ORIGINAL ARTICLE

Rheological analysis of asphalt binders modifiedwith Elvaloy� terpolymer and polyphosphoric acidon the multiple stress creep and recovery test

Matheus David Inocente Domingos •

Adalberto Leandro Faxina

Received: 21 August 2013 / Accepted: 13 December 2013

� RILEM 2013

Abstract This study presents the results of a series

of creep-recovery experiments that were conducted on

asphalt binders modified with polyphosphoric acid

(AC?PPA) and Elvaloy� terpolymer combined with

polyphosphoric acid (AC?Elvaloy?PPA) at the tem-

peratures of 52, 58, 64, 70 and 76 �C. The multiple

stress creep and recovery (MSCR) test was used to

determine the percent recoveries (R), the non-recov-

erable compliances (Jnr) and the percent differences in

non-recoverable compliances (Jnr, diff) of these mod-

ified asphalt binders and the ones of the 50/70-

penetration grade base material. The base material is

graded as PG 64-xx and the formulations with PPA

and Elvaloy?PPA are graded as PG 76-xx. Two pairs

of loading–unloading times were used in the MSCR

tests, namely, 1/9 s and 2/18 s. The AC?Elva-

loy?PPA shows not only the highest R values and

the lowest Jnr values, but also the lowest sensitivity to

a sudden increase in the stress level inside the asphalt

mixture. Although the results are not as promising as

the ones found in the AC?Elvaloy?PPA, the

AC?PPA may be taken as an alternative to replace

the unmodified asphalt binder when the loading and/or

the climate conditions are more severe.

Keywords Creep and recovery �Polyphosphoric acid � Elvaloy� terpolymer �Asphalt binders � Performance grade

Abbreviations

AASHTO American Association of state Highway

and Transportation Officials

AC Asphalt cement or asphalt binder

ASTM American Society for Testing and Materials

DSR Dynamic shear rheometer

FHWA United States Federal Highway

Administration

MSCR Multiple stress creep and recovery

PG Performance grade

PPA Polyphosphoric acid

RCRT Repeated creep and recovery test

RTFO Rolling thin-film oven

SARA Saturates, aromatics, resins and asphaltenes

1 Introduction

Modified asphalt binders are commonly used for

paving applications as an interesting alternative to deal

with severe traffic and environmental conditions on

roads and highways. This is accomplished by improv-

ing the rheological properties of the bituminous

material with the addition of one or more modifiers,

which in turn results in pavements with greater

resistance to rutting, fatigue cracking and thermal

M. D. Inocente Domingos (&) � A. L. Faxina

Department of Transportation Engineering, Sao Carlos

School of Engineering, University of Sao Paulo, Avenida

Trabalhador Sao-Carlense, 400, Parque Arnold Schimidt,

Sao Carlos, Sao Paulo 13566-590, Brazil

e-mail: matheusdavid@sc.usp.br

Materials and Structures

DOI 10.1617/s11527-013-0242-y

cracking and improved field performance [7, 20, 23,

24]. Certain degrees of improvement in the rheological

properties of the asphalt binder can also be achieved

by selecting better crude oils or tailoring the refinery

process, but these techniques are limited by the

number of starting crudes and actions that can be

carried out during the distillation of petroleum [7, 20].

Therefore, asphalt binder modification is still one of

the best alternatives to obtain better pavement

performance.

One of the modifiers that can be added to the asphalt

binder is the Elvaloy� reactive ethylene terpolymer,

which is commercialized by DuPontTM. The designa-

tion ‘‘reactive ethylene terpolymer’’ comes from (a) the

degree of reactivity of the material when blended with

the asphalt binder; (b) the main component of the

polymer chain, which is the ethylene molecule; and

(c) the presence of three components in this chain:

ethylene, butyl acrylate and glycidylmethacrylate [23].

It is believed that the glycidylmethacrylate molecule is

responsible for the reaction between Elvaloy� and the

asphalt binder when they are mixed at high tempera-

ture, and this reaction leads to the formation of a stable

asphalt-polymer system with enhanced properties [7,

22, 23]. To keep the forming polymer network below

the chemical gel point and avoid the risk of producing

an absolutely useless asphalt gel, the modifier content

usually varies from 1.5 to 2.5 % by weight [22, 23].

In addition to Elvaloy�, polyphosphoric acid (PPA)

can also be used as an asphalt binder modifier. PPA is a

medium strong acid with no free water and is typically

a mixture of other acids, i. e., pyrophosphoric acid,

triphosphoric acid and higher acids. It is a nonoxidant

compound and is highly soluble in organic compounds

[6]. PPA can be used in the formulation either alone or

in combination with another additive, which makes it

possible to reduce modification costs and also provides

flexibility in meeting the Superpave� specification

criteria such as the rotational viscosity at 135 �C and

the elastic recovery [8]. The addition of PPA has

marked effects on the high PG grade of the asphalt

binder; however, its impact on the low PG grade is null

or very small depending on the crude source [6, 8, 16].

The characterization of the resistance of asphalt

binders to the pavement distress mechanisms is

typically made by means of laboratory tests that are

supposed to adequately simulate the actual tempera-

ture and loading conditions in the pavement. With

respect to rutting, one notable advance was the

development of the repeated creep and recovery test

(RCRT) by Bahia et al. [5], in which subsequent

loading–unloading cycles at a predefined stress level

are applied on a 25-mm asphalt binder sample that is

sandwiched between the two parallel plates of a

dynamic shear rheometer (DSR) and the resulting

strain levels are continuously monitored. This test was

later refined by the United States Federal Highway

Administration (FHWA) through the introduction of a

new rutting parameter—the non-recoverable (creep)

compliance Jnr—and the addition of stress levels

ranging from 0.25 to 25.6 kPa in the same procedure

to determine the stress dependency of the material. It

was then renamed as multiple stress creep and

recovery test, or simply MSCR test [14].

Currently, the MSCR test is the most recent

innovation in the characterization of the resistance of

asphalt binders to rutting and in the study of the

rheological behavior of these materials at high tem-

peratures. Many researchers [2, 9, 11–14, 16, 17, 25]

examined the data obtained from this test and, among

other conclusions, they observed that the non-recov-

erable compliances of the asphalt binders have good

correlations with rutting measurements on asphalt

mixtures [2, 9, 12, 14, 17], that the MSCR test can

distinguish between formulations prepared with PPA

and one more additive and the corresponding ones

without PPA [9, 11, 12] and that the stress sensitivity

can vary significantly among the asphalt binders,

especially the modified ones [14, 25].

More recently, some research studies [1, 15, 18]

involved modifications in the standard MSCR test

protocol—stress levels of 0.1 and 3.2 kPa, 1-s creep

time, 9-s recovery time and ten loading–unloading

cycles at each stress level—due to some major

limitations. One of these limitations is the 9-s recovery

time used in the test, which may not be long enough to

allow full recovery of all modified asphalt binders,

especially the ones with high levels of delayed

elasticity [14, 15]. In addition, the two stress levels

used in the test do not necessarily reflect the actual

stresses of the bituminous material inside the pave-

ment structure. With respect to the creep–recovery

behavior of asphalt binders, extremely long recovery

times [15] and plasto-viscoelastic approaches com-

bined with nonlinear viscoelastic theories [19, 21]

were used by researchers to accurately identify the

unrecovered strain of modified asphalt binders at each

loading–unloading cycle.

Materials and Structures

In this paper, two different pairs of creep and

recovery times—1-s creep time and 9-s recovery time

(1/9 s) and 2-s creep time and 18-s recovery time

(2/18 s)—were used in the MSCR tests to quantify and

analyze the impact of a change in the loading–

unloading times on the rheological properties of

asphalt binders modified with PPA (AC?PPA) and

Elvaloy� terpolymer combined with PPA (AC?Elva-

loy?PPA). The results were compared with each other

at all test temperatures and stress levels, and the best

formulations were reported.

2 Materials and methods

2.1 Materials, preparation of samples and

short-term aging procedure

One base asphalt binder from the Replan-Petrobras

refinery (Paulinia, Sao Paulo, Brazil) was used to

prepare the AC?PPA and the AC?Elvaloy?PPA

formulations with the continuous grades of 77.8 and

81.2 �C, respectively. The unmodified material is

graded as 50/70 in the Brazilian penetration grade

specification and is graded as PG 64-xx (continuous

grade of 67.0 �C) in the revised version of the

Superpave� asphalt binder specification (AASHTO

M320-09, Table 3). The 4170 Elvaloy� terpolymer

was supplied by DuPontTM and the Innovalt� E200

PPA was supplied by Innophos Inc (US). Some typical

properties of this terpolymer include a melting point of

72 �C, density of 0.94 g/cm3 and a maximum pro-

cessing temperature of 280 �C.

The Elvaloy� and the PPA contents were chosen to

obtain modified asphalt binders with the same high-

temperature performance grade in the Superpave�

specification (PG 76-xx). This average 7-day maxi-

mum expected pavement temperature is 12 �C or two

grades above the one of the unmodified material. As

stated by AASHTO M320-09 in Table 3, the high PG

grade of the asphalt binder is the one at which the

parameter G*/sin d (G star divided by sine delta) is

higher than or equal to 1.0 kPa in the unaged

condition. The AC?PPA and the AC?Elvaloy?PPA

were prepared on a Fisatom 722D low-shear mixer.

Table 1 gives the modifier contents, the processing

variables and the results of some conventional binder

tests—ring-and-ball softening point (ASTM D36-06),

penetration at 25 �C (ASTM D5-06) and rotational

viscosity at 135 �C (ASTM D4402-06)—that were

carried out to first characterize the rheological prop-

erties of the unaged and short-term aged materials.

Short-term aging of asphalt binders was performed

on the rolling thin-film oven (RTFO) by following the

Table 1 Variables of the formulations and results of some conventional binder tests

Variable or rheological property Base binder (AC) AC?PPA AC?Elvaloy?PPA

Asphalt binder content (percentage by mass) 100.0 98.8 98.4

PPA content (percentage by mass) – 1.2 0.3

Elvaloy� content (percentage by mass) – – 1.3

Continuous grade (�C) 67.0 77.8 81.2

Mixing temperature (�C) – 130 190

Mixing time (min) – 30 120a

Rotation speed (rpm) – 300 300

Penetration at 25 �C, unaged (dmm)b 58.0 36.5 52.0

Penetration at 25�, short-term aged (dmm)b 30.8 23.8 31.8

Softening point, unaged (�C)b 49.4 56.8 63.6

Softening point, short-term aged (�C)b 56.1 67.2 70.9

Rotational viscosity at 135 �C, unaged (Pa s)c, d 0.36 0.72 1.71

Rotational viscosity at 135 �C, short-term aged (Pa s)c, d 0.59 1.94 3.67

a The polyphosphoric acid was added to the AC?Elvaloy after 60 min of mixing timeb Results are the average of four replicatesc Results are the average of two replicates and ten values were obtained at each single testd Viscosity measurements were made with the spindle 21

Materials and Structures

procedures established in the ASTM D2872-04 stan-

dard. In this test, samples of 35 ± 0.5 g are poured

into standard cylindrical glass bottles and placed in a

rolling oven at 163 �C for 85 min. Then, these

samples are cooled at room temperature and the

differences between their masses before and after

short-term aging are used to calculate the mass loss

ML. The ML values are shown in Table 2 together with

the results of other aging indexes. As it can be seen, the

effects of volatilization of the light fractions of the

binder were more pronounced than the effects of

oxidation for the unmodified and modified materials,

since all the ML values are negative. The results also

suggest that the AC?PPA is the most sensitive

formulation to short-term aging (highest increase in

softening point, viscosity aging index and mass loss

values), followed by the AC?Elvaloy?PPA and the

50/70 unmodified asphalt binder. In other words,

asphalt binder modification had a marked effect not

only on the rheological properties of the base material,

but also on its sensitivity to aging.

2.2 Multiple stress creep and recovery tests

The MSCR tests (ASTM D7405-10a) were conducted

on an AR-2000ex DSR supplied by TA Instruments.

Two pairs of creep and recovery times—1/9 s and

2/18 s—were used in the tests and two replicates were

performed for each short-term aged asphalt binder at

the temperatures of 52, 58, 64, 70 and 76 �C. The

equations written in Fig. 1 were used to calculate the

percent recoveries R and the non-recoverable (creep)

compliances Jnr at each temperature and stress level,

and their final results were determined by means of the

average of these two replicates. Although there are

recommendations for performing MSCR tests up to

70 �C based on the climatic conditions of the US [4],

the temperature of 76 �C was selected by the authors

because it is representative of the maximum expected

pavement temperature that can be observed in some

regions of Brazil. With exception of the creep and

recovery times, the other testing variables (stress

levels of 0.1 and 3.2 kPa, 10 loading–unloading cycles

at each stress level and the five above-mentioned

temperatures) remained the same in all MSCR tests.

The stress sensitivity of the asphalt binders was

evaluated by means of the percent differences in non-

recoverable compliances (Jnr, diff). This parameter

shows the percentage of increase in the Jnr value of the

asphalt binder when the stress level is increased from

0.1 to 3.2 kPa, as it can be seen in the equation in

Fig. 1. In practical terms, it evaluates the susceptibil-

ity of the asphalt binder to rutting when unexpected

heavy traffic loadings are applied on the pavement

structure or unusually high temperatures are observed

in the field [3, 10]. The Jnr, diff values were determined

at all MSCR test temperatures, and then they were

compared with the upper limit of 75 % found in the

Superpave� asphalt binder specification. This limit

was set by the specification with the aim of discon-

sidering materials that are overly stress sensitive and

potentially susceptible to rutting, even though the

other PG criteria are met [3].

2.3 Percent recovery and non-recoverable

compliance ratios

Numerical ratios were used to evaluate the effects of

longer creep and recovery times on two of the final

outcomes of the MSCR test (R and Jnr). The variations

in the R values were analyzed by means of the ratio of

the percent recovery of the asphalt binder at 1/9 s to

the one at 2/18 s, and therefore the percent recovery

ratio RP was obtained. The variations in the Jnr values

were analyzed by means of the ratio of the non-

recoverable compliance of the asphalt binder at 2/18 s

to the one at 1/9 s yielding the non-recoverable

compliance ratio RC. These two parameters were

calculated for all asphalt binders at each test temper-

ature and stress level.

Three distinct cases can be identified on the MSCR

test when the creep and recovery times are increased

from 1/9 s to 2/18 s. In the first case, the percent

Table 2 Aging indexes

Aging index Base

binder

(AC)

AC?PPA AC?

Elvaloy?

PPA

Retained penetrationa (%) 53.0 65.1 61.1

Increase in softening

pointb (�C)

6.8 10.4 7.2

Viscosity aging indexc 1.63 2.68 2.15

Mass loss (%) -0.1094 -0.2263 -0.0421

a Penetration after aging divided by the one before agingb Softening point before aging subtracted from the one after

agingc Rotational viscosity after aging divided by the one before

aging

Materials and Structures

recoveries are higher and the non-recoverable com-

pliances are lower at 2/18 s than at 1/9 s, which

indicates that the RP and the RC values are lower than

one. The second case is the opposite of the first one,

i. e., the percent recoveries are lower and the non-

recoverable compliances are higher at 2/18 s than at

1/9 s (the RP and the RC values are greater than one). In

the third and last case, the percent recoveries and the

non-recoverable compliances do not significantly

change when the loading–unloading times are

increased (the RP and the RC values are approximately

equal to one). Since higher R values and lower Jnr

values are favorable to the resistance of the asphalt

binder to rutting, it is highly desirable to obtain RP and

RC values lower than one for all materials.

3 Analysis and discussion of results

3.1 Multiple stress creep and recovery tests

at 1/9 s

Figure 2 displays the percent recoveries of the asphalt

binders at 1-s creep time and 9-s recovery time. The

addition of PPA and Elvaloy?PPA to the 50/70 base

material increased its R values at common high

pavement temperatures, especially for the AC?Elva-

loy?PPA. Higher percent recoveries indicate that the

asphalt binder can recover a higher portion of its total

strain at the end of each loading–unloading cycle,

which is favorable to the resistance of the material to

rutting. In rheological terms, it is also possible to

increase the percent recovery of some modified

materials by increasing the recovery time until no

significant variations in the unrecovered strain are

observed anymore. This technique was applied in a

research study carried out by Delgadillo et al. [15],

according to which a full recovery of the Elvaloy-

modified asphalt binders was possible only when

recovery times of more than 1,000 s were used after a

creep time of 1 s in the MSCR test.

The percent recoveries are all between 60 and 81 %

for the AC?Elvaloy?PPA and are no greater than

64 % for the AC?PPA. These R values are non-null

for the unmodified asphalt binder only at lower test

temperatures, namely, up to 64 �C at 0.1 kPa and up to

58 �C at 3.2 kPa. Good results can be observed for the

AC?Elvaloy?PPA in the whole temperature range

and at both stress levels, and this may be explained by

the formation of a stable asphalt-polymer system in the

formulation after the addition of the Elvaloy� ter-

polymer [7, 22, 23]. Although major increases in the

R values were experienced by the AC?PPA, no

considerable variations were found at 76 �C and

3.2 kPa. Two possible reasons may be set out to

explain this surprising finding: (a) the quantity of PPA

was not sufficient to significantly increase the percent

recoveries of the material in such test conditions, even

though the PG grade of 76-xx was achieved; or (b) the

individual properties of the modifier and the ones of

the unmodified asphalt binder led to this result.

The non-recoverable compliances of the asphalt

binders at 1-s creep time and 9-s recovery time are

shown in Fig. 3. The effect of the addition of modifiers

on the Jnr values of the asphalt binders is the opposite of

the one observed for the R values, i. e., there were

0.00E+00

2.00E-01

4.00E-01

6.00E-01

8.00E-01

1.00E+00

1.20E+00

1.40E+00

0.00E+00

2.00E+00

4.00E+00

6.00E+00

8.00E+00

1.00E+01

1.20E+01

1.40E+01

1.60E+01

1.80E+01

2.00E+01

0 5 10 15 20 25 30 35 40 45 50

Acc

umul

ated

Stra

inat

100

Pa

Acc

umul

ated

Stra

inat

3,20

0 P

a

Time (seconds)

Strain Data - 3,200 Pa

Strain Data - 100 Pa

Initial Strain(ε0)

MaximumStrain (εC)

Residual Strain (εR)

100

100

Fig. 1 Plot of the MSCR

test and its final outcomes

Materials and Structures

marked reductions in Jnr under all MSCR test condi-

tions when PPA or Elvaloy?PPA was added to the

base material. The results are much better for the

AC?Elvaloy?PPA than for the AC?PPA at temper-

atures higher than or equal to 64 �C, and the ones of the

base asphalt binder can overcome 5.0 kPa-1 at 70 and

76 �C. The non-recoverable compliances of the

AC?Elvaloy?PPA are all lower than 1.0 kPa-1 and

the ones of the AC?PPA are no greater than 2.4 kPa-1.

As previously observed for the percent recovery,

lower non-recoverable compliances also suggest that

the modified asphalt binders are less prone to rutting

after the application of loading–unloading cycles at

high pavement temperatures. In terms of the asphalt

mixture, asphalt binders with lower Jnr values and/or

higher R values will less contribute to the

accumulation of unrecovered strain in the asphalt

layer. From the point of view of rheology, these lower

Jnr values may be obtained by decreasing the amount

of unrecoverable strain at the end of the creep-

recovery cycle, for a particular stress level. Since the

stress levels are the same for all MSCR tests, it can be

inferred that the amount of permanent strain is lower

for the modified asphalt binders than for the unmod-

ified one.

In addition to the reduction in the non-recoverable

compliances, it can be observed that the Jnr values are

fairly similar for the two modified asphalt binders at

the temperatures of 52 and 58 �C. This similarity also

exists between the non-recoverable compliances of the

AC?Elvaloy?PPA at 0.1 and 3.2 kPa in the whole

temperature range. With respect to the results at the

12.4

5.8

1.0

0.0

0.0

8.9

0.5

0.0

0.0

0.0

63.7

55.2

44.6

34.1

24.0

62.4

49.8

31.3

12.1

0.8

80.3

79.8

77.1

72.5

66.0

80.7

79.8

77.2

71.4

60.5

0

20

40

60

80

100

120

52 58 64 70 76

Per

cen

t R

eco

very

(%

)

Temperature (°C)

0.1 kPa - base binder (AC) 3.2 kPa - base binder (AC) 0.1 kPa - AC+PPA

3.2 kPa - AC+PPA 0.1 kPa - AC+Elvaloy+PPA 3.2 kPa - AC+Elvaloy+PPA

Fig. 2 Percent recoveries of asphalt binders at 1-s creep time and 9-s recovery time

0.33

0

0.92

0 2.43

5

5.95

5

13.5

25

0.35

0

1.02

0 2.77

5

6.79

0

15.2

30

0.03

0

0.08

0

0.22

5

0.58

5

1.47

5

0.03

0

0.09

0

0.28

5

0.85

5 2.38

0

0.03

0

0.07

0

0.14

5

0.31

0

0.67

5

0.03

0

0.06

5

0.13

5

0.28

5

0.65

5

0

3

6

9

12

15

18

52 58 64 70 76

No

n-R

eco

vera

ble

Co

mp

lian

ce (

kPa-

1 )

Temperature (°C)

0.1 kPa - base binder (AC) 3.2 kPa - base binder (AC)

0.1 kPa - AC+PPA 3.2 kPa - AC+PPA

0.1 kPa - AC+Elvaloy+PPA 3.2 kPa - AC+Elvaloy+PPA

Fig. 3 Non-recoverable

compliances of asphalt

binders at 1-s creep time and

9-s recovery time

Materials and Structures

lowest temperatures, they may be explained by the

very low strain levels found in the modified asphalt

binders under such test conditions, and therefore a

clear distinction between the creep and recovery

responses of these materials cannot be accurately

drawn. As the temperature and the strain level

increase, this distinction can be more easily made

and the rheological response will be dependent on the

modifiers that were added to the asphalt binder. The Jnr

values of the AC?Elvaloy?PPA indicate that an

increase in the stress level from 0.1 to 3.2 kPa did not

greatly affect its rheological response, and the same

can be said for the R values (Fig. 2).

3.2 Multiple stress creep and recovery tests

at 2/18 s

Figure 4 shows the percent recoveries of the unmod-

ified asphalt binder, the AC?PPA and the AC?Elva-

loy?PPA at 2-s creep time and 18-s recovery time.

Again, the 50/70 base material experienced substantial

increases in its R values at both stress levels after the

addition of PPA (test temperatures up to 64 �C) or

Elvaloy?PPA (whole temperature range), and the

differences between the values at 0.1 and 3.2 kPa are

greater for the AC?PPA than for the AC?Elva-

loy?PPA. These results suggest that, even when the

loads are applied for a longer period of time (longer

creep time) and they are more spaced in time (longer

recovery time), the presence of modifiers can still be

detected in the MSCR test. Although the recovery time

increased from 9 to 18 s, it is not possible to say that

the percent recoveries in Fig. 4 are equal to the

maximum ones for the asphalt binders because the

creep time is also longer in the MSCR test (2 s instead

of 1 s) and recovery times other than 9 and 18 s were

not investigated in the study.

With respect to the numerical values, it can be seen

that they are all lower than 8, 60 and 81 % for the base

asphalt binder, the AC?PPA and the AC?Elva-

loy?PPA, respectively. There is an approximately

linear decrease in the results of the AC?PPA at 0.1

and 3.2 kPa with increasing temperature, and this is

also valid for the ones of the AC?Elvaloy?PPA at

0.1 kPa. The percent recoveries of the AC?Elva-

loy?PPA are all high within the temperature interval

considered in the study (between 50 and 81 %), and

the differences between the values at 0.1 and 3.2 kPa

are relatively small. As previously found in the creep

and recovery times of 1/9 s, no variations were

observed for the percent recovery of the AC?PPA at

76 �C and 3.2 kPa, and this could be explained by an

insufficient quantity of modifier or by the individual

characteristics of the components of the formulation

(asphalt binder and PPA).

Figure 5 presents the non-recoverable compliances

of the AC?PPA, the AC?Elvaloy?PPA and the base

asphalt binder at 0.1 and 3.2 kPa and the five test

temperatures considered in the study. As with the 1-s

creep time and the 9-s recovery time, major reductions

in the Jnr values can be noticed when PPA or

Elvaloy?PPA is added to the asphalt binder, and this

is especially remarkable for the AC?Elvaloy?PPA.

This means that, although both formulations have

lower non-recoverable compliances than the 50/70

unmodified material, the results are better for the

7.9

2.9

0.0

0.0

0.03.

4

0.0

0.0

0.0

0.0

59.4

50.2

38.8

27.1

17.7

56.4

40.0

19.0

3.3

0.0

80.0

79.5

75.5

69.0

61.2

80.5

78.9

74.6

65.6

50.7

0

20

40

60

80

100

120

52 58 64 70 76

Per

cen

t R

eco

very

(%

)

Temperature (°C)

0.1 kPa - base binder (AC) 3.2 kPa - base binder (AC) 0.1 kPa - AC+PPA

3.2 kPa - AC+PPA 0.1 kPa - AC+Elvaloy+PPA 3.2 kPa - AC+Elvaloy+PPA

Fig. 4 Percent recoveries

of asphalt binders at 2-s

creep time and 18-s recovery

time

Materials and Structures

AC?Elvaloy?PPA than for the AC?PPA. By taking

into account only the variables that are related to the

asphalt binder, asphalt mixtures prepared with the

AC?Elvaloy?PPA will show lower susceptibility to

rutting than the ones prepared with the AC?PPA.

The differences between the Jnr values of the

AC?Elvaloy?PPA and the AC?PPA are relatively

small at temperatures lower than or equal to 64 �C,

and they become significant at 70 and 76 �C. This may

be explained by the extremely low strain levels found

in the two formulations at test temperatures much

lower than their high-temperature performance grade

of 76 �C. As the temperature and the strain level

increase, the effects of the addition of the modifiers

(PPA or Elvaloy?PPA) are more apparent. The non-

recoverable compliances of the AC?Elvaloy?PPA at

0.1 kPa do not considerably differ from the ones at

3.2 kPa within the temperature interval of 52–76 �C,

and the same can be said for the percent recoveries of

the material at temperatures up to 70 �C (Fig. 4). This

is not observed for the AC?PPA at higher tempera-

tures (70 and 76 �C), at which the non-recoverable

compliances can differ by more than 50 % when the

stress level is increased from 0.1 to 3.2 kPa.

3.3 Stress sensitivity of the asphalt binders

Table 3 depicts the results of the percent differences in

non-recoverable compliances (Jnr, diff) of the base

asphalt binder, the AC?PPA and the AC?Elva-

loy?PPA at the creep and recovery times of 1/9 s

and 2/18 s. Mathematically, the parameter Jnr, diff

shows the percentage of increase in the non-

recoverable compliance of the material when the

stress level is increased from 0.1 to 3.2 kPa at a

predefined temperature. This parameter is used as an

indicator of the stress sensitivity of the asphalt binder,

and the Superpave� specification (AASHTO M320-

09, Table 3) sets a maximum value of 75 % for

materials tested at the high PG grade, the creep and

recovery times of 1/9 s and the stress level of 3.2 kPa.

According to this specification, asphalt binders with

Jnr, diff values higher than 75 % are too stress sensitive

and highly susceptible to the accumulation of perma-

nent (or unrecovered) strain in the field when the

climate and/or the loading conditions are adverse, and

0.63

2

1.75

4 4.69

0

11.7

05

26.4

87

0.67

9

2.00

8 5.42

7

13.3

35

30.1

36

0.05

2

0.14

5

0.40

6

1.10

0

2.78

0

0.05

5

0.17

4

0.56

6

1.74

0 4.77

0

0.05

3

0.10

5

0.23

2

0.53

1

1.18

8

0.05

1

0.10

2

0.21

5

0.49

1

1.20

0

0

5

10

15

20

25

30

35

52 58 64 70 76

No

n-R

eco

vera

ble

Co

mp

lian

ce(k

Pa-

1 )

Temperature (°C)

0.1 kPa - base binder (AC) 3.2 kPa - base binder (AC)

0.1 kPa - AC+PPA 3.2 kPa - AC+PPA

0.1 kPa - AC+Elvaloy+PPA 3.2 kPa - AC+Elvaloy+PPA

Fig. 5 Non-recoverable

compliances of asphalt

binders at 2-s creep time and

18-s recovery time

Table 3 Percent differences in non-recoverable compliances

(Jnr, diff, %) of asphalt binders

Temperature and

loading–unloading

condition

Base

binder

(AC)

AC?PPA AC?

Elvaloy?

PPA

52 �C, 1/9 sa 6.1 0.0 0.0

58 �C, 1/9 s 10.9 12.5 -7.1

64 �C, 1/9 s 14.0 26.7 -6.9

70 �C, 1/9 s 14.0 46.2 -8.1

76 �C, 1/9 s 12.6 61.4 3.0

52 �C, 2/18 sb 7.5 5.6 -4.0

58 �C, 2/18 s 14.5 20.0 -2.2

64 �C, 2/18 s 15.7 39.5 -7.6

70 �C, 2/18 s 13.9 58.2 -7.4

76 �C, 2/18 s 13.8 71.6 1.0

a 1/9 s = 1-s creep time and 9-s recovery timeb 2/18 s = 2-s creep time and 18-s recovery time

Materials and Structures

therefore their use for paving applications is not

recommended at the high PG grade [3].

As shown in Table 3, the highest Jnr, diff values are

found in the AC?PPA and the lowest ones are found

in the AC?Elvaloy?PPA at several temperatures and

stress levels. This indicates that the AC?Elva-

loy?PPA is less prone to rutting than the AC?PPA

and the unmodified asphalt binder when unusually

high pavement temperatures and/or unforeseen traffic

loadings are observed in the field. None of the Jnr, diff

values exceeded the upper limit of 75 % set by the

Superpave� specification at the high-temperature

performance grades of 64 �C (50/70-penetration grade

base material) and 76 �C (modified asphalt binders),

and the one which got closer to this limit is the

AC?PPA at both pairs of loading–unloading times:

61.4 % at 1/9 s and 71.6 % at 2/18 s. The results vary

from 7 to 16 % for the unmodified material and are no

greater than 10 % in modulus for the AC?Elva-

loy?PPA. On the other hand, substantial increases in

these percent differences are observed for the

AC?PPA with increasing temperature, either at

1/9 s or 2/18 s.

A careful analysis of the Jnr, diff values of asphalt

binders shows that slight variations in this parameter

are observed for the 50/70 base asphalt binder and the

AC?Elvaloy?PPA as the temperature increases. In

other words, the effect of temperature on the stress

sensitivity of the AC?Elvaloy?PPA and the unmod-

ified material is not as profound as the one found in the

AC?PPA. The AC?Elvaloy?PPA also shows nega-

tive Jnr, diff values at many temperatures and for both

pairs of loading–unloading times, but these values are

all close to zero. A negative result for this parameter

indicates that the non-recoverable compliance

decreased with increasing stress level from 0.1 to

3.2 kPa, i. e., the asphalt binder is less susceptible to

rutting at 3.2 kPa than at 0.1 kPa.

Despite the presence of some negative values for

the percent differences in non-recoverable complianc-

es (Jnr, diff \ 0) of the AC?Elvaloy?PPA, this curious

phenomenon cannot be interpreted as a decrease in the

susceptibility of the asphalt binder to rutting at higher

stress levels (lower Jnr values). This may be explained

by factors such as the following: (a) the variability in

the results of each single laboratory test, all of them

within a margin of tolerance; and (b) the observation

of a fairly short interval of results that includes

positive, null and negative Jnr, diff values, which makes

it difficult to draw a reasonable conclusion about the

findings. What is possible to say is that the formation

of a stable asphalt-polymer system in the AC?Elva-

loy?PPA after the addition of Elvaloy� terpolymer

and PPA contributed—at least in some extent—to the

reduction in the influence of the stress level and the

temperature on the resistance of the material to the

accumulation of unrecoverable (or permanent) strain.

3.4 Percent recovery and non-recoverable

compliance ratios

The percent recovery ratios RP, i. e., ratios of the

percent recoveries of the asphalt binder at 1/9 s to the

ones at 2/18 s, are summarized in Table 4. Some of

these ratios could not be found because the percent

recovery of the material at 2-s creep time and 18-s

recovery time is equal to zero. As the results show, the

increase in both creep and recovery times led to a

reduction in the R values of the unmodified and

modified asphalt binders (RP [ 1.0), especially for the

AC?PPA and the 50/70 base material. On the other

hand, this increase slightly affected the R values of the

formulation with Elvaloy� terpolymer and PPA, since

the RP values are all lower than 1.10 for the stress level

of 0.1 kPa and are no greater than 1.20 for the stress

level of 3.2 kPa. This means that the application of

repeated traffic loads for a longer period of time

(longer creep time) and more spaced in time (longer

recovery time) will not cause a huge impact on the

Table 4 Percent recovery ratios (RP)

Temperature

and stress

level (kPa)

Base

binder

(AC)

AC?PPA AC?

Elvaloy?

PPA

52 �C, 0.1 1.57a 1.07 1.00

58 �C, 0.1 1.99 1.10 1.00

64 �C, 0.1 NC 1.15 1.02

70 �C, 0.1 NC 1.26 1.05

76 �C, 0.1 NC 1.35 1.08

52 �C, 3.2 2.62 1.10 1.00

58 �C, 3.2 NC 1.24 1.01

64 �C, 3.2 NC 1.65 1.04

70 �C, 3.2 NC 3.67 1.09

76 �C, 3.2 NC NC 1.19

NC not possible to be calculateda RP value = ratio of the percent recovery at 1/9 s to the one at

2/18 s

Materials and Structures

percent recovery of the AC?Elvaloy?PPA when

compared with the base material and the AC?PPA.

Although the RP values of the AC?PPA are higher

than the ones of the AC?Elvaloy?PPA, they are

lower than the ones of the unmodified asphalt binder at

the temperature of 52 �C. The results vary from 1.0 to

3.7 for the formulation with PPA and are equal to 1.99

and 2.62 for the 50/70 base binder at the stress levels of

0.1 and 3.2 kPa, respectively. This means that the

percent recoveries of the AC?PPA at 2/18 s are up to

3.7 times lower than the ones obtained at 1/9 s.

Similarly, the percent recoveries of the unmodified

material at 52 �C and 1/9 s are 1.99 and 2.62 times

higher than the ones at 52 �C and 2/18 s for the stress

levels of 0.1 and 3.2 kPa, respectively. With respect to

the AC?Elvaloy?PPA, the percent recoveries of the

formulation at 1/9 s are up to 20 % greater than the

ones at 2/18 s at all MSCR test temperatures.

The non-recoverable compliance ratios RC, i. e.,

ratios of the non-recoverable compliance of the

asphalt binder at 2/18 s to the one at 1/9 s, are given

in Table 5. The effect of longer creep and recovery

times on the Jnr values of the asphalt binders is the

opposite of the one observed for the R values, that is,

the non-recoverable compliances are greater at 2/18 s

than at 1/9 s (RC [ 1.0) for all materials. These

increases are fairly bigger for the unmodified asphalt

binder (RC values between 1.9 and 2.0) at many

temperatures and stress levels, followed by the

AC?PPA (RC values between 1.7 and 2.1) and the

AC?Elvaloy?PPA (RC values between 1.5 and 1.9).

The results indicate that the addition of PPA or

Elvaloy?PPA to the asphalt binder diminished the

impact of longer creep and recovery times on the Jnr

values of the material at high pavement temperatures.

Other than highlighting the effects of the addition

of different types of modifiers on the creep and

recovery properties of the asphalt binder, the results

shown in Tables 4 and 5 reveal that the AC?Elva-

loy?PPA has the lowest sensitivity to the increase in

both creep and recovery times (lowest RP and RC

values), and that the base material has the highest ones

(highest RP and RC values) under several MSCR test

conditions. The addition of PPA alone also reduced the

impact of these longer creep and recovery times on

two of the outcomes of the test (R and Jnr), but the rate

of reduction is lower than the one found in the

AC?Elvaloy?PPA. By analyzing the results of the

parameters RP and RC, two conclusions can be

reached: (a) the AC?Elvaloy?PPA is the best

formulation among the three asphalt binders studied

in the paper; and (b) the AC?PPA may be taken as a

possible alternative to replace the unmodified material

when the loading–unloading conditions are a crucial

factor.

4 Summary and conclusions

In this study, a 50/70-penetration grade base asphalt

binder (PG 64-xx) was modified with polyphosphoric

acid—PPA (AC?PPA) and Elvaloy� terpolymer

combined with PPA (AC?Elvaloy?PPA) to obtain

formulations with the same high PG grade (PG 76-xx)

in the Superpave� specification. Two creep and

recovery times—1/9 s and 2/18 s—were used on the

multiple stress creep and recovery tests (MSCR tests)

to perform a rheological analysis of the materials at

temperatures ranging from 52 to 76 �C. The following

conclusions can be reached with respect to the creep-

recovery behavior of these asphalt binders:

• The impact of the addition of one (PPA) or two

modifiers (Elvaloy?PPA) to the base asphalt

binder was highly beneficial to the percent recov-

eries (R) and the non-recoverable compliances

(Jnr) of the material at typical high pavement

temperatures, especially for the AC?Elva-

loy?PPA; in a more practical and simplistic

approach, asphalt mixtures prepared with the

Table 5 Non-recoverable compliance ratios (RC)

Temperature

and stress

level (kPa)

Base

binder

(AC)

AC?PPA AC?

Elvaloy?

PPA

52 �C, 0.1 1.91a 1.73 1.76

58 �C, 0.1 1.91 1.81 1.50

64 �C, 0.1 1.93 1.80 1.60

70 �C, 0.1 1.97 1.88 1.71

76 �C, 0.1 1.96 1.89 1.76

52 �C, 3.2 1.94 1.82 1.69

58 �C, 3.2 1.97 1.93 1.58

64 �C, 3.2 1.96 1.99 1.59

70 �C, 3.2 1.96 2.04 1.72

76 �C, 3.2 1.98 2.00 1.83

a RC value = ratio of the non-recoverable compliance at

2/18 s to the one at 1/9 s

Materials and Structures

AC?PPA or the AC?Elvaloy?PPA would be less

susceptible to the appearance of rutting than the

ones prepared with the unmodified binder;

• The AC?Elvaloy?PPA shows not only higher

percent recoveries and lower non-recoverable

compliances than the AC?PPA, but also lower

values for the percent differences in non-recover-

able compliances (parameter Jnr, diff), the percent

recovery ratios (RP) and the non-recoverable

compliance ratios (RC) under the MSCR test

conditions used in the paper; this can be translated

into a less stress sensitive formulation and into a

reduction in the susceptibility of the modified

material to rutting at longer loading–unloading

times;

• Although the Jnr values of the AC?PPA at 76 �C

and 3.2 kPa are much lower than the ones of the

base asphalt binder, no signs of improvement in the

percent recoveries of the modified material could

be observed in these same test conditions; this may

be explained by the insufficient quantity of mod-

ifier (even though the high PG grade of 76 �C was

achieved) or by the individual characteristics of the

components of the formulation (SARA fractions of

the asphalt binder and the grade of PPA); and

• The results of the parameter Jnr, diff include positive,

null and negative values for the AC?Elvaloy?PPA

within the temperature interval from 52 to 76 �C;

however, the rutting performance of the Elvaloy-

modified asphalt binder is not inherently better at

3.2 kPa than at 0.1 kPa in some test conditions due

to the wide variety of Jnr, diff values and the natural

variability of the results of each laboratory test.

What is possible to say is that the stable-asphalt

polymer system in the AC?Elvaloy?PPA contrib-

uted—at least in some extent—to the reduction in

the stress sensitivity of the asphalt binder.

The use of different creep and recovery times in the

MSCR test is an attempt to better simulate the actual

loading–unloading conditions that can be observed in

the field. As it is known, the vehicles do not travel at

the same speed and do not apply equivalent loads on

the pavement structure. In addition, the time interval

between vehicles—which is referred to as ‘‘gap’’ in the

technical literature—is not necessarily the same

throughout the day. Although the present study was

restricted to only two pairs of creep and recovery

times, the results support the idea that the analysis of

the creep-recovery behavior of modified asphalt

binders is a key feature to accurately predict the

response of the bituminous material in a real pave-

ment, as well as its contribution to the resistance of the

asphalt mixture to rutting in different loading and/or

climate conditions.

Acknowledgments The first author acknowledges the

Brazilian Federal Research Agency (CAPES) for providing a

scholarship. The second author gratefully thanks the Research

Agency of the Sao Paulo State (FAPESP) for providing financial

funds (FAPESP process number 2006/55835-6).

References

1. Adorjanyi K, Fuleki P (2011) Performance evaluation of

bitumens at high temperature with multiple stress creep

recovery test. Hung J Ind Chem 39(2):195–199

2. Anderson RM (2011) Understanding the MSCR test and its

use in the PG asphalt binder specification. Asphalt Academy

Webinars, Lexington. http://www.asphaltinstitute.org/dotAsset/

48a2c0b0-d756-4b8c-913f-e7aaf91ef03d.pdf. Accessed 25

Jul 2013

3. Anderson M, D’Angelo J, Walker D (2010) MSCR: a better

tool for characterizing high temperature performance

properties. Asphalt—The Magazine of the Asphalt Institute,

Lexington. http://www.asphaltmagazine.com/news/detail.

dot?id=d90e7ce8-f127-4617-ac1d-7d788e2df710. Acces-

sed 28 Jul 2013

4. Asphalt Institute (2010) Guidance on the use of the MSCR

test with the AASHTO M320 specification. Asphalt Insti-

tute Technical Advisory Committee, Lexington. http://

www.asphaltinstitute.org/dotAsset/9007b936-5596-456f-

844b-2fc317700873.pdf. Accessed 28 Jul 2013

5. Bahia HU, Zhai H, Zeng M, Hu Y, Turner P (2001)

Development of binder specification parameters based on

characterization of damage behavior. J Assoc Asph Paving

Technol 70:442–470

6. Baumgardner GL (2012) Why and how of polyphosphoric

acid modification: an industry perspective. Transp Res Circ

E-C160:14–26

7. Becker Y, Mendez MP, Rodrıguez Y (2001) Polymer

modified asphalt. Vis Technol 9(1):39–50

8. Buncher M, D’Angelo J (2012) Establishing a baseline of

knowledge through 2005 by reviewing AI IS-220: poly-

phosphoric acid modification of asphalt. Transp Res Circ

E-C160:12–13

9. D’Angelo JA (2009) The relationship of the MSCR test to

rutting. Road Mater Pavement Des 10(Suppl 1):61–80.

doi:10.1080/14680629.2009.9690236

10. D’Angelo J (2010) The multiple stress creep recovery

(MSCR) procedure. Federal Highway Administration,

Washington. http://www.fhwa.dot.gov/pavement/materials/

pubs/hif11038/hif11038.pdf. Accessed 28 Jul 2013

11. D’Angelo J (2010) New high-temperature binder specifi-

cation using multistress creep and recovery. Transp Res

Circ E-C147:1–13

Materials and Structures

12. D’Angelo JA (2012) Effect of polyphosphoric acid on

asphalt binder properties. Transp Res Circ E-C160:27–39

13. D’Angelo J, Dongre R (2009) Practical use of multiple

stress creep and recovery test: characterization of styrene–

butadiene–styrene dispersion and other additives in poly-

mer-modified asphalt binders. Transp Res Rec 2126:73–82.

doi:10.3141/2126-09

14. D’Angelo J, Kluttz R, Dongre R, Stephens K, Zanzotto L

(2007) Revision of the superpave high temperature binder

specification: the multiple stress creep recovery test. J Assoc

Asph Paving Technol 76:123–162

15. Delgadillo R, Bahia HU, Lakes R (2012) A nonlinear con-

stitutive relationship for asphalt binders. Mater Struct

45(3):457–473. doi:10.1617/s11527-011-9777-y

16. Fee D, Maldonado R, Reinke G, Romagosa H (2010)

Polyphosphoric acid modification of asphalt. Transp Res

Rec 2179:49–57. doi:10.3141/2179-06

17. Gierhart D (2013) Multiple stress creep recovery (MSCR):

why it should be implemented. Louisiana Transportation

Conference, Baton Rouge. http://www.ltrc.lsu.edu/ltc_13/

pdf/presentations/S43_Why%20the%20Multiple%20Stress

%20Creep%20Recovery%20(MSCR)%20Test%20Should

%20be%20Implemented_LTC2013.pdf. Accessed 25 Jul

2013

18. Golalipour A (2011) Modification of multiple stress creep

and recovery test procedure and usage in specification.

Thesis, University of Wisconsin-Madison, Madison

19. Huang C-W (2008) Development and numerical imple-

mentation of nonlinear viscoelastic–viscoplastic model for

asphalt materials. Dissertation, Texas A&M University,

College Station

20. Kalantar ZN, Karim MR, Mahrez A (2012) A review of

using waste and virgin polymer in pavement. Constr Build

Mater 33:55–62. doi:10.1016/j.conbuildmat.2012.01.009

21. Masad E, Huang C-W, D’Angelo J, Little D (2009) Charac-

terization of asphalt binder resistance to permanent defor-

mation based on nonlinear viscoelastic analysis of multiple

stress creep recovery (MSCR) test. J Assoc Asph Paving

Technol 78:535–566

22. Polacco G, Stastna J, Biondi D, Antonelli F, Vlachovicova Z,

Zanzotto L (2004) Rheology of asphalts modified with glyc-

idylmethacrylate functionalized polymers. J Colloid Interface

Sci 280(2):366–373. doi:10.1016/j.jcis.2004.08.043

23. Polacco G, Stastna J, Biondi D, Zanzotto L (2006) Relation

between polymer architecture and nonlinear viscoelastic

behavior of modified asphalts. Curr Opin Colloid Interface

Sci 11(4):230–245. doi:10.1016/j.cocis.2006.09.001

24. Yildirim Y (2007) Polymer modified asphalt binders. Constr

Build Mater 21(1):66–72. doi:10.1016/j.conbuildmat.2005.

07.007

25. Zoorob SE, Castro-Gomes JP, Pereira Oliveira LA,

O’Connell J (2012) Investigating the multiple stress creep

recovery bitumen characterisation test. Constr Build Mater

30:734–745. doi:10.1016/j.conbuildmat.2011.12.060

Materials and Structures