Cement and Concrete Composites 73 (2016) 241e250

Contents lists available at ScienceDirect

Cement and Concrete Composites

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

Influence of leaching on the local mechanical properties of an

aggregate-cement paste composite
M. Jebli a, b, c, d, *, F. Jamin a, b, E. Garcia-Diaz d, M. El Omari c, M.S. El Youssoufi a, b
a
LMGC, CNRS, Universit
e de Montpellier, Cc 048, 163 rue Auguste Broussonnet, 34090, Montpellier, France
b
Laboratoire de Microm
ecanique et d’Int

egrit

e des Structures MIST, IRSN-CNRS-Universit

e de Montpellier, France
c
LAEPT, Facult

e des Sciences Semlalia, Universit

e Cadi Ayyad, B.P 2390, 40000, France
d
C2MA, Ecole des Mines dAl es, 6 Avenue de Clavi
eres, 30319, Al
es Cedex, France

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

Article history: This paper presents an experimental study to characterize the mechanical behaviour, at the local scale, of
Received 10 September 2015 sound and degraded concrete after leaching. An experimental protocol was developed to study the ef-
Received in revised form fects of the calcium leaching mechanism at the interfacial transition zone (ITZ) on the mechanical
13 April 2016
properties of the cement-aggregate interface and bulk paste of model material. The process of lixiviation
Accepted 3 May 2016
Available online 10 May 2016
with deionised water occurs very slowly. The experimental study in the laboratory was accelerated by
replacing the water with an ammonium nitrate solution. To quantify the development and kinetics of
degradation at the cemented bond, the concrete leaching fronts were characterized at different levels of
Keywords:
Concrete
degradation using phenolphthalein and local mechanical tests (compression and tensile) and performed
Leaching on samples consisting of two aggregates bound by cement paste. Tests were performed on sound
Cement-aggregates scale samples at different hydration times and on chemically degraded samples. The results show the effect of
Interfacial transition zone (ITZ) leaching on the mechanical properties of the samples and making it possible to correlate the progress of
Mechanical tests leaching to the evolution of these locally considered properties. The experimental results show that there
is an ITZ effect on the alteration of the mechanical properties due to leaching.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction process of material transport and the dissolution of hydrates,

leading to modification of the microstructure of concrete. The
The chemical degradation of concrete by leaching is the result of physicochemical and mechanical properties of the concrete are
the liquid attack of (water with a low pH compared with the pore gradually altered with the evolution of the degradation rate, so a
fluid) leading to the dissolution of the cement paste hydrates [1]. loss of stiffness and strength of the concrete are observed [28]. Most
This is particularly true when the hydrates contain calcium (Ca2þ) experimental studies of the leaching effect on the mechanical
as the main chemical component of the cement paste. For this properties of cementitious materials are investigated for bulk ma-
reason, simplified approaches avoid the need to consider all the terials [20,24,34] and show that the strength of the samples is
elementary chemical mechanisms and focus on the evolution of significantly reduced after dissolution of the hydrates. It is also
calcium ion concentration in the liquid and solid phases. Cement noted that the change in microstructure increases the ductility of
degradation is produced by several dissolution fronts depending on chemically degraded material. However, the effect of leaching on
the relative solubility of each hydrate. The portlandite (CH) con- the interfacial transition zone remains little studied in the scientific
tained in cement paste is the first mineral to be completely dis- literature despite the fact that the presence of aggregates should
solved, followed by the progressive decalcification of (CeSeH) over influence the leaching process and mechanical behaviour of
time [1,10,32]. This degradation, which develops from the surface leached concrete [29]. So it is of interest to conduct investigations at
by the diffusion of products dissolved towards the outside, leads to the local scale in order to provide a better understanding of the
a significant increase in porosity. This, in turn, accelerates the mechanical behaviour of leached concrete and to characterize the
effect of leaching at the Interfacial Transition Zone (ITZ) on the
mechanical concentration at this scale.
* Corresponding author. The Interfacial Transition Zone (ITZ) is considered an additional
E-mail address: mouad.jebli@umontpellier.fr (M. Jebli). phase introduced by the presence of aggregates [27]. It is

http://dx.doi.org/10.1016/j.cemconcomp.2016.05.001
0958-9465/© 2016 Elsevier Ltd. All rights reserved.
242 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250

characterized by a higher porosity than the bulk paste and a high the same diameter (8.3 mm), and were bound by a cement paste
concentration of portlandite [4,11,16,31,40] which increases calcium (cement CEM II/B 32.5 N) with a ratio water/cement of 0.5 with a
leaching. The work of Zheng and al. [39] and Winslow and al. [37] circular cross section. Aitcin [2] found a reduction in the porosity
have also shown that ITZ provides a preferential path for the and thickness of the interfacial transition zone in the High-
transport of water and ions. In order to characterize the lixiviation Performance Concrete (HPC). Generally, the thickness of the ITZ
effect of the ITZ on mechanical properties, a series of experiments of the HPC is limited to 12 mm, whereas that of ordinary concretes is
was carried out by Carde and François [9] on a cement paste and about 50 mm. The distance between two aggregates was fixed at
mortar, with or without silica fume, demonstrating the effect of the 1.7 mm, this being the average distance between two large grains
ITZ on the mechanical properties of concrete. A recent experi- observed in a concrete [12]. According to the definition of De Lar-
mental study by Buzzi and al. [8] showed that a leached rock- rard [23], the distance between these two large adjacent aggregates
concrete interface at fixed leaching depth presents a radical separated by the cement paste is called the Maximum Paste
change of behaviour because of the local loss of mechanical Thickness (MPT). The MPT affects the compressive strength of
strength. concrete. When MPT decreases, the compressive strength in-
The process of lixiviation with deionised water occurs very creases. the interstitial space between aggregates mobilized by the
slowly [1]. It has been shown that for concrete with a water/cement ITZ also increase. Another explanation is crystal orientation of
ratio of 0.4, a degradation depth of 4 cm would necessitate a CeSeH, more the MPT is lower, more the hydrates are oriented,
leaching time about of 300 years and very few data are available in what improve resistance to compressive [23].
the literature about long term degradation of concrete by pure The geometric configuration of the sample is presented in Fig. 1
water [35]. In order to collect experimental data in a reasonable and the “material” data in Table 1. Simplified form of aggregates is
period of time, it is necessary to use an accelerated procedure. used. These were perfectly spherical and of the same size, as orig-
There are three methods to accelerate calcium leaching of cemen- inally proposed by Bisschop and Van Mier [5] to study the effect of
titious materials: application of an electric field [17,33], using aggregates on drying shrinkage in cement-based composites and
temperature [21] and using different chemical solutions to increase by Burlion and al. [7] to study microcracking in leached cementi-
concentration gradients between the interstitial solution and the tious materials. Demoulding was carried out after 24 h. The sample
aggressive environment [13,32]. Ammonium nitrate (NH4NO3) has obtained had a diameter of 8 mm and a height of about 18 mm: This
been chosen by several authors as a representative aggressive so- sample was regarded as representative. The samples were stored in
lution for these accelerated experiments [3,10,19,25,26,30,34]. The water saturated with lime at 21  C to reduce the leaching with
principle of accelerated degradation with ammonium nitrate is the water and the effects of drying shrinkage.
penetration of NH4þ and NO3 ions stemming from the dissolution The specific sample geometry selected allowed us to consider
of ammonium nitrate (Eq. (1)), in the interstitial solution by the concrete as a three-phase material: aggregates, cementitious
diffusion. The NH4þ ion reacts with the OH of the interstitial so- matrix and ITZ (Fig. 2). This allowed us to follow the direct effect of
lution (Eq. (2)) and the NO3 ion reacts with the calcium ions (Eq. leaching at the ITZ interface and also in the bulk paste. Two sets of
(3)). This consumption of Ca2þ ions disturbs the chemical balance of samples were examined: the first was kept in water saturated with
the calcium between the liquid and solid phases, causing the lime at 21  C (control series) and the second was immersed in the
dissolution of hydrates to free new calcium ions (Eq. (4)). aggressive ammonium nitrate solution (degraded series), after

NH4 NO3 ⇔NH4þ þ NO

3 (1)

NH4þ þ OH ⇔NH4 OH⇔NH3 þ H2 O (2)

Ca2þ þ 2NO
3 ⇔CaðNO3 Þ
2
(3)

CaðOHÞ2 ⇔Ca2þ þ 2OH (4)

These accelerated degradation tests make sense when it comes
to studying coupled processes: degradation-cracking [6,34],
degradation-permeability [15,38] or degradation-mechanical
behaviour [10,17,18,36]. This paper presents the effect of degrada-
tion on mechanical behaviour.
The proposed study is a contribution to the experimental
characterization of the impact of chemical degradation by leaching
of cementitious materials on mechanical properties at the cement
paste-aggregate interface and the cement bond. In this context, a
preliminary study was carried out on sound samples to characterize Fig. 1. Description of concrete samples at local scale.
the behaviour of concrete during the hydration process. In-
teractions on this scale were experimentally analysed through
mechanical testing of compression and tensile by means of a spe- Table 1
Material data of components used.
cific device.
Aggregates
Type Calcite CaCO3
1.1. Sample design and materials selected Density 2.6 g/cm3
Diameter 8.26±0.09 mm
For this study, the specimens consisted of two spherical “calcite” Cement paste
aggregates, the calcareous aggregates having been chosen for their Formula CEM II/B-LL 32.5 N
Water/Cement ratio 0.5
high adhesion with the cement paste [22]. The aggregates were of
 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250 243

of ammonium nitrate solution with a concentration of 6 mol/litre.

The leaching process resulted from the high gradients of concen-
tration between the pore solution in the cement paste and the
aggressive solution surrounding the samples. The pH of the solu-
tion was monitored and the solution was renewed when the
pH  9.25 [20]. The pH of the initial solution NH4NO3 concentrated
to 480 g/l was approximately equal to 6. Once the sample was
immersed in the solution, the pH quickly increased to pH 7 then
slowly increased and stabilized at pH 8 (Fig. 3). Thereafter, at each
time (respectively 1, 6, 24, 48, 72, 96 and 101 h), the samples were
extracted to measure the depth of chemical degradation and test
their mechanical behaviour.
The degraded depth was determined using the phenolphthalein
test on sectioned samples. The work presented by Le Belle
go [24]
described the successful use of pH indicators to characterize the
state of the chemical degradation of samples. Phenolphthalein
changes colour when the pH is about 9 during the degradation of
concrete, in the course of the alteration the pH varies from the
initial value of the cement paste (pH ¼ 13) to that of ammonium
nitrate solution (pH ¼ 6) characterising the acid environment.
Fig. 2. Schematic diagram of the three-phase composite sphere for concrete.
Therefore, phenolphthalein does not give the exact position of the
dissolution front of portlandite. By comparing the measurement
being kept in water saturated with lime for 40 days. using pH indicators with that of SIMS microprobe analysis [24], has
After preparation, the samples were subjected to tensile or shown that for cement paste the total degraded depth et can be
compressive tests at various times of hydration (from 1 to 40 days) corrected from the degraded thickness measured with phenol-
to study the local behaviour of a cemented bond. The evolution of phthalein eph by multiplying by a coefficient of 1.17.
the mechanics of compression and tensile tests on samples was
recorded with a high speed camera (Vision Research Phantom v12). et ¼ 1:17  eph (5)

1.2. Accelerated leaching process

1.3. Mechanical tests
The objective of this experiment was to analyse the different
states of chemical degradation of samples. This was achieved by an Both the sound and degraded samples were subjected to
accelerated test able to replicate what is happening in reality to compression and tensile tests at constant loading speed of
characterize the long-term behaviour of cement based materials. 0.01 mm/s. The test monitoring in displacement seems better to get
An accelerated test with ammonium nitrate solution was chosen. more information about the damage mechanisms involved. To
After preparation, the samples were kept in water saturated measure their stiffness and force at rupture. All tests were per-
with lime for the first 40 days and then immersed in a high volume formed on the same Mechanical Testing System (MTS) machine

Fig. 3. Evolution of pH of the solution during the leaching test.

244 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250

equipped with a force sensor of 5000 ± 10 N for compression tests shown in Fig. 6 at different stages of hydration between 1 and 40
and 500 ± 1 N for tensile tests. The force applied to the sample and days.
the longitudinal displacements were measured in the axial direc- These figures show a significant increase in the force at rupture
tion during the test. and stiffness of the samples at the beginning of hydration, with
To perform mechanical tests, specific accessories were pro- almost constant values after more than three weeks of hydration.
duced, adapted to local testing. Fig. 4 shows the experimental de- The examined samples in the compression test show ductile
vice of the compression and tensile tests. Before each test, the behaviour at a very early age, which become increasingly brittle.
dimensions and weight of each sample were measured. For each The values of the force at rupture noted f ðjÞ (in Newton) and
duration of hydration and degradation, four samples are subjected stiffness kðjÞ, were estimated from the linear part of the slope of
to compression or tensile mechanical testing. force-displacement curve calculated as shown in Fig. 7, depending
on hydration time (in days); the compressiontest results are shown
2. Results in Figs. 8 and 9. There was an increase of the forces at rupture and
stiffness in compression during hydration with a drop in speed of
In this section, all the results obtained during mechanical testing evolution after 28 days.
are presented and discussed, for compression tests and tensile tests The evolution of the force was approximated by an exponential
for sound samples and chemically degraded samples by leaching. A function, based on the Eq. (6) presented in Eurocode 2 [14] to es-
preliminary study was carried out on sound samples to characterize timate the macroscopic resistance of concrete.
the behaviour of concrete during the hydration process at the local
scale in the cement-aggregate interface and the cement paste bond. fc ðjÞ ¼ bcc  fc ð28Þ (6)
Interactions on this scale were analysed experimentally.
Where fc ð28Þ ¼ 1500 N is the force at rupture in compression of the
2.1. Sound samples

2.1.1. Compression tests

Mechanical tests were used to characterize the mechanical
behaviour and fracture on the scale of the cement bond. The me-
chanical compressive test characterizes the mechanical properties
of the cementitious binding. The maximum load was governed by
the compressive strength of the bulk paste as shown in Fig. 5,
where cracking develops along a generatrix in the cement bond.
Force - displacement curves during compression tests are

Fig. 6. Force - displacement Curves during compression tests.

Fig. 4. Pictures of supports for (a) the compression test and (b) the tensile test.

Fig. 5. Pictures of the sample during compression test (23 days of hydration). Fig. 7. Method of measuring stiffness.
 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250 245

The compression test characterizes the cement bond between

two spherical aggregates, which is observed by the evolution of the
crack all along the cement bond of the composite.

2.1.2. Tensile tests

The tensile force was derived from the maximum force recor-
ded. Direct tensile testing is the most representative of this stress
mode, but as it requires more preparation and carries a greater risk
of failure, it is less used. The tensile tests in our study were used to
characterize mechanical behaviour of the bulk paste and ITZ. While
the cracking observed occurred only at the interface between
aggregate and cement paste (weak area), no failure occurred in the
cement paste bond or aggregate (Fig. 10).
The force - displacement curves during tensile tests were
recorded as shown in Fig. 11 at different stages of hydration.
These figures show a significant increase in the force at rupture
noted ft ðjÞ (in Newton) at the beginning of hydration, which
remained constant after more than three weeks of hydration. The
values of the force at rupture noted ft ðjÞ (in Newton), depending on
hydration time (in days) of tensile tests are shown in Fig. 12.
The evolution of the force was approximated, like the
Fig. 8. Evolution of the force at rupture as a function of hydration time during me- compressive test, by an exponential function, based on the Eq. (9)
chanical tests. presented in Eurocode 2 [14].

ft ðjÞ ¼ ðbcc Þa  ft ð28Þ (9)

samples at 28 days and bcc is a coefficient which depends on the age
of the concrete according to the equation: Where ft ð28Þ ¼ 150 N is the force at rupture in tensile of the
"  1=2 !# samples at 28 days and bcc defined in Eq. (7). For the results ob-
28 tained in this study, a ¼ 1. It may also be noted that the value of
bcc ¼ exp s 1  (7)
j suggested by Eurocode 2 [14] for values of j  28 days is equal to 1.
As at macroscopic scale, the composite samples at local scale
Where s is a coefficient that depends on the type of cement, which showed very good compression behaviour and poorly resisted
is determined by linear interpolation. For the results obtained in tensile testing. There was a ratio of ten between the forces at
this study s ¼ 1. rupture of the compressive test fc ðjÞ and the tensile test ft ðjÞ which
The stiffness kðjÞ of the samples was estimated based on a slope is generally accepted in the literature at the macroscopic scale. In
of the linear part of the force-displacement curve. A similar expo- the tensile test in our study the crack grew at the ITZ, so the tensile
nential function (8) was used to approximate the evolution of the test characterizes ITZ behaviour is concluded.
stiffness kðjÞ, based on the equation proposed in Eurocode 2 [14] in
order to estimate the evolution of the elastic modulus with hy- 2.2. Altered samples
dration time of concrete.
2.2.1. Chemical degradation tests: leaching
kc ðjÞ ¼ kc ð28Þ  ðbcc Þ0:3 (8) Having put the phenolphthalein in the transverse and radial
sample cross-section, two distinct zones could be observed: an area
Where kc ð28Þ ¼ 9 kN=mm is the stiffness at 28 days and the same in pink (sound zone) and an area in grey (degraded zone). Illus-
bcc expression used before with s ¼ 1. trations are presented in Fig. 13 for different degradation times. On
the sample surface, the thickness of degradation is measured. As we
have seen, the degradation thickness was homogeneous. There was
almost no dispersion, the results with very little dispersion of about
±0.1 mm are obtained.
For each degradation duration, the sound surface Ss is measured
and the degraded surface Sd ¼ S0  Ss is deduced, where S0 is the
initial surface. To follow the progress of the degradation process, we
define a degradation rate d corresponding to the ratio of the
degraded surface by the initial surface:

Fig. 9. Evolution of stiffness as a function of hydration time during mechanical tests. Fig. 10. Pictures of the sample during tensile test (21 days of hydration).
246 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250

Fig. 11. Force - displacement Curves during tensile tests.

Using the p value, the other degradation time which be used for
qualitative estimates can be predicted, depending on the cement
type and characteristics of the acceleration method of degradation
used.
The evolution of the degradation kinetics has been shown
experimentally by other authors [20,24,29,34] using the same
method of accelerated degradation at the macroscopic scale. In our
case,
pffiffiffiffiffiffiffiffiat the local scale, the value of the slope p was equal to 2 mm by
day.
Chemical degradation of the concrete due to dissolution of the
hydrates caused a loss mass in the samples. Fig. 15 shows the loss of
mass in relation to the degraded rate d.

2.2.2. Mechanical tests: compression and tensile tests

At each time interval of chemical degradation, the concrete
samples were subjected to compression and tensile tests to char-
acterize the effect of leaching on the mechanical properties at
different chemical degradation rates respectively for 6, 24, 48, 72,
96 and 101 h. The case of the sound sample, taken as a reference,
was stored in water saturated with lime for 40 days.
Fig. 12. Evolution of the force at rupture as a function of hydration time during me- The evolution of the forces at rupture (for compression and
chanical tests. tensile tests) is then represented according to rate d (Eq. (10)) to
highlight the effect of the degradation on forces. Considering these
results, it is possible to define a critical degradation rate dcrit
(Figs. 17 and 20). This critical rate corresponds to the intersection of
Sd S0  Ss
d¼ ¼ (10) the linear segments at the beginning and at the end of the degra-
S0 S0
dation curve (Figs. 17 and 20).
Figs. 14 and 15 show the evolution of degraded thickness and For the compression test: Figs. 16 and 17 show that at the start of
mass loss as a function of degradation time (1 h, 6 h, 1days, 2 days chemical attack (for 1 and 6 h) there was no effect on the
and 3 days) and degraded rate d. Beyond 3 days, the samples are compressive force at rupture. The decrease of stiffness and force at
totally degraded. rupture of the samples began after the degraded rate d has excee-
Degradation is governed by a diffusion process and can be ded dcrit , as shown in Fig. 17, i.e. a significant dissolution of the
described by Fick’s law relating degraded thickness to the square cement paste to observe the decrease of the force at rupture. Also
root of the time. Fig. 14 shows a linear evolution of the degraded the behaviour of the degraded concrete became more and more
thickness(e) measured in relation to the square root of immersion ductile.
time in the aggressive solution with the formula: When the cement paste was completely leached, the force at
rupture was about three times lower than the reference value.
pffiffi
e¼p t (11) Similar results were obtained by Le Belle
go [24] and Nguyen [29].
The decrease in forces at rupture with the increasing of the
 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250 247

Fig. 14. Evolution of degraded thickness as a function of square root immersion

pffiffiffiffiffiffiffiffi
time.ð day Þ

Fig. 15. Variation of the loss mass in relation to degraded rate.

between chemical degradation by leaching and the mechanical

behaviour of the composite in compression is noted.
For the tensile test: very few tests have been performed on the
effect of chemical degradation on the mechanical properties of
concrete by direct tensile testing, due to the difficulty of realization,
so we were unable to find results in the literature. Figs. 19 and 20
show the fast drop in tensile force at rupture and the stiffness of
degraded samples by about 50% after just 1 h of degradation in the
aggressive solution.
A strong impact of chemical degradation on mechanical
behaviour is noted, more marked in the case of tensile soliciting the
paste-aggregate interface. In the case of the compression test, the
force at rupture fc is less affected when a degradation rate d lower
than dcrit . Beyond dcrit , the force at rupture drops suddenly. Indeed
the sound transverse surface of the sample supports applied forces
Fig. 13. Measurement of degraded thickness using phenolphthalein, on samples hav-
ing undergone respectively 0, 6 h, 24 h, 72 h chemical degradation by NH4NO3. (a)
until a nearby degradation rate of dcrit (Fig. 17). However, in the case
transverse cross-section (b) radial cross-section. of the tensile test, the force at rupture ft is affected for low values of
d (Fig. 20). This result, shown experimentally in this study can be
correlated to the theoretical results, showing that the degradation
degradation rate results from the increase of porosity of the kinetics of ITZ are accelerated, due to the higher porosity and
degraded zone. This leads to the initiation of a cracks series in the concentration of portlandite at the interface. Indeed, these two
damaged area as shown in Fig. 18. A strong impact coupling factors work to accelerate the leaching process and the transport of
248 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250

Fig. 18. Pictures of crack evolution of degraded sample during compression test after
72 h of degradation.

facies in the case of tensile tests show that cohesion loss occurs at
the cement paste-aggregate interface. Besides, the degradation of
the tensile strength properties occurs for relatively low degradation
rate. This suggests that the interface properties are more sensitive
Fig. 16. Mechanical behaviour of sound and partially degraded samples under
compression tests. to the degradation than those of the bulk paste. It is well known
that the porosity of the ITZ is always greater than the porosity in the
bulk paste. So, the chemical attack front, controlled by the diffusion
material, leading to the loss of cohesion between the bulk paste and of the ions, is naturally accelerated at the interface, more porous
aggregates as shown in Fig. 21. This results could not be observed and thus more permeable.
with phenolphthalein du to the weak thickness of ITZ. Fig. 22 illustrates this statement. It shows how the degradation
of the interface could be much faster than the degradation of the
3. Discussion bulk paste, and how it could be widely underestimated by the
overall degradation rate.
The chemical attack by leaching induces the total dissolution of The mechanical compression test showed the effect of leaching
the calcium hydroxide and progressive decalcification of CSH, in the cemented bond between the two aggregates. There was a
leading to an increase in porosity and a decrease of mechanical drop in force at rupture and there were changes in behaviour with
properties. Most experimental studies on the effect of the cemen- the increase of leaching, as is observed at macroscopic scale. This is
titious material leaching on mechanical properties are carried out explained by the increase in porosity due to dissolution of the hy-
on a macroscopic and mesoscopic scale. So it was of interest to drates. The chemical alteration of the interface did not affect the
perform experiments at the local scale to understand the degra- results of the compression test.
dation mechanism on this scale. As shown above, all the rupture The tensile test directly characterized the ITZ in our chosen

Fig. 17. Evolution of the force at rupture as a function of degraded rate during mechanical tests.
 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250 249

Fig. 22. Diagram of the degradation evolution of the composite at the ITZ and the bulk
paste.

cohesion between the cement paste and aggregate, and a signifi-

cant decrease in the mechanical properties of this zone.

4. Conclusion

An original experimental programme of the chemical mechan-

Fig. 19. Mechanical behaviour of sound and partially degraded samples under tensile
ical behaviour of leached concrete at cement-aggregate interface
tests.
scale is presented. Leaching at this scale clearly showed the
important role of the paste-aggregates interface in the concrete
leaching process. Mechanical compressive tests showed the
chemo-mechanical behaviour of leached concrete, and a strong
coupling between calcium leaching and mechanical behaviour is
noted. As the duration of leaching increased, a loss of stiffness and
force at rupture was observed: the mechanical behaviour was
modified by leaching, and it became more ductile as leaching
increased. The tensile test showed the effect of ITZ degradation by
calcium leaching on the change in mechanical properties. The ki-
netics of chemical degradation by leaching were greater for the ITZ
than for other areas, as seen by the abrupt loss of tensile strength
after just a few hours of immersion in the aggressive solution
(NH4NO3). More research needs to be performed to fully analyse
these experimental results. Modelling on this scale with appro-
priate numerical tools should allow investigations at the Repre-
sentative Elementary Volume scale and identification of the
behaviour laws, in the context of the durability and aging of con-
crete issue.

References

[1] F. Adenot, Durabilite
du be
ton: caracte
risation et mode

lisation des processus
Fig. 20. Evolution of the force at rupture as a function of degraded rate during me-
gradation du ciment, 1992 (Ph.D. thesis, France).
physiques et chimiques de de
chanical tests. [2] P. Aıtcin, The durability characteristics of high performance concrete: a re-
view, Cem. Concr. Compos. 25 (2003) 409e420.
[3] A. Bajza, I. Rousekeva, O. Vrana, Corrosion of hardened cement paste by
ammonium nitrate, in: 8th International Congress of the Chemistry of Cement,
vol. 5, 1986, pp. 99e103.
[4] D. Bentz, E. Garboczi, Guide to Using HYDRA3D: a Three-dimensional Digital-
image-based Cement Microstructural Model, Technical Report, National
Institute of Standards and Technology, Gauthersburg, MD, 1992.
[5] J. Bisschop, J. Van Mier, Effect of aggregates on drying shrinkage microcracking
in cement-based composites, Mater. Struct. 35 (2002) 453e461.
[6] N. Burlion, D. Bernard, D. Chen, X-ray microtomography: application to
microstructure analysis of a cementitious material during leaching process,
Cem. Concr. Res. 36 (2006) 346e357.
[7] N. Burlion, T. Rougelot, D. Bernard, Les apports de la microtomographie et
fissuration: cas d’un composite ciment-verre en cours de lixiviation, in:
18e me Congre s Français de Me
canique (Grenoble 2007), 2007.
[8] O. Buzzi, M. Boulon, M. Herve
, K. Su, Leaching of rock-concrete interfaces, Rock
Fig. 21. Pictures of crack evolution of degraded sample during tensile test after 72 h of Mech. Rock Eng. 41 (2008) 445e466.
degradation. [9] C. Carde, R. François, Effect of the leaching of calcium hydroxide from cement
paste on mechanical and physical properties, Cem. Concr. Res. 27 (1997)
539e550.
[10] C. Carde, R. Francois, J. Torrenti, Leaching of both calcium hydroxide and csh
geometry of samples. Very interesting results are obtained in terms
from cement paste: modeling the mechanical behavior, Cem. Concr. Res. 26
of the kinetics of ITZ degradation. There was a very fast decrease of (1996) 1257e1268.
tensile force at rupture after just one hour of degradation. This was [11] S. Care
, Influence of aggregates on chloride diffusion coefficient into mortar,
not the case in the compression test. This can be explained by the Cem. Concr. Res. 33 (2003) 1021e1028.
[12] F. De Larrard, Concrete mixture Proportioning: a Scientific Approach, CRC
higher degradation kinetics of ITZ, resulting from the loss of Press, 1999.
250 M. Jebli et al. / Cement and Concrete Composites 73 (2016) 241e250

[13] A. Delagrave, Mechanisms of Chloride Ion Penetration of Normal and High 266e272.
Performance Cement Based Systems, Ph.D. thesis, Universite
de Laval, Canada, [26] F. Lea, The action of ammonium salts on concrete, Mag. Concr. Res. 17 (1965)
1996. 115e116.
[14] B. EN, 1e1: 2004 Eurocode 2: Design of Concrete Structures, 1992. [27] J. Maso, La liaison entre les granulats et la p^ ate de ciment hydrate
, in: 7e
me

, V. LHostis, H. Peycelon, M. Helie, A. Millard, P. Lachaize, Durabilite
[15] C. Galle
du Congre s International de la chimie des ciments, Rapport principal, 1980.
be
ton des infrastructures et des colis pour le stockage graphite, 2004. Rapport [28] P. Mason, The effects of aggressive water on dam concrete, Constr. Build.
Technique RTDPC. Technical Report SCMME/04-672-A, CEA. Mater. 4 (1990) 115e118.
[16] E. Garboczi, L. Schwartz, D. Bentz, Modeling the influence of the interfacial [29] V. Nguyen, Couplage de
gradation chimique-comportement en compression
zone on the dc electrical conductivity of mortar, Adv. Cem. Based Mater. 2 du be
ton, Ecole des Ponts ParisTech, France, 2005. Ph.D. thesis.
(1995) 169e181. [30] V. Nguyen, B. Nedjar, J. Torrenti, Chemo-mechanical coupling behaviour of
[17] B. Ge
rard, O. Didry, J. Marchand, D. Breysse, H. Hornain, Modelling the Long- leached concrete: part ii: modelling, Nucl. Eng. Des. 237 (2007) 2090e2097.
term Durability of Concrete Barriers for Radioactive Waste Disposals, Tech- [31] J. Ollivier, J. Maso, B. Bourdette, Interfacial transition zone in concrete, Adv.
nical Report, Electricite
de France (EDF), Direction des Etudes et Recherches, Cem. Based Mater. 2 (1995) 30e38.
Paris, France, 1996. [32] E. Revertegat, C. Richet, P. Gegout, Effect of ph on the durability of cement
[18] E. Guillon, F. Benboudjema, M. Moranville, Modelling the mechanical evolu- pastes, Cem. Concr. Res. 22 (1992) 259e272.
tion of a chemically degraded cement paste at the microstructure scale, in: [33] H. Saito, S. Nakane, S. Ikari, A. Fujiwara, Preliminary experimental study on the
FRAMCOS, vol. 5, 2004, pp. 12e16. deterioration of cementitious materials by an acceleration method, Nucl. Eng.
[19] F. Heukamp, Chemomechanics of Calcium Leaching of Cement-based Mate- Des. 138 (1992) 151e155.
rials at Different Scales: the Role of CH-dissolution and CSH Degradation on [34] C. Tognazzi, Influences couple
es de la fissuration et de la de
gradation sur
Strength and Durability Performance of Materials and Structures, Ph.D. thesis, transferts dans les betons, in: Congre s Universitaire de Ge
nie Civil: Concours
Massachusetts Institute of Technology, 2003. Jeunes Chercheurs, 1998.
[20] F. Heukamp, F.-J. Ulm, J. Germaine, Mechanical properties of calcium-leached €gårdh, B. Lagerblad, Leaching of 90-year Old Concrete Mortar in Contact
[35] J. Tra
cement pastes: triaxial stress states and the influence of the pore pressures, with Stagnant Water, Technical Report, Swedish Nuclear Fuel and Waste
Cem. Concr. Res. 31 (2001) 767e774. Management Co., Stockholm (Sweden), 1998.
[21] S. Kamali, B. Ge
rard, M. Moranville, Modelling the leaching kinetics of cement- [36] F. Ulm, J. Torrenti, F. Adenot, Chemoporoplasticity of calcium leaching in
based materials influence of materials and environment, Cem. Concr. Compos. concrete, J. Eng. Mech. 125 (1999) 1200e1211.
25 (2003) 451e458. [37] D.N. Winslow, M.D. Cohen, D.P. Bentz, K.A. Snyder, E.J. Garboczi, Percolation
[22] S. Kamali-Bernard, D. Keinde, F. Bernard, Effect of aggregate type on the and pore structure in mortars and concrete, Cem. Concr. Res. 24 (1994)
concrete matrix/aggregates interface and its influence on the overall me- 25e37.
chanical behavior. A numerical study, in: Key Engineering Materials, vol. 617, [38] I. Yurtdas, S. Xie, J. Secq, N. Burlion, J. Shao, M. Sibai, P. Brossolet, B. Fraboulet,
2014, pp. 14e17. Trans Tech Publ. Couplage comportement me
canique et perme
abilite

: cas d’une p^ ate de ciment
[23] F. de Larrard, A. Belloc, The influence of aggregate on the compressive pe
trolier de
grade

e chimiquement a  90 C, Rev. Eur. Ge
nie Civ. 11 (2007)
strength of normal and high-strength concrete, ACI Mater. J. 94 (1997). 827e837.

go, Couplage chimie me
[24] C. Le Belle
canique dans les structures en be
ton arme
[39] J. Zheng, C. Li, X. Zhou, Thickness of interfacial transition zone and cement
attaque
es par l’eau-Etude expe
rimentale et analyse nume
rique, Ph.D. thesis, content profiles around aggregates, Mag. Concr. Res. 57 (2005) 397e406.
ENS de Cachan, France, 2001. [40] R. Zimbelmann, A contribution to the problem of cement-aggregate bond,
[25] C. Le Belle
go, B. Ge
rard, G. Pijaudier-Cabot, Chemo-mechanical effects in Cem. Concr. Res. 15 (1985) 801e808.
mortar beams subjected to water hydrolysis, J. Eng. Mech. 126 (2000)