Low Temperature Fracture Properties of PolyphosphoricAcid Modified Asphalt Mixtures

Eyoab T. Zegeye1; Ki H. Moon2; Mugur Turos3; Timothy R. Clyne4; and Mihai O. Marasteanu5

Abstract: The low temperature fracture properties of polyphosphoric acid (PPA) modified mixtures are evaluated and compared with those ofpolymer modified mixtures. The main objective is to determinewhether PPA can partially or completely substitute traditional polymer modifiers,without adversely affecting the mixture’s resistance to thermal cracking. Laboratory compacted and field cored test samples from the MinnesotaRoad Research Project (MnROAD) were tested using traditional methods as well as newly developed fracture testing protocols: indirect tensiletest (IDT), semicircular bending test (SCB), and disk-shaped compact tension test (DCT). The effects of temperature, air-void content, commaand long-term aging on low temperature fracture properties were analyzed; and field performance observations of the test cells from MnROADwere discussed. On the basis of the analysis, the fracture resistance of the PPA modified mixture is less than that of the SBS modified mixture.However, when PPA is used to substitute part of the SBS, a mixture with fracture resistance comparable to that of the mixture modified with SBSalone is produced. DOI: 10.1061/(ASCE)MT.1943-5533.0000488. © 2012 American Society of Civil Engineers.

CE Database subject headings: Temperature effects; Cracking; Acids; Bending; Asphalts; Mixtures.

Author keywords: Low temperature cracking; Polyphosphoric acid; Semicircular bending tests; Disk-shaped compact tension tests;Low temperature fracture properties; Modified asphalt mixtures.

Introduction

The increasing traffic loads to which asphalt pavements are sub-jected have accelerated the need to explore new paving materialsand new technologies that improve the performance of pavements.A common method used to improve the performance of asphaltbinders consists of increasing the high temperature limit of the per-formance grade (PG), thus expanding the temperature range with-out affecting the low temperature limit. For this reason, specialpolymer additives have been used to produce modified asphaltbinders with improved resistance to rutting and unaltered resistanceto thermal cracking. Recent investigations indicate that asphaltbinders modified with PPA, alone or in combination with tradi-tional polymers, can yield similar gains in PG limit for lower costof modification (Clyne et al. 2009). However, because of prematurefailures observed in some asphalt pavements containing PPA modi-fied binders, its application has been restricted by some highwayagencies (Arnold et al. 2009).

In the research described in this paper, the effect of PPAmodification on the low temperature fracture properties of asphalt

mixtures was investigated. Traditional and new fracture tests wereused to determine fracture properties for a set of laboratoryprepared specimens and field samples cored from the MinnesotaRoad Research Project (MnROAD).

Background on Use of Acid Modifiers in AsphaltMixtures

One of the earliest documented uses of PPA as an asphalt bindermodifier was in altering the viscosity–penetration relationship ofasphalt [S. Alexander, “Method of treating asphalt,” U.S. Patent3,751,278 (1973)]. Unlike the air-blown oxidation process, whichstiffens the binder at high temperatures but severely damages itslow temperature properties, adding PPA increases resistance torutting while still meeting the low temperature requirements.

In past decades, several researchers (Dickinson 1974; Giavariniet al. 2000; Bishara et al. 2001; Ho et al. 2001; Maldonado et al.2006; Arnold et al. 2009) have documented the mechanisms under-lying PPA modification. The strength of the effect may vary frombinder to binder, but it is generally observed that the PG higherservice temperature limit increases almost linearly with the amountof added PPA (Maldonado et al. 2006; Arnold et al. 2009).

Giavarini et al. (2000)andArnoldet al. (2009)performed in-depthrheological andphysiochemical analysis, includingbinder fractiona-tion, to investigate whether the chemical composition of the binderaffects the outcome of PPA modification. Both studies reported anincrease in asphaltene and decrease in resin fraction with increasingPPA concentrations. Stiffening is more likely to occur in binders thatexhibit the high asphaltene-to-resin ratios caused by adding PPA.

Kodrat et al. (2007) investigated the low temperature fractureproperties of PPA modified binders and compared them with thoseof plain and polymer modified (SBS, Elvaloy) binders of similargrades. Tests were performed on a large number of binder typesusing the extended bending beam rheometer (EBBR), compacttension (CT), and double edge-notched tension (DENT). The re-sults confirmed that addition of PPA increases the PG span withoutaffecting the lower temperature limit. The effects of PPA on fracture

1Ph.D. Candidate, Univ. of Minnesota, 500 Pillsbury Dr. SE,Minneapolis, MN 55455 (corresponding author). E-mail: zegey001@umn.edu

2Ph.D. Candidate, Univ. of Minnesota, 500 Pillsbury Dr. SE,Minneapolis, MN 55455.

3Scientist, Univ. of Minnesota, 500 Pillsbury Dr. SE, Minneapolis,MN 55455.

4MnROAD Operations Engineer, Mn/DOT Office of Materials andRoad Research, 1400 Gervais Ave., Maplewood, MN 55109.

5Associate Professor, Univ. of Minnesota, 500 Pillsbury Dr. SE,Minneapolis, MN 55455.

Note. This manuscript was submitted on March 31, 2011; approved onJanuary 24, 2012; published online on January 26, 2012. Discussion periodopen until January 1, 2013; separate discussions must be submitted for in-dividual papers. This paper is part of the Journal of Materials in CivilEngineering, Vol. 24, No. 8, August 1, 2012. ©ASCE, ISSN 0899-1561/2012/8-1089–1096/$25.00.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 1089

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

properties in the brittle state and on the reversible aging processwere found to be insignificant. However, PPA modified binderssubjected to the DENT test at ambient temperature had significantlyreduced strain tolerance, thus increasing the probability of fatiguecracking in service.

The combined effects of PPA and other additives such as hy-drated lime, SBS, and Elvaloy on some properties of asphalt mix-tures were also extensively discussed by Fee et al. (2010). Theseresearchers investigated the moisture susceptibility of mixturescomposed of limestone aggregates and PPA modified asphaltbinder. Hydrated lime used in combination with PPA did not affectthe PG grade gain and significantly improved mixture performance.However, they noted that if used in great amounts, the lime mayreact with the PPA and reduce or neutralize the expected gain inPG high temperature limit. Furthermore, the use of PPA in conjunc-tion with Elvaloy or SBS results in a greater improvement in mix-ture performance than the use of PPA alone.

The considerably lower cost of PPA compared with polymermodifiers has made PPA modification a popular choice for pave-ment applications. In addition, using PPA in combination withother polymers can reduce the cost of modification (Clyne et al.2009) and improve the mixing and compaction characteristics ofthe mix (Arnold et al. 2009).

Materials

Four Superpave asphalt mixtures were used in this investigation (seeTable 1). The mixtures were used in Cells 33, 34, 35, and 77 of

MnROAD and varied in the type and amount of binder modificationemployed; PPA (Cell 33), SBS (Cell 34), PPAþ SBS (Cell 35), andPPAþ Elvaloy (Cell 77). These modifications were used to up-grade a base binder of about PG 49–34 to a binder of PG 58–34.Accordingly, 0.75 PPA, 1% SBSþ 0:3 PPA, 2% SBS, and 1.1%Elvaloyþ 0:3 PPA were used for Cells 33, 34, 35, and 77, respec-tively. The amounts of modifiers used were varied to obtain the samegain in high temperature PG limit. The effects of the modificationson the rheological properties of the base binder are illustrated inFig. 1 (Clyne et al. 2009). The figure reports the direct shear rhe-ometer (DSR) and bending beam rheometer (BBR) test results. Priorto the BBR testing, the binders had been aged in a pressure agingvessel (PAV). It can be observed that the four different types andamounts of modification yielded similar gains in high temperaturePG without deteriorating the low temperature limit.

The mixtures have a maximum aggregate of 12.5 mm, limited(< 10%) limestone, and 1% hydrated lime. In addition, three of themixtures also contained a liquid phosphate ester antistrip, as indi-cated in Table 1. The decision whether to add the liquid antistripwas based on Hamburg wheel track testing on laboratory mixturesperformed during the mix design stage.

Sample Preparation

The test samples for this study were prepared from loose mix com-pacted in the laboratory using the Superpave gyratory compactor(SGC). Mixtures were compacted at the University of Minnesota(UMN) with two target air-void contents: 4 and 7%. The latter

Table 1. Description of Asphalt Mixtures Used

MnRoad test cell Construction date Binder grade Asphalt modifiers Innovalt W liquid antistrip

33 September 2007 PG 58–34 0.75% PPA 0.5%

34 September 2007 PG 58–34 1:0%SBSþ 0:3% PPA 0.5%

35 September 2007 PG 58–34 2.0% SBS None

77 September 2007 PG 58–34 1:1%Elvaloyþ 0:3%PPA 0.5%

Fig. 1. Performance grade (PG) of modified binders

1090 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

is assumed to be more representative of typical asphalt pavementsimmediately after construction.

In addition, field cores taken from MnROAD Cells 33, 34, 35,and 77 were also investigated. The cores were taken in June 2010,approximately 2 1/2 years after their construction. Eleven cylindri-cal field cores for each mixture, sampled from between wheelpaths, were delivered to UMN. The thickness of the cores rangedfrom 100 to 150 mm (4 to 6 in.). The cells from which the sampleswere collected had two identical 50-mm (2-in.) lifts, but only thetop layers were made of the mixtures of interest to this researchstudy. Therefore, the bottom layers were cut and discarded. In ad-dition, the upper 5 mm (0.20 in.) was also cut and discarded toproduce a smoother surface. Air-void contents of the field cores,determined according to AASHTO T166 (2005) are listed inTable 2.

Experimental Variables

Cylindrical SGC compacted and field cores were cut into indirecttensile (IDT), semicircular bending (SCB), and disk-shaped com-pact tension test specimens. Three replicates were used for eachfactor/level combination.

The IDT, SCB, and DCT laboratory specimens obtained fromthe 4 and 7% SGC cylinders were tested at two temperatures:binder PG low temperature limit (PGLT), and PGLTþ 10°C.

An additional set of the 7% void content SGC cylinders waslong-term aged by placing the SGC cylinders an oven for 5 daysat 85°C, according to the AASHTO R30-02 protocol (AASHTO2006). The long-term aged and field samples were tested onlyat PGLT.

Test Methods

The IDT and SCB tests were performed at the UMN, and the DCTtests were performed at the University of Illinois, Urbana-Champaign (UIUC).

The IDT test was used to determine the creep compliance andstrength of the mixtures according to AASHTO T322(AASHTO 2003).

The SCB test was used to determine the fracture toughness(KIC) and fracture energy (GF) of the mixtures. The test procedure,as well the computation of the SCB fracture parameters, is reportedelsewhere (Li and Marasteanu 2004, 2010). The SCB specimenwas a 150-mm-wide, 25-mm-thick semicircular slice. A verticalnotch 15 mm long and 2 mm wide was cut in the middle of thespan between the supports. Tests were performed in a MTS (EdenPrairie, MN) servohydraulic testing system equipped with an envi-ronmental chamber. The loading was controlled by the crack mouthopening displacement (CMOD) and varied to maintain a constantCMOD rate of 0:0005 mm∕s for the entire duration of the test. Thisslow loading rate was selected to reflect the rate of temperaturechange that occurs in real field conditions.

The DCT test was performed to determine the fracture energy ofthe mixtures according to ASTM D7313-07 (ASTM 2007). Thetest is performed on DCT specimens 150 mm wide, 145 high,and 50 mm thick. The initial fracture ligament length was82.5 mm. Tensile loading was applied at the holes at a constantCMOD rate of 0:017 mm∕s.

Test Results of Laboratory Compacted Samples

Results of the IDT strength test are illustrated in Fig. 2. The figureshows the mean strength values calculated from three test repli-cates. The strength of the mixtures decreases as the temperaturedrops; however, the SBS and PPAþ Elvaloy modified mixtureswith 7% air-void content appear to be less sensitive to the changein temperature. As expected, mixtures with lower void content havehigher strengths than mixtures with higher void content. With re-spect to the effect of modification, several observations can bedrawn: SBS and PPAþ SBS modified mixtures have higherstrength values at all air-void and test temperature levels;

Table 2. Air-Void Contents for Field Cores

Cell Mean CV (%)

33 5.3 1

34 5.9 2

35 6.4 2

77 5.1 13

Fig. 2. Results of IDT strength tests

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 1091

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

þElvaloy modified mixtures have about the same strength asPPA-modified mixtures.

Fig. 3 shows the IDT stiffness test results at 60 and 500 s. It isshown that for all mixtures considered, stiffness increases dramati-cally when temperature drops. Similarly, stiffness is much higher inthe 4% void content mixtures. Overall, the highest increase in stiff-ness resulting from a temperature change is observed in the PPAþElvaloy modified mixture. The PPA modified mixture is relativelyless responsive to change in temperature at 4% void content. On thecontrary, at 7% void content, the PPA modified mixture has thesecond highest increase in stiffness. With respect to the effect ofmodification, overall SBS alone and PPþ SBS modified mixtureshave higher stiffness values compared PPA modified mixtures.

However, when only the highest test temperature is considered,the difference between the mixtures is reduced.

Results of the SCB fracture tests are illustrated in Fig. 4. Thehighest fracture toughness values are observed in SBS and PPAþSBS modified mixtures, regardless of the air-void content level.From Fig. 4, it can also be inferred that PPA and PPAþElvaloy mixtures are basically equivalent with respect to KIC val-ues. The SBS and PPAþ SBS modified mixtures, tested at thehighest test temperature, have higher SCB fracture energy valuesthan the other mixtures. For all mixtures, GF decreases with tem-perature and with increase in void content. Interestingly, the rates ofdecrease in GF resulting from the decrease in temperature are sim-ilar for PPA and PPAþ SBS modified mixtures. Both fracture

Fig. 3. Results of IDT creep stiffness tests

Fig. 4. Results of SCB fracture tests

1092 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

parameters, KIC and GF , are negatively affected by increase in air-void content. An exception to this trend is seen in SBS modifiedmixtures tested at the highest test temperature (PGLTþ 10°C).

Fracture energy was also computed from the DCT test; the re-sults are shown in Fig. 5. The DCT fracture energy results areslightly higher than those obtained for SCB because of the differentspecimen geometry and the much higher loading rate used in DCTtesting. Nevertheless, the DCT results also confirm that SBS andPPAþ SBS modified mixtures have higher fracture energy thanPPA and PPAþ Elvaloy modified mixtures.

Statistical Analysis of Results from LaboratoryCompacted Samples

The MacAnova statistical software package was used to performstatistical analysis, and a three-way analysis of variance (ANOVA)was used to examine the differences among the mean response val-ues of the different treatment groups. Results for each test param-eter were separately analyzed. Experimental factors were: testtemperature (two levels), air-void content (two levels), and typeof asphalt modification (four levels).The significance of the differ-ences was tested at the 0.05% level of error. In running the statis-tical tests, particular attention was paid to some aspects of theANOVA models. The methods of analysis of experimental results(comparing the average responses of different treatment groups us-ing ANOVA or contrasts and pairwise comparison) are based on theassumption that the errors in the data are independent normals with

constant variance. If these model assumptions do not hold, theinference may be misleading. Therefore, for each of the ANOVAmodels used in this study, the nature of and degree to which themodel assumptions were violated was checked, and correctivemeasures were taken if needed. Graphical and numerical tools (pro-vided in MacAnova) such as normal probability plots and residualplots were used to assess the violation of the assumption. The pri-mary tool used to deal with the violation of assumptions was trans-formation of the response to a different scale (square root,logarithm, etc.). In addition, the occurrence of unbalanced data(not all factor/level combinations have the same amount of repli-cation) necessitated ANOVA models adjusted to Type III sum ofsquares. All four treatments are comparable, in that none of themrepresents a control group (no treatment). Therefore, the statisticalanalysis is intended to evaluate whether the difference in responseamong the four differently modified mixtures is significant.

The ANOVA results for all test parameters are summarized inTable 3. Significant treatment effects are indicated in the table.

According to the statistical analysis, the effects of temperatureand air-void content observed for all test parameters, and discussedbefore, are statistically significant. The effect of modifiers, which isthe subject of this research work, is found to be statistically signifi-cant for the following test parameters: DCT fracture energy, SCBfracture toughness, SCB fracture energy and IDT strength. Thechange in stiffness caused by modifiers was found to be not sig-nificant. This means that the changes in mixture stiffness attributedto the four different types of modification are comparable.

Fig. 5. Results of DCT fracture tests

Table 3. Summary of ANOVA

Factor

DCT SCB SCB IDT IDT IDT

GF GF KIC Strength S at 60 s S at 500 s

Mixture type a a a a b b

Void content b a a a a a

Test temperature a a b a a a

Mixture void b b b b b b

Mixture temperature b a b b b b

Mixture void temperature b b b b b b

Note: S = IDT stiffness.aHighly significant.bNot significant.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 1093

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

In addition, multiple comparisons at the 5% level of significancewere performed to compare and rank the tested mixtures accordingto the different test methods. The outcomes are reported in Table 4,in which statistically similar mixtures are grouped together. Theletter ‘A’ is used to indicate the best performing group of mixtures;the letter ‘B’ refers to the second best performing group of mix-tures; and so on. Mixtures that are difficult to distinguish are in-dicated by two or three group letters. Accordingly, SBS andPPAþ SBS modified mixtures have generally higher strengthand higher fracture resistance than mixtures containing PPA orElvaloyþ PPA. In the case of IDT strength, the difference betweenPPAþ SBS and SBS is also considered to be statistically signifi-cant; SBS alone provides higher material strength.

Test Results for Field Cored Samples

In Figs. 6 and 7, the test results for field and long-term aged labo-ratory samples are plotted along with values for the correspondingunaged laboratory compacted samples. The laboratory samplesplotted all have an air-void content of 7%. Both field samplesand aged samples were tested only at PGLT test temperature.

With respect to the stiffness of field cores (see Fig. 6), it can beobserved that mixtures containing PPA or Elvaloyþ PPA havehigher stiffness at 60s than mixtures containing SBS orSBSþ PPA. However, at 500s, the difference is decreased. Onthe basis of the strength and fracture parameter results, it appearsthat generally, SBS and PPAþ SBS modified mixtures performbetter. This finding is similar to what was observed for the labo-ratory compacted mixtures.

Statistical Analysis of Results for Field CoredSamples

Analysis of variance was used to test whether the differences in theresponses of the four field mixtures were statistically significant. Itwas found that the field mixtures were similar with respect to IDTstiffness, IDT strength, SCB KIC, and DCT GF. The differences inSCB fracture energy results were small but statistically significant.In Table 5, the mixtures are ranked with respect to SCB fractureenergy value.

Test Results for Long-Term Aged LaboratoryCompacted Samples

Finally, the effect of long-term laboratory aging on the modifiedmixtures was investigated. Test results are summarized in Figs. 6and 7. Long-term aging increases the stiffness of all mixtures, ex-cept the SBS modified at 500s. In the IDT strength plot, agingcauses a 9% increase in strength for all mixtures except that modi-fied with PPAþ Elvaloy. According to the SCB fracture toughnessplot, long-term aging has almost no effect on PPA and SBS modi-fied mixtures. On the contrary, PPAþ polymer modified mixturesexhibit almost a double-digit percentage increase because of long-term aging. SCB fracture energy always increases with the long-term aging to which the mixtures were subjected. The highestincrease is observed in PPA and SBS modified mixtures. Finally,DCT fracture energy results show no difference between long-termaged and unaged mixtures. The PPAþ SBS modified mixture ex-hibits a steep decrease in DCT GF as a result of long-term aging.

Table 4. Statistical Grouping and Ranking of Modified Mixtures—Lab Compacted

Modifier

IDT strength SCB K IC SCB GF DCT Gf

(MPa) (MPa �m0:5) (J∕m2) (J∕m2)

Group mean Rank Group mean Rank Group mean Rank Group mean Rank

SBS 5.1 A 0.96 A 451.57 A 561 A/B

PPAþ SBS 4.57 B 0.89 A/B 428.87 A 584 A

PPAþ Elv 4.19 C 0.81 B 390.13 A/B 496 B

PPA 4.01 C 0.80 B 323.67 B 490 B

Fig. 6. Results of IDT creep stiffness tests for unaged, aged, and field samples

1094 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

On the basis of the ANOVA, the effect of long-term aging onIDT, SCB, and DCT results was not statistically significant.The results suggest that the current long-term aging procedureneeds to be investigated further to determine if this method is ap-propriate for simulating field aging with respect to low temperatureproperties.

Test Section Field Performance

After 4 years in service, MnROAD Cells 33, 34, and 35 show nothermal cracking. The only test section with transverse cracking isCell 77 (Elvaloyþ PPA), which has two cracks in 350 ft. There areseveral construction-related reasons why this section may havesome cracking, so there is little concern that the low temperaturecracking is excessive. Overall, the performance of each of the PPAmodified binders has been very good.

Conclusions and Remarks

The low temperature properties of asphalt mixture modified withPPA and PPA in combination with traditional polymer modifierswere investigated using traditional and newly developed test meth-ods. All mixtures had the same mix design except for the type andamount of modifiers used to upgrade the base binder to PG 58–34.The modifications considered were: PPA, SBS, PPAþ SBS, andPPAþ Elvaloy. Tests were performed on both laboratory com-pacted and field cored test specimens.

On the basis of the analyses performed on the experimentalresults, the following conclusions were drawn:• Indirect tensile stiffness does not statistically significantly differ

between the modifications. Therefore, the four modificationsconsidered in the present work have a similar effect on mixturestiffness.

• For IDT strength, PPAþ SBS and SBS modified mixtures aresignificantly stronger than mixtures modified with PPA aloneand PPAþ Elvaloy.

• Comparison of the mixtures with respect to SCB and DCT frac-ture parameters confirms that PPAþ SBS and SBS modifiedmixtures have better resistance to cracking than PPA and PPAþElvaloy modified mixtures.

• For field cored samples, SCB fracture energy does not statisti-cally differ with mixture. The mixtures containing PPA alonehave worse fracture properties than the other mixtures.

• Long-term aging method does not cause statistically significantdifferences among the mixtures. This result brings into questionthe validity of using this method to simulate field aging on lowtemperature properties.

• Field performance of the cells investigated, including allPPA-modified mixtures, is very good.Overall, it can be concluded that the low temperature fracture

properties of mixtures modified with 0.7% PPA are reasonablygood. The SBS and PPAþ SBS modified mixtures have the high-est resistance to low temperature cracking of all mixtures investi-gated. The results suggest that using PPAþ SBS modifiedmixtures, in which the amount of SBS is half of the amount usedin typical SBS modified mixtures, may provide a cost-effectivealternative for improving the cracking resistance of asphalt pave-ments. These results are confirmed by the field performance of theMnROAD test sections that have not thermally cracked.

Acknowledgments

This research was sponsored by Federal Highway AdministrationNational Pooled Fund Study TPF-5(132) led by the Minnesota De-partment of Transportation. This support is gratefully acknowledged.

Table 5. Statistical Grouping and Ranking of Modified Field Mixtureswith Respect to SCB GF

Modifier Mean (J∕m2) Rank

SBS 421.29 A

PPAþ Elvaloy 301.46 A/B

PPAþ SBS 288.36 A/B

PPA 246.34 B

Fig. 7. Results of IDT strength, SCB, and DCT tests for unaged, aged, and field samples

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 1095

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

The authors also acknowledge Dr. Eshan Dave for providing DCTtest results for all asphalt mixtures. The results and opinions pre-sented are those of the authors and do not necessarily reflect thoseof the sponsoring agencies.

References

AASHTO. (2003). “Standard method of test for determining the creep com-pliance and strength of hot-mix asphalt (HMA) using indirect tensiletest device.” T322-03, Washington, DC.

AASHTO. (2005). “Standard test method for the bulk specific gravity ofcompacted bituminous mixtures using saturated surface-dry speci-mens.” AASHTO T166, Washington, DC.

AASHTO. (2006). “Standard specification for mixture conditioning of hotmix asphalt (HMA).” AASHTO R30-02, Washington, DC.

Arnold, T. S., Youtcheff, J. S., Jr., and Needham, S. P. (2009). “Use ofphosphoric acid as modifier for hot-mix asphalt.” Polyphosphoric AcidModification of Asphalt Binders: A Workshop, Circular E-C160,Transportation Research Board, Washington, DC, 40–51.

ASTM. (2007). “Standard test method for determining fracture energy ofasphalt–aggregate mixtures using the disk-shaped compact tensiongeometry.” D7313-07, West Conshohocken, PA.

Bishara, S. W., Mahoney, D., and McReynolds, R. L. (2001). “Modificationof binder with acid: Advantages and disadvantages.” TRB SuperpaveAbstracts 2001: 80th Transportation Research Board Annual Meeting,Transportation Research Board, Washington, DC.

Clyne, T., Johnson, E., McGraw, J., and Reinke, G. (2009). “Fieldinvestigation of polyphosphoric acid modified binders at MnROAD.”Polyphosphoric Acid Modification of Asphalt Binders: A Workshop,

Circular E-C 160, 115–130.Dickinson, E. J. (1974). “The constitution and quality of paving grade

asphalts produced by air blowing distillation residues of Kuwait and‘light Arabian’ crude oil.” Conf. Proc., Association of Asphalt PavingTechnologists, Transportation Research Board, Washington, DC,132–161.

Fee, D., Maldonado, R., Reinke, G., and Romagosa, H. (2010). “Polyphos-phoric acid modification of asphalt.” Transp. Res. Record: J. Transp.Res. Board, 2179, 49–57.

Giavarini, C., Mastrofini, D., Scarsella, M., Barré, L., and Espinat, D.(2000). “Macrostructure and rheological properties of chemicallymodified residues and bitumens.” Energy Fuels, 14(2), 495–502.

Ho, S., Zanzotto, L., and MacLeod, D. (2001). “Impact of chemicalmodification on composition and properties of asphalt binders.” Proc.,Canadian Technical Asphalt Association, Vol. 46, Polyscience,Quebec, Canada, 153–170.

Kodrat, I., Sohn, D., and Hesp, S. A. M. (2007). “Comparison of polyphos-phoric acid-modified asphalt binders with straight and polymer-modifiedmaterials.” Transportation Research Record 1998, TransportationResearch Board, Washington, DC, 47–55.

Li, X., and Marasteanu, M. O. (2004). “Evaluation of the low temperaturefracture resistance of asphalt mixtures using the semi circular bend test.”J. Assoc. Asphalt Paving Technol., 73, 401–426.

Li, X., and Marasteanu, M. O. (2010). “Using semi circular bending test toevaluate low temperature fracture resistance for asphalt concrete.” Exp.Mech., 50(7), 867–876.

Maldonado, R., Falkiewicz, M., and Bazi, G. (2006). “Asphalt modificationwith polyphosphoric acid.” Proc., Canadian Technical Asphalt Associ-ation, 51st Annual Meeting, 163–178.

1096 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng. 2012.24:1089-1096.

Dow

nloa

ded

from

asc

elib

rary

.org

by

DA

YT

ON

, UN

IVE

RSI

TY

OF

on 0

8/13

/14.

Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.