You are on page 1of 10

Construction and Building Materials 68 (2014) 667676

Contents lists available at ScienceDirect

Construction and Building Materials


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

Flexural beam fatigue strength evaluation of crushed brick as a


supplementary material in cement stabilized recycled concrete
aggregates
Mahdi M. Disfani a,, Arul Arulrajah a, Hamed Haghighi a, Alireza Mohammadinia a, Suksun Horpibulsuk b
a
b

Swinburne University of Technology, Melbourne, Australia


Suranaree University of Technology, Nakhon Ratchasima, Thailand & CSI Distinguished Geotechnical Engineering Fellow, Swinburne University of Technology, Melbourne, Australia

h i g h l i g h t s
 Crushed brick was cement stabilized with recycled concrete aggregate.
 Evaluation of the fatigue life and fatigue modulus of cement stabilized brick blends.
 Flexural beam, Repeated Load Triaxial, unconned compression and other tests.
 Cement stabilized blends with crushed brick comply with pavement requirements..

a r t i c l e

i n f o

Article history:
Received 19 March 2014
Received in revised form 10 June 2014
Accepted 3 July 2014
Available online 26 July 2014
Keywords:
Recycled concrete aggregate
Crushed brick
Cement stabilization
Repeated Load Triaxial
Fatigue strength

a b s t r a c t
In recent years, efforts have been made by various researchers to explore the sustainable use of Construction and Demolition (C&D) materials as a construction material in civil engineering applications. Recycled
crushed brick is a commonly found material from demolition activities and works to date on this material
in pavement applications have been limited to its usage in unbound pavement layers. This research was
undertaken to evaluate the performance of crushed brick as a supplementary material in cement stabilized recycled concrete aggregates. An extensive suite of tests were undertaken on the crushed brick and
recycled concrete aggregate blends stabilized with 3% cement. The laboratory evaluation comprised pH,
plasticity index, foreign materials content, particle size distribution, linear shrinkage, California Bearing
Ratio, modied Proctor compaction, Repeated Load Triaxial test, Unconned Compressive Strength Test
and Flexural Beam Tests. The cement stabilized blends with up to 50% crushed brick content and 3%
cement were found to have physical properties, which comply with the local state road authority requirements. The results of Repeated Load Triaxial tests indicated the Recycled Crushed Aggregate/Crushed
Brick (RCA/CB) blends performed well with 50% Crushed Brick (CB) content just on the border line for
bound pavement material. Unconned Compression Strengths met the minimum requirement for 7 days
of curing for all blends, while the 28 day strength of the blends also improved signicantly. The results of
the exural beam tests were noted to be consistent with past works with cement stabilized quarry produced crushed rock products. The modulus of rupture and exural modulus for all the cement-stabilized
blends were found to be consistent with the previous works, which indicate that these blends are suitable
for applications such as cement-stabilized pavement subbases. The fatigue life was also within the range
that has been previously reported for quarry materials. The cement-stabilized blends with crushed brick
as a supplementary material with up to 50% brick content and 3% cement were found to have physical
and strength properties, which would comply with road authority requirements.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Traditional quarry materials for construction are becoming
scarce in many developed and developing countries. In recent
Corresponding author.
E-mail address: mmiridisfani@swin.edu.au (M.M. Disfani).
http://dx.doi.org/10.1016/j.conbuildmat.2014.07.007
0950-0618/ 2014 Elsevier Ltd. All rights reserved.

years, there has been strong environmental move to reduce the


expansion of quarries, as our major cities continue to grow, as well
as to explore the sustainable use of Construction and Demolition
(C&D) materials in construction applications. Signicant inroads
have been made in recent years in researching the use of C&D
materials as a valuable resource in applications such as pavement
base/subbase layers, embankments, footpath and other civil

668

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676

engineering infrastructure applications. The reuse of these recycled


materials in civil engineering infrastructure applications will result
in a low carbon solution, considering that recycled materials have
signicant carbon savings, compared with virgin quarried materials. C&D materials are increasingly being used in pavement applications, particularly as a base/subbase material. C&D materials that
have been evaluated and successfully implemented in recent years
in roads, footpath and pipe-bedding applications include recycled
concrete aggregate [1,2], Crushed Brick [2,3], Reclaimed Asphalt
Pavement [47], waste rock [8,9] and waste glass [1013].
In Australia alone, approximately 8.7 million tons of recycled
concrete aggregate (RCA) and 1.3 million tons of Crushed Brick
(CB) are stockpiled annually [14]. Cement stabilization is a popular
option in lightly stabilized pavement bases/subbases in metropolitan roads in major cities worldwide. Approximately 50,000 km of
road network is located in metropolitan Melbourne, in the state of
Victoria, Australia in which the pavement bases/subbases have
been cement stabilized. Traditionally in the past, only high quality
cement-stabilized crushed rock has been used for cement-stabilized pavement bases/subbases in Melbourne.
While research has been undertaken in recent years particularly
with cement-stabilized reclaimed asphalt pavement in pavements
[4], the usage of supplementary materials such as CB in combination with other recycled aggregates has not been studied. Arulrajah
et al. [3] reported that recycled CB performs satisfactorily only at
low moisture levels and suggested blending recycled brick with
binders or other durable recycled materials to enhance its performance in base/subbase applications.
Aside from determination of elastic modulus of cement-stabilized materials, the fatigue properties of cement-stabilized materials are also of importance. Fatigue damage leads to a reduction in
modulus of the cement-stabilized layers, thus affecting pavement
response [15]. Potential methods that can be used for determining
the modulus of cement-stabilized materials include the exural
beam, direct tension, indirect tensile, longitudinal vibration and
the direct compression tests [16]. The longitudinal vibration and
the direct compression tests have also been attempted but were
found to be unsuitable for determining the fatigue properties of
cement-stabilized materials [17].
The indirect tensile test and the exural beam test have been
used in past studies [1821]. However, due to the lack of established test protocols in Australia and other countries to determine
the modulus and fatigue properties of cement-stabilized materials,
the exural beam test is the preferred method for evaluation of
cement-stabilized granular materials. It was also recognized as a
proper design parameter for Australian environment by Austroads
[22]. The exural beam test is a practical test method for the determination of strength, modulus and fatigue life of cement-stabilized
materials. The number of load cycles to reduce exural modulus to
half of the initial modulus is an accepted denition for fatigue life
of laboratory samples [17]. Flexural beam specimens typically rupture shortly after the number of cycles to attain half the initial
modulus is reached. The initial modulus can subsequently be
dened as the mean modulus for the rst 50 cycles of applied load
during the fatigue test. The initial strain is also considered as the
mean strain during the rst 50 load cycles applied during the fatigue test [17].
The prime objective of this research is to evaluate the performance of CB as a supplementary material with RCA in lightly
cement-stabilized pavement base/subbase applications. The development of a laboratory evaluation procedure for these recycled
products as a pavement base/sub-base material would result in
an increased level of condence within industry as to their likely
in-service performance and appropriate application as well as
result in a higher uptake of recycled materials in urban areas
where cement-stabilized subbase pavements are commonly used.

2. Materials and methods


Samples of recycled concrete aggregate (RCA) and crushed brick (CB) were
obtained from a recycling site in the state of Victoria. RCA and CB used in this study
typically comprise graded aggregates up to 20 mm in size. Laboratory tests were
undertaken on prepared samples of cement-stabilized RCA blended with various
contents of CB (RCA/CB). 3% General Portland (GP) cement was used in the
cement-stabilized RCA/CB blends (i.e. 3 g GP cement was added to 100 g of dry
RCA and CB blend). The engineering properties of the cement-stabilized RCA/CB
blends investigated were: 100% RCA (100RCA), 85% RCA blended with 15% CB
(85RCA/15CB), 70% RCA blended with 30% CB (70RCA/30CB) and 50% RCA blended
with 50% CB (50RCA/50CB). All the blending percentages are based on dry mass of
each material.
The laboratory evaluation was conducted to determine the engineering properties of blends included pH, plasticity index, foreign materials content, particle size
distribution, linear shrinkage, California Bearing Ratio (CBR), modied Proctor
compaction, Repeated Load Triaxial (RLT) test, Unconned Compressive Strength
(UCS) Test and Flexural Beam Tests.
pH tests were undertaken in accordance with Australian protocols [23]. Both
samples consisted of material passing 2.36 mm sieve. Plastic limit, liquid limit
and plasticity index tests were undertaken in accordance with Australian standard
[24]. Linear shrinkage of RCA and CB were undertaken according to Australian standard test method [25]. To determine the percentage by mass of foreign material in
the fraction of RCA retained on a 4.75 mm sieve, visual categorization was carried
out according to Victorian state road authority specication [26]. In this method
the foreign material (glass, plastic, wood pieces, ceramic etc.) is carefully separated
manually from a specied dry mass of RCA and then the ratio between mass of foreign material and mass of initial specimen is reported as foreign material content.
Particle size distribution tests were performed with standard sieves with the
aperture sizes of 19 mm, 13.2 mm, 9.5 mm, 6.7 mm, 4.75 mm, 2.36 mm, 1.18 mm,
600 lm, 425 lm, 300 lm, 150 lm and 75 lm [27]. A hydrometer was used to
determine the particle size distribution for particles ner than the 75 lm sieve [28].
Modied Proctor compaction tests were undertaken to determine the maximum dry density and optimum moisture content [29]. Soaked CBR tests were performed with samples prepared at their optimum points (Optimum Moisture
Content, OMC, Maximum Dry Density, MDD) using modied Proctor compactive
effort and tested upon completion of four days soaking condition [30]. Due to high
strength of the cement-stabilized blend samples, the CBR test was carried out using
a high capacity 250kN universal testing machine.
UCS tests were conducted with samples using split molds to ensure the specimens were not disturbed during removal and parallel end faces were maintained
[29,31,32]. UCS specimens were compacted in ve layers of pre-determined mass
using a Proctor compaction machine and a one-piece split mold (modied compaction as per AS1141.51). A portion of the remaining material was dried in an oven for
the determination of moisture content, MC of the sample at the time of compaction.
Three specimens of each blend were kept in a fog chamber for a curing period of
7 days. Additionally, four specimens of each blend were kept in a fog chamber for
a curing period of 28 days. All samples were subjected to 4 h of immersing in water
prior to the UCS test according to Australian Standard test method for UCS [32].
RLT tests were undertaken in accordance with the Australian test protocol [33].
The samples were compacted in a 105 mm diameter mold with the height of
200 mm in 8 layers. Four specimens were prepared for RLT testing with dynamic
compaction method [29]. The automatic (mechanical) compaction apparatus,
which permits a continuous and even compaction mode, was used to produce uniform specimens to specied density and moisture content. All the specimens were
compacted to the target density of 100% MDD and target moisture content of 100%
OMC. The specimens were then air dried back to the target moisture of 70% of the
OMC. After reaching the target moisture content, specimens were wrapped and left
for additional 24 h to assure moisture uniformity within test specimen. The previous experience of authors in RLT testing of C&D material shows that for these coarse
aggregates a period of 24 h is more than adequate to ensure uniform moisture distribution in the specimen [3,8]. The moisture contents at different parts of the specimen were measured and found to be similar. The RLT testing procedure consists of
a permanent strain test followed by a resilient modulus test. The permanent deformation determination characterizes the vertical permanent strain with multiple
loading stages (at different stress conditions) to enable quantication of the effects
of vertical stress on permanent strain in a single test. For the cement-stabilized
RCA/CB blends, a constant 50 kPa conning stress, and three different loading
stages (at specied deviator stresses of 350 kPa, 450 kPa and 550 kPa) were used,
each loading stage involved 10,000 repetitions. A trapezoidal repeated deviator
stress with a total period of 3 s with rise and fall times of up to 0.3 s and load pulse
width of 1 s was used. The resilient modulus determination characterizes the vertical resilient strain response over sixty ve stress conditions using combinations of
applied dynamic vertical and static lateral stresses in the ranges of 100500 kPa and
20150 kPa, respectively. Each stress condition involved 200 load repetitions. The
stresses and stress ratios are increased in small increments to avoid early failure,
which is probable to occur at high stress ratios. Table 1 summarizes the target (at
compaction) and actual (after testing) sample degree of compaction and moisture
content values for each specimen for the RLT tests. Generally, it was possible to
prepare the specimens within the tolerance of 0.9% for density ratio using the

669

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676


Table 1
RLT specimen conditions for cement stabilized RCA/CB blends.
Blend

OMC %

MDD t/m3

Target MC,
(% of OMC)

Actual MC after RLT


test (% of OMC)

Target degree
of compaction (%)

Actual degree
of compaction (%)

100RCA
85RCA/15CB
70RCA/30CB
50RCA/50CB

11.7
11.7
11.7
12.0

2.039
2.007
1.991
1.996

70
70
70
70

56
56
79
70

100
100
100
100

99.7
99.1
99.9
100

dynamic compaction method targeted at 100% OMC. However, it was difcult to


obtain the target moisture contents using the dry-back method. This inaccuracy
in the dry-back method for demolition wastes was also observed in previous studies
by the authors, and is an accepted feature of RLT testing.
Flexural beam testing consisted of 3 stages of testing to determine the exural
strength, exural modulus and fatigue life of the cement-stabilized materials. Four
pairs of beams in total, one for each blend, were prepared for exural strength, exural modulus and fatigue life determination [17,34]. Flexural beam tests were performed on rectangular samples with the dimensions of 400 mm long  100 mm
wide  100 mm high. A rectangular mold with internal dimensions of 400 mm
long  320 mm wide  145 mm high was used to compact the slabs. The compacted slabs were left in the closed mold and covered with a wet cloth and lid to
minimize moisture loss and stored at 23 C for a minimum of 2 days before being
removed from the mold and cured in a fog room at an external laboratory facility.
Each slab was subsequently cut into two beams after a minimum curing period of
14 days to ensure that the slab was strong enough to be cut. All the beams were
cured in a fog room for a total of 28 days.
Fatigue testing was conducted in a controlled stress mode. The rst beam from
the same slab (Beam A) was used to determine the peak load required to break the
beam using the Modulus of Rupture test method [34]. On completion of the modulus of rupture test on the rst beam of each blend, the second beam from the same
slab (Beam B) was used for modulus testing. 40% of the peak load from the previous
test was applied on this second paired beam from the same slab with a haversine
pulse of 1 Hz comprising 250 ms for loading and 750 ms for resting for 100 cycles.
This load was selected to be low enough so as not to damage the sample but high
enough to produce sufcient displacement at the middle of the beam in order to
accurately estimate the tensile strain and consequently resilient (exural in this
case) modulus. In the next stage, beam B was subsequently used for fatigue testing.
The load was increased to 70% of the peak load with an increased frequency of 2 Hz
comprising 250 ms for loading and 250 ms for resting. By denition, the number of
cycles to achieve half the initial modulus is termed as fatigue life. The number of
load cycles to half initial modulus which is dened as fatigue life is usually very
close to the number of load cycles to ultimate failure of the samples for the exural
beam fatigue test [35]. The fatigue life is highly sensitive to the applied load; i.e., the
lesser the applied load, the greater the fatigue life. Therefore, some samples may fail
after a few hundred cycles while some other samples may last for several thousand
cycles.

3. Results and discussion


The pH value for the blends ranges from 11.3 to 12.0, which
indicates that the RCA/CB blends are alkaline by nature. As the clay
content in all the blends was low, the plastic limit and liquid limit
could not be obtained as Atterberg limit is directly related to clay
mineralogy. Thus, the RCA/CB blends are classied as non-plastic
and possess insignicant linear shrinkage.
The foreign material content in RCA was 3.56%, including 2.62%
CB and 0.94% asphalt. The CB samples were visually assessed to
contain up to 70% brick component, with the balance proportions
comprising predominantly of RCA and other foreign materials.
The particle size distributions of RCA/CB blends as determined
from sieve analysis and hydrometer tests along with coefcients
of uniformity and curvature are shown in Fig. 1. The particle size
of CB is slightly larger than that of RCA. As such, when these two
materials are blended, the particle size distribution curves of the
blends are eventually the same. The grading limits of all blends,
except 100% CB, were found to be within the local road authority
specied limit for pavement subbase materials. The RCA/CB blends,
which classied as well graded gravel (GW), were found to comply
with the acceptable gradation envelope for pavement base/
subbase materials
Fig. 2 presents the compaction curves showing optimum moisture content (OMC) and maximum dry density (MDD) of the
RCA/CB blends with 3% GP cement binder. The OMC and MDD
values were found to be consistent for the blends, with only minor
variations. This is because the specic gravities and particle size
distribution (Fig. 1) of CB and RCA, which typically control the
compaction curve, are essentially the same.

100
90
80

100RCA
RCC-Fine
85RCA/15CB
70RCA/30CB

70

50RCA/50CB

60

100CB-Fine

50

Base/Subbase Upper Limit

Passing (%)

100CB
Base/Subbase Lower Limit

40
30
20
10
0
0.001

0.01

0.1

10

Sieve Size (mm)


Fig. 1. Particle size distribution curves of the cement stabilized RCA/CB blends.

100

670

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676


2.05

8
100 RCA

28 curing days

100RCA

7 curing days

2.03

85RCA/15CB

2.01

50RCA/50CB

Average UCS values, MPa

Dry Density, Mg/m3

70RCA/30CB
85RCA/15CB

1.99
50RCA/50CB

1.97
70RCA/30CB

1.95
1.93

Minimum road authority requirement for 7 days cured


cement stabilized samples

1.91
8.5

9.5

10.5

11.5

12.5

13.5

14.5

Moisture Content, %
0

100RCA

Fig. 2. Compaction curves of cement stabilized RCA/CB blends.

85RCA/15CB

70RCA/30CB

50RCA/50CB

Fig. 3a. UCS results of cement stabilized RCA/CB blends.

Table 2 and Fig. 3a summarize the UCS results for all the
cement-stabilized RCA/CB blends. For 7 day curing period, the
mean UCS value obtained was between 4 and 4.63 MPa for the
cement-stabilized RCA/CB blends. The cement-stabilized RCA
blends were therefore found to meet the minimum 7 day mean
UCS value of 4 MPa specied in the local road authority specication [36]. The 28 day curing period was found to lead to an increase
in the mean UCS value for all the blends. This is consistent with
expectations that a longer curing period in a fog chamber would
result in a higher mean UCS value of cement-stabilized materials
[3739]. Although only the 7 day curing period of UCS value is
specied by the local road authority specication [36], the 28 day
curing period tests were undertaken as an extra measure to determine the strength of the cement stabilized RCA/CB blends under

increased curing period. The strength values of samples with


7 days of curing are almost the same for RCA/CB blends. The
stressstrain behavior of selected samples during the UCS test
are also included in Fig. 3b. These samples demonstrated strength
close to 28 day mean strength as reported in Table 2. The axial
strain which was determined by using the initial height was
3.38% for 100RCA which declined to 2.04 for 85RCA/15CB. The
other two samples of 70RCA/30CB and 50RCA/50CB experienced
the axial strain of 2.5% and 2.8% respectively. Similar to the UCS,
the soaked CBR values of cement-stabilized RCA/CB blends are
practically the same, varying between 370 and 505 kPa as shown
in Fig. 4. The high CBR values of cement-stabilized RCA/CB blends
are due to the increase in cementation bond caused by the addition

Table 2
UCS results of cement stabilized RCA/CB blends.
Blend

100RCA

85RCA/15CB

70RCA/30CB

50RCA/50CB

Sample ID

100RCA-28-1
100RCA-28-2
100RCA-28-3
100RCA-28-4
100RCA-7-5
100RCA-7-6
100RCA-7-7
85RCA-28-1
85RCA-28-2
85RCA-28-3
85RCA-28-4
85RCA-7-5
85RCA-7-6
85RCA-7-7
70RCA-28-1
70RCA-28-2
70RCA-28-3
70RCA-28-4
70RCA-7-5
70RCA-7-6
70RCA-7-7
50RCA-28-1
50RCA-28-2
50RCA-28-3
50RCA-28-4
50RCA-7-5
50RCA-7-6
50RCA-7-7

Curing days

28

28

28

28

Strength, MPa
Individual

Mean

StdDeva

6.0
4.9
5.6
4.9
3.8
4.2
4.0

5.35

0.47

4.0

0.17

7.02

0.76

4.0

0.29

5.6

0.63

4.63

0.17

5.2

0.54

4.2

0.28

6.0
6.6
7.9
7.6
3.8
3.6
4.3
4.6
5.8
6.3
5.9
4.8
4.7
4.4
5.6
5.8
4.5
4.8
4.6
4
4

Std Dev: Standard Deviation (how much each individual value differs from the average value).

MC (% of OMC)

Dry density (% of MDD)

92.8
97.7
96
100
100.2
103.1
98.7

97.1
95.4
97.6
96.1
96.1
96.0
96.1

92.3
94.3
97.1
95.7
101.5
100.6
101.8

97.8
95.8
98.0
97.7
97.6
97.7
98.2

96.4
95.7
94.5
95.4
93.2
95.7
98.3

97.0
97.1
99.1
98.1
98.9
99.8
98.3

89.1
93.0
94.8
95.9
96.2
99.9
97.1

98.7
98.5
97.6
98.2
98.2
97.9
97.8

671

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676


8
7

Axial Stress, MPa

6
5
4

50RCA/50CB
70RCA/30CB
85RCA/15CB
100RCA

3
2
1
0

Axial Strain, %
Fig. 3b. Stressstrain behavior of selected 28 days UCS samples.

of 3% GP cement. The UCS and CBR values of a stabilized material at


particular cement content are generally controlled by particle size
and intact strength of unstabilized material. The ne particle needs
more cement volume per contact material grains for the same
cement content, hence the strength of stabilized materials is low
for ner particles [3840]. Arulrajah et al. (2014) reported that
CBR values of RCA and CB are essentially the same, which are
118160 for RCA and is 123138 for CB. Due to similar particle size
distribution and CBR (intact strength of unstabilized materials), the
UCS values of stabilized RCA/CB blends are almost the same for
different RCA/CB ratios.
Fig. 5 presents the permanent deformation and resilient strain
results of the cement stabilized RCA/CB blends. There is an increasing trend for permanent deformation with the increase of CB
replacement. The resilient strain increases sharply with increasing
deviator stress, while the permanent strain insignicantly changes
with loading cycles for a particular deviator stress. This response is
different from the unstabilized material in which permanent strain
increases with increasing loading cycles [41]. In other words, the
stabilized materials have higher durability than the unstabilized
material. Similarly, the resilient (elastic) strain increases as CB content increases. However, the permanent strain is almost constant
with number of cycles after the rst deviator stress. Based on the

method proposed by Vuong and Arnold [42] for assessing the performance of materials from RLT tests, the behavior of the materials
can be dened as stable, as the blends are seen to exhibit constant
permanent strain rate and resilient modulus.
Fig. 6 presents the resilient modulus versus stress stage of the
cement-stabilized RCA/CB blends. The resilient modulus of all
RCA/CB blends increases with increasing stress stage, which is typical for various pavement materials. In general, increasing the CB
replacement ratio causes a reduction of resilient modulus of
cement-stabilized RCA blends. The permanent strain and resilient
modulus indicate the durability of the pavement materials, which
is typically governed by Los Angeles (LA) Abrasion of the unstabilized material. It was reported that the LA abrasion of the RCA is
higher than the CB (Arulrajah et al., 2014). It is 28 and 36 for RCA
and CB, respectively. Consequently, the permanent strain increases
and resilient modulus decreases with increasing CB content even
though the UCS and CBR of the stabilized RCA/CB blends remain
almost constant as seen from Figs. 3a, b and 4. In other words, the
UCS and CBR are not the only controlling parameters for cyclic
(long-term) response specially permanent strain and resilient
strain which in return affect resilient modulus. Table 3 summarizes
the permanent strain and resilient modulus test results of cement
stabilized RCA/CB blends. Typical quarry aggregates would exhibit
resilient modulus values of between 225 and 400 MPa at 70% of
the OMC based on the computation of resilient modulus from the
permanent deformation testing phase [41]. The RCA/CB blends are
found to perform within the ranges expected of bound quarry subbase materials, with 50RCA/50CB on the borderline.
Fig. 7 shows the resilient modulus of different RCA/CB blends
against the maximum axial stress (deviator stress) for three different conning pressures of 20, 30 and 40 kPa. The increase of conning pressure resulted in a fractional increase of resilient
modulus in 100RCA and 70RCA/30CB blends. However such a trend
can only be partially established for 85RCA/15CB and 50RCA/50CB
blends. Fatigue life of pavement materials is said to be the number
of cycles needed for 50% reduction in stiffness compared to the initial stiffness. Austroads assumes this to be 1 million cycles for satisfactory performance. Further investigations with a wider range of
blends and conning pressure levels is required to condently
establish the trend for all the blends of RCA and CB.
Fig. 8 shows various stages of the four point beam fatigue test
and the modulus of rupture of each blend is presented in Table 4.
The modulus of rupture tends to decrease with CB content due to

Recycled concrete aggregate content (%)


100

90

80

10

20

70

60

50

30

40

50

40

600
575
550
525

achieved CBR range

CBR (%)

500
475
450
425
400
375
350
325
300

Crushed brick content (%)


Fig. 4. CBR results of cement stabilized RCA/CB blends.

60

672

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676


2000

2.5
Stage 2
deviator stress: 450 kPa
confining pressure: 50 kPa

Stage 1
deviator stress: 350 kPa
confining pressure: 50 kPa

Stage 3
deviator stress: 550 kPa
confining pressure: 50 kPa

1800

Permanent Strain, %

1400
1.5

1200
1000

Permanent Strain
Resilient Strain

800

1
100RCA
85RCA/15CB
70RCA/30CB
50RCA/50CB

600

0.5

Resilient Strain (microstrain)

1600

400
200

0
5000

10000

15000

20000

0
30000

25000

Loading Cycles
Fig. 5. RLT Permanent deformation results of cement stabilized RCA/CB blends.

lower durability of the materials as evident from the UCS and CBR
tests presented earlier in Figs. 3a, b and 4. The modulus of rupture
is identical for 100RCA and 85RCA/15CB, indicating that a small CB
content has only minimal effect on this modulus. 70RCA/30CB has
a smaller modulus of rupture than 50RCA/30CB. The peak load was
similarly identical for 100RCA and 85RCA/15CB as well as for
70RCA/30CB and 50RCA/30CB. According to these results, the
increase in CB content in the cement-stabilized RCA blends results
in a corresponding decrease in the peak load and modulus of rupture.
Fig. 9 shows an example of 4 point beam fatigue test results of a
selected 85RCA sample. At the very early stage of the test, the average resilient modulus of rst 50 cycles was 10505.14 MPa which
was considered as the initial modulus. By repeating the application
of the 70% of the peak load, the resilient modulus gradually
declined until it reached the 50% of the initial gure at 6597 cycles.
From this point, the declining rate of the resilient modulus

drastically dropped leading to the sample failure. Table 5 summarizes the results of exural modulus and exural fatigue beam
tests. As evident, the fatigue life for 100RCA was 130 cycles while
it was 29401 cycles for 50RCA/50CB. This variation in cycles could
be due to the fact that 80% of peak load was applied on 100RCA
while 70% of peak load was applied on 50RCA/50CB. To obtain a
wider spectrum of fatigue life versus the applied load for each
blend, it is recommended to undertake testing on additional beams
to determine true fatigue life characteristics as this is highly sensitive to load variations. A shift factor of about one third of the initial
exural modulus may give a rough estimation of eld design modulus [43]. Assuming a shift factor of one third of exural modulus, a
minimum design modulus is 3700 MPa for the cement-stabilized
RCA/CB blends.
The results of the exural beam tests were compared with previous works on cement-stabilized local base materials, including

1200
100RCA

85RCA/15CB

70RCA/30CB

50RCA/50CB

Confining Stress

Deviator Stress

500

1000

400

800

300

600

200

400

100

200

0
0

10

20

30

40

50

60

Stress Stage Number


Fig. 6. RLT Resilient modulus values of cement stabilized RCA/CB blends.

70

Applied Stress, kPa

Resilient Modulus, MPa

600

673

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676


Table 3
Results of permanent strain testing and resilient modulus for cement stabilized RCA/CB blends.
Blend

Actual dry density (% of MDD)

Actual moisture content (% of OMC)

100RCA
85RCA/15CB
70RCA/30CB
50RCA/50CB

99.7
99.1
99.9
100

56
56
79
70

Permanent strain at the end of each stage


(microstrain)
Stage 1

Stage 2

Stage 3

1180
940
1450
2570

1340
950
1450
2640

1500
1230
1550
2820

Resilient modulus range (MPa)

324.5498.3
231.1447.9
272.5435.6
201.3375.7

500
480

c = 20 kPa

100RCA

c = 30 kPa

Resilient Modulus, MR (MPa)

460

c = 40 kPa

85RCA/15CB

440
420

70RCA/30CB

400
380
50RCA/50CB

360
340
320
250

300

350

400

450

500

550

600

650

Max. Axial Stress (kPa)


Fig. 7. RLT Resilient modulus versus maximum axial stress for selected conning pressures.

Fig. 8. Different Stages of four point beam fatigue test.

hornfel and siltstone [17]. The modulus of rupture and exural


modulus for all the cement-stabilized blends were found to be consistent with the previous works, which indicates that these blends

are suitable for cement-stabilized subbases. The fatigue life was


also within the range that has been previously reported [17],
though additional testing would appropriately determine the true

674

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676

Table 4
Modulus of rupture results of cement stabilized RCA/CB blends.
Specimen

100RCA

85RCA/15CB

70RCA/30CB

50RCA/30CB

Width (mm)
Height (mm)
Moisture content (%of OMC)
Dry density, (% of MDD)
Peak load (kN)
Modulus of rupture (MPa)
Tensile strain at 95% of peak load (microstrain)

100.9
101.66
99.1
96.5
4.29
1.23
94.22

101.5
101.4
96.5
96.5
4.28
1.23
93.08

100.23
101.08
106
96
3.6
0.88
57.85

100.6
100.7
100.3
95.4
3.7
1.09
87.48

12000

1.4
100RCA
85RCA/15CB

1.2

70RCA/30CB
50RCA/50CB

Modulus of Rupture (MPa)

Resilient Modulus, Mpa

10000

8000

6000

4000

2000
half initial modulus line

0.8

0.6

0.4

0.2

0
0

1000

2000

3000

4000

5000

6000

7000

Cycles

Unconfined Compressive Strength (MPa)

Fig. 9. Four point beam fatigue test results of 85RCA blend.

fatigue life as just one test per blend would not determine this sufciently. Since limited data is available on the in-service and laboratory [17] fatigue life of cemented materials used in Australia
[22], broader range of tests on wider range of materials need to
be carried out in order to establish an acceptable exural fatigue
life for Australian pavement materials.
As UCS testing is much simpler and well-known than exural
beam testing, a relationship between the two tests can facilitate
the estimation of the exural strength of cement-stabilized materials using the UCS data. It was suggested that exural tensile
strength of cement-stabilized materials is about one-third of the
UCS for low-strength materials and about one-fth of the UCS for
high-strength materials [44]. Kersten [45] suggested an approximately linear relationship for various cement contents at all curing
periods for hardened cement-stabilized soils and showed a nearly
linear relationship at all cement contents and at all curing times.

Fig. 10. Modulus of rupture and unconned compressive strength of RCA/CB


blends.

The exural strength was approximately 20 percent of the compressive strength. Mandal [46] suggested a non-linear model
between exural strength and UCS for stabilized materials which
is similar to ndings of Kersten [45] for lower strength materials
however this model predicts lower exural strength values as the
UCS value grows. Fig. 10 shows the established relationship
between modulus of rupture (or exural tensile strength) and
UCS by department of transport, South Africa [44], Kersten [45]
and Mandal [46] along with the results of this study for different
RCA/CB blends. As evident, the results are reasonably in agreement
with the discussed models. Values reported in Fig. 10 from this
study are the average value for at least 4 different tests on each
blend.

Table 5
Four point exural beam fatigue test results of cement stabilized RCA/CB blends.
Specimen

100RCA

85RCA/15CB

70RCA/30CB

50RCA/50CB

Width (mm)
Height (mm)
Moisture content (% of OMC)
Dry density (% of MDD)

100.62
101.68
100.8
95.5

102.3
100.3
97.4
96.2

101.05
100.6
107
96.1

101.1
100.6
103.5
95.4

Flexural modulus test


Applied load (% of peak load)
Mean exural modulus; cycles 50100 (MPa)
Tensile stress for modulus test (kPa)

40
11463.25
494.15

40
11846.20
496.79

40
11351.33
420.26

40
11780.75
432.91

Flexural beam fatigue test


Applied load (% of peak load)
Tensile stress for fatigue test; mean of rst 50 (kPa)
Tensile strain; mean of rst 50 (microstrain)
Initial modulus; rst 50 (MPa)
Stress ratio
Strain ratio
Cycles to half initial modulus

80
978.28
102.05
9601.05
0.79
1.08
130

70
872.18
83.06
10505.14
0.70
0.89
6597

70
737.711
78.26
9435.30
0.84
1.35
396

70
758.21
70.95
10700
0.69
0.81
29401

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676

4. Conclusions
The results of RLT test were used to ascertain the performance
of cement-stabilized RCA/CB blends under simulated trafc loading conditions. Based on the results, 100RCA demonstrated the
highest resilient modulus range and 50RCA/50CB demonstrated
the lowest resilient modulus range. All blends were found to
perform within the ranges expected of bound quarry subbase
materials with 50RCA/50CB on the borderline.
Mean UCS values met the minimum requirement of 4 MPa for
minimum of 7 days of curing for all blends. 20% and 24% increase
in strength was observed for 28 days cured samples of 70RCA/
30CB and 50RCA/50CB respectively compared to the 7 days cured
samples, while the average 28-day strength of 85RCA/15CB samples improved signicantly from 4 to 7.02 MPa (by almost 75%).
The notable increase in strength after 28 days of curing compared
to just 7 days of curing is as expected. The achieved mean UCS values, particularly after 7 days of curing were slightly lower than
anticipated. The water/cement ratio in the prepared UCS samples
were commonly around 4.0 or above, whereas industry practice
is to target a maximum water/cement ratio of 3.5 as this provides
certainty that the achieved UCS values will comfortably achieve
the specication limits.
The modulus of rupture (or exural tensile strength) varied
from 0.88 MPa to 1.23 MPa while the exural modulus ranged
from 11351.33 MPa to 11846.20 MPa. Assuming a shift factor of
0.3, the design modulus was estimated based on the exural modulus, which ranged from 3405.39 MPa to 3553.86 MPa. The modulus of rupture and exural modulus for all the cement-stabilized
blends were found to be consistent with the previous works, which
indicates that these blends are suitable for cement-stabilized
subbases.
The range of exural fatigue life was between 130 and 29401
cycles. The wide range of fatigue life is due to the fact that exural
fatigue test is highly sensitive to the applied load. The results of the
exural beam tests were noted to be consistent with past works
with cement-stabilized quarry produced crushed rock products.
The cement-stabilized RCA/CB blends with up to 50% CB content
and 3% GP cement were found to have physical and strength properties, which comply with the road authority requirements for
pavement base/subbase applications.
Acknowledgements
This research was supported under Sustainability Victoria
(Contract No. 7303), Australian Research Councils Linkage Projects
(LP120100107) and Australian Research Councils Linkage Infrastructure, Equipment and Facilities (LE110100052) funding
schemes. The 5th author acknowledges the support from the
Suranaree University of Technology and the Thailand Research
Fund under the TRF Senior Research Scholar program Grant No.
RTA5680002.
References
[1] Poon CS, Chan D. Feasible use of recycled concrete aggregates and crushed clay
brick as unbound road sub-base. Constr Build Mater 2006;20:57885.
[2] Azam A, Cameron D. Geotechnical properties of blends of recycled clay
masonry and recycled concrete aggregates in unbound pavement construction.
J Mater Civ Eng 2013;25(6):78898.
[3] Arulrajah A, Piratheepan J, Bo MW, Sivakugan N. Geotechnical characteristics
of recycled crushed brick blends for pavement sub-base applications. Can
Geotech J 2012;49(7):796811.
[4] Puppala AJ, Hoyos LR, Potturi AK. Resilient moduli response of moderately
cement-treated reclaimed asphalt pavement aggregates. J Mater Civ Eng ASCE
2011;23(7):9908.
[5] Taha R, Al-Harthy A, Al-Shamsi K, Al-Zubeidi M. Cement stabilization of
reclaimed asphalt pavement aggregate for road bases and subbases. J Mater Civ
Eng 2002;14(3):23945.

675

[6] Hoyos LR, Puppala AJ, Ordonez CA. Characterization of cement ber-treated
reclaimed asphalt pavement aggregates: preliminary investigation. J Mater Civ
Eng ASCE 2011;23(7):97789.
[7] Arulrajah A, Piratheepan J, Disfani MM. Reclaimed asphalt pavement and
recycled concrete aggregate blends in pavement subbases: laboratory and eld
evaluation. J Mater Civ Eng 2014;26(2):34957.
[8] Arulrajah A, Ali YMM, Piratheepan J, Bo MW. Geotechnical properties of waste
excavation rock in pavement sub-base applications. J Mater Civ Eng ASCE
2012;24(7):92432.
[9] Akbulut H, Gurer C. Use of aggregates produced from marble quarry waste in
asphalt pavements. J Build Environ 2007;42:192130.
[10] Grubb DG, Gallagher PM, Wartman J, Liu Y, III MC. Laboratory evaluation of
crushed glass-dredged material blends. J Geotech Geoenviron Eng ASCE
2006;132(5):56276.
[11] Wartman J, Grubb DG, Nasim ASM. Select engineering characteristics of
crushed glass. J Mater Civ Eng 2004;16(6):52639.
[12] Arulrajah A, Ali M, Disfani M, Piratheepan J, Bo M. Geotechnical performance of
recycled glass-waste rock blends in footpath bases. J Mater Civ Eng
2013;25(5):65361.
[13] Disfani MM, Arulrajah A, Bo MW, Hankour R. Recycled crushed glass in road
work applications. Waste Manage 2011;31(11):234151.
[14] Arulrajah A, Piratheepan J, Disfani M, Bo M. Geotechnical and geoenvironmental
properties of recycled construction and demolition materials in pavement
subbase applications. J Mater Civ Eng 2013;25(8):107788.
[15] Yeo R, Vuong B, Alderson A. Contract report-towards national test methods for
stiffness and fatigue characterisation of stabilised materials. ARRB Transport
Research, RC2028-002. 2002. p. 23.
[16] Austroads. Pavement design: a guide to the structural design of road
pavements. vol (loose leaf). Sydney, New South Wales, Australia: Austroads;
Standards Australia; 2004. p. 1.
[17] Yeo R. The development and evaluation of protocols for the laboratory
characterisation of cemented materials. Austroads, 2008. p. 89.
[18] Andrews RC, Brimble RM. Mechanistic characterisation of cementitiously
stabilised pavement materials. Adelaide, South Australia, Australia: Transport
SA. Materials Technology Section, 1998. p. 1.
[19] Bullen F. The resilient moduli of cement treated materials. Road Trans Res
1994;3(2):94104.
[20] Litwinowicz A, Brandon A. Dynamic exure testing for prediction of cement treated pavement life. In: Australian Road Research Board Ltd (ARRB)
Conference, 17th, 1994, Gold Coast, Queensland, Australia. Vermont South,
Victoria, Australia: Australian Road Research Board Ltd (ARRB); 1994. p. 22947.
[21] Otte E. A structural design procedure for cement treated layers in
pavements. Faculty of Engineering, University of Pretoria; 1978.
[22] Jameson G. Guide to pavement technology Part 2: pavement structural
design: Austroads Incorporated, 2008.
[23] Standards Australia, Methods of testing soils for engineering purposes Method
431: Soil chemical tests Determination of the pH value of a soil
Electrometric method, AS 1298.4.3.1, 1997.
[24] Standards Australia, Method of testing soils for engineering purposes, Method
311: Soil classication tests-Determination of the liquid limit of a soil-Four
point Casagrande method, AS 1289.3.1.1-2009, 2009.
[25] Standards Australia, Method of testing soils for engineering purposes, Method
341: Soil classication tests-Determination of the linear shrinkage of a soilStandard method, AS 1289.3.4.1-2008, 2009.
[26] VicRoads, Manual of Testing: Foreign Materials in Crushed Concrete, RC
372.04, 2008.
[27] Standards Australia, Methods for sampling and testing aggregates, Method
111: Particle size distribution-Sieving method, AS 1141.11.1-2009, 2009.
[28] Standards Australia, Method of testing soils for engineering purposes, Method
363: Soil classication tests Determination of the particle size distribution of
a soil Standard method of ne analysis using a hydrometer, AS 1289.3.6.32003, 2003.
[29] Standards Australia, Methods of testing soils for engineering purposes, Method
521: Soil compaction and density tests-Determination of the dry density/
moisture content relation of a soil using modied compactive effort, AS
1289.5.2.1-2003, 2003.
[30] Standards Australia, Methods of testing soils for engineering purposes, Method
611: Soil strength and consolidation tests Determination of the California
Bearing Ratio of a soil Standard laboratory method for a remoulded
specimen, AS 1289.6.1.1-1998, 1998.
[31] Standards Australia, Method of testing soils for engineering purposes, Method
121: Sampling and preparation of soils-Disturbed samples-Standard method,
AS 1289.1.2.1-1998, 1998.
[32] Standards Australia, Methods for preparation and testing of stabilized
materials, Method 4: Unconned compressive strength of compacted
materials, AS 5101.4 - 2008, 2008.
[33] Austroads, Commentary to AG:PT/T053 determination of permanent
deformation and resilient modulus characteristics of unbound granular
materials under drained conditions, 2007.
[34] Standards Australia, Methods of testing concrete, Method 11: Determination
of the modulus of rupture, AS 1012.11 2000, 2000.
[35] Yeo R. Construction report for cemented test pavements inuence of vertical
loading
on
the
performance
of
unbound
and
cemented
materials. Sydney: Austroads; 2008. p. 96.
[36] VicRoads, Cementitious treated crushed concrete for pavement subbase,
Section 821, 2011.

676

M.M. Disfani et al. / Construction and Building Materials 68 (2014) 667676

[37] Horpibulsuk S, Katkan W, Sirilerdwattana W, Rachan R. Strength development


in cement stabilized low plasticity and coarse grained soils: laboratory and
eld study. Soils Found 2006;46(3):35166.
[38] Horpibulsuk S, Rachan R, Chinkulkijniwat A, Raksachon Y, Suddeepong A.
Analysis of strength development in cement-stabilized silty clay from
microstructural considerations. Constr Build Mater 2010;24(10):201121.
[39] Miura N, Horpibulsuk S, Nagaraj TS. Engineering behavior of cement stabilized
clay at high water content. Soils Found 2001;41(5):3345.
[40] Horpibulsuk S, Miura N, Nagaraj TS. Assessment of strength development in
cement-admixed high water content clays with Abrams law as a basis.
Geotechnique 2003;53(4):43944.
[41] Arulrajah A, Piratheepan J, Disfani M, Bo M. Resilient moduli response of
recycled construction and demolition materials in pavement subbase
applications. J Mater Civ Eng 2013;25(12):19208.

[42] Vuong B, Arnold G. Predicting in-service performance of alternative pavement


materials for New Zealand conditions. Wellington, NZ: Land Transport New
Zealand; 2006. p. 180.
[43] Choummanivong L, Yeo R, Lourensz S. Laboratory assessment of cemented
materials. arrb; 2006. p. 68.
[44] Department of Transport. Cementitious Stabilizers in Road Construction South
Africa. In: Technical recomendations for highways, editor. Pretoria, Republic of
South Africa: Department of Transport; 1985.
[45] Kersten MS. Soil stabilization with Portland cement. Washington, DC: National
Academy of Sciences National Research Council; 1961.
[46] Mandal T. Fatigue behavior and modulus growth of cementitiously stabilized
pavement layers. University Of Wisconsin-Madison; 2012.