Construction and Building Materials 34 (2012) 236–242

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Dynamic properties of stone mastic asphalt mixtures containing wasteplastic bottles

Taher Baghaee Moghaddam ⇑, Mohamed Rehan Karim, Tamalkhani SyammaunCenter for Transportation Research, University of Malaya, 50603 Kuala Lumpur, Malaysia

a r t i c l e i n f o

Article history:Received 29 November 2011Received in revised form 3 February 2012Accepted 25 February 2012Available online 30 March 2012

Keywords:Waste polyethylene terephthalateReinforced asphalt mixtureStiffness modulusFatigue life

0950-0618/$ - see front matter � 2012 Elsevier Ltd. Adoi:10.1016/j.conbuildmat.2012.02.054

⇑ Corresponding author.E-mail address: p.baghaee@gmail.com (T. Baghaee

a b s t r a c t

Fatigue failure is a common problem of asphaltic concrete which can lead to pavement damage. Manystudies have been conducted to find ways for increasing fatigue life of asphalt concrete mixtures. Thisstudy investigates effects of adding waste polyethylene terephthalate (PET) on stiffness and fatigue prop-erties of SMA mixtures at optimum asphalt contents. Different percentages of waste PET with maximumsize of 2.36 mm were added to SMA mixtures. Indirect tensile stiffness modulus test and indirect tensilefatigue test were conducted at temperature of 20 �C and at three different stress levels (250, 350,450 kPa). The results showed that stiffness modulus of mixture increased at lower amount of PET con-tent; however, adding higher amount of PET made mixture less stiff. In addition, PET reinforced mixturesexhibit significantly higher fatigue lives compared to the mixtures without PET.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Asphalt Concrete (AC) mixture is subject to many externalforces during its service life which could eventually lead to failure.Different types of failures have been observed in asphalt mixturessuch as permanent deformation (rutting), fatigue failure, and lowtemperature cracking. Fatigue failure is a common damage in ACmixtures which appears in the form of cracking (alligatorcracking).

Fatigue resistance is the ability of the asphalt mixture to resistrepeated bending forces without fracture and cracking. In asphaltconcrete pavement fatigue cracking is caused by successive tensilestrains due to repeated traffic loading. According to structural anal-ysis fatigue cracks are produced at the bottom of asphalt layerwhere the maximum tensile strains accrue, thereafter these crackspropagate to the surface of asphalt mixtures. Fatigue life of AC mix-tures has a negative correlation with the loads applied by vehicleson road pavements. Besides, fatigue life differs significantly amongtypes of AC mixtures.

Stone Mastic Asphalt (SMA) is a type of asphaltic concretewhich consists of more coarse aggregate content and filler, andhas better characteristic against permanent deformation comparedto the conventional dense graded mixture. SMA was developed inGermany in 1960s, and was used in Europe for years. Because ofSMA success in Europe, it has been used in the United States since1991 [1,2]. Previous studies showed that SMA mixture tends tohave lower fatigue life in comparison with dense graded mixture

ll rights reserved.

Moghaddam).

because inherent structure of dense graded mixture provides bet-ter interlock between the aggregate particles [3,4].

Using additives is a common way to improve fatigue life of ACmixtures. Different types of fibers and polymers can be used inAC mixtures. In a study, effects of adding polyester, polyacryloni-trile, lignin and asbestos fibers with different percentages wereinvestigated by Xu et al. [5]. It was shown that fatigue life improvedby adding fibers, and polyester and polyacrylonitrile which are con-sidered as polymer fibers had the best effect on fatigue properties ofAC mixtures. It is also reported that adding polypropylene fiber en-hanced the fatigue resistance of asphalt mixtures, while fatigue lifeincreased 27% by adding 1% polypropylene fiber [6].

Although utilization of virgin additives in asphalt mixture canimprove fatigue properties of AC, in many cases road constructioncost increases considerably. Thus, many investigations were con-ducted on the mixtures containing waste materials as additivesto improve asphalt mixture characteristics and prevent from im-posed additional charges due to usage of virgin materials. Further-more, this would be an alternative solution for environmentalpollution by utilizing waste materials as secondary materials inroad construction projects. Waste glass, steel slag, tires and plastics(polymers) are examples of waste materials which have been usedin AC mixtures in previous studies [7]. Among waste materialswaste tire and recycled polymer have a prominent utilization[8–13]. The use of glass fiber has also been found to improve fati-gue life of SMA mixes according to studies by Mahrez and Karim[14].

The main objective of this study is to investigate stiffness andfatigue properties of SMA mixtures containing different percent-ages of waste polyethylene terephthalate (PET).

T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242 237

2. Experimental procedure

2.1. Materials

SMA is gap-graded AC which is used in this study. Particle size distribution ofthe gradation is presented in Fig. 1. Granite-rich aggregate particles were obtainedfrom Kajang Rock Quarry in Malaysia. Because of the importance of aggregate qual-ity in SMA mixtures, several tests were done on coarse and fine aggregate particles,and the results are listed in Table 1. Furthermore, in SMA mixture amount of aggre-gate passing sieve 0.075 mm (filler) is higher than the amount used for conven-tional Hot Mix Asphalt, and is between 8% and 10% by weight of aggregateparticles [1]. In this study 9% filler was used.

In order to prepare AC mixtures, 80–100 penetration-grade virgin asphalt hasbeen utilized. Table 2 illustrates some properties of asphalt cement which is usedin this research.

PET is a type of polyester material, and is often used for packing in food and bev-erage industries. Waste PET was obtained from PET bottles. For utilization of PETbottles as additive in AC mixtures the bottles were cut to small parts, thereaftercrushed by crushing machine. The crushed PET particles were sieved, and the par-

Fig. 1. Particle size distribution for stone mastic asphalt mixture.

Table 1Properties of coarse and fine aggregate.

Property Value Standard test method

Coarse aggregateL.A. Abrasion 19.45% < 30% ASTM C 131Flakiness index 2% < 20% BS 812 Part 105.1Elongation index 11% < 20% BS 812 Part 105.2Specific gravity ASTM C 127

Bulk 2.60SSD 2.62Apparent 2.65

Absorption 0.72% < 2% ASTM C 127

Fine aggregateSpecific gravity ASTM C 128

Bulk 2.63SSD 2.64Apparent 2.66

Absorption 0.4% < 2% ASTM C 128Soundness loss 4.1% < 15% ASTM C 88

Table 2Properties of asphalt cement.

Properties Value Standard test method

Penetration at 25 �C (0.1 mm) 87 ASTM D 5Softening point (�C) 46 ASTM D 36Viscosity at 135 �C (mPa s) 325 ASTM D 4402Viscosity at 170 �C (mPa s) 62.5 ASTM D 4402Specific gravity 1.03 ASTM D 70

ticles passing sieve 2.36 mm were used for this investigation (see Fig. 2). Table 3depicts some properties of PET material.

2.2. Specimen fabrication

The specimens were prepared at optimum asphalt content (OAC) usingMarshall Method. All together six different amounts of OACs have been obtainedfor six different PET contents, 6.77%, 6.45%, 6.43%, 6.29%, 6.36% and 6.51% of OACeach for 0%, 0.2%, 0.4%, 0.6%, 0.8% and 1% (all by weight of aggregate particles) ofPET content, respectively.

For preparing AC mixtures, 1100 g of mixed aggregate was placed in the oven at160 �C between 3 and 4 h. Asphalt binder was also heated at 130 �C before mixingwith aggregate particles. PET particles, with the maximum size of 2.36 mm, wereadded directly to the mixture as the method of dry process. Mixing temperaturewas kept constant at the temperature between 160 and165 �C. The loose mixturewas placed in the preheated mold and 50 blows of compaction were applied byMarshall Hammer on each side of specimen at temperature of 140 �C.

2.3. Test method

2.3.1. Indirect tensile stiffness modulus testNon-destructive indirect tensile stiffness modulus test is used to obtain stiff-

ness of AC mixtures. Indirect tensile stiffness modulus gives the relationship be-tween stress and strain of AC mixtures at specific load and temperature. This testwas carried out by Universal Testing Machine (UTM) according to AASHTO TP31.UTM is a computer controlled system and operates automatically. During the test,specimens were subjected to compressive haversine waveform loads across the ver-tical section across the thickness of specimen, and deformation of specimen wasmeasured by linear variable differential transducers (LVDTs) along diametrical sec-tion of specimen. In order to obtain uniform temperature the specimens wereplaced at controlled temperature environmental chamber for 4 h prior to test.The test was conducted at three different stress levels (250, 350 and 450 kPa) andat temperature of 20 �C.

Horizontal tensile stress and stiffness modulus of AC mixtures can be obtainedby the following equations:

rxðmaxÞ ¼ 2� Pp� d� t

ð1Þ

Sm ¼P � ðmþ 0:27Þ

H � tð2Þ

where rx(max) is the maximum horizontal tensile stress in middle of specimen; Sm isthe stiffness modulus; P, applied vertical peak load, H; amplitude of horizontal defor-mation, t; average thickness of specimen; d, average diameter of specimen and m,Poisson’s ratio.

Each specimen was tested twice. After the first load repetition was applied (thefirst 10 cycles) the specimen was rotated around 90� and another cyclic load wasapplied on the specimen. The final stiffness modulus was calculated as the averagevalue of first and second loading repetition.

2.3.2. Indirect tensile fatigue testFatigue generally is expressed as the fracture under repeated stress with a max-

imum value generally less than tensile strength of material [6]. Fatigue life of as-phalt mixture depends on different mixture properties such as type and amountof asphalt binder, air voids, and mix gradation [3]. Fatigue test can be conductedat controlled stress (load) or strain modes. During the controlled stress modeamount of applied stress is kept constant, and amount of strain increases. In con-trolled strain mode the strain value (deformation) is kept constant while the stressdecreases during the test [15].

In this study, Indirect Tensile Fatigue Test (ITFT) was carried out in controlledstress mode according to EN 12697. For running ITFT, UTM was used, and compres-sive cyclic load was applied along with diametrical section of specimen in the formof haversine waveform with 500 ms repetition time and 100 ms pulse width. ITFTwas conducted at the same stress levels and temperature as were used for stiffnessmodulus test.

Vertical deformation of specimen was monitored during the test. Fatigue lifewas defined as the number of load repetitions reached when the specimen splits,or deformation reaches to the maximum value of 9 mm [16].

Horizontal tensile strain also can be obtained as the function of stress and stiff-ness of mixture by using Eq. (3):

exðmaxÞ ¼ rxðmaxÞð1þ 3mÞSm

ð3Þ

ex(max) is the Maximum tensile strain at the center of specimen, rx(max) is the Max-imum tensile stress at the center of specimen, Sm is the Stiffness modulus of speci-men and m, Poisson’s ratio (0.35 at temperature of 20 �C).

Fig. 2. Procedure to prepare crushed PET particles from PET bottles.

Table 3Properties of PET.

Property Used standard Value

Density (g/cm3) ASTM D 792 1.35Water absorption (%) ASTM D 570 0.1Tensile strength (psi) ASTM D 638 11,500Approx melting temperature (�C) – 250

238 T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242

3. Results and discussion

3.1. Indirect tensile stiffness modulus results

Stiffness modulus of mixtures with different percentages of PETwas obtained at different stress levels at 20 �C. For each percentage

Fig. 3. Stiffness modulus test result for each

of PET the mean value of three tested specimens was obtained andis shown in Fig. 3. As depicted in Fig. 3, stiffness modulus decreaseswith an increase in applied stress, and in the same stress levelsstiffness of mixtures increases initially with addition of PET fol-lowed by a decreasing trend. Besides, it was shown that 0.2% isthe optimum content of PET to achieve maximum stiffness value.The trend shows adding higher amount of PET results in less stiffermixtures. Besides, it can be illustrated that stiffness values differslightly among the stress levels at higher PET content (e.g. 1%PET). The changes in stiffness values cannot be referred to the per-centage of air voids in the mixtures because all the samples werefabricated at optimum asphalt content and had the same air voids.It was reported that higher asphalt content makes mixture lessstiffer [3,8], so the decrease in stiffness values for asphalt mixturecontaining waste PET cannot be attributed to the percentages of

percentage of PET at three stress levels.

T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242 239

asphalt content while the optimum asphalt content for controlmixture is higher than the mixtures containing PET. In addition,although the mixture reinforced by 0.2% PET has higher asphaltcontent than mixtures containing 0.4%, 0.6% and 0.8% PET, it hashigher stiffness value in comparison with those mixtures. The pos-sible reason for this result is attributed to mechanical characteris-tics of PET in the mix. In fact because of high melting point of PET(approximately 250 �C) the PET particles do not melt during mixfabrication which is around 160 �C, so during the mix fabricationand after heating PET properties change more or less into crystalproperties [17]. Thus, the rigid PET can make mix more flexibleand cause higher deformation under load application in themixture.

3.2. Indirect tensile fatigue test results

Results of ITFT for each percentage of PET content are depictedin Fig. 4. Fig. 4 presents relationship between stress, or strain, andfatigue life (Nf) in logarithm scale with linear relationships. Test re-sults indicate that PET reinforced asphalt mixtures have consider-ably higher fatigue lives in comparison with control mixtures (themixture without PET), and adding higher amount of PET results inhigher fatigue life. For instance, fatigue life for the mixture contain-ing 1% PET increased by 124.8% at 250 kPa stress level when Nf in-creased from 27,571 to 61,981 cycles. It can be seen that ACmixtures tends to have lower fatigue lives at higher stress levels.

Some literatures indicate that there was a direct relationshipbetween fatigue life and stiffness of mixtures [3,14]. In otherwords, stiffer mixtures had better characteristics against fatiguedamage. Nevertheless, there is not any exact correlation betweenthese two AC characteristics. As the result obtained in this studyshowed however PET reinforced mixtures are less stiff; they hadlonger fatigue lives compared to the mixture without PET. The in-crease in fatigue life is consistent with previous studies on using

Fig. 4. Relationship between cycles t

fibers in SMA mixtures conducted by Ye et al. [18]. Higher fatiguelife of PET reinforced mixtures may be attributed to improvementof elastic property of asphalt mixtures containing PET which canpostpone fatigue damage due to numerous numbers of load appli-cations, or because of distraction of stress, which is generated inmixture, by PET particles.

3.3. Prediction models for fatigue characteristic of PET reinforcedmixtures

Usually using fatigue prediction models are helpful in order topredict fatigue characteristics of asphalt mixture at specific tem-perature, asphalt binder, air voids and specific gradation. In orderto predict fatigue characteristics of control and PET reinforcedSMA mixtures two models are proposed in this study. These mod-els are created based on the relationship existing between stress,or strain, and fatigue life as are presented in Eqs. (4) and (5):

Nf ¼ A1r

� �n ð4Þ

Nf ¼ a1e

� �b ð5Þ

where Nf is the number of load cycles to failure, r is applied stress, eis initial strain, and A, n, a and b are regression coefficients (fatigueparameters) which are related to mixture properties.

Fatigue prediction models for mixtures with different percent-ages of PET content were calculated by SPSS software using regres-sion analysis, and are presented in Table 4. Data obtained indicatethat there are strong correlations between stress, or strain, and fa-tigue lives of mixtures. Also, there are substantial differences incase of fatigue parameters among control and PET reinforced mix-tures (see Figs. 5 and 6).

o failure and Stress (and Strain).

Table 4Fatigue prediction models for control and PET reinforced mixtures.

PET (%) r (kPa) e(10�6) Nf (Cycles) Fatigue model Correlation coefficient

0 250 118.41 27,571 Nf ¼ 9� 1017 � 1r� �5:645 0.987

350 176.90 3011 Nf ¼ 1:5� 1013 � 1e� �4:251 0.969

450 257.25 1031

0.2 250 113.99 34,371 Nf ¼ 5� 1018 � 1r� �5:935 0.983

350 167.13 3191 Nf ¼ 9� 1013 � 1e� �4:624 0.96

450 250.75 1091

0.4 250 119.08 46,771 Nf ¼ 4:5� 1019 � 1r� �6:247 1.00

350 175.47 5641 Nf ¼ 9� 1014 � 1e� �4:968 0.996

450 249.59 1191

0.6 250 121.82 56,031 Nf ¼ 9:5� 1019 � 1r� �6:352 1.00

350 176.64 6551 Nf ¼ 3:5� 1015 � 1e� �5:197 0.995

450 250.70 1341

0.8 250 122.46 57,781 Nf ¼ 7:5� 1019 � 1r� �6:303 1.00

350 176.59 6981 Nf ¼ 3� 1015 � 1e� �5:156 0.998

450 251.36 1421

1 250 134.83 61,981 Nf ¼ 4� 1019 � 1r� �6:178 1.00

350 191.49 7761 Nf ¼ 1� 1017 � 1e� �5:747 1.00

450 253.85 1641

Fig. 5. Coefficients of A and n in fatigue prediction equations for PET reinforced andcontrol mixtures. Fig. 6. Coefficients of a and b in fatigue prediction equations for PET reinforced and

control mixtures.

240 T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242

3.4. Deformation (Displacement)

In asphalt mixtures crack formation created at the center of thespecimen where the maximum tensile stress and strain occur, thenpropagates along the two sides of specimen. Fig. 7 illustrates dis-placement of control mixes compared to those containing 1%PET. As can be seen there are three phases for the displacementof the specimens as defined by law of displacement or deforma-tion. These three stages included rapid displacement phase whichis normally because of compacting the mixture due to existing airvoids in the mixture. The second stage is elastic zone with low lev-

els of displacement and usually called as a stable phase. The thirdstage is called plastic zone and is an unstable phase for crack-growth in the mixture.

As can be seen in Fig. 7 section a, control mixture and the mix-ture reinforced by 1% PET reach nearly the same displacement atthe beginning of second phase which is around 2 mm; however,more differences can be observed among these two values at high-er stress levels. Further, displacement for control and PET rein-forced mixture have the same value (5 mm) at 250 stress leveland at the end of elastic phase and this value decreased for control

Fig. 7. Displacement vs Fatigue life at different stress levels.

T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242 241

mixture at higher stress levels and reach to the values of 4 and3 mm for 350 and 450 Kpa stress, respectively. It is good to notethat at higher stress levels higher deformation is required for PETreinforced mixture to go to plastic zone compared to the controlmixtures, and the values remain constant at the value of 5 mm atdifferent stress levels.

In addition, as can be illustrated in Fig. 7 length of second stage(elastic phase) for the mixture containing 1% PET is considerablylonger compared to the control mixture. From the result it can beconcluded that PET improved elastic property of mixture. Basedon the displacement crack propagation law and characteristics ofthe mixtures containing PET particles it is concluded that PET rein-forced mixture is more likely to have plastic fracture in comparisonwith the control mixture which is more likely to have brittlefracture.

4. Conclusion

In this study, effects of adding different percentages of wastePET particles with maximum size of 2.36 mm were investigatedon stiffness and fatigue properties of SMA mixtures. Further fatigueprediction models were proposed. The results are summarized asfollows:

1. Stiffness was changed by adding PET particles. It was shownthat although stiffness of SMA mixture initially increased byadding lower amount of waste PET into the mixtures, itdecreased at higher amount of PET content (e.g. 1% PET).

2. Fatigue properties of SMA mixture improved significantlyfor PET reinforced mixture compared to the control mixture,and mixtures containing higher amount of PET had higherfatigue lives.

3. This study implies that elastic property of asphalt mixtureimproved by adding PET that makes mixture more flexible,and prevents from crack initiation and propagation in mix-tures due to cyclic load application.

Acknowledgment

The authors gratefully acknowledge Mr. Mohd Khairul Anwarfor his supporting role in the implementation of this researchprogram.

References

[1] Asi Ibrahim M. Laboratory comparison study for the use of stone matrixasphalt in hot weather climates. J Constr Build Mater 2006;20:982–9.

[2] Schmiedlin RB. Stone matrix asphalt: the Wisconsin experience. Trans ResRecord: J Trans Res Board 1988;1616:34–41.

[3] Nejad FM, Aflaki E, Mohammadi MA. Fatigue behavior of SMA and HMAmixtures. J Constr Build Mater 2010;24:1158–65.

[4] Moghaddam TB, Karim MR, Mahrez A. A review on fatigue and ruttingperformance of asphalt mixes. J Sci Res Essays 2011;6(4):670–82.

[5] Xu Q, Chen H, Prozzi JA. Performance of fiber reinforced asphalt concrete underenvironmental temperature and water effects. J Constr Build Mater2010;24(10):2003–10.

[6] Tapkın S. The effect of polypropylene fibers on asphalt performance. J BuildEnviron 2008;43:1065–71.

[7] Huang Y, Bird RN, Heidrich O. A review of the use of recycled solid wastematerials in asphalt pavements. J Resour Conserv Recy 2007;52:58–73.

[8] Arabani M, Mirabdolazimi SM, Sasani AR. The effect of waste tire thread meshon the dynamic behaviour of asphalt mixtures. J Constr Build Mater2010;24:1060–8.

[9] Katman HY, Karim MR, Mahrez A, Ibrahim MR. Performance of wet mixrubberised porous asphalt. Proc Eastern Asia Soc For Transport Stud2005;5:695–708.

[10] Katman HY, Karim MR, Ibrahim MR, Mahrez A. Effect of mixing type onperformance of rubberised porous asphalt. Proc Eastern Asia Soc For TransportStud 2005;5:762–71.

242 T. Baghaee Moghaddam et al. / Construction and Building Materials 34 (2012) 236–242

[11] Casey D, McNally C, Gibney A, Gilchrist MD. Development of a recycledpolymer modified binder for use in stone mastic asphalt. J Resour ConservRecy 2008;52:1167–74.

[12] Hınıslıoglu S, Agar E. Use of waste high density polyethylene as bitumenmodifier in asphalt concrete mix. J Mater Lett 2004;58:267–71.

[13] Li J, Zhang Y, Zhang Y. The research of GMA-g-LDPE modified Qinhuangdaobitumen. J Constr Build Mater 2008;22:1067–73.

[14] Mahrez A, Karim MR. Fatigue characteristics of stone mastic asphalt mixreinforced with fiber glass. Int J Phys Sci 2010;5(12):1840–7.

[15] Arabani M, Mirabdolazimi SM. Experimental investigation of the fatiguebehaviour of asphalt concrete mixtures containing waste iron powder. J MatSci Eng A-Struct 2011;528:3866–70.

[16] Yan J, Ni F, Yang M, Li J. An experimental study on fatigue properties ofemulsion and foam cold recycled mixes. J Constr Build Mater 2010;24:2151–6.

[17] Ahmadinia E, Zargar M, Karim MR, Abdelaziz M, Shafigh P. Using waste plasticbottles as additive for stone mastic asphalt. J Mater Design 2011;32:4844–9.

[18] Ye Q, Wu S, Li N. Investigation of the dynamic and fatigue properties of fiber-modified asphalt mixtures. Int J Fatigue 2009;31:1598–602.