Summary Report AGING OF...
Transcript of Summary Report AGING OF...
Summary Reporton
AGING OF ASPHALT-AGGREGATE SYSTEMS
SR-OSU-A-003A-89-2
by
Chris A. BellAssociate Professor of Civil Engineering
Oregon State UniversityCorvallis, OR 97331
November 1, 1989
ABSTRACT
Asphaltpaved roads have been used in the United States for about 100 years. They have
been used in Europe since the 1850's. No doubt,pioneeringpavement engineers soon realized that
in the shortterm asphalt hardenedafter heating, mainly due to volatilization,and, in the long term it
hardened, mainlydue to oxidation.
Hardening is primarilyassociated with lossof volatile componentsin asphalt during the
constructionphase (short-term aging), and progressiveoxidation of the in-placematerial in the field
(long-term aging). Bothfactors cause an increase in viscosityof the asphaltand a consequent
stiffeningof the mixture. This may cause the mixtureto become hard and brittleand susceptibleto
disintegrationand crackingfailures. Also, the productsof oxidationmay renderthe mixture less
durable than the odginal mixture, in terms of wear resistanceand moisturesusceptibility. However,
"aging" is not necessarilya negative phenomenon,since some aging may help a mixtureachieve
optimumproperties.
Compared to researchon asphaltcement, there has been little research on the agingof
asphalt mixtures,and, to date, there is no standard test. Pavement engineers understandthe need
to model the effects of short- and long-term aging of asphalt-aggregatemixtures in structuraldesign
procedures,and while some researchhas addressedthis need, as yet no standardprocedurehas
emerged to address it.
This report presents a state of the art on researchdirectedto understandingthe phenomenon
of aging of asphalt-aggregatemixtures. Recommendationsare made for aging procedureswhich
show promisefor laboratoryinvestigation. Test methodsto evaluate agingare also considered.
At this point, extended heating proceduresshow the most promisefor short-termaging and
pressure oxidationand/or extended heating the most promisefor long-term aging. Such procedures
will also be considered for long-termaging, togetherwith oxidationtechniques. Promisingtest
methods to evaluatethe effectsof aging of mixturesinclude: resir_entmodulus,indirect tensile test,
and dynamic modulustest. Tests on recoveredasphalt will also be necessary. Microviscosityand
ductilitytests will be consideredtogetherwith chemical fractionationtests.
ACKNOWLEDGEMENTS
The work reportedherein has been conductedas a part of projectA-O03Aof the
Strategic HighwayResearch Program (SHRP). SHRP is a unit of the National Research
Councilthat was authorized by section128 of th Surface Transportationand Uniform
RelocationAssistanceAct of 1987. This projectis entitled,"PerformanceRelated Testing
and Measuringof Asphalt-AggregateInteractionsand Mixtures,"and is being conductedby
the Instituteof TransportationStudies,Universityof California, Berkeley, with Carl L.
Monismithas PrincipalInvestigator. The support and encouragementof Dr. lan Jamieson,
SHRP ContractManager, is gratefullyacknowledged. Mai-Ly Christi providedvaluable
assistancewith literaturesearching.
The draft of this report was reviewedby an Expert Task Group (ETG) who provided
many valuable comments and will continueto provideguidance throughoutthe contract.
The members are:
EmestG. Bastian EricE. HarmFederalHighwayAdministration IllinoisDepartmentof Transportation
CampbellCrawford CharlesS. HughesNationalAsphaltPavingAssocation VirginiaHighwayandTransportationResearch
CouncilWilliamDearasaughTransportationResearchBoard DallasN. Little
TexasA&MUniversityFrancisFeeELFAsphalt KevinStuart
FederalHighwayAdministrationDouglasI. HansonNewMexicoState HighwayDepartment RogerL. Yarbrough
UniversityAsphaltCompany
Finally, ! am indebted to Nancy Brickmanand Laurie Dockendorfwhose considerable
word-processingskills,patience and rapidturnaroundtime have enabled this report to be
completedon time.
DISCLAIMER
The contents of this report reflect the views of the author,who is solely responsible
for the facts and accuracyof the data presented. The contentsdo not necessarilyreflect
the officialview or policiesof the Strategic Highway ResearchProgram (SHRP) or SHRP's
sponsors. The results reported here are not necessarilyin agreementwith the resultsof
otherSHRP researchactivities. They are reportedto stimulatereview and discussionwithin
the researchcommunity. This reportdoes not constitutea standard,specification,or
regulation.
iii
TABLE OF CONTENTS
1.0 INTRODUCTION ............................................ 1
1.1 Problem Definition ....................................... 1
1.2 Purpose ............................................. 2
1.3 Scope ............................................... 3
1.4 Organization .......................................... 3
2.0 LITERATURE REVIEW ....................................... 4
2.1 Introduction ........................................... 4
2.2 Factors Influencing Aging .................................. 4
2.3 Binder Studies ......................................... 8
2.3.1 Extended Heating Procedures .......................... 8
2.3.2 Oxidation Tests ................................... 20
2.3.3 Ultraviolet and Infrared Light Treatment ................... 27
2,3.4 Steric Hardening ................................... 30
2.4 Mixture Studies ......................................... 32
2.4.1 Extended Heating Procedures .......................... 32
2,4.2 Oxidation Tests ................................... 38
2,4.3 Ultraviolet/Infrared Treatment .......................... 42
2.4.4 Stedc Hardening ................................... 43
2.5 Relationship Between LaboratoryAgingTests and Field Performance ..... 46
2.5.1 CaliforniaStudies .................................. 46
2.5.2 MichiganTest Road ................................ 52
2.5.3 Texas Test Roads ................................. 54
2.5.4 PennsylvaniaTest Roads ............................. 59
iv
2.5.5 Iowa Study ..................... ................. 59
2.5.6 Oregon Study .................................... 59
2.6 RelationshipBetween Chemical Compositionand Field Performance ..... 61
2.6.1 Tests for AsphaltComposition.......................... 61
2.6.2 Value of CompositionalAnalyses ........................ 63
2.6.3 SignificantCompositionalParameters ..................... 64
2.6.4 Other PertinentStudies .............................. 70
3.0 PROMISING AGING AND TEST METHODS ......................... 72
3.1 PromisingAging Methods .................................. 72
3.1.1 Short-TermAging Methods ............................ 72
3.1.2 Long-Term Aging . ................................. 74
3.2 PromisingTest Methods ................................... 79
3.2.1 Tests on Aged MixtureSamples ........................ 79
3.2.2 Tests on RecoveredAsphalts .......................... 83
3.3 Evaluationof Aging Methods ................................ 84
3.4 Evaluationof Test Methods ................................. 86
4.0 GAPS IN THE KNOWLEDGE ......................... .......... 89
4.1 Contributionof Various Factorsto Aging ........................ 89
4.2 Aging of LaboratoryMixturesversus Field Aging ................... 90
5.0 CONCLUSIONS AND RECOMMENDATIONS ......................... 92
5.1 Conclusions ........................................... 92
5.2 Recommendations....................................... 93
6.0 REFERENCES .............................................. 95
APPENDIX A Aging Procedure Using Pressure Oxidation
APPENDIX B A Brief Summary of the Electromagnetic Spectrum Components:Ultraviolet and Infrared Radiation
APPENDIX C Outline of Test Plan: Aging of Asphalt-Aggregate Systems
vi
LIST OF FIGURES
Figure
1 Effect of Asphalt Film Thicknesson Level of Oxidationat 130°C (after Petersen,1989).
2 Effectof Asphalt Film Thicknesson Weight Loss in RTFO Bottleat 130°C (afterPetersen, 1989).
3 Effectof Bottle OpeningSize on AsphaltVolatile Loss, Level of Oxidation,andViscosity Increaseat 130°C (after Petersen, 1989).
4 Effect of Aging Time on Asphalt VolatileLoss, Level of Oxidation,and ViscosityIncrease at 113°C (after Petersen, 1989).
5 Viscosity Change During PavementService of three Zaca-Wigmore Asphalts(after Petersen, 1989).
6 Flowchartof Testing Procedures(after Lee, 1973).
7 Penetrationversus Time of Aging (after Lee, 1973).
8 Viscosityat 77°F versus Time of Aging (after Lee, 1973).
9 Viscosityat 140°F versusTime of Aging(after Lee, 1973).
10 Time-EquivalencyCorrelationCurvesby Voids Level (after Lee, 1973).
11 Effect of Lime Additionon the Viscosityof the 80/100 Pen Asphalt (after Edler,1985).
12 Effect of Pressure Oxidation (POV) Aging on Fraass Temperature of AsphaltCement (after Kim et al., 1986).
13 Change in Asphalt Properties During Construction and With Different LaboratoryHeating Time Intervals for the MI-0021 and TX-0021 Mixtures (after Von Quintaset al., 1988).
14 Aging Modulus Ratio (after Kim et al., 1986).
15 Fatigue Life of Aged Specimens (after Kim et al., 1986).
16 Pavement Aging Index Compared with MicrofilmAging Index (after Simpson etal., 1959).
17 Increase of Viscosity with Time (after Simpson et al., 1959).
18 Relationship Between Phenol Interaction Coefficient and Pavement SurfacePerformance Rating (after Petersen, 1984).
vii
19 Effect of Climate on Hardening (4 Climatic Sites) Combined Aggregates andVoids (after Kemp and Predoehl, 1981).
20 Age Hardening Involves Consistency Changes (after Corbett and Schweyer,1981).
21 Typical Hardening Curves (after Benson, 1976).
22 PenetrationHardeningCurves for Nine Test Sites (after Benson, 1976).
23 Viscosityversus Time Relationships,1964 Pavements (after Kandhal andKoehler, 1984).
24 FractionationProceduresfor AsphaltCement.
25 RelationshipBetween AbsoluteViscosityand Gaestel Index (after Tuffour et al.,1989).
26 Possible Interactive Effects of Temperature and Pressure Oxidation on Aging.
27 Dynamic Mechanical Analysis (after Coffman et al., 1964).
viii
LIST OF TABLES
Table
1 LaboratoryAcceleratedAging and EvaluationMethods.
2 EffectsWhich May Reducethe BindingPropertiesof Asphalt (after Traxler,1963).
3 Level of Aging of BOSCAN AC-10 Asphalt B-3051 Using Different Methods(after Petersen, 1989).
4 Aging Index of AC-10 Asphalts from DifferentSources UsingTFAAT (afterPetersen, 1989).
5 RelationshipsBetween Pavement Void Contentand Viscosityof AsphaltsAfter11-13 Years of PavementService (after Petersen,1989).
6 Effectof ERTFOT and PressureOxidation(POV) on Viscosity (after Edler et al.,1985).
7 Effect of the Weatherometer on Viscosity (after Edler et al., 1985).
8 Binder Viscosity--Beforeand After AcceleratedAging (after Hugo and Kennedy,1985).
9 Summary of StrengthData for SpecimensConditionedUsing DifferentAc-celerated Age HardeningTechniques(after Von Quintas et al., 1988).
10 Aging Results of Cores from Newly ConstructedRoad Surfaces and LaboratoryPrepared BriquettesUsing54 HoursUV-Radiationby AdaptedTTI Method*(after Hugo and Kennedy, 1985).
11 Aging Resultsof Cores from Newly ConstructedRoad Surfaces and LaboratoryPrepared Briquettes Using336 HoursAtlas Weatherometer UV-Radiation(afterHugo and Kennedy, 1985).
12 Evaluationof Aging Methods.
13 Evaluationof Test Methods.
ix
1.0 INTRODUCTION
1.1 Problem Definition
Asphalt paved roads have been used in the United States for about 100 years
(Krchma and Gagle, 1974). They have been used in Europe since the 1850's (Croney, "
1977). No doubt, pioneeringpavementengineerssoon realized that in the short term
asphalthardened after heating,mainlydue to volatilization,and, in the long term it
hardened,mainly due to oxidation. Becauseof the many ways asphalt can age, it seems
necessaryto clarify the terminologyused by pavement engineersregarding hardeningof
asphaltand asphalt mixtures. The terms "age hardening• and •aging"are regularlyused to
describethe phenomenonof •hardening•. The term •embrittlement• may also be used.
As implied above, hardeningis primarilyassociatedwith loss of volatilecomponentsin
asphaltduringthe constructionphase, and progressiveoxidationof the in-place material in
the field. Both factors Causean increasein viscosityof the asphalt and a consequent
stiffeningof the mixture. This may cause the mixtureto become excessivelyhard and
brittleand susceptibleto disintegrationand crackingfailures (Vallerga, i981). Also, the
productsof oxidationmay render the mixture less durable than the originalmixture, in terms
of wear resistanceand moisturesusceptibility(Barth, 1962). However, •aging• is not
necessarilya negativephenomenon,since some aging may help a mixtureachieve optimum
properties, it is appropriateto note the Webster'sdictionarydefinitionsfor aging which
support the above discussion,viz:
• becomingold
• showing the effects or characteristicsof increasingage
• acquiringa desirablequalityby standingundisturbedfor some time
• becoming mellowor mature
• bringingto a state fit for use or to maturity.
Since the 1930's research has continuedto develop an understanding of the factors
contributingto short-and long-termaging (Welborn, 1979). Much of thiswork has been
directedtoward the asphaltcement rather than the asphaltaggregate mixture,such that in
1989 we have standardizedtests to determine and control the short-termaging of the neat
asphalt. The present thin film oven test (TFOT) and rollingthin film oven tests (RTFOT)
are an outcomeof some of this research.
Compared to researchon asphalt cement, there has been little research on the aging
of asphalt mixtures,and, to date, there is no standard test. Pavement engineers under-
stand the need to model the effects of short-and long-termaging of asphalt-aggregate
mixturesin structuraldesign procedures,and while some research has addressed this need,
as yet no standard procedurehas emerged to address it.
1.2 Purpose
The purposeof this report is to presenta state of the art on research directedto
understandingthe phenomenonof aging of asphalt-aggregatemixtures. Short-term and
long-termeffectswill be considered. Also included is a review of test methods that have
been used to assess the effects of aging. Recommendationswill be made for aging
proceduresand test methods which show promisefor laboratoryinvestigation. Following
preliminarytests, more detailed testing will be done, which will ultimately lead to develop-
ment of standard proceduresfor aging mixtures. Such proceduresmay be incorporated in
a mix design or mix analysissystem so that the effects of aging on the fundamental
- propertiesof asphalt mixturescan be assessed.
It is not the purposeof this report to formulatea detailedwork plan for the laboratory
studiesthat will follow. The testing requiredfor developingstandard tests and identifying
limitationsof suchtests will be determined only after thoroughstudy of the promisingaging
and test methods. Similarly, the details of mixturesto be tested will not be given here. A
range of mixturesrepresentingcontemporarypractice willbe tested in the detailed test
program. This will includemixturesdesigned for heavy duty pavementssuch as "large
stone"mixtures.
1.3 Scope
As indicatedabove, this study will involveevaluationof methodsconsideringthe
short- and long-termaging of asphaltmixtures. An extensiveliterature search was
conductedsuchthat the author believesthat the majorityof relevantstudieshave been
considered. In doing so, it was clear that much of the literaturewas published20 to 50
years ago. However,contemporaryresearchhas yieldedthe promisingmethodsof aging
whichwill be recommendedfor furtherevaluation.
1.4 Organization
After this introduction,Chapter2 presentsa literaturereview which concentrateson
laboratorytest proceduresfor aging asphaltand asphalt-aggregatemixtures. However,
because of the need to relate laboratoryaging to field aging, studiesthat have attemptedto
establishsuch relationsare also reviewed. Similarly,since thoroughevaluationof the
effectsof aging could includeconsiderationof chemicalproperties,the final sectionof
Chapter2 reviewsstudiesthat have consideredthe relationbetween these propertiesand
field performance.
Chapter 3 identifiespromisingaging and test methodsand presentsa preliminary
evaluationof them. Chapter4 discussesgaps in the knowledgeregardingaging of asphalt-
aggregatemixtures. Finally,Chapter5 presentsConclusionsand Recommendations. A
substantiallist of referencesis also included,along with three Appendicespresenting
specificdetails for test methodsand for a proposedwork plan for laboratorytests.
3
2.0 LITERATURE REVIEW
2.1 Introduction
A literature searchwas conductedthrough the TransportationResearch Information
Services ('rRIS) unit of the TransportationResearch Board. In addition,DIALOG, and the
Universityof California(Berkeley) MELVYL systemwas used. Also, the author conducted
an informalsearch and received many helpful suggestionsfrom various colleagues.
The literatureis extensive and, by necessity,only selected referenceswill be cited.
Table 1 presents a selectionof referencesdealingspecificallywith laboratoryaging and
summarizes the methodsused, togetherwiththe test or tests used to evaluate the extent
and/or effects of aging. Discussionof the work representedin Table 1 is presented below
in two sections,one relatingto binder studiesand another to mixturestudies. Other work
addressingthe aging of mixtures in the fieldwill also be reviewed,and a section is included
that reviews researchthat has attempted to relate asphalt chemical compositionto the field
performance of asphaltaggregatemixtures. Before the discussionof aging and test
methods,the factors influencingaging are summarized.
2.2 Factors Influencing Aging
Traxler (1961) listedfive factors influencinghardeningof asphaltcements in ap-
proximateorder of importance,viz:
1) Oxidation
2) Volatilization
3) Time (developmentof internalstructureon aging)
4) Polymerizationinducedby actiniclight (free radical reactions)
5) Condensationpolymerization(by heat)
He indicatedthat "a completeunderstandingof hardeningwill require extensiveexperimen-
tation."
4
Table 1. LaboratoryAcceleratedAging and EvaluationMethods.
Date Investigator(s) Aging Method Evaluation Method
1903 Dow 18, 24 hrs, 325"F (163"C) Change in weight, penetration of residueMixture aged for 30 min,300"F (149"C) Recovered asphalt - change in penetration
1937 Nicholson Air blowing, 15 min, Penetration, ductility425"F (229"C)
Raschig & Duyle Air blowing, 15 min, Change in penetration400"F (204"C)
Hubbard& 5ollomb Ottawa Sand mixture,time Recoveredasphalt- change in penetrationand temperaturevaried
1939 Lang & Thomas Ottawa Sandmixture, aging Change in mix properties, abrasion,oven, outdoor exposure strength, etc.
1940 Shattuck Mixture oven aging 30 min Recovered asphalt - penetration,325"F (163"C) ductility, softening point
Lewis & Welborn I/B-in.film oven test Change in weight,penetration,ductility5 hr, 325"F (163"C)TFOT
1946 Lewis & Halstead I/B-in.film oven test Change in weight,penetration,ductility5 hr, 325"F (163"C)
1952 Pauls & Welborn Ottawa Sand mixture oven Con_)ressive strength, recovered asphalt.aged 325"F (163"C) TFOT TFOTresidue
1955 Griffin,Miles, & Shell microfilmtest - Viscositybeforeand after aging - agingPenther 5 micron (.0002-in.)film, index
2 hr, 225"F (I07"C)
1957 Vallerga,Monismith& Ultravioletand infrared PenetrationGranthem weathering Softeningpoint
Ductility
Brown, Sparks & Rapid chilling of asphalt Tensile test on asphalt sampleSmith sample
1958 Heithaus & Johnson Road tests - laboratory Recovered asphaltsaging - microfilm test Microfilm Agin9 Index
1961 Traxler TFOT and microfilm Microviscosityat 77"F (25"C)compared15 micron (.0006-in.)film, 2 hr, 225"F (107"C)
Halstead & Zenewitz TFOTand 15 micron film Microviscosity at 77"F (23"C) co_ared(.0006 in.) film, 2 hr, k shear rate of 0.05 sec-" was used225"F (107"C)
1963 Hveem. Zube & Skog Shell microfilm test ,mdi- Microviscosity at 77"F (25"C) before andfled - 20 micron (.0008- after agingin.), 24 hr, 210"F (99"C)Rolling TFOTand TFOT Viscosities of RTFOT,TFOT, and recovered325"F (163"C), $0 min asphalts comparedCohesiogreph test
1968 Lee TFOTat 325"F then PO6at Microvtscosity at 77"F150"F 24, 48, 96. 240 hr - limitingviscosityat 29 pstg and 132 psig - time to harden to 30 megapotsesAsphalt end sand-asphalt - shear indexmixes Asphaltene content
1969 Schmidt& $antucci Rollingmicrofilm test Mlcroviscosityof residue20 micron {.O008-in.bottle) 210"F (99"C)
5
Table I. LaboratoryAcceleratedAging and EvaluationMethods (Cont.)
Date Investigator(s) Aging Method EvaluationMethod
1973 Lee TFOT Microviscosity at 77"FPOB at 150"F and 20 atm Capillary.viscosityat 140"FRecovery of field aged Microductility, Fraass testmaterials Asphaltene content and oxygen percentage
Rostler analysis
1976 Senson TFOT Mtcroviscostty at 77"FActinic ltght Penetration at 77"FMixtures weathered in thefield
Plancher, 6reen RTFOT, RMFO, Mtcrovtscoatty @77"F (25"C)& Petersen Colummoxidation Asphaltene determination
Mixture oven aging for Chemical analysis5 hr g 302"F (150"C) Resilient modulusLima and nontreated asphaltand mixtures
1977 Kumar & 6oetz Permeation by air at 140"F Creep Test @7O'F ± 3" (Zl"C ± 2")(60"C) at a head of 0.02 - repeated load conditioningin. of water (0.5 mm) for - then 5 psi for 5 min1, 2, 4, 6, 10 days Slope and intercept of creep curve used
indicate progressive oxidationRatio of slope or intercept at X days toinitialslope : DurabilityIndex
1981 Kemp& Predoehl Actinic Light Weathering Penetration 77"F (ZS"C)test, Rolling microfilm Ductility 77"F (25"C)test, Ottawa sand mixture Resilient Modulus (MR)aging,ModifiedShell Microviscosity@ 77"FMicrofilmtest CapillaryViscosity @ 140"FMixtures weathered infield
Santucci, 6oodrich Tilt Oven Durability test Viscosity @6O'C and 135"C& Sundberg @235"F (113"C) for 168 hr Penetration @4"C and 25"C
and @239"F (I15"C) for Ductility @Z5"C100 hr
1983 McHattie Extended RTFOT Penetration 77"F (25"C)100 hr, Z39"F (115"C) Kinematic-viscosity Z75"F
Resilient ModulusEAL (Equivalent Axle Load) Life
1985 Edler, et al. Weatherometer - lOO micron Viscosity at 113"F (45"C) by slidingasphalt film, RTFOT plate mic¢oviscometer at a shear rate ofextended to 8 hr; Pressure 0.05 sac-"oxidation for 96 hr, 149"F Oxidation level - infrared spectra(65"C), 300 psi High molecular weight constituentsModified TFOT - 100 micronfilm, 24 hr
Hugo & Kennedy Oven age hardeningof mix- Mlcroviscosity@ 77"F (25"C)tures @ IOO'C Shrinkageof beam of mixtureUltravioletexposureofmixtures for 54 hr and 14
days
1986 Kim et al. Pressureoxidation@ 14O'F Capillaryviscosityat 140 and 275"F{6O'C)and I00 psi Fraassand penetrationtests0 to S days ResilientModulus and fatigueRecoveryof fieldaged Corhett-Swerbrickanalyses:materials
I
° i
Table 1. Laboratory Accelerated Aging and Evaluation Methods (Cont.)
Date Investigator(s) Aging Method Evaluation Method
1988 Von Quintas et al. Short-term oven aging @ Resilient modulus275"F (135"C) for 8, 16. Indirect tensile strain24, & 36 hr CreepLong-term pressure oxida-tion @135"F (60"C and 100psiOven aging for 2 days @135"F (6O'C) then 5 days@225"F (107"C)
Tia et al. Convection oven aging @ Resilient modulus140"F (60"C) for 1, 7. 28, Indirect tenstle strengthand 90 days Recovered asphalt properties includingForced draft oven aging @ viscosity @ 140"F (60"C) penetration @140"F (60"C) for 1, 7. 28, 77"F (25"C)and 90 days Schweyer rhecmeter at 77"F and 59"F {15"C)Ultravioletlightaging@ Infraredspectralanalyses140"F (60"C)for 1, 7, 28. Corbett-Swarbrickanalysesand 90 daysAging under natural condi-tions for 1, 2, and 3 years
1989 Petersen Thin film accelerated Weight loss due to volatilizationaging test Ketone content
Viscosity
In a subsequentpaper Traxler (1963), listed 15 effectswhich may reduce the binding
propertiesof asphalt. Table 2 is reproducedfrom this paper. It shouldbe noted that the
effects listedare not necessarilyin order of severity and may not apply when considering
asphaltmixtures. Traxler providessome experimentaldata to support the listof effects,
however,he statesclearly in the paper that "someeffects have not been given experimental
consideration.
More recently,Peterson (1984) has listedthree major factors causinghardeningof
asphalt in asphalt mixtures,viz:
1) lossof oily components by volatilityor absorption
2) changes in compositionby reactionwith atmosphericoxygen
3) molecularstructuringthat producesthixotropiceffects (sterichardening)
It shouldbe noted that these include four of Traxler's 15 effects.
The majorityof researchers consideringaging of asphalt and asphalt mixtureshave
limitedtheir investigationsto those factorsgiven by Petersen. Significantdevelopmentsare
presentedbelow.
2.3 Binder Studies
Table 1 shows that eady work on laboratoryaging proceduresemphasizedbinder
studiesand, in particular,extended heating. Much of thiswas done by or for the Bureauof
PublicRoads (BPR). Many of these studiesand others relatingto asphaltdurabilitywere
summarizedby Welbom (1979).
- 2.3.1 Extended Heating Procedures
a) Thin Rim Oven Test (TFOT)
Welbom (1984) includesa referenceto Dow (1903) indicatingthe use of an
• early extended heatingtest. Nearly 40 years later, Lewis and Welbem (1940)
introducedthe TFOT for differentiatingamong asphalts in terms of volatilityand
hardeningcharacteristics. Further workwas reported by Lewisand Halstead (1946).
8
®_- ,...- -; =® - _ o I ®
__ .._,,. ._ O..--rc _ <_o o
_'_ I I :_ I I :_( :_ :_ I X X
6
_ , , , , , , , , , , ,
.p
;_ _ _: , × , _ ,< x , , ,, , ,:_ _•_ =u_ _= _ x x , x , , , , x , x
=E
o o _ _ o o_
-o oo-0 4),_ .o "0 0
9
A 50 ml sample of asphalt was heated in a 1/8 in. (3 mm, i.e. 3000 microns)
film in a 5.5 in. (140 mm) diameter flat container for 5 hrs at 325°F (163°C). This
test was adoptedby AASHTO in 1959, and by ASTM in 1969. Residue from the
TFOT was tested for penetration,ductilityand softening point.
Welbom (1979) notes that the test is primarilyused to predict relativechanges
that occur in asphalts duringhot-plantmixing. Several researchershave made
significantmodificationsto the TFOT, as will be describedbelow. One minorchange
was done by Edler et al (1988), who reducedthe film thicknessto 0.004 in. (100
microns)and increasedthe exposureto 24 hrs.
b) Shell Microfilm Test
Griffin, Miles and Penther (1955) reportedthis test for 0.0002 in. (5 micron)film
of asphalt aged for 2 hrs on glass plates. The "aging index" parameterwas used to
evaluatedasphalts, i.e.:
Aging Index = viscosityafter aqinqviscositybefore aging (2.1)
Welborn (1979) notes that no data were reported correlatingfield and laboratory
aging, but that the study indicatedthe importanceof volatilizationfrom the standpoint
of hardening.
Hveem, Zube and Skog (1963) proposeda modificationto the Shell microfilm
test, by increasingthe film thickness to 0.0008 in. (20 microns)and the exposuretime
to 24 hrs. They demonstratedindirectlya relationshipbetween field and laboratory
hardening.
Traxler (1961), and Halsteadand Zenewitz (1961) also presentedslight
variationsof the microfilmtest as noted in Table 1.
10
c) Rolling Thin Film Oven Test (RTFOT)
The RTFOT was developed by the CalifomiaDivisionof Highwaysin an effort
to age asphalt in thinnerfilms than the 1/8 in. used in the TFOT. Hveem, Zube and
Skog (1963) reportedthe procedure. Bottlescontaining35 g samples are rotated in
an oven at 325°F (163°C) for 75 minutes. This causes a 0.05 in. (1.25 mm or 1250
microns)film to flow around the glass jar.
This test was adoptedby ASTM in 1970 and is used by several Pacific Coast
states for routinetesUng.
Several researchershave made slightmodificationsto the RTFOT, for example,
Eclleret al. (1985) used a time period of 8 hrs. More significantmodificationsare
presentedbelow.
d) Rolling Microfilm Oven Test (RMFO)
Schmidt and Santucci(1969) modifiedthe RTFOT procedureby dissolving
asphalt in benzene, coatingthe insideof the bottle, then allowingthe benzene to
evaporate and leave a 20 micron film of asphalt. This procedurewas necessary to
create a 0.008 in. (20 micron)film of asphalt. The asphalt is then rotated at 210°F
(99°C) for 24 hrs. A disadvantageof this test is the very small amount of asphalt
obtainedfrom each bottle(0.5 g).
e) Tilt-Oven Durability Test (TODT)
Kemp and Predoehl(1981) adapted the RTFOT test by tiltingthe oven by 1.06°
(higher at the front). The tiltingpreventedasphalt migrationfrom the rotatingbottles.
The test is continuedfor 168 hrs at 235°F (113°C). Penetration,viscosity,and
ductilitytests were run on the residue. The comparisonof data obtainedwith field
data will be discussedin a subsequentsection.
McHattie (1983) has reported a similarextensionof the RTFOT. He used a
temperatureof 239°F (115°C) for a period of 100 hours.
11
Santucci et al. (1981) evaluated the TODT at the conditionsnoted above (168
hrs at 235°F), and at the conditions used by McHattie. They note that these latter
conditions were formerly recommended by the Caltrans researchers. Santucci et al.
evaluated several asphalts with both test procedures, and concluded that 168 hrs at
235°F (113°C) was more severe. They measured penetration and viscosity of the
aged asphalt, and also concluded that viscosity was a more sensitive measure of
changes in the TODT. An alarming point raised by these authors is that two ovens
produced by different manufacturers were used in the study and generated
significantly different results.
f) Thin Rim Accelerated Aging Test (TFAAT)
Petersen (1989a) and his coworkershave developed a modificationof the
RMFO to providea 4 g sample of asphalt;a practical size for further testing. A
temperatureof 235°F (113°C) was used for 72 hrs. This test was developed to
complementa columnoxidationproceduredeveloped by Davis and Peterson (1966) in
which asphalt is coated on Teflon particlesto film thicknesses of about 15 microns
and oxidativelyaged in a gas chromatographiccolumn at 266°F (130°C) for 24 hrs by
passing air throughthe column.
As backgroundto the test development,Petersen (1989a) notes that many
asphalts exhibitvolatileloss in the TFOT and RTFOT in excess of what is typically
lostduring the lowertemperature long-termaging in the field. He notes that Corbett
and Merz (1975), in analysisof asphalts used in the MichiganTest Road observed
virtuallyno change in the asphalt saturatefraction (whichcontains the potentially
volatileasphaltcomponents) after 18 years of pavement service. He also notes that
with regard to the TFOT and RTFOT, "the level of oxidativeaging and hardening in
the tests is much less than what occursduringfield aging, reflectingonly the aging
12
- j
that occursduring hot-plantmixing." With this backgroundin mind, the TFAAT was
developedto producea representativelevel of volatilizationand oxidation.
Figures1 and 2 show the effect of asphaltfilm thicknesson ketone contentand
weight loss of the sample. Ketonesand anhydridesare formed on oxidation. Based
on data from studiesby Schmidt (1973) and Petersen et al. (1983), the 4 g sample
was found to have a representativelevel of ketones after 24 hrs aging at 266°F
(130°C). However,the volatile losswas excessive (Figure 2) and therefore studieson
the effect of bottle openingsize on volatile loss were conducted. Figure 3 showsthe
resultsand the openingsize of 3 mm selected for use.
Subsequentstudieswere done at 235°F (113°C) for comparisonwith the
California tilt-ovendurabilitytest (TODT). Figure4 showsthe data obtainedand the
period for testing of 3 days that shouldbe used at this temperature.
Petersen (1989a) presentsdata comparing the aging index of asphalts aged
usingthe TFAAT and TFOT (Table 3). He also presentsTFAAT data for several
asphalts (Table 4) and data for field aged asphalts (Table 5). Table 3 shows that, for
the asphaltevaluated,the TFAAT is much more severe than the TFOT. Aging is
describedby the aging indexas defined in Eq. 2.1. Comparisonof the aging indices
in Table 4, representingTFAAT data, and Table 5, representingfield data, shows that
the TFAAT causes a similar level of aging to that occurringin thefield. The field
data was obtained from an extensiveFHWA-fundedstudyconductedby Vallergaet al.
(1970). Petersen cautionsthat there is a differencebetween the "kinetics"of
oxidationin the field and in the TFAAT. He illustratesthis pointwith data given by
Zube and Skog (1969) for the Zaca-Wigmoretest road (Figure 5) which showsthat
field oxidationand viscosityincreasefall off with time in the field, whereas they do not
(Figure 4) in the TFAAT. This is associatedwith temperaturedifferencebetween the
two conditionsand its effect on molecularstructuringand steric hardening. Petersen
13
14
0
t_
G)
0C:O
:. °°..J
4..)
__ °o_..c
-,-o_LL ,,.._
G)
d
I I I ! 0vl 0 m 0 u_ =+ ' ,-;. ,.;. _..
N '_)SUDN_) |tlSIE) M
15
16
17
Table 3. Level of Aging of BOSCANAC-]O Asphalt 13-3051 UsingDifferent Methods (after Petersen, 1989)
Method of Aging Aging IndexI Log Aging IndexI
Unaged I. 0 0
TFOT 3.0 0.48
TFAAT 214.0 2.33
1Viscosity ratio at 140"F (60"C)
Table 4. Aging Index of AC-IO Asphalts from Different SourcesUsing TFAAT(after Petersen, 1989)
Asphalt Source Grade Aging Index1 Log Aging Index
California Valley AC-IO 10 1.00
Midcontinent AC-IO 32 1.51
Nichigan AC-IO 74 1.87
North Slope-Haya Blend AC-IO 90 1.95
California Coastal AC-IO 134 2.13
Boscan AC-IO 214 2.33
West Texas-Naya Blend AC-IO 338 2.53
1Viscosity ratio at 140"F (60"C)
Table 5. Relationships Between PavementVoid Content and Viscosity ofAsphalts After 11-13 Years of Pavement Service (afterPetersen, 1989)
Average Zl_itial Average Viscositye LogViscosity," poise Recovered Asphalt, _ Average _oid Aging Aging
140"F (60"C) poise, 140'F (60"C) Content,_ % Index Index
1.9x103 8.0x103 2 4.2 0.62
1.9x103 3.1xi04 4 16 1.20
1.9x103 8.0x104 6 42 1.62
1.9x103 1.6x105 8 84 1.92
1.9x103 3.1x105 10 163 2.21
Averoge for 18 initial asphaltsFrom best fit line of viscosity versus void content for 63 projects
18
19
suggeststhat these two phenomenasignificantlyreduce the rate of field hardening
after the first two or three years of service.
2.3.2 Oxidation Tests
a) Air Blowing
Nicholson(1937) utilizedair blowingat 425°F (229°C) at a rate of 1/3 _/min
(0.0091 m3/min)for 15 min. As noted in Table 1, penetrationand ductilitywere
measuredbefore and after aging, and asphalts retaininghighervalues were judged
superior. Another approachwas to age asphalts to a penetrationof 20 to 25 and
those retaininghighestductilitywere judged to be best.
Raschigand Doyle (1937) air-blew asphalts at 400°F (204°C) at the same rate
as Nicholson,and determinedthe change in penetration.
b) Pressure Oxidation
D.Y. Lee initiated a five-year study for the Iowa State Highway Commissionin
1966 entitled,"Developmentof a LaboratoryDurabilityTest for Asphalts." Lee (1968)
reportson the initialdevelopmentsof the study and recognizes the need to developa
two-stage test to simulate:
1) hardeningduring mixing
2) hardeningduringservice life
Lee adopted the TFOT withoutmodificationto simulatethe first stage and pressure
oxidationfor the secondstage. Lee notes that other researchers had used pressure
oxidationproceduresand that the BritishRoad Research Laboratory(RRL) developed
a device for use with road tars (HMSO, 1962). The RRL included the Fraass brittle
point as a means of evaluatingthe extent of aging.
Lee (1973) reported the major findingsof the Iowa study and presentedthe
followingprocedure for the "Iowa DurabilityTest" (IDT):
1) Use of BPR thin film oven test (TFOT) on the originalasphalt.
20
I
2) Applicationof pressure-oxidationtreatment to the residue from the
TFOT (film thickness1/8 in.) for up to 1000 hrs at 150°F (65°C) and
a pressure of 20 atm (300 psi) of oxygen.
3) Evaluationof the physicaland chemical changesin asphalt during
the artificial aging process in relationto the originalproperties.
Lee's pressure oxygenvesselswere 7.5 in. diameter by 7.5 in. high and rated to 450
psi.
The test proceduresused by Lee (1973) are summarized in Figure6. Figures
7, 8 and 9 show sample data from his study and the correlationbetween laboratory
data and field data. Lee found that the developmentof aging followeda hyperbolic
model as suggestedby Brownet al. (1957):
Ay_ Ta + bT (2.2)
where Ay = change of propertywith time T or the differencebetween the zero-
life value and the value at any subsequenttime
a = constant;the value of the propertyat zero time
b = rate of change of the property
1/'o = the ultimatechange (limitingvalue) of the property
Figures8 and 9 show highcorrelationcoefficientsbetween curves developedwith this
model and the observed data from the laboratoryand field tests. Only in the case of
ductilitydata, was this model inadequate.
Correlationof field and laboratorydata also resulted in hyperbolic relationships,
implyinga different rate of aging in the field, as suggestedby Petersen (1989a).
Figure 10 shows several master curves,one of which represents all nine projects
evaluated by Lee. From this curve, Lee concludesthat 46 hrs of aging with the IDT
is equivalentto 60 months field aging for Iowa conditions.
21
ORIGINAL t !
ASPHALT (o) PLANT (p) PAVEMENT (f)
TFOT RECOVERED ASPHALTS ]ASTM(d-O)D-1754 , (f-0,6,12,etc.)T
IOWA DURABILIIY TESTI DENSITY AND VOIDS
20 arm., 150°F I DETERMINATIONS
(d-24,_,_tc.) J
PENETRATION ! VISCOSITY AT 77°FASTM D-S _ SLIDING PLATE
! -
ASTM D-36. _ ASTM 0-2171
1 -
t FLASH ANO FIRE ] CHEMICAL ANALYSES
ASTM D-92
ASTM D- 113 SPECTROSCOPY
l -
SPECIFIC GRAVITY FRASS BRITTLEPOINT
ASTM D-70 TEST
SPOT TEST ] SPECIAL TESTS
AASHO T-102
Figure 6. Flowchartof Testing Procedures(afterLee, Ig73).
22
LABORAIC)RY AGEING (d), hr
0 100 _'00 300 400 500 600 700 800 900 10001 1 I I 1 I I I 1
ASPHALT CEMENT No. 1! :_o- ,KEY
o laboratoryaged100--Z o fieldaged (wheeltracks)L) • fieldaged (betweenwheel tracks)
8C PREDICTIVECURVES
LtJ
Z 60 0=. • T
=_401
I
Y- 1.0_+o I I I
1 1 I 1 1 I I I I I
120 -- ASPHALT CEMENT No. 2KEY
i O0 -- o laboratoryaged
Z I o field aged (wheeltracks)0 • field aged (between wheel tracks)
80 PREDICTIVECURVES
t--t=l=t
Z 60 TI=1=1
=" _ _._=_... _&Y=_ 0.353+0-.018T40
20emmm_
o +1°" I I 1a 0 10 20 30 40 50
FIELD AGEING (f), months
Figure7. PenetrationversusTimeof Aging(afterLee, 1973)
23
TIME, hourilm-,_.'_.'°_o_ ,.oo 6oo 8oo_ooo
_os i I I I I8
T
6 55.63= 106 3).=43.100+0.718 TR = 0.998
4 T-- t_y- 26.627 + 0.893
R-- 0.892
=-
_ 2
< />.
107 '_0u
8 T• = i"6.505 + I. 866 T
6 R = 0.972
4
ASPHALT CEMENT No. 1
KEY
a labo_ory aged
2 • _ aged(wheel_cks)• gad aged(betweenwheatracks).... PREDICTIVE CURVES
I 1 1 1• 0 !0 20 30 40 SO
TIME, moe_)_
Figure8. Viscosityat 77°F versusTime of Aging(afterLee, 1973)
24
Figure4. Viscosityat 140 F versustime of aging.TIME,
0 2OO 4O0 6O0 8OO IOO0
I l I I I
od 78 - o fA• fB /
6 - f/ T
_ -Ay = 143.815 + 0.523T'4 -- R = 0.999
0O.
- /o 2 -
° />.
104 -- /
/ /__i _ \ T. i ] /_f - ' Ay = 30.207- 1.0_f
"F? //2_- / f o _V,_d .
]I/ • eek_aged_ va,e__acks)
.... PP_VE _$
1o3 1 I I I I I ,, Ia 0 10 20 30 40 _0
TIMEo meml_e
Figure 9. Viscosity at 140°F versusTime of Aging(after Lee, 1973)
25
100 Tf
80 -- fog TI - 3.762 "_0.473 Tf
60 -- R = 0.954
46 .-
40--
f
log TI -1 2.599 + 0.541 TF.c:, 20 -- R = 0.865 A.C. No. 6-9
O IfZ
---, TI - 3.664 + 0.536 IfO< TO R = 0.849 A.C. No. 1-9
" EO 8 TF.< log TI•,, 8. 395"+ 0.494 TfO 6co R = 0.761 A.C. No 1,2< •
d--
2 - ASPHALT CEMENTSNO. 3 and No. 4
i _______J_ J 1 !0 10 20 30 40 50 60 70
FIELD AGEING, (TF), months
Figure 10. Time Equiivalency Correlation Curves by Voids Level (after Lee, 1973)
26
Edleret al. (1985) utilizeda similarapproachto Lee (1973) in a study to
evaluate proceduresto retard oxidativehardeningof asphalticsurfacingsin South
Africa. They utilizeda two-stage procedurewith extended RTFOT (ERTFOT) aging
done first for 8 hrs, followed by pressureoxidationat 300 psi and 149°F (65°C) for 96
hrs. Evaluationof the extent of asphaltaging was by slidingplate microviscometerat
a shear rate of 0.05 sec1 and at 113°F (45°C). Oxygen absorptionmeasurements
were also made, and the molecularweightcompositionof originaland aged asphalts
were determinedby gel permeationchromotography.Table 6 and Figure 11 show
viscositydata for lime and nontreatedasphalt. Note that the curves are of a
hyperbolicform, and that the levelsof aging index are similar to those reportedby
Petersen (1989a). ExtendedTFOT work was also attemptedon 100 micron (0.004
in.) films for 24 hrs followingthe ERTFOT. The viscositylevelsachieved with this
approachwere outsidethe range of the microviscometer.Also, the levels of oxygen
absorbed and high molecularweight fractionswere higherthan when the pressure
oxygen treatmentwas used.
Kim et al. (1986) utilizedpressureoxidationto age asphalton Fraass test
plaques. Asphaltsamples were subjectedto 100 psi oxygenpressureat 140°F
(60°C) for two and five days. Figure 12 shows the data for three different asphalts.
These authors note that an advantage of thisaging approachis that the asphalt is
aged on the "container"used for the test methodto evaluate the extent of aging. The
asphalt film on a Fraass plaque is 0.5 mm thick (500 microns). This is much thicker
than any of the thin film tests describedabove. However, Figure 12 shows that the
Fraass breakingpoint is significantlyalteredby pressureoxidation.
2.3.3 Ultraviolet and Infrared Light Treatment
Vallergaet al. (1957) reportedon studiesusingboth ultraviolet(UV) and infrared (IR)
light to age asphalt films in TFOT containers. The ultraviolettreatmentwas found to be
27
Table 6. Effect of ERTFOT and PressureOxidation (POV) on Viscosity(afterEdler et al., 1985).
Viscosity (Pa.s) Aging Index I
Asphalt Lime Added ERTFOT& POV AqedPenetration (%) Original Residue Original
4" ' '
0 4.37x103 7.50 x 105 171.6,,.,,,
60/70 6 6.40xi03 3.64 x 105 56.9
12 1.18xlO4 4.64 x 105 3g.3
0 1.91xIO3 3.44 x 105 179.8
80/100 6 3.40xi03 1.33 x 105 39.1,,.,
12 5.73xi03 1.58 x 105 27.5
28
• " I0'_ f f f / j/
OO $ *ill[ AoO¢,.Jii-m_l •
!
i0 _
S
l
iOI ! ,Z_lGl_t. I_TFOT I[II_rlL1)T'*'IDOl
AGI[iNG ¢ONl_t'lrtOmS
Figure 11. Effect of Lime Addition on the Viscosityof the 80/100 Pen Asphalt (afterEdler, 1985)
• /- / I
r_ _01 * I I I .
0 l 2 3 4 S
Days in the POV
Figure 12. Effect of POV Aging on Fraass Temperature of Three Asphalt Cements (afterKim et al., 1986)
29
most effective in terms of changesin penetrationsofteningpointand ductilityof the asphalts
treated.
Traxler (1963) presentsdata to show the effects of "actinic"light and refers to the
effect as a photochemicalreaction. His data clearly shows that there is a significanteffect
on thin films of asphalt (3 microns)but with thicker films the effect was small.
Edler et al. (1985) used a weatherometer(ASTM, 1979) to age 0.004 in. (100 micron)
films of asphalt. A temperatureof 149°F (65°C) was used, and a 102 min cycle of UV,
followed by 18 min of UV and water spray (300 psi). Samples were aged for a total of
32.5 hrs, 73.5 hrs, 7 days, and 14 days. The effects of the weatherometer on viscosityare
shown in Table 7. it was also shownthat the oxidationlevels and changes in high
molecularweight constituentswere comparablewith the otheraging proceduresused by
these investigators.
2.3.4 Steric Hardening
Traxler (1963) identified molecularstructuring(thixotropy)as one of his 15 effects (see
Table 2). The resultof this effect is steric hardening. He notes that this effect is mostly
reversedby applicationof heat or mechanicalworking, but that a portionis permanent,and
the extent to which it occursdependson asphaltcomposition.
Brownet al. (1957) presentsome data from a study to demonstrate that more steric
hardeningoccurs in asphaltswhen they are cooled slowly. The effect was demonstratedby
carryingout simple tensiletests on the asphaltsamples. Petersen (1984 and 1989a) has
also emphasizedthe role of steric hardening. The discussionof the TFAAT presented
above outlines Petersen's(1989a) "model"of the role of steric hardening,but as yet no test
has been developedto quantify its preciserole.
30
I
Table 7. Effect of the Weatherometeron Viscosity(afterEdler et al., 1985).
Viscosity(Pa.s) Aging Index
AgedWeatherometerResidue Original
Hours Days HoursAsphalt Lime Added
Penetration (%) Original 32.5 73.5 7 14 32.5 73.5
0 4.37x103 1.35xi056.43x105UTT* UTT 30.9 147.1
60/70 6 6.40xi03 1.25xi056.58xi05UTT UTT 19.5 102.8
12 1.18xi042.24x105g.lgx105UTT UTT 19.0 77.9
0 1.glx1035.54xi043.08xi05UTT UTT 29.0 161.3
80/100 6 3.40xi038.40x1042.75xi05UTT UTT 24.7 80.9
12 5.73xi03g.OSx10413.31x105UTT UTT 15.8 57.8
*UTT - unable to test; residuestoo hard
31
2.4 Mixture Studies
Less work has been done with laboratoryaging of mixtures than with binder studies.
However, as early as 1903, Dow reportedextended heating of mixtures,and the evaluation
of the effects by recoveryof asphalt and comparisonof penetrationbefore and after aging.
The developmentof an acceptedprocedurefor extractingand recoveringasphalt from
mixtures(Abson, 1933) no doubt influencedsubsequentresearch,since many of the early
studieson mixture aging (see Table 1) involvetests on recoveredasphalt, for example,
Hubbardand Gollomb(1937) and Shattuck(1952).
Lang and Thomas (1939) used a mixture of Ottawa sand and several asphalts in their
extensivestudy. This provided a standard"aggregate"which the authorsfelt would be
reproducible. Several other researchershave used Ottawa sand mixturesas will be
reportedbelow. Lang and Thomas made about 12,000 mix specimensin their study.
Some aging tests were done with ultravioletlight, and, some with extended exposure to
heat and air. Evaluation of the extent of aging was by abrasion tests on the mixture,
consistencytests and chemical compositiontests on Abson recoveredasphalts.
The above backgroundshows that significantcontributionsto understandingmixture
aging were made over 50 years ago. Subsequentcontributionswill be presentedbelow.
2.4.1 Extended Heating Procedures
Pauls and Welbom (1952) exposed2 in. by 2 in. cylindersof Ottawasand mixtureto
325°F (163°C) for varioustime periods. The compressivestrength of the cylinderswas
determined. Also, the consistencyof the recoveredasphalt was comparedwith that of the
originalasphaJL Asphalts representingmajorsources produced in the mid-1930's were
used in the study. The conclusionsof the study includedthe following:
1) The hardeningpropertiesof asphalt cements can be determinedeither by
measuringthe compressivestrengthof laboratoryoven-aged, molded
32
specimens,by tests on asphalt recoveredfrom the laboratory-aged
specimensor by the TFOT.
2) Because the TFOT procedureis relativelysimple, it is highlyvaluable for
predictinghigh temperaturehardeningof asphaltcements.
It shouldbe noted that there is no suggestionthat the TFOT was suitable for predicting
long-termhardeningdue to field weathering.
Plancheret al. (1976) also used an oven aging procedureon 1 in. thick by 1-1/2 in.
diametersamples (25 mm by 40 mm) as a part of a study to evaluatethe effect of lime on
oxidativehardeningof asphalt. It was found that the resilientmodulus(measuredwith a
diametraltest configuration)of lime-treatedmixtureswas changed less than nontreated
mixturesby the aging process. It shouldbe noted that Plancher et al. presentan
explanationof the chemistryof lime action. This study shouldalso be consideredwith that
by Edler et al. (1985), whichfoundthat lime had a considerableeffect in retardingaging in
asphalt samples.
Kemp and Predoehl (1981) utilized Ottawa sand mixturesin their work. A planetary
oven at 140°F (60°C) was used, with various times of exposureup to 1200 hours. The
asphalt was then recoveredand could be subjected to "micro-tests". These authors favored
the TODT since it generatesa much largerquantityof asphalt.
Hugo and Kennedy (1985) describea methodof oven aging of mixturebriquettes at
212°F (100°C). They note that this procedureis similar to an Aus_alian standard
(StandardsAssociationof Australia, 1980). This procedurewas carded out for 4 and 7
days in a dry atmosphereand in an atmosphereof 80% relativehumidity,due to the need
to assess a project located close to the ocean. Asphaltwas recovered for viscosity
determinationfrom 4 in. (100 mm) samples cored from laboratoryproducedslabs. Table 8
showsthe data obtained. Also, the samples were weighed before and after aging, and the
weight loss used to indicateloss of volatiles. Finally,beam sampleswere cut from the
slabs and the shrinkageduring the aging test determined.
33
Table 8. Binder ViscositymBefore and After AcceleratedAging(afterHugo and Kennedy, 1985).
Viscosityof RecoveredBinder at 25"C (Log Pa.S)AI
Aging Aging AfterBitumen Relative Lime 4 Days Aging 7 Days Aging 7 DaysContent Density Mix Content Before With 4 Days With 7 Days Dry
Mix No. % % Grading % Aging Humidity - Dry Humidity - Dry Aging
1 5.5 95.2 CONT - 5.16 5.51 5.32 5.62 5.67 3.23
2 5.5 93.0 CONT - 5.16 5.47 5.46 5.62 5.60 2.75
3 6.5 90.4 GAP - 5.16 5.80 5.89 5.95 6.10 8.70
4 7.0 90.4 GAP - 5.16 5.84 5.84 5.96 5.90 5,50
5 6.0 90.6 CONT 2 5.16 5.61 5.59 5.81 5.81 4.47
B 7.0 97.5 GAP 1 5.16 5.46 5.46 6.54 5.59 2.69
7 7.0 90.5 GAP* - 5.16 5.74 5.74 5.88 5.80 4.37+ chips
8 7.0 92.7 GAP 2 5.16 5.70 5.73 5.86 5.83 4.68+ chips
9 5.0 92.8 OPEN - 5.16 5.72 5.77 5.72 5.91 5.62
10 5.0 92.5 OPEN 1.5 5.16 5.75 5.76 5.92 5.92 5.75
*Appliedat 10 kg/m2.
34
Von Quintas et al. (1988) have published the findings from the second phase of the
study to develop an Asphalt Aggregate Mixture Analysis System (AAMAS). They
investigated the use of forced-draft oven aging to simulate "production hardening", viz short-
term hardening. They compared the recovered versus initial penetration and viscosity ratios
for asphalts used in each of five projects, for both field and laboratory aging. The
laboratory method involved oven heating loose mixture samples for periods of 8, 16, 24 and
36 hours at a temperature of 275°F (135°C). Figure 13 shows data for two of the projects
and shows that similar levels of aging are obtained in the laboratory and the field, but there
is also considerable scatter in the laboratory data.
Von Quintas et ai. (1988) discounted the possibility of using the TFOT or RTFOT to
age the asphalt first, and then prepare laboratory mixtures, because this would be time
corlsuming. Such an approach was used by Vallerga et al. (1967).
Von Quintas et al. (1988) also investigated "long-term environmental aging". A forced
draft oven was investigated with the following conditions:
1) place six compacted specimens in the oven at 140°F (60°C) for 2 days,
then remove three specimens.
2) increase the temperature to 225°F (107°C) and age the remaining three
specimens for 5 days.
Also, a pressure oxidation treatment was investigated, with three compacted specimens
conditioned for 5 or 10 days at 140°F (60°C) and 100 psi. The data for the oven aging
and oxidationtreatmentsare shown in Table 9. It was foundthat indirecttensile strengths
were higher for the oven-aged mixturesand failurestrainswere lower than for pressure
oxidizedmixturesimplyingthat the oven aging was more severe. Initialand recovered
asphalt propertiesare also shown, also indicatingthat the oven aging was more severe for
one mixturebut less severe for the other two. However, it shouldbe noted that the data
for the oxygen-treatedsamplesappears to be transposedfor Texas projectsand disordered
35
_. o. ,_. ,, N -- o
C
o = co o i
" - 7_
. o _ "g
= __ "_ _._._ _.._
x>, ,, _ " "'O=_) 2¢e-g,4_
: "- " sgI°. i , _-
_[- (,.) .
E . " 4...1
-- -- o o_ _ _'_ ¢_"
- j -._-- _==., k I= 0
. o ,_ o _ . g
_." = ® ,o
?J I i s-,.,,I _ A I :_
" Ii li la..
"" 'i _.s \•.x / &
- o
Ik* _ N
36
Table g. Summaryof StrengthData for SpecimensConditionedUsingDifferentAcceleratedAge HardeningTechniques (afterVon Quintaset al., 1988)
IndirectAccelerated Tensile Strainat Resilient Viscosity
State/ Aging No. of Penetration Viscosity Strength Failure Modulus AgingProject Method Days (77"F) (140"F) psi mils/in, ksi Index
Michigan Unaged 0 60 2144 84 14.56 482 1.00 "MI-OO21
OxygenBomb 5 63 3844 111 14.04 480 1.79
10 43 7542 123 12.13 582 3.52
Forced Draft 2 76 2512 "120 9.88 m 1.17Oven
7 49 5897 139 6.59 _ 2.75
Texas Unaged 0 30 4388 129 9.01 601 1.00TX-O021
OxygenBomb 5 55 3488 151 8.84 738 0.79
10 64 2822 167 5.98 683 0.64
Forced Draft 2 37 5654 200 5.11 _ 1.29Oven
7 30 9904 241 2.69 -- 2.26
Virginia Unaged 0 37 6000 114 7.97 758 1.00VA-0621
Oxygen Bomb 5 32 28021 128 10.23 569 4.67
10 35 25021 121 10.66 431 4.17
Forced Draft Z 47 4401 146 5.37 _ 0.73Oven
7 27 7910 177 2.69 -- 1.32
37
for the Virginiaproject. The viscositydata for the oven aged Virginiaproject also appear to
be disordered, and there is no modulusdata for the oven aged samples.
Von Quintas et al. (1988) recommendthe oven aging approachover the pressure
oxidationapproach. However, this authorfeels that more extensive researchis needed
before a methodshouldbe selected,particularlysince the data is very limitedand
questionable. The future researchshould includesome evaluation of the oxidationlevels
achievedby chemicaltests rather than exclusivelyby physicaltests. Temperature levels
used should be carefully evaluated, since these may cause specimen disruptionin some
cases. Such a problem may have occurredin the AAMAS study and contributedto the
unusualtrendsin the data. In addition, the levelsof aging need to be examined closely.
The viscositydata in Table 9 were used by this authorto determinean aging index (see
Eq. 2.1) for each aged sample and it can be seen that the highest levels achieved are low
comparedto those achieved in the field, as reportedby Petersen (Table 5). Also, follow-up
studiesshould be done on the five field projectsutilized in the AAMAS study.
An excellent feature of the AAMAS study is their emphasisthat the tensile strain at
break is a better indicator of the effect of aging than the tensile strength. This is logical
since the dominanteffect of aging is embrittlementand failure to accommodatetrafficand
environmentallyinducedstrains.
2.4.2 Oxidation Tests
Kumar and Goetz (1977) describe a study of the effects of film thickness,voids and
-. permeabilityon asphalt hardeningin an asphalt-aggregatemixtures. Their methodof
hardeningthe mix involved"pulling"air through a set of compacted specimensat a constant
head of 0.02 in. of water. The low head was used to avoid turbulence in the air flow
through the specimen. The specimenswere maintainedat 140°F (60°C). The test was
interruptedat periodsof 1, 2, 4, 6 and 10 days and the samples tested in simple creep.
The creep testdata were the only measures of aging in this study. There was no recovery
38
of asphalt from the aged mixtures in this study, therefore,it is very difficult to assess the
extent of aging achieved. "
A valuable feature of the Kumar and Goetz study is the quantifyingof film thickness
and permeability. They evaluatedopen-graded and dense-gradedmixturesproducedwith a
range of air voids, permeabilities,and film thicknesses. It was concludedthat for open-
graded mixes, the ratio of a film thicknessfactorto permeability is the best predictorof
resistanceto hardening. However, for dense mixtures,permeabilityis the best predictor. It
shouldbe notedthat Goode and Lufsey(1965) also concludedthat permeabilitywas a
better indicatorof aging susceptibilitythan voidscontent.
Kim et al. (1986) utilizedpressure oxidationto age laboratorycompactedsamples
representativeof Oregon mixtures. Samples were subjectedto oxygenat 100 psi and
140°F (60°C) for 0, 1, 2, 3 and 5 days. The effects of aging were evaluated by resilient
modulusand fatigue life determinedby the diametraltest. The modulusresults for the
three mixturesevaluatedare shown in Figure 14. Data are plottedin terms of modulus
ratio, viz:
modulus ratio = modulusof aqed mixture (2.3)modulusof unagedmixture
Modulusratiosgenerally increased with aging time and more rapily for the poorly
compactedmixtures (88% compactionlevel). However, a problemexhibitedby one of the
mixtureswas a loweringof modulusin the earlypart of the aging procedure. This was
attributedto a loss of cohesionin the samples at the temperature of 140°F (60°C) used in
the aging procedure. Similar resultswere found with pressure oxidationtests done by Von
Quintas et al. (1988) as shownin Table 9. This is a potentialproblemwith any aging
procedurefor compactedsamples, if an elevatedtemperature is used. Some confinement
of the samples may be desirable or it may be necessary to use a lowertemperature.
The fatigue data obtainedby Kim et al. is shownin Figure 15 and shows a steady
increase with aging with the more poorlycompactedmixturesshowinglonger fatigue lives.
39
2.0
D ID-ST
1.8 [] PR-DR
A Am-HR
o 1.6
_q
= 1.40z
1.2
1.0
0 1 2 3 4 5
Days in the PO_
(a) AI: 88_ Compaction Level
1.6
I PR-DR
E! ID-S-r1.4
• AIC-HR
O._ 1.2
-- 1.0
Oz
0.8
0.6
0 1 2 3 4 5
Days in the P0V
(b) At 94_ Compaction Level
Figure 14. Aging Modulus Ratios for Three Asphalt Mixtures (after Kim et al., 1986)
40
104
S
o 1°__" .
,o $
[7 •ID-ST
I PR-DR
& AIC-H_102 _ = I t t t
0 1 2 3 4 S
Days in the PO_
. (a) At 884 Compaction Level
10 4D ID-ST
s • PR-DR
= -°_,',1
0
_ 10 30
4,0
"_ Si,o
0.
l:g
102
0 1 2 3 4 5
Days in the PO¥"
(b) At 944 Compaction Level
Figure 15. Fatigue Ufe of Three Mixture Specimens (after Kim et al., 1986)
41
These testswere carriedout at a fixed level of applied tensilestress and therefore, the
stiffer mixes experiencelowerstrain levels. A differenttrendwould be expected if a fixed
tensilestrainhad been used.
2.4.3 Ultraviolet/Infrared Treatment
Hveem, Zube and Skog (1963) presenteda comprehensivedescriptionof various
tests and specificationsfor pavinggrade asphalts. This includedan infrared weatheringtest
for Ottawasand mixtures. The mixturesare producedfrom sand of No. 20 to No. 30 size
and 2% asphalt. This gives a "statisticallyuniformfilm of asphalt" of about 5 to 7 microns
(i.e..005 to .007 mm). The mixtureis tested in a "semi-compactedstate'. The infrared
radiationwas controlledto give a constantmass temperatureof 140°F (60°C) and an air
stream at 105°F (41°C) was maintainedacrossthe specimen.
Hveem et al. describea calibrationprocedureto determine the number of hours
requiredin the weatheringtest to correspondto field aging. This procedureutilized a shot
abrasiontest to evaluatethe aged mixtures. The authors "statewith some confidence"that
1000 hrs of exposurein the weatheringmachineis approximatelyequal to 5 yrs field aging.
Kempand Predoehl (1981) also utilizedan "actiniclightweathering test', but do not
includeany data in the paper reviewed. The test was conductedwith the following
conditions:
1) 35°C (95°F)
2) 18 hrs
3) 1000 MW/cm2of 3660 Angstromactinicradiafion
The authorsnote that tests run with differentasphalt film thicknessesindicatethat the tests
measure the hardeningwithinthe outer5 micronsof the asphalt film.
Hugo and Kennedy(1985) evaluatedthe effect of UV-radiationon mixturesobtained
from laboratoryprepared slabs and freshlyconstructedfieldprojects. Two approacheswere
used. The firstwas similar to that used by Traxler (1963), and used 54 hoursof UV
42
exposure. The secondwas by 0se of an Atlasweatherometer for a periodof 14 days.
Tables 10 and 11 show the resultsof these tests,expressed in terms of the viscosityat
77°F (25°C) of the recoveredasphalt. Note that the levels of aging index are verysmall
comparedto similar testsconductedon pure asphaltby Edler et al. (1985), see Table 7.
Tia et al. (1968) conductedan extensivelaboratorystudy which includeddeveloping
aging methodsby heat and ultravioletlight. Table 1 includesa summary of the tests done.
Their tests showedthat similarlevelsof aging were achieved in mixturesamples aged by
either ultravioletlight or oven aging. They recommendedthat an improvedprocedure
shouldbe developed incorporatingboth ultravioletlightand forceddraft oven heating. An
operatingtemperature of 140°F (60°C) was recommended. In identifyingultravioletlight as
a major cause of mixtureaging, Tia et al. note that the resultanteffect is a surface one.
However,they cite data developedby Coons and Wright (1968) which showed much higher
viscositiesfor recoveredasphaltfrom the top quarter inch of cores than from depths of one-
half inch. This data suggeststhat althoughthe phenomenonoccursat the "surface', it
does affect a significantdepth of a mixture.
2.4.4 Steric Hardening
Hveem et al. (1963) describea cohesiographtest to measure the "setting"quality of
paving grade asphalts. The test involvesmakingfour Ottawa sand semi¢ylindrical mixture
specimens. Two of the specimensare tested immediatelyin the cohesiograph,and two
after 24 hrs storage at 140°F (60°C). The cohesiographsamples are approximately12 in.
long and slender. They are extrudedout of a supportso that they act as a cantilever and
will break into short sections at the test temperature. The "lengthof break"determinedis
the average length of the brokensectionsof the samples.
Hveem et al. note that if there is a differencebetween the two sets of samples, this
reflectsthe tendency of an asphalt to "structure'. They note that this is not substantially
43
Table 10. Aging Resultsof Cores from Newly ConstructedRoad Surfacesand LaboratoryPreparedBriquettesUsing 54 HoursUV-Radiationby AdaptedTTI Method* (afterHugo andKennedy, 1985).
Mpacha,Southwest Lab Mix Lab Mix
Johannesburg Pretoria Cape Town Africa No. 4 No. 5
Binder, % 5.9 6.8 5.6 9.6 7.0 6.0
Binder Penetration 40/50 60/70 80/100 60/70 80/100 80/100
Voids, % 13.2 2.2 4.9 11.0 12.0 13.0
Mix CompositionR Semigap Semigap Cont. Sand Gap Cont.Reference Table No. 15 15 15 15 12 12
Viscosity % 0_Log Pa.S @ 25 C
Surface: 0-5 mm
Aged 6.46 6.27 5.90 5.42 6.08 6.13
Unaged 6.05 5.87 5.16 5.14 5.16 5.16
Aging Index 2.57 2.51 5.50 1.91 8.32 9.33
*Tests were done at University of Stellenbosch. Bitumen extracted from 5 mmslice off top of sample.
44
Table 11. Aging Results of Cores from Newly ConstructedRoad Surfacesand LaboratoryPreparedBriquettesUsing 336 HoursAtlas WeatherometerUV-Radiation*(afterHugo andKennedy, 1985).
Mpacha,Southwest Lab Mix Lab Mix
Johannesburg Pretoria Cape Town Africa No. 4 No. 5
Binder,% 5.9 6.8 5.6 9.6 7.0 6.0
BinderPenetration 40/50 60/70 80/100 60/70 80/100 80/100
Voids,% 13.2 2.2 4.9 11.0 '12.0 13.0
Mix Composition-- Semigap Semigap Cont. Sand Gap Cont.ReferenceTable No. 15 15 15 15 12 12
Viscosity _n 0_Log Pa.S @ 25 C
Surface: 0-5 mm
Aged 6.64 6.53 6.05 6.56 6.18 5.86
2 mm Below Surface
Aged 6.18 6.25 5.93 5.6? 5.48 5.47
RelativeAgingIndex 2.88 1.91 1.32 7.7B 5.01 2.45
*Testswere done at Council for Scientific& IndustrialResearchLaboratories,Pretoria,by A. Jurriaanse,A. Edler.C.M. MacCarronand M.M. Hattingh. Bitumenwas extractedand testedby the same method as thatused for the BKS survey.
45
due to the 24 hr curing period, because remoldingthe samples reduces the 24 hr reading
and, in some cases, reduces it to the originalstate.
The author is not aware of any otherstudies that have investigatedsteric hardening
of mixes.
2.5 Relationship Between Laboratory Aging Tests and Field Performance
As noted at the beginningof this chapter, the majorityof researchershave considered
aging of the binder ratherthan that of the mixture. Throughoutsections2.3 and 2.4 very
littleof the informationpresentedincludesfield data. However, there are notable
exceptions,such as Lee (1973) who presenteddata such as shown in Figures7, 8, 9, and
10. Also, Von Quintas et al. (1988) includedsome data relatingshort-termfield aging to
laboratoryaging (see Figure 13), and Petersen (1989a)compared TFAAT aging indiceswith
those from field data reported by Vallerga and Halstead (1971). In each of these studies,
asphaltwas recovered from the field, consistencydata obtainedand comparedwith
laboratoryaged asphaltproperties. The Von Quinatas, et al. (1988) study recovered
asphalt from the laboratoryaged mixture.
The resultsfrom several test roads willbe reviewedbelow, where controlledprojects
provideexcellent data. However,there are several cases where insufficientdata are
available to draw meaningfulconclusions. Welbom (1979) has previouslydone an excellent
job of reviewingmost of thiswork.
2.5.1 California Studies
a) Zaca-Wigmore Test Roads
Two test roadswere constructedin 1954 and 1955, with 10 test sections using
differentasphalts. Several researchershave evaluated the results.
Simpsonet al. (1959) compared the viscosityof asphalt recoveredfrom various
sectionsof the two test roads (referred to as period1 and period2) after 16 months
and three years servicewith that of asphaltaged in the Shell microfilmtest for two
46
i
hours. Viscositywas determinedat 77°F (25°C) with a sliding plate viscometer at a
shear rate of 0.05 sec1. Figures16 and 17 show selectedresults. Note the
hyperboliccurves in Figure 17 which showstwo sets of data representingthe two
constructionperiodswhere differentprocedureswere used. Note that there is a
definite correlationbetween fieldand laboratorydata.
Zube and Skog (1969) publisheda final reporton the test road. The final
report includedthe following:
1) An excellentcorrelationwas found for the hardeningof asphaltsduring
plant mixingand duringthe TFOT.
2) The rate of hardeningunderequivalentweatheringand pavement
conditionswas influencedby the source of the asphalt. The hardening
can be attributedmainlyto the initial voidcontentand rate of change in
void contentsduringpavement life.
3) The amountof fatigue crackingappeared to be related to the consistency
of the recovered asphalt as measuredby penetration or viscosity. Other
forms of crackingappeared to be relatedto the gain in shear susceptibility
of the asphalt duringservice life. This was also indicatedby loss of
ductility.
This last conclusionregardinga relationshipwith ductility lossis also noted by
Kandhaland Koehler(1984) from observationsin Pennsylvania. These findingsare
discussed later.
Davis and Peterson (1967) used inversegas liquidchromatography(IGLC) to
study Zaca-Wigmoreasphalts. The data for phenol retentioncoefficientfor laboratory
aged asphaltcorrelatesextremelywell with performance ratingfor eight of the
sectionsafter 51 monthsservice. Petersen (1984) explains that phenol retentiontime
is a measure of the concentrationof polar functionalgroups in the asphalt.
Furthermore,the concentrationof these polar groups is related to potential for
47
"3
63 _. G80
E 60 S 60 OH oB.:)
40
c_'/_ v _ _ oe-z
Jj._,.,_ Do2
0 O"_'_'Z ' ' ' _ _ ' "_ ' ' '0 20 40 60 80 _C " 20 40 60 8C
'_2 hr of 225 F/% _z hr ot 22.5 F'/%
a) 16 Month Service b) 30 Month Service
Figure 16. Pavement Aging Index Compared with Microfilm Aging Index. (after Simpson et al., 1959)
.CO - ' " E
T
! T e° !o 6
w
,o ,., i -
_ .
_ _0o
= e-/
// i "o
i
0PuQmill0 IO 20 30 40 5C 910 I 2 3 4
F_.emenl AOe.month I_- _ Ct 225 F. 5 Film
a) Test Se_on Data b) M_rofilm Aging Test Data
Figure 17. Increase Viscosity with Time (after Simpson et al., 1959)
48
crackingor fracturing. Figure 18 shows the excellentcorrelationbetween phenol interaction
time and performancerating.
b) Other California Studies
Followingthe Zaca-Wigmoretest road, other test sectionswere installedin the
1960's. Kemp (1973) has describedthe relationshipbetween laboratorydurability
tests and field hardeningas measured by recoveredviscosityat 77°F (25°C). Kemp
and Predoehl(1981) indicatethat these tests failed to provideconsistentdata due to
many uncontrolledvariables. The CaliforniaTransportationLaboratorytherefore
pursueda new approachwhere most variablesand constantswould be controlled.
Kemp and Predoehl (1981) report on a studyconducted between 1974 and
1980, where laboratoryproducedspecimens3.5 in. highby 4 in. in diameter were
aged in four distinctclimates in the field. Three asphalts of high, medium, and low
temperaturesusceptibilitywere evaluated,with two aggregate types (absorptiveand
nonabsorptive)prepared at three voidcontents. Sampleswere exposedto the field
conditionsfor 1, 2, and 4 yrs. The propertiesof the aged samples, and asphalt
recoveredfrom them, were comparedwith those obtainedfrom laboratory tests.
Figure 19 shows data for the microviscosityat 77°F (25°C) for each asphalt
recoveredfrom the field aged samples. Note in particularthe characteristichyperbolic
shape of the aging curves. The majorconclusionsdrawn from this study were as
follows:
1) High average temperature(thermal oxidation)is the most significant
factoraffectingthe rate and amount of asphalthardeningin hot
climates.
2) Void content is a contributingfactor and its effect was similar among
all asphalts.
49
z=
:_ 210
190
'_ 170
- G
-_,
UC
o J•_ 130
t_
Ot,.Q
el.C
_ II0 -a
0
_L
90 I i I t i I i iFoiled 7 6 5 4 3 2 I
Surface Performance Rating After 51 Months
Figure 18. RelationshipBetween Phenol InteractionCoefficientand Pavement SurfacePerformanceRating (after Petersen, 1984)
50
_1000 ( s 4, I ! 1 ,1000 : ; ' 1 i
I I I
J
tOOO_ 1000
&"I
- ° Iw
:= ii ,T6 .
• w
: so o so '
• , jWI[.ATNI[IIINGsIT[¢i+liMATI[ AVGilrTIPMil W(ATkl(IIIING &VGIrT(MP"_S___ STMIOIr SJTI[ ¢LlUAT( *
O ;GqT |_G; ¢OAs'rAL '_2 t O S'OmT8_¢_ COASTAL $2 i
OG $OUT(4L&IJ[Sd_llM (11TO ¢J_&!alIC_VII'L[T G3 1,0 • SACIIAM( IITO (Jl( IT YALL[T 63 O "_t+"*°(-u_,_t+,,,--'t* 0 SOur.LAIn( m_. STII40( -*lO_ T++J_-- 4 I6 llllOlO I,OwO_[S(IIT ?3 0 I
6 INOIO LOW0(5.[111+ "P_0r
Io _ I I l I I0 I I I ] t .10 _0 _1 40 S_ 60 |0 _g _ IJ0 _'_
WtATI+arIIlIeG tll[RlOO (IdONTNS) W(ATI4(III*NG I_RIOD (IdONT+,IS)
a) Valley Asphalt b) LA Basin Asphalt
)000_ ! i I l i
LlO00
_ w
_o
S
mCa_Tm(ltmO av4 TIW_ll'r f Ckllill_PIl_ ell
41 s_m4alltlto laliT v_U[_ 63.0
O ll_rll L,I+_[ 16110(-iiit_m_llm -- 41 . I41, *ll*O idw IJl_nv IP3.0
tC I I I I I_0 _0 )0 40 SO 60
WI[AT_II_Ul Plll*OO (IIO*ITNS I
c) Santa Maria Asphalt
Figure 19. Effect of Climate on Hardening (4 Climatic Sites) Combined Aggregates andVoids (after Kemp and Predoehl, 1981)
51
• 3) Aggregate absorptionis a contributingfactor and is more significant
with more volatileasphalts.
4) The tilt oven durabilitytest can be used to predicthot climate
hardeningof asphalt.
5) The followingfactorswill improvedurability:
a) adherenceto specified compactionrequirementsto reduce
voids
b) avoid use of absorptive aggregates
c) use the softest grade of asphaltconsistentwith curingand
stabilityconstraints
d) insulatethe asphalt concretesurface with a cover suchas a
reflectivechip seal in hot areas.
2.5.2 Michigan Test Road
Michigan State Highway Departmentcompleted constructionof the test road in 1954.
Six differentasphaltswere used in six test sections. Welborn (1979) has summarized the
major findingsof various researcherswho have studiedthis test mad over a period of over
25 years, in a study by Corbett and Merz (1975), the major conclusionwas that,
consideringthe age of the road and the currentlevel of hardnessof the binders,no
distinctioncould be made among the binders used.
Corbett and Schweyer (1981) use data from the test road to illustratesome concepts
regardingcompositionand rheoiogyeffects on age hardeningof bitumen. They present
Figure20, which is presumablybased on average data from the test road, and make the
followingpoints:
1) there is a two- to fourfoldviscosityincrease and a 30% penetration
decrease by the time the pavement is a year old.
52
J
VIS. ",.. 60C (140F)
PEN. '_ 25C (77F)
DUCT. _ 25C (77F)
"22,000 ilc 110 1"),, -
00(3 1_X _ '_,,_F) -
18,000 - 9(
16,000 - 8(
14,000 - 7(
12,000 6_
10,000 5 :.,gO:
80O(3 41
600C 3 30
4OOO 2
2000 I 1 •
J l l0 5 10 i 5 20 25
YEARS
Figure 20. Age Hardening InvolvesConsistencyChanges (after Corbett and Schweyer,1981)
53
2) thereafter the viscositylevel increasesmore slowly and typically reaches about
20,000 poisesat 140°F (60°C) in about 25 years, based on an originalAC-20
asphalt.
3) ductilityat 77°F (25°C) is, in part, a flow test and it drops off rapidlyin
later life.
Goodrich(1985) has summarizeddata from the Michiganand other test roads. It is
interestingto note that he found that the performance ratingsof the Michigantest roads
best correlatedwith:
1) percentductilityretained
2) softeningpointof originalasphalt
3) viscosityaging ratio
(18 yr viscosityat 140°F/originalviscosityat 140°F)
4) percent asphaltenes
5) percentair voids after 52 monthsservice.
The analysis presentedby Goodrich(1985) does not clarifywhether the percent
asphaltenes is that measured on originalasphalt,or on asphalt-extractedfrom the test road.
2.5.3 Texas Test Roads
Gallaway (1959) reportson a studyof surface treatmenttype pavements built in
Texas in 1954. Eleven test roads were builtand performance monitored. One of the
conclusionsgiven by the author is to restrictthe oxidationsusceptibilityof asphalts in the
specificationrequirements.
Benson (1976) addressed the relationbetween low temperaturecrackingof asphalt
pavements in Texas, and asphalt hardening. Nine test sectionswere evaluated in detail
and relationshipsbetween asphaltphysicalproperties and performance established. Good
correlationswere found between the viscosityat 77°F (25°C), aging index, and air voids
with performance. Laboratorymixturesamples 17 in. in diameter by 2 in. deep were
54
J
compacted using a Texas gyratory compactor. These laboratorysamples were called
"pizzas'. Pizzas were aged in the field in two locationswith differentlevels of solar radia-
tion and precipitation. Controlsets were stored in the laboratory. After three years the
pizzas were cored, tested for resilientmodulus,and penetralJonand viscositytests, both at
77°F (25°C) were carriedout on the recoveredasphalt. The majorityof pizzas were regular
dense mix. However,a set of 15 was made using an "antihardening"additive. This was a
combinationof paraphenyclenediamineantiozonatesand ultravioletlight inhibitors,at a
treatmentlevel of 1%. Bensonalso conductedthe "ActinicLightTest" on asphaltsamples.
The testwas conductedon 10 micronfilms with the followingconditions:
1) 95°F (35°C)
2) 18 hours
3) 1000 microwatts/cm2 of 3660 Angstromwavelength radiation
These are identicalto the conditionsreported by Kemp and Predoehl (t981). The thin film
oven test was also conducted. Asphaltpropertiesafter these two aging tests were
comparedwith recovered asphaltproperties. No significantcorrelationwas established.
An excellentfeature of Benson'sstudy is the hardeningmodelsdeveloped. These
were developedfor viscosityand penetrationand take the form:
V = a t b (2.4)
and, P = a + b In(t) (2.5)
where, V = viscosityat 77°F
(megapoises,shear rate = 0.05 sec" in a slidingplate viscometer)
P = penetrationat 77°F
t = time from laydown(months)
a+b = nondimensionalcoefficients derivedfrom least squares regression
analysis.
55
It shouldbe noted that the curves for viscosityshown in Figure 21 are very similar to those
obtained by Lee (1973), as presented earlier, and illustrated in Figures 7 through 10.
However, Benson used an exponential model rather than a hyperbolic one such as used by
Lee. Benson found better correlation coefficients with the penetration data and attributed
the poorer correlation for the viscosity model to the complexity of the microviscosity test.
Perhaps better correlationwould have been achieved with a hyperbolicmodel. An example
of the data is shown in Figure21, and a collectionof curves for several sites in Figure 22.
The penetrationmodel was chosenbecause of 14 nine-year-oldpavements studied,this
model fitted the data with a correlationcoefficientgreater than 0.95.
Benson notes that the key to applyingthis model is to predict the parameters "a" and
"b', and that "a" is a measure of short-termhardening,since when t is equal to one month:
P=a+bln(1)=a (2.6)
The parameter "b" relates to the curvatureand representslong-termaging susceptibility. A
predictiveequation for "a"was developed from multipleregressionanalysis,yielding:
a = 0.52(ORP) - 2.0 (2.7)
where, ORP = OriginalPenetration.
This results in "a" values of about half the originalpenetration. It was not possibleto
developan equation for "b". However, Benson notes that this shouldbe related to
environmentalfactors and long-termchemicalreactions.
Other major findingsfrom Benson'sstudy are as follows:
1) There was not a significantrelationshipbetween laboratorymix density
and hardeningsusceptibility. However, field data did show a significant
correlationbetween cracking,hardening,and void content.
2) The additivesinvestigatedwere ineffectiveas hardeninginhibitors.
56
V = 8.4 t °'_5 P = 34- 4.2 Intn =7 r=0.98 n= 7 r=-0.99
70_ 70
E60- E 6°
; h,E50- u. 50-
h-_" 40 h- 40
;_o _so_8 2o- -_ 2o-
0 2b 40 60 8b 160 120 0 2b 4'0 60 go Ibo li-c
Time (Months) Time (Months)
a) Site 40
V = 6.6t °'zl P= 43- 3.0 In 1'n=6 r= 0.98 r=6 r= -0.98
,o !_6o _ 6050 _ 50
_ _o. _ _o"- --'_ 20o 20- _.
_ _ ,o_ 1o-h| , i .
0 zo 40 _0 8'0 t60 Izo 0 io 4'0 6b 8b ,_,..Time (Months) Time (Months;
b) Site 41
Figure 21. Typical Hardening Curves for Two Different Sites (alter Benson, 1976)
57
70-
E 60-E
v
5OLL0p..
TRAN.E.9 CRACKo FREQ.
I0
0 60 I00 120
Time (Months)
Figure22. PenetrationHardeningCurvesfor NineTestSites (afterBenson,1976)
58
3) The Actinic Ught Test and Thin Film Oven Test were not reliable
indicators of hardening susceptibility.
2.5.4 Pennsylvania Test Roads
Kandhal and Koehler (1984) reported major findingsfrom test roads constructed in
1961, 1962, 1964, and 1976. Several test sectionswere constructedand their performance
monitored. A very thoroughset of tests was completedon originaland recovered asphalts.
Figure23 shows viscositydata for recovered asphalt. The authors point out that these are
hyperboliccurves, similarto those observedby Lee (1973). However, they note that low
temperatureductility is an importantfactor,with pavementscontainingasphalt of low
ductilityshowinggreater tendencyto crack.
2.5.5 Iowa Study
As noted earlier, a significant study was conductedby Lee (1973)relating the
propertiesof laboratoryaged asphaltsto those of recoveredasphaltsfrom nine test
sections. Hyperbolicmodels were developed for field and laboratorydata, as illustratedin
Figures 7, 8, and 9. Field and laboratorydata were related with "time-equivalency
correlationcurves" as shownin Figure 10.
2.5.6 Oregon Study
Kim et al. (1986) attempted long-termaging of asphalt mixturesusingpressure
oxidation. It was shownthat this approachdid producesignificantaging as shown in
Figures14 and 15.
Thenoux et al. (1988a) in a companionstudy attemptedto relate asphalt properties to
field performance. The properties includedthe commonphysicaldata as well as the
Corbett-Swarbrickfractionationdata. The attemptto developrelationshipswas frustrated by
the lack of controlof the field test sections. Eightsites, dispersedaround the state were
selected. These were nominallyfive to 10 years old, and asphalt had been stored in the
59
/°
// s
i " -°" ..
8 ..
e_6 . .
g_4 /
0
iAC S
_Z' AC 2 ..................S__ Ac s
AC 4
ACS
! AC6
IZO 40 60 80 KX) 120
Tlmo m Months
a) Viscosityat 77"F4
3
o_
_ ./_j°'°'°" --3
3 2 " "=" Acu ........
::',-:: .........."_11" A_.,.........
E#
_ _ _ .......L.
8_ . AC 6 ......
I I n I I I )
_.I,,,.,._L2o ,o..__ ,o ,oo ,__.. Viscosity._t]A_'K._
Figure 23. Viscosity versus Time Relationships, 1964 Pavements (aider Kandhal andKoehler,1984
60
state materials laboratoryfor each project. Since only the initialcondition and end condition
(i.e. at the time of the study) were i_nown,no rate of change informationcould be
established.
The Oregon study wouldhave been more valuable had more recoveredasphalt
propertiesbeen obtained,and if historicalpavementconditiondata had been available.
2.6 Relationship Between Chemical Composition and Field Performance
Many of the studiesdescribedearlier in thischapter have involveddeterminationof
the compositionof an asphalt,as indicatedin Table 1. Changes in compositionwere an
indicationof the extent of aging in a laboratoryor field aged sample. Some studies (e.g.
Vallergaand Halstead, 1971) includeda comparisonof a compositionalparameterwith the
observed performanceof a field project. This section presentsa review of several studies
that specificallyaddress the questionof whether there is a relationshipbetween composition
and field performance. An excellentpaper by Goodrichet al. (1986) addresses this topic
and forms the basisfor what is presentedbelow.
2.6.1 Tests for Asphalt Composition
Goodrichet ai. (1986) indicatedthat the procedurescommonlyused fall into six
categoriesas follows:
1) Fractionationby precipitation:solventprecipitation,chemicalprecipitation.
2) Fractionationby distillation:vacuumdistillation,thermogravimetric
analysis.
3) Chromatographicseparation: gas chromatography,inversegas-liquid
chromatography,liquidchromatography(adsorption,ion exchange,
coordination,thin layer, size exclusion).
4) Chemicalanalysis: spectrophotometrictechniques(infraredultraviolet,
nuclear magneticresource,X-ray fluorescence,emission,neutron
activation),titrimetricand gravimetrictechniques,elemental analysis.
61
5) Molecularweight analysis by mass spectrometry,vapor pressure
osmometry,and size exclusionchromatography.
6) Indirectcompositionalanalysisby internaldispersionstabilitytests.
Of these tests, fractionalseparationhas been used most. Four types of procedureare
commonly used:
1) Chemical Precipitation,e.g. the RosUer-Stembergapproach, standardized
in ASTM D2006, but discontinuedin 1976.
2) Solvent Fractionation,e.g. the Traxler-Schweyerapproach.
3) Adsorption-DesorptionChromatography,e.g. the Clay-gel procedure,
standardizedin ASTM D2007, and the Corbett-Swarbrickprocedure,
standardized in ASTM D4124.
4) Size ExclusionChromatography,e.g. the gel permeationchromatographic
(GPC) techniquestandardizedin ASTM D3593.
A brief descriptionof each of these proceduresis given by Goodrichet al. (1986). The first
three types of procedureinvolveprecipitatingan asphaltenefraction from a solutionof an
asphalt in a solvent. It shouldbe noted that the Corbett-Swarbdckapproach uses a
differentsolvent to the other methods,and that the "asphaltenes"fraction obtainedby this
approachwill be differentto that obtainedby other approaches. This reinforcesthe point
that fractionalseparationdoes not separate the asphalt into unique chemicalcompounds,
but rather into groupsof compoundsof a similar type.
The GPC separationapproach has been used by a number of researchersduringthe
1970's and 1980's. In particular,a group at Montana State Universityhas used this
techniqueextensively(Jennings, 1985). The asphalt is separated accordingto the size of
molecularassociations,and indeed,Jennings(1985) has likened this techniqueto a sieve
size analysisfor aggregates.
2.
2.6.2 Value of Compositional Analyses
Goodrichet al. (1986) assessedthe value of compositionalanalysesby fractional
separationtechniques. They concludethat suchanalysesprovidea methodfor following
changesin an asphalt,whether during laboratorystudies, refinery processing,field mixingor
field aging, or recycling. They also conclude that compositionalparametersobtainedhave
notcorrelated well with field performance,nor have ratiosbased on the fractions. However,
they indicatethat physicaltests have correlatedwell. Goodrich et al. also emphasize a
point that is frequentlyoverlookedwhen consideringthe durabilityof asphalt pavements;
namely,that construction,mix design,and climaticvariables have an overridinginfluence.
The asphalt content, film thickness,air voids,and permeability are very important. As
notedin an earlier section,Kumar and Goetz (1977) showedthat film thicknessand
permeabilitywere the mainfactors controllingagingof open-gradedmixtures. However, for
dense mixtures,permeabilitywas the major factor. The Californiadurabilitystudy (Kemp
and Predoehl, 1981) is a clear indicatorthat climatic effects dominateaging of mixturesin
the field. The data in Figure 19 representsthree distinctlydifferentasphaltsand a wide
range of air void contentsbut these have much less effect on aging than the climate.
With the conclusionsof Goodrichet al. (1986) in mind, it shouldbe noted that if the
constructionand climaticvariableswere similar,compositionalparametersmay be an
indicatorof potentialperformanceamong projectsusingdifferentasphalts: In a thorough
review of the relationof asphalt chemistryto physicalpropertiesand specifications,Halstead
(1985) notes that althoughthe value of compositionalanalysis has been questioned,work in
South Africa reported by Jamieson and Hattingh(1970) showed reasonable correlation
between compositionalparametersand road performance. Halstead emphasizesthat such
parametersare one indicationof performance, but other factors must be considered.
Clearly good mix design and constructionare required. Goodrichet al. (1986) note that
63
good correlationswere obtained with Australiantest roads (see Dickinson, 1980), which
involved some surface treatments as well as dense mixtures.
2.6.3 Significant Compositional Parameters
A simple "model" of asphalt is that it consistsof a colloidaldispersionof
"asphaltenes"in a dispersionmediumof "maltenes"(Barth, 1962). The asphaitenes have
the highest polarityand tendencyto interactand-associate. They are also insoluble in a
nonpolarsolventsuch as pentane, hexane or heptane. The maltenes are soluble in such
solvents. This broad distinctionenables fractionationof asphalts. Figure24 shows
schematicsof two of the procedurenoted earlier. It may be seen that the maltenes are
further separated. The four fractionsobtainedwith the Corbett-Swarbrickprocedure or the
five from the Rostler-Sternbergprocedureform the basis for most comparisonsof compo-
sitionof asphalt with performance. Three parametersderived from the fractions obtained
from these proceduresare describedbelow.
a} Rostler Durability Ratio (RDR)
The fractionsobtainedfrom the Rostler-Stembergproceduremay be used to
develop a Rostlerdurability ratioas follows:
N +AtRDR = (2.8)
P+A=
where N, A1, P, A=are the nitrogenbases, first acidaffins,paraffins,and second
acidaffins,respectively.
Rostlerand White (1962) presentedextensivedata relatingthis ratio to percent
weight loss from pellet abrasion tests on Ottawa Sand mixturesbefore and after
aging. The pellet abrasion test was an adaptationof a shot abrasiontest described
by Hveem et al. (1963). Agingwas accomplishedin an oven for 7 days at 140°F
(60°C). The RDR is the ratio of the most reactive fractionsto the least reactive.
64
I ASPHALTI
n-PentaneI
I IPrecipitate Solution
i ASPHALTEN,=SJI NITROGENBASES1 H'S0"85%
1STACIDAFFINS I H_S0Zl98 %
NDACIDAFFINS ] H2 S0'4S04
L PARAFFINS
a) Chemical Precipitation by Rostler-Stemberg, ASTM D2006
I ASPHALT I1
n-HeptaneI
�I
MALTENES I Precipitatei J'Chromatography
on Alumina
n-Heptane ,[ SATURATES ]
Co
. ,o,°.e ,L"_'"_"_l.,o..,,°_
; Methanol-Toluene f POLAR ITrichloroethglene AROMATICS
b) Adsorption/DesorptionChromatographyby Corbett-Swarbrick,ASTM D4124
Figure24. FractionationProceduresfor Asphalt Cement
65
Rostler and White indicated that asphalts with RDR less than 1.14 had excellent
abrasion resistance.
Vallerga and Halstead (1971) presented an overview of an extensive BPR
study, which sampled more than 300 asphalts from projects 11 to 13 years old. This
included Rostler-Stemberg analyses, and it was concluded that for pavements with
less than 2'/o air voids there was a direct relationshipbetween the RDR and aging
measured by the viscosityof recoveredasphalt. For higher air voidsthe RDR was
overshadowedby other effects, but the ratio for minimumhardening appears to be in
the range 1.0 to 1.4. It was significantthat mostof the pavements containing
asphalts originallyfoundto have high RDR values did not survive to be 11 to 13
years old.
Jamiesonand Hattingh (1970) conducted an extensivestudy involving field
projectsusing 18 bitumensin "premix"and evaluationof asphaltswhich includedthe
Rostler-Stembergapproach. They concludedthat there was good correlationbetween
the RDR and road performanceand that appropriatelimits for coastal South Africa
wouldbe 1.0 to 1.7, which is a similarrange to that suggestedby Vallerga and
Halstead(1971).
Goodrichet ai. (1986) pointout that the RDR does not consider the asphaltene
content and that it has limited correlationwith field data (overlookingthe findingsof
Jamiesonand HatUngh, 1970). They note that the Gotolskiratio does consider the
asphaltenes.
b) Gotoiski Ratio (GR)
Gotolskiet al. (1965) proposedthis as an aitemative to the RDR:
GR=N +A I+A_ (2.9)P+A
where N, AT,P and A= are as before and A is the percent asphaltenes.
66
This parameter was usedby Jamiesonand Hattingh (1970) and by Anderson and
Dukatz (1980). The former found a range of 0.8 to 3.4 for the GR for 18 bitumens
investigatedand suggestedlimitsof 1.3 to 2.6 to ensure satisfactoryperformance.
The latter authors addressedthe questionof whether asphaltshad changed over a
30-year period, 1950-80. They concludedthat there had been statisticallysignificant
changesand that the Rostlerand Gotolskiparametershad increasedduring that time.
They also indicatethat high and low Rostlerparameters indicatepotentiallypoor
asphaltperformance,which is in agreementwith Vallerga and Halstead (1971), and
that the RDR is mostcloselyassociatedwithtemperaturesusceptibility. Anderson
and Dukatz also indicatethat a high GR is an indicatorof poor performance and that
this ratio is more closelyrelatedto aging effects.
In discussingthe Andersonand Dukatz (1980) paper, Puzinauskasstates that
there is no substantiationof the claim that the RDR is more closelyassociatedwith
temperaturesusceptibilityof an asphalt and the GR relates more closely to aging.
Furthermore,he states that in his opinionthe RDR and GR are associatedwith a
variable reactivityof asphalts.tosulfuricacid and nothingmore. These comments
appear to ignorethe work reportedby Vallergaand Halstead (1971) men_onedearlier
in this section.
In a subsequentextensionof the workat PennsylvaniaState University,
Andersonet al. (1983) concludedthat, "Exceptfor temperature susceptibility,there
have been no long-term,significantchangesin asphaltpropertiesthat can be
identifiedas significantby the analyticaltechniquesused in the project." This study
includedaddi_onaldata to that reported in 1980 by Andersonand Dukatz. Only brief
mentionof the RDR is made to the effect that it "remainedrelativelyunchangedover
the period 1950-81," and there is no mentionof the GR. The authors obviously
reconsideredthe conclusionsmade in 1980, and there is a clear indicationin the
67
discussionto the 1983 paper that the Rostler analysis was used because it was the
only analysis used in the data available from 1950 and 1960.
c) Gaestel's Colloidal Instability Index (1(3)
Bn31_et al. (1986) summarized the Gaestel ColloidalInstabilityIndex (IC) which
was defined by Gaestel et al. (1971) as follows:
IC - Asphaltenes+ Flocculants(Saturated Oils) (2.10)Peptizers(Resins) + Solvents(AromaticOils)
BrOl_et al. presentthe followingexplanationof IC:
"The higherthe ratio, the more will the asphalt cement be a geltype and the lower will be its colloidalstability. Gaestel also notesthat all the propertiesof the binder (softeningpoint,ductility,embdttlementtemperature,thermalsusceptibility,elastic recovery,shearingsusceptibility,etc.) vary significantlywith the colloidalinstabilityindex and hence with composition."
Br01_et al. did not developany data for IC, because fractionationwould be
required,and a faster GPC techniquewas preferrecl.
Tuffouret al. (1989) presentedexcellentdata relating IC to changes in asphalt
propertiesdue to aging. They expressthe index as follows:
IC = Asphaltenes+ Saturates (2.11)Naphthene Aromatics+ Polar Aromatics
Tuffour et al. carded out clay-gelcompositionalanalyses on two sets of asphalt
samples: one set recoveredfrom cores from roads showingsigns of durability and
aging problems,the otherset were laboratoryaged using the TFOT at three
- temperature levels (163, 140, and 120°C) and variousperiods of aging. Figure 25
shows the data developed for absoluteviscosity (60°C) versus IC. The authors
developed regressionequationswhich confirmedthe strongcorrelationapparent from
the figure. They concludethat high values of IC are associatedwith instabilityand
that the data suggestthere is a uniquerelationshipbetween IC and viscosity.
68
107 I i | i I I I
-- A Road Core SampleB
O Test Sample
I0 4 --
0
3o I I 1 I I I II 0 0.2 0.4 0.6 0.8 L0 1.2 1.4 1.6Goestel Index (IC)
Figure 25. Relationship between Absolute Viscosity and Gaestel Index (after Tuffour et al.,1989)
69
2.6.4 Other Pertinent Studies
Buttonet al. (1984) addressedthe influenceof temperaturesusceptibilityon pavement
constructionand performance. They conductedan extensivereview of previouswork, and
a significantlaboratorytest program,whichincludedsome fractionationtests. They make a
numberof significantconclusions,includingthe following:
1) Asphaltscontainingless than 10% asphaltenes,particularlythe softer
grades, appear to have a greater probabilityof producingslow setting
paving mixtures. However,an asphaltwill manifest itself as slow setting
only if the aggregate type and/or gradationis such that a criticalpaving
mixtureis produced(even throughthe aggregate may meet specifications)
or possiblyif densificationof the pavement is inadequate.
2) There is no correlationbetween asphalt temperaturesusceptibilityand
asphaltene content. There is no relationshipbetween asphalt temperature
susceptibilityand other chemicalconstituentsof asphaltsas determined by
the Rostler-Stemberganalysisor the Rostlerparameter.
More recently,Thenoux et al. (1988b) publisheddata, from Corbett-Swarbrick
analysesof field and laboratoryaged asphalts,which showed reasonablecorrelation
between varioustemperaturesusceptibilityparametersand the compositionalparameters,
particularlyasphaltenecontent. This conflict with Buttonet ai. (1984) may be due to the
difference in the two fractionationprocedures,and reflects a general trend in the literature
of sporadiccorrelationswhen the Rostler-Stembergparametersare considered but generally
good correlationswhen the Corbett-Swarbrickor clay-gelparameters are considered.
Thenouxet al. (1988a) also notedthat since asphalt temperature susceptibilityhas a
significantinfluenceon field performance,their correlationof compositionwith temperature
susceptibilityimpliesthat composition relates to performance.
The work of Jenningsand his associatesat MontanaState Universityhas special
significance,since it has led to widespreadadoptionof the GPC techniqueduringthe
70
1980's. Jennings(1985) reviewedthe techniqueand stated that while GPC is not the
answer to everything,it can be used to predictthe crackingpotentialof roadways.
Jenningset al. (1988) present data to supportthis predictionpotential. They also included
Corbett-Swarbrickanalysesand noted that test sectionsthat were ruttingmost severely had
lower asphaltenecontentsand that resistanceto cracking is associatedwith relatively lower
concentrationof naphthenearomatics. The first observationagrees with the first conclusion
given above for Buttonet al.
71
3.0 PROMISING AGING AND TEST METHODS
The previouschapter outlineda variety of laboratoryaging methodswhich have been
used by variousresearchers. Thischapterwill discussthe most promisingmethods for
aging asphalt-aggregatemixtures. It will also includetest methodsused to establishthe
effects of the aging procedure. This may includetests used to evaluate the extent of aging
in field test sectionsas well as laboratoryaged mixtures.
3.1 Promising Aging Methods
The literaturereview clearlyshows that there are two componentsof aging to
consider,viz:
1) short term
2) long term
The short-termcomponentoccursduringthe constructionphase, while the mix is hot. This
is probably caused by volatilizationalthoughthere may be some steric hardeningand
oxidationinvolvedand volatilizationof oxidationproducts. The long-termcomponentoccurs
while the mixtureis in place. A time periodof about 10 years is a reasonableperiod of
interest for this component. This long-termaging is predominantlycaused by oxidation,
although there may be some sterichardeninginvolvedand volatilizationof oxidation
products. Actiniclight may serve to acceleratethe aging at the surface of a pavement.
3.1.1 Short-Term Aging Methods
Candidate methods shouldsimulateconditionsduringmixing,storage, and
transportation. Therefore, conditionssimilar to the TFOT or RTFOT are appropriate. The
mixture should be loose and possiblyagitated continuouslyor periodically at a temperature
of about 275°F (135°C) to simulatethe precompactionphase of construction. Possiblythe
mix shouldbe compactedpriorto cooling to simulatefield conditions.
72
Only two short-termmethodsemerged in the literature. The first is the oven aging
method used by Von Quintas et al. (1988). This methodis describedbriefly below. Other
methodsinvolvedOttawa Sand mixtures. Such mixturesare consideredinappropriatefor
this study (e.g., Pauls and Welbom (1952), and Kemp and Predoehl (1981)) since they are
not representativeof typicalmixtureswith regard to intemalvoidsand permeability. Two
additionalheatingapproachesare worthyof consideration,viz:
1) extended mixing
2) microwaveheating
a) Oven Heating
The Von Quintas et al. (1988) procedureconsistedof heating the mixture
sample at 275°F (135°C) for 0, 8, 16, 25 and 36 hours,prior to compactionand
cooling. The effect of temperaturewas also studiedbefore adopting275°F.
b) Extended Mixing
Since mixingis an inevitablepart of the aging process, it is logicalto take
advantage of extendingthis in an aging method. An added advantage of this is that
since the mix is continuouslyagitated,all the mix is given equal opportunityto age. it
is probablybest to combinethis approachwith oven heating.
c) Microwave Heating
Microwaveovens may providea controlledmethod of extended heating. A
thorough evaluationof this techniqueis required,since there is littleknowledge
regardingits use. AI-Ohaly and Terrel (1988) providesome insightregardingthe
mechanism of microwaveeffects,and this willprovidea startingpointfor subsequent
work.
73
3.1.2 Long-Term Aging
Candidatemethodsshouldsimulatedconditionsin the field and emphasizeoxidation.
Therefore, the mixturemust be compactedand aged at a moderate temperature,probably
no more than 140°F (60oc). Several methodsemerged from literature,viz:
1) pressureoxidationtreatment
2) extended oven aging
3) infrared/ultraviolettreatment
An additionalapproachcould be used, which wouldattempt to pass osygenat low
temperaturethrougha sample in a triaxiaicell. This configurationhas promise for
evaluatingthe effects of other fluids,such as air, water, or water vapor.
a) Pressure Oxidation Treatment
Kim et al. (1986) applieda pressure of 100 psi at 140°F (60°C) withina
pressure-safevessel for periodsof 0, 1, 2, 3, and 5 days. The completeprocedureis
given in AppendixA. The pressurelevel used shouldprobablybe raised to 300 psi,
since a thoroughevaluationof asphaltaging by Lee (1973) suggeststhat such a
pressure is necessary.
Petersen (1989a and 1989b) proposesthat molecular structuringand steric
hardeninghave a major effect on levelsof field aging. He cites data such as are
shownin Figures5, 8, 9, 19, and 23, to illustratean aging model. This model I
suggests that aging progresses rapidly during the early life of a pavement (the first
two to three years), and then slowsconsiderablydue to the progressof molecular
structuringand steric hardening. The extent to which this occurs is dependenton the
propertiesof the asphalt, the climate,and the mixtureproperties (air voids). Most of
the data for field aging showsthat the asphaltapproaches an ultimatelevel of
consistency,typicallyexpressedas a viscosity. However, some field data (e.g. Figure
19) and limited laboratorydata (Figures8 and 9) suggestthat the ultimatevalue will
74
not be reached for some time. This may be explained by the model, in terms of
lower level of structuringand steric hardeningwhich exists due to the type of
"quenching"used to prepare the samples priorto testing,and/or the temperature
levelsprevailing. Petersen (1989b) suggeststhat higher temperaturesreduce the
ability of an asphalt to form molecularassociations,and therefore aging can continue
for a longerperiod and/or to a higher level than at lower temperatures.
The above model has significantimplicationsin the development of an aging
method for asphaltaggregatemixtures. Figure26 has been developed to illustrate
the possibleinteractionof temperatureand pressure oxidationwith regard to aging.
This illustratesthe need to includetemperatureas a variable in evaluation of aging
methods,and it is proposedto include77°F (25°C) in the preliminary laboratory
studiesfor thisproject. Further,Figure26 illustratesthat to use a temperaturewhich
is too high may lead to unrealisticlevels of aging when comparedto field aging. This
is significantwhen evaluatingoven aging techniqueswhich rely on high temperatures.
Kumar and Goetz (1977) passed air throughsamples at 140°F (60°C). This
approachusing air is not regardedas suitable,since the extent of aging achieved
appeared to be quite low. However, a similar configuration,using oxygenunder low
pressures,may be appropriate,particularlywhen consideringcyclesof aging and
moisturetreatment,as discussedbelow.
b) Extended Oven Aging
Von Quintaset al. (1988) used the followingapproach:
1) heat six compacted specimensin an oven at 140°F (60°C) for 2
days, then remove three.
2) increasethe temperatureto 225°F (107°C) and age the remaining
three specimensfor 5 days.
75
High Temperature
==Field Aging
___- Low Temperature
Low Temperature withPressure Oxidation
Age
Figure 26. Possible Interactive Effects of Temperature and Pressure Oxidation on Aging
76
The specimenswere not supportedthroughoutthisprocessand no collapseof
specimenswas detected. Hugo and Kennedy (1985) heated specimensin an oven at
212°F (100°C) for 0, 4 and 7 days. Dry and moistatmosphereswere used. As
suggestedabove, these oven aging methodsare operating at highertemperatures
than desirable to achieve aging similarto field aging.
"l'iaet al. (1988) used forced draft and conventionaloven aging at 140°F (60°C).
The forceddra_.oven showedpromise for futuredevelopmentbut c_earlythe effect of
temperaturelevel on the integrityof the compactedsamples needs to be evaluated.
This author is concernedthat the internalstructureof a sample may be disruptedat
temperaturessuchas 140°C. In extreme cases the samplesmay "slump".
¢) Ultraviolet/Infrared Treatment
Hugo and Kennedy (1985) utilized"ultraviolet"treatment. Benson (1976) used a
level of "actinicradiation"apparently identicalto that used in the Californiastudies
describedby Kempand Predoehl(1981). Californiaweatheringtests have been
describedpreviouslyby Hveem et al. (1963) as infraredtreatment.
Clearly there is a need to define the differencebetween ultravioletand infrared
radiationand its likelyeffects. To this end, AppendixB has been prepared. This
shows that the wavelengthof light used by Benson(1976) and Kemp and Predoehl
(1981) is in the ultravioletrange. In their procedure,sampleswere subjectedto
actiniclight of 1000 MW/cm2 of 3660 Angstroms(i.e. 366 nanometers) for 18 hrs at
95°F (35°C).
Vallerga et al. (1957) utilizedboth infrared and ultravioletand found that more
aging occurredwith ultravioletlight. However,their explanationof the rationale for
usinginfrared is in keepingwith the informationin AppendixB, namely that infrared
causes molecularvibrationof materials. This resultsin a radiant heatingeffect. It is
most likely that aging of asphalt exposedto infrared is the same as if it was heated
77
directly (Petersen, 1989b). However,ultravioletlightis recognizedas havingpotential
to promote chemicalreactions,and in the case of asphalts,will do so, but only to a
depth of about 3 microns(Traxler, 1963).
Tia et al. showed significantaging of compactedsamples in an ultraviolet
chamber, and recommendedthat it shouldbe used togetherwith forced draft oven
aging, although they did not age sampleswith such a combination.
Ultravioletlightwill be considered further in this study, possibly in combination
with one or more of the other aging methods. There is potential for using an Atlas
Weatherometer or similardevice for this purpose.
d) Trlaxlal Cell Conditioning
A similar approachto that used by Kumar and Goetz (1977) but with
modificationshas great promise. It shouldbe possibleto "force" a fluid through a
mixture specimenwhile it is contained in a triaxial cell. Thus the cell would act as
the conditioningvessel,but would also enable measurements to be made on the
specimen, such as permeabilityand modulusdeterminations. This idea is described
more fully by Terrel (1989).
As noted in the Introduction,aged asphalt materialsmay be more moisture
susceptiblethan unaged materials. Since asphalt surfaces are subject to alternate
cycles of dry and wet conditionsin the field, it is importantto establish if cycles of
aging and moisturetreatments are more damaging than the additiveeffects of each
.. treatment done separately. A method of altemativelycycling oxidationwith moisture
treatment will be developed. Such a procedurecould be done quite easily in a
modified triaxial test cell. This configurationcould also be used to determinethe
modulusof the test specimen withoutits removal from the cell. The permeability can
also be measured in a triaxialcell.
78
3.2 Promising Test Methods
A procedureusing oxygenas the fluid will be comparedwith one using air. The
effects of temperature shouldalso be evaluated.
The LiteratureReview revealedthat a wide variety of test methods have been used to
evaluate the effects of aging of asphaltmixtures. These fall into two categories:
1) Tests on aged mixturesamples
2) Tests on asphalt recoveredfrom aged mixturesamples.
3.2.1 Tests on Aged Mixture Samples
Table 1 includesthe followingtest proceduresused to evaluate aging:
1) ResilientModulus,MR(e.g. Kemp and Predoehl, 1985; Kim et al., 1986;
Von Quin.taset al., 1988; and "ria et al., 1988).
2) Fatigue (e.g. Kim et aJ.,1986).
3) Creep Test (e.g. Kumar and Goetz, 1977).
4) IndirectTensile Test (strainat break) (e.g. Von Quintas et al., 1988; and,
Tia et al., 1988).
In addition,a dynamicor complexmodulushas promise, for reasonsoutlinedbelow.
a) Resilient Modulus (MR)
Kemp and Predoehl (1985), Kim et al. (1986), Von Quintas et al. (1988) and Tia
et al. (1988) used the diametralapproachto determineMR. Kim et al. and Tie et al.
used the approachdescribedin ASTM D-4123, and presumably this was also used by
Kempand Predoehl,and Von Quintaset al. The general trend observedby all these
researchersis an increaseof Mn as the age of samples increases. This test can be
performedquite quicklyand is nondestructive.
79
b) Fatigue
Kim et al. conductedfatigue tests on aged samples and noted that fatigue life
increasedas aging progressed. As notedin the previouschapter, this is not
surprisingsince the modulusof samplesincreased, and the applied stress was kept
the same. Fatigue tests are potentiallyvery time consumingand destroythe samples
used.
c) Creep
Kumar and Goetz (1977) conductedcreep tests at intermediatetime periodsas
their aging test progressed, in order to establisha creep curve at each incrementof
aging, a test periodof about 25 minuteswas necesary. This is longer than the time
required to conducta modulustest.
d) Indirect Tensile Test
This test was utilizedby Tia et al. (1988) and Von Quintuset al. (1988). Von
Quintas et ai. (1988) utilizedthis test and recommendedthe strain at break as the
parameter to evaluate the extent of aging rather than the tensilestrength. As a
sample ages and its modulusincreases,so does its strength,in the same way as its
fatigue life at a particularstress level does. However, the strain at break decreases,
since the sample becomesbrittleas it ages. Strain at break is therefore a good
representationof the deteriorationin flexibilityas a sample ages. Althoughthis test is
destructive,it is performedquickly.
e) Dynamic Modulus
The principlesof dynamicanalysis of asphaltsand asphalt-aggregatemixtures
were discussedby Goodrich(1988). Mamloukand Sarofim (1988) have also
describedmethodsto obtain the modulusof asphalt mixtures. Many other
researchersbefore them have described the concepts, for example, Coffman et al.
8O
I
(1964) and Sisko and Brunstrum(1968). Figure27 shows a typical plot of stress
versus strain from a repeated loadingtest capable of monitoringthe stress and strain
pulses.
The followingsummaryis based on the referencescited above. The complex
modulus(E*) resultingfrom sucha test is:
oosin (otE* = (3.1)
¢osin((ot-_)
where, o) = frequency of appliedstress or strain
$ = phase differencebetweenstress and strain
t = elapsed time
ao and G = maximumvalues of stress and strain pulses,
respectively.
The absolutevalue of the complexmodulusIE*I is commonlyreferred to as the
dynamicmodulus,and is found from:
ao (3.2)IE'I =
Other parametersthat may be obtained are the storagemodulus (E'), the loss
modulus(E'), and the losstangent. These are defined as follows:
E'= IE*Icos¢ (3.3)
E" = IE*I sin $ (3.4)
Losstangent = tan $ = E'/E' (3.5)
For a test on asphalt,the dynamicviscosity(In*l)may be expressed as:
I_*1 = IE*l/¢o (3.6)
81
E"_o"v : O"osl. I.t
, f _, J
q q MTM,i i *
Figure27. DynamicMechanicalAnalysis(afterCoffmanet al., 1964)
82
Goodrich(1988) notes that the loss tangent correlateswell with the performanceof an
asphalt in mixes. For this reason, dynamic"analysiswill be consideredin thisprojectas
having the promise for evaluatingthe aging effects in asphalt-aggregatemixtures.
None of the tests outlinedabove have showna strongcorrelationwith field data.
However,the resilientmodulus, indirecttensionand dynamicmodulustests are worthy of
use in subsequentstudies,and perhaps the creep test, since it is somewhat analogousto
viscositytests used for asphaltcements.
3.2.2 Tests on Recovered Asphalts
Table 1 shows that numeroustestshave been cloneon samples of asphalt recovered
from laboratoryand field aged mixtures. The most commonare:
1) Penetration
2) Softening Point
3) Viscosity
4) Ductility
5) Corbett-SwarbrickFractionation
6) Rostler-SternbergFractionation
No attempt will be made to describethese tests, since they are either currently
standardizedby ASTM, or have been in the past. A commonelement in all studiesis the
use of the Abson approachfor recoveringasphalt. Several studiesutilized the slidingplate
microviscometer,e.g. Hugo and Kennedy(1985), Kemp and Predoehl (1981), Benson
(1976), and Simpson et al. (1959).
Anothertest worthy of considerationis a cone and plate viscositydetermination,which
requiresa very small asphalt sample and is capable of operatingat higher temperatures
than a slidingplate device. Other testswill probablyemerge once the laboratory
investigationsare underway, and where appropriatethese will be investigated.
83
A hot centrifuge recovery procedure will also be considered. This latter method of
recovery avoids the use of solvent extraction, but does require heating of a mixture so that
small amounts of asphalt can be centrifuged from a mixture. The method has been
described by Cahill, Jamieson and Sheedy (1989), and was developed for use with surface
dressings. "
3.3 Evaluation of Aging Methods
To establish which of the aging methodsdescribed above are the most promising for
future study, a number of criteriahave been selected as follows:
1) simulationof fieldconditions
2) simplicityof use
3) cost
4) existingexperience
5) reliability
6) sensitivityto mix variables
7) other, i.e. any other relevantfactors.
No attempt has been made to weight the criteria; they are all important. However, it
is vital that the methods simulate field conditions and that they are viable for routine
laboratory use. Hence, the first three c,riteria are probably more important than the others.
Also, the first three ¢riteda are the only ones for which reasonable estimates can be made.
_ Tables 12(a) and 12(b) present an evaluation of the promising short-term and long-term
methods described earlier in this chapter. Of the three short-term aging methods, extended
heating and mixing of loose mixtures are the most promising. However, the use of micro-
wave heating should not be discounted until more detailed evaluation has been carded out.
Of the long-term methods, pressure oxidation and oven aging are the most promising and
84
Table 12. Evaluationof Aging Methods.
a) Short Term
Criterion Oven Heating ExtendedMixing MicrowaveTreatment
I. Simulation Good based on data Simulatesplant Not the sameof field from Von Quintas mixingconditions et el. (1988)
2. Simplicity Easy to use Easy to use Easy to useof use -no special equip-could use lab mixers
ment needs -or modifiedRTFOT
3. Cost of Moderate Moderate Moderate
equipment
4. Existing Very littlewith None Very littleexperience mixtures
5. Reliability Not established Not established Not establishedor accuracy -may requirea -would require
standard oven standardization ofequipment
6. Sensitivity Not established Not established Not establishedto mix variables
7. Other Analogouswith the Analogousto RTFOT May pro_w_testructuringTFOT
b) LongTerm
PressureOxygen Ultraviolet InteractionWithCriterion Treatment Oven Aging Treatment MoistureConditioningi
I. Simulation Preliminarytests Preliminarytests Difficultto assess True representationof field show that similar show that significant-in servicepavements of climatic cyclesconditions levels of aging aging can be achieved are subject to heat,
:are achieved -but at higher temp. light, and oxidationthan field
2. Simplicity Moderate Easy to use Moderate Difficultof use -needscareful -no specialequipment-coulduse a weather-
attention to safe needs ometer or lampshandlingofoxygen
3. Cost of Moderate to high Moderate Moderate to high Moderate to highequipment
A. Existing Very little Very little Littlewith mixtures Noneexperience
Reliability Questionable Questionable Not established Not establishedor accuracy -based on data -based on data fromfrom AAMAS study AAMSA study
_B.Sensitivity Preliminarytests Not established Not established Not establishedto mix indicatepromisingvariables performance
7. Other 6ood experienceof Analogouswith anseveralstudies extendedTFOT orwith asphalt indi-RTFOTcates potentialfor this method
85
will be thoroughly evaluated in a laboratorystudy. However, ultraviolettreatment and
interactionwith moistureconditioningwill be evaluated also to providefurther insight.
Of the aging methods identified,the literaturedoes not lead to a conclusionthat one
method is clearlysuperiorto others,or that an as yet untriedmethod is available. Indeed,
the literatureis very sparse with regard to laboratoryaging of asphalt-aggregatemixtures.
At the beginningof the literaturesearch, the author did not expect to find evidence that
ultravioletlightwould have a significanteffect. However, several researchers found a
significanteffect and one group ('ria, et al., 1988) stronglyrecommend further investigation
of such an approach in conjunctionwith a forced draft oven. Clearly, the effects of
ultravioletlight must be includedin a laboratorystudy of promisingmethods.
3.4 Evaluation of Test Methods
The criteriaused to evaluate test methodsare similar to those for aging methods, viz:
1) comparisonwith field data
2) simplicityof use
3) cost
4) existingexperience
5) reliability
6) sensitivityto mix variables
7) sample size
8) destructiveversus nondestructive
9) other, i.e. any other relevant factors.
Tables 13(a) and 13(b) presentan evaluationof the variousmethods for mixturesand
recoveredasphaltsdescribedearlier in this chapter. Of the tests on mixtures,the modulus
and indirect tensiletest should definitelybe included in a thorough laboratorystudy. The
dynamicmodulustest shows particularpromise,since it has the potentialto identify the
86
Table 13. Evaluation of Test Hethods
. a) Htxture Tests
Iteslltont Indirect TonalieCriterion I_dulus Test Oymtc lindulos Creep
11. Conpartson vlth Net eat_bltshed Not eatabllshnd Not establishnd Not establlehndfieldData
Z. $iapllcltyof Use _ estobltshede.dAn established Notstg.lflcantlyAn establlshedtest,tondordlzedtest test differentto -etralghtforverd-_retely dtfft--stratghtforverdlestltentIkxlulus buttakestloncult
3. Cost Of (qutponnt High Roderete _ltgh Moderate
4. Existing (xperlonca [jd_tstve Extensive Imdorato 6ond
S. Reliability or faries with Udum, Ysrlea with Unkno__uratcy* eqvlpmnt equipment
6. Sensitivity to Excellent Good Hat established GoodMix Yartables
7. Samle Size Varies with aodo Usuelly 4" die. Often 4" diameter Varies with modeof testing by 2-]lZ" high by 2-1/2" high. but of testing
)roferabiy higher
8. 9estructlve vs. Nondestrvcttvt Oestructive Ikmdostructive IlondrstructlveNondestruct Ivs
9. Other Represe_tattvs of Haspotential to /V_lo_us to are_esttng loeding establish the of- viSCosity testIn the field feet of the 4spha|t
I,
• NOTE: Only the resilient aodu|us test Is standsrdlznd by the ASTNas aethod 94123 (ATe. |gM). As of1988. no precision and bias statements haysbu_ established for this test..
b) Tests on RecoveredAsphalts
StandardCriterion ConsistencyTests buctl 1try Frnct I_t Io_ Ntcrovisceslty
J
1. Comparison 6ncd for viscosity ,_,.... Not yell estoblished 6ncdwtth Field Date andpenetration -c_fllcts in the
date | Iterature
Z. Simplicity Standard tests, Standardtest Standard tests Requires anof Use easy to do requiring experl- experienced technician
encedtechnician
3. Cost of Lov to anderato Hoderste Hnderote Itocllrst.i to highEquipment
(. Existing High Good GoodExperience
,S. Reliability High High Ooubtfulwith Inea- 9oubtful with Inexper-or Accuracy* perlencedtechnlctom tencadtechnician
6. Sensitivity IliA IU'A N/A IU'Ato Nix Verloble$
7. S_ple Size Requiresfstrly Aatct,_._-_--_Atlity Thesemethodsneed Needonly a m'111argose-,ples test can be used relya mull Lanple @le-.difficult to to Binlaize sonple 4.obtain size needed
8. Destructive vs I_A II/A I/A II/ANondestructlye
9. Other The I¢_ondurts & none andplata pro-avaliobte are slew endure Is • better
elternato to • slidingplata N_resch -- Itcan be usedover ewider rangeoftonperatures
• NOTE: The ASTNprovides precision statem_ts for rest of the tests felling Into the generic grtmpeOftests coveredby this teb|e.
87
effect of the asphaltin an asphalt-aggregatemixture.The creep test will not be usedsince
it is the least promising.
No attempthas been made to weigh the criteria. However, perhapsthe most
importantcriterionfor mixturetests is whetherthe test is nondestructive. For example, the
resilientmodulustest, being nondestructive,enables the modulusof a sample to be
obtainedthroughoutthe entire aging processto which a sample is subjected. Alternatively,
the indirecttensile test requires destructionof the sample, and therefore requires a large
numberof duplicatesamples in order to establishthe incrementaleffects of an aging
method. The first criterionis intendedto indicateif data from the test has shown any
correlationwith field aging. As with the aging methods,viabilityfor routine use is vital, and
therefore,criteria2 and 3 are also extremelyimportant. Reliabilityand accuracyare also
critical and a note is includedin the table regardingtest precisionwhere the test is
standardizedby ASTM. In the case of mixturetests, there is insufficientdata from previous
research to make any evaluationbased on this criterion.
For the tests on recoveredasphalt, the most importantcriterionis probably regarding
the specimen size. Tests requiringsmall specimens(say less than 5 g) are preferable,
since the effort requiredto recover asphalt is minimized. An overridingconcern with this
categoryof test is the effect of the recoveryprocess itself, and the po_ibility that the
compositionof the asphalt may change. Thenouxet al. (1988b) evaluated four alternatives
of recoveringasphalt,and determinedthe compositionof the recoveredasphalt by the
Corbett-Swarbrickapproach. It was concludedthat the compositionwas sensitiveto the
recoverymethod used. A recovery approachwhich will be evaluated in this study is a hot
cent_uge procedurereported by Cahillet al. (1989). This approachwill still change the
compositionof an asphalt because of the need for heat, and it must be establishedwhether
this change is more or less than with solventextractionand recovery.
88
4.0 GAPS IN THE KNOWLEDGE
The focus of the precedingchapters has been to describemethods of laboratory
aging of asphaltsand asphalt mixtures. Studiesrelatinglaboratorymethodsand field aging
were also described, and methodsof aging and testingmixturesin the laboratorywere
evaluated. Two major gaps in the knowledgeare particularlyevident:
1) There is little definitionof the relativecontributionof the various factors
that cause aging.
2) There is littledata relatingaging of laboratorymixturesto field aging.
These will be discussedfurther below.
4.1 Contribution of Various Factors to Aging
In Chapter 2, three major factors identifiedby Peterson (1984) that contributeto aging
were presented,viz:
1) loss of oilycomponentsby volatilizationor absorption
2) changes in compositionby reactionwithatmosphericoxygen
3) molecularstructuringthat producesthixotropiceffects (sterichardening)
In Chapters2 and 3, aging and test methodswere identifiedfor asphalts and asphalt
mixturesthat take into account each of these factors. The effects of volatilizationare
probablythe easiest to reproduce,and best defined. For example, Benson (1976)
developed a model for penetrationchange in the early life of a mixturewhich shows that a
typicaldrop in penetrationis of the order of 50'/=. It should also be noted that a study
conductedin Oregon (e.g. Lundand Wilson, 1984) suggeststhat a similar level of
hardeningis experiencedwhen expressed as a change in viscosity.
The extent of aging resultingfrom oxidationis not as clearlydefined as that resulting
from volatilization. This is hardlysurprisingsince it is a long-termeffectwith many more
89
variablescontributingto the observedperformancewith time. However, aging methods fall
into three main groups as follows:
1) Extendedheating
2) Pressureoxidation
3) Ultravioletlighttreatment
These are all exaggerationsof what a pavementexperiences in the field, and althougheach
type of test has been shownto producechanges in propertiessimilar to those occurringin
the field, it is not clear which is most appropriate,or what the contributionof each effect is
in field aging. In other words, if possible,it shouldbe clarified which approach, if any, most
closelyreproduces"oxidation'.
The steric hardeningphenomenonis probablythe most difficultto define and
reproduce. Any test which woulddetect a contributionto aging by this phenomenon must
be a mixturetest since recoveryof the asphaltwoulddestroy the associated molecular
structuring(Petersen, 1984). A repeated loadtest on mixtures,such as a resilientmodulus
test, may also destroythe structuring. Therefore, a rapid "one-shot"test would be suitable,
such as a creep test, or, the dynamic modulustest approach which tracks the changesin
several parametersas loadingprogresses.
4.2 Aging of Laboratory Mixtures versus Field Aging
Several studieshave showna good correlationbetween propertiesof laboratoryaged
asphaltand asphalt recoveredfrom the field,e.g. Lee (1973) and Benson (1976). However,
no such data is available for mixtures,other than to suggest that property levels of
laboratoryaged mixturesare similar to those in the field. Clearly, a requirementof this
study will be to establishthe needed correlations.
One cause for concem with laboratoryaging of mixturesto simulate long-termeffects
is that a compactedsample is requiredand its integritymay be affected by the elevated
temperatureused in the aging methods. It should be noted that of the three general
90
methodsof long-termaging, i.e. extended heating, pressureoxidation,and ultravioletlight
exposure,all known applicationsuse high temperature. In the case of the extended heating
approachrecommendedby Von Quintas et al. (1988), the secondstage of testing requires
a temperatureof 225°F (107°C) which could cause collapse of the sample unless it is
confinedin some way. The data presentedby Kim et al. (1986) also suggestsa problem
with maintainingthe integrityof a sample. This problem needs to be thoroughly
investigated.
91
5.0 CONCLUSIONS AND RECOMMENDATIONS
5.1 Conclusions
Based on the review of available literature,and to some extent, on engineering
judgment,the followingconclusionsare made.
1) Aging of asphalt mixturesoccursin essentiallytwo phases, short-termand
long-term. Short-term aging is primarilydue to volatilizationof the asphalt
cement duringconstruction,whereas long-termaging is due to oxidation
and stedc hardening in the field.
2) There is a wealth of data on laboratoryaging of asphalt cements, and
several studieshave developedrelationshipsbetween propertiesof
laboratoryaged asphalts and those recoveredfrom in service projects.
The studieson asphalt aging provideconsiderableguidance for
developmentof aging methods for asphalt mixtures.
3) The mostpromisingmethodsfor short-term aging of mixturesare
extended heatingand extendedmixing. Microwaveheating should also
be considered.
4) The most promisingmethodsfor long-termaging of mixturesare:
pressureoxidation,extended oven aging, ultravioletlighttreatment, and
alternate aging and moisturetreatment.
5) Test methodsfor evaluatingthe effects of aging fall into two groups: test
on mixtures,and, tests on recoveredasphalts.
6) Tests to evaluate mixturesmay be divided into nondestructiveand
destructivegroups. Nondestructivetests, in particularresilientmodulus
and dynamicmodulus,show promisefor future use. However, the indirect
tensiontest and the resultingtensile strain at break should also be
evaluated.
92
7) Tests requiringsmall specimensizes are preferable for tests on recovered
asphalt. Therefore, microductilityand microviscositytests have potential.
Fractionationtests shouldalso be. includedin subsequentstudies. A hot
centrifugeasphalt recoverymethodshouldbe investigatedas an alternate
to solventextractionand recovery.
5.2 Recommendations
A testingplan has been developed to enable evaluationof the promisingaging and
test methods. An outlineof this is presentedin AppendixC. The key elements of the plan
are given below.
1) The followingmethods of aging shouldbe investigated:
a) Short Term
• oven aging of loose mixtures
• extended mixingof loosemixtures
• microwaveheating
b) LongTerm
• pressure oxidation
• oven aging
• ultravioletlighttreatment
• cyclesof aging and moisture
2) The followingtests to evaluate laboratoryaging shouldbe investigated:
a) Tests on Mixtures
• ResilientModulus
• IndirectTensile
• Dynamic Modulus
93
b) Tests on RecoveredAsphalt
• Viscosityat 25°C and 60°C
• Penetrationat 25°C
• Ductility
• Corbett-Swarbrickand/or Rostler-Stembergfractionation
3) Field data need to be obtainedto demonstratethe validity of the aging
methods and to calibratethem. It is recommendedthat on-goingfield
studies,such as those being carried out for the AAMAS projectare used
for the purpose. The study team needs to establishothersources of field
data. in all cases, a supply of the originalmaterialsused is required,and
data for mixtureand asphaltpropertiesrecoveredfrom the pavement at
regular time intervalsduring and after construction.
94
6.0 REFERENCES
ASTM (1979), "1979 Annual Book of AmericanSocietyfor Testing Material Standards," Part35, D2565-76, "OperatingXenon Arc-Type(Watercooled)Light -and Water ExposureApparatusfor Exposureof Plastics,"724-727.
ASTM (1988), "1988 Annual Book of ASTM Standards,"Vol. 04.03, Road and PavingMaterials,Traveled Surface Characteristics,American Society for Testing Mateirals,Philadelphia.
Abson, G. (1933), "Apparatusfor the Recoveryof Asphalts,"Proceedings, American Societyfor Testing Materials, Part II, Vol. 33.
AI-Ohaly, A.A. and R.L. Ten'el (1988), "Effect of MicrowaveHeating on AdhesionandMoistureDamage of Asphalt Mixtures,"TransportationResearch Record 1171,TransportationResearch Board,27-36.
Anderson, D.A. and E.L. Dukatz (1980), "AsphaltProperties and Composition: 1950-1980,"Proceedings, Association of Asphalt Paving Technologists, Vol. 49, 1-29.
Anderson, D.A., E.L. Dukatz, and J.L. Rosenberger(1983), "Propertiesof Asphalt Cementand AsphalticConcrete,"Proceedings, Association of Asphalt Paving Technologists,Vol. 52, 291-324.
Barth, E.J. (1962), "Asphalt",Gordon& Breach, New York.
Benson, P.E. (1976), "LowTemperature TransverseCrackingof AsphaltConcretePavements in Central and West Texas," Texas TransportationInstitute& StateDepartment of Highwaysand PublicTransportation,CooperativeResearch Project,Report 175-2F.
Brown, A.B., J.W. Sparks, and F.M. Smith (1957), "Stedc Hardeningof Asphalts,"Proceedings, Association of Asphalt Paving Technologists, Vol. 26, 486-494.
BrOl6,B., G. Ramond, and C. Such (1986), "RelationshipsBetween Composition, Structure,and Propertiesof Road Asphalts: State of Researchat the French Public WorksCentral Laboratory,"TransportationResearchRectord 1096, Transportation ResearchBoard, 22-34.
Button,J.W., D.N. Little,B.M. Gallaway, and J.A. Epps (1983), "Influenceof AsphaltTemperature Susceptibilityon PavementConstrucutionand Performance," NationalCooperativeHighway Research ProgramReport 268, Washington,D.C.
Cahill, M.C., I.L. Jamieson and J.P.M. Sheerly (1989), "Surface DressingFailures: AReview of Studiesin Ireland," Acceptedfor the 1989 EurobitumeSymposium,Madrid, Spain.
Coffman, B.S., D.C. Craft, and J. Tamayo (1964), "Comparisonof Calculated and MeasuredDefelectionsfor the AASHO Road Test," Proceedings, Association of Asphalt PavingTechnologists, VoI. 33, 54-91.
95
Coons, R.F. and P.H. Wright (1968), "An Investigationof the Hardening of AsphaltRecovered from Pavements of Various Ages," Proceedings, Association of AsphaltPaving Technologists, Vol. 37, 510-528.
Corbett, L.W. and H.E. Schweyer (1981), "Composition and Rheology Considerations in AgeHardeningof Bitumen," Proceedings, Association of Asphalt Paving Technologists,Vol. 50, 571-582.
Corbett, L.W. and R.E. Merz (1975), "AsphaltBinder Hardening in the MichiganTest Road,After 18 Years of Service," TransportationResearch Record 554, TransportationResearch Board,27-34.
Croney, D. (1977), "The Designand Performance of Road Pavements," Her Majesty'sStationeryOffice, London,England.
Davis, T.C. and J.C. Petersen (1967), "An Inverse GLC Study of Asphalts Used in theZaca-Wigmore ExperimentalTest Road, Proceedings, Association of Asphalt PavingTechnologists, Vol. 36, 1-10.
Dickinson,E.J. (1980), "The Hardeningof Middle East Petroleum Asphalts in PavementSurfacings,"Proceedings, Association of Asphalt Paving Technologists, Vol. 49, 30-62.
Dow, A.W. (1903), "Asphalt Experimentsat Washington,"Engineering Record, Vol. 47, No.18.
Edler, A.C. et al. (1985), "Use of Aging Tests to Determinethe Efficacyof Hydrated LimeAdditionsto Asphalt in RetardingIts OxidativeHardening," Proceedings, Associationof Asphalt Paving Technologists, Vol. 54, 118-139.
Gaestel, C., R. Smadja andK.A. Lamminan (1971), "Contribution._ la Connaisance desPropd_tdsdes BitumesRoutiers,"Rev. G(_ndraledes Routes et Adrodromes,466,85-94.
Goode, J.F. and L.A. Lufsey, "Voids, Permeability,FilmThicknessvs. Asphalt Hardening,"Proceedings, Association of Asphalt Paving Technologists, Vol. 35, 430-463.
Goodrich,J.L. (1985), "StudiesRelating Asphalt Compositionto Road Performance,"Chevron,USA, Asphalt Division.
Goodrich,J.L., J.E. Goodrich,and W.J. Karl (1986), "Asphalt CompositionTests: Their_ Applicationand Relation to Field Performance," Transportation Research Record
1096, TRB, Washington, D.C., 146-167.
Goodrich, J.L (1988), "Asphalt and Polymer Modified Asphalt Properties Related to thePerformance of Asphalt Concrete Mixtures," Proceedings, Association of AsphaltPaving Technologists, Vol. 57, 116-175.
Gotolski, W.H., J.M. Lucas, S.K. Ciesielski, and J.A. Kofalt (1965), "A Study of PhysicalFactors Affecting the Durability of Asphaltic Pavement," Interim Report No. 3, ThePennsylvania State University.
Griffin, R.L., T.K. Miles, and C.J. Penther (1955), "Microfilm Durability Test for Asphalt,"Proceedings, Association of Asphalt Paving Technologists, VoL 24, 31-62.
96
Halstead, W.J. and J.A. Zenewitz (1961), "Changesin Asphalt ViscositiesDuring Thin-FilmOven and MicrofilmDurabilityTests," Public Roads, Vol. 31, No. 11, 211-218.
Halstead,W.J. (1985), "Relationof Asphalt Chemistryto PhysicalPropertiesandSpecifications,"Proceedings, Association of Asphalt Paving Technologists, Vol. 54,91-117.
Heithaus,J.J. and R.W. Johnson(1958), "A MicroviscometerStudy of Road AsphaltHardeningin the Field and Laboratory,"Proceedings, Association of Asphalt PavingTechnologists, Vol. 27, 17-34.
Her Majesty's Stationery Office (1962), "BituminousMaterials in Road Construction,"RoadResearch Laboratory,London,204-205.
Hubbard, P. and H. Gollomb(1937), "The Hardeningof Asphaltwith Relation toDevelopmentof Cracks in Asphalt Pavements,"Proceedings, Association of AsphaltPaving Technologists, Vol. 9, 165-194.
Hugo, F. and T.W. Kennedy (1985)i "SurfaceCrackingof Asphalt Mixturesin SouthernAfrica,"Proceedings, Association of Asphalt Paving Technologists, Vol. 54, 454-501.
Hveem, F.N., E. Zube, and J. Skog (1963), "ProposedNew Tests and SpecificationsforPaving Grade Asphalts," Proceedings, Association of Asphalt Paving Technologists,Vol. 32, 247-327.
Jamieson, I.L. and M.M. Hattingh,"The Correlationof Chemical and Physical PropertiesofBitumenswith Their Road Performance,"Proceedings, Australian Road ResearchBoard, Vol. 5, Part 5.
Jennings,W. (1985), "Predictionof Asphalt Performanceby HP-GPC," Proceedings,Association of Asphalt Paving Technologists, Vol. 54, 635-646.
Jennings,W., J.A.S. Pribanic,J. Smith,and T. Mendes (1988), "Predictingthe Performanceof Montana Test Sectionsby Physicaland ChemicalTesting,"TransportationResearch Record 1171, TransportationResearchBoard, 59-65.
Kemp, G.R. (1973), "AsphaltdurabilityTests and Their Relationshipto Field Hardening,"American Societyfor Testing Materials,Special TechnicalPublication532, 79-99.
Kemp, G.R. and N.H. Predoehl (1981), "A Comparisonof Field and LaboratoryEnvironmentson Asphalt Durability,"Proceedings, Association of Asphalt PavingTechnologists, Vol. 50, 492-537.
Kim, O-K., C.A. Bell, J. Wilson, and G. Boyle(1986), "Effect of Moistureand Aging onAsphalt Pavement Life, Part 2 - Effectof Aging,"FHWA-OR-RD-86-01-2, FinalReport to Oregon Departmentof Transportationand the Federal HighwayAdminisb'ation.
Krchma,L.C. and D.W. Gagle (1974), "A USA Historyof Asphalt Refined from Crude Oiland Its Distdb_Jtion,"Proceedings, Association of Asphalt Paving Techrtologists 50thAnniversary Historical Review, Vol. 43A, 26-88.
97
Kumar, A. and W.H. Goetz (1977), "Asphalt Hardening as Affected by Film thickness, Voids,and Permeability in Asphaltic Mixtures," Proceedings, Association of Asphalt PavingTechnologists, Vol. 46, 571-605.
Lang, F.C. and T.W. Thomas (1939), "LaboratoryStudies of Asphalt Cements," University ofMinnesotaEngineeringExperimentStation, Bull.55, Vol. XLII.
Lee, D.Y. (1973), "Asphalt DurabilityCorrelationin Iowa,"Highway Research Board, Record468, 43-60.
Lee, D.Y. (1968), "Developmentof a Laboratory.DurabilityTest for Asphalts," HighwayResearch Board, Record 231, 34-49.
Lewis,R.H. and J.Y. Welbom (1940), "Report on the Propertiesof the Residues of 50-60and 85-100 PenetrationAsphaltsfrom Oven Tests and Exposure,"Proceedings,Association of Asphalt Paving Technologists, Vol. 11, 86-157.
Lewis, R.H. and W.J. Halstead (1946), "Behavior of Asphalts in Thin Film Oven Test,"Public Roads, Vol. 24, No. 8, 220-226.
Lund, J.W. and J.W. Wilson (1984), "Evaluationof Asphalt Aging in Hot Mix Plants,"Proceedings, Association of Asphalt Paving Technologists, Vol. 53, 1-18.
Mamlouk, M.S. and R.S. Sarofim(1988), "Modulusof Asphalt Mixtures-An UnresolvedDilemma,"TransportationResearch Record 1171, TransportationResearch Board,193-198.
McHattie, R.L (1983), "Estimatingthe Durability of Chem-Crete Modified Paving Asphalt,"Alaska Departmentof Transportation.
Nicholson,V. (1937), "A LaboratoryOxidationTest for Asphaltic Bitumens," Proceedings,Association of Asphalt Paving Technologists, Vol. 9, 208-214.
Pauls, J.T. and J.Y. Welbom (1952), "Studiesof the Hardening Propertiesof AsphalticMaterials,"Proceedings, Association of Asphalt Paving Technologists, Vol. 21, 48-75.
Petersen, J.C., E.K. Ensley,H. Rancher, and W.E. Haines (1976), "PavingAsphalts:Asphalt-AggregateInteractionsand Asphalt IntermolecularInteractions,"FHWA-RD-77-25, Federal HighwayAdministration,InterimReport.
Petersen,J.C. (1984), "ChemicalCompositionof Asphalt as Related to Asphalt Durability:State of the Art,"TransportationResearch Record 999, TransportationResearchBoard, 13-30.
Petersen,J.C. (1989a), "A Thin-Rim AcceleratedAging Test for EvaluatingAsphaltOxidativeAging," Proceedings, Association of Asphalt Paving Technologists, PrepdntVolume.
Petersen,J.C. (1989b), PersonalCommunication,May 1989.
Plancher, H., E.L Green, and J.C. Peterson(1976), "Reduction of Oxidative Hardening ofAsphaltsby Treatmentwith HydratedUme-A MechanisticStudy," Proceedings,Association of Asphalt Paving Technologists, Vol. 45, 1-24.
98
Raschig, F.L. and P.C. Doyle (1937), "A LaboratoryOxidationTest," Proceedings,Association of Asphalt Paving Technologists, Vol. 9, 215-217.
Rostler,F.S. and R.M. White (1962), "Compositionand Changes in Compositionof HighwayAsphalts,85-100 PenetrationGrade," Proceedings, Association of Asphalt PavingTechnologists, Vol. 31, 35-89.
Santucci,LE., J.E. Goodrichand J.E. Sundberg(1981), "The Effect of Crude Source andAdditiveson the LongTerm Oven Agingof Paving Asphalts,"Proceedings,Association of Asphalt Paving Technologists, Vol. 50, 560-571.
Schmidt,R.J. (1973), "LaboratoryMeasurementof the Durabilityof Paving Asphalts,"AmericanSociety for Testing Materials,SpecialTechnical Publication532, 79-97.
Schmidt,R.J. and L.E. Santucci(1969), "The Effectof Asphalt Propertieson the FatigueCrackingof Asphalt Concreteon the Zaca-WigmoreTest Project,"Proceedings,Association of Asphalt Paving Technologists, Vol. 38, 39-64.
Shattuck,C.L. (1940), "Measurementof the Resistanceof Oil Asphalts (50-60 Pen) toChanges in Penetrationand Ductilityat Plant MixingTemperatures," Proceedings,Association of Asphalt Paving Technologists, Vol. 11, 186-203.
Simpson,W.C., R.L. Griffin,and T.K. Miles (1959), "Correlationof the Micro-FilmDurabilityTest with the Field HardeningObserved on the Zaca-Wigmore ExperimentalRoadProject,"ASTM STP 277.
Sisko,A.W. and L.C. Brunstrum(1968), "The RheologicalPropertiesof Asphalts in Relationto Durabilityand Pavement Performance,"Proceedings, Association of AsphaltPaving Technologists, Vol. 37, 448-475.
StandardsAssociationof Australia (1980), "Methods of Testing Bitumenand RelatedRoadmaking Products,"AS 2341.
Terrel, R.L and J.W. Shute (1989), "Water Sensitivity"Summary Report for the StrategicHighway Research Program,SR-OSU-A-003A-89-3.
Thenoux,G., C.A. Bell, J.E. Wilson, D. Eakin, and M. Schroeder (1988a), "EvaluationofAsphalt Propertiesand Their Relationto Pavement Performance,"Final Report toOregon Departmentof Transportationand U.S. Department of Transportation,Federal HighwayAdministration,FHWA-OR-RD-88-02-1.
Thenoux,G., C.A. Bell, and J.E. Wilson(1988b), "Evaluation of Physical and FractionalPropertiesof Asphaltand Their Interrelationship,"TransportationResearch Record1171, TransportationResearch Board, 82-97.
Tia, M, B.E. Ruth, C.T. Charai, J.M. Shiau, D Richardson,and J. Williams (1988),"Investigationof Originaland In-ServiceAsphalt Propertiesfor the DevelopmentofImprovedSpecifications- Final Phase of Testingand Analysis," Final Report,Engineeringand IndustrialExperimentStation, Universityof Florida, Gainesville, FL.
Traxler, R.N. (1961), "RelationBetween AsphaltCompositionand Hardeningby Volatilizationand Oxidation,"Proceedings, Association of Asphalt Paving Technologists, VoL 30,359-377.
99
Traxler, R.N. (1963), "Durability of Asphalt Cements," Proceedings, Association of AsphaltPaving Technologists, Vol. 32, 44-58.
Tuffour, Y.A., I. Ishai, and J. Craus (1989), "Relating Asphalt Aging and Durability to ItsComposition Changes," Proceedings, Association of Asphalt Paving Technologists,Preprint Volume.
Vallerga, B.A., F.N. Finn, and R.G. Hicks (1967), "Effect of Asphalt Aging on the FatiguePropertiesof AsphaltConcrete,"Proceedings, Second International Conference onthe Structural Design of Asphalt Pavements, Ann Arbor, Michigan,595-617.
Vallerga, B.A. and W.J. Halstead (1971), "Effectsof FieldAging on FundamentalPropertiesof Paving Asphalts," Highway ResearchRecord 361, Highway Research Board,Washington,D.C., 71-92.
Vallerga, B.A. (1981), "Pavement DeficienciesRelated to Asphalt Durability,"Proceedings,Association of Asphalt Paving Technologists, Vol. 50, 481-491.
Vallerga, B.A., C.L Monismith,and K. Granthem (1957), "A Study of Some FactorsInfluencingthe Weatheringof Paving Asphalts," Proceedings, Association of AsphaltPaving Technologists, Vol. 26, 126-150.
Von Quint.as,H., J. Scherocman,T. Kennedy,and C.S. Hughes (1988), "Asphalt AggregateMixture AnalysisSystem," Final Report to NCHRP.
Welbom, J.Y. (1979), "Relationshipof Asphalt Cement Propertiesto Pavement Durability,"National CooperativeHighway Research Program, Synthesis59, Washington,D.C.
Zube, E. and J. Skog (1969), "Final Report on the Zaca-Wigmore Asphalt Test Road,"Proceedings, Association of Asphalt Paving Technologists, Vol. 38, 1-38.
100
APPENDIX A
AGING PROCEDURE USING PRESSURE OXIDATION
The modified pressure oxidationvessel,which was developed originally in England
(HMSO, 1962), consists of a cylindrical pressure vessel fitted with end plates. The upper
plate contains a safety rupture disc and a pressure gauge. The lower plate contains a plug
valve that connects to the oxygen supply. Figure A1 shows a diagram of the pressure
oxidation vessel (POV) used in the study by Kim et al. (1986).
The following are the main steps in the use of the POV:
1) Samples (asphalt mixtures or asphalt cement) are prepared.
2) The samples are placed in a POV.
3) A vacuum [26 in. (66 cm) Hg] is applied for 20 minutes.
4) The POV is filled via the valve from an oxygen cylinder to 300 psi
(2070 kPa). This pressure is held for 30 minutes to ensure leak-free
joints.
5) The oxygen is then disconnectedand the POV is then placed in an
oven maintainedat 77°F (25°C) or 140°F (60°C) for a periodof time
such as 1, 2, 3, and 5 days.
6) At the conclusionof the test the valve is opened,the cover is
removed,and the aged mixturesare cooled to room temperatureprior
to storageand subsequenttesting.
A-1
ASME Pressure Relief
Pressure GaugeB 8
Top Plate
I
Stainless SteelI CylinderII
, Steel Rod
' O-Ring Seal
I
Bottom Plate
- Pressure Port
FigureA1. PressureOxidationVessel (POV)
A-2
APPENDIX B
A BRIEF SUMMARY OF THE ELECTROMAGNETIC SPECTRUM
COMPONENTS: ULTRAVIOLET AND INFRARED RADIATION
Ultraviolet(UV) and infrared radiationare two componentsof the electromagnetic
spectrum. Referringto Figure B1 visiblelight lies between these two regions. This
appendixwill briefly summarizeeach of these types of radiation.
Ultravioletradiationwas discoveredby JohannWilhelm Ritter in 1801, while
investigatingthe effects of lighton chemicalsubstances. From this experiment, he
hypothesizedthat energy existed beyondthe violet region of visible lightspectrum(see
FigureB2). The wavelengthof UV radiationis longer than x-rays but shorterthan visible
light. Visible light and UV light are divided into regions. Ultraviolet radiationhas three
regions. They are:
1) Near or black light- 4000A-3000A (where 4000/i, is the boundary)
2) Far - 3000A-2000A
3) Vacuum - below 2000A
where 1/k = 1 Angstrom and 1A = 0.10 nanometer (lxlQ "° meters).
The chief source of UV radiation is the sun. Most of the UV generated is from the
"black light" category, while less than 14% has a wavelength of less than 3000A. More
... than 50% of UV radiation is absorbed by the atmosphere, the majority of which is the
shorter wavelength radiation. UV radiation follows the laws of reflection and refraction. It is
absorbed by substances which are transparent to visible light and transmitted through
quartz, fluorite and distilled water (Koller_1989). Ozone in the atmosphere is made by the
absorption of the shorter UV wavelength by oxygen. Aluminum is used as a reflecting
B-1
o0
_mo-'1-"0
f,n...I .T.I-
•- (.0
o m "_m
2-_. >-
I- cO 0 m"-- Z o.._ .-_ ,,, oor'r o Z) ._o-- 0 "S
ooILl
I--
.__
_0
B-2
Frequency (hz) Wavelength ( Z )
1 018- 0.3 nm
101_7 VACUUMUV 2,,,nm3 nm (7.5 x 1014hz)//400
16 /10 30 nm t -- violet/
/..... s - Blue/
1015 FARUV 300 nm s 500 •NEAR UV..... t _ -- GreenVISIBLE UGHT
10 TM3000 nm " - Yellow
NEARIR _ 600-- Orange
1013 0.03 mm
- Red1012
FARIR 0.3 mm • 700(4.28 1014hz)x
1011 , 3 mm
FigureB2. SpectralRegions(afterKoller, 1989)
B-3
device due to its high reflectivity and resistance to tarnishing. Other reflecting materials
include sand, snow and water.
UV radiation "promotes chemical reaction", which is the study of photochemistry. An
example of this would be the bleaching and folding of dyes due to the exposure of the sun.
Its biological effects include suntans and sunburns. This radiation type is a source of
vitamin D, which aids in healing wounds in children. Another unique characteristic is its
ability to kill bacteria; therefore, UV lamps are used to sterilize air in hospitals, etc. UV
radiation is also an efficient method in making fluorescent light.
Infrared radiation was discovered by Sir William Herschel (1738-1822) in 1800. With
the aid of a thermometer, he realized that radiant heat was greatest outside the "red" region
of the visible spectrum (see Figure B2). Sources of infrared radiation are incandescent
solids or electrical discharges in gases. They are detected by the use of photographic
plates (Andrew, 1988). Infrared radiation is composed of frequencies (a range of) which
make molecules of substance vibrate internally. Because it is unique (like a set of
fingerprints), this can be used to identify a compound. Also, similar to UV radiation, certain
portions of the spectra are absorbed in the atmosphere. The study of infrared radiation is
important because it is a method in probing atomic and molecular structure of solids.
B--4
REFERENCES
Andrew, K.L. (1988), "Spectra," Colliers Encyclopedia, Vol. 21,416-419.
Brady and Holum (1981), Fundamentals of Chemistry, John Wiley & Sons, 188-191.
Killer, L.R. (1989), "Ultraviolet Radiation," Encyclopedia Americana, Vol. 27, 353c-353d.
B-5
APPENDIX C
OUTLINE OF TEST PROGRAM
AGING OF ASPHALT-AGGREGATE SYSTEMS
The following is a summary of a laboratory test program that has been presented in
a separate document for the SHRP project A-003A (see Hicks et al., 1989). It should be
noted that at the time that this report is in preparation (October 1989), a subcontract is
being negotiated between Oregon State University and the Research Triangle Institute (RTI),
in North Carolina. It is intended that RTI conduct studies regarding the role of ultraviolet
light on aging and also some fundamental studies regarding steri¢ and oxidative aging.
1. Purpose
To evaluate the most promising aging method(s) which simulate short-term and long-
term aging effects.
2. Test Program
The program to be undertaken at OregonState Universitywill be done in three
phasesas follows:
a. PreliminaryTest Program
b. ExpandedTest Program
c. Field Validation
Only the preliminary programwill be outlinedhere. This programwill evaluate the aging
methods presented as the most promisingin the body of this report. The evaluationwill be
done with a limited number of materialand test variables. The expanded test programand
field validationphases will considermore variables.
The preliminary programwill be done in two groups of tests as follows:
C-1
a. Conventional Procedures
b. Modified Triaxial Procedures
Each of these is described below.
3. Conventional Procedures
Aging Methods. Based on the literature review, the following "conventional" methods
will be evaluated:
Short-Term Lon.q-Term
• Oven Aging " • Pressure Oxidation
• MicrowaveAging • Oven Aging
• ExtendedMixing
EvaluationMethods. The tests proposed at this time are as follows:
Short-Term Long-Term
• ResilientModulus • Microviscosity
• DynamicModulus • Microductility
• Tensile Test • Fractionation
Other tests may be used, such as infrared spectrometryand size exclusion
chromatographyon recoveredasphalt.
4. Variables Considered for Conventional Aging
The same variableswill be used for each of the five aging methods (short- and long-
term). These are:
C-2
2 asphalts
2 aggregates
2 air void levels
2 test temperatures
3 time periods
All mixtureswill be prepared using the mix design asphaltcontentand gradations,and
standard compactionproceduresfor the Californiakneadingcompactor.
A 3/4 fractionof the completefactorialwith no replicatetestswill providethe
preliminaryinformationrequired. The followingtable presentsthe combinationsof the
variablesto be used for each method:
Air Voids Air VoidsLow High
I Temperature Temperature
Asphalt (AC) Level 1 Level 2 Level i Level 2and
Aggregate (AC) Time Period Time PeriodCo_inations
aSc aSc abe abc
ACI + Agg 1 X X X X X X X X X
AC2 + Agg 1 X X X X X X X X X
ACI + Agg 2 X X X X X X X X X
AC2 + Agg 2 X X X X X X X X X
For each aging method,36 specimenswill be preparedand tested accordingto the
combinationsof variables shown in the table.
5. Modified Trlaxlal Cell
This approachwill try an untestedprocedure,consistingof forcing fluids to flow
through a mixturespecimen. At this time it is not knownwhat pressure level will be
needed. The main tests to be done will involveoxygenand air as two additionalvariables
to those listedabove. A 1/2 fractionof the complete factorialwith no replicatetests will
C-3
provide the preliminary information required. The following table presents the combinations
of the variables to be used:
ATMOSPNERE
OXYGEN AIR
Air Voids Air Voids Air Voids Air Voids
Low Medium Low Medium
Temp Level Temp Level Temp Level Temp Level
A,ph,,Acl 112 I 1 2 1 2 112and
Aggregate (AC) Time Period Time Period Time Period Time PeriodCombinations
a b c abc abc abc ab c a b c a b c a b c
AC 1 + Agg 1 X X X X X X X X X X X X
AC 2 + Agg 1 X X X X X X X X X X X X
AC 1 2 X X X X X X X X X X X X
AC 2 + Agq 2 X X X X X X X X X X X X
In this case, 48 specimens will be required. A secondary series of tests will be done to
cycle conditioning of specimens with both water and oxygen or air. Only one combination
of asphalt and aggregate will be used. The experiment design has not yet been developed
for these tests.
C-4