bly, an interface failure. A sudden and explosive collapse
was sometimes observed for samples prepared with
M20.
The obtained results showed that the amount of deg-
radation of all the adhesive resins in the presence of
water reflects on their joint performance, in terms of
bond strength as well as the kind of failure mechanism,
particularly when concretes with high strength are em-
ployed. Similar behavior was observed by employing
adhesives with comparable water uptake at saturation.
However, water has a slightly higher detrimental effect
on the joint produced with the unfilled adhesive (S50)
compared with the filled one (M20); this compares with
other researchers. It has been reported that, after a 135
days immersion in distilled water, a heavily filled epoxy
resin presents a lower bond strength loss (21%) com-
pared with the un-immersed samples than an unfilled
epoxy adhesive (strength loss 50%), in both cases using
the same kind of concrete [12].
5. Conclusions
Rather than being the universal solution for any
kind of deterioration in civil infrastructure, thermoset-
ting materials and their composites could provide
alternatives for rehabilitation and renewal not possible
with conventional materials. However, these materials
can degrade when water is present. In particular, the
resin matrix allows moisture adsorption and this can
lead to a variety of mechanisms, some of which result
in deterioration of the polymer and, in turn, to a
18
22
26
30
34
0 5 10 15 20 25 30
Immersion time (days)
Bsdnotretgn(hMPa)
t=5.0
t=2.0
Fig. 4. Bond strength vs. immersion time for joints obtained with f
ck
50 concrete and M20 epoxy adhesive (t= thickness of the adhesive
layer).
968 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
decay of the effectiveness of the performance of the
restoration.
The reported study, in particular, investigated the
mechanical performances of concrete structures repaired
by using epoxy adhesives, as for injecting cracks, for
anchoring steel reinforcements and finally for the
strengthening of concrete structures using FRPs. The
most critical aspect when bonding different materials is
the interface behavior, that influences in to a great
extent the effectiveness of the bonded system both under
service and ultimate stages. Some key parameters,
involved in the interface behavior, have been investi-
gated in this paper, such as: (a) the properties of adhe-
sives and concrete, (b) the thickness of the adhesive
layer and (c) the presence of environmental agents, in
particular the presence of water. Epoxy adhesives-con-
crete joints, in fact, can be considered, only to a certain
degree, suitable for load bearing applications when the
requirement is a constant exposure to water.
On the basis of obtained results, the following consid-
erations can be made:
The performances of the system concrete-resin is sig-
nificantly influenced by the mechanical and physical
properties of both materials. An important aspect
to be considered is the strength and the modulus of
the concrete substrate in relation to that of the
adhesive.
When increasing the thickness of the adhesive layer,
the joint effectiveness generally decreases. On the
other hand, a very thin layer of adhesive involves a
carefully manufactory procedure to prevent problems
relating to the joint quality. Therefore, an appropri-
ate range of adhesive thicknesses should be defined
from a design point of view.
Environmental agents, in particular water, influence
the joint performances causing a decay of mechanical
properties. Similar studies supported the idea that the
loss in joint strength is primarily due to degradation
of interfacial region through water–substrate interac-
tion. As a consequence, service conditions in terms of
both applied load and environmental agents have to
be considered when adhesives are used for repairing
concrete structures.
The lack of standard tests for durability investiga-
tion of adhesive–concrete joints makes the assess-
ment of reliable theoretical models difficult, due to
the difficulty of comparing results obtained from dif-
ferent test procedures and apparatus. Referring to
adhesives, the available standard tests generally refer
to resins cured at elevated temperatures, and do not
consider the specific properties of adhesives cured at
low temperatures. Durability in severe environments
is one of the key issues in broadening the applica-
tion of structural joints beyond the aerospace
industry.
Accelerated durability tests are generally accepted to
provide an indication of the long-term behavior of
such systems. However, a deeper insight into the
behavior of materials and structures exposed to envi-
ronmental agents would require a proper experimen-
tal investigation, in particular under real conditions,
in order to properly define the relationships between
results obtained under accelerated and long-term
exposure tests.
References
[1] Zhou J, Lucas JP. Hygrothermal effects of epoxy resins. Part I: the
nature of water in epoxy. Polymer 1999;40:5505–12.
[2] Mays GC, Hutchinson AR. Adhesives in civil engineering. Cam-
bridge: Cambridge University Press; 1992.
[3] Shaw SJ. Epoxy resin adhesives. In: Ellis B, editor. Chemistry and
technology of epoxy resins. London: Blackie Academic & Pro-
fessional; 1994. p. 206–55.
[4] Antoon MK, Koenig JL. Irreversible effects of moisture on the
epoxy matrix in glass-reinforced composites. J Polym Sci – Polym
Phys 1981;19:197–212.
[5] Ivanova KI, Pethrick RA, Affrossman S. Investigation of hydro-
thermal ageing of a filled toughened epoxy resin using dynamic
mechanical thermal analysis and dielectric spectroscopy. Polymer
2000;41:6787–96.
[6] Sung N-H. Moisture effects on adhesive joints. In: Adhesives and
sealants. Engineered materials handbook, vol. 3. U.S.A.: ASM
International; 1990. p. 622–7.
[7] Kajorncheappunngam S, Gupta RK, GangaRao HVS. Effect of
aging environment on degradation of glass-reinforced epoxy. J
Compos Constr 2002;6:61–9.
[8] Karbhari VM. Durability of FRP composites for civil infrastruc-
ture – myth, mystery or reality. In: Shenoi RA, Moy SS, Hollaway
LC, editors. Proceedings of the first international conference of
advanced polymer composites for structural applications in
construction. Thomas Telford, UK: Southampton University;
2002. p. 33–43.
[9] McMaster MG, Soane DS. Water sorption in epoxy thin films.
IEEE Trans Compon Hybr 1989;12:373–86.
[10] Carpinteri A. Materiali per lÕIngegneria Civile. Roma: CNR
(Consiglio Nazionale delle Ricerche) Publication; 1989.
[11] Collepardi, M. In: Tintoretto, editor, Il Nuovo Calcestruzzo;
2002.
[12] Tu L, Kruger D. Engineering properties of epoxy resins used as
concrete adhesives. ACI Mater J 1996;93(1):26–35.
[13] Neville AM. Properties of concrete. London: Pitman Publishing
Ltd; 1981.
[14] ASTM D 790-92. Standard test method for flexural properties of
unreinforced and reinforced plastics; 1992.
[15] ASTM D 570-81. Standard test method for water absorption of
plastics; 1988 [Re-approved].
[16] Wright WW. The effect of diffusion of water into epoxy resins and
their carbon-fibre reinforced composites. Composites
1981;12:201–5.
[17] Carfagna C, Apicella A, Nicolais L. The effect of the prepolymer
composition of amino-hardened epoxy resins on the water
sorption behavior and plasticization. J Appl Polym Sci
1982;27:105–12.
[18] Ellis B, Rashid HU. Water absorption isotherms for cured
bisphenol A type epoxy resins: effects of isotherm temper-
ature and resin hardener. J Appl Polym Sci 1984;29:
2021–38.
M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970 969
[19] ASTM C 882-91. Standard test method for bond strength of
epoxy resin system used with concrete by slant shear; 1991.
[20] Frigione M, Naddeo C, Acierno D. Caratterizzazione di
prodotti epossidici per impieghi nellÕingegneria civile. In:
Associazione Italiana di Scienza e Tecnologia delle Macro-
molecole, editor. Proceedings of XIII Convegno Italiano di
Scienza e Tecnologia delle Macromolecole, Genova, Italy;
1997. p. 437–40.
[21] Struik LCE. Physical aging in amorphous polymers and other
materials. New York: Elsevier Scientific Publishing Company;
1978.
[22] Nichols ME, Wang SS, Geil PH. Creep and physical aging in a
polyamideimide carbon fiber composite. J Macomol Sci – Phys
1990;B29(4):303–36.
[23] Clough RL, Gillen KT. Accelerated aging methods for predicting
long-term mechanical performance of polymers. In: Clegg DW,
Colleyer AA, editors. Irradiation effects on polymers. Amster-
dam: Elsevier Applied Science; 1991 [chapter 4].
[24] Morgan RJ. The effect of thermal history and strain rate on the
mechanical properties of diethylenetriamine-cured bisphenol-A-
diglycidyl ether epoxies. J Appl Polym Sci 1979;23:2711–7.
[25] Kong ES-W. Physical aging in epoxy matrices and composites. In:
Dusek K, editor. Advances in polymer science, vol. 80. Berlin,
Heidelberg: Springer-Verlag; 1986. p. 126–71.
[26] Berens AR, Hodge IM. Effects of annealing and prior history on
enthalpy relaxation in glassy polymers. 1. Experimental study on
poly(vinyl chloride). Macromolecules 1982;15:756–61.
[27] Hodge IM. Effects of annealing and prior history on enthalpy
relaxation in glassy polymers. 4. Comparison of five polymers.
Macromolecules 1983;16:898–902.
[28] Cook WD, Mehrabi M, Edward GH. Ageing and yielding in
model epoxy thermosets. Polymer 1999;40:1209–18.
[29] Frigione M, Lettieri M, Mecchi AM, Santamaria U. Durability of
epoxy resins employed for restoration of historical buildings. In:
Proceedings of ICOM-CC 13th Triennial meeting, Rio de Janeiro,
Brazil. London: James & James; 2002. p. 888–93.
[30] Ellis B. The kinetics of cure and network formation. In: Ellis B,
editor. Chemistry and technology of epoxy resins. Lon-
don: Blackie Academic & Professional; 1994. p. 72–116.
[31] Li C, Dickie RA, Morman KN. Dynamic mechanical
response of adhesively bonded beams – effect of environmen-
tal exposure and interfacial zone properties. Polym Eng Sci
1990;30:249–55.
[32] DeÕNeve B, Shanahan MER. Water absorption by an epoxy resin
and its effect on the mechanical properties and infra-red spectra.
Polymer 1993;34:5099–105.
[33] Zhou J, Lucas JP. Hygrothermal effects of epoxy resins. Part II:
variations of glass transition temperature. Polymer 1999;40:
5513–22.
[34] Gupta VB, Drzal LT, Lee CY-C, Rich MJ. The temperature
dependence of some mechanical properties of a cured epoxy resin
system. Polym Eng Sci 1985;25:812–23.
[35] White JR, Turnbull A. Weathering of polymers – mechanisms of
degradation and stabilization, testing strategies and modeling. J
Mater Sci 1994;29:584–613.
[36] Cotter JL. In: Wake WC, editor. Developments in adhe-
sives. London: Elsevier Applied Science; 1977. p. 1.
[37] Minford JD. In: Kinloch AJ, editor. Durability of structural
adhesives. London: Applied Science; 1983. p. 135–214.
[38] Tiwari RK, Balan M, Stumpff P, Soni SR, Mijovic J., Miura N.
The influence of moisture on the strength and durability of
adhesively bonded joints. In: Repecka L, Saremi FF, editors.
Proceedings of the SAMPE 2001 (Society for the Advancement of
Materials and Process Engineering), Long Beach, CA, USA; 2001.
[39] Bowditch MR. The durability of adhesive joints in the presence of
water. Int J Adhes Adhes 1996;16:73–9.
[40] Kinloch AJ. Interfacial fracture mechanical aspects of adhesion
bonded joints – review. J Adhes 1979;10:193.
970 M. Frigione et al. / Construction and Building Materials 20 (2006) 957–970
