Fatigue performance evaluation of SBS modified mastic asphalt mixtures published on CBM

Construction and Building Materials 48 (2013) 908–916

Contents lists available at ScienceDirect

Construction and Building Materials

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

Fatigue performance evaluation of SBS modified mastic asphalt mixtures

0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.07.100

⇑ Corresponding author. Tel.: +82 2 3408 3812; fax: +82 2 3408 4332.E-mail address: [email protected] (H.J. Lee).

Tae Woo Kim a,b, Jongeun Baek a,b, Hyun Jong Lee b,⇑, Ji Young Choi a

a Highway Pavement Research Division, SOC Research Group, Korea Institute of Construction Technology, 2311 Daewha-Dong, Ilsan-Gu, Goyang-Si, Gyeonggi-Do 411-712, Republicof Koreab Department of Civil and Environmental Engineering, Sejong University, 98 Gunja-Dong, Gwangjin-Gu, Seoul 143-747, Republic of Korea

h i g h l i g h t s

� SBS modifiers enhanced low-temperature crack resistance of asphalt binder.� High strain fatigue life of the SBS modified mastic asphalt mixtures was improved.� The SBS modified mixtures had greater flexural toughness than control mixtures.� The SBS modifiers was more substantial in retarding fatigue crack growth.

a r t i c l e i n f o

Article history:Received 14 February 2013Received in revised form 23 July 2013Accepted 25 July 2013

Keywords:Mastic asphaltSBS (Styrene–Butadiene–Styrene) modifierFatigue performanceTLA (Trinidad Lake Asphalt)Bridge deck pavement

a b s t r a c t

This study evaluated the fatigue performance of Styrene–Butadiene–Styrene (SBS) modified masticasphalt mixtures used for bridge deck pavements. The effect of the type and content of newly developedSBS modifiers was investigated using typical binder tests. Four-point bending beam fatigue and indirectstrength tests were conducted to examine fatigue and fracture behaviors of the SBS modified masticasphalt mixtures. The SBS modifiers without C@C double bonds enhanced the mechanical properties ofthe mastic asphalt binder and mixtures: lower stiffness after short- and long-term oxidation, significantlygreater fatigue resistance at a higher strain level, and higher crack development resistance.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Recently long-span bridges are constructed considerably in theworld. Orthotropic steel bridge decks have been popularly used forthe long-span bridges in order to reduce self weight of the bridgedeck [1–4]. The orthotropic steel bridge deck plate and deck pave-ments are designed as thin as possible, which leads to large deflec-tions and to reduce fatigue life of deck pavements. For example, ifthickness of asphalt concrete deck pavements is reduced by 50%,tensile strain at the bottom of the deck pavement can be doubledapproximately, resulting in considerable reduction of the fatiguelife of the deck pavements [5,6]. Hence, high performance pavingmaterials needs to be developed for deck pavements in long-spanbridge decks.

Since 1960s mastic asphalt concrete pavements have been usedin European countries and Japan to enhance fatigue performance ofthe bridge deck pavements [7,8]. For example, 70 mm thick and50 mm thick mastic asphalt pavements have been used popularly

in Germany and the Netherlands, respectively. In Japan, a combina-tion of a 35 mm thick mastic asphalt base layer and a 35 mm thickmodified asphalt wearing surface layer has been applied to deckpavements. The main reason of the popular usage of the masticasphalt concrete in bridge deck pavements is that the masticasphalt could provide superior waterproofing and higher flexuralresistance than other deck pavement materials. Mastic asphaltpavements contain higher binder contents, typically more than8.0%, resulting in superior durability and flexibility. Also, masticasphalt is mixed and constructed at a very high temperature ofapproximately 230 �C using a special mixer, namely cooker whichenables to maintain high temperature during production andtransportation. Due to superior fluidity at the high temperature,mastic asphalt does not need a compaction procedure during theconstruction; nonetheless, air void of the mastic asphalt concreteis almost close to zero.

In general, Trinidad Lake Asphalt (TLA) occupies 30–50% ofasphalt binder in mastic asphalt mixtures. TLA is a high viscousnatural asphalt, providing higher durability, rut-resistance, andlow-temperature cracking resistance. In Germany and the Nether-lands, it is recommended to use the ratio of TLA to asphalt binder

Table 1Components of the five types of SBS modifiers.

ID SBS type Antioxidant content (wt.%)

SBS A SBS with C@C double bonds 0.4SBS B 1.0

SBS C SBS without double bonds 0.4SBS D 2.5SBS E 5.0

T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916 909

of 30% by weight; the ratio increased to 50% in the UK. Despite ofgood performance of the TLA, it is a natural product so that it is dif-ficult to modify its original characteristics as required further.

Recently a couple of European countries led by Germany devel-oped Styrene–Butadiene–Styrene (SBS) modified mastic asphaltbinders to improve the fatigue performance of mastic asphaltpavements. SBS is a synthetic rubber composed of three longchains of polystyrene, polybutadiene, and polystyrene which havecompletely opposite characteristics. Polystyrene is hard plastic tomake SBS durable at high temperature; polybutadiene behaves likea soft rubber at a room temperature leading SBS to be more flexi-ble. SBS modified mastic asphalt mixtures have good fatigueperformance at a wide temperature range of �30 to 20 �C [9,10].Thus, the use of the SBS is beneficial to improve fatigue perfor-mance of mastic asphalt concrete. However, the use of modifiedasphalt concrete in deck pavements led to the increase of initialconstruction cost. Life cycle cost (LCC) analysis showed that theLCC of modified asphalt concrete, Gussasphalt, mastic asphalt,and epoxy asphalt concrete could be lower than that of conven-tional deck pavements because of longer service life and lower usercost [11].

Mastic asphalt concrete is suitable for 50–80 mm thick deckpavements [7]. Furthermore, thinner deck pavements, e.g.,40 mm or less, are needed for super long-span bridges as a wayof significantly reducing self-weight of the deck pavements [6].Among several paving materials for long-span bridge deck pave-ments, SBS modified mastic asphalt concrete could be appropriatebecause of various types of SBS modifiers and successful previousstudies. However, in order to apply the SBS modified masticasphalt concrete into relatively thin deck pavements, it requiressuperior fatigue performance at a higher strain level of 1000 leor even greater [5,12].

100

2. Research objective and scope

In this study, SBS modified mastic asphalt mixture was devel-oped to enhance the fatigue resistance of conventional masticasphalt mixture where TLA is used as a part of asphalt binder. Inorder to evaluate the fatigue performance of the SBS modified mas-tic asphalt mixture, various laboratory tests for asphalt binder andmixtures were conducted. First, physical and mechanical bindertests were conducted to examine the characteristics of the SBSmodified asphalt binder and to determine the proper type and con-tent of the SBS modifiers and TLA in the SBS modified mastic as-phalt mix design. The physical tests include penetration, ring andball temperature, and ductility tests before and after rolling thinfilm oven aging. Second, four-point bending (FPB) fatigue testsand indirect tensile (IDT) strength tests were conducted to evalu-ated the fatigue resistance of the SBS modified mastic asphalt mix-tures in terms of fatigue life and crack development, respectively.

0

20

40

60

80

0.01 0.1 1 10 100

Pas

sing

(%

)

Seive size (mm)

Lower limit

Upper limit

This study

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

3. Experimental program

3.1. Materials

3.1.1. SBS modified asphalt binderControl asphalt binder consists of AC 60-80 base asphalt and 30% of TLA. In SBS

modified asphalt binder, a part of the TLA was substituted by SBS modifiers to im-prove fatigue crack resistance. The type and content of the SBS modifiers will bedetermined based on the performance of the SBS modified asphalt binder. Table 1lists the five types of SBS modifiers evaluated in this study.

The SBS modifiers mainly contain SBS rubbers and antioxidants. Type A and BSBS modifiers have double bonds (C@C) in molecular structures which can be bro-ken at a high temperature and result in thermal instability of the SBS rubbers. Intype C, D, and E SBS rubbers, the double bonds were cut by hydrotreating process[13] to improve the high-temperature thermal stability. Also, two antioxidantswere added in the SBS modifiers to mitigate the mastic asphalt from aging due tooxidation during mixing and transportation at a high temperature of 200 �C or

greater. The antioxidants were Irganox-1010 produced by Ciba and Irgafos 168 pro-duced by Gumho Petrochemical Co. The amount of the antioxidant ranged from0.4% to 5.0% by weight of the SBS modifier where the antioxidants were usedequally.

3.1.2. AggregatesAggregates used for the SBS modified mastic asphalt mixtures were based on

the Gussasphalt specifications used in Germany and Japan [14,15]. The aggregateswhose nominal maximum aggregate size of 4.75 mm were used and consisted of46.5% of coarse aggregates, 30.0% of fine aggregates, and 23.5% of fillers. Fig. 1shows the distribution of aggregates of the mastic asphalt mixture with upperand lower limits of Gussasphalt used in Japan [14]. In order to minimize the vari-ance of aggregate gradations, the quantity of aggregates was measured precisely(±0.1 g) to meet the target gradation according to ASTM C136 [16]. The maximumpercentage of flat and elongated particles whose dimensional ratio is greater than3:1 was limited to 10%, which is typically used for high-volume four-lane roadsin Korea.

3.1.3. Mix design for mastic asphalt mixturesIn this study, the mastic asphalt mixtures were manufactured as follows: first

blended aggregates were heated up in a controlled-temperature oven at 200 �Cfor at least 12 h. Asphalt binder was also heated up in another oven at 150 �C for1–2 h. Then, the aggregates, asphalt binder, and TLA were mixed in a small-size coo-ker equipped with a self heating system that could control the cooker temperatureup to 300 �C as shown in Fig. 2(a). It took approximately 20 min until the mixturetemperature reached the target temperature of 240 ± 5 �C. During the mixing, mix-ture temperature was periodically monitored and adjusted slightly to keep the tar-get temperature. For the SBS modified mastic asphalt, the SBS modifiers were addedduring the mixing procedure.

A different mix design procedure was used for the mastic asphalt mixtures inorder to consider unique characteristics of the mastic asphalt mixtures, zero airvoids and no compaction. The Luer fluidity, indentation, and wheel tracking testswere conducted on the control and SBS modified mastic asphalt mixtures to checkconsistency and rut-resistance at high temperatures. In the Luer fluidity test, a 995-g weight was released to loosen 10-kg mastic asphalt mixtures at various temper-atures of 200–260 �C and penetration depth and time were periodically recordedduring the test as shown in Fig. 2(b). For each temperature, the penetration timevalue corresponding to 5 cm of penetration was determined. The mastic asphaltmixtures can be mixed and placed properly if the penetration time is in the rangeof 4–20 s.

The indentation test [17] was performed to measure indentation depth of a 7-cm3 cubic specimen by a 52.5-kg rod at 40 �C during 30 min as shown in Fig. 2(c). Ifthe indentation value ranges 1–4 mm, the mastic asphalt is regarded to be stable atthe high temperature. The wheel tracking test [18] was also conducted to measurerut depth by repeated wheel loads at 60 �C as shown in Fig. 2(d). Deformation of a

Fig. 2. Laboratory test setup for mix design: (a) the cooker, (b) the Luer fluidity test, (c) the identification test, and (d) the wheel tracking test.

910 T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916

specimen under 686 N of the wheel load was recorded and the number of load rep-etitions per 1 mm of deformation was computed as dynamic stability. The mini-mum dynamic stability for mastic asphalt mixtures is typically 300 cycles/mm.

Based on the mix design test results for various contents of asphalt binder, SBSmodifier, and TLA, the optimum SBS modified asphalt binder content was deter-mined as 8.3% by weight of the mixture which includes 15% of TLA and 8.5% ofSBS modifier by weight of the asphalt binder. On the other hand, the asphalt binderin the control mastic asphalt mixture was 8.5% by weight of the mixture where 30%of TLA by weight of the asphalt binder was used. Table 2 shows the contents of baseasphalt, SBS, and TLA and mix design test results.

3.2. Laboratory tests

3.2.1. Asphalt binder testsA series of physical and mechanical binder tests was conducted on the control

and five SBS modified asphalt binders to evaluate the effect of the SBS modifier onthe behavior of the asphalt binder in terms of penetration, softening point, ductility,and stiffness before and after short-term and long-term aging. The rolling thin filmoven (RTFO) and pressure aging vessel (PAV) tests were used to simulate the short-term and long-term aging of the asphalt binder. In addition, the morphology of theSBS modified asphalt binder was investigated using optical microscopy.

3.2.2. Mixture testsIn this study, fatigue crack resistance of the mastic asphalt mixture was the

main concern because fatigue cracking has been considered as the main failure cri-terion for deck pavements [19]. The fatigue performance of the mastic asphalt mix-tures was evaluated using four-point bending (FPB) fatigue tests [20,21]. Inaddition, crack resistance was evaluated using FPB and indirect tensile (IDT)strength tests.

The FPB fatigue tests were performed at a frequency of 10 Hz and at a temper-ature of 20 �C under the controlled-strain mode of loading which is more preferredfor thin flexible pavements such as bridge deck pavements where the elastic recov-ery properties of the material can affect its fatigue life [22]. Only a minor permanentdeformation could be developed in the specimen during the fatigue tests since thestrain applied to the specimen must be recovered to the original position in each

Table 2Mix design test results for the control and SBS modified mastic asphalt mixtures.

Asphalt binder Control

AC (%)a 8.5Component Base SBSContent (%)a 70.0 0.0

Test Targetb

Fluidity time (sec) 4–20Indentation depth (mm) 1–4Dynamic stability (cycles/mm) >300

a Asphalt binder and component contents are represented by weight of the mixture ab The target values were adapted from previous studies [4,14].

cycle. Initial flexural stiffness of a beam specimen (380 mm long � 63 mmwide � 500 mm high) was calculated at the 50th loading cycle. According to ASTMD7460-10 testing standard, fatigue failure is defined as the number of loading cy-cles when the initial stiffness is reduced by 50%. In this study, however, the 60%of stiffness reduction was also tried as a fatigue criterion because some specimenssuch as SBS modified mixtures did not show any significant fracture failure evenafter 50% stiffness reduction.

A haversine loading without rest period was repeatedly applied to a specimenuntil the 60% of stiffness reduction was obtained. Relatively high levels of strainamplitudes, 800, 1000, and 1200 le, were used in the fatigue tests to simulate highlevel of bending developed in long-span bridge deck pavements [6]. It was alsofound from preliminary tests that the fatigue life of the mastic asphalt mixtureswas too long to be measured properly at an initial strain level from 300 to 600 le.

The FPB strength test was conducted at �10 �C and 20 �C of temperatures fol-lowing KS (Korean Standard) F 2395 testing procedure [23]. A compressive loadingwith the constant speed of 50 mm/min was applied to a specimen (300 mm long,50 mm wide, and 50 mm high) until the compressive load was reduced by 20%from its peak. Then, bending strength was calculated using the peak load devel-oped during the strength test. Failure strain corresponding to the peak load wasalso determined. Using the FPB strength tests data, toughness values were calcu-lated to evaluate crack resistance of the specimens at low and intermediatetemperatures.

The IDT strength tests were conducted on the control and SBS modified masticasphalt mixtures at 20 �C [24]. The IDT tests were finished once horizontal tensilestress was reduced by 20% from its peak. Tensile strain and stress at the center ofthe specimen were calculated. Since the IDT strength test was conducted to evalu-ate the crack development, a relatively slower loading rate of 1.0 mm/min was ap-plied to capture crack images and only horizontal strain at the center of thespecimen was measured. Crack development during the IDT test was monitoredusing two digital cameras. The cameras were placed in front and back of an IDTspecimen. During the tests, digital images were captured every second. The cracklength of each crack was measured using a commercial image processing program,Scion Image. Each crack was manually marked from the beginning to the end of thecrack path and then the crack length was automatically calculated. An averagevalue of crack lengths obtained from both sides was used in the analysis. Detailedtest procedures were described in a previous study [25].

SBS modified

8.3TLA Base SBS TLA30.0 76.5 8.5 15.0

Control SBS modified

19 61.86 1.80340 321

nd by weight of the asphalt binder, respectively.

T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916 911

4. Binder test results

4.1. Effect of SBS type

The effect of the SBS type on the behavior of the asphalt binderwas evaluated. The content of the five SBS modifiers was fixed at8.0% by weight of the asphalt binder and no TLA was used. Penetra-tion, ring and ball temperature, and ductility test results before andafter thin film oven aging of the base and SBS modified asphalt bin-der were summarized in Table 3. Targets in the table were given formastic asphalt binder used in the wearing surface and base layersof deck pavements [14,26].

When the SBS modifiers were added to the base asphalt binder,the penetration depth before the RTFO was reduced by approxi-mately 30% as shown in Table 3. On the other hand, the penetrationratio was increased by approximately 35% regardless of the SBStypes. Ductility of the SBS modified asphalt binder before and afterthe RTFO was lowered by 40–50%, respectively and still satisfiedthe target requirements. Overall, the addition of the SBS modifiermade the binder consistency harder.

The performance of type C, D, and E SBS modifiers having nodouble bonds was better than the A and B type SBS modifiers: lowerpenetration at 25 �C before and after the RTFO and low-tempera-ture stiffness after the PAV. Especially, the effect of the antioxidantin type C, D, and E SBS modifiers on the asphalt binder propertiesafter aging was meaningful. When type A and B SBS modifiers areused, the low-temperature stiffness was greater than its limit of350 MPa. However, the low-temperature stiffness of type C, D,and E SBS modified asphalt was reduced by approximately 25%.

4.2. Effect of SBS modifier content

The effect of the SBS modifier content on the asphalt binderproperties was evaluated. Only type C SBS modifier was selectedfor this evaluation and SBS modifier contents were 8%, 10%, and12%. Fig. 3 shows the asphalt binder properties with respect to typeC SBS contents and corresponding limits which were adapted fromprevious studies [4,14,26]. Parenthesis in Fig. 3(a)and (b) repre-sents penetration ratio (%) and softening point change after theRTFO, respectively. When 8% of the SBS content was used, someof the binder requirements were not satisfied. However, increasingthe SBS content up to 10% and 12% improved the asphalt binderproperties satisfying all the requirements except the softeningpoint change as shown in Fig. 3(b).

Among the binder properties, the effect of the SBS content onthe softening point before the RTFO, the ductility after the RTFO,and the low-temperature stiffness after the PAV was relatively sig-nificant. The increase of the SBS content from 8% to 12% resulted inthe increase of the original softening point from 63 �C to 87 �C(38%), the increase of the RTFO ductility from 52 cm to 68 cm(31%), and the decrease of the PAV low-temperature stiffness from358 MPa to 150 MPa (�58%). Thus, the effect the use of 10% and

Table 3Binder test results for the base and SBS modified asphalt binder.

Tests Targeta Base

Original Penetration @ 25 �C (dm) 20–40 47Softening point (�C) 70–90 51Ductility @ 25 �C (cm) P50 140

RTFO Penetration ratio (%) P65 57Softening point change (�C) �2–8 4Ductility @ 25 �C (cm) P10 100

PAV Stiffness @ �16 �C (MPa) 6350 361

a The target values were adapted from previous studies [4,14,26].

12% of type C SBS modifiers in the asphalt binder was the most sig-nificant in improving crack resistance at low temperature.

The compatibility of the SBS modifiers in the asphalt matrix wasinvestigated using optical microscopy. Fig. 4 shows the morphol-ogy development of the SBS modified asphalt (4%, 8%, and 12%)at a temperature of 160 �C. When 4% and 8% SBS modifiers areused, the SBS modifiers (white dots in the figure) were dispersedin the asphalt matrix. A large network structure was constructedat a magnification of 500 when 12% of the SBS modifiers are used.When 30% of TLA is added to the SBS modified asphalt binder, itseemed that the SBS modifiers were distributed uniformly with amedium network structure while TLA could not be observed inthe fluorescence method.

4.3. Effect of TLA contents

A certain amount of TLA can be also used in the SBS modifiedasphalt binder. The effect of the TLA on the binder propertieswas evaluated to determine a proper amount of TLA. The TLA con-tents evaluated were 0%, 10%, 20%, and 30% by weight of the as-phalt binder where the C type SBS content was fixed as 10% byweight of the base asphalt. When the TLA content is 30%, no SBSmodifier was used, i.e. it is conventional TLA content in mastic as-phalt. Binder test results were summarized in Table 4.

As the TLA content increased and the SBS modifier content de-creased, binder properties changed gradually in most tests. Withthe increase of TLA contents, original penetration decreased; origi-nal ductility increased while ductility after the RTFO decreased;penetration ratio and softening point change after the RTFO in-creased. From these observations, the effect of TLA on the penetra-tion before and after the RTFO was significant due to higherviscosity of TLA. When 30% of TLA is used without the SBS modifier,its original softening point was 56 �C which is 14 �C lower than thelower limit of 70 �C and its softening point change exceeded therequirement. It means that the asphalt mixtures with 30% TLAcan be vulnerable to permanent deformation at high temperature.In addition, as the TLA content increased, ductility reduction afterthe RTFO was significant, e.g., 79–62 cm for 0% of TLA and 99–27 cm for 30% of TLA. Thus, the mastic asphalt mixtures with only30% of TLA can susceptible to cracking due to aging. Thus, it is bet-ter to replace a part of TLA by the SBS modifier to improve ruttingand cracking resistance of mastic asphalt mixtures.

5. Mixture test results

5.1. Four-point bending (FPB) test

5.1.1. FPB strength testThe FPB strength tests were conducted to measure flexural

strength and toughness at a low temperature of �10 �C and inter-mediate temperature of 20 �C. Fig. 5 shows flexural stress versusbending strain of the control and SBS modified mastic asphalt

SBS type (8%)

A B C D E

34 34 33 33 3380 80 63 63 6371 71 80 80 80

73 76 78 84 72�8 �6 1 3 861 77 52 61 60

448 468 358 332 340

0

20

40

60

80

0 8 10 12

Pen

etra

tion

at

25o C

(dm

)

SBS content (%)

Original RTFO

(57) (78) (71)(67)

(a)

0

30

60

90

120

0 8 10 12

Soft

enin

g po

int

(o C)

SBS content (%)

Original RTFO

Upper limit

Lower limit

(4)

(1)

(-2)

(-4)

(b)

0

40

80

120

160

0 8 10 12

Duc

tilit

y at

25

o C (

cm)

SBS content (%)

Original RTFO

50

10

(c)

0

100

200

300

400

0 8 10 12

Stif

fnes

s at

- 16

o C (

MP

a)

SBS content (%)

PAV350

(d)

Fig. 3. Asphalt binder properties with respect to SBS C modifier contents: (a) penetration, (b) softening point, (c) ductility, and (d) stiffness.

Fig. 4. Morphology development of asphalt binder with (a) 4%, (b) 8%, (c) 12% SBS modifiers, and (d) 10% SBS modifiers with 30% TLA.

912 T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916

mixtures. Compared to the control mastic asphalt mixture, the SBSmodified mastic asphalt mixture have 14.6 MPa, 20% greater flex-ural strength at �10 �C but 5.4 MPa, 40% lower flexural strengthat 20 �C. However, bending strain at the failure of the SBS modifiedmastic asphalt mixture was greater than that of the control mastic

asphalt mixture: 4% at �10 �C and 42% at 20 �C. Thus the SBSmodified mastic asphalt mixture can be considered to have higherflexural resistance at �10 �C and comparable at 20 �C.

In addition, flexural toughness was used to evaluate the fractureresistance of the SBS modified mastic asphalt mixture. Fracture

Table 4Binder test results for various TLA contents.

Category Targetb Base TLA content (%)

0 10 20 30a

SBS modifier (%)

10 9 8 0a

OriginalPenetration @ 25 �C (dm) 20–40 33 39 27 20 24Softening point (�C) 70–90 63 82 78 80 56Ductility @ 25 �C (cm) P50 80 79 79 90 99

RTFOPenetration ratio (%) P65 78 71 78 83 100Softening point change (�C) �2–8 +1 �12 �1 5.3 10Ductility @ 25 �C (cm) P10 52 62 48 54 27

PAVStiffness @ �16 �C (MPa) 6350 361 173 – – –

a Conventional TLA/asphalt binder ratio where no SBS modifier was used in theasphalt binder.

b The target values were adapted from previous studies [4,14,26].

T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916 913

energy or toughness is composed of the total dissipated energy andelastic strain energy. The strain energy represents the energy con-sumed to deform the material while the total dissipated energy isthe energy used to initiate cracks in the material. Therefore, the to-tal dissipated energy may be proportional to the fracture resistanceof the asphalt concrete. To calculate the strain energy, one mayneed viscoelastic properties of asphalt mixtures. However, in thisstudy, the viscoelastic properties were not measured. Thus, thetoughness was used to relatively evaluate crack resistance.

Toughness was calculated based on the area under the strain–stress curves until its strength, marked in Fig. 5(b). Fig. 6 showsthe toughness of the control and SBS modified mastic asphalt mix-tures. Toughness at �10 �C and 20 �C are 9.1 N m and 29.3 N m,respectively, for the control mastic asphalt mixture; 12.0 N mand 34.6 N m, respectively, for the SBS modified mastic asphalt

-10oC

Control

SBS modified

Bending strain (μμε)

Fle

xura

l str

ess

(MP

a)

20

15

10

5

0 0 2.0 4.0 6.0 8.0 10.0

Fle

xura

l str

ess

(MP

a)

20

15

10

5

0

(a)

(b)

Bending strain (με)

20oC

Control

SBS modified

Toughness

0 10.0 20.0 30.0 40.0 50.0

Fig. 5. Bending strain–flexural stress of the FPB strength test at: (a) �10 �C and (b)20 �C.

mixture. The use of the SBS modifier improved the toughness ofthe mastic asphalt mixture by 31.4% at �10 �C and 18.0% at 20 �Cdue to higher ductility. Hence, it may be concluded that the SBSmodifier gives more flexibility to the mastic asphalt mixture whichenables to resist against cracking due to bending.

5.1.2. FPB fatigue testWith the increase of load repetitions in the FPB fatigue tests,

flexural stiffness drops suddenly at the beginning, tends to de-crease gradually, and finally drops to be fractured. As shown inFig. 7(a), the control mastic asphalt mixture followed the typicalfatigue behavior at the lowest strain level of 800 le. The SBS mod-ified mastic asphalt mixtures followed the similar behavior of thecontrol one until 0.3 million cycles at the fatigue failure of the con-trol one. After then, the stiffness of the SBS modified mastic asphaltmixtures maintained around 50% stiffness up to 1.0 million cyclesat the 800 le level and 40% stiffness up to 0.5 million cycles at thehigher strain levels. At the end of the fatigue tests, no visible crackwas observed in the beam specimens of the SBS modified masticasphalt mixtures while the beam specimen of the control onewas broken completely by a distinct single crack. This implies thatthe SBS modifier could prevent rapid fatigue crack development atthe high strain levels.

Based on two flexural stiffness levels of 40% and 50%, the num-ber of loading cycles to failure, Nf was determined. As shown inFig. 8, the SBS modified mastic asphalt mixtures had greater fatiguelife than the control ones at both stiffness levels. For example atthe 1000 le level, Nf of the control and SBS modified mastic asphaltmixtures are 17,072 and 80,361 at the 50% stiffness level; 21,665and 639,903 at the 40% stiffness level. The Nf of the SBS modifiedmastic asphalt mixtures at the 40% stiffness level was significantlygreater than that at the 50% stiffness level. Also, as the strain levelincreased from 800 to 1200 le, the Nf ratio of the SBS modified tocontrol mastic asphalt mixtures increased significantly from 19 to273. Thus the SBS modifier could enhance the fatigue life of the SBSmastic asphalt mixture considerably, especially in controlling fati-gue crack development in a secondary or tertiary fatigue stagerather than in a primary fatigue stage.

5.2. Indirect tensile strength test

Since a visible crack was not observed in the SBS modified mas-tic asphalt mixtures from the FPB fatigue tests, IDT strength testswere conducted on the two mixtures until complete failure to eval-uate crack initiation and propagation behavior of the mixtures.

5.2.1. Strength and toughnessThe IDT strength test results at 20 �C were shown in Fig. 9. The

maximum tensile strength of the SBS modified mastic asphalt

9.1

29.3

12.0

34.6

31.4

18.0

0

10

20

30

40

50

0

10

20

30

40

50

-10 20

Toug

hnes

s im

prov

emen

t (%

)

Toug

hnes

s (N

·m)

Temperature (oC)

Control SBS modified

Improvement

Fig. 6. Comparison of toughness of the control and SBS modified mastic asphaltmixtures at �10 �C and 20 �C.

Nor

mal

ized

fle

xura

lst

iffn

ess

Nor

mal

ized

fle

xura

lst

iffn

ess

Nor

mal

ized

fle

xura

lst

iffn

ess

(a)

(b)

(c)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0

0.2

0.1

Control

Control

Control

0.4

0.2

SBS modified

SBS modified

SBS modified

0.6

0.3

0.8

0.4

Number of loading cycles (×10 )6

Number of loading cycles (×10 )6

Number of loading cycles (×10 )6

1.0

0.5

800

1000

1200

0.0 1.2

με

με

0.6

0.0 0.1 0.2 0.3 0.4 0.5 0.6

με

Fig. 7. Normalized flexural stiffness reduction with respect to the number ofloading cycles at the strain levels of (a) 800, (b) 1000, and (c) 1200 le.

1.0E+02

1.0E+03

1.0E+04

1.0E+00 1.0E+02 1.0E+04 1.0E+06

Number of loading cycles (103)

T = 20 oC50% stiffness

(a)

SBS modified (R2= 0.949)

Control (R2 = 0.809)

1.0E+02

1.0E+03

1.0E+04

1.0E+00 1.0E+02 1.0E+04 1.0E+06

Number of loading cycles (103)

T = 20 oC40% stiffness

(b)

SBS modified (R2 = 0.821)

Control (R2 = 0.827)

Init

ial s

trai

n (μ

)εIn

itia

l str

ain

(μ)ε

Fig. 8. Fatigue life versus initial strain of the control and SBS modified masticasphalt mixtures at (a) 50% and (b) 40% stiffness.

0

200

400

600

800

0 100 200 300 400 500 600

Tens

ile S

tres

s (k

Pa)

IDT strength testT = 20 oC

Control

SBS modified

914 T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916

mixture was 635.5 kPa (2.9% in coefficient of variation (CV)) whichis 36% lower than that of the control mastic asphalt mixture,407.3 kPa (4.1% in CV). Unlike the FPB strength test results, IDTtoughness computed with horizontal tensile stress and strainwas 37% lower than that of the control one. It means that the useof the SBS modifier in the mastic asphalt mixtures was not benefi-cial in retarding fatigue crack initiation.

Loading time (sec)

Fig. 9. Horizontal tensile stress variations with respect to loading time in the IDTstrength test.

5.2.2. Crack growthDuring the IDT strength test, a crack monitoring system pro-

posed in a previous study [25] was used to measure crack lengthin a designated area whose size is 130 mm in height and 50 mmin width as shown in Fig. 10(a). Fig. 10(b) shows crack images forthe control and SBS modified mastic asphalt mixtures at the peakload of the IDT strength test. In the control mastic asphalt mixture,an apparent single crack was observed through the center line ofthe designated area. In the SBS modified mastic asphalt mixture,relatively tiny and unlinked cracks were distributed in the upperand lower designated area. Based on these different crack patternsobserved in the control and SBS modified mastic asphalt mixtures,one can conclude that the SBS modifier spreads fatigue damagesover larger area, leading to less stress concentration and higherfatigue crack resistance.

Fig. 11 shows crack length variations with loading time for thecontrol and SBS modified mastic asphalt mixtures. Wherein, thecrack length is a sum of length of visible individual cracks devel-oped in the designated area. Until a loading time of 300 s, the cracklength of the SBS modified mastic asphalt mixtures is approxi-mately 2–4 mm, close to that of the control mix. It implies thatthe SBS modifier does not affect fatigue crack initiation of the mas-tic asphalt mixtures at 20 �C.

After that, crack growth rate, crack length increase per second,of the control mix became faster than that before. For example,the crack growth rate of the control mix was less than 0.01 mm/suntil 300 s and jumped to 0.23 mm/s at 350 s where micro-cracks

Control

130 mm Designatedarea

(b)(a)

SBS modified

50 mm

Fig. 10. (a) Designated area in the IDT specimen and (b) crack images at the peak load the IDT strength test [25].

0

5

10

15

20

25

30

100 200 300 400 500 600

Cra

ck le

ngth

(cm

)

Loading time (s)

IDT strength testT = 20 oCLoading rate : 1mm/min

Control

SBS modified

Fig. 11. Fatigue crack growth with respect to loading time during the IDT strengthtest on the control and SBS modified mastic asphalt mixtures.

T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916 915

began to merged and structural integrity of the IDT specimen gotloosen significantly by the fatigue damage. For the SBS modifiedmastic asphalt mixtures, the crack growth rate was less than0.1 mm/s until 500 s and became two-folds at 550 s. It could beresulted from the SBS modifiers enhancing the fatigue crack resis-tance of the mastic asphalt mixtures by reducing the accumulationrate of fatigue damage [10]. Therefore, the benefit of the SBSmodifier in the mastic asphalt mixtures was insignificant in delay-ing the initiation of fatigue cracks but substantial in retarding thefatigue crack growth under indirect tensile mode of loading. It isreminded that a similar behavior was observed in the FPB fatiguetest results shown in Fig. 7.

6. Conclusions

In this study, several laboratory tests were conducted to evalu-ate the effect of SBS modifiers on the fatigue performance of masticasphalt mixtures. The mechanical properties of the SBS modifiedasphalt binder and mixtures were improved, especially in low-temperature crack resistance and fatigue life. Besides, importantconclusions were achieved as follows:

– The SBS modifiers could reduce TLA contents in mastic asphaltmixtures by 50% and improve physical and mechanical proper-ties of the mastic asphalt mixtures.

– It was observed from the binder tests that removal of the C@Cdouble bonds in the SBS modifier was helpful for enhancinganti-aging performance of the mastic asphalt mixture: lowerlow-temperature stiffness at �10 �C after the PAV and higherductility at 25 �C after the RTFO.

– The SBS modified asphalt mixtures had better crack resistanceby as factor of 1.32 at �10 �C and 1.18 at 20 �C and significanthigher fatigue life, more than 100 times at a higher strain levelof 800–1200 le.

– The benefit of the SBS modifier in the asphalt mixtures wasmore substantial in retarding fatigue crack growth than crackinitiation under indirect tensile mode of loading.

– This enhancement of the SBS modified asphalt mixtures canlead to the reduction of the thickness and self-weight of thedeck pavement, resulting in saving construction cost for thebridge and cables as well as the deck pavement.

It is noted that these conclusions were based only on a limitednumber of samples and conditions. Further studies will be fol-lowed to investigate the effect of the SBS modifiers on permanentdeformation, moisture damage and long-term performance of mas-tic asphalt mixtures in the field.

Acknowledgements

The authors gave appreciation on a partial support by SamsungC&T for a project entitled ‘‘A Study on the Development of Syn-thetic Gussasphalt Deck Pavements to Reduce Construction Costof Cables in Long-Span Bridges’’ and by the Carbon Neutral RoadTechnologies Development Research Program through the KoreaInstitute of Construction & Transportation Technology Evaluationand planning (KICTEP) and the Minister of Land, Transport andMaritime Affairs (MLTM).

References

[1] Suldien C. The design of modern steel bridges. Boston: London Edinburgh;1992.

[2] Li X-L, Chen Y-L. New composite pavement system for orthotropic steel bridgedecks. In: Proceedings of 2009 GeoHunan international conference, Hunan,China; 3–6 August 2009. p. 75–84.

916 T.W. Kim et al. / Construction and Building Materials 48 (2013) 908–916

[3] Park H-M, Choi J-Y, Lee H-J, Hwang E-Y. Performance evaluation of a highdurability asphalt binder and a high durability asphalt mixture for bridge deckpavements. Constr Build Mater 2009;23(1):219–25.

[4] Yang Y-K, Yang J-H, Kim K-M, Kim S-J. Bridge surface pavement of the Incheonbridge. In: Proceedings of international commemorative symposium for theInchon bridge, Korea; 23 September 2009.

[5] Medani TO. Design principles of surfacings in orthotropic steel bridge decks[dissertation]. The Netherlands: Delft University of Technology; 2006.

[6] Kim T-W, Baek J, Lee H-J. A study of orthotropic steel bridge deck pavementbehavior using 3D finite element analysis. In: Proceedings of 14th KoreaSociety of Road Engineers Conference, Seoul, Republic of Korea; 26 September2012. p. 41–6.

[7] Medani TO. Asphalt surfacing applied to orthotropic steel bridge decks-Aliterature review. Report 7–01-127-1. The Netherlands: Delft University ofTechnology; 2001.

[8] Connor R, Fisher J, Gatti W, Gopalaratnam V, Kozy B, Leshko B, et al. Manual fordesign, construction, and maintenance of orthotropic steel deckbridges. Washington (DC): FHWA; 2012.

[9] Edwards Y, Westergren P. Polymer modified waterproofing and pavementsystem for the high coast bridge in Sweden. Swedish: NRTRI; 2001.

[10] Kim B, Roque R, Birgisson B. Effect of styrene butadiene styrene modifier oncracking resistance of asphalt mixture. Transpor Res Rec 2003;1829(2):8–15.

[11] Kim TW, Baek J, Lee HJ, Lee SY. Effect of pavement design parameters on thebehaviour of orthotropic steel bridge deck pavements under traffic loading. IntJ Pavement Eng 2013 [in press].

[12] Hicks RG, Dussek IJ, Seim C. Asphalt surface on steel bridge Decks. TransportRes Rec 2000;1740(17):135–42.

[13] Matar S, Hatch LF. Chemistry of petrochemical processes. 2nded. Houston: Gulf Publishing; 2000.

[14] Chen J-S, Liao M-C, Huang C-C. Evaluation of Gussasphalt applied to steel decksurfacing. In: Proceedings of T&DI congress 2011, Chicago, USA; 13–16 March2011. p. 462–71.

[15] Kim W-S, Lee S-H, Lee G-G, Yang Y-K. Case study of Gussasphalt constructionof seongsu grand bridge. Int J Highway Eng 2004;4(1):247–52 [Korean].

[16] American Standard. Standard test method for sieve analysis of fine and coarseaggregates. ASTM C1360-06.

[17] European Standard. Bituminous mixtures – test methods for hot mix asphalt,Part 20: Indentation using cube or cylindrical specimens. EN 12697-20.

[18] European Standard. Bituminous mixtures – test methods for hot mix asphalt,Part 22: Wheel tracking. EN 12697-22.

[19] Castro M. Structural design of asphalt pavement on concrete bridges. Can JCivil Eng 2004;31(4):695–702.

[20] American Standard. Standard test method for determining fatigue failure ofcompacted asphalt concrete subjected to repeated flexural bending. ASTMD7460-10.

[21] European Standard. Bituminous mixtures – test methods for hot mix asphalt,Part 26: Resistance to fatigue. EN 12697-26.

[22] Artamendi I, Khalid H. Different approaches to depict fatigue of bituminousmaterials. In: Proceedings of the 15th European conference of fracture –advanced fracture mechanics for life and safety assessments, Paper ECF-15,Sweden; August 11–13 2004.

[23] Korean Standard. Standard testing method for bending strength of asphaltmixtures. KS F 2395. [Korean].

[24] American Standard. Standard method of test for determining the creepcompliance and strength of hot-mix asphalt using the indirect tensile testdevice. AASHTO T 322.

[25] Nguyen MT, Lee H-J, Baek J. Fatigue analysis of asphalt concrete under indirecttensile mode of loading using crack Images. J Test Eval 2013;43(1):1–11.

[26] Lee K-H, Lee K-H. Laboratory investigation into mix design procedure andcriteria of Gussasphalt. Proc Korea Soc Road Eng, Republic of Korea2000;2(1):69–73 [Korean].