l07-082

Mechanical evaluation of asphaltaggregatemixtures prepared with fly ash as a fillerreplacement

Serkan Tapkn

Abstract: The objective of this study is to investigate the effect of fly ash as a filler replacement on the mechanical prop-erties of asphaltaggregate mixtures. Utilization of fly ash, which is the by-product of coal-fired power generation, is ofgreat importance from an environmental and economical point of view. In this study, a dense bituminous mixture com-posed of calcareous aggregate was selected as the reference mixture. It was observed that there was a definite increase inMarshall stability and decrease in flow values, especially when calcareous filler was replaced by Soma-type fly ash, whichwas one of the three types of fly ashes used. The mechanical properties, namely elastic strain, elastic modulus, and perma-nent strain, of the asphalt mixtures were determined by carrying out fatigue tests with a UMATTA tester for three types offly ashes, portland cement, lime, and control specimens. The changes in mechanical properties are important in the sensethat they affect the behavior of asphalt concrete pavement under applied loads. This mechanism can be explained basicallyby bitumen extension. The fatigue life of fly ash specimens, especially Soma fly ash, was found to be considerably higherthan that of calcareous filler specimens. Based on this study, it is demonstrated that fly ash can be used effectively in adense-graded wearing course as a filler replacement.

Key words: fly ash, filler replacement, Marshall method, mechanical properties, fatigue test, performance, bitumen exten-sion, fatigue life.

Resume : Lobjectif de cette etude etait dexaminer limpact des cendres volantes comme matie`re de charge de remplace-ment sur les proprietes mecaniques des melanges asphalte-granulats. Lutilisation des cendres volantes, le sous-produit descentrales thermiques alimentees au charbon, est dune grande importance environnementale et economique. Dans la pre-sente etude, un melange bitumineux dense, compose dagregat calcaire, a ete choisi comme melange de reference. La stabi-lite Marshall a definitivement augmentee et une diminution des valeurs decrasement a ete remarquee, surtout lorsque lamatie`re de charge calcaire etait remplace par les cendres volantes de type Soma, lun des trois types de cendres volantesutilises. Les proprietes mecaniques, dont la deformation elastique, le module elastique et la deformation plastique des me-langes asphaltiques ont ete determinees en effectuant des essais de resistance a` la fatigue par un UMATTA pour trois typesde cendres volantes, du ciment Portland, de la chaux et des echantillons temoins. Les changements aux proprietes mecani-ques sont importants en ce quils ont un impact sur le comportement du revetement en beton asphaltique sous les chargesqui y sont appliquees. Ce mecanisme peut etre explique a` la base par lextension du bitume. La longevite a` la fatigue desechantillons de cendres volantes, surtout la cendre de type Soma, sest averee considerablement plus elevee que celle desechantillons comportant des matie`res de charge calcaires. La presente etude a demontre que les cendres volantes peuventetre utilisees de manie`re efficace comme matie`re de charge dans une couche dusure de categorie dense.

Mots-cles : cendres volantes, matie`re de charge de remplacement, methode Marshall, proprietes mecaniques, essai de fati-gue, rendement, extension du bitume, longevite a` la fatigue.

[Traduit par la Redaction]

1. Introduction

With development of infrastructure and increasing trans-portation demands, construction of better pavements withlonger service lives is required. The development of new

modified paving materials and use of them in constructionresults in high performance pavements to meet the needs ofthe communities. While developing these new modifiedpaving materials, attention should be paid to using wasteindustrial materials effectively in construction to addressenvironmental and economic concerns.

In the last few decades, research findings have shown thatthe use of new paving materials requires more sophisticatedapproaches for pavement and material mix design processes.Pavement design and mix design should be handled togetherand performance evaluation techniques should accompanythe classical mix design methods.

The use of modified asphaltaggregate mixtures in surface-bound layers creates a demand for evaluation of the per-formance of those modified mixtures. Asphalt modification

Received 22 January 2007. Revision accepted 5 June 2007.Published on the NRC Research Press Web site at cjce.nrc.ca on30 January 2008.

S. Tapkn. Civil Engineering Department, Faculty ofEngineering and Architecture, Iki Eylul Campus, AnadoluUniversity, 26555 Eskis ehir,Turkey (e-mail: [email protected]).Written discussion of this article is welcomed and will bereceived by the Editor until 31 May 2008.

27

Can. J. Civ. Eng. 35: 2740 (2008) doi:10.1139/L07-082 # 2008 NRC Canada

is accomplished by using polymer modifiers in the asphaltaggregate mixtures. This modification can be done both bymodifying bitumen with the polymer or by adding themodifier directly in to the mixture of bitumen and aggre-gate during mixing process. The first technique needs spe-cialized equipment and therefore brings additionalexpenditure. Asphalt modification by using polymers,whether bitumen-based or mix-based, is an expensive proc-ess and needs skilled workmanship and special equipment.Also, for developing countries, asphalt modification withpolymers means more dependence on developed countriesfor the transfer of technology, know-how, and, most impor-tantly, importation of patented modifiers. The importanceof using waste materials, such as fly ash, in the modifica-tion of asphaltaggregate mixtures comes into the scene atthis point. Together with the economic considerations, us-ing fly ash in asphaltaggregate mixtures alters the mixturebehavior in a beneficial way (Ali et al. 1996; Churchilland Amirkhanian 1999; Asi and Assaad 2005).

The oldest application of asphaltaggregate mix modifica-tion is the replacement of normal filler material by anothersuitable material available in nature or easily found on themarket. This process is called filler replacement. In practice,filler replacement is put into application for cases wherenormal filler produced during the crushing process of rocksfrom quarries is scarce or aggregate is not clean and inher-ent mud or clay constituents are present; thus, it becomesnecessary to wash out the aggregate, which causes the totalloss of filler.

There are several materials available on the market and innature that may be used in place of normal filler. Fly ash, aby-product of coal-fired power generation, is an example ofsuch a material. Because of its origin, there is a rising inter-est in fly ash among scientists and civil engineers (Sobolevand Naik 2005). The importance of fly ash as a waste mate-rial, from the economic point of view, draws major attentionto this material by pavement engineers. The vast amount ofwaste material produced daily is one of the major worldwideproblems in waste management. Developing countries likeTurkey produce a considerable amount of fly ash in coal-fired power plants every year and this huge amount of wastecreates a significant problem with respect to handling andstorage, which are important both from the economic andenvironmental point of view. Nations are forced to searchfor more suitable ways to recycle these waste materials, themain reason being primary environmental concerns and adecreasing number of landfill sites. For a considerableamount of time, researchers have been investigating the useof fly ash in the construction industry, to enable better man-agement of this important waste material and to improve theproperties of construction materials. Fly ash has been usedextensively in concrete production for many years, but thereare limited applications where fly ash has been used inpavement engineering when compared with concrete appli-cations. This study aims to provide a suitable means for theutilization of fly ash in dense bituminous mixtures.

The first aim of this study was to review available litera-ture on the use of fly ash in asphalt concrete mixtures. Sec-ond, the possibilities of improving the mechanical propertiesof asphalt mixtures by using fly ash as mineral filler wasexplored. Finally, the effect of the use of fly ash as a filler

material was evaluated for performance improvement ofasphalt concrete pavements.

2. Historical background and definition offiller replacement

Normally, stone dust produced during the aggregatecrushing process is used as filler in asphaltaggregate mix-tures. It is common practice to use hydrated lime, portlandcement, and some other materials in place of stone dustwhen it is not available in the required amounts. The influ-ence of mineral filler comes from the fact that althoughsmall in weight with respect to the total mix, its surfacearea is quite large. In some cases, a reduction in the demandfor bitumen can be accomplished by the addition of mineralfiller because it fills the voids in the aggregate mass. How-ever, in other cases, due to the increase in surface area, alarger demand for bitumen may be necessary. In this study,various filler materials were selected to investigate thisimportant effect on the characteristics of the mixture.

Paving mixtures are composed of mineral aggregates heldtogether by an asphalt binder or bitumen. The mineralaggregates are distributed throughout the mixture in sizesranging from coarse to fine. Properly compacted asphaltmixtures produce a structure whose stability, stiffness, andwearing properties are dependent on the interlocking of theaggregate and the cohesiveness of the binder. Based on theabove rationale, mineral filler is defined as finely dividedmineral matter and includes material such as rock dust, slagdust, portland cement, hydrated lime (used also as an anti-stripping agent), fly ash, waste glass (Sobolev 2003), andloess (Anderson et al. 1982). If a more formal definition issought to be given, mineral filler is the material passing aNo. 200 (0.074 mm) US standard sieve. In usual practice,the mineral filler used in asphaltaggregate mixtures is thetail-end product obtained during the crushing process of nat-ural rock that conforms with aggregate specifications. Theuse of fly ash, portland cement, or other suitable materialsin place of natural mineral filler is universally accepted.

A controversial and unresolved question regarding fillermaterials concerns the mechanisms by which fillers changethe properties or behavior of paving mixtures. Brief consid-eration of such mechanisms may be helpful for the sake ofcompleteness of the filler replacement theory.

The mineral filler, whether artificial or natural, can affecthot-mix asphalt concrete in a number of ways. It may:

. stiffen the asphalt cement

. extend the asphalt cement (it can be called the bitumenextension)

. alter the moisture resistance of the mix

. affect the aging characteristics of the mix

. affect the workability and compaction characteristics ofthe mix (Anderson et al. 1982).

The mineral filler fraction is often referred to as theamount of material filling the voids between the coarsersand and aggregate particles. This concept sounds very ele-gant until an attempt is made to define the size range of thecoarser fraction that creates the voids. Puzinauskas (1983)arrived at the conclusion that there is no single size that can

28 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

be considered as filling the voids or creating the voids. Moreappropriately, the filler, the dust or simply the minus No.200 (US standard sieve) fraction should be considered anextension of the mineral aggregate framework. It is this min-eral aggregate framework that is primarily responsible forthe stability and load carrying properties of hot-mix asphaltconcrete (Puzinauskas 1983).

The portion of the filler with particles thicker than theasphalt film contributes to the interlocking of the aggregate.The other portion of the filler, with particles smaller thanthe thickness of the asphalt film, is suspended in the asphaltand constitutes the binder in the mixture. The filler may ex-tend the asphalt, thereby increasing the effective asphaltcontent in the mixture (Dukatz and Anderson 1980). Twotypes of stiffening effects may result from the portion of themineral filler that is suspended in the asphalt:

. a relatively small stiffening resulting from the volumefilling of the filler

. a relatively large stiffening resulting from the physico-chemical interaction between the asphalt and the surfaceof the mineral filler (Puzinauskas 1983).

3. Use of fly ash in asphalt concrete mixturesThe idea of using fly ash as mineral filler in asphalt mix-

tures is not a new concept. Its use in asphalt concrete mix-tures has been investigated by many scientists and theresearch findings that have been obtained by these scientistscan be summarized as follows.

Carpenter (1952) determined that a class F fly ash re-sulted in an excellent effect on the retained compressivestrength for asphalt concrete specimens immersed in water.Warden et al. (1952) stated that fly ash was a suitable fillermaterial in terms of mixing, placing and compaction, stabil-ity, resistance to water damage, and flexibility. Zimmer(1970) analyzed the effect of carbon content on fly ash. Hisresults revealed that specimens that are prepared by usingfly ash had higher retained strength after immersing them inwater.

Of significant difference from Carpenters research, Hen-ning (1974) investigated the effect of a class C fly ash onasphalt mixture properties. He concluded that the additionof 4% fly ash resulted in the highest stability and flow, butended up with lower air voids. Henning also stated that flyash created an improvement in the stability after immersionin water. Sankaran and Rao (1973) made a comparison offly ash with other fillers, such as kaolin clay and crusheddust. They pointed out that fly ash at 2% filler content pro-vided the highest stability among the other fillers.

Rosner et al. (1982) used fly ash as mineral filler andanti-stripping agent for asphalt concrete mixtures. Theyshowed that the retained strength of the samples increasedas additional fly ash was used in the prepared mixtures. Inmost of their cases, the retained strengths of fly ash mixtureswere considerably greater than those using natural fillermaterial.

In a study carried out by Suheibani (1986), fly ash wasevaluated as an asphalt extender. Suheibani analysed howfly ash particle size affects the viscosity of the asphalt towhich it was added. He also examined the effect of fly ash

as an extender on density measurements, voids, and mechan-ical properties of asphalt. Indirect tensile strength, creep,and resilient modulus test results showed that the additionof class F fly ash provided superior fatigue life, rut depthresistance, and tensile strength.

Also, Tons et al. (1983) investigated the use of class F flyash as a bitumen extender. Bitumen was replaced by variouspercentages of different fly ashes. Tests were carried out onasphalt specimens to determine resistance to moisture dam-age, thermal cracking, rutting, fatigue life, and asphalt hard-ening in mixtures. Noticeable improvements were observedfor asphalt hardening, moisture and freezethaw damageresistance, rutting resistance, fatigue life, density, and tensilestrength.

Cabrera and Zoorob (1994) established that, based on aworkability index at various temperatures, the pulverized flyash filler hot-mixed asphalt could be mixed and compactedat temperatures as low as 110 and 85 8C, respectively, with-out any detrimental effects on engineering and performanceproperties. They stated that there might have been consider-able savings in energy without an additional asphalt cementrequirement.

4. Experimental program

4.1. Material propertiesMixture characteristics are directly dependent on the

properties of the aggregate and bitumen that constitute thepaving mixture. As the filler, in general, is an integral partof the aggregate used in bituminous mixtures, its character-istics and amount contained in the mix play an importantrole in modifying the mixture characteristics. The optimumbitumen content and air voids is influenced by the filleramount and eventually all mechanical mixture characteris-tics are affected.

In the laboratory test program, a single aggregate grada-tion was selected to suit the wearing course type 3 gradationlimits set by the General Directorate of Highways of Turkey(2000). Calcareous aggregates obtained from a native quarryand 60/70 penetration bitumen obtained from a nearby refin-ery were used in the preparation of all specimens. The phys-ical properties of the bitumen are stated in Table 1.Modified mixtures used three types of fly ash, namelySoma (F type), Cayirhan (F type), and Kangal (C type) andtwo other filler materials, namely lime and portland cementin place of the calcareous filler. This replacement was madesolely on a weight basis. Some of the major properties ofthe coarse aggregates, fine aggregates, and six types of filler(namely calcareous, three types of fly ash, lime, and port-land cement) are stated in Tables 2, 3, 4, and 5.

Aggregate gradation for the bituminous mixtures tested inthe laboratory has been selected as an average of the wear-ing course type 3 gradation limits given by the GeneralDirectorate of Highways of Turkey (2000). The mixturegradation and gradation limits can be seen in tabulatedform in Table 6. Throughout the study, six different typesof filler were used, namely three types of fly ash, calcare-ous filler, lime, and portland cement. The fly ash usedthroughout the study was obtained from Soma, Cayirhan,and Kangal thermal power plants located in different re-gions of Turkey. Soma fly ash and Cayirhan fly ash are F

Tapkn 29

# 2008 NRC Canada

type and Kangal fly ash is C type. The lime and portlandcement were obtained from local manufacturers. Calcareousfiller was the tail-end product of calcareous aggregate thatwas obtained from the local quarry.

The chemical composition of the fly ash was supplied bythe research laboratories of the relevant thermal powerplants. The compositions are given in Table 4 and the appa-rent specific gravity values are given in Table 5.

4.2. Hot-mixture design using Marshall test methodTo determine the optimum bitumen content, it is required

to perform Marshall tests and carry out some analysis onthese test results. The relevant Marshall test data are statedin a tabulated manner in Tables 7 through 12. These valuesare the average of three different sets of specimens. There-fore, each table represents the test results of 30 differentspecimens. The standard Marshall specimens were preparedby applying 50 blows on each face. Also, the optimum bitu-men contents obtained for six different types of filler materi-als at the end of the Marshall design are given in Fig. 1.Throughout the tests that were carried out, 7% of filler wasreplaced solely on a weight basis by the three types of flyash, portland cement, and lime.

By using results given in Tables 712, namely for thethree different types of fly ash, portland cement, lime, andcalcareous filler, the optimum bitumen contents have beencalculated and presented in a graphical form in Fig. 1.

The bitumen contents corresponding to the mixtures withmaximal stability and unit weight, 4% air voids and 80%

voids filled with asphalt, were found and averaged accordingto the limits given by the General Directorate of Highwaysof Turkey (2000) to find the optimum value. These optimum

Table 1. Physical properties of the asphalt cement.

Property Test value StandardPenetration at 25 8C (1/10 mm) 62.0 ASTM D573 (ASTM 1973)Penetration Index +1.0 Ductility at 25 8C (cm) >100 ASTM D 11379 (ASTM 1979a)Loss on heating (%) 0.053 ASTM D 680 (ASTM 1980)Specific gravity at 25 8C (kg/m3) 1033 ASTM D7076 (ASTM 1976b)Softening point (8C) 57 ASTM D3676 (ASTM 1976a)Flash point (8C) 257 ASTM D 9278 (ASTM 1978)Fire point (8C) 295 ASTM D 9278 (ASTM 1978)

Table 2. Physical properties of coarse aggre-gates, tested in accordance with ASTM C127-80 (ASTM 1980b).

Property Test valueBulk specific gravity (kg/m3) 2754Apparent specific gravity (kg/m3) 2821Water absorption (%) 0.26

Table 3. Physical properties of fine aggregates,tested in accordance with ASTM C128-79(ASTM 1979b).

Property Test valueBulk specific gravity (kg/m3) 2741Apparent specific gravity (kg/m3) 2766Water absorption (%) 1.43

Table 4. Chemical composition of fly ash (FA) samples. LoI, losson ignition.

Oxide (%)FA Soma(class F)

FA Cayirhan(class F)

FA Kangal(class C)

SiO2 50.48 49.74 27.92Al2O3 27.64 14.70 11.96Fe2O3 4.80 9.04 5.14SiO2 + Al2O3 + Fe2O3 82.92 73.48 45.02CaO 13.08 13.64 37.86MgO 1.30 5.10 2.60SO3 0.97 3.64 12.10Na2O 0.30 2.10 0.40K2O 2.00 1.20 0.80Na2Oeq 1.62 2.89 0.93LoI 1.07 2.44 3.15

Table 5. Apparent specific gravities offiller materials, tested in accordancewith ASTM D854-83 (ASTM 1983).

Filler type Test valueCalcareous filler (kg/m3) 2632Soma fly ash (kg/m3) 2105Cayirhan fly ash (kg/m3) 2194Kangal fly ash (kg/m3) 2525Lime (kg/m3) 2315Portland cement (kg/m3) 3060

Table 6. Type 3 wearing course gradation accord-ing to the General Directorate of Highways ofTurkey (2000).

Sieve size(mm)

Gradationlimits (%)

Passing(%)

Retained(%)

12.7 100 100 09.52 87100 93.5 6.54.76 6682 74 19.52.00 4764 55.5 18.50.42 2436 30 25.50.177 1322 17.5 12.50.074 410 7 10.5Pan 7

30 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

bitumen contents are represented in Fig. 1. Voids in mineralaggregate (V.M.A.) must be greater than 12%13% accord-ing to the same criteria set forth by this agency. The V.M.A.results tabulated in Tables 712 confirm this fact.

The particle size distribution graphs of six different fillerspecimens, which are grouped into two parts as fly ashesand the others, are given in Figs. 2 and 3. The particle sizedistribution analyses of filler materials were carried out byMalvern Mastersizer (Malvern Instruments Ltd. 1997) work-ing in a dry basis. All of the tests were carried at the samelaboratory with the same test conditions. Figure 2 shows theparticle size distribution of three types of fly ashes. It canbe observed that the finest of these three samples is Kan-gal fly ash (class C) and the Soma fly ash (class F) is the

coarsest. The values of specific surface areas are stated inTable 13. These fly ash samples had approximately equalspecific surface area values (Table 13). Kangal fly ash hasthe highest specific surface area, therefore, resulting in thehighest optimum bitumen content (highest bitumen absorp-tion capacity). The research results confirm this assump-tion. As can be seen in Fig. 1, the optimum bitumencontent of Kangal fly ash is 6.28%, which is the largestamount of all fillers. Soma fly ash has a specific gravityof 2105 kg/m3. It is the lightest among the fillers and itcan affect some of the mixture properties. The amount andthe characteristics of the filler material can considerably al-ter the behavior of bituminous mixtures (Puzinauskas 1983;Dukatz and Anderson 1980; Carpenter 1952; Warden et al.

Table 7. Marshall test results for calcareous filler. V.M.A., voids in mineral aggregate.

Bitumencontent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 17.875 11.882 2235 1254 2.01 623.93.0 17.241 10.546 2298 1496 2.14 699.13.5 16.640 9.136 2329 1850 2.28 811.44.0 15.474 6.762 2373 2109 2.46 857.34.5 14.063 4.084 2424 2396 2.60 921.55.0 14.317 3.249 2429 2034 2.75 739.65.5 14.234 2.037 2443 1926 3.65 527.76.0 14.752 1.516 2439 1574 3.89 404.66.5 15.141 0.858 2440 1345 5.11 263.27.0 15.905 0.653 2429 1087 6.15 176.7

Table 8. Marshall test results for Soma (class F) fly ash filler. V.M.A., voids inmineral aggregate.

Bitumencontent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 17.680 12.155 2247 1388 2.27 611.53.0 17.573 10.979 2261 1717 2.32 740.13.5 17.201 9.512 2282 2207 2.40 919.64.0 16.724 7.919 2306 2385 2.31 1032.54.5 16.180 6.238 2332 2649 2.16 1226.45.0 15.987 4.941 2349 2476 2.41 1027.45.5 15.520 3.325 2373 2432 2.40 1013.36.0 15.813 2.578 2376 2119 2.68 790.76.5 15.982 1.691 2383 1755 3.50 501.47.0 16.414 1.121 2381 1358 4.33 313.6

Table 9. Marshall test results for Cayirhan (class F) fly ash filler. V.M.A., voids inmineral aggregate.

Bitumencontent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 16.080 10.749 2290 1524 2.27 671.43.0 15.874 9.447 2307 1834 2.32 790.53.5 15.931 8.427 2317 1909 2.40 795.44.0 15.430 6.792 2342 2123 2.31 919.04.5 14.919 5.134 2367 2062 2.16 954.65.0 14.846 3.957 2381 1902 2.41 789.25.5 14.227 2.154 2410 2047 2.40 852.96.0 14.544 1.416 2412 1597 2.68 595.96.5 14.880 0.708 2414 1342 3.50 383.47.0 15.686 0.563 2402 1122 4.33 259.1

Tapkn 31

# 2008 NRC Canada

1952). During the mixing process at high temperatures, thebitumen and filler come together to form a paste. The ulti-mate physical and chemical characteristics of this pastedetermine the mixture behavior. Bitumen extension andstiffening effect controlled by the filler content and itscharacteristics are two phenomenons that are frequentlymentioned in published literature (Puzinauskas 1983;Dukatz and Anderson 1980; Carpenter 1952; Warden et al.1952). Although the experimental results provide ade-quate guidance for the mix design process, to further clar-ify fillerbitumen interaction, more complete studies areneeded.

The effect of filler replacement with fly ash, lime, andportland cement is given in Tables 14 and 15. The most im-

pressive and interesting observation is the increase in theMarshall stability values and decrease of flow values forSoma fly ash specimens (versus control specimens with cal-careous filler).

Table 14 reports the maximum stability values obtainedfrom each data set of different fillers. It can be seen thatthe average stabilities of Soma fly ash specimens are app-roximately 11% higher and the average flow values areapproximately 17% lower than that of calcareous filler(control) specimens (Table 14). Also, the average air voidvalues of Soma fly ash specimens are approximately 53%higher than that of the control specimens. Finally, the aver-age specific gravity values of Soma fly ash specimens areapproximately 4% lower than that of calcareous filler

Table 10. Marshall test results for Kangal (class C) fly ash filler. V.M.A., voids inmineral aggregate.

BitumenContent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 18.240 12.991 2257 1499 3.15 475.93.0 18.277 11.966 2267 1815 2.97 611.13.5 18.577 11.207 2270 1930 2.85 677.24.0 18.242 9.801 2290 2091 2.37 882.34.5 18.849 9.413 2283 1989 2.60 765.05.0 18.339 7.782 2309 2073 2.67 776.45.5 17.330 5.566 2348 2219 2.83 784.16.0 16.889 3.980 2372 2268 2.74 827.76.5 16.251 2.153 2402 2231 3.53 632.07.0 16.867 1.791 2395 1863 3.46 538.4

Table 11. Marshall test results for lime filler. V.M.A., voids in mineral aggregate.

Bitumencontent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 17.933 12.597 2252 1225 2.88 425.33.0 17.825 11.417 2266 1577 2.84 555.33.5 18.417 10.999 2261 1687 2.20 766.84.0 17.434 8.859 2299 2065 2.04 1012.34.5 17.609 7.985 2305 1889 2.21 854.85.0 16.999 6.230 2334 1824 2.12 860.45.5 16.474 4.556 2360 1774 2.44 727.06.0 16.295 3.267 2376 1675 2.51 667.36.5 16.105 1.962 2392 1579 3.38 467.27.0 16.528 1.376 2392 1329 4.57 290.8

Table 12. Marshall test results for portland cement filler. V.M.A., voids in mineralaggregate.

Bitumencontent (%)

V.M.A.(%)

Air void(%)

Unit weight(kg/m3)

Stability(kg)

Flow(mm)

Marshallquotient

2.5 17.681 12.285 2303 1722 2.71 635.43.0 17.082 10.552 2331 2014 2.70 745.93.5 15.945 8.217 2375 2232 2.57 868.54.0 15.681 6.817 2394 2151 2.25 956.04.5 14.777 4.692 2431 2280 2.43 938.35.0 14.171 2.882 2460 2420 2.37 1021.15.5 14.308 1.907 2468 2162 2.97 727.96.0 14.741 1.278 2467 1860 3.97 468.56.5 15.391 0.914 2460 1549 4.25 364.57.0 16.092 0.627 2451 1327 4.58 289.7

32 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

specimens. The reasons why Soma fly ash specimens per-form better than the other two types of fly ashes can bestated as follows:

(1) Soma fly ash (class F) is the coarsest of the three typesof fly ashes (Fig. 2). Therefore, the stiffening effect aris-

ing from the fillerbitumen interaction is realized in theasphalt concrete specimens prepared with Soma fly ash.

(2) The aluminium oxide (Al2O3) percentage of Soma flyash specimens is 27.64%, which is more than twice thatof the other types of fly ash. The aluminium oxide por-tion of fly ash can be considered to be responsible for

Fig. 1. Optimum bitumen content for different filler materials.

Fig. 2. Particle size distributions of fly ash.

Fig. 3. Particle size distribution of portland cement, lime, and calcareous filler.

Tapkn 33

# 2008 NRC Canada

the bond with bitumen and, therefore, for the higherstrength and stiffness of the Marshall specimens.

(3) The apparent specific gravity of Soma fly ash is2105 kg/m3, which is the smallest of the three types offly ashes. Also, this value is much smaller than thespecific gravity of the calcareous filler. Therefore, itcan be concluded that the fillerbitumen paste of Somafly ash specimens shows better bitumen extension whencompared with the specimens prepared with other typesof fillers.

These results conform well with the results obtained byTons et al. (1983) in the sense that class F fly ash acts as asuperior bitumen extender. Two other fly ash materials andlime did not show superior properties versus calcareousfiller, with respect to stability. But, on the other hand, limehas an average flow value of approximately 22% lower thancalcareous filler. This interesting behavior of lime comesfrom the fact that it has the highest specific surface area of582.2 m2/kg. However, lime filler specimens also have thesmallest average stability values. Another major drawbackis related to the behavior of the portland cement specimens(Table 14). As can be seen, these samples also show supe-rior properties because of the high specific gravity of port-land cement. The highest average Marshall quotient value,which gives important information about the stiffness modu-lus of the mixture, was obtained for Soma fly ash speci-mens. This gives an important indication for fatigue lifeprediction of these types of specimens (i.e., the higher thestiffness of the specimens, the longer the fatigue life is).The fatigue life of Soma fly ash specimens is the highestamong all specimens. The minor variations in the specificgravity values of the Marshall specimens can be explainedby the changes in the specific gravity values of filler ma-terials.

From the point of view of economics, the aim of the de-signer is to minimize the cost of the bitumen that is beingused in the design. In the case where fly ash or other wastematerial is used as filler material, there is no extra cost topay for the replacement operation. The most important pointthat has to be emphasized is that fly ash replacement pro-vides real economy. This is the major difference from thepolymer modification option. The optimum bitumen contentfor calcareous filler is approximately 5%. Comparisons be-tween calcareous filler and the other filler materials are

shown in Table 15. In this table, the stability values aretaken as the reference point as these are important mechani-cal parameters of asphalt. Also, the stability values act as adominant figure in the design process of asphaltaggregatemixtures. Therefore, the bitumen content values that producea stability value similar to that of the control specimens(with calcareous filler) have to be represented. The maxi-mum stability values obtained for different types of fly ashare greater than those of the control specimens. A represen-tative bitumen content value that gives approximately equalstability values to the optimum bitumen content of 5%, as itwas defined for the control specimens, is accepted as theadjusted bitumen content of the different fillers. Thesevalues are presented in Table 15.

The Marshall quotient values in Table 15 are presented inFig. 4. This figure shows the favourable behavior of fly ashspecimens when compared with a calcareous type of filler.The fatigue life of fly ash specimens is relatively higherwhen compared with the fatigue life of the control mix.

4.3. Experimental methods

4.3.1. Fatigue life analysis of Marshall specimensFatigue can be defined as the phenomenon of fracture

under repeated or fluctuating stress having a maximumvalue generally less than the tensile strength of the material(Austin and Gilchrist 1996). Another definition for fatigue isthat it is a process of progressive localized permanentstructural change occurring in a material subjected to condi-tions that produce fluctuating stresses and strains at somepoint or points and which may culminate in cracks or com-plete fracture at a sufficient number of fluctuations (ASTM1963). Fatigue in asphalt pavement surfaces is caused by therepetitive application of vehicular loads, which induce fluc-tuating stresses and strains in the bituminous surface layer.Under traffic loading, the layers of a flexible pavementstructure are subjected to continuous flexing. The distress,which is caused by the fatigue phenomenon, begins withcracking that shows up as map or alligator cracking patternson the surface.

4.3.2. Repeated load indirect tensile testThe repeated load indirect tensile test is a commonly used

tensile test for stabilized materials. Most of the early re-ported tests have been reported for concrete or mortar; how-ever, this test is currently applied to cement-treated gravel,lime-soil mixtures, and asphalt-stabilized materials. Thistest involves loading a cylindrical specimen with a compres-sive load along two opposite sides, which results in a rela-tively uniform tensile stress acting perpendicular to andalong the diametral plane of the applied load. Finally, asplitting failure generally occurs along the diametral plane(Austin and Gilchrist 1996).

Under the applied force, perpendicular and horizontalstresses, elastic modulus and strain values can be easily cal-culated. The fatigue life can be determined from the reduc-tion of the elastic modulus value to a limiting value, definedas a fraction of the initial value (generally 50%) or as anumber of load repetitions that cause the first crack to arisein the specimen (Brown 1990).

Table 13. Specific surface areas of dif-ferent types of filler materials, tested inaccordance with ASTM C204-00(ASTM 2000).

Filler type (m2/kg) Test valueCalcareous filler 322.9Soma fly ash 249.4Cayirhan fly ash 242.7Kangal fly ash 277.0Lime 582.2Portland cement 265.3

34 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

4.3.3. Elasticity modulus, permanent and elasticdeformations

The UMATTA tester (ELE-UMATTA 1994) is a testingsystem that is used to find both the elastic modulus and thepermanent and elastic deformations of the previously pre-pared Marshall specimens (Wallace and Monismith 1980).The system operates automatically and is controlled withthe help of a personal computer and software called UMAT(ELE-UMATTA 1994). The parameters, such as appliedload level, load repetition, and time to reach maximum load-ing level, are given as an input before starting the test.

While the test is being carried out, with predefined intervals,the elastic and plastic deformations are recorded and tensilestress, resilient strain, and elastic modulus values are calcu-lated. The experiment is conducted in a temperature-controlled unit and the interior and surface temperature ofthe specimen is continuously recorded.

Based on these experiment results, the fatigue life ofasphalt pavements have been modelled in laboratory condi-tions and extensive tests have been conducted to investi-gate the effect of fly ash replacement on mixtureproperties. These tests were carried out with a UMATTA

Table 14. Volumetric mix properties of mixtures with maximum stability based ondifferent types of filler materials.

Filler typeAir void(%)

Unit weight(kg/m3)

Stabilitymax(kg)

Flow(mm)

Marshallquotient

Calcareous filler 4.084 2424 2396 2.60 921.5Soma fly ash 6.238 2332 2649 2.16 1226.4Cayirhan fly ash 6.792 2342 2123 2.31 919.0Kangal fly ash 3.980 2372 2268 2.74 827.7Lime 8.859 2299 2065 2.04 1021.3Portland cement 2.882 2460 2420 2.37 1021.1

Table 15. Mixture properties and the adjusted bitumen content values.

Filler typeAdjustedbitumen content

Stability(kg)

Flow(mm)

Marshallquotient

Gain inbitumen (%)

Calcareous filler 5.0 2034 2.75 739.6 0Soma fly ash 3.5 2207 2.40 919.6 +30Cayirhan fly ash 4.0 2123 2.31 919.0 +20Kangal fly ash 4.0 2091 2.37 882.3 +20Lime 4.0 2065 2.04 1012.3 +20Portland cement 3.0 2014 2.70 745.9 +40

Fig. 4. Marshall quotient values for different filler specimens.

Tapkn 35

# 2008 NRC Canada

tester (ELE-UMATTA 1994). The aim of these experi-ments was to model the fatigue or alligator cracking onthe pavement structure.

In this study, repeated indirect loadings were applied onMarshall specimens and lateral deformations were measured.The experiment was terminated when visible cracks startedto occur on the specimen surface.

5. Test results and discussionThe aim of this study was to investigate the effect of dif-

ferent types of fillers on properties of dense bituminousmixes. For the calcareous aggregate, calcareous type offiller, and 60/70 penetration bitumen used throughout theentire study, the optimum bitumen content was found to beapproximately 5.0%. The six sets of specimens, each com-posed of three different samples, were prepared using the5.0% bitumen content (the optimum bitumen content forcalcareous filler specimens). Although the air void contentsof the Marshall specimens prepared with Kangal fly ashand lime are above 6%, these specimens were tested todetermine their fatigue lives. These types of mixes per-formed very well in the fatigue tests when compared withcontrol specimens. The fatigue tests were carried out in atemperature-controlled unit. The test temperature was40 8C to resemble the in situ conditions that initiate fati-gue cracking. The Poissons ratio was taken as 0.35 andthe pulse period was selected as 500 ms. Loading timewas 100 ms, the rest period was 400 ms, and the appliedload was 1000 N. All of the tests were carried out withthese testing parameters and the test results presented inFigs. 5, 6, and 7 are the average values of the three setsof asphalt concrete specimens. Figure 5 represents the elas-tic strain versus pulse count of six different types of speci-mens prepared with different fillers. Figure 6 represents theelastic modulus versus pulse count of the same specimens.Figure 7 represents the permanent strain versus pulse countof these specimens.

It can be observed that different types of specimens showquite different patterns with respect to their fatigue lives. Ascan be seen from the graphs in these figures, Soma fly ashspecimens have the longest fatigue life and calcareous fillerspecimens have the shortest. When these graphs are ana-lysed, it can be seen that there are some peak points in theelastic strain versus pulse count graphs. After these pointsare reached, if the test is continued, a strain-hardening por-tion with a second peak value can be observed (Tapkn1998). In this study, the tests were terminated at the pointwhere the first visible crack was seen on the specimen sur-face. Therefore, the fatigue lives were calculated on thisbasis. Initial strain values, maximum strain values, finalstrain values, and the pulse counts corresponding to thesefinal strain values are presented in Table 16.

There are various approaches for the determination offatigue life of dense bituminous mixtures in published litera-ture (Tapkn 1998). In this study, the first visible crack ap-proach was used. According to this, an average fatigue lifeof the Soma fly ash specimens is approximately 36.8%higher than that of the control specimens.

The average initial resilient strains of the Soma fly ashspecimens are 97.4% lower than that of calcareous fillerspecimens. This is a clear indicator for the anticipated lon-ger fatigue life of this material. The average maximum resil-ient strain of Soma fly ash specimens is 26.8% less than thatof the control specimens. This means that Soma fly ashspecimens are exposed to a considerably lower amount ofresilient strain and, therefore, such materials would havelonger fatigue lives. The final average resilient strain of theSoma fly ash specimens is 5.6% less than the referencespecimen. This is shown in the strain hardening portion ofthe graphs in Fig. 5, which comes after the first peaks. Withthis redistribution of the load in the specimen body, theSoma fly ash specimens show longer fatigue lives.

The elastic modulus versus pulse count of six differenttypes of filler specimens are presented in Fig. 6. The mini-

Fig. 5. Elastic strain versus pulse count of specimens prepared with different fillers.

36 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

mum value of the resilient modulus values is the point thatis believed to be the end of the fatigue life. From this pointof view, Table 17 shows the behavior difference betweenthe six different types of specimens.

As can be seen from Table 17, Soma fly ash specimenshave the highest initial resilient modulus (Eo). On the otherhand, the calcareous filler specimens have the lowest initialresilient modulus. The initial resilient modulus of Soma fly

Fig. 6. Elastic modulus versus pulse count of specimens prepared with different fillers.

Fig. 7. Permanent strain versus pulse count of specimens prepared with different fillers.

Table 17. Resilient modulus of specimens withdifferent fillers.

SpecimenEo(MPa)

Emin(MPa)

Pulsecount

Calcareous filler 350.1 101.1 48 222Soma fly ash 580.0 321.4 65 947Cayirhan fly ash 401.9 137.5 51 336Kangal fly ash 466.2 263.3 58 520Lime 372.8 115.8 50 558Portland cement 433.3 174.5 54 223

Table 16. Resilient strains of specimens with different fil-lers.

Specimeneo(e)

emaximum(e)

efinal(e)

Pulsecount

Calcareous filler 300 766 540 48 222Soma fly ash 152 561 510 65 947Cayirhan fly ash 254 711 560 51 336Kangal fly ash 174 641 601 58 520Lime 289 742 587 50 558Portland cement 201 674 604 54 223

Tapkn 37

# 2008 NRC Canada

ash specimens was 65.7% higher than the control specimensand the minimum resilient modulus (Emin) was 217.9%higher than the control specimens. This represents the signif-icant behavioral difference between these two types ofspecimens with respect to their mechanical performance.Therefore, it is also clear from Table 17 and Fig. 6 that theSoma fly ash specimens will have the longest fatigue life.This supports the results of the resilient strain analysis.

Figure 7 shows the permanent strain versus pulse count ofsix different types of filler specimens. Permanent strainsrecorded during the tests show that as the number of loadrepetitions increase, permanent strains also increase. At theinitial stages of the test, this increase is gradual and nearlylinear. When Fig. 7 is examined, it can be observed that,after a number of load repetitions, the rate of increase inpermanent strains gets higher and, at later stages, againdrops. Before visible cracks are observed on the samples,the increase in permanent strains with load repetitions againbecomes nearly linear. Crack initiation, which is not visibleand not easily detectable, most probably will occur in be-tween the two linear portions of permanent strain plots(Tapkn 1998). This is shown in Fig. 8 for the permanentstrain versus pulse count relation.

It is believed that crack initiation starts before pulse countreaches the point where the rate of increase in permanentstrain is at its maximum. From each permanent strain pulse count plot, these points are determined by a singleprocedure. To each permanent strain pulse count relation,a regression curve is fitted and the pulse counts at whichrate of increase in permanent strains start to change abruptlyare determined. According to these discussions, the fatiguelives of specimens can be thought of as the number of loadrepetitions at which the rate of increase in permanent strainsis highest (Tapkn 1998). These critical pulse counts areshown in Fig. 9 for the entire series of specimens tested.

When the values of these six sets of specimens are exam-ined (which change between 38 325 and 50 344 pulsecounts), it can be seen that the difference between the fati-gue lives of the calcareous filler and Soma fly ash fillerspecimens with respect to permanent strain measurements is31.4%. This value conforms well with the fatigue life differ-

ence value of 36.8% obtained from the analysis of elasticstrain and elastic modulus graphs of Fig. 5 and Fig. 6.

6. Conclusions and recommendations

Filler replacement with fly ash provides a considerableeconomy of bitumen in asphaltaggregate mixtures designedfor the equivalent stability. From the physical interactionof fly ash and bitumen paste, a more preferable asphaltaggregate mixture behavior is obtained. This asphaltaggregatemixture behavior was investigated from the point of viewof rheological properties of the mixtures, namely by carry-ing out fatigue tests. The fatigue life of fly ash specimens,especially Soma fly ash, was found to be considerablyhigher than that of the control specimens. This is a verypromising result that can be explained mainly by the stiff-ening and void-filling effects of the fly ash filler acting asa bitumen extender in the asphaltaggregate mixture. Somafly ash is the coarsest among the investigated types of flyash it has the highest Al2O3 content and has the lowestspecific gravity therefore, the specimens prepared withSoma fly ash had the longest fatigue lives.

Fly ash filler replacement will draw the interest of pave-ment engineers as utilization of fly ash is an important prob-lem both from an economic and environmental point ofview. Also, another alternative, polymer modification, is amore expensive process that brings additional cost to theoverall design. Fly ash can be obtained easily with littlecost and the addition of fly ash to the asphaltaggregatemixtures does not need any specialized equipment or skilledworkmanship.

The physico-chemical interaction between fly ash and thebitumen should be studied with the help of techniques usedin materials science. The use of scanning electron micro-scopy can resolve many questions about the interaction offly ash and bitumen paste. Also, fatigue tests were carriedout using the 5% optimum bitumen content, which was thecalculated value for calcareous filler specimens. The sametests can be carried out with the adjusted bitumen contentsobtained for the different types of fillers, to validate eco-nomic considerations with respect to the use of these fillers.

Fig. 8. Region of crack initiation for permanent strain versus pulse count (Tapkin 1998).

38 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada

These recommendations are beyond the scope of this study,but they form a good basis for further research.

ReferencesAli, N., Chan, J.S., Simms, S., Bushman, R., and Bergan, A.T.

1996. Mechanistic evaluation of fly ash asphalt concrete mix-tures. Journal of Materials in Civil Engineering, 8(1): 1925.doi:10.1061/(ASCE)0899-1561(1996)8:1(19).

Anderson, D.A., Tarris, J.P., and Brock, D. 1982. Dust collectorfines and their influence on mixture design. Proceedings of theAssociation of Asphalt Paving Technologists, 51: 363374.

Asi, I., and Assaad, A. 2005. Effect of Jordanian oil shale fly ashon asphalt mixes. Journal of Materials in Civil Engineering,17(5): 553559. doi:10.1061/(ASCE)0899-1561(2005)17:5(553).

ASTM. 1963. A guide for fatigue testing and the statistical analysisof fatigue data. ASTM Special Publication 91-A. American So-ciety for Testing and Materials, Philadelphia, Pa.

ASTM. 1973. Standard Test Method for Penetration of BituminousMaterials. Standard ASTM D5-73. American Society for Testingand Materials, West Conshohocken, Pa.

ASTM. 1976a. Standard Test Method for Softening Point of Bitu-men (Ring-and-Ball Apparatus). Standard ASTM D36-76.American Society for Testing and Materials, West Consho-hocken, Pa.

ASTM. 1976b. Standard Test Method for Specific Gravity andDensity of Semi-Solid Bituminous Materials (PycnometerMethod). Standard ASTM D70-76. American Society for Test-ing and Materials, West Conshohocken, Pa.

ASTM. 1978. Standard Test Method for Flash and Fire Points byCleveland Open Cup Tester. Standard ASTM D92-78. AmericanSociety for Testing and Materials, West Conshohocken, Pa.

ASTM. 1979a. Standard Test Method for Ductility of BituminousMaterials. Standard ASTM D113-79. American Society for Test-ing and Materials, West Conshohocken, Pa.

ASTM. 1979b. Standard Test Method for Density, Relative Density(Specific Gravity), and Absorption of Fine Aggregate. StandardASTM C128-79. American Society for Testing and Materials,West Conshohocken, Pa.

ASTM. 1980a. Standard Test Method for Loss on Heating of Oil

and Asphaltic Compounds ASTM D6-80. American Society forTesting and Materials, West Conshohocken, Pa.

ASTM. 1980b. Standard Test Method for Density, Relative Density(Specific Gravity), and Absorption of Coarse Aggregate. Stan-dard ASTM C127-80. American Society for Testing and Materi-als, West Conshohocken, Pa.

ASTM. 1983. Standard Test Methods for Specific Gravity of SoilSolids by Water Pycnometer. Standard ASTM D854-83. Ameri-can Society for Testing and Materials, West Conshohocken, Pa.

ASTM. 2000. Standard Test Methods for Fineness of HydraulicCement by Air-Permeability Apparatus. Standard ASTM C204-00. American Society for Testing and Materials, West Consho-hocken, Pa.

Austin, R.A., and Gilchrist, A.J.T. 1996. Enhanced performance ofasphalt pavements using geocomposites. Geotextiles and Geo-membranes, 14: 175186. doi:10.1016/0266-1144(96)00007-6.

Brown, S. 1990. The Shell bitumen handbook. Shell Bitumen, UK.Cabrera, J.G., and Zoorob, S. 1994. Design of low energy hot

rolled asphalt. In Proceedings of the 1st European Symposium.The Civil Engineering Materials Unit, Department of CivilEngineering, University of Leeds, Leeds, UK. pp. 289-308.

Carpenter, C.A. 1952. A comparative study of fillers in asphalticconcrete. Public Roads, 27: 101110.

Churchill, E.V., and Amirkhanian, S.N. 1999. Coal ash utilizationin asphalt concrete mixtures. Journal of Materials in Civil Engi-neering, 11: 295301. doi:10.1061/(ASCE)0899-1561(1999)11:4(295).

Dukatz, E.L., and Anderson, D.A. 1980. The effect of various fil-lers on the mechanical behaviour of asphalt and asphaltic con-crete. Proceedings of the Association of Asphalt PavingTechnologists, 49: 530543.

ELE-UMATTA. 1994. Universal materials testing apparatus forasphalt and unbound specimens. Reference and operating man-ual. ELE International Ltd., Hertfordshire, UK.

General Directorate of Highways of Turkey. 2000. Highway techni-cal specifications, Item No. 170/2. General Directorate of High-ways of Turkey, Ankara, Turkey.

Henning, N.E. 1974. Evaluation of lignite fly ash as a mineral fillerin asphaltic concrete. Twin City Testing and EngineeringLaboratory, St. Paul, Minn. Report No. 2(73).

Fig. 9. Fatigue lives of specimens prepared with different fillers according to permanent strain records.

Tapkn 39

# 2008 NRC Canada

Malvern Instruments Ltd.. 1997. Mastersizer Series Software Suite,version 2.19 [computer program]. Malvern Instruments Ltd.,Worcestershire, UK.

Puzinauskas, V.P. 1983. Filler in asphalt mixtures. The AsphaltInstitute, College Park, Md. Research Report No. 692..

Rosner, J.C., Chehovits, J.G., and Morris, G.R. 1982. Fly ash as amineral filler and anti-strip agent for asphalt concrete. In Pro-ceedings of the 6th International Ash Utilization Symposium.Vol. 1. United States Department of Energy, Morgantown, W.Va. Report No. DOE/METC/82/2.

Sankaran, K.S., and Rao, D.R. 1973. The influence of the qualityof filler in asphaltic paving mixtures. Indian Roads Congress,35: 141151.

Sobolev, K. 2003. Recycling of waste glass in eco-cement. Ameri-can Ceramic Society Bulletin, 82: 95019507.

Sobolev, K., and Naik, T.R. 2005. Performance as a factor forsustainability of the cement industry. In Proceedings of theCANMET/ACI Three-Day International Symposium on Sustain-able Development of Cement and Concrete, 57 October 2005,Toronto, Ontario. Department of Civil Engineering and Mech-anics, College of Engineering and Applied Science, The Univer-

sity of Wisconsin, Milwaukee, Wis. Report No. CBU-200517.pp. 295-312.

Suheibani, A.R.S. 1986. The use of fly ash as an asphalt extender.Ph.D. thesis, University of Michigan, Ann Arbor, Mich.

Tapkn, S. 1998. Improved asphalt aggregate mix properties byportland cement modification. M.Sc. thesis, Middle East Techni-cal University, Civil Engineering Department, Ankara, Turkey.

Tons, E., Goetz, R.O., and Razi, M. 1983. Fly ash as asphalt redu-cer in bituminous base courses. Project Report. University ofMichigan, the Board of Water and Light, Consumer Power Co.,and Detroit Edison Co., Detroit, Mich.

Wallace, K., and Monismith, C.L. 1980. Diametral modulus testingon nonlinear pavement materials. Proceedings of the Associationof Asphalt Paving Technologists. 49: 633652.

Warden, W.B., Hudson, S.B., and Howell, H.C. 1952. Evaluationof mineral filler in terms of practical pavement performance.Proceedings of the Association of Asphalt Paving Technologists.27: 101110.

Zimmer, F.V. 1970. Fly ash as bituminous filler. In Proceedings ofthe 2nd Ash Utilization Symposium, Pittsburgh, Pa., 1011March 1970. United States Department of Interior, Bureau ofMines, Washington, D.C.

40 Can. J. Civ. Eng. Vol. 35, 2008

# 2008 NRC Canada