superpave mixture design · pdf fileaccount for this in the mix design, ... extremes for which...

NoticeThis document is disseminated under the sponsorship of the Department of Transportation inthe interest of information exchange. The United States Government assumes no liability for itscontents or use thereof. This report does not constitute a standard, specification, or regulation.

The United States Government does not endorse products or manufacturers. Trade and manu-facturer's names appear in this report only because they are considered essential to the object ofthe document.

WesTrack was the Federal Highway Administration's (FHWA) test facility in Nevada fordeveloping performance-related specifications for hot-mix asphalt pavement construc-tion. It also provided some of the earliest data on the performance of Superpave asphaltmixture designs under high rates of heavy truck loading. When Superpave-designed testsections placed at the track in June 1997 had very rapid rutting failures, the highway com-munity was concerned that the mixture design and construction procedures might bemissing important, but unknown, constraints. A forensic team composed of academicians,asphalt industry representatives, and State highway agency engineers was assembled tostudy the early failures and, if appropriate, to make recommendations for revising theSuperpave procedures. Their examination of the failures resulted in Report No. FHWA-RD-99-134, Performance of Coarse-Graded Mixes at WesTrack—Premature Rutting, which is avail-able from FHWA or on the Internet at the Turner-Fairbank Highway Research Centerhomepage at www.tfhrc.gov.

During the team's investigation, its members concluded that the asphalt paving commu-nity needed a good guide on the design of Superpave mixtures. Such a guide would sup-plement existing specifications and supporting literature and would incorporate the expe-rience of engineers across the country, including the WesTrack designers, in the initialyears of Superpave mixture design and placement. It would be a useful companion to theNational Asphalt Pavement Association's Superpave Construction Guidelines. This publica-tion, Superpave Mixture Design Guide, was prepared by the forensic team. Its contents arethe views of the team and do not necessarily reflect the views of the U.S. Department ofTransportation.

Note that this version of the guide is not expected to be the final word on Superpave mix-ture design. Both current research studies and additional field experience are likely toyield refinements in the future.

T. Paul Teng, P.E.Director, Office of Infrastructure

Research and Development

WesTrack Forensic Team Consensus Report

February 2001Washington, DC

SUPERPAVE MIXTURE

S U P E R PAV E M I X T U R E D E S I G N G U I D E 1

Introduction

Superpave design methods and tools arebeing implemented by many State agenciesto replace the Marshall and Hveem designmethods. In 1999, 2,515 projects, specifyingsome 73 million metric tons of Superpave,were let.[1] The majority of the projects in1999 and in previous years were constructedwith little or no difficulty. On several proj-ects, there were some problems during thisinitial implementation. For the most part, thecauses of the problems have been identifiedand have been solved. In 2000, estimateswere that more than 3,900 projects, specify-ing some 134 million metric tons ofSuperpave, would be let; this would repre-sent 62 percent of the total hot-mix asphalt(HMA) tonnage expected to be contracted forby State agencies during 2000 in the UnitedStates.[1] Superpave has become the mixturedesign method of choice by most State trans-portation departments across the country.

This document, intended as a companion tothe National Asphalt Pavement Association's(NAPA) Superpave Construction Guidelines,[2]

is a guide for the HMA designer to maximizethe benefits of Superpave while avoidingpotential problems. The Superpave designprocess is part of a total pavement designsystem. Superpave is a system of compo-nents that work together to provide aSUperior PERforming asphalt PAVEment. Astraffic levels and loading conditions increaseabove 1 million 80-kN (18,000-lb) equivalentsingle-axle loads (ESALs) during a pave-ment's design life, some design areas are notadequately addressed by the currentSuperpave specifications.

This guide discusses several issues thatshould be considered during the mixturedesign process to maximize the benefits ofthis method. The Superpave design processis documented in publications from theFederal Highway Administration (FHWA),the Strategic Highway Research Program(SHRP), the Asphalt Institute (AI), and theAmerican Association of State Highway and

Transportation Officials (AASHTO).[3-8] Thosepublications should be used for detaileddesign information. This guide is intended toserve as a bridge between existing knowl-edge and additional tools being developed tomeasure and predict Superpave mixture per-formance under traffic.

A Superpave mix design includes severalprocesses and decision points. First, designcompaction levels are established and mate-rials are selected and characterized. Then,mixture samples are prepared and laborato-ry test results are compared to design crite-ria. However, the existing Superpave designsystem does not properly address perform-ance prediction testing on mixture samplesor decision-making during the designprocess. This guide will address both ofthese areas.

Superpave Mixture DesignCompaction Level Determination

Prior to 1999, the design ESALs shown in theGyratory Compaction Criteria table of PP-28(in AASHTO Provisional Standards[8]) did notclearly indicate that they represent the pave-ment's cumulative ESALs for a 20-yeardesign life, rather than the cumulativeESALs for a shorter or longer design life. TheWesTrack Forensics Team and the LeadStates Team both recently reminded usersthat, regardless of the actual design life ofthe pavement, the user should determinethe expected ESALs for 20 years and selectthe design level for that traffic and loading.For example, a project with a 5-year intend-ed life may have a 5-year cumulative ESALcount of 2.9 million. This corresponds to a

This document is a guide for the HMA designer

to maximize the benefits of Superpave while

avoiding potential problems.

2 S U P E R PAV E M I X T U R E D E S I G N G U I D E

20-year cumulative ESAL count of 11.6 mil-lion (neglecting traffic growth compound-ing), and this latter ESAL count should beused in the design. The 1999 version of PP-28 includes a footnote to the GyratoryCompaction Criteria table with the appropri-ate guidance.[8]

Experience has shown that rutting damageoften occurs in the first few years of a pave-ment's life; therefore, the design should bebased on the rate of loading. To properlyaccount for this in the mix design, the mixdesigner should always use 20-year designESALs, essentially converting total loads to arate of loading. Estimating ESALs over a 20-year life, instead of the actual design life,may affect the mixture design compactionlevel, the performance-graded (PG) binderselection, and the aggregate consensus prop-erties specified for the project. Compactioncriteria, aggregate properties, and volumet-ric properties are all more stringent at high-er ESAL levels.

Superpave Performance-GradedBinder Selection

The Superpave Performance-Graded BinderSpecification (AASHTO MP-1) is based onproviding a binder that is resistant to rutting,fatigue cracking, and low-temperature crack-ing at specific pavement temperatures. Thebinder temperature ranges in the specifica-tion are based on the high and low tempera-tures at which a binder reaches critical val-ues of distress-predicting properties.Reliability factors included in the designmethod account for normal pavement tem-perature variations and allow the designer tomake a rational decision regarding the rangeof temperature extremes for which to design.Binder grade is selected based on designhigh and low pavement temperaturesexpected at the construction site and ondesired reliability.

The most common method of selecting abinder grade is to determine the design air

temperature range for the specific project andthen to establish the corresponding designpavement temperatures. Before selecting thegrade to be used, the designer must also con-sider traffic volume and traffic speed.

The owner should consider factors such ascost and traffic levels in establishing reliabil-ity, and hence, the final binder grade selec-tion, for a specific project. For example, if aPG 64 binder provides 94 percent reliabilityfor high temperatures, it may not be cost-effective to specify a PG 70 binder to obtain98 percent reliability. However, if a PG 64binder only provides 52 percent reliability, itwould probably be reasonable to specify aPG 70 binder to obtain 98 percent reliability.

With respect to traffic volume, when thedesign traffic exceeds 10 million ESALs,Superpave suggests that an increase in thehigh-temperature binder grade be consid-ered. When design traffic is more than 30million ESALs, Superpave requires a one-grade increase in the high-temperaturebinder grade. With respect to traffic speed,Superpave recommends increasing the high-temperature binder grade by one grade forslow transient traffic (20 to 70 km/h) and bytwo grades if standing traffic conditions(<20 km/h) exist. The binder specifiershould increase the high-temperature gradefor traffic volume or traffic speed, but not forboth. If the system is used correctly, a pave-ment with high design ESALs with stoppedtraffic conditions will require an asphaltbinder that is two high-temperature gradeshigher than that required by the pavementtemperature alone.

It should be realized that when the high-temperature grade is increased by one grade,the stiffness of the binder will approximate-ly double. In other words, a PG 70 binder willbe twice as stiff as a PG 64 binder at a tem-perature of 64°C. Furthermore, a PG 76binder will be four times stiffer than a PG 64binder at a temperature of 64°C. Trafficspeed will also have an effect on binder stiff-ness in the pavement. At 50 km/h, a binder


will have a lower apparent stiffness than itdoes when carrying traffic at 100 km/h. Inother words, a mixture containing PG 70binder in a pavement with traffic moving at50 km/h will have roughly the same stiff-ness as a mixture containing PG 64 binder ina pavement with traffic moving at 100 km/h;thus, the increased high-temperature gradeof the binder effectively offsets the effect ofslower traffic speeds.

Consideration should be given to the impactthat increasing the binder grade will have onthe construction process. Depending on thegrade, such an increase could require mixingand compaction temperatures beyond rea-sonable construction temperatures.

Only strong aggregate skeletons can experi-ence significant performance increases withincreased asphalt binder stiffness. Thestiffer binder locks the aggregate particles inplace to prevent rutting. The binder cannotcarry the load alone and cannot overcome apoor aggregate skeleton by itself.

The final step before selecting the bindergrade to be specified is to compare the gradebeing considered with grades historicallyused in the area. If the binder seems unrea-sonably soft for preventing rutting based onpast history, or unreasonably stiff for con-struction purposes, the selected gradeshould be reconsidered.

Superpave Aggregate Selection

Aggregates are the largest component ofHMA, making up 80 to 85 percent of themixture by volume and roughly 95 percentof the mixture by weight. Aggregate charac-teristics and quality are major factors in theperformance of HMA. As part of its focus onbinder and mixture properties, in the early1990's SHRP convened an expert panel todetermine which aggregate properties weremost important for pavement performance.The properties selected included coarseaggregate angularity, fine aggregate angular-

ity, flat and elongated particles, clay con-tent, and gradation. Aggregate source prop-erties, such as soundness, toughness, anddeleterious materials, were also found to beimportant. However, the criteria applied tothe source properties were found to reflectregional differences in aggregate quality,and were usually based on aggregate avail-ability. The panel determined that thesource properties were best left for eachState or local agency to establish. The fol-lowing discussion addresses various aggre-gate properties (consensus and source) andtheir effect on the Superpave design process.

Coarse Aggregate Angularity Mixtureswith crushed coarse aggregate with sharp,angular shapes will usually have the greatestshear resistance and, hence, the highestresistance to rutting. These materials createHMA mixtures with the highest voids in themineral aggregate (VMA). Coarse aggregateangularity is defined as the percentage byweight of the aggregate with one or morefractured faces according to AmericanSociety for Testing and Materials (ASTM)D5821. Superpave requires increased per-centages of crushed faces as the design ESAL

AGGREGATE PROPERTIES

CONSENSUS PROPERTIES(required)

• coarse aggregate angularity (CAA)

• fine aggregate angularity (FAA)

• flat and elongated particles

• clay content

SOURCE PROPERTIES(agency option)

• toughness

• soundness

• deleterious materials


level increases. VMA increases somewhat ascoarse aggregate angularity increases.

Uncompacted Void Content of FineAggregate (Fine Aggregate Angularity)Similar to coarse aggregate, crushed angularfine aggregate will usually have the greatestshear resistance. The use of crushed angularfine aggregate typically increases the mix-ture VMA. Fine aggregate angularity is estab-lished by AASHTO T304, Method A, whichmeasures the percentage of air voids presentin loosely compacted aggregate that passesthe 2.36-mm sieve. More fractured facesgenerally result in higher uncompacted voidcontents in this test. Superpave specifiesuncompacted void contents of at least 45percent on high-volume roads (>3 x 106

ESALs). Crushed manufactured fine aggre-gates generally have uncompacted void con-tents of at least 44.5 percent, while roundednatural sands typically are less than thatvalue. When a fine aggregate known to beangular has test results lower than expected,the aggregate's bulk specific gravity shouldbe verified since the test result is sensitive tothis property; a significant change in thebulk specific gravity should trigger aredesign of the mixture.

Particle shape can also influence the uncom-pacted void content. Some very cubical man-ufactured fine aggregates, especially some

limestones, have had less than 45 percent(but more than 40 percent) uncompactedvoid contents, but still have provided goodperformance in pavements. If the perform-ance has been satisfactory, the cubical man-ufactured fine aggregate may be used (withcaution).

The fine aggregate's uncompacted void con-tent significantly influences the VMA. Theuse of cubical angular fine aggregate is rec-ommended to increase the VMA. Careshould be taken when using aggregate withuncompacted void contents higher than 47percent; use of these aggregates may resultin mixtures with excess VMA, which leads,in turn, to a very high binder content.

Flat and Elongated Particles The per-centage of flat and elongated particles (notflat or elongated) in coarse aggregate isanother important aggregate parameter. Flatand elongated particles can break during theconstruction process, changing the mixturegradation and the overall mixture properties.Soft aggregate has a greater tendency tobreak than hard aggregate. Flat, sliveredaggregate particles also have a tendency tolie flat in the pavement, creating slippageplanes and reducing aggregate interlock. Asmall percentage of flat and elongated parti-cles in the mixture may increase the VMA inthe laboratory-designed mix. A furtherincrease may, however, decrease the VMA inthe plant-produced mixture because ofaggregate breakage during mixing.

The critical measurement for a flat and elon-gated particle is the ratio of its maximumand minimum dimensions. CurrentSuperpave standards allow no more than 10percent of the coarse aggregate particles tobe flat and elongated (i.e., a ratio greaterthan 5:1). Testing is performed according toASTM D4791, "Flat Particles, ElongatedParticles, or Flat and Elongated Particles inCoarse Aggregate." Superpave establishesthat testing be done on material retained onthe 4.75-mm sieve, instead of on the 9.5-mmsieve as specified in the ASTM method.

CONTRASTING STONE SKELETONS

Angular Aggregate

Rounded Aggregate


Testing aggregate particles passing the 9.5-mm sieve and retained on the 4.75-mm sievewill be more difficult and results may bemore variable.

Sand Equivalent Sand equivalent, asmeasured by AASHTO T176, "Plastic Finesin Graded Aggregates and Soils by Use of theSand Equivalent Test," identifies the pres-ence of clay in the fine aggregate. Clay canmake the mixture moisture sensitive and/orcombine with moisture to cause the mixtureto act "tender" (i.e., to lose density with con-tinued compaction in the field). Clay con-tent must be controlled by satisfying theminimum sand equivalents specified in theSuperpave standards.

Aggregate Toughness Typically, mixturescontaining very hard aggregate (i.e., a Mohshardness of 7 or greater) do not have a prob-lem meeting VMA criteria. A very hardaggregate, such as basalt, does not easilycrush or degrade during laboratory com-paction or during mix production in an HMAplant. These aggregates can produce mix-tures that have an adequate VMA.

Soft aggregates, such as some types of lime-stone having a Mohs hardness of about 5, areoften abraded during the gyratory com-paction process; this can make it difficult tomeet VMA criteria during the design phase.During production, the aggregates are oftenabraded in the hot-mix plant to an evengreater degree than in the laboratory designusing a gyratory compactor. When plant-produced material is compacted in a gyrato-ry compactor, the aggregate is abraded fur-ther and even more fines are generated inthe mixture; this further reduces the VMA.Mixtures designed with soft aggregates oftenhave a problem meeting VMA criteria in thedesign stage and, particularly, during pro-duction. It is extremely important that theplant-produced mixture satisfy the mini-mum VMA requirement.

Mixtures designed with a blend of hard andsoft aggregate could have difficulty meeting

VMA specifications. The addition of hard,coarse, or fine aggregate to these types ofaggregate blends will usually increase theVMA.

Superpave Mixture DesignConsiderations

Superpave mixture design criteria includeair voids, VMA, and voids filled with asphalt(VFA). Meeting the VMA minimum criterionis usually difficult to achieve during mixdesign and typically even more difficult toachieve in the plant-produced material. Thisdocument will only discuss VMA.

Voids in the Mineral Aggregate In manycases, achieving minimum VMA require-ments during the design phase can be diffi-cult. Many factors affect VMA. The most crit-ical of these are aggregate characteristicssuch as gradation, surface texture, andshape. If the design VMA is close to (i.e., nomore than 0.6 percent above) the minimum,aggregate properties may change during pro-duction and cause the VMA to drop belowthe minimum during mixture production.Differences between as-designed and plant-produced properties and other field prob-lems are documented in the NAPA publica-tion, Field Management of Hot-Mix Asphalt.[9]

As noted above, VMA is affected by both theaggregate gradation (relationship to theaggregate maximum density line) and theaggregate's characteristics and properties.For all designs, VMA should be plotted as afunction of binder content and the resultinggraph should be evaluated to check the VMA.Typically, VMA will decrease with increasingbinder content to some minimum, thenincrease as binder content continues toincrease. The design binder content, selectedat 4 percent air voids, should be near theminimum of the plotted curve or preferablyon the lean binder content side of the curve.If the VMA at the design binder content is onthe rich side of the VMA curve, adjustmentsto the gradation should be considered; these


are discussed later in this section. InSuperpave mixtures, however, VMA is some-times insensitive to binder content andshows little change. If the VMA at the designbinder content is close to the minimumallowable VMA value and the curve is rela-tively flat, the mixture should be redesigned.

There are two competing demands duringthe mix design process: sufficient inter-par-ticle space must be available for a minimumamount of binder, but, at the same time, theaggregate must have a sufficiently strongskeleton to carry the traffic loads. Superpavemixture design specifications require thatadequate VMA be obtained without weaken-ing the aggregate skeleton.

Having representative aggregate bulk specif-ic gravity values is necessary in order toaccurately calculate a mixture's VMA duringdesign and production. For this reason,aggregate bulk specific gravity should bedetermined at a frequency appropriate forthe variability of the source.

Mixtures with high VMA need to bereviewed for possible performance prob-lems. The WesTrack Forensic Team recom-mended that the VMA of coarse-gradedSuperpave mixtures be no more than 2.0percent above the minimum requiredvalue.[10] Furthermore, the Team recommend-ed running a draindown test (AASHTOT305-97) on these mixtures if the VMA is 1.5percent or more above the minimum value.If a gradation yields a mixture with too highof a VMA and, consequently, too high of abinder content, the mixture design shouldbe repeated with a new gradation with lowerVMA.

• Gradation Effect Problem mixes typical-ly will have a low VMA and may not beresponsive to changes in gradation (whenaggregate sources are not changed). Usually,however, changing the gradation of a mixturewill influence the amount of void space in theaggregate skeleton. The effect of gradation isseparate from the shape and surface texture

effects if all particles have the same shapeand texture. If the stockpiles in the blend areof dissimilar materials, changing the stockpilepercentages will change the gradation, but itwill also influence the shape and texture ofthe aggregate blend. Thus, VMA will changenot only because of gradation changes, butalso because of shape and texture changes.

Research papers published by Nijboer in the1940's, Goode and Lufsey in the 1960's, andthe Asphalt Institute in the 1980's provide abasis for the 0.45 power chart. Nijboer inves-tigated aggregate gradations plotted as thelog percent passing versus log particle size.He showed a maximum packing density forboth gravel and crushed aggregates whenthe slope was 0.45. Goode and Lufsey con-firmed Nijboer's results on gravel aggregates.Work by the Asphalt Institute evaluated themaximum density line on a 0.45 power chartfor both gravel and crushed limestone mix-tures and reconfirmed the previous results.

Moving the gradation away from the 0.45power maximum density line generallyincreases the VMA for a fine gradation, i.e.,when the gradation is above the maximumdensity line. For a coarse gradation, VMAmay decrease slightly and then increase asthe gradation moves away from the maxi-mum density line. For hard aggregates, theJob Mix Formula (JMF) should be parallel tothe maximum density line until after pass-ing the restricted zone, i.e., for aggregateretained on the 4.75-mm sieve (for 25-mm orlarger mixes) or on the 2.36-mm sieve (for19-mm or smaller mixes). Then the grada-tion line should be taken to the desired -0.075-mm content. Therefore, with fine gra-dations, the JMF should be above and paral-lel to the maximum density line. For acoarse gradation, it should be below and par-allel to the maximum density line.

Many coarse Superpave mixes have an "S"shape, starting on the fine side of the maxi-mum density line and finishing the S on thecoarse side. If the same particle shape andtexture are used (same aggregate source, dif-

ferent sizes), the highest VMA that can beachieved is that for the gradation that is thefarthest from the maximum density line. Tominimize the chance of mix tenderness dur-ing construction, the JMF should be 3 to 4percent above the lower control point for the2.36-mm sieve.

If the aggregate is obtained from a gravelsource, normally the fine aggregate must beremoved before the coarse aggregate entersthe crusher. The crushed material should bedivided into three or more stockpiles thatcan then be blended into a combination thatmeets the minimum VMA requirements. Allof the aggregate processed may not beusable; it may be necessary to waste some ofthe material in order to meet the require-ments of the mixture design.

If the aggregate is obtained from a quarriedsource, the crushed material should not beplaced into a single stockpile, but should bedivided into at least three separate sizeranges, depending on the nominal maxi-mum size of the aggregate required in themix. The use of multiple stockpiles allowsmore flexibility to change gradation and,thus, VMA. In addition, it may still be neces-sary to incorporate another size of aggregatefrom the quarry or from a different source.

The VMA of coarse-graded mixes can gener-ally be increased by reducing the amount ofmaterial passing the 4.75- and 2.36-mmsieves. The reason has to do with packing—smaller particles fill the spaces between larg-er ones. By reducing the amount of materialpassing the 4.75- and 2.36-mm sieves, inter-mediate material is removed and more spaceis created between the coarse aggregate par-ticles. Hence, the mixture cannot compactas tightly, i.e., VMA is increased.

In fine-graded mixes, VMA is created by fineaggregate—the material passing the 2.36-mmsieve. To increase the VMA in fine-gradedmixes, the percentage of material passingthe 2.36-mm sieve should be increased. Careshould be taken not to create a hump in the

gradation on the 0.6- to 0.3-mm sieves usingan aggregate that has a low fine-aggregateangularity value.

The dust content (i.e., the amount of materi-al finer than 0.075 mm) in a mixture has asignificant effect on the VMA. Lowering thedust content will increase the VMA. Thiseffect may not be entirely due to the grada-tion, but may also be due to characteristicsof the dust, such as shape and size. In gen-eral, reducing dust content to the extent thatthe dust-to-binder ratio will allow will maxi-mize the amount of VMA that can beobtained for the specific gradation.

If the dust content is coming from the addi-tion of mineral filler, adjusting the dust con-


SUPERPAVE AGGREGATE GRADATION

100

0.075 .3 2.36 12.5 19.0

Sieve Size (mm) Raised to 0.45 Power

Perc

ent P

assi

ng

Design Aggregate Structure

FHWA 0.45 POWER GRADING CHART

100

80

60

40

20

0.075 .3 .6 1.18 2.36 4.75 9.5 12.5 19.0

Sieve Size (mm) Raised to 0.45 Power

Perc

ent P

assi

ng

MaximumSize

Maximum Density Line


tent can be simply a matter of reducing theamount of filler being used. If the dust is pre-dominantly from one of the aggregate stock-piles, e.g., screenings, reducing the percent-age of that stockpile used in the blend shouldbe tried. If the screenings are the only man-ufactured fines coming into the mix, usingwashed screenings or blending with awashed screening may be necessary.

If baghouse fines will be introduced back intothe mix during production, some of the samefines should be added during the mix design.During the design, adding half of the quanti-ty of baghouse fines expected to be addedduring production is an appropriate proce-dure. These fines should be obtained fromthe actual plant that will be used for produc-tion; otherwise, mineral filler or an alternatesource of baghouse fines could be used.These fines will reduce the VMA of the mix-ture. If the aggregate in the mix contains fri-able particles, a greater quantity of dustshould be used in the laboratory mix designsince the friable particles tend to create moredust during mix production. A mix designthat includes baghouse fines will be morerepresentative of the mix as produced. Theaddition of baghouse fines during the mixdesign will better simulate the reduction inVMA that typically occurs during production.

• Surface Texture Effect The way in whichaggregate particles pack together for anygiven gradation is influenced by the surfacetexture of the particles. Rougher texture gen-erates more friction between aggregate par-ticles and the mixture therefore resists com-paction. Hence, for a given number ofdesign gyrations, the mixture will not com-pact as much and the VMA will be higher.Smooth texture, by contrast, does not gener-ate as much friction between aggregate par-ticles. For a given number of design gyra-tions, the mixture containing smoother par-ticles will compact more easily and the VMAwill be lower.

Typically, crushed faces have more texturethan uncrushed faces. In the case of gravel

aggregate, the uncrushed portion of theparticles tends to have a smooth texture.The greater the percentage of each individ-ual particle surface area that is fractured,the more surface texture that will be pres-ent. Usually, the more a gravel is crushed,the more surface texture it will have.Particles with two crushed faces tend tohave a greater percentage of surface areawith rough texture than will particles withonly one crushed face. However, crushingwill not always increase texture, becausesome aggregates fracture with very smoothfaces.

If manufactured sand and natural sand arebeing used together in a mix design, the per-centage of manufactured sand can beincreased to increase surface texture.Substituting 20 percent washed manufac-tured sand (with good "bite") for an equiva-lent amount of natural sand can increase theVMA substantially. (What is good bite?Squeeze a handful of angular manufacturedsand, then a handful of rounded naturalsand, and feel the difference in the way theparticles bite into one another.) If the manu-factured sand contains more dust than thenatural sand, gains in VMA from the surfacetexture may be decreased by the increase indust content. For example, if the naturalsand is relatively clean and the manufac-tured sand has a high minus 0.075-mm dustcontent, the benefit of increased surface tex-ture may be partially or completely offset bythe increased dust content.

• Shape Effect For any given gradation,the density to which aggregate particles willpack is influenced by the shape of the parti-cles. Angular particles (i.e., those withsharp, defined edges) tend to produce mix-tures with a higher VMA than mixtures con-taining rounded particles. Cubical particlesthat retain a sharp, angular edge tend to cre-ate a higher VMA than particles with round-ed edges.

The effect of flat and elongated particlesdepends on the laboratory compaction


method. Under Marshall compaction, theparticles were not as free to rotate as theyare in a Superpave gyratory compactor. Infact, flat particles tend to bridge in aMarshall mold and give a high VMA.Therefore, flat and elongated particles tendto increase asphalt content in a Marshall mixdesign. In the Superpave gyratory com-pactor, where the particles are kneaded intoa more stable condition, flat and elongatedparticles tend to lie horizontally; thisreduces the VMA and the optimum bindercontent.

During construction, rollers tend to orientflat and elongated particles horizontally. AMarshall mix design containing excess flatand elongated particles could compact veryeasily or be compacted to a lower air voidcontent than desired during the roadwaycompaction process. A Superpave mixdesign will have a more appropriate bindercontent since gyratory compaction bettersimulates compaction during constructionthan does Marshall compaction. Therefore,the influence of particle shape must be con-sidered when comparing the VMA ofMarshall specimens to that of Superpavespecimens.

If a mix design has a low VMA, the amount offlat and elongated particles must be deter-mined. Superpave specifications limit thepercentage of particles with a maximum-to-minimum dimension ratio of greater than 5.If flat and elongated particles are contributingto a low VMA in a mixture, the percentage ofparticles that exceed a 3:1 ratio should bedetermined. If the percentage of particlesexceeding the 3:1 ratio is high (i.e., greaterthan 40 percent), material from a coarseaggregate stockpile that has a lower percent-age should be added. It may be possible tochange one of the coarse aggregate stockpilesfor another that contains more cubical andangular aggregate particles. Adding an inter-mediate-size coarse aggregate with cubicaland angular shapes will prevent the largerparticles from lying flat. Thus, VMA willincrease.

Crushing operations influence the amountof flat and elongated particles produced. Ifexcess flat and elongated particles are beingproduced, the crushing operation should beevaluated. In some instances, the amount offlat and elongated particles produced can bereduced by changing the aggregate feed rate,or by changing the opening of the cone orjaw crushers. In some cases, it might be nec-essary to modify the crushing operation byadding to or changing the equipment used.Vertical-shaft impact crushers, for example,tend to produce more cubical particles thando some cone crushers (especially oldermodels).

In summary, VMA depends on the grada-tion, surface texture, and particle shape ofthe aggregate. In designing a mix, all of thesecharacteristics must be considered. Whenthere is difficulty in meeting the minimumVMA requirements, some or all of the abovecharacteristics must be adjusted. It should beremembered that the VMA of a plant-pro-duced mixture is typically lower than theVMA of the laboratory trial mix formula.Allowances should be made for the reduc-tion in VMA that will occur between the lab-oratory-designed and the plant-producedmixtures.[9]

Mixing and Compaction Temperatures For unmodified binders, the mixing andcompaction temperatures used during thedesign process should be established with arotational viscometer. If the binder is modi-fied, the binder supplier must provide rec-ommended mixing and compaction temper-atures. If the binder content determined inthe mix design process seems unrealistic,the supplier should be consulted to deter-mine whether the mixing and compactiontemperatures being used are still appropriatefor the material being delivered. The com-paction temperature used during designshould also be used in plant production qual-ity control and quality assurance testing.

The laboratory mixing and compaction tem-peratures may not be appropriate for use in

1 0 S U P E R PAV E M I X T U R E D E S I G N G U I D E

actual plant production and laydown.Environmental conditions at the time of con-struction and other factors, such as haullength and lift thickness, need to be consid-ered in establishing the actual mixing tem-perature used at the plant. These tempera-tures should not be any greater than thosenecessary to ensure complete mixing of theHMA while minimizing premature aging ofthe binder and providing for adequate com-paction in the field.

If the quality control and/or quality assur-ance sample is taken behind the paver andthe sample requires reheating before com-paction, a comparison should be madebetween the properties of the reheated sam-ples and samples that are compacted beforecooling. Detailed procedures for qualityassurance sampling and testing should beestablished before construction begins.

Dust-to-Binder Ratio Superpave calcu-lates the dust-to-binder ratio using the effec-tive binder content. Using the effectivebinder content rather than the total bindercontent will normally result in a higher dust-to-binder ratio because of binder absorptioninto the aggregate. To account for absorp-tion, the limit for the dust-to-binder ratioshould be increased. In the originalSuperpave specification, the dust-to-binderratio was 0.6 to 1.2 by weight. FHWA'sAsphalt Mixture Expert Task Group and theAASHTO Lead States recommended chang-ing the limit to 0.6 to 1.6; AASHTO subse-quently added a note to the mix design spec-ification suggesting that agencies considerchanging the limit for coarse-graded mixes to0.8 to 1.6.

During design, mixtures that are above themaximum density line at the 2.36-mm (for19-mm or smaller mixtures) or 4.75-mm (for25-mm or larger mixtures) critical sieveshould have a dust-to-binder ratio of no morethan 1.4. For mixtures that pass below themaximum density line at the critical sievesize, the ratio should not exceed 1.6.Characteristics of the fines will control the

amount that can be added to a mixture.Changing the fines source or productionprocess will change how the fines affect themixture characteristics. For fine-graded mix-tures (above the maximum density line at thecritical sieve), a ratio of about 1.0 has provid-ed satisfactory performance. For coarse-grad-ed mixtures (below the maximum densityline at the 4.75-mm sieve), as the VMAincreases the dust-to-binder ratio shouldincrease toward 1.6. If the mixture VMA ishigh (more than 2.0 percent above the mini-mum), the ratio should approach 1.6.

High dust-to-binder ratios will typically stiff-en the mixture and improve permanentdeformation resistance. However, if theVMA is more than 1.5 percent above theminimum, it is preferable to adjust theaggregate properties to reduce the VMAinstead of increasing the dust content.

Performance Indicator Tests No test iscurrently available that is satisfactory, byitself, as a performance predictor for mix-tures generated by Superpave volumetricprocedures. Appendix A contains a discus-sion of various tests that may be used to indi-cate the relative performance of differentmixtures. The designer should have experi-ence with one of the tests before assumingthat the test results will actually predict fieldperformance. Criteria developed elsewheremay not apply to a particular combination ofmaterials, environmental conditions, pave-ment structure, and traffic.

Completing the Design After the Super-pave design is completed, the designer needsto ask two final questions:

• Is this HMA design reasonableand logical?

• Is the binder content reasonablefor the type of aggregate, the nom-inal maximum aggregate size, theVMA, and the gradation used inthe mixture?

S U P E R PAV E M I X T U R E D E S I G N G U I D E 1 1

If the answer to either question is "No," thedesign should be re-evaluated and/orredone. Once the answers to both questionsare "Yes" and mixing plant operation hasbegun, the mixture volumetrics of the plant-produced mixture must be checked andthose volumetrics must meet the minimumsrequired in the design. In addition, perform-ance tests should be repeated on the plant-produced mixture if these tests were per-

formed during the original design phase.During plant verification, enough mixtureshould be produced to ensure that the plantis operating uniformly. The designer shouldbe prepared to make mixture adjustments toaccount for changes caused by plant produc-tion. If changes are made, the mixtureshould be re-verified. A key to good mixtureperformance is to verify the HMA plant-pro-duced mixture properties.

DESIGN CHECK LIST

Use a performance-graded (PG) binder and an N-design valueappropriate for the weather, traffic level, and traffic speed for theproject under consideration. Heavy, slow traffic will require astiffer PG binder than may have been used in the past.

Check that a complete mix design has been done in accordancewith specifications and that it meets all of the aggregate con-sensus property requirements and specified volumetric criteria.

Check that the submitted design contains a reasonable bindercontent for the materials used and the design level specified.

Generally, more dust (material passing the 0.075-mm sieve) isneeded for coarse-graded mixtures. The character of the dustwill control how much can be added to the mixture. Laboratorysamples should contain the expected plant-produced amount ofmaterial finer than 0.075 mm.

In coarse-graded mixtures, if the VMA is more than 1.5 percentabove the specified minimum, check for binder draindown.Excessive draindown is an indication that the binder content istoo high for the binder grade, aggregate type, and/or gradationbeing used.

Evaluate the mixture with a performance indicator test that hasworked satisfactorily based on local experience (until a univer-sally acceptable test is included in Superpave). Does the mixtureperform as expected?

Verify the properties of the plant-produced mixture to check vol-umetric properties. Repeat the performance test on the plant-produced mixture if the test was run during the mixture design.

➊

❷

❸

❹

➎

❻

❼


Introduction

Many different types of performance testsare currently available for assessing a mix-ture's ability to resist permanent deforma-tion (commonly referred to as "rutting").These tests, which include Marshallflow/stability, Hveem stability, the gyratorytesting machine, wheel-track testers, theSuperpave Shear Test device, and triaxialtesters, generally attempt to quantify mix-ture strength and/or stiffness. The individ-ual tests have shown varying levels of suc-cess in capturing a mixture's ability to resistrutting. Therefore, the designer must knowthe limitations of each test and how to incor-porate test results into mixture design selec-tion. This appendix describes each test andexamines how suitable each is for assistingengineers in designing rut-resistant mix-tures. At the same time, mixture designersare reminded that a mixture that is resistantto rutting will not necessarily resist thermalor fatigue cracking, moisture damage, ordurability problems such as raveling.

Marshall

The Marshall mixture design process seeksto optimize a mixture's performance withregard to fatigue cracking, rutting, and dura-bility by determining the optimum bindercontent for the gradation selected. Once theoptimum binder content is selected, themixture must meet minimum stability val-ues and maximum flow values. A number ofEuropean countries have modified the spec-ification criteria to use a stability quotient(stability/flow) criterion in lieu of the mini-mum stability and maximum flow values.

Many mixtures have stability values that aretwo or three times the minimum, but alsoexceed the maximum flow value. TheEuropean approach appears more logicalbecause it normalizes the stability/flow val-ues. Marshall flow does provide an indica-tion when a mixture is over-asphalted—highflow values indicate excess binder content.

The Marshall test conditions may signifi-cantly affect the test's value in predictingrutting performance. First among these isthe ratio of the test specimen's size to thenominal maximum aggregate size. A 100-mm- (4-in.-) diameter specimen thatincludes a large nominal maximum aggre-gate size (37.5 mm) or a more open-gradedmixture (one containing little intermediate-size material) does not provide good-qualitytest data. The effects of the specimen edgesare amplified and the assumption that theMarshall breaking head is applying a uni-form load across the specimen is no longervalid. The effective load on the specimen(load divided by the contact area) is higherfor larger nominal maximum aggregate sizemixtures. Another shortcoming of the proce-dure is the 60°C (140°F) temperature atwhich the Marshall test is conducted. Themixture may encounter temperatures 5 to10°C (9 to 18°F) higher in place in someparts of the country.

Hveem

The Hveem Stabilometer is a mixture designtool used primarily in the western UnitedStates. The concept behind the HveemStabilometer is an empirical measurementof the internal friction within a mixture,

P E R FO R M A N C E T E S TS


resulting from application of a vertical axialload. Like the Marshall Method, Hveem test-ing is conducted on 100-mm- (4-in.-) diame-ter specimens at 60°C (140°F). As notedabove, this temperature does not always rep-resent the highest temperature a mixturewill experience in the field. Furthermore,stabilometer values are measurements ofinternal friction, which is more a reflectionof the properties of the aggregate than of thebinder. As with Marshall flow values, Hveemstability does provide an indication when amixture is over-asphalted—low stability val-ues indicate excess binder content.

Gyratory Testing Machine

The gyratory testing machine (GTM), devel-oped by the U.S. Army Corps of Engineers,measures the increase in the angle of gyra-tion during compaction. The gyratory shearindex, a measure of a mixture's stability, isthe initial angle of gyration divided by themaximum angle. Shear indices above 1.1usually indicate poor mixture stability, whilevalues nearer to 1.0 are more stable.

Wheel-Track Testers

Currently, three wheel-track testers areavailable commercially—the French LCPC[Laboratoire Central des Ponts et Chaussees]Rutting Tester, the Georgia Loaded-WheelTester (marketed as the Asphalt PavementAnalyzer), and the Hamburg Wheel-TrackingDevice. Conceptually, the three devices arethe same (a rolling load is applied to labora-tory-scale specimens), but they differ signif-

icantly in design, load configuration, andtest conditions. To complicate the compari-son, each device has a different recom-mended pass/fail criterion for mixtures. Themachine design for each of the devices sig-nificantly affects how well its results can becorrelated with field performance.[11]

The French LCPC Rutting Tester uses a 90-mm-wide pneumatic tire to test specimensthat are 180-mm wide. This specimen widthand the closeness of the confining rigid spec-imen holder to the location of repeated load-ing distorts the development of the mixture'sshear plane, especially for mixtures contain-ing larger aggregate. As a result, poor mix-tures tend to perform better than expectedin the French device, and discriminatingbetween good- and poor-performing mix-tures becomes difficult. The device shouldnot be used to test mixtures that have aggre-gate larger than 20-mm.

The Georgia Loaded-Wheel Tester (GLWT)runs a concave steel wheel over a pressur-ized 29-mm-wide hose to apply loads onspecimens. Testing can be conducted on dryspecimens or underwater. For mixtures con-taining a larger size of aggregate, aggregatebridging becomes a problem. The appliedfootprint from the pressurized hose is muchnarrower than the footprint of a vehicle tirethat the mixture will be subjected to underfield conditions. As a result, the GLWT testcriteria may allow for some poor mixtures tobe placed.

The Hamburg Wheel-Tracking Device(HWTD) applies a sinusoidal load on speci-mens using a steel wheel underwater at anelevated temperature. The HWTD measuresa mixture's ability to resist rutting and strip-


ping. The probability that these same testconditions will coincide in the field is unlike-ly. The use of a steel wheel further increas-es the severity of the test. Because a steelwheel does not deform under the test condi-tions like a pneumatic tire, the effective loadper unit area is much higher than that occur-ring during actual field loading. A mixturethat survives the HWTD test should be rut-resistant in the field; however, mixtures thatdo not survive the test may also performwell in the field. Use of this device in mix-ture pass/fail situations can result in therejection of acceptable mixtures.

FHWA's Asphalt Mixture Expert Task Grouprecommends the following cautionary notefor wheel-track testers:

Rut testers, properly calibrated,have been utilized by some agenciesas effective proof testers. However,they should not be used to predictactual pavement performancebecause of differences in in-servicetemperature and loading conditions.The devices use empirical evalua-tion of some measured response to aloaded wheel as an indicator of per-formance. Local criteria from oneregion are not applicable in another.As such, each potential user needsto develop his/her own evaluationof wheel test results using local con-ditions.

Superpave Shear Tester

The Superpave Shear Tester (SST) can beoperated in any of six different modes: volu-metric, uniaxial strain, repeated shear atconstant stress ratio, repeated shear at con-stant height, simple shear at constant height,

and frequency sweep at constant height. Allbut the repeated shear at constant heighttest were included in the original Superpaveperformance testing program. The report,Background of SUPERPAVE Asphalt MixtureDesign and Analysis,[4] describes the testmodes in detail. Problems have beenencountered in interpreting data from therepeated shear at constant stress ratio test,the simple shear at constant height test, andthe frequency sweep at constant heighttests.[12-14] As a result of these problems, noattempt was made to link the predicted per-formance from the laboratory tests to thefield performance.

Romero and Mogawer presented additionalSST results and compared the results ofrepeated shear at constant height tests withthose from full-scale accelerated tests.[15]

They stated that the repeated shear at con-stant height test mode is able to rank mix-tures with different binders, but with highvariability in mixture stiffness. This variabil-ity often makes it impossible to place eachmixture into statistically different groups.SST results have shown significant variabili-ty between laboratories for the simple shearat constant height test mode. Until this vari-ability can be reduced, it will not be possibleto adopt universally acceptable criteria. Insummary, the SST is still being studied todetermine the usefulness of the results fromeach of its six test modes; work with thedevice has not reached a point where itsresults can be used in any standard mode topredict rutting performance.

Creep Tests

Triaxial testing equipment has been used formany years in soil mechanics and onasphalt materials. The creep test and, to a


lesser extent, the creep-creep recovery(CCR) test have been used for HMA undervarious triaxial stress states. The creep andCCR tests are used to estimate ruttingpotential. Most commonly, a uniaxial statictest is used in either a confined or anunconfined mode. The unconfined test doesnot simulate field conditions. The appliedpressure cannot exceed 207 kPa (30 psi)without specimens failing, and the test tem-perature is kept at 40°C (104°F) in theunconfined test, well below actual fieldloading conditions that often reach 830 kPa(120 psi) and 60°C (140°F). The confinedcreep test can be run at higher pressuresand temperatures, with a confining pressureof 138 kPa (20 psi). Research has shown thatconfined creep testing has a higher correla-tion to permanent deformation than uncon-fined testing.[16] A viscoelastic layered pave-ment performance system that uses creepand CCR testing to estimate the permanentdeformation in asphalt mixtures subjectedto repeated haversine loading of in-serviceloading frequencies already exists.[17]

However, this CCR testing does not exist asinput directly into constitutive models forasphalt pavements. Research is underway toexamine the ability of this equipment tomeasure "fundamental" material propertiesand to include these measurements in con-stitutive modeling. Currently, the equip-ment and procedures to help engineersmake rational mixture design decisions arenot available in the context of measuringengineering properties as input to constitu-tive models.

One test that is being recommended for per-formance evaluation of HMA by theresearchers on NCHRP Project 9-19 is theStatic Creep/Flow-Time test. In this test, acylindrical sample of bituminous pavingmixture is subjected to a static axial load.The test can be performed without confine-

ment or with a confining pressure applied tobetter simulate in situ stress conditions. Theflow time is defined as the time after initialload application when shear deformation,under constant volume, starts. The appliedstress and the resulting permanent and/oraxial strain response of the specimen aremeasured and used to calculate the flowtime. Using this test, the selection of thedesign binder content and aggregate struc-ture can be fundamentally enhanced by theevaluation of the mix's resistance to shearflow (flow time). This fundamental engi-neering property can be used as a perform-ance criteria indicator for permanent defor-mation resistance of the asphalt concretemixture, or can simply be used to comparethe shear resistance properties of variousbituminous paving mixtures.

Conclusions

Currently, no single test is suitable as anational standard for predicting rutting. Thedevelopment of such a procedure is urgentlyneeded, but a satisfactory procedure may beyears away. In the meantime, if an agencyhas extensive experience with a particulartest over a range of materials typical of its

geographic area, it should consider using thetest to predict rutting performance. Each ofthe devices outlined here has difficulty inpredicting the true performance of anasphalt mixture and should be used withgreat caution.

Currently, no single test is suitable as a

national standard for predicting rutting.


1. Superpave 1999-2000 NationalImplementation, Report to theAASHTO Task Force on SHRPImplementation, May 2000.

2. Superpave ConstructionGuidelines, Special Report 180,National Asphalt PavementAssociation, Lanham,Maryland, 1997.

3. McGennis, R.B.; Shuler, S.; andBahia, H.U. Background ofSUPERPAVE Asphalt Binder TestMethods, Report No. FHWA-SA-94-069, Federal HighwayAdministration, Washington,D.C., 1994.

4. McGennis, R.B.; Anderson,R.M.; Kennedy, T.W.; andSolaimanian, M. Background ofSUPERPAVE Asphalt MixtureDesign and Analysis, Report No.FHWA-SA-95-003, FederalHighway Administration,Washington, D.C., 1995.

5. The SUPERPAVE Mix DesignSystem Manual of Specifications,Test Methods, and Practices,Report No. SHRP-A-379,National Research Council(Strategic Highway ResearchProgram), Washington, D.C.,1994.

6. Superpave Performance-GradedAsphalt Binder Specification andTesting, Superpave Series No. 1(SP-1), The Asphalt Institute,Lexington, Kentucky, 1995.

7. Superpave Mix Design,Superpave Series No. 2 (SP-2),The Asphalt Institute,Lexington, Kentucky, 1996.

8. AASHTO Provisional Standards,Interim Edition, AmericanAssociation of State Highwayand Transportation Officials,Washington, D.C., May 1999.

9. Field Management of Hot MixAsphalt, Report No. IS-124,National Asphalt PavementAssociation, Lanham,Maryland, 1997.

10. WesTrack Forensic Team.Performance of Coarse-GradedMixes at WesTrack—PrematureRutting, Report No. FHWA-RD-99-134, Federal HighwayAdministration, Washington,D.C., June 1998.

11. Williams, R.C., and Stuart, K.D."Evaluation of LaboratoryAccelerated Wheel TestDevices," Proceedings, The 9thRoad Engineering Associationof Asia and AustralasiaConference, Wellington, NewZealand, May 1998.

12. Hicks, R.G., and Finn, F.N. Stage1 Validation of the RelationshipBetween Asphalt Properties andAsphalt-Aggregate MixPerformance, Report No. SHRP-A-398, National ResearchCouncil (Strategic HighwayResearch Program), Washington,D.C., 1994.

13. Zhang, X. "EvaluatingSuperpave PerformancePrediction Models Using aControlled LaboratoryExperiment," Journal of theAssociation of Asphalt PavingTechnologists, Volume 66, 1997.

14. Romero, P., and Mogawer, W.S."Evaluation of the SuperpaveShear Tester's Ability toDifferentiate Between MixturesWith Different Aggregate Size,"Transportation Research Record1630, National ResearchCouncil, Washington, D.C.,1998, pp. 69-76.

15. Romero, P., and Mogawer, W.S."Evaluation of the SuperpaveShear Tester Using 19-mmMixtures From the FederalHighway Administration'sAccelerated Load Facility,"Journal of the Association ofAsphalt Paving Technologists,Volume 67, 1998, pp. 573-601.

16. Roberts, F.L.; Kandhal, P.S.;Brown, E.R.; Lee, D.Y.; andKennedy, T.W. Hot Mix AsphaltMaterials, Mixture Design, andConstruction, National AsphaltPavement AssociationEducation Foundation, Lanham,Maryland, 1991.

17. VESYS 3am User's Manual,Office of Research andDevelopment, Federal HighwayAdministration, McLean,Virginia, 1996.



Team Members

Ray Brown, DirectorNational Center for Asphalt Technology211 Ramsay HallAuburn University, AL 36849Tel.: (334) 844-6228Fax: (334) 844-6248

Larry Michael, Regional EngineerMaryland Department of Transportation528 East Main StreetHancock, MD 21750Tel.: (301) 678-6134Fax: (301) 678-5190

Erv Dukatz, V.P., Materials & ResearchMathy Construction CompanyP.O. Box 563915 Commercial CourtOnalaska, WI 54650Tel.: (608) 779-6392Fax: (608) 781-4694

Gerald HuberHeritage Research Group7901 W. Morris StreetIndianapolis, IN 46231Tel.: (317) 243-0811Fax: (317) 486-5095

Ron Sines, Director, QC/QA OperationsP.J. Keating Co.P.O. Box 367Fitchburg, MA 01420Tel.: (978) 582-5200Fax: (978) 582-7130

Jim Scherocman, Consulting Engineer11205 Brookbridge DriveCincinnati, OH 45249Tel.: (513) 489-3338Fax: (513) 489-3349

Liaison Members

John D'Angelo, Sr. Pavement Mtls. Engineer Office of Pavement Technology (HIPT-10)Federal Highway AdministrationRoom 3118, Nassif BuildingWashington, D.C. 20590Tel.: (202) 366-0121Fax: (202) 366-7909E-mail: john.d'[email protected]

Chris Williams (formerly Research HighwayEngineer, Federal Highway Administration),Asst. Prof., Dept. of Civil & Env. Eng.Michigan Technological University870 Dow Env. Sci. & Eng. BuildingHoughton, MI 49931Tel.: (906) 487-1630Fax: (906) 487-2943

Editor

Terry Mitchell, Research Mtls. EngineerOffice of Infrastructure Res. & Dev. (HRDI-11)Federal Highway Administration6300 Georgetown PikeMcLean, VA 22101Tel.: (202) 493-3147Fax: (202) 493-3161E-mail: [email protected]

FHWA-RD-01-052

V I S I T U S O N T H E W E B A T :

www.tfhrc.gov

http://www.tfhrc.gov/

superpave mixture design · pdf fileaccount for this in the mix design, ... extremes for which...

Documents