Construction and Building Materials 44 (2013) 798811

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

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

Flexural, shear and bond strength of polymer concrete utilizing recycled

resin obtained from post consumer PET bottles
Fareed Mahdi a, H. Abbas b,, Asif Ali Khan c
a
Department of Civil Engineering, Z.H. College of Engineering and Technology, Aligarh Muslim University, Aligarh 202 002, U.P., India
b
Specialty Units for Safety and Preservation of Structures, Department of Civil Engineering, King Saud University, Riyadh, Saudi Arabia
c
Analytical and Polymer Research Laboratory, Department of Applied Chemistry, Z.H. College of Engineering and Technology, Aligarh Muslim University, Aligarh 202 002, U.P., India

h i g h l i g h t s

 Polymer concrete (PC) is produced using waste PET as an alternate binding material.
 Flexure/shear and bond behaviour of plain and reinforced PC are studied.
 Properties of PC are compared with plain and reinforced cement concrete.
 PC produced is quite strong in exure and shear and bond with steel is also good.

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

Article history: The present investigation aims at studying the engineering properties of concrete produced by using
Received 25 August 2011 waste polyethylene terephthalate (PET) bottles as an alternate binding material in place of the commonly
Received in revised form 19 March 2013 used ordinary Portland cement. The recycled PET plastic waste was depolymerized through glycolysis to
Accepted 22 March 2013
produce unsaturated polyester resin which was used as a binder to produce polymer mortar (PM) and
Available online 25 April 2013
polymer concrete (PC). Four sets of PM/PC were produced with PET to glycol ratio of 1:1 and 2:1. The ini-
tiator promoter combinations taken were Methyl ethyl ketone per oxide (MEKP) and cobalt naphthanate
Keywords:
(CoNp) in one group while Benzoil per oxide (BPO) and N,N-diethyl aniline (NNDA) were used in another
PET
Polymer concrete
group.
Flexural strength The experimental investigation involves studying the engineering properties of PM, plain and rein-
Shear strength forced PC viz. modulus of rupture, modulus of elasticity, stressstrain behaviour, loaddeection charac-
Bond strength teristics, exural/shear strength and bond with reinforcing steel. The test results obtained for the four
sets of PC have been compared for assessing the performance of PET to glycol ratio and the initiator-pro-
moter combinations. The test results have also been compared with equivalent grade of plain and rein-
forced cement concrete.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
The enormous production of MSW, of which the major compo-
nent is the plastic product has led to an environmental catastrophe
in the urban regions of India and other developing countries which
has spread into the rural areas as well. The production of large
amount of PET has created an environmental problem of gigantic
proportions as it does not decompose readily in nature due to its
slow biodegradability [13] and hence could be thought of as a
noxious material.
During the last many years research has been carried out to
study the effect of PET wastes in concrete and construction mate-
Corresponding author.
E-mail address: abbas_husain@hotmail.com (H. Abbas).
rial either as a binder or simply as a ller material [46]. Choi et al.
[7] reported the use of PET wastes as ne aggregate to develop
lightweight aggregate concrete with a density of 1390 kg/m3. The
compressive strength of concrete got reduced by 530% depending
upon the quantity of PET waste used in the mix. However the
structural efciency (compressive strength/density ratio) of such
concrete was higher as compared to the conventional concrete
without PET. Ochi et al. [8] investigated the use of PET waste as
concrete reinforcing bre and found it to be useful for arresting
cracks in concrete. Remadnia et al. [9] studied the effect of PET
waste and animal protein in the form of powdered Haemoglobin
on the workability and mechanical properties of concrete. The re-
sults showed improved workability but the compressive strength
was reduced by up to 54%. Marzouk et al. [10] studied the effect
of PET waste on the density and compressive strength of concrete.

0950-0618/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.03.081
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 799

Table 1
Details of optimal sets.

Set No. Chemical composition

Group I
Type of glycol Diethylene glycol
Dibasic acid Maliec anhydride and phthalic anhydride
Initiator Benzoil per oxide (BPO)
Promoter N,N-diethyl aniline (NNDA)
1 PET glycol ratio 1:1
2 PET glycol ratio 2:1
Group III
Type of glycol Diethylene glycol
Dibasic acid Maliec anhydride and phthalic anhydride
Initiator Methyl ethyl ketone per oxide (MEKP)
Promoter Cobalt naphthanate (CoNp)
5 PET glycol ratio 1:1
6 PET glycol ratio 2:1

The study demonstrates that shredded plastic bottles may be used

successfully as aggregate in place of sand in cementitious concrete Fig. 1. Stressstrain curves for different sets of polymer mortars.
composites. It was found and compressive strength decreased
when the PET aggregate exceeded 50% by volume of sand.
It has been reported in the past [1117] that plastic waste can
be used in concrete or mortar as a binder thus replacing cement.
Most of the studies carried out for the depolymerization of PET
have discussed the role of various types of glycols and metal ace-
tates to be used as catalyst in the process [1722], whereas an
exhaustive study was carried out by the authors [16] so as to ascer-
tain the appropriate PET to glycol ratio and initiator-promoter
combinations. It was emphasized that such PC may be effectively
used in various structural pre-cast applications such as drains for
acid waste, underground vaults and junction boxes, sewer pipes,
power line transmission poles, transportation related components
such as median barriers, bridge panels and rail road ties. Further,
the use of PET waste as binder in the production of PC considerably
reduces the cost as compared to the PC produced using virgin resin.
Hisham and Moetaz [23] conducted experiments to study the
exural behaviour of polymer concrete (PC) made with epoxy resin
and a polyester with varying percentages. The results show that
the modulus of rupture and ultimate compressive strain for PC
were much higher thus improving the ductility as compared to
the ordinary Portland cement concrete. Ribeiro et al. [24] found
that the exural properties of unsaturated polyester and epoxy
mortars are strongly affected by the environmental temperatures.
However, except within a limited temperature range, the exural
strength of these mortar formulations decreases drastically with Fig. 2. Stressstrain curves for different sets of polymer concrete.
the rise in temperature. Rebeiz and Fowler [25,26] studied exural
behaviour of steel-reinforced PC beams using unsaturated polyes-
ter resins based on recycled PET waste. The studies indicated very
good exural and shear strength of reinforced PC so produced. ACI et al. [29] studied the mechanical properties of PC made with
method was found to yield conservative results when applied to recycled PET and recycled concrete aggregates. The PC at a resin
reinforced PC and alternative design equations were developed content of 9% was nearly unaffected by HCl, whereas the PC with
for predicting the shear and exural strength of steel-reinforced 100% recycled aggregate showed poor acid resistance. Unlike acid,
PC beams. alkali compounds did not attack the PC with 100% recycled aggre-
Jo and Kim [27] investigated the mechanical properties of PC gate as observed from the weight change and the compressive
produced using resin based on recycled PET and achieved com- strength.
pressive strength of 73.7 MPa, exural strength of 22.4 MPa, split- In the present study, stressstrain characteristics, exural and
ting tensile strength of 7.85 MPa, and elastic modulus of 27.9 GPa, shear strength of polymer concrete (PC), exural and bond strength
at 7 days. The authors later studied uniaxial creep behaviour of of reinforced PC (RPC) have been studied to understand the exural
the PC [28]. The creep strain at early ages was found to increase behaviour of PC beams. The optimal proportions of PET to glycol ra-
in PC more rapidly than in ordinary cement concrete. Creep in tio and initiator-promoter combinations obtained in the previous
PC was the result of molecular movement in the viscoelastic resin studies of the authors [16,17] have been used in the production
binder. The creep values increased with an increase in applied of PC. The optimal proportions in the previous study [16] were ob-
stress, although the values were not proportional to the stress ra- tained based on the compressive and tensile strength of PM and
tio, because of the nonlinear viscoelastic behaviour of the PC. Joe concrete, thermo-gravimetric and SEM analysis.
800 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811

Table 2
Mechanical properties of polymer mortar and concrete for different sets.

Set No. Compressive strength (MPa) Ultimate crushing strain (%) Modulus of elasticity of Modulus of
polymer concrete (GPa) rupture (MPa)
Polymer mortar Polymer concrete Polymer mortar Polymer concrete
5 22.0 42.2 1.64 1.68 3.33 9.31
6 26.0 35.6 2.69 1.26 3.33 7.45
1 18.0 25.0 6.90 3.27 2.22 8.83
2 28.0 30.0 4.26 4.35 1.67 6.34

Table 4
Table 3 Flexural/shear test results of PC for different a/d ratios.
Unit weight of different materials. Set Specimen Span to Failure Ultimate Shear
Material Unit weight (kN/m3) No. ID depth ratio, load (kN) deection strength
a/d (mm) (MPa)
Coarse aggregate (fully compacted) 15.5
Coarse sand (fully compacted) 17.8 5 SB 1 1 1.0 7.633 4.60 1.53
Polymer resin 11.5 6 SB 2 1 1.0 8.920 3.80 1.78
Polymer concrete 21.0 1 SB 3 1 1.0 8.310 4.13 1.66
2 SB 4 1 1.0 8.290 2.80 1.66
5 SB 1 2 1.5 6.710 2.20 1.34
6 SB 2 2 1.5 7.130 2.60 1.43
1 SB 3 2 1.5 3.740 1.73 0.75
2 SB 4 2 1.5 5.500 1.53 1.10
P
5 PCB1 2.0 3.880 2.30 0.78
6 PCB2 2.0 3.103 1.90 0.62
1 PCB3 2.0 3.676 2.40 0.74
2 PCB4 2.0 2.640 1.97 0.53
b
5 SB 1 3 2.5 3.000 1.40 0.60
d
6 SB 2 3 2.5 3.670 1.30 0.73
1 SB 3 3 2.5 2.914 1.26 0.58
2 SB 4 3 2.5 3.115 1.00 0.62

a L-2a a
Shear span Flexure span Shear span

Fig. 3. Schematic diagram for test setup for exure/shear test.

2. Experimental work

2.1. Preparation of polymer resin

The resin was prepared by the depolymerization of PET through Glycolysis.

Diethyl Glycol and zinc acetate as a catalyst were used for the purpose. Maleic
anhydride and phthalic anhydride or only Maleic anhydride were used to carry
out non-catalyzed melt polyesterication reaction. The styrene monomer obtained
from GSC (India) was used to reduce the viscosity of the resin and to allow the cross
linking of the molecular chains. All chemicals and reagents were of analytical re-
agent grade. Different combinations of initiator/promoter, dibasic acids and PET
to glycol ratio were used for the formation of four different groups. For the prepa-
ration of cured resin, benzoyl peroxide as initiator and N,N-diethyl aniline as pro-
moter were used in groups I and II, whereas in the remaining groups III and IV,
MEKP and CoNp were used as initiator and promoter systems respectively. These
chemicals were obtained from CDH (India). The PET to glycol ratio was varied for Fig. 5. Variation of shear strength with span to depth ratio.

(a) Unreinforced beam (b) Reinforced beam

Fig. 4. Beams under exural test.
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 801

4.0 Specimen 1
5.0
Specimen 2
Specimen 3

4.0
3.0

Load (KN)
Load (kN)
3.0
2.0

2.0

1.0
1.0 Specimen 1
Specimen 2
Specimen 3
0.0 0.0
0.0 1.0 2.0 3.0 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Deflection (mm) Deflection (mm)
(a) PCB1 (b) PCB2

4.0 4.0
Specimen 1
Specimen 2
Specimen 3

3.0 3.0

Load (kN)
Load (kN)

2.0 2.0

1.0 Specimen 1 1.0

Specimen 2
Specimen 3

0.0 0.0
0 1 2 3 4 0 0.5 1 1.5 2 2.5 3
Deflection (mm) Deflection (mm)
(c) PCB3 (d) PCB4
Fig. 6. Loaddeection curves for un-reinforced polymer concrete beams.

Fig. 8. Cross-section of reinforced beams.

Table 5
Fig. 7. Modulus of rupture and compressive strength of cement-concrete and Flexural strength and ultimate deection of CC and RPC beams.
polymer concrete.
Set Designation of Failure Ultimate Bending moment
No. beams load (kN) deection (mm) (kN-m)
CCB 0.768 1.1 0.040
each group leading to eight different sets [16] and out of these eight sets, four opti- 5 RPCB1 8.881 7.5 0.444
mal sets viz. Set Nos. 1, 2, 5 and 6 were obtained on the basis of compressive and 6 RPCB2 3.154 8.2 0.158
tensile strength, thermal stability and morphology of mortar and concrete. For 1 RPCB3 5.628 6.2 0.282
these optimal sets, maliec and phthalic anhydride were used as dibasic acids to pro- 2 RPCB4 6.604 6.2 0.331
duce resin as a binding agent. The studies reported in this paper are limited only to RCCB 2.601 7.5 0.130
these four optimal sets whose details are given in Table 1.
802 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811

10 oxide (BPO) and N,N-diethyl aniline as initiator and promoter system, these were
mixed rst and then the resin was mixed with it prior to its mixing with aggregates.
In the case of Methyl ethyle ketone per oxide (MEKP) and Cobalt naphthanate
Un-reinforced
8 (CoNp), resin was rst poured on the aggregates and then MEKP was added in
Reinforced
the resin followed by CoNp and then the entire system was mixed manually.
Ultimate Load (kN)

6
2.3. Preparation of specimens

4 The conventional tools normally employed in the preparation of cement con-

crete such as spatula and trowel were used in the preparation of PC. After usage,
the spatula and trowel were immediately washed with water and soap for avoiding
2 the hardening of sticking material on their surfaces.
After proper mixing, the freshly prepared material was put in moulds for the
casting purpose. During mixing, it was ensured that the materials do not come in
0 contact with water because it interferes with the polymerization process. During
CCB/RCCB

PCB1/RPCB1

PCB2/RPCB2

PCB3/RPCB3

PCB4/RPCB4
Fig. 9. Ultimate exural load for reinforced and plain cement concrete and polymer
concrete beams.
2.2. Production of PC
The PC test specimens for compression, exure, shear and bond were prepared
by using unsaturated polyester resin (UPER) obtained from PET waste. UPER was di-
luted with styrene in order to reduce its viscosity. Then this diluted resin was mixed
with free radical initiator and promoter before mixing it with inorganic aggregate.
The weight of resin taken was 10% by weight of aggregate. In case of Benzoil per
the mixing process especially at the time when initiator and promoter were added,
facemask was used to avoid the inhalation of toxic fumes. In order to avoid the con-
tact with material, gloves were also used so that the free radicals that are being gen-
erated when the promoter reacts with the initiator do not affect the skin.
In the previous study [17], curing of the material was carried out at room tem-
perature for the preparation of PC. It was found that the PC does not gain any appre-
ciable strength even after 45 days of curing. It was, therefore, decided that the
curing be carried out at some elevated temperatures so as to facilitate the release
of the free radicals. It was observed in previous study [16,17] that proper hardening
was not taking place and the material remained somewhat sticky even after seven
days, suggesting that proper polymerization has not taken place, which in turn may
be due to the inadequate presence of free radicals. It was observed that PC test spec-
imens gained signicant strength at high temperatures.
The specimens were demoulded after 6 h in the case of BPO and within 15 min
in the case of MEKP. The demoulded material was then left at room temperature for
12 h, after which they were kept in an oven at 150 C for 72 h for the curing
purpose.

12.0 4.0

3.5
10.0
3.0
8.0
2.5
Load (kN)
Load (kN)

6.0 2.0

1.5
4.0
1.0
Specimen 1 Specimen 1
2.0
Specimen 2 0.5 Specimen 2
Specimen 3
Specimen 3
0.0 0.0
0.0 2.0 4.0 6.0 8.0 0.0 4.0 8.0 12.0
Deflection (mm) Deflection (mm)
(a) RPCB1 (b) RPCB2

7.0 8.0

6.0 7.0

6.0
5.0
Load (kN)
Load (kN)

5.0
4.0
4.0
3.0
3.0
2.0 Specimen 1 2.0
Specimen 2 Specimen 1
1.0 Specimen 3 1.0 Specimen 2
Specimen 3
0.0 0.0
0.0 4.0 8.0 12.0 16.0 0.0 2.0 4.0 6.0 8.0 10.0
Deflection (mm) Deflection (mm)
(c) RPCB3 (d) RPCB4
Fig. 10. Loaddeection curves for reinforced polymer concrete beams.
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 803

1
CCB RCCB
3
Specimen 1
0.8
Specimen 2

Load (KN)
0.6 2

Load (KN)
0.4
1
Specimen-1
0.2 Specimen-2
Specimen-3

0 0
0 0.5 1 1.5 2 0 2 4 6 8 10 12 14
Deflection (mm) Deflection (mm)
(a) Un-reinforced (b) Reinforced
Fig. 11. Loaddeection curves of cement concrete beams.

2.4. Compression testing of PC

The stressstrain characteristics of PM and PC were established on the basis

of compression testing of cubes. Three cubes each for the optimal sets were
prepared. The size of PM and PC cubes were 70.6 and 150 mm respectively.
The cubes were tested under direct compression using 200 t compression test-
ing machine.
The stressstrain curves for PM and PC cubes under compression are plotted in
Figs. 1 and 2 respectively. The compressive strength, modulus of elasticity and ulti-
mate crushing strains after applying zero correction are given in Table 2.
The unit weights of different materials used in the preparation of PC test spec-
imens are given in Table 3. It is observed that the unit weight of the composite
material is more than the unit weight of the ingredients, suggesting that some of
the resin has been absorbed by the aggregates.

2.5. Modulus of rupture and shear strength of PC

In order to determine the modulus of rupture and shear strength of PC, the test
specimens of 50  50  305 mm size were prepared. The maximum size of coarse
aggregate was taken as 10 mm. The shear span to depth ratios, a/d, considered
Fig. 12. Test Specimen for pull out test. for testing were 1.0, 1.5, 2.0 and 2.5. Thus twelve test specimens were prepared

Mould with steel

bar in the centre Cast iron mould
with top cover

Bottom plate
Top cover

Fig. 13. Mould for pull out test specimens.

804 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811

Failed steel with neck

formation

Dial gauge

Steel

(a) Under test (b) After test

Fig. 14. Testing of pull out specimens.

for each of the four optimal sets as given in Table 1 giving the total number of test
specimens as 48. The modulus of rupture of PC was determined using the test
results for a/d = 2.0, wherein the shear span was one-third of the effective span of
beam. The mix proportion adopted for preparing the PC beams was 1:4:6 (resin:
ne aggregate: coarse aggregate) by weight.
The tests were carried out at Institute Instrumentation Centre (IIC) at Indian
Institute of Technology (IIT), Roorkee on a 10 t INSTRON universal testing machine.
The beams were tested under four point loading and the deection was given at a
uniform rate of 0.5 mm/min. The crack pattern and mode of failure was recorded.
The schematic diagram of the set up for exural testing of beams is shown in
Fig. 3. One of the beams under test is shown in Fig. 4.
The shear strength results with failure load and deection at failure are gi-
ven in Table 4. The variation of shear strength with a/d ratio is plotted in Fig. 5.
The loaddeection curves for the four sets of PC are plotted in Fig. 6. The ulti-
mate exural load for the four sets of PC is plotted in the form of bar chart in
Fig. 7.
2.6. Flexural testing of reinforced PC beams
Three reinforced PC beams of 50  50  305 mm size were also prepared for the
four sets given in Table 1. The mix proportion was kept same i.e. 1:4:6 (resin: ne
aggregate: coarse aggregate). For the sake of comparison, the reinforced cement con-
crete beams of the same size were also prepared using the same aggregates. The mix
proportion for the cement concrete beams was taken as 1:2:4 (cement: sand: coarse
aggregate). The beams were reinforced with two plain steel bars of 4 mm diameter
at top as well as at bottom at a clear cover of 3 mm. The shear stirrups were in the form
of hoops of 2 mm diameter at a spacing of 30 mm c/c. The reinforcement details of the
beams are shown in Fig. 8. One of the beams under test is already shown in Fig. 4.
The tests were carried out as per the specications laid down in Indian stan-
dards [30] in Institute Instrumentation Centre (IIC) at Indian Institute of Technology
(IIT) Roorkee, Roorkee using 10 t INSTRON universal testing machine. The beams
were tested under four point loading and the deection was given at a uniform rate
of 0.5 mm/min. The crack pattern and mode of failure was recorded.

Movable clamping
jaw of UTM
Steel bar Upper clamped plate
with central hole

Tightening
screw

Cube

Clamping rod

Dial gauge

Fixed lower
plate

Fixed end
Fig. 15. Schematic diagram for pull out test set up.
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 805

Table 6
Bond strength of rebar embedded in polymer concreteb.

Set Failure load Bond strength of polymer concrete, fb fb

 100
fc
No. (kN) (MPa)
(%)
5 68.1a 12.04 28.52
6 50.4 9.07 25.50
1 3.0 0.54 2.16
2 16.9 2.98 9.93
a
Steel bar fails with neck formation.
b
fc = Compressive strength of PC.

The modulus of rupture of PC, fr, was determined using simple bending
theory:

pl
fr 2
1
bd
where l is the effective span; p is the total applied load; b is the width of the section;
d is the depth of the section.
The failure load, ultimate deection after applying zero correction, exural
strength, and maximum bending moment of RPC beams and cement concrete
beams are given in Table 5. The modulus of rupture of different sets of PC and
cement concrete is plotted in Fig. 9. The compressive strength is also shown in
Fig. 9. Fig. 16. Load slip curves for various combinations.
The loaddeection curves for the exural test specimens of four sets of
RPC, CC and RCC are plotted in Figs. 10 and 11. The ultimate exural load for
four sets of RPC and cement concrete are already plotted in the form of bar
chart in Fig. 9.
Cross Linked Chains
2.7. Pull out test

The pull out test was performed on three specimens of the four optimal sets of
PC to determine the bond strength of PC. The pull out test was conducted on
150 mm cube with centrally placed deformed steel bar of 12 mm diameter as
shown in Fig. 12. The moulds used for casting the specimens are shown in
Fig. 13. The pull out test was carried out as per Indian standards [31]. The tests were
conducted in the concrete laboratory of the Department of Civil Engineering at
Indian Institute of Technology Roorkee, Roorkee on a 10 t universal testing machine.
One of the specimens under test is shown in Fig. 14 and the schematic diagram of
the test set up is shown in Fig. 15. The average bond strength fb of PC was calculated
using the relation:

P
fb 2
p/h
where P is the pull applied on the steel bar; / is diameter of the steel bar; h is the
embedded length of steel bar = size of the cube. Fig. 17. Scanning electron microphotographs (SEM) of hardened polymer resin
The observed failure load and the calculated values of bond strength for differ- using MEKP as initiator (Set No. 5) at a magnication of 2.00 Kx.
ent combinations are given in Table 6. The load slip curves are plotted in Fig. 16.

3. Results and discussion

Straight Chains
3.1. Stressstrain characteristics of polymer mortar

The observations made from the stressstrain diagrams plotted

in Fig. 1 are given in the following:

(i) The ultimate crushing strain of PM produced with MEKP as

initiator (i.e. group III comprising of Set numbers 5 and 6)
is less than the crushing strain of PM produced with BPO
as initiator (i.e. group I comprising of Set numbers 1
and 2). This is also conrmed from the SEM photographs,
wherein the linkages shown in the case of MEKP are more
in numbers.
(ii) The ultimate crushing strain of PM produced with BPO as
Fig. 18. Scanning electron microphotographs (SEM) of hardened polymer resin
initiator (i.e. Set Nos. 1 and 2) with PET to glycol ratio of
using BPO as initiator (Set No. 2) at a magnication of 2.00 Kx.
2:1 is less than that of 1:1. Whereas it is vice versa in the
case of PM produced with MEKP as initiator (i.e. Set Nos. 5
and 6).
(iii) Amongst all the sets, Set 1 has the highest ultimate crushing (iv) All the stressstrain curves may be categorized under the
strain of 6.90% followed by Set 2 which has an ultimate following two categories:
crushing strains of 4.26%. (a) Linear in the beginning followed by parabolic curve
806 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811
This type of curve is observed in Set 1 and 6. The curve is linear
up to 94.40% and 58.30% of ultimate compressive strength for Set 1
and 6 respectively. The beginning of parabolic curve indicates the
initiation of breaking of linkages, thus leading to the development
of micro-cracks.
(b) Parabolic curve sandwiched between two linear curves
The stressstrain curves of rest of the sets i.e. Set numbers 2 and
5 are of this type. The beginning of parabolic curve indicates the
initiation of breaking of linkages, thus leading to the development
of micro-cracks. But the subsequent linear portion of stressstrain
curve may be due to the reason that other linkages become active
in sharing the load along with the partially broken linkages. It is
due to this reason that the slope of the second linear portion of
stressstrain curve is less than the slope of the rst linear portion.
3.2. Stressstrain characteristics of PC
The observations made from the stressstrain diagrams of PC
plotted in Fig. 2 are given in the following:
(i) The ultimate crushing strain of PC produced with MEKP as
initiator (i.e. Set Nos. 5 and 6) is less than the crushing strain
of PC produced with BPO as initiator (i.e. Set Nos. 1 and 2).
This is also conrmed from the SEM photographs, wherein
the linkages shown in the case of MEKP are more in numbers.
(ii) The ultimate crushing strain for PC produced with BPO as ini-
tiator (i.e. Set Nos. 1 and 2) with PET to glycol ratio of 2:1 is
less than that of 1:1. Whereas it is vice versa in the case of
PC produced with MEKP as initiator (i.e. Set Nos. 5 and 6).
a/d = 1
Set No.5
(a)
a/d = 1
Set No. 1
(c)
Fig. 19. Tested unreinforced polymer concrete beams for a/d = 1.
(iii) Amongst all the sets, Set 2 has the highest ultimate crushing
strain of 4.35% followed by Set 1 and Set 5 which have a ulti-
mate crushing strains of 3.27% and 1.68% respectively.
(iv) All the stressstrain curves are Linear in the beginning fol-
lowed by parabolic curve. The beginning of parabolic curve
indicates the initiation of breaking of linkages, thus leading
to the development of micro-cracks.
3.3. Modulus of elasticity
The modulus of elasticity of different sets of PC given in Table 2
is the initial tangent modulus worked out from stressstrain curves
presented in Fig. 2 after applying zero correction. The observations
made from these results are:
(i) The modulus of elasticity of PC produced with MEKP as ini-
tiator (i.e. Set Nos. 5 and 6) is more than the modulus of elas-
ticity of PC produced with BPO as initiator (i.e. Set Nos. 1 and
2).
(ii) There is no denite trend of the variation of modulus of elas-
ticity when the PET to glycol ratio is varied.
(iii) Amongst all the sets, Set 5 and 6 have the highest modulus
of elasticity of 3.33 GPa followed by Set 1 which has the
modulus of elasticity of 2.22 GPa. Both the Sets 5 and 6 have
MEKP as initiator and Maliec and phthalic anhydride as
dibasic acids. The higher value of modulus of elasticity of
both Sets 5 and 6, corroborates that material produced using
MEKP is relatively rigid as compared to Set 1 and 2. The
modulus of elasticity of PC is lower than the cement con-
crete of same grade due to elasto-plastic nature of resin bin-
der in PC.
a/d = 1
Set No6
(b)
a/d = 1
Set No.2
(d)
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 807

a/d = 1.5 a/d = 1.5

Set No.5 Set No.6

(a) (b)
a/d = 1.5 a/d = 1.5
Set No.1 Set No.2

(c) (d)
Fig. 20. Tested unreinforced polymer concrete beams for a/d = 1.5.

a/d = 2.5 a/d = 2.5

Set No.5 Set No.6

(a) (b)

a/d = 2.5 a/d = 2.5

Set No.1 Set No.2

(c) (d)
Fig. 21. Tested unreinforced polymer concrete beams for a/d = 2.5.
808 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811

(iv) The modulus of elasticity of PC obtained in this study is
much lower than the value of 27.9 GPa obtained in an earlier
study [27] which is due to relatively lower compressive
strength of PC and larger quantity of resin used in this study.
3.4. Modulus of rupture of PC
The observations made from the results given in Table 2 are gi-
ven and discussed in the following:
(i) The modulus of rupture of PC produced with MEKP as ini-
tiator (Sets 5 and 6) is more than that of un-reinforced PC
produced with BPO as initiator (Sets 1 and 2). This is con-
sistent with the compressive strength results obtained for
PC cubes (Table 2), and may be due to the cross linking
sites being provided by phthalic anhydride [32] leading
to higher strength due to intense cross linking of polymer
chains just like the structure of a fabric as shown in SEM
photographs of hardened polymer resin shown in Figs. 17
and 18
(ii) The modulus of rupture of PC having PET to glycol ratio of
1:1 is more than that of 2:1 for both the cases of MEKP
and BPO.
(iii) For the same PET to glycol ratio of 1:1, the modulus of rup-
ture of the beam PCB1 produced with MEKP as initiator was
about 5.2% more than the modulus of rupture of the beam
PCB3 produced with BPO as initiator. Whereas, for the same
PET to glycol ratio of 2:1, modulus of rupture of beam PCB2
produced with MEKP as initiator was 14.9% more than that
of beam PCB4 produced with BPO as initiator.
(iv) Amongst all the four sets, the modulus of rupture of the
beam PCB1 is highest at 9.31 MPa, whereas, it is minimum
for the beam PCB4.
(v) The modulus of rupture of PC beams varies from 21% to 35%
of its compressive strength. On the other hand, the modulus
of rupture of CC beams of the same grade varies from 10% to
14% of its compressive strength (Fig. 7). The higher value of
modulus of rupture for PC is due to the elasto-plastic nature
of polymer binder as compared to the cement concrete
wherein the cement binder is brittle.
(vi) The highest value of modulus of rupture of PC for beam PCB1
is about 42% of the value obtained in an earlier study [27].
This is due to relatively lower compressive strength of PC
in this study which is about 57% of the strength reported
in Ref. [27].
3.5. Loaddeection behaviour of un-reinforced PC beams
The observations made from the Loaddeection curves of ex-
ural testing are given and discussed in the following:
(i) The ultimate deection of the un-reinforced PC beams pro-
duced with MEKP as initiator (Sets 5 and 6) after applying
zero correction is less than the ultimate deection of the

9
14
SB 1-1 SB 2-1
8 a/d =1 a/d = 1
MEKP 1:1 12 MEKP 2:1
7
10
6
Load (kN)

Load (kN)

5 8

4
6

3
4
2
Specimen 1
Specimen 2 2 Specimen 1
1
Specimen 3 Specimen 2
Specimen 3
0 0
0 1 2 3 4 5 6 0 1 2 3 4 5
Deflection (mm) Deflection (mm)

10
10 SB 4-1
SB 3-1 9 a/d = 1
9 a/d = 1 BPO 2:1
BPO 1:1 8
8
7
7
6
Load (kN)

6
Load (kN)

5 5

4 4

3 3

2 2
Specimen 1
Specimen 1 Specimen 2
1 Specimen 2
1
Specimen 3
Specimen 3
0 0
0 1 2 3 4 5 6 0 1 2 3 4
Deflection (mm) Deflection (mm)

Fig. 22. Loaddeection curves of unreinforced polymer concrete beams for a/d = 1.
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811 809

beams produced with BPO as initiator (Sets 1 and 2). The
results are in conformity with the crushing strain obtained
for PC cubes as given in Table 2.
(ii) The ultimate deection for PC beam produced with BPO as
initiator with PET to glycol ratio of 2:1 is less than that of
1:1. Whereas it is vice versa in the case of PC beams pro-
duced with MEKP as initiator. This result also corroborates
the strain results obtained for PC cubes.
(iii) All the loaddeection curves are linear in the beginning fol-
lowed by parabolic curve with large curvature. There is sud-
den drop in load at failure.
(iv) The failure of all the beams of different sets of PC occurred by
exural cracking in the middle third portion with almost a
vertical crack. No cracks were observed outside the middle
third portion of the beam. The failure occurred by the forma-
tion of only one crack, thus dividing the beam in two pieces,
but the two pieces were still connected at failure. No local
crushing of the material was observed at the four load points.
3.6. Loaddeection behaviour of reinforced PC beams
The observations made from the Loaddeection curves of ex-
ural testing are given and discussed in the following:
(i) All the loaddeection curves are linear in the beginning,
and then there is a small drop in the load followed by a par-
abolic curve. The ultimate deection of reinforced PC beams
is 313486% higher than the ultimate deection of un-rein-
forced PC beams. A comparison of loaddeection curves of
un-reinforced and reinforced PC beams indicate that the
provision of reinforcement has introduced considerable duc-
tility. The only exception to it is beam PCB2, which shows
that the reinforcement does not increase the ultimate ex-
ural load, but deection gets considerably increased. It
may be due to the bond slip between the reinforcing bar
and PC.
(ii) The load carried by reinforced PC beams (except PCB2) is
53150% higher than the load carried by un-reinforced PC
beams (Fig. 9). Whereas, the load carried by reinforced CC
beam is 238% higher than the load carried by un-reinforced
CC beam.
(iii) The load carried by reinforced PC beams (except PCB2) is
116241% higher than the load carried by reinforced CC
beam, which may be due to higher compressive strength
of PC as compared to CC taken in the study (Fig. 9).
(iv) The failure of reinforced PC beams is through the develop-
ment of distributed cracks in the middle third portion of
the beam. The nal failure of the beam is by widening of
one of these cracks.
3.7. Bond strength of PC
The observations made from the results given in Table 6 are
given and discussed in the following:

8
9
SB 1-2
7 a/d = 1.5 SB 2-2
8
MEKP 1:1 a/d = 1.5
MEKP 2:1
6 7

5 6
Load (kN)

Load (kN)

5
4
4
3
3
2
2
Specimen 1
1 Specimen 2 Specimen 1
1
Specimen 3 Specimen 2
Specimen 3
0 0
0 1 2 3 0 1 2 3 4
Deflection (mm) Deflection (mm)

7
5
SB 4-2
SB 3-2
a/d = 1.5
a/d = 1.5 6
BPO 2:1
BPO 1:1
4
5
Load (kN)
Load (kN)

3 4

3
2

1
Specimen 1 1 Specimen 1
Specimen 2 Specimen 2
Specimen 3 Specimen 3
0 0
0 1 2 3 0 0.5 1 1.5 2
Deflection (mm) Deflection (mm)

Fig. 23. Loaddeection curves of unreinforced polymer concrete beams for a/d = 1.5.
810 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811

(i) It is observed that the PC produced using MEKP as initiator
and having PET to glycol ratio of 1:1 (Set 5) has the highest
bond strength as compared to other combinations. For this
set of PC, the steel bar got fractured without bond slip fail-
ure, which indicates that the actual bond strength is much
higher.
(ii) The bond strength of PC produced with MEKP as initiator is
higher than the bond strength of PC produced with BPO as
initiator. This is valid for both the PET to glycol ratios consid-
ered in the study.
(iii) The bond strength of PC with PET to glycol ratio of 1:1 is
greater than that of 2:1 in the case of MEKP as initiator (Sets
5 and 6)and it is vice versa in the case of BPO as initiator
(Sets 1 and 2).
3.7.2. Bond Strength v/s Compressive Strength
The comparison of bond strength with the compressive strength
of PC is given in Table 6. It is observed that the bond strength of PC
produced with MEKP as initiator is 25.5% and 28.5% of the com-
pressive strength for the two PET to glycol ratios of 1:1 (Set 5)
and 2:1 (Set 6) respectively, which is much higher than the bond
strength of PC produced with BPO (Sets 1 and 2), wherein the cor-
responding values are 2.16% and 9.93%.
When the PET to glycol ratio is increased from 1:1 to 2:1, the
bond strength decreases marginally in the case of PC produced
with MEKP as initiator. Whereas, in the case of PC produced with
BPO as initiator, the increase in PET to glycol ratio from 1:1 to
2:1, results in considerable increase (360%) in the bond strength.

3.8. Shear strength of PC

3.7.1. Load slip curves
In the case of PC made of BPO as initiator, it is observed from the
load slip curves (Fig. 16) that the PC with PET to glycol ratio of 1:1
(Set 1) has higher slips as compared to the slip observed in the PC
with PET glycol ratio of 2:1 (Set 2) at the same pull, and conse-
quently has a lesser bond strength. But on the contrary the PC
made with MEKP as initiator and having PET glycol ratio of 1:1
(Set 5) has a smaller slip than the slip observed in PC with PET gly-
col ratio of 2:1 (Set 6) at the same pull and therefore has higher
bond strength. The load slip pattern for all the four sets of PC is al-
most linear.
The failure patterns of the beams are shown in Figs. 1921 for a/
d ratios of 1.0, 1.5 and 2.5. The failure pattern of a typical beam for
a/d ratio of 2 is already shown in Fig. 4. The loaddeection plots
for a/d ratios of 1.0, 1.5 and 2.5 are shown in Figs. 2224 respec-
tively, whereas, these plots for a/d = 2.0 are already shown in Fig. 6.
It is observed from Fig. 5 that the shear strength is decreasing
with increase in the a/d ratio for the PC of Sets 1 and 5 having
PET to glycol ratio of 1:1. Whereas, for PET to glycol ratio of 2:1
for both the cases of BPO and MEKP (Sets 2 and 6), it is found that
the shear strength decreases up to a/d ratio of 2 and thereafter it
increases slightly.

3.5 6
SB 1-3 Specimen 1 SB 2-3 Specimen 1
Specimen 2 Specimen 2
a/d = 2.5 a/d = 2.5
Specimen 3 Specimen 3
3 MEKP 1:1 5 MEKP 2:1

2.5
4
Load (kN)
Load (kN)

2
3
1.5

2
1

0.5 1

0 0
0 1 2 3 0 1 2 3
Deflection (mm) Deflection (mm)

3.5 4
Specimen 1 SB 4-3 Specimen 1
SB 3-3 Specimen 2
Specimen 2 a/d = 2.5
a/d = 2.5 Specimen 3 3.5 Specimen 3
3 BPO 2:1
BPO 1:1
3
2.5
2.5
Load (kN)
Load (kN)

2
2
1.5
1.5
1
1

0.5
0.5

0 0
0 1 2 3 0 1 2 3
Deflection (mm) Deflection (mm)

Fig. 24. Loaddeection curves of unreinforced polymer concrete beams for a/d = 2.5.
 F. Mahdi et al. / Construction and Building Materials 44 (2013) 798811
During testing it was observed that for a/d ratio of 1.0, the beams
failed without giving any prior warning i.e. the failure was sudden
and brittle and the beams split into two portions as shown in
Fig. 19. As the a/d ratio increases, the beams start showing ductile
behaviour. For a/d ratio of less than or equal to 1.5, the beams does
not seem to show any deection at failure (Figs. 22 and 23). But for
higher values of a/d ratio, there is signicant deection at failure be-
cause of predominant exural behaviour, as shown in Fig. 24.
For a/d ratio of 1.0 (Fig. 19), beams generally failed by the for-
mation of cracks near the mid span, except one or two beams in
each set, where the splitting has occurred under the load. It is seen
from Fig. 20 that most of the beams with a/d ratio of 1.5 have failed
near the load, except the beams made of MEKP with PET to glycol
ratio of 1:1 (Set 5), which failed by the formation of crack near mid
span. The same failure trend was observed in the case of all the sets
at a/d ratio of 2 and 2.5 (Figs. 4 and 21).
It is observed from Fig. 5 that the shear strength seems to become
asymptotic for lower values of a/d ratios. Therefore the shear
strength of PC of all the four sets may be taken as that corresponding
to the a/d ratio of 1.0. Thus it is observed from Table 4 that the shear
strength of PC varies from 1.53 MPa for Set 5 to 1.78 MPa for Set 6.
4. Conclusions
The compressive strength, modulus of elasticity and modulus of
rupture of PC produced with MEKP as initiator (i.e. Set Nos. 5 and
6) is more than the modulus of elasticity of PC produced with BPO
as initiator (i.e. Set Nos. 1 and 2). This may be due to the cross link-
ing sites being provided by phthalic anhydride leading to higher
strength due to intense cross linking of polymer chains just like
the structure of a fabric. The modulus of rupture of PC having
PET to glycol ratio of 1:1 is more than that of 2:1 for both the cases
of MEKP and BPO. The modulus of rupture of PC beams varies from
21% to 35% of its compressive strength. The highest value of mod-
ulus of elasticity and modulus of rupture were 3.33 GPa and
9.31 MPa respectively.
The bond strength of PC produced with MEKP as initiator is
higher than the bond strength of PC produced with BPO as initiator.
The PC produced using MEKP as initiator and having PET to glycol
ratio of 1:1 (Set 5) has the highest bond strength as compared to
other combinations. The bond strength of PC produced with MEKP
as initiator is 25.5% and 28.5% of the compressive strength for the
two PET to glycol ratios of 1:1 and 2:1 respectively, which is much
higher than the bond strength of PC produced with BPO. When the
PET to glycol ratio is increased from 1:1 to 2:1, the bond strength
decreases marginally in the case of PC produced with MEKP as ini-
tiator. Whereas, in the case of PC produced with BPO as initiator,
the increase in PET to glycol ratio from 1:1 to 2:1, results in consid-
erable increase (360%) in the bond strength.
The shear strength of PC seems to become asymptotic for lower
values of a/d ratios. The PC is quite strong in shear with shear
strength varying from 1.53 to 1.78 MPa.
References
[1] Chen CH. Study of glycolysis of poly(ethylene terephthalate) recycled from
post consumer soft drink bottles, further investigations. J Appl Polym Sci
2003;87(12):200410.
811
[2] Pimpan V, Sirisook R, Chuayjuljit S. Synthesis of unsaturated polyester resin
from post consumer PET bottles: effect of type of glycol on the characteristics
of unsaturated polyester resin. J Appl Polym Sci 2003;88:78892.
[3] Kint D, Munoz GS. A review of the potential biodegradability of PET. Polym Int
1999;88:34652.
[4] Kardon J. Polymer-modied concrete: review. J Mater Civ Eng
1997;9(2):8592.
[5] Ohama Y. Recent progress in concrete-polymer composites. Adv Cem Based
Mater 1997;5(2):3140.
[6] Siddique R, Khatib J, Kaur I. Use of recycled plastic in concrete: a review. Waste
Manage 2008;28(10):183552.
[7] Choi YW, Moon DJ, Kim YJ, Lachemi M. Characteristics of mortar and concrete
containing ne aggregate manufactured from recycled waste polyethylene
terephthalate bottles. Constr Build Mater 2009;23:282935.
[8] Ochi T, Okubo S, Fukui K. Development of recycled PET bre and its application
as concrete-reinforcing bre. Cem Concr Compos 2007;29:44855.
[9] Remadnia A, Dheilly RM, Laidoudi B, Queneudec M. Use of animal protein as
foaming agent in cementitious concrete composites manufactured with
recycled PET aggregates. Constr Build Mater 2009;23:311823.
[10] Marzouk OY, Dhilly RM, Queneudec M. Valorization of post consumer waste
plastic in cementitious concrete composites. Waste Manage
2007;27(2):3108.
[11] Rebeiz KS, Serhal SP, Fowler DW. Structural behaviour of polymer concrete
beams using recycled plastics. ASCE J Mater Civil Eng 1994;6(1):15065.
[12] Rebeiz KS, Yang S, David WF. Polymer mortar composites made with recycled
plastics. ACI Mater J 1994;91(3):3139.
[13] Rebeiz KS. Pre-cast use of polymer concrete using unsaturated polyester resin
based on recycled PET waste. Constr Build Mater 1996;10(3):21520.
[14] Vaidya UR, Nadkarni VM. Unsaturated polyester resin from polyethylene
terephthalate waste. Ind Eng Chem Res 1987;26(2):1948.
[15] Rebeiz KS, Fowler DW, Paul DR. Time and temperature dependent properties
of polymer concrete made with resins using recycled PET. Annual tech conf
proc, soc. of plastic engineers, Brookeld, CT, USA; 1991, vol. 37. p. 21469.
[16] Mahdi F, Abbas H, Khan AA. Strength characteristics of polymer mortar and
concrete using different compositions of resins derived from post-consumer
PET bottles. Constr Build Mater 2010;24:2536.
[17] Mahdi F, Khan AA, Abbas H. Physiochemical properties of polymer mortar
composites using resins derived from post-consumer PET bottles. Cem Concrte
Compos 2007;29(3):2418.
[18] Baliga S, Wong WT. Depolymerization of poly(ethylene terephthalate) recycled
from post-consumer soft drink bottles. J Polym Sci A: Polym Chem
1989;27(6):207182.
[19] Chen JW, Chen LW. The glycolysis of poly(ethylene terephthalate). J Appl
Polym Sci 1999;73(1):3540.
[20] Guclu G, Kasgoz A, Oabudak S, Oagumus S, Orbay M. Glycolysis of polyethylene
terephthalate waste in xylene. J Appl Polym Sci 1998;69(12):23119.
[21] Kao CY, Cheng WH, Wan BZ. Investigation of catalytic glycolysis of
polyethylene terephthalate by differential scanning calorimetry.
Thermochim Acta 1997;292(12):95104.
[22] Vaidya UR, Nadkarni VM. Polyester polyols from glycolysed PET waste: effect
of glycol type on the kinetics of polyesterication. J Appl Polym Sci
1989;38(6):117990.
[23] Hisham AF, Moetaz MH. Flexural behaviour of Polymer concrete. Constr Build
Mater 1999;13:25362.
[24] Ribeiro MCS, Novoa PR, Ferreira AJM, Marques AT. Flexural performance
polyester and epoxy polymer mortars under severe thermal conditions. Cem
Concr Compos 2004;26:8039.
[25] Rebeiz KS, Fowler DW. Flexural strength of reinforced polymer concrete made
with recycled plastic waste. ACI Struct 1996;93(5):52430.
[26] Rebeiz KS, Fowler DW. Shear and exure behavior of reinforced polymer
concrete made with recycled plastic wastes. ACI Special, Publication SP166-04;
1996, vol. 166. p. 6178.
[27] Jo B-W, Park S-K, Kim C-H. Mechanical properties of polyester polymer
concrete using recycled polyethylene terephthalate. ACI Struct J
2006;103(2):21925.
[28] Jo B-W, Tae G-H, Kim C-H. Uniaxial creep behavior and prediction of recycled-
PET polymer concrete. Constr Build Mater 2007;21(7):15529.
[29] Jo B-W, Park S-K, Park J-C. Mechanical properties of polymer concrete made
with recycled PET and recycled concrete aggregates. Constr Build Mater
2008;22(12):228191.
[30] IS516. Methods of tests for strength of concrete. Bureau of Indian Standard
India; 1959.
[31] IS2770, Part I. Methods of testing bond in reinforced concrete. Bureau of Indian
Standard, India; 1967.
[32] Brydson JA. Plastic materials. Oxford, UK: Butterworth-Heinemann; 1999.