See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/287508539

Interfacial properties of modeled recycled

aggregate concrete modified by carbonation
Article in Construction and Building Materials February 2016
Impact Factor: 2.3 DOI: 10.1016/j.conbuildmat.2015.12.077

READS

19

4 authors, including:
Jianzhuang Xiao

Li Long

Tongji University

Tongji University

97 PUBLICATIONS 1,326 CITATIONS

5 PUBLICATIONS 17 CITATIONS

SEE PROFILE

SEE PROFILE

Available from: Li Long

Retrieved on: 04 July 2016

Construction and Building Materials 105 (2016) 307320

Contents lists available at ScienceDirect

Construction and Building Materials

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

Interfacial properties of modeled recycled aggregate concrete modified

by carbonation
Chunhui Wang, Jianzhuang Xiao , Guanzhen Zhang, Long Li
Department of Structural Engineering, College of Civil Engineering, Tongji University, Shanghai 200092, PR China

h i g h l i g h t s

Modeled recycled coarse aggregate (MRCA) was designed and modified by carbonation.

Interfacial properties of modeled recycled aggregate concrete (MRAC) were studied.

Both test and simulation results confirmed the effect of carbonation modification.

a r t i c l e

i n f o

Article history:
Received 29 September 2015
Received in revised form 12 November 2015
Accepted 11 December 2015

Keywords:
Modeled recycled coarse aggregate (MRCA)
Modeled recycled aggregate concrete
(MRAC)
Carbonation
Interfacial properties
Push-out test
Simulation study

a b s t r a c t
Modeled recycled coarse aggregates (MRCAs) were designed with different water to cement ratios
(w/c) of old hardened mortar (OHM). Each type of MRCAs was modified by carbonation. All carbonated
and uncarbonated MRCAs were used to cast modeled recycled aggregate concrete (MRAC) specimens.
Push-out tests were conducted to investigate the influence of MRCA carbonation modification on interfacial properties of MRAC. Based on the loaddisplacement curves, the peak load and peak displacement
were mainly analyzed to assess the effect of carbonation modification. The experimental results show
that the peak load increased while peak displacement decreased as the MRCA was intensified by carbonation. The effect of carbonation modification was more obvious at a higher w/c. With the increase of w/c
of new hardened mortar (NHM) or OHM, the corresponding peak load decreased, whereas the peak displacement tended to increase at first and then decrease. A simulation study was also conducted by software ABAQUS and verified by the experimental results. The parametric study confirms that interfacial
properties of MRAC significantly depend on carbonation depth, distribution of OHM and shape of MRCA.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Gravel and sand, which account for most composites in concrete, have been exploited and used in construction industry
widely. However, with the economic development of China, production of concrete is consuming lots of natural resources every
year. Besides, China is also faced with the problem of waste concrete. Most of these waste concrete results from demolition of
old buildings or disasters like earthquakes. It is reported that
approximately 200 million tons of waste concrete are currently
produced annually in the mainland of China [1], which leads to
encroachment of lands and potential soil damage. To resolve the
issue of waste concrete and shortage of natural resources, recycled
aggregate concrete (RAC) technique comes into being as an ecological method.

Corresponding author.
E-mail address: jzx@tongji.edu.cn (J. Xiao).
http://dx.doi.org/10.1016/j.conbuildmat.2015.12.077
0950-0618/ 2015 Elsevier Ltd. All rights reserved.

Compared with natural aggregate concrete (NAC), the

mechanical properties of RAC are more difficult to predict. RAC is
characterized as a multi-phase material with several different representative scales: at macroscopic scale, RAC could be regarded as
a homogeneous material; whereas at mesoscopic scale, it is composed of coarse aggregates, old interfacial transition zones (ITZs),
OHM, new ITZs and NHM [2], as shown in Fig. 1. Compared with
natural coarse aggregates (NCAs), due to the existence of adhered
OHM, recycled coarse aggregates (RCAs) have the characteristics
of high porosity, low density, high water absorption and low
strength, which are traditionally considered to lead to a reduction
of the mechanical properties of RAC [35]. Whats more, by existing investigations, ITZ is even more porous than OHM. It has been
proved that old ITZs between aggregates and OHM, as well as new
ITZs between OHM and NHM have a great influence on the
mechanical properties of RAC [6,7]. Therefore, it is an effective
way to improve the mechanical behaviors of RAC by improving
the properties of ITZs. In recent years, many investigators have

308

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

Table 1
Mix proportion of mortar.

NCA
Old hardened
mortar
New hardened
mortar

Unit mass (kg/m3)

Strength grade

w/c

M20
M30
M40

0.68
0.45
0.37

Cement

Water

Sand

529
680
835

353
333
316

1224
1098
974

Old ITZ
New ITZ
Fig. 1. Variant phases of RAC.

NCA
Old hardened
mortar
New hardened
mortar
Old ITZ

(Ca(OH)2) dissolved in the concrete pore water, producing calcium

carbonate (CaCO3) and water (H2O). In addition, hydrated
calcium silicate (CSH), un-hydrated tri-calcium-silicate (C3S) and
bi-calcium-silicate (C2S) consume carbon dioxide (CO2) as well.
Carbonation can effectively improve the compactness of cement
mortar, decrease the porosity and water absorption, and increase
the density and strength of concrete containing RCA [18,19].
Furthermore, the carbonation approach has the advantage of lower
cost and being more environmental friendly compared with other
approaches [20]. Besides, although there are variations in different
calculating models of carbonation depth for concrete proposed by
many researchers [2123], the basic fact is widely accepted that
carbonation occurs only in the surface layer under atmospheric
conditions during the life of structures. Carbonation depth can be
predicted by equation proposed by Jiang et al. [23]:

s
p
W=cc C  0:34
xc 8391  RH
n0  t
cHD cc C

New ITZ

1:1

Fig. 2. Schematic of MRAC.

engaged in studies on the modification of RCA. Du et al. [8]

reported that specific chemical grout intensified the RCA, proving
that it is feasible to modify the RCA. Wan et al. [9] concluded that
the compactness of ITZ can be improved by decreasing the w/c and
adding appropriate amount of fly ash. Tam et al. [10] firstly proposed that the two-stage mixing approach (TSMA) could increase
the compressive strength and decrease the strength variability.
Kong et al. [11] reported that three-stage mixing approach, namely
enveloping volcanic ash around the RCA, improved the microstructure at ITZ of RAC. Zhu et al. [12] found that the mechanical
properties of particle shaped RCA were much higher than those
of simply-crushed RCA. Noguchi et al. [13] confirmed that microheating technique effectively removed the adhered OHM to
improve the quality of RCA. Shah and Hou et al. [14,15] revealed
that nanoSiO2 could speed up the hydration process of cement
and therefore increase the early-age strength of cement paste.
Shah and Konsta-Gdoutos et al. [16,17] reported that elastic modulus and fracture energy were significantly increased if carbon
nano-tube was added to mortar.
Carbonation is a neutralizing procedure, in which carbon
dioxide (CO2) in the atmosphere react with calcium hydroxide

where xc is the predicted carbonation depth, RH is the relative

humidity, W and C are unit amount of water and cement, cc is an
adjusting coefficient about type of cement, cHD is an adjusting coefficient about degree of cement hydration, no is the CO2 concentration, and t is carbonation time. For concrete with 0.45 w/c, the
average carbonation depth is about 7.2 mm in 50 years under atmospheric conditions. Considering the average dimension of concrete
structure elements, the carbonation depth can be ignored. It means
that most mortar, part of which becomes OHM of RAC, remains
uncarbonated. Therefore it is predictable that carbonation might
be an effective approach to improve the mechanical properties of
RCA by improving the interfacial properties.
MRAC proposed in 2012 by Xiao et al. [24] is a simplified analysis model of RAC. The existing researches have demonstrated that
MRAC can be used to investigate the relation between the
mesostructure of each phase and mechanical behavior of RAC
[2,24]. In this model, the NCA is simplified as a regular cylinder
while the OHM is simplified as thin cylindrical vessel surrounding
the NCA. The cylindrical NCA is called as modeled natural coarse
aggregate (MNCA), and the cylinder with adhered OHM is named
as modeled recycled coarse aggregate (MRCA). Specimens that
are cast by MRCA and mortar are termed as MRAC [6], as shown
in Fig. 2.

New hardened
mortar

0
R2
R15

120

Old hardened
mortar

Modeled
aggregate

30

Mortar
40

(a) MRCA

(b) Mortar specimen

30
120

40

(c) MRAC

Fig. 3. Geometric dimensions of specimens.

309

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) Molds

(b) Poured plates

Fig. 4. Molds and poured plates used in producing MRCA.

Table 2
MRCAs and corresponding MRAC specimens.

No. of
MRCA

w/c of
OHM

Whether the OHM

was carbonated

No. of
MRAC

w/c of
NHM

Quantity of
specimens

MRCA1
MRCA2
MRCA3
MRCA4
MRCA5
MRCA6
MRCA7
MRCA8
MRCA9
MRCA10

0.37
0.37
0.68
0.68
0.45
0.45
0.45
0.45
0.45
0.45

Carbonated
Uncarbonated
Carbonated
Uncarbonated
Carbonated
Uncarbonated
Carbonated
Uncarbonated
Carbonated
Uncarbonated

MRAC1
MRAC2
MRAC3
MRAC4
MRAC5
MRAC6
MRAC7
MRAC8
MRAC9
MRAC10

0.45
0.45
0.45
0.45
0.37
0.37
0.45
0.45
0.68
0.68

4
4
4
4
4
4
4
4
4
4

Fig. 7. Push-out test device.

by researchers for its advantage that in the test the interfacial

strength from actual composite can be measured directly. In this
study, MRAC was used both in push-out test and numerical simulation to investigate the effect of carbonation modification on
interfacial properties of RAC.
2. Research significance

Fig. 5. The accelerated carbonation test device. 1-Carbonation chamber 2-MRCAs

and mortar specimens 3-Desiccant 4- CO2 gas storage 5-Value 6-Flow regulator 7Safety value 8-Temperature and humidity inductor 9-Piezometer 10-Vacum pump
11-Data processor.

It is well known that the mechanical properties such as compressive strength and elastic modulus of RAC are lower than those
of NAC with the same w/c. One of the major reasons for the lower
strength exhibited by RAC is the increase of weak ITZs. How to
improve the interfacial properties effectively and environmental
friendly becomes a great concern in this field. By the new concept
of MRAC, push-out test and numerical simulation, this paper tries
to reveal the effect of RCA carbonation modification on the interfacial properties of RAC. This investigation will be useful for the
improvement of RAC and its application in civil engineering.
3. Experimental program
3.1. Specimen design
MRCA displayed in Fig. 3(a) and mortar specimen as a reference group for
measuring the carbonation depth and verifying the carbonation degree of OHM displayed in Fig. 3(b) were utilized in this study. Mortar with different mix proportions
was used to cast MRAC and mortar specimens. The mortar was marked with M20,
M30 and M40, of which the w/c was 0.68, 0.45 and 0.37 respectively. Based on previous study [6], MRACs were designed as square slabs with dimensions of
120 mm  120 mm  30 mm. Every MRAC had only one MRCA locating in the center of the specimen as shown in Fig. 3(c). The thickness of the OHM was 5 mm.

Fig. 6. Interfacial transition zones in MRAC.

Push-out test is one of meso-mechanical laboratory methods

used in measuring the interfacial properties between different
kinds of materials [25,26]. The test is more and more recognized

3.2. Materials and mix proportions

NCAs with diameters of 30 mm and heights of 30 mm used in the tests were
obtained by drilling granite stones. The fine aggregate used in the tests was river
sand. Mixing water was tap water. An ordinary Portland cement with grade of
42.5 was supplied for this experimental investigation. Table 1 lists the mix proportions of mortar which were designed according to JGJ55-2001 [27].

310

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) 0 day

(b) 7 days

(c) 14 days

(d) 21 days

Fig. 8. Carbonation depths of mortar at different carbonation ages.

Table 3
Mean carbonation depth at different ages.
Age

0 day

Location

Lateral

Bottom

Lateral

Bottom

Lateral

Bottom

Lateral

Bottom

Mean carbonation depth (mm)

10

15

15

(a)

(b)

7 days

14 days

(c)

21 days

(a)

Fig. 9. Crack type I.

(b)

(c)

Fig. 11. Cracks type III.

Before MRCAs were cast, the NCAs were fixed and placed vertically in wood
molds and the mortar intended to represent OHM was then filled around the NCAs.
Mortar specimens were cast simultaneously. After being placed in the laboratory for
24 h, the poured plates were cured in a curing room with 20 2 C temperature and
95% relative humidity for 28 days until testing. Following that, MRCAs with diameters of 40 mm were cored from the plates with mortar adhered to NCAs, as shown
in Fig. 4. Detailed information about all the specimens is presented in Table 2.
3.3. Testing procedure
After curing, both MRCAs and mortar specimens were taken out from the curing
room and were kept in laboratory conditions for a week. Then the samples were dried
in a drying oven under 40 C for 24 h. Following that the specified samples were cured
in an accelerated carbonation test chamber with (20 2)% CO2 concentration,
30 3 C temperature and 70 5% relative humidity [28,29]. The carbonation device
is displayed in Fig. 5. The MRCAs which were not carbonated after being cured for
28 days would be put in the laboratory until other MRCAs were carbonized.

(a)

(b)

Fig. 12. Cracks type IV.

Finally, both carbonated and uncarbonated MRCAs were used to cast MRAC
specimens accordingly. Push-out tests were conducted to investigate the effect of
carbonation modification of MRCA on mechanical response of specimens which
represented the interfacial properties of MRAC, as shown in Fig. 6.

(c)
Fig. 10. Cracks type II.

(d)

311

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(b) MRAC3 and MRAC4

(a) MRAC1 and MRAC2

(c) MRAC5 and MRAC6

(d) MRAC7 and MRAC8

(e) MRAC9 and MRAC10

Fig. 13. Influence of carbonation modification on loaddisplacement curves.

4. Experimental results and discussions

4.1. Carbonation tests results

Fig. 14. Simplified model loaddisplacement curves.

All the tests were conducted in Tongji Structural Laboratory. The measurement
and loading systems are displayed in Fig. 7. A YHD-30 displacement transducer was
installed at the bottom of MRCA which has the accuracy of 0.003 mm. The loading
rate was controlled at 0.05 mm/min. MRCA at the center of MRAC specimen was
pushed out under vertical loading. Failure process was observed. In addition, load
and displacement were monitored and recorded by a computer acquisition system.

After the mortar specimens being split, the fresh broken surface
was sprayed with phenolphthalein which made carbonated zone
colorless and turned uncarbonated zone red to measure the carbonation depths. The carbonation depths for 0 day, 7 days, 14 days
and 21 days were measured, as shown in Fig. 8.
The average carbonation depths at different ages are given in
Table 3. It can be concluded that the carbonation depth of
mortar specimen at 21 days is much larger than the thickness of
adhered OHM, indicating that the OHM of MRCA has been totally
carbonated at 21 days. Modification time taken to make OHM
completely carbonated can be predicted by Eq. (1) proposed by
Jiang et al. [23]. For instance, for MRCA5 with 0.45 w/c it takes
about 13.2 days to carbonize its OHM completely. Under atmospheric conditions, modification time can be roughly estimated
by equation proposed by Gong et al. [22]:

D2 D1

r
t 2  c2
t 1  c1

or simplified as

t2

t1 c1
c2

312

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) Carbonated

(b) Uncarbonated

Fig. 15. Loaddisplacement curves with different w/c of NHM.

c2 are those under atmospheric conditions. It can be inferred that

it should take about 8793.5 days to carbonize OHM of MRCA5,
and it is unpractical in engineering. Therefore accelerated carbonation process is necessary for carbonation modification of RCA.
4.2. Push-out test results
In order to investigate the effect of MRCA carbonation modification on the interfacial properties, push-out tests were carried out.
During the push-out tests, the loaddisplacement curves were
mainly recorded. The experimental analysis was conducted mainly
according to the failure processes and loaddisplacement curves,
especially the peak load and the peak displacement. Peak load is
the maximum push-out load, and the peak displacement is the displacement at the peak load.

Fig. 16. Influence of w/c for NHM on peak load.

Fig. 17. Influence of w/c for NHM on peak displacement.

where D1, t1, c1 are carbonation depth, carbonation time and

CO2 concentration in accelerated carbonation depth while D2, t2,

4.2.1. Failure patterns

The failure patterns are displayed from Figs. 912. In push-out
tests, cracks appeared on the surface of some specimens with the
increasing load. Most cracks initiated at old ITZ and connected with
each other within OHM, and then propagated into mortars until
the specimen failed. However, a small quantity of cracks initiated
within the NHM. Four major types of cracks can be found: crack
type I, II, III and IV, as shown from Figs. 912.
4.2.2. The effect of carbonation modification
Fig. 13 shows the influence of carbonation modification on the
loaddisplacement curves. Although the experiments reveal
variations in the peak load and the corresponding displacement
for both carbonated and uncarbonated MRCA, the similarities of
these curves seem remarkable, as displayed in Fig. 14. In addition,
loaddisplacement curves of MRAC present obvious nonlinear
characteristics. Generally, these curves are composed of ascending
portion, descending portion and residual portion. In this paper,
deformation stiffness is represented as the slope of ascending portion of loaddisplacement curve. It should be mentioned that there
is a relative low stiffness stage in these curves at a low load level.
This is because that the displacement measured by transducer consists of displacement of specimens and additional displacement
which could be induced by the deformation of loading system. It
is illustrated that for MRAC with carbonated MRCA (termed as
MRAC-CA), both peak load and stiffness are higher while the peak

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) Carbonated

313

(b) Uncarbonated

Fig. 18. Load-displacement curves with different OHM w/c.

indicates that MRCA carbonation modification increases both bond

strength and brittleness of old ITZ and OHM. The reason for higher
peak load of MRAC-CA may lie in the higher strength of OHM and a
reduction of porosity around interfaces provoked by carbonation. It
has been proved that a reduction of open porosity of concrete, positive effect of water which is loosed during the carbonation and
acts in the process of hydration and variation in chemistry increase
the strength of mortar [30].

Fig. 19. Influence of OHM w/c on peak load.

Fig. 20. Influence of OHM w/c on peak displacement.

displacement is relatively low when compared to MRAC with

uncarbonated MRCA (termed as MRAC-UA). This phenomenon

4.2.3. The effect of NHMs w/c

Fig. 15 demonstrates the effect of NHMs w/c on loaddisplacement curves. In this section, the w/c of OHM was 0.45, and the w/c
of NHM varied from 0.37 to 0.45 to 0.68. The test results illustrate
that an increase of the w/c of NHM leads to a reduction of peak load
as well as the deformation stiffness for both MRAC-CA and MRACUA.
Figs. 16 and 17 demonstrate the variation of the peak load and
peak displacement being as a function of w/c for NHM respectively.
It can be seen that the increasing w/c of new mortar leads to a linear reduction of the peak load for MRAC-CA and MRAC-UA. The
peak load of MRAC-CA is always higher than that of MRAC-UA.
Furthermore, with the increase of w/c of NHM, the effect of
carbonation modification on peak load of MRAC is becoming more
notable. For instance, for MRAC with 0.68 w/c of NHM, the peak
load of MRAC-CA is 16.1% higher than that of MRAC-UA; whereas
for MRAC with 0.37 w/c of NHM, the peak load of MRAC-CA is only
1.2% higher than that of MRAC-UA. It means carbonation can
undoubtedly improve the interfacial properties of MRCA, especially
when the w/c is relatively high. This phenomenon may be caused
by improvement of compactness and porosity at interface between
MRCA and NHM, and this improvement may be more remarkable
when the w/c of new mortar is higher. From Fig. 17, the peak displacement of MRAC-UA is always larger than that of MRAC with
carbonated MRCA. For MRAC-UA, the increase of w/c for NHM
increases the peak displacement. For MRAC-CA, when the w/c of
new cement mortar is less than 0.45, the peak displacement
increases with the increase of the w/c of new cement mortar. When
the w/c of new cement mortar exceeds 0.45, the tendency of variation of peak displacement is in the opposite case. This may result
from the significant decrease of peak load of MRAC when the w/c of
NHM varies from 0.45 to 0.68.

314

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

3.5

50
45

3.0

Mortar matrix
ITZ

2.0
1.5
1.0

Mortar matrix
ITZ

40
35

Stress (MPa)

Stress (MPa)

2.5

30
25
20
15
10

0.5

0.0
0.0000 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006

Strain

(a) Uniaxial tensile

0
0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014

Strain

(b) Uniaxial compression

Fig. 21. Constitutive relation of ITZ and mortar matrix [33].

Table 4
Mechanical parameters of various phases of MRAC.
Material

Thickness (lm)

Elastic modulus (GPa)

Poisson ratio (m)

NCA
OHM
NHM
Old ITZ
New ITZ

50
60

70.0
25.0
23.0
20.0
18.0

0.16
0.22
0.22
0.20
0.20

Fig. 23. Mesh generation.

according to the w/c, namely, with higher w/c the influence is more
notable. Peak displacement increases at first and then decreases
with the increasing w/c of OHM for MRAC-CA and MRAC-UA. The
peak displacement of MRAC-CA is smaller than that of MRAC-UA,
indicating that brittleness was improved by carbonation.
Fig. 22. FEM of MRAC.

5. Finite element method (FEM) analysis and parameter study

4.2.4. The effect of OHMs w/c
Fig. 18 shows the effect of OHMs w/c on loaddisplacement
curves. In this section, the w/c of OHM varied from 0.37 to 0.45
to 0.68. It is observed that not only the peak load but also the
deformation stiffness of MRAC decreases as the w/c of OHM
increases for both MRAC-CA and MRAC-UA.
Figs. 19 and 20 demonstrate the variations of the peak load and
peak displacement with w/c of OHM respectively. The peak load
decreases with the increasing w/c of OHM. The peak load of
MRAC-CA is always higher than that of MRAC-UA within the range
of w/c of OHM from 0.37 to 0.68. Furthermore, with the increase of
w/c of OHM, the effect of carbonation on peak load of MRAC is
becoming increasingly obvious. In detail, when the w/c of OHM
equals 0.37, the peak load increases by 2.7% due to modification,
while in the condition of 0.68 w/c of OHM, the increasing rate is
20.9%. It should be mentioned that this feature is consistent with
results for different w/c of NHM. The phenomenon can be
explained by the experimental results of Kim et al. [31] that the
influence of carbonation on the strength varies considerably

5.1. FEM simulation and test verification

5.1.1. Constitutive relationships of materials
In this numerical study, MRAC is considered as a cementitious
composite including NCA, old ITZ, OHM, new ITZ and NHM. The
constitutive relationship of each phase was defined separately.
5.1.1.1. Mortar matrix. Mortar matrix shows softening behavior
after reaching its peak tensile stress and compressive stress. For
describing this behavior, former investigators have developed a
damage plasticity model for mortar matrix [3234] which is
widely used in simulating the mechanical properties of concrete
under monotonic loading, cyclic loading and dynamic loading. In
this simulation, damage plasticity model was used for the constitutive relationship of mortar matrix. Elastic modulus, the strength
and the deformation capacity are obtained according to test data
from Refs. [3537]. The uniaxial tensile and compressive stress
strain relationships of mortar matrix are shown in Fig. 21 [33].

315

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) MRAC1

(b) MRAC2

(c) MRAC3

(d) MRAC4

(e) MRAC5

(f) MRAC6

(g) MRAC7

(h) MRAC8

(i)

(j)

MRAC10

Fig. 24. Comparison of loaddisplacement curves in the test and simulation.

MRAC9

316

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

Table 5
Comparison of test results and simulation results.
No.

Peak load

MRAC1
MRAC2
MRAC3
MRAC4
MRAC5
MRAC6
MRAC7
MRAC8
MRAC9
MRAC10

Peak displacement

Test results (kN)

Simulation results (kN)

Variation rate (%)

Test results (mm)

Simulation results (mm)

Variation rate (%)

11.815
11.500
9.065
7.500
11.644
11.502
10.991
10.534
9.171
7.898

12.286
11.700
9.526
7.914
12.296
11.677
11.768
11.135
8.989
8.589

3.99
1.74
5.09
5.52
5.60
1.52
7.07
5.71
1.98
8.75

0.303
0.377
0.399
0.414
0.328
0.369
0.407
0.430
0.357
0.430

0.291
0.335
0.375
0.355
0.301
0.335
0.373
0.414
0.341
0.410

3.96
11.14
13.53
14.25
8.23
9.21
8.35
3.72
4.48
4.65

(a) Numerical (low load level)

(b) Experimental (failure pattern)

(c) Numerical (failure)

Fig. 25. Simulated concentration of plastic strain and cracks of MRAC8.

OHM

NHM

Aggregate

Old ITZ New ITZ

Fig. 26. Minimum principal stress contour map of MRAC8.

5.1.1.2. ITZs. Because of the difficulties in determining the properties of ITZ with traditional testing method, the comprehension of
the mechanical properties and thickness of ITZ was confined. Based
on the nanoindentation test results reported by Xiao et al. [38], the

Fig. 27. Influence of carbonation depth.

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) = 0.224 mm

317

(b) = 0.420 mm

Fig. 28. Plastic strain contour map of MRAC (r = 0.5).

A region

B region

(a) MRAC-H

(b) MNAC

Fig. 29. Modeled concrete schematics.

(a) = 0.246mm

MRAC8
20

MRAC-H
MNAC

Load (kN)

15

10

(b) = 0.400mm
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2

Fig. 31. Plastic strain contour maps of MRAC-H.

Displacement (mm)
Fig. 30. Influence of distribution of OHM.

elastic modulus and strength of old ITZ and new ITZ are assumed to
be 80% and 85% relative to those of related mortar matrix respectively. Damage plasticity model was used for the constitutive relationship of ITZ in this simulation. It is similar to mortar matrix, but
the strength and elastic modulus are lower relative to related mortar matrix, as shown in Fig. 21. Material parameters of carbonated
OHM and ITZs are adjusted based on the test results by Kim et al.
[31] which revealed effect of carbonation on the rebound number
of concrete. Since it is generally assumed that there is a proportional relation between the rebound number measured by a Schmidt rebound hammer and compressive strength of concrete, in
this study, when w/c is 0.68, 0.45 and 0.37, the increase of compressive and tensile strength of mortar and ITZ after carbonation
are 22%, 10% and 3%, respectively.

5.1.1.3. NCA. The compressive strength of NCA exceeds 150 MPa in

general. As no crack or damage was observed in the NCA during
loading in laboratory tests, in this study, NCA was modeled as a
linear-isotropic material. The mechanical parameters of each phase
in MRAC, which were obtained in the experimental results, are
listed in Table 4. The Poissons ratios of old ITZ and new ITZ are
determined according to the study by Ramesh et al. [39].

5.1.2. FEM establishment

The software ABAQUS was used to establish the FEM. Due to the
symmetry of MRAC specimens, the model was simplified as one
quarter of MRAC in order to reduce the calculation time without
losing accuracy. In order to compare with the test results, each
phase in FEM had the same geometric dimensions with experimental MRAC specimens. The ITZ was modeled by a thin cylindrical
vessel, and the thickness of old ITZ and new ITZ were defined as

318

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

Fig. 32. Plastic strain contour map of MNAC.

Old hardened
mortar

Fig. 34. Influence of aggregate shape.

New hardened
mortar

120

Modeled
aggregate
5 24 5

Fig. 33. MRAC-C.

50 lm and 60 lm respectively according to the former study by

Xiao et al. [38]. FEM of MRAC is displayed in Fig. 22. 8-Node linear
brick, reduced integration elements (C3D8R) were used to mesh all
the phases of FEM, as shown in Fig. 23. ABAQUS/explicit solver was
adopted for calculation.
The model was subjected to a uniformly distributed displacement at the top surface of NCA. A displacement-controlled loading
scheme was used, and the maximum displacement was set as
2 mm in order to obtain the entire loaddisplacement curves.
The boundary conditions of X-plain and Y-plain were defined as
XSYMM (U1 = UR2 = UR3 = 0) and YSYMM (U2 = UR1 = UR3 = 0) in
ABAQUS.
5.1.3. Test verification
According to above parameters and descriptions, push-out tests
were simulated by above FEM and loaddisplacement curves were
obtained. In order to validate the FEM, the loaddisplacement
curves of simulation results were compared with those of test
results, as demonstrated in Fig. 24. In addition, the peak load and
peak displacement of test results are compared with those of simulation results in Table 5. It can be illustrated that the numerical
results agree well with the test results. Being similar to experimental curves, the simulation curves also include ascending portion,
descending portion and residual portion. However, the simulated
peak load is slightly higher than the test results. Furthermore,
the numerical peak displacement was smaller than that of test
results, perhaps mainly due to the absence of low stiffness stage
in test curves at low load level.
In addition, the model also simulated crack initiation and propagation correctly. In this study, maximum principal plastic strain

was utilized as damage parameter to analyze the crack generation

and propagation as well as failure patterns. Fig. 25 displays the
comparison of plastic strain distribution of numerical results and
cracks in the test for MRAC8. It is observed that the micro-cracks
occurred at old ITZs, and then connected with each other within
OHM, which agrees well with the experimental observation. With
the increase of load level, the cracks propagated into new cement
mortar until the specimen was split or the bond around ITZs totally
failed. Minimum principal stress contour map of MRAC8 along the
section (y = 0) is presented in Fig. 26. It can be seen that the stress
was transmitted to supports through old ITZ, OHM, new ITZ and
NHM. Low bond strength of ITZs may be the fundamental cause
of low peak load of MRAC. Overall, the FEM can simulate the
push-out tests correctly and predict remarkably well the failure
pattern of MRAC.
5.2. Parametric analysis
In this section, a parametric analysis was performed to investigate the effect of carbonation depth, distribution of OHM and
shape of MRCA on interfacial properties. MRAC8 was used as the
base-case.
5.2.1. Effect of carbonation depth
In practical engineering, due to the limitation of modification
time, not all RCA can be modified by carbonation completely. Therefore, it is necessary to investigate the effect of carbonation depth on
interfacial properties of RAC. For partly carbonated MRCA, old ITZ
can be considered as uncarbonated, whereas the OHM should be
separated as uncarbonated region (close to aggregate) and carbonated region (close to new mortar). As the thickness of OHM is 5 mm
in MRCA, the ratio (r) of carbonation depth to the thickness of OHM
varied from 0 to 0.5 to 1. Simulated loaddisplacement curves are
displayed in Fig. 27. It is demonstrated that the curve with r = 0.5
has a discrepancy when compared with the curves with r = 0 and
1 that there is a yielding portion. Model predictions show that the
peak load of MRAC increases as the ratio (r) increases. When r
equals 0, 0.5 and 1, the peak load are 11.135 kN, 11.479 kN and
11.768 kN, respectively. However, the peak displacement with
r = 0.5 was close to that with r = 0.
Plastic strain contour maps when r = 0.5 are shown in Fig. 28. In
the case of r = 0.5, cracks firstly initiated at old ITZ, and connected
with each other within uncarbonated region of OHM. Following
that the cracks propagated through carbonated region and

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(a) = 0.224 mm

319

(b) = 0.380 mm

Fig. 35. Plastic strain contour maps of MRAC-C.

extended into NHM. As the OHM was partially modified by carbonation, unlike the MRAC with r = 0, the specimen with r = 0.5 had
only several macro-cracks in carbonated region. Overall, simulation results prove that carbonation depth influences the interfacial
properties and crack propagations of RAC, and the interfacial properties can be improved even though the adhered OHM is partly
carbonated.
5.2.2. Effect of distribution of OHM
During the production of RCA, inhomogeneity of adhered OHM
is inevitable. For instance, some RCAs are all covered by OHM,
whereas some are partly covered. In this section, in order to investigate the effect of this difference, MRAC with MRCA half covered
by OHM (MRAC-H) and concrete with only MNCA (MNAC) are
designed and push-out tests are simulated. And then the calculated
results are compared with those of MRAC8.
Fig. 30 shows the loaddisplacement curves of different specimens. Compared to MRAC8, the tendency of loaddisplacement
curve of MRAC-H is obviously different. It is observed that the peak
load of MRAC-H is lower than that of MRAC8. It means that inhomogeneity of adhered OHM has adverse influence on the interfacial
property. For better discussion, in MRAC-H, the side covered by
OHM is termed as A region, and no covered side is termed as B
region, as presented in Fig. 29. For MRAC-H, when the displacement exceeded 0.246 mm, the load at first decreased and then
increased with the increase of displacement. As the displacement
reached 0.400 mm, the load of MRAC-H reached its peak and then
fell down rapidly. From the plastic strain contour map shown in
Fig. 31, it can be observed that plastic strain concentration generated at ITZs in both A and B region and propagated in the direction


of h 45 and h 45 respectively. Internal stress was redistributed as the cracks propagated. With the increase of displacement, a new crack appeared near the border between A and B


region and propagated in the direction of h 0 when the specimen came to fail.
MNAC behaved similar to MRAC8, but a 76% higher peak load
appeared. Besides, the initial deformation stiffness was also higher
when compared with MRAC8. Plastic strain contour map of MNAC
before failure is presented in Fig. 32. The continuous mortar matrix
of MNAC provided a continuous path that transmitted the external
load from MNCA to supports continuously, which led to the higher
peak load of MNAC. It is obvious that distribution of mortar matrix
plays a role in determining the mechanical properties of RCA.
5.2.3. Effect of MRCA shape
Considering the wide variety of source of RCA, there are manifold shapes of RCAs in RAC. MRAC with cuboid aggregate (MRACC) was designed to investigate the effect of RCA shape on the

push-out results, as shown in Fig. 33. The area of old ITZ of

MRAC-C was designed close to that of MRAC8 for better
comparison.
Compared with MRAC8, the peak load and peak displacement of
MRAC-C are both lower, as shown in Fig. 34. Furthermore, the tendency of loaddisplacement curve of MRAC-C has a remarkable discrepancy from that of MRAC8. For MRAC-C, there exists an obvious
yielding stage when the displacement exceeds the peak displacement. It can be observed from Fig. 35 that cracks along the direction


of h 45 occurred earlier at the corner of MNCA than MRAC8, prior

to connecting with each other within OHM. Lateral deformation of
mortar relative to aggregate resulted in tensile and tensile-shear
bond failure at the corner of the aggregate. The load reached its
peak value at the displacement of 0.224 mm. At that time, no obvious strain concentration was observed except the cracks along


h 45 and around old ITZ, indicating that the specimen partly

remained its carrying capacity. This explains the existence of yielding stage of loaddisplacement curve. The MRAC-C lost its carrying
capacity when the cracks in OHM connected with each other
because the load transmission path was cut off. Therefore, it can
be concluded that the interfacial properties of MRAC and failure
process significantly depends on the MRCA shape.
6. Conclusion
Modeled recycled aggregate concrete (MRAC) specimens were
designed to investigate the effect of modeled recycled coarse
aggregate (MRCA) carbonation modification on interfacial properties of MRAC by push-out tests. Failure patterns and loaddisplacement curves were analyzed, and the effect of MRCA carbonation
modification on loaddisplacement curves were investigated. A
simulation study was also undertaken by software ABQUS and verified by the experimental results. The parametric study was conducted to investigate the influence of the carbonation depth,
distribution of old hardened mortar (OHM) and shape of recycled
coarse aggregate (RCA) on interfacial properties. Based on the
experimental and numerical study, some conclusions are summarized as follows:
(1) Carbonation can improve the interfacial properties of MRAC.
Particularly, the effect of modification is more significant
when the w/c of both new hardened mortar (NHM) and
OHM is higher. On the other hand, carbonation can decrease
the peak displacement of MRAC.
(2) The variation of interfacial properties is analyzed with the
change of w/c of NHM and OHM. For both carbonated and
uncarbonated specimens, an increase of w/c of NHM and
OHM leads to a reduction of peak load.

320

C. Wang et al. / Construction and Building Materials 105 (2016) 307320

(3) Most cracks initiated at ITZs and propagated into mortars.

There are mainly four types of failure patterns in the pushout tests.
(4) The finite element method (FEM) is capable of simulating
the loaddisplacement curves correctly, as well as the plastic strain concentration and failure mechanism. The parametric study confirms that interfacial properties of MRAC
significantly depend on carbonation depth, distribution of
OHM and shape of MRCA.
(5) This carbonation modification approach can provide an
effective, environmental friendly and low-cost method for
enhancing the interfacial properties of RAC, and thus, the
approach opens up a wider scope of RAC applications.

Acknowledgments
The authors wish to acknowledge the financial support from the
National Natural Science Foundation of China (NSFC) (Project No:
51325802, 51438007).
References
[1] J.Z. Xiao, W.G. Li, Y.H. Fan, X. Huang, An overview of study on recycled
aggregate concrete in China (19962011), Constr. Build. Mater. 31 (6) (2012)
364383.
[2] J.Z. Xiao, Q. Liu, W.G. Li, On the micro- and meso-structure and failure
mechanism of recycled concrete, J. Qingdao Technol. Univ. 30 (4) (2009) 2430
(in Chinese).
[3] T.C. Hansen, Recycled aggregate and recycled aggregate concrete. Second state
of art report, developments from 19451985. RILEM Technical Committee 37
DRC, Mater. Struct. 19 (3) (1986) 201246.
[4] P.J. Nixon, Recycled concrete as an aggregate for concrete a review, Mater.
Struct. 11 (5) (1978) 371378.
[5] J.Z. Xiao, Experimental investigation on complete stress-strain curve of
recycled concrete under uniaxial loading, J. Tongji Univ. (Nat. Sci.) 35 (11)
(2007) 14451449 (in Chinese).
[6] W.G. Li, J.Z. Xiao, J.Q. Yuan, Stress distribution characteristics of modeled
recycled aggregate concrete under uniaxial compression, J. Tongji Univ. (Nat.
Sci.) 40 (6) (2012) 906913 (in Chinese).
[7] C.S. Poon, Z.H. Shui, L. Lam, Effect of microstructure of ITZ on compressive
strength of concrete prepared with recycled aggregates, Constr. Build. Mater.
18 (6) (2004) 461468.
[8] T. Du, H.Q. Li, Investigation on mechanical properties of intensified recycled
aggregate concrete, China Concr. Cem. Prod. 2 (2003) 1920 (in Chinese).
[9] H.W. Wan, J.L. Xu, Z.H. Shui, J. Jiang, Study on the structure and properties of
interfacial transition zone (ITZ) of the regenerated concrete, J. Wuhan Univ.
Technol. 26 (11) (2004) 2932 (in Chinese).
[10] V.W.Y. Tam, X.F. Gao, C.M. Tam, Microstructural analysis of recycled aggregate
concrete produced from two-stage mixing approach, Cem. Concr. Res. 35 (6)
(2005) 11951203.
[11] D.Y. Kong, T. Lei, J.J. Zheng, C.C. Ma, J. Jiang, J. Jiang, Effect and mechanism of
surface-coating pozzalanics materials around aggregate on properties and ITZ
microstructure of recycled aggregate concrete, Constr. Build. Mater. 24 (5)
(2010) 701708.
[12] Y.G. Zhu, J.Q. Han, Q.Y. Li, X.Q. Dai, Experimental research on the influence by
particle shaping on mechanical performance of recycled aggregate concrete, J.
Qingdao Technol. Univ. 30 (4) (2009) 115118 (in Chinese).
[13] T. Noguchi, R. Kitagaki, Environmental impact minimization and performance
maximization in concrete recycling, in: 2rd International Conference on Waste
Engineering and Management, ICWEM 2010 RILEM Proceedings, 73, pp. 60
17.

[14] S. Kawashima, P. Hou, D.J. Corr, S.P. Shah, Modification of cement-based

materials with nanoparticles, Cement Concr. Compos. 36 (1) (2013) 815.
[15] P. Hou, K.J. Wang, J.S. Qian, S. Kawashima, D.Y. Kong, S.P. Shah, Effects of
colloidal nanoSiO2 on fly ash hydration, Cement Concr. Compos. 34 (10) (2012)
10951103.
[16] S.P. Shah, M.S. Konsta-Gdoutos, Z.S. Metaxa, Highly dispersed carbon
nanotude-reinforced cement-based materials, Cem. Concr. Res. 40 (7) (2009)
10521059.
[17] M.S. Konsta-Gdoutos, Z.S. Metaxa, S.P. Shah, Multi-scale mechanical and
fracture characteristics and early-age strain capacity of high performance
carbon nanotube/cement nanocomposites, Cement Concr. Compos. 32 (2)
(2010) 110115.
[18] B.J. Zhan, C.S. Poon, C.J. Shi, CO2 curing for improving the properties of concrete
blocks containing recycled aggregates, Cement Concr. Compos. 42 (9) (2013)
18.
[19] Y.H. Fang, Y.T. Zhang, X.Y. Mo, Influence of carbonation on the microstructure
and permeability of hardened cement paste and mortar, J. Hohai Univ. (Nat.
Sci.) 33 (1) (2005) 104107 (in Chinese).
[20] S. Monkman, Y. Shao, Integration of carbon sequestration into curing process
of precast concrete, Can. J. Civ. Eng. 37 (2) (2010) 302310.
[21] V.G. Papadakis, C.G. Vayenas, M.N. Fardis, Fundamental modeling and
experimental investigation of concrete carbonation, ACI Mater. J. 88 (4)
(1991) 363373.
[22] L.S. Gong, M.Q. Su, H.L. Wang, Experimental research on multi-factor equation
of concrete carbonation, Build. Sci. 03 (1986) 2938 (in Chinese).
[23] L.X. Jiang, Y. Zhang, Y.Q. Liu, et al., Calculating method of concrete carbonation
depth and experimental research, Concrete 04 (1996) 1217 (in Chinese).
[24] J.Z. Xiao, W.G. Li, Z.H. Sun, S.P. Shah, Crack propagation in recycled aggregate
concrete under uniaxial compressive loading, ACI Mater. J. 109 (4) (2012) 451
461.
[25] S. Caliskan, Aggregate/mortar interface: influence of silica fume at the microand macro-level, Cement Concr. Compos. 25 (4) (2003) 557564.
[26] Y.F. Zhang, H. Liu, L. Qi, Simulation of fiber pushing-out test, J. Wuhan Univ.
Technol. (Mater. Sci. Ed.) 6 (2009) 965969.
[27] MOHURD, JGJ55-2011 Specification for Mix Proportion Design of
Ordinary Concrete, China Architecture and Building Press, Beijing, 2011 (in
Chinese).
[28] Y.M. Zhang, J. Ming, Accelerated carbonation test of concrete at 3% and 20%
CO2 volume fraction, J. Build. Mater. 15 (5) (2012) 684689 (in Chinese).
[29] J. Ming, Y.M. Zhang, Accelerated carbonation test of concrete at low CO2
volume fraction, China Concr. Cem. Prod. 3 (2010) 58 (in Chinese).
[30] J. Jan, Physico-mechanical properties of carbonated concrete, Constr. Build.
Mater. 18 (9) (2004) 645652.
[31] J.K. Kim, C.Y. Kim, S.T. Yi, Y. Lee, Effect of carbonation on the rebound number
and compressive strength of concrete, Cement Concr. Compos. 31 (2) (2009)
139144.
[32] U. Cicekli, G.Z. Voyiadjis, R.K.A. Al-Rub, A plasticity and anisotropic damage
model for plain concrete, Int. J. Plast 23 (1011) (2007) 18741900.
[33] S.M. Kim, R.K.A. Al-Rub, Meso-scale computational modeling of the plasticdamage response of cementitious composites, Cem. Concr. Res. 41 (3) (2011)
339358.
[34] J. Lee, G.L. Fenves, Plastic-damage model for cyclic loading of concrete
structures, J. Eng. Mech. 124 (8) (1998) 892900.
[35] D. Corr, M. Accardi, L. Graham-Brady, S.P. Shah, Digital image correlation
analysis of interfacial debonding properties and fracture behavior in concrete,
Eng. Fract. Mech. 74 (1) (2007) 109121.
[36] N. Tregger, D. Corr, L. Graham-Brady, S.P. Shah, Modeling the effect of
mesoscale randomness on concrete fracture, Probab. Eng. Mech. 21 (3) (2006)
217225.
[37] W.G. Li, J.Z. Xiao, Z.H. Sun, S.P. Shah, Failure processes of modeled recycled
aggregate concrete under uniaxial compression, Cement Concr. Compos. 34
(10) (2012) 11491158.
[38] J.Z. Xiao, W.G. Li, D.J. Corr, S.P. Shah, Effects of interfacial transition zones on
the stress-strain behavior of modeled recycled aggregate concrete, Cem. Concr.
Res. 52 (10) (2013) 8299.
[39] G. Ramesh, E.D. Sotelino, W.F. Chen, Effect of transition zone on elastic moduli
of concrete materials, Cem. Concr. Res. 26 (4) (1996) 611622.