Construction and Building Materials 24 (2010) 1193–1200

Contents lists available at ScienceDirect

Construction and Building Materials

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

Evaluating permanent deformation in asphalt rubber mixtures

Liseane P.T.L. Fontes a, Glicério Trichês a, Jorge C. Pais b,*, Paulo A.A. Pereira b

a University Federal of Santa Catarina, Department of Civil Engineering, Rua João Pio Duarte, 88040-970 Florianópolis, Santa Catarina, Brazilb University of Minho, Department of Civil Engineering, Campus of Azurém, 4800-058 Guimarães, Portugal

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

Article history:Received 27 April 2009Received in revised form 7 December 2009Accepted 16 December 2009Available online 13 January 2010

Keywords:Permanent deformationAsphalt rubberShear testWheel tracking

0950-0618/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.conbuildmat.2009.12.021

* Corresponding author. Tel.: +351 253 510 208; faE-mail address: jpais@civil.uminho.pt (J.C. Pais).

Permanent deformation or rutting, one of the most important distresses in flexible pavements, has longbeen a problem in asphalt mixtures, mainly in countries with high temperatures such as Brazil. Through-out the years, researchers have used different test methods to estimate the rutting performance ofasphalt mixtures. One of the alternatives to reduce permanent deformation in asphalt pavement layersis through the use of mixtures produced with asphalt rubber. Crumb rubber from waste tires introducedinto the asphalt is one of suitable application to dispose the tires and used as an additive to enhance theproperties of the conventional asphalts. This work aims at comparing the rutting performance of asphaltrubber mixtures (with dense and gap-graded aggregate gradation) with the conventional dense gradedmixture most used in Brazil. The asphalt rubber mixtures were produced by the wet process using con-tinuous blend and terminal blend asphalt rubber. To study their performance, two laboratory tests, theRepeated Simple Shear Test at Constant Height (RSST-CH) and the Accelerated Pavement Testing Simu-lator Test (wheel tracking) were carried out. The testing results confirmed that the use of asphalt rubberbinder improves significantly the resistance to rutting. The highest resistance is presented by the mix-tures produced with continuous blend binders and gap-graded aggregate gradation. The results of bothtesting apparatus can be correlated by a linear relationship. The testing results allowed concluding thatthe characteristics of the asphalt rubber binders cannot be used to predict the permanent deformationresistance of the asphalt rubber mixtures.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Waste tires constitute a serious environmental problem thatmany countries have to face as they accumulate rapidly and theyare not easily disposed off. Many approaches have been consideredin recent years for treating and improving the conventional as-phalts, such as the introduction of the additives in order to improvetheir properties. The use of crumb rubber from waste tires in as-phalt, giving origin to the term asphalt rubber, has been an alterna-tive to minimize their ecological impact and, simultaneously, toimprove the mechanical properties of the asphalt mixtures.

The recycling of crumb rubber-modified materials has been anarea of interest since crumb rubber modifier was first used in as-phalt paving materials over 40 years ago in the United States. Somestate departments of transportation have used rubberized asphaltconcrete materials in limited recycling experiments or demonstra-tion projects. The respective studies include different types of wetprocess binders and/or various gradations of crumb rubber-modi-fied as an aggregate substitute (dry process). The results show that

ll rights reserved.

x: +351 253 510 217.

a wide range of crumb rubber-modified paving materials can besuccessfully reclaimed and recycled [1].

To produce asphalt rubber, the crumb rubber has to be cuttingand scraping into small sizes down to powder particles and then,added into the conventional asphalt. Processing scrap tires intocrumb rubber can be accomplished through the ambient grindingand the cryogenic grinding technologies. In the ambient ground-rubber processing, scrap tire rubber is ground or processed at orabove ordinary room temperature. Cryogenic processing uses li-quid nitrogen to freeze tire chips or rubber particles prior to sizereduction (�120 �C). A very fine ground crumb rubber modifier istypically used in crumb rubber asphalt [2].

Besides the size, the main difference between ambient andcryogenic rubber is the morphology of the particles. Rubber parti-cles from ambient process generally have a porous or fluffy appear-ance, whereas produced by the cryogenic process, the surface ofthe particles are glasslike [2].

Shen et al. [3] evaluated many variables of crumb rubber thataffecting their interaction with the binder, including the surfacearea, and they observed that the surface area of the ambient crumbrubber was twice as large as that of the cryogenic one, leading to amuch higher complex modulus and phase angle of the asphalt rub-ber binders. Also, the phase angle and complex modulus were

1194 L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200

affected by both the surface area and average particle size. How-ever, the average size is the predominating factor.

Crumb rubber can be added into asphalt through the dry pro-cess or the wet process. In the dry process, the crumb rubber ismixed together with the aggregates prior to the addition of the as-phalt. In this process, the crumb rubber is used as an aggregate.

The wet process, the only one used in this work, includes anymethod by which crumb rubber is blended with conventional as-phalt before incorporating the binder into the asphalt paving mate-rials. Asphalt rubber binder results from the chemical reaction of amix of liquid asphalt binder with 15–22% crumb rubber obtainedfrom used tires and added to liquid asphalt. It reacts at high tem-peratures prior to being mixed with the aggregate. Potential bene-fits of asphalt rubber binders obtained through the wet processinclude improvement of the fatigue life of a pavement, enhancedresistance to permanent deformation and reduction of crack prop-agation when compared to other binders.

This work intends to evaluate the permanent deformationperformance of gap and dense graded mixtures containing an as-phalt rubber binder prepared through wet processes (continuousblend and terminal blend) and crumb rubber obtained from thecryogenic and ambient processes. The behaviour of asphalt rub-ber mixtures was compared to that of a conventional mixture.The tests were carried out through two rutting tests, such asRSST-CH (Repeated Simple Shear Test at Constant Height) andAccelerated Pavement Testing Simulator (wheel tracking). Theasphalt rubber binders were evaluated through characterizationtests (penetration, softening point, resilience) and viscosity(Brookfield viscometer) as well. From this, the tested resultsare compared and analyzed statistically in order to elucidatethe various effects and relationships of the rutting laboratoryperformance.

2. Permanent deformation in asphalt pavements

The permanent deformation (rutting) of asphalt pavements hasa major impact on the performance of a pavement throughout itslife. Rutting not only reduces the useful service life of pavements,but it may also affect basic vehicle handling manoeuvres, whichcan be hazardous to highway users. Rutting develops graduallyas the number of load applications increases. Rutting appears aslongitudinal depressions in the wheel paths and small upheavalsto the sides. It is caused by a combination of densification andshear deformation. These depressions or ruts are important be-cause, if the surface is impervious, the ruts trap water causinghydroplaning (particularly for passenger cars), which is extremelydangerous and as the ruts progress in depth, steering becomesincreasingly difficult, leading to added safety concerns [4].

Zaniewski and Harsha [5] assert that densification is the furthercompaction of asphalt mixtures pavements by traffic after con-struction. When compaction is poor, the channelized traffic pro-vides a repeated kneading action in the wheel track areas andcompletes the consolidation. A substantial amount of rutting canoccur if thick layers of asphalt are consolidated by the traffic.

The lateral plastic flow of the asphalt mixtures due to wheeltracks results in rutting. The use of excessive asphalt cement inthe mix causes the loss of internal friction between the aggregateparticles, what provokes that traffic loads are supported by the as-phalt cement rather than by the aggregate structure. Plastic flowcan also occur due to a lack of angularity of the aggregates andto an insufficient surface texture that is needed for inter-particlefriction. Plastic flow can be minimized by using large size aggre-gates, angular and rough textured coarse and fine aggregate, stifferbinders, as well as by providing suitable compaction during con-struction [6].

Additionally, for the assessment of the rut depth it is also nec-essary to recognize the evolution of the void content in a pavementasphalt layer. When the air-void content drops below 2–3%, thebinder acts as a lubricant between the aggregates and reducespoint-to-point contact pressures. Permanent deformation of as-phalt-aggregate mixes is strongly controlled by the plastic compo-nent due to the aggregate skeleton. This causes permanentdeformation changes either in volume or in shear, what mostly oc-curs in hotter days or because of heavy loads [7].

Under hot conditions or under sustained loads, asphalt cementsbehave like viscous liquids and flow. Viscous liquids, such as hotasphalt, are sometimes called plastic because, once they start flow-ing, they do not return to their original position. This is why, in hotweather, some asphalt pavements flow under repeated wheelloads and wheel path ruts appear. However, rutting in asphaltpavements during hot weather is also influenced by the propertiesof aggregates and it is probably more correct to say that the asphaltmixture is behaving like a plastic mixture [8].

According to Bennert et al. [9], in general, the addition of crumbrubber to asphalt mixtures and the proper design and field imple-mentation of the asphalt rubber mixtures expand the workingrange of the conventional mixtures providing reduction of ruttingat high temperatures, reduction of fatigue cracking at intermediatetemperatures, reduction of thermal cracking and minimization ofthe potential for age hardening.

At high temperatures, the asphalt binder tends to flow easierdue to the natural decrease of viscosity associated with highertemperatures. This condition creates a ‘‘softer” asphalt mixture,which is prone to rutting. The addition of crumb rubber to the as-phalt mixture provides extra viscosity, what contributes to thestiffening of the HMA at higher temperatures [10].

Asphalt rubber mixtures generally have greater fatigue livesdue to the higher binder contents and higher rutting resistanceas a result of their higher binder viscosity. They are also generallymore permeable than conventional mixes, what reduces the splashand spray during periods of rain [11].

A study about the aging of the asphalt rubber was carried out byLee et al. [12], in which the rolling thin-film oven test (RTFOT;163 �C for 85 min) and the short-term oven aging (STOA) wereused in the laboratory to represent the aging of an asphalt binderduring plant mixing, transportation and paving. The authors com-pared the aging effects of the RTFOT and the STOA methods usingthe gel permeation chromatography (GPC), concluding that theRTFOT method has less aging effect on the binders than the STOAmethods for asphalt mixtures prepared in the laboratory. The long-er the aging time in the RTFOT led to an increase in the high tem-perature viscosity and the high failure temperature of asphaltbinders, except for rubber-modified binders.

Doh et al. [13] conducted a study in which they found out thehigh correlation between rutting resistance and the strength prop-erty of asphalt mixtures and assert that it might be a simple test topredict the deformation resistance of asphalt mixtures at hightemperatures.

3. Mixtures characterization

3.1. Crumb rubber and asphalt rubber

Two types of rubber, obtained through the ambient and thecryogenic processes, were used to produce asphalt rubber bind-ers for this work. The ambient crumb rubber was producedin Brazil, whereas the cryogenic crumb rubber was produced inPortugal. The use of cryogenic crumb rubber from Portugal wasintended to be compared to the crumb rubber produced inBrazil.

Table 1ADOT A-R specifications and rubber gradations.

Sieves(mm)

ADOT A-R (% passing) Ambient (% passing) Cryogenic (% passing)

2.00 100–100 100 1001.18 65–100 99 990.60 20–100 96 900.30 0–45 44 200075 0–5 4 3

70

80

); ); 4000

4500

P)

Table 4Binder and void content of the asphalt mixtures.

Name Asphalt Gradation Binder content (%) Void content (%)

MGTB1 ARTB1 Gap 8.5 6.0MDTB2 ARTB2 Dense 7.0 5.0MGCB1 ARCB1 Gap 8.0 6.0MDCB2 ARCB2 Dense 7.0 5.0MDCO 50/70 pen Dense 5.5 4.0

L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200 1195

The gradation analysis was carried out in accordance with therequirements of the ASTM C 136, amended by the Greenbook[14] recommendations. The rubber gradations followed the Ari-zona Department of Transportation (ADOT) requirements type B,as presented in Table 1.

Four asphalt rubbers were produced using ambient and cryo-genic rubbers. 50/70 and 35/50 pen asphalts were used to producethe asphalt rubber binders. The asphalt rubber from the continu-ous blend was produced with 17% of rubber content at a tempera-ture of 180 �C during 90 min. The terminal blend asphalt rubberswere produced at an industrial plant with 15% and 20% rubber con-tent. Table 2 presents the designations and the summary of eachasphalt rubber.

The asphalt rubber binders were characterized by penetration,softening point (ring and ball test), resilience and apparent viscos-ity tests. The hardening properties were also evaluated using therolling thin-film oven test (RTFOT). Considering that the conven-tional asphalt 50/70 pen (ACO) was used to produce the conven-tional mixture, this binder was tested as well. The results of theasphalt characterization tests can be observed in Fig. 1. Table 3 pre-sents the results of RTFOT.

The asphalt rubber binders produced by the continuous blend(ARCB1 and ARCB2) with 35/50 pen asphalt presented a highersoftening point temperature than the terminal blend asphalt rub-ber binders (ARTB1 and ARTB2). The amount of crumb rubber didnot influence these results. The same conclusion was drawn inthe resilience. As expected, the conventional asphalt (ACO) pre-sented a lower softening point and resilience, what indicates thatthis type of asphalt could produce mixtures with great thermalsusceptibility at high temperatures and lesser elastic properties.

Table 2Asphalt rubber features.

Designation Baseasphalt

Rubbertype

Rubber content(%)

Process

ARTB1 50/70 pen Ambient 20 Terminal blendARTB2 50/70 pen Ambient 15 Terminal blendARCB1 35/50 pen Cryogenic 21 Continuous

blendARCB2 35/50 pen Ambient 21 Continuous

blend

Table 3Results of RTFOT (ASTM D2872).

RTFOT 163 �C, 85 min ARTB1 ARTB2 ARCB1 ARCB2 ACO

Change of mass (%) 0.3 0.3 0.9 0.2 0.3Softening point elevation (�C) 1.0 2.9 17.2 11.2 4.3Penetration 25 �C, 100 g, 5 s

(0.1 mm)28.8 25.3 15.5 19.5 22.3

Apparent viscositya (cP), 175 �C 5350 1962 3025 8813 95.8Retained penetration (%) 72.0 60.2 92.2 99.0 43.3

a Brookfield viscometer, spindle number 27, 20 rpm.

The ARTB2 presented lower apparent viscosity, followed by ARTB2and ARCB1. The ARCB2 showed the highest viscosity.

3.2. Mixtures and specimens

Asphalt rubber mixtures were produced using dense and gapgradations, whereas the conventional mixture was produced witha dense graded, as presented in Table 4. The use of dense gradationin the asphalt rubber mixtures was due to the fact that only finecrumb rubber gradation was used to modify the asphalt. The useof dense graded gradation in asphalt rubber mixtures is a currentpractice in Brazil.

The binder content of the mixtures was evaluated according tothe Marshall method (ASTM D1559) which has been the mostwidely applied for designing and controlling paving mixtures inBrazil. Fig. 2 presents the gradation curves of the studied mixtures.

The dense asphalt rubber mixtures follow the aggregate grada-tion defined by the Asphalt Institute (mix type IV) in The AsphaltHandbook Manual Series no. 4, whereas the gap graded asphaltrubber mixtures follow that defined by the California Departmentof Transportation (Caltrans) SSP39-400 – ARHM-GG mixture (As-phalt Rubber Hot Mix Gap Graded). The conventional mixture fol-lowed the DNIT gradation (Brazilian Road Department, inPortuguese) specifications, which manage and establish the roadtechnical specifications in Brazil.

After being designed, the mixtures were produced and com-pacted in slabs using a cylinder with up-vibration to achieve

0

10

20

30

40

50

60

ARTB1 ARTB2 ARCB1 ARCB2 ACO

Pen

etra

tion

(0,0

1 m

mSo

ften

ing

poin

t (ºC

Res

ilien

ce (%

)

0500100015002000250030003500

App

aren

t vis

cosi

ty (c

Penetration Softening point Resilience Apparent viscosity

Fig. 1. Characterization tests of the asphalts.

0102030405060708090

100

0.01 0.1 1 10 100Sieves size (mm)

% p

assi

ng

Conventional

Caltrans (ARHM-GG)

Dense graded

Fig. 2. Aggregate gradation of studied mixtures.

Fig. 4. Wheel tracking slab (in A, dimensions in plant view; in B, the cross section view showing the configuration adopted for asphalt rubber mixtures in the test).

Fig. 3. RSST-CH specimen dimensions and glued to caps.

1196 L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200

the apparent density of the mixtures defined in the design.The slabs of asphalt mixtures were sawed to produce eightcylindrical specimens for RSST-CH tests. In this type of test,the specimens are glued to aluminium caps, as depicted inFig. 3. For wheel tracking tests, specimens were extracted fromslabs produced with two layers. The first layer, with an asphaltrubber mixture, was 3.0 cm thick and the remaining 5.0 cmwere built with the conventional mixture (MDCO), as illus-trated in Fig. 4.

Fig. 5. RSST-CH and Whee

4. Testing program

In this work the mechanical performance of the studied mix-tures was evaluated through the Repeated Simple Shear Test atConstant Height (RSST-CH) and the Accelerated Pavement TestingSimulator (wheel tracking). The tests were conducted at a tem-perature of 60 �C, to simulate the worst climate conditions towhich the mixtures are subjected when applied in pavementrehabilitation in the South of Brazil. That temperature is widely

l tracking equipments.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 1000 2000 3000 4000 5000Period

Pla

stic

she

ar s

trai

n

MDCO MGCB1 MGTB1 MDCB2 MDTB2

Fig. 6. Development of the plastic shear strain in the RSST-CH tests.

0

1

2

3

4

5

6

7

8

9

0 20 40 60 80 100 120

Def

orm

atio

n (m

m)

Time (minutes)

MDCO MDTB2 MGCB1 MGTB1 MDCB2

Fig. 7. Evaluation of the deformation in the wheel tracking test.

1E+05

1E+06

1E+07

ESA

L (8

0 kN

)

L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200 1197

used to perform wheel tracking tests and it is very close to therecommended temperature proposed by SHRP program to per-form the RSST-CH test, i.e. the average maximum temperaturefor seven consecutive days at 5 cm which, for the South of Brazil,is approximately 55 �C. The tests were performed at the sametemperature to allow the comparison of the results and the estab-lishment of a correlation between both.

4.1. Repeated Simple Shear Test At Constant Height (RSST-CH)

Shear deformations in pavements that have been appropriatelycompacted, caused primarily by large shear stresses in the upperportions of the asphalt-aggregate layer(s), are frequent [4]. Repeti-tive loading in the shear is required in order to accurately measure,in laboratory, the influence of the mixture composition on perma-nent deformation resistance. As the rate at which permanent defor-mation accumulates increases rapidly with higher temperatures,laboratory testing must be conducted at temperatures that simulatethe highest levels expected in the paving mixture in service [7].

The RSST-CH test applies a repeated haversine shear stress of1218 N to test cylindrical specimens. The applied load has a dura-tion of 0.1 s, with an unload time of 0.6 s. The test follows theAASHTO TP7-01 test procedure C. The results of the RSCH-CH testare expressed in terms of the number of passes of the equivalentstandard axle load of 80 kN (ESAL 80 kN) as a function of the num-ber of applied load cycles in the RSST-CH. The RSST-CH equipmentis presented in Fig. 5 (left side).

Wheel tracking is used to assess the permanent deformationresistance of asphalt mixtures under conditions which simulatethe effect of traffic. A loaded wheel tracks a specimen under spec-ified conditions of load, speed and temperature, while the develop-ment of the rut profile is monitored and continuously measuredduring the test.

The equipment used in this work is presented in Fig. 5 (rightside) and consists of a wheel that moves forward and backward(frequency of 1 Hz) and of a device that monitors the rate at whicha rut develops on the surface of the test specimen. The deformationis measured in established intervals of time, up to 120 min.

A steel mould has been used to provide confinement to the30 � 25 � 8 cm specimens. The specimens were subjected to a500 kPa pressure. The rut depth (permanent deformation) is re-corded as a function of the number of wheel passes in the follow-ings intervals of time: 1, 3, 5, 10, 15, 20, 25, 30, 35, 40, 45, 60, 75,90, 105 and 120 min. Testing was finished after 120 min.

The tests results, given by the following expression, are ex-pressed in terms of the velocity of deformation (v):

v t2=t1 ¼dt2 � dt1

t2 � t1ð1Þ

where vt2/t1 is the velocity of deformation between time 1 and 2, dt1

and dt2 is deformation or rut depth in time 1 and 2; t1 and t2 is thetime 1 and 2, respectively.

The results presented in this work correspond to velocities ofdeformation verified between the 120 and 105 (v105/120) min.The limit allowable depends on the intensity of the traffic and the cli-matic area where the pavement is located. For the most unfavour-able conditions, the limit for v105/120 is 1.5 � 10�2 mm/min.

Eight specimens were tested for each mixture using the RSST-CH, resulting in 40 samples, and three slabs for each mixture weretested in the wheel tracking device, resulting in 15 samples.

1E+03

1E+04

MDCO MGTB1 MDCB2 MDTB2 MGCB1

Fig. 8. RSST-CH test results.

5. Testing results

The development of the plastic shear strain deformation in theRSST-CH tests is presented in Fig. 6. Fig. 7 represents the evolution

of the vertical deformation in the wheel tracking tests. Most of thepermanent deformation verified in these tests occurs at the firstloading cycles following a slightly increase of the permanent defor-mation during the tests.

In both tests, the results showed that the conventional mixturepresents the highest deformation and, therefore, the lowest resis-tance to permanent deformation, whereas the asphalt rubber mix-tures show more permanent deformation resistance.

The RSST-CH results, expressed in terms of number of cycles ofthe equivalent standard axle (ESAL 80 kN), are presented in Fig. 8.The wheel tracking test results, expressed in terms of velocity ofdeformation (v105/120) of the slabs, are presented in Fig. 9. Theranking obtained from both tests is not identical. However, the bestmixture in both tests is an asphalt rubber mixture produced withasphalt rubber binder from continuous blend. Once the wheeltracking test results of the continuous blend asphalt rubber mix-tures is identical, the asphalt rubber mixture with gap-gradedaggregate gradation and continuous blend asphalt rubber binderpresents the best permanent deformation behaviour.

MDCO

MGTB1

MDCB2MDTB2

MGCB1

y = 6.1435x-0.419

R² = 0.6552

1E-02

1E-01

1E+071E+061E+05

V10

5/12

0 (m

m/m

in)

ESAL (80 kN)

Fig. 10. Comparison between RSST-CH and wheel tracking tests.

0,00E+00

5,00E-03

1,00E-02

1,50E-02

2,00E-02

2,50E-02

3,00E-02

3,50E-02

4,00E-02

4,50E-02

MDCB2 MGCB1 MDTB2 MGTB1 MDCO

Ave

rage

v10

5/12

0(m

m/m

in)

Fig. 9. Wheel tracking test results.

1198 L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200

The comparison between both tests is illustrated in Fig. 10, fromwhich it can be concluded that a linear trend, in log–log scale,exists between both tests. Mixtures with low permanent deforma-

MGCB1MDTB2

MGTB1 MDCB2

MDCO

y = 5E+06e-0.045x

R² = 0.4096

1E+05

1E+06

1E+07

10 20 30 40 50 60

ESA

L (

80 k

N)

Penetration (0,01 mm)

Fig. 11. Relationship between penetration and

MDCO MGTB1 MDCB2

MDTB2

MGCB1

y = 4909.7e0.0823x

R² = 0.4324

1E+05

1E+06

1E+07

50 55 60 65 70 75

ES

AL

(80

kN

)

Softening point (ºC)

Fig. 12. Relationship between softening point an

tion resistance in the RSST-CH test exhibit high deformation in thewheel tracking test. Mixtures with low deformation in the wheeltracking test exhibit high permanent deformation resistance inthe RSST-CH test. Four of the five mixtures tested follow the trend-line presented in this case.

In this study the influence of the properties of the asphalt bin-der on the permanent deformation resistance of asphalt mixtureswas evaluated by a series of graphics in which the permanentdeformation, expressed in terms of ESALs and velocity of deforma-tion (v105/120), is related to the asphalt binder penetration(Fig. 11), softening point (Fig. 12), resilience (Fig. 13) and apparentviscosity (Fig. 14).

The relationship between permanent deformation resistance,in both tests, and penetration (Fig. 11) shows that the increaseof the penetration (softer binders) will decrease the permanentdeformation resistance (decrease the ESALs and increase thevelocity of deformation). Despite a reduced correlation, mainlybecause the permanent deformation resistance is influenced bythe properties of mixtures (the most important are the aggre-gate gradation, binder content and void content) it is evidentthe influence of binder penetration on permanent deformationresistance.

The analysis of Fig. 12 allows concluding that the softeningpoint can be an indicator of the permanent deformation behav-iour. It is noticeable that asphalts with high softening point makemixtures with better permanent deformation resistance, ex-pressed in terms of low velocity of deformation and high resis-tance to plastic deformation in the RSST-CH test. In general, ahigh softening point conducts to a better permanent deformationresistance.

The relationship between the resilience and the permanentdeformation expressed in terms of ESALs and the velocity of defor-mation (Fig. 13) allows concluding that the increase of the resil-ience enhances the resistance to permanent deformation.

10 20 30 40 50 60Penetration (0,01 mm)

MGCB1

MDTB2

MGTB1

MDCB2

MDCO

y = 0.0065e 0.0301x

R² = 0.6855

1.0E-02

1.0E-01

v 10

5/12

0(m

m/m

in)

ESAL (80 kN) and velocity of deformation.

50 55 60 65 70 75

Softening point (ºC)

MDCO

MGTB1

MDCB2 MDTB2

MGCB1

y = 0.6265e -0.054x

R² = 0.6941

1.0E-02

1.0E-01

v 10

5/12

0(m

m/m

in)

d ESAL (80 kN) and velocity of deformation.

MDCO 1

MGTB1

MGCB1

MDCB2

MDTB2

y = 312575e0.0383x

R² = 0.5676

1E+05

1E+06

1E+07

0 10 20 30 40 50 60

ES

AL

(80

kN

)

Resilience (%)

0 10 20 30 40 50 60

Resilience (%)

MDCO

MGTB1

MGCB1

MDCB2MDTB2

y = 0.0407e-0.025x

R² = 0.8905

1.0E-02

1.0E-01

v 10

5/12

0(m

m/m

in)

Fig. 13. Relationship between resilience and ESAL (80 kN) and velocity of deformation.

MDCO

MDCB2

MGCB1

MGTB1

MDTB2

1E+05

1E+06

1E+07

ESA

L (8

0 kN

) MDCO

MGTB1

MGCB1MDTB2 MDCB2

y = 0.0333e-3E-04x

R² = 0.5748

1.0E-02

1.0E-01

0 1000 2000 3000 4000 5000

v 10

5/12

0(m

m/m

in)

Apparent viscosity (cP)

0 1000 2000 3000 4000 5000

Apparent viscosity (cP)

Fig. 14. Relationship between apparent viscosity and ESAL (80 kN) and velocity of deformation.

L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200 1199

The analysis of the results presented in Fig. 14 indicates that theapparent viscosity of the binder cannot be used as an indicator ofthe permanent deformation resistance.

The best correlation between the asphalt characteristics and thepermanent deformation resistance was obtained by the penetra-tion and the resilience in the wheel tracking test. An adequate cor-relation was also obtained between the permanent deformationresistance in the RSST-CH test and resilience, allowing definingthe resilience as the best indicator for the permanent deformationresistance of the asphalt mixtures.

However, these comparisons are not valid if only asphalt rubberbinders are considered, concluding that the asphalt rubber charac-teristics cannot be used to predict the permanent deformationresistance of asphalt rubber mixtures.

6. Conclusions

The test results showed that asphalt rubber mixtures improvedtheir resistance to permanent deformation in relation to the mostused conventional Brazilian asphalt mixture, independently fromthe type of asphalt rubber or gradation adopted. It was observedthat the asphalt rubber binder with the highest softening pointproduced asphalt rubber mixtures with better resistance to perma-nent deformation.

The results of the permanent deformation evaluation seem toconclude that the asphalt rubber mixtures with gap-graded aggre-gate gradation and continuous blend asphalt rubber binder presentthe best performance.

Based on the testing results, a linear trend exists between thepermanent deformation evaluated from the RSST-CH and thevelocity of deformation from the wheel tracking test, which allowsthe use of both tests in the permanent deformation prediction.

The correlation between the asphalt binder characteristics andthe asphalt mixtures properties exists mainly between the resil-ience and the permanent deformation. However, this correlationis not valid for asphalt rubber binders.

Acknowledgements

The first author was supported by the Programme Alban, theEuropean Union Programme of High Level Scholarships for LatinAmerica, Scholarship No. E04D040507BR from 2004 to 2006. Cur-rently, the first author is supported by the Brazilian National Coun-cil of Scientific and Technological Development (in Portuguese)(CNPQ).

References

[1] Lee S-J, Akisetty CK, Amirkhanian SN. Recycling of laboratory-prepared long-term aged binders containing crumb rubber modifier. Constr Build Mater2008;22(9):1906–13.

[2] Baker TE, Allen TM, Jenkins DV, Mooney T, Pierce LM, Christie RA, et al.Evaluation of the use of scrap tires in transportation related applications in theState of Washington. Washington (DC): Washington State Department ofTransportation; 2003.

[3] Shen J, Amirkhanian S, Xiao F, Tang B. Influence of surface area and size ofcrumb rubber on high temperature properties of crumb rubber modifiedbinders. Constr Build Mater 2009;23(1):304–10.

[4] Sousa JB, Craus J, Monismith CL. Summary report on permanent deformation inasphalt concrete. Strategic Highway Research Program, Report no. SHRP-A/IR-91-104, Washington, DC; 1991.

[5] Zaniewski JP, Sri Harsha N. Asphalt technology program evaluation of bindergrades on rutting performance. Morgantown, West Virginia: Department ofCivil and Environmental Engineering; 2003.

[6] Roberts FL, Kandhal PS, Brown ER, Lee D, Kennedy TW. Hot mix asphaltmaterials, mixture design and construction. Lanham (MD): NAPA EducationFoundation; 1996.

[7] Sousa JB, Solaimanian M, Weissman SL. Development and use of the repeatedshear test (constant heigh): an optional Superpave mix design tool. Strategic

1200 L.P.T.L. Fontes et al. / Construction and Building Materials 24 (2010) 1193–1200

Highway Research Program, Report no. SHRP-A-698, Washington,DC; 1994.

[8] FHWA. Background of superpave asphalt mixture design and analysis. FederalHighway Administration, FHWA-SA-95-003. Lexington, KY; 1994.

[9] Bennert T, Maher A, Smith J. Evaluation of crumb rubber in hot mix asphalt.Final Report, USDOT, Research and Special Programs Administration,Washington, DC, USA; 2004.

[10] Takallou HB, Bahia HU, Perdomo D, Schwartz R. Use of superpave technologyand construction of rubberized asphalt mixtures. Transportation ResearchRecord 1583, TRB, National Research Council, Washington, DC; 1997. p. 71–81.

[11] Hicks RG. Asphalt rubber design and construction guidelines, volume I –design guidelines. Sacramento, California, USA: Northern California

Rubberized Asphalt Concrete Technology Center (NCRACTC) and CaliforniaIntegrated Waste Management Board (CIWMB); 2002.

[12] Lee S-J, Amirkhanian SN, Shatanawi K, Kim KW. Short-term agingcharacterization of asphalt binders using gel permeation chromatographyand selected Superpave binder tests. Constr Build Mater 2008;22(11):2220–7.

[13] Doh YS, Yun KK, Amirkhanian SN, Kim KW. Framework for developing a staticstrength test for measuring deformation resistance of asphalt concretemixtures. Constr Build Mater 2007;21(12):2047–58.

[14] Greenbook. Standard specifications for public works construction. 2000Edition, Anaheim; 2000.