See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/324877373

Comparison of New Test Methods and New Specifications for Rutting Resistance
and Elasticity of Modified Binders

Article · November 2017

CITATION READS

1 393

3 authors, including:

Raquel Moraes
National Center for Asphalt Technology (NCAT)
40 PUBLICATIONS   159 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

WHRP (0092-14-20) TPF-5 (302) Pooled Fund Project on "Modified Binder (PG+) Specifications and Quality Control Criteria" View project

All content following this page was uploaded by Raquel Moraes on 01 May 2018.

The user has requested enhancement of the downloaded file.

 Comparison of New Test Methods and New Specifications for Rutting
Resistance and Elasticity of Modified Binders

Raquel Moraes, Ph.D.

Asphalt Research Engineer
Department of Civil and Environmental Engineering
University of Wisconsin – Madison
Madison, Wisconsin

Daniel Swiertz, P.E.

Director of Mix Design Laboratories Bitumix Solutions
a Division of Henry G. Meigs, LLC.
Portage, Wisconsin

Hussain Bahia, Ph.D.

Professor
Department of Civil and Environmental Engineering
University of Wisconsin – Madison
Madison, Wisconsin

Acknowledgements

This study was funded in part by the Transportation Pooled Fund “Modified Binder (PG+) Specifications
and Quality Control Criteria” (TP05-302). Support of the partner States is acknowledged. Support of the
Western Cooperative Testing Group and Rocky Mountain Asphalt User/Producer Group is also
acknowledged. The results and opinions presented are those of the Authors and do not necessarily reflect
those of the sponsoring agencies or organizations.

© Canadian Technical Asphalt Association 2017

204 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

ABSTRACT

There is no consensus among state highway agencies as to the appropriate binder specifications required for
adequate quality control and acceptance of modified binders. Supplemental tests have been adopted in
addition to standard Performance Grade (PG) tests and are often referred to as “PG+” procedures. The
Multiple Stress Creep and Recovery (AASHTO M332) test has been proposed to replace the AASHTO
M320 for PG grading of binders. However, some agencies are concerned that asphalt binder formulation
will change after adopting the new system, while others are unsure how to relate current M320 grades to the
new S, H, and V grades.

In this study, testing of a large number of binders was completed and correlations between results of “PG+”
tests and MSCR tests were performed. Regarding the MSCR test, results show that the %R parameter is a
good candidate to detect the presence, and potentially the quantity, of elastomeric modification. However,
using universal limits for the MSCR %R parameter that are dependent on J nr values is not practical nor
useful since current binder formulations are controlled differently by agencies. Furthermore, no logical
equivalency of M320 PG grades to M332 traffic grades was identified, so changes to binder formulations
are anticipated.

RÉSUMÉ

Il n'y a pas de consensus parmi les différents ministères et départements de transport quant aux normes
appropriées requises pour un contrôle de qualité adéquat et l'acceptation des bitumes modifiés. Des essais
supplémentaires ont été adoptés en plus des essais standards de qualité de performance [PG] et sont souvent
appelés essais «PG+». La norme MSCR [Déformations Répétées & Recouvrance au Fluage - AASHTO
M332] a été proposée pour remplacer la norme AASHTO M320 pour les bitumes PG. Cependant, certaines
agences craignent que la formulation des bitumes ne changent après l'adoption du nouveau système, alors
que d'autres ne savent pas comment relier les classes de bitume actuelles aux nouvelles classes S, H, & V.

Dans cette étude, un grand nombre de bitumes ont été analysés et des corrélations entre les résultats des
tests «PG+» et des tests MSCR ont été vérifiées. En ce qui concerne l’essai MSCR, les résultats montrent
que le paramètre %R est un bon candidat pour détecter la présence et possiblement la quantité d’élastomère
ajoutée. Cependant, l'utilisation de limites universelles pour le paramètre %R du MSCR qui dépend des
valeurs Jnr n'est pas pratique ni utile, car les formulations des bitumes actuelles sont contrôlées
différemment par les agences. En outre, aucune équivalence logique n’a été identifiée entre les classes PG
de M320 et celles de M322 liées au trafic, de sorte que des modifications de formulation des bitumes sont
anticipées.

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 205

1.0 INTRODUCTION

1.1 Background

In 1987, the Strategic Highway Research Program (SHRP) was established to improve the performance and
durability of roads in the United States, and the Superpave™ (Superior Performing Asphalt Pavements)
asphalt binder grading system was a product of this research effort. The Superpave Performance Graded
(PG) asphalt binder specification is a performance related specification based on the rheological properties
of the asphalt binder, as well as on the climate and loading conditions of the pavement where it is to be
placed.

Today, the PG asphalt binder specification is used throughout the U.S., with many state agencies adopting
a version of the Standard Specification for Performance-Graded Asphalt Binder (AASHTO M320) [1].

Although Superpave was a significant improvement over earlier approaches for asphalt binder evaluation,
one of the concerns related to this specification is that it was developed primarily for unmodified asphalt
binders. Therefore, the applicability of the existing AASHTO M320 specification to modified binders has
been questioned by industry and many of the state highway agencies. Furthermore, there is no consensus
among state highway agencies as to the appropriate binder specifications required for adequate quality
control and acceptance of modified binders. For this reason, supplemental tests have been adopted in
addition to the standard PG tests, and are often referred to as PG+ (PG plus) procedures. However,
differences exist between the PG+ test methods, test conditions, and performance limits specified among
regions. In addition to these highlighted differences, the majority of the “PG+” test methods have not been
verified to relate to asphalt pavement performance, and most are used to ensure the presence of specific type
of modifiers in the binder. In fact, most PG+ tests focus on measuring elastic behaviour, which can be
achieved by only one group of polymers. This trend was changed as the Multiple-Stress Creep-Recovery
(MSCR) test was introduced to amend the PG requirements and to focus on rutting performance through
simulation of traffic loading and climate more directly.

1.2 Multiple-Stress Creep-Recovery (MSCR) Test

The AASHTO M332 (Performance-Graded Asphalt Binder Using the Multiple-Stress Creep-Recovery
(MSCR) Test) [2] has been proposed to replace the AASHTO M320 (Standard Specification for
Performance-Graded Asphalt Binder) [1] in the U.S. In part, this proposed replacement is to control the
proliferation of the PG+ convention tests, such as the Elastic Recovery (AASHTO T301) [3], Phase Angle
(AASHTO T315) [4], Ductility (AASHTO T51) [5] and others, which are used today to amend the
AASHTO M320 standard, and better control the use of modified binders since the MSCR test is designed
to be performance related and unbiased to the type of modifier used.

The MSCR test was developed to better capture the contribution of asphalt binder to permanent deformation
resistance of asphalt mixtures relative to the Superpave G*/sin(δ) parameter used in the AASHTO M320
grading system. The concept was first introduced in NCHRP report 459 [6] in which it was observed that
G*/sin(δ) could not differentiate between polymer modified binders and their contribution to mixture rutting
resistance. A significant amount of research followed the NCHRP 459 report and has been devoted to
relating the non-recoverable creep compliance parameter (Jnr) measured during this test to laboratory and
field measurements of permanent deformation. In general, there is a consensus in the literature that the J nr
parameter does a better job of predicting deformation resistance of polymer modified asphalts [7].

© Canadian Technical Asphalt Association 2017

206 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

Although northern climate regions do not appear to have significant problems with permanent deformation
of the asphalt mixture layer, an attractive aspect of the MSCR test is its ability to eliminate the need for
traditional “PG+” elasticity tests through the Percent Recovery (%R) parameter, which measures average
creep strain recovery. For example, many state agencies have a version of the AASHTO T301/ASTM D6084
[8] Elastic Recovery test in their specification to ensure modified binders contain elastomeric or partial-
elastomeric (hybrid) polymer modification. Although there is no specific research that conclusively relates
AASHTO T301/ASTM D6084 Elastic Recovery to performance, it is an easily interpreted test to ensure
binders have been modified with elastomeric polymers, which has been at least subjectively linked to better
deformation resistance, crack resistance, and durability [7, 9]. From a practical standpoint, it should be noted
that the traditional T301/D6084 procedures take several hours to run and require the use of a ductility bath,
whereas the MSCR test is performed using a Dynamic Shear Rheometer (DSR), and takes about 20 minutes
to be run. Both test methods use Rolling Thin-Film Oven (RTFO) residue.

Research suggests that the %R parameter may better capture the quality of modification relative to
T301/D6084, since the MSCR test is run at higher temperatures relative to the traditional elastic recovery
test, which is run at 25°C. Literature suggests that at 25°C the asphalt binder is sufficiently stiff to act as
reinforcement to the polymer network, potentially moderating the effects of poor quality modification [7].
The ability of the MSCR test to identify the extent and strength of the polymer network in a binder can be
demonstrated through a comparison of different blending techniques in one base binder.

Table 1 [9] shows two binders modified with the same amount of Styrene-Butadiene-Styrene (SBS), but
using two different SBS types and blending techniques. The binders show similar T301 elastic recovery
results (i.e., 73.8 and 86 percent), yet the MSCR %R results are considerably different when tested at high
temperature (i.e., 19.2 and 40.3 percent). In other words, testing at higher temperatures can better
differentiate between modified binders with respect to the interaction between the polymer and the base
binder. However, continuous debate exists about the capability of the MSCR %R parameter to replace tests
such Elastic Recovery and Force Ductility.

Table 1. Example of ability of MSCR %Recovery parameter in capturing modification quality.

Comparison of binders with the same base and different polymers and blending process [9].

Continuous Temperature Jnr Elastic % Recovery

Polymer
Grade (°C) @ 3.2 kPa-1 Recovery @ 3.2 kPa-1
66.7-24.1 N/A 64 3.12 5 0
75.7-22.3 4% SBS 70 1.85 73.8 19.2
76.6-25.2 4% SBS 70 1.18 86 40.3

AASHTO M332 does not provide limits for the %R parameter (or T301 recovery, for that matter), instead
a chart is provided in the M332 specification that is intended to provide “an indication of elastic response.”
The chart is shown in Figure 1. According to the standard, if a binder tested according to AASHTO T350
[10] produces a Jnr and associated %R value that when plotted on the chart falls above the solid line, the
binder is said to be “modified with an acceptable elastomeric polymer”; and “if the point falls below the
line on the graph, the indication is that the asphalt binder is not modified with an elastomeric polymer.” In
the U.S., several states that have implemented the MSCR grading procedure specify a minimum limit of
%R based on this chart directly (i.e., the equation is used with the binder Jnr and if the calculated %R is
below the measured %R, the result is acceptable). Others use a ‘tiered’ approach in which a single minimum
%R is used for a given range of Jnr values. The latter is the approach used currently by the Combined State
Binder Group (Iowa, Minnesota, North Dakota, South Dakota, Wisconsin, and Nebraska).

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 207

Figure 1. Indication of elastomeric modification, from AASHTO M332 [2].

Another positive attribute of the AASHTO M332 standard (MSCR Test) that should be considered is the
specified testing temperature. The AASHTO M320 grading procedure uses ‘grade bumping’ to adjust the
high temperature binder grade to account for heavier and/or slower moving traffic. In truth, the pavement
will never see the high temperatures of 76oC and above listed in the Superpave PG specification; it is just
an artificial way of requiring a stiffer binder by testing at a higher temperature [9].

It is well-established that the integrity of some polymer systems can be highly temperature dependent,
meaning that although a certain polymer or polymer concentration may provide adequate performance
within a certain climatic zone, it may be eliminated from usage consideration because it is tested at
significantly higher temperature than that climatic temperature. The AASHTO M332 procedure specifies
testing only at the high pavement temperature for the climatic zone of the project and adjusting for traffic
amount/speed by lowering the maximum limit on Jnr. The result is expected to be a better representation of
actual field conditions, and correlations of Jnr with field data suggest this to be true. However, some agencies
are concerned that formulation of binders will change, while others are not sure about how to relate the PG
grades to the new S (Standard), H (Heavy), V (Very heavy), and E (Extreme) grades for the expected traffic
level and traffic load rate.

2.0 RESEARCH OBJECTIVES

This study was performed with the following objectives: (a) Establish a correlation between the MSCR
Percent Recovery (%R) parameter with existing elasticity test methods, and comment on the ability of the
%R parameter to detect the presence of elastomeric modification and potentially provide indication of the
level of elastomeric modification; (b) Comment on existing limits for %R based on available databases,
published literature, and ongoing research projects (such as the Pooled Fund TPF-5/302) [11] relating %R
to performance; and (c) Provide recommendations for implementation, if applicable, of the J nr, % Jnr diff.,
and %R parameter based on the needs of the governing agency.

© Canadian Technical Asphalt Association 2017

208 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

3.0 MATERIALS AND TESTING PROCEDURE

3.1 Materials

The modified and unmodified asphalt binders used in this study were obtained from the Western States
Cooperative Testing Group (WCTG) and Rocky Mountain Asphalt User/Producer Group (RMAUPG),
representing the western United States; the Combined State Binder Group (CSBG), representing the upper
Midwest; the State of Kansas Department of Transportation (DOT); the State of Ohio DOT; and the Pooled
Fund TPF-5 (302) [11] state partners (Colorado, Idaho, Kansan, Ohio, and Wisconsin). The PG
classification of each modified asphalt binder is described in Table 2. In order to avoid bias during the
analysis of the results, no information about the type of the modifier used in each asphalt binder was
provided from the partner states.

Table 2. Performance Grade (PG) classification of the asphalt binders selected in this study.

Combined Western States Cooperative

Pooled
Performance State Testing Group (WCTG) &
Kansas Ohio Fund
Grade (PG) Binder Rocky Mountain Asphalt
DOT DOT TPF-5
Classification Group User/Producer Group
(302)
(CSBG) (RMAUPG)
52-34 X
58-34 X X X
64-22 X X
64-24 X
64-28 X X X X X
64-34 X X X X
70-22 X X X X
70-28 X X X X
70-34 X X
76-22 X X X X
76-28 X X X

3.2 Testing Procedures

The AASHTO M320 specification and associated laboratory testing procedures were utilized to grade the
asphalt binders at high temperature. As a first step, the original (or unaged) binders were graded to study
their properties during the transportation and handling steps (i.e., before mixing). As a second step, the
binders were exposed to a short-term aging process in the RTFO in accordance with the AASHTO T240
procedure [12], in an attempt to simulate asphalt mixture properties during production and construction.

3.2.1 Multiple-Stress Creep-Recovery (MSCR) Test Procedure

Using a DSR, the MSCR test was performed on RTFO-aged binders in accordance to AASHTO T350 at
different temperatures, depending on the asphalt binder grade. The test was conducted by applying two
stress levels of 0.1 kPa (for twenty cycles) and 3.2 kPa (for ten cycles). Each cycle consisted of 1 second
shear creep followed by a recovery period of 9 seconds. The MSCR test generates two key parameters that
are used to determine the binder grade as per AASHTO M332. These values are: non-recoverable creep

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 209

compliance (Jnr) and percent recovery (%R). Based on the value of Jnr, an asphalt binder can get a grade of
S (Standard), H (Heavy), V (Very heavy) and E (Extreme) heavy to account for the traffic level. The average
non-recoverable creep compliances at 0.1 kPa (Jnr0.1) and 3.2 kPa (Jnr3.2) are expressed in Equations 1 and 2.

∑20
𝑁=11[𝐽𝑛𝑟 (0.1,𝑁)]
𝐽𝑛𝑟 0.1 = 10
(1)

∑10
𝑁=1[𝐽𝑛𝑟 (3.2,𝑁)]
𝐽𝑛𝑟 3.2 = 10
(2)

Where: Jnr (0.1, N) and Jnr (3.2, N) are the non-recoverable creep compliances at 0.1 kPa and 3.2 kPa
at cycle number N, respectively, and
N is the cycle number at each stress level.

The percent difference in non-recoverable creep compliance between 0.1 kPa and 3.2 kPa is calculated using
Equation 3.

[𝐽𝑛𝑟3.2 −𝐽𝑛𝑟 0.1 ]×100

𝐽𝑛𝑟 𝑑𝑖𝑓𝑓 = 𝐽𝑛𝑟 0.1
(3)

The average percent recovery at 0.1 kPa (R0.1) and 3.2 kPa (R3.2) are, respectively, expressed using Equations
4 and 5.

∑20
𝑁=11[∈𝑟 (0.1,𝑁)]
𝑅0.1 = 10
(4)

∑10
𝑁=1[∈𝑟 (3.2,𝑁)]
𝑅3.2 = 10
(5)

Where: ∈𝑟 (0.1, N) and ∈𝑟 (3.2, N) are the percent recovery at 0.1 kPa and 3.2 kPa at cycle number N,
respectively; and
N is the cycle number at each stress level.

3.2.2 Elastic Recovery - AASHTO T301

The Elastic Recovery (ER) tests were performed in accordance with AASHTO T301 [3] using a
ductilometer and briquette specimens. This test method is useful in confirming that a material has been
added to the asphalt binder to provide a significant elastomeric characteristic. However, it does not
necessarily identify the type or amount of elastomeric modifier added. In this study, the RTFO-aged asphalt
binder specimens were conditioned in a water bath at 25°C for one hour. The test is displacement-controlled
and the asphalt binder specimens were pulled apart at a constant speed of 5 cm/min to an elongation of 20
cm, and usually held for 5 min; however, some state agencies vary the hold period and/or elongation limit.
After the hold period, a cut is made in the middle of the specimen and allowed to remain in the ductilometer
for recovery. After usually one hour, the percent elongation recovery is determined using Equation 6.
20−𝑥
% 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 = 20
× 100 (6)

Where: x is the final reading in centimeters after the two severed ends of the specimen are brought
back together.

© Canadian Technical Asphalt Association 2017

210 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

3.2.3 Repeated Creep Mixture Testing Procedure

To simulate high temperature climactic conditions, the AASHTO TP79 [13] standard was followed to test
Hot Mix Asphalt (HMA) samples at 46°C in an environmental chamber. After the conditioning, samples
were loaded in creep cycles until failure; where one creep cycle consisted of a 150 psi load being applied
for 0.1 seconds, followed by a rest period of 0.9 seconds. The Flow Number (FN) of each HMA sample was
then calculated in accordance to the Francken model [14].

4.0 RESULTS

4.1 Correlation of MSCR Percent Recovery (%R) Parameter with Elastic Recovery

Two US states (Ohio and Kansas) have provided databases of MSCR testing conducted at the intended
pavement temperature for their given region (predominately 64°C). Plots of the correlation between Elastic
Recovery (PG+) measured in accordance with the ASTM D6084 and MSCR %R for Kansas DOT (KDOT)
and Ohio DOT (ODOT) are shown in Figures 2 (a) and (b), respectively. The plots are intentionally kept
separate for three reasons: (1) Kansas and Ohio each run the current PG+ elasticity (both use ASTM D6084)
test differently, with Ohio opting for a 5 minute rest after elongation and Kansas not specifying a rest (more
commentary will be provided on this topic later); (2) the two states have different limits for the PG+
elasticity tests and have other PG+ tests (phase angle, ductility, etc.) not conducted by the other state that
may influence the type and amount of modification that takes place in the respective state; and (3) ODOT
only allows modification with SBR, SB, SBS, or Elvaloy®, whereas Kansas does not directly specify the
type of additive. It is important to note that specification drives formulation, meaning whatever specification
a given state or combined state group adopts, regional formulations will adapt to meet that specification. It
is, therefore, expected that comparing binders produced in different regions by different suppliers will not
be completely analogous.

Several important observations can be made from the plots. First, the explained variance (R 2) when
separating test method and MSCR test temperature becomes much more reasonable than when combining
all data; interestingly, the R2 values for MSCR tests conducted at 64°C are approximately equal between
the two states. Second, the ranking of %R values appears to be much more logical in terms of increasing %
R with increasing grade. One might expect a similar base binder be used in a given region based on refinery
availability and modified accordingly to meet specification. Thus, a similar PG 64-22 base was likely used
in Kansas to produce PG 70-22 and PG 76-22 binders and it is logical to produce these binders by slightly
increasing the level of modification to achieve the respective grade bump. If an elastomeric modifier is used
(based on specification), it stands to reason that the PG 76-22 should generally produce a higher %R than
the PG 70-22, and this appears to be the case. It should be also noted that the range in MSCR %R values
changes from less than 10 percent to more than 95 percent for both states, whereas the relative change in
PG+ elastic recovery (ASTM D6084) measured at 25°C changes from about 50 to nearly 100 percent, with
ODOT values more spread out. The ASTM D6084 procedure uses 10 cm elongation, the sample is cut
immediately, and a recovery time of 60 minutes is used, whereas KDOT opts to use a total elongation of 10
cm, a hold time of five minutes, and a recovery time of 60 minutes. Based on this information, one might
expect elastic recovery values to be higher for given binder tested in the ASTM D6084 without a hold time
for a given binder for which there is no stress relaxation period. This appears to be the case when comparing
the KDOT and ODOT data (as shown in Figure 3). In addition, each state has other PG and PG+
requirements that must be met, and this could influence the elastic recover values as well.

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 211

100
90
80 64-28
MSCR %R, 3.2 kPa

70 64-34
60 70-22
50 70-28
40 76-22
30 76-28

20 all

10 R² = 0.73
0
40 50 60 70 80 90 100
ASTM D6084 (KDOT), %Recovery

(a)

100
R² = 0.74
90

80 (R2 = 0.78 with 88-22 removed)

70
MSCR %R, 3.2 kPa

60 64-28
50 70-22
76-22
40
88-22
30
64-22
20

10

0
40 50 60 70 80 90 100
ASTM D6084 (ODOT), %Recovery

(b)

Figure 2. Correlation of Elastic Recovery (ASTM D6084) to Multiple Stress Creep Recovery
(MSCR) %R for (a) KDOT and (b) ODOT MSCR performed at 64°C.

© Canadian Technical Asphalt Association 2017

212 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

100

90

80

70
MSCR %R, 3.2 kPa

60

50 ODOT

40 KDOT

30

20 R² = 0.74
R² = 0.73
10

0
40 50 60 70 80 90 100
PG+ % Recovery

Figure 3. Comparison of KDOT and ODOT data showing effect of PG+ elasticity method.

From this analysis, it can be concluded that any specification for MSCR %R intended to replace an existing
PG+ elastic recovery specification must: (1) be based on a uniform MSCR test temperature for a given set
of binders (most likely the high pavement temperature for a given region), and (2) be correlated to the actual
PG+ test used in that region. For agencies interested in keeping similar formulations to what they currently
utilize, it appears that implementing a uniform specification for %R across all regions (as in adopting the
M332 %R outright) will likely result in formulation change within a given region, with the magnitude of
the change depending on the current PG+ specifications used. For example, the Asphalt Institute
recommended to the Southeast Asphalt User Producer Group that an appropriate %R value when moving
from T301 to %R at 64°C is 15 percent less than the corresponding T301 value. As shown in Figure 2(b),
for the ODOT data, this number is actually closer to a 25 percent reduction at 60 percent ASTM D6084 and
about a 5 percent reduction at 80 percent ASTM D6084 recovery, suggesting limits should be based on the
binder grade itself rather than a standard reduction for all grades. For KDOT on the other hand, since the
elastic recovery values are higher, and the slope of the correlation is steeper, the corresponding limits for
%R to achieve the same elastic recovery would be lower. For example, 80 percent elastic recovery correlates
to an MSCR %R at 64°C of only 30 percent.

4.2 Correlation of MSCR Percent Recovery (%R) Parameter with Phase Angle

Phase angle measured at high grade temperature is another PG+ used in conjunction with AASHTO
T301/ASTM D6084 to measure elasticity of binders. The perceived benefit of running the phase angle is
that it is a quick and repeatable test method, and may be done on original binder; also, for daily quality
control this test can reduce testing time considerably.

Figure 4 shows the correlation for ODOT binders between MSCR %R and phase angle (both original binder
and RTFO).

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 213

80.0

75.0
(R2 = 0.90 with 88-22 removed)
70.0
Phase Angle (RTFO), °

65.0
70-22

60.0 76-22

55.0 88-22
R² = 0.79
64-22
50.0
64-28
45.0

40.0
0 20 40 60 80 100
MSCR %R, 3.2 kPa

(a)
85

80
Phase Angle (Unaged binders), °

75

70 70-22

65 76-22

60 88-22
R² = 0.68
55 (R2 = 0.80 with 88-22 removed) 64-22

64-28
50

45

40
0 20 40 60 80 100
MSCR %R, 3.2 kPa

(b)

Figure 4. ODOT Multiple Stress Creep Recovery (MSCR) %R at 64C vs. Phase Angle at grade
temperature for (a) RTFO-aged binders and (b) original binders.

© Canadian Technical Asphalt Association 2017

214 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

As shown, the correlation is strongly linear for the RTFO aged binders, with an explained variance (R2) of
nearly 80 percent. When the PG 88-22 binders are removed, the correlation increases to 90 percent.
Interestingly, the correlation of the MSCR %R with the phase angle taken for the original binders is also
strong, particularly if the PG 82-22 binders are removed from consideration. The overall spread in phase
angle measurements (i.e., not grouped by binders known to have varying levels of modification) suggests
that the phase angle test may have utility as a quality control point, but is probably not sufficient to determine
amount or quality of modification.

4.3 Feasibility of Using AASHTO M332 to Replace PG+ Methods

The same data from the ODOT and KDOT databases can be plotted on the M332 chart (Figure 5).
Delineations for AASHTO M332 Jnr requirements for H, V, and E grades are shown on the chart with %R
limits currently in use by the Combined State Binder Group (CSBG) for MSCR tests at 64°C for reference.
The data in Figure 5 show logical trends with respect to increasing %R and reduced Jnr with higher PG
grads, particularly for KDOT (Figure 5a). For example, producers would likely need to use more modifier
in the PG 64-34 than the PG 64-28, since they would likely start with a softer base asphalt (i.e., a higher
Useful Temperature Interval, UTI). The result is a shift to higher %R and lower Jnr values for the PG 64-34
binders, which is in agreement with the data. The same can be said of the PG 70-28 vs. PG 70-22 binders,
where the PG 70-28 binders show a shift in %R higher and a shift in J nr lower (data moves up and to the
left).

If these states agencies wish to implement M332 as a grading procedure to replace grade bumping, the data
in Figure 5 suggests that changes in formulation will likely occur over time. Using the Kansas data set as
an example, if a region requires a PG 64-22 for climatic reliability, a AASHTO M320 grade bumped PG
70-22 would be approximately equivalent to a AASTHO M332 64H-22. However, the PG 70-22 binders
fall in the V and E range, likely because of the higher level of modification used to achieve the requirements
of the AASHTO M320. Since the V and E grades require higher minimum %R values than H grade,
suppliers producing PG 70-22 can actually drop the concentration of elastomer, adjust reaction/cross-linking
times, or utilize a hybrid modification system to produce a binder that is actually softer with a lower elastic
response and pass specification for the H grade. In other words, formulations of modified binders are likely
to change if the AASHTO M332 is used to replace the AASHTO M320.

Similar conclusions can be made when looking at a subset of Combined State Binder Group (CSBG) data.
Qualification and quality control samples from projects specifying polymer modified asphalt were collected
to develop (or validate) limits on %R. An example of this data is shown in Figure 6, where each point
represents a quality control or verification sample taken for a given project, not necessarily an individual
binder. Several members of the CSBG (Wisconsin, for example) have made the decision to ‘phase in’ the
M332 grading procedure over the course of two to three years. Originally Wisconsin utilized the AASHTO
T301 elastic recovery and phase angle requirements for all polymer modified asphalts, with the MSCR %R
as a ‘report only’ parameter. The T301 recovery was dropped (phase angle remained), and the MSCR %R
minimum requirement was implemented (the binder was still ‘graded’ by M320, but the MSCR test was
run at 58 or 64 °C). In Wisconsin, two asphalt grades are used: PG 58-28 and PG 58-34. A PG 64-28 and
PG 70-28 would, therefore, be approximately equivalent to a PG 58H-28 and PG 58V-28, respectively.
Examining Figure 6, one can see that these equivalencies do not necessarily hold true, and in fact for all but
the PG 58-34 binders, the Jnr classifications are one traffic grade higher than expected (Jnr is lower).

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 215

KDOT Prequalification and Verification MSCR - Tests at 64 °C

100
90 E V H
M332
80
E Grade %R min.
70
MSCR %R, 3.2 kPa

Effect of Increased V Grade %R min.

60 Elastomeric Modification H Grade %R min.
50 64-28
40 64-34
70-22
30
70-28
20
76-22
10
76-28
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

MSCR Jnr, 3.2 kPa

(a)

ODOT Prequalification and Verification MSCR - Tests at 64 °C

100
90 E V H
80 M332
E Grade %R min.
MSCR %R, 3.2 kPa

70
V Grade %R min.
60
H Grade %R min.
50
64-22
40
70-22
30
76-22
20 88-22
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

MSCR Jnr, 3.2 kPa

(b)

Figure 5. Multiple Stress Creep Recovery (MSCR) data (a) KDOT binders, and (b) ODOT binders.

© Canadian Technical Asphalt Association 2017

216 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

100

90 E V H M332
PG 58-34
80
PG 64-28
70
MSCR %R, 3.2 kPa

PG 64-34
60 PG 70-22

50 PG 70-28
PG 70-34
40
PG 76-28
30
E Grade %R min.
20
V Grade %R min.
10 H Grade %R min.

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
MSCR Jnr, 3.2 kPa

Figure 6. Combined State Binder Group example data. MSCR at 58C.

Table 3 shows a breakdown of statistics for the binders shown in Figure 6. It is clear that the majority of
currently produced binders pass the proposed specification, many by a wide margin, particularly in J nr
values. Binders that pass specification by a wide margin are likely overdesigned for the AASHTO M332
specification, meaning they cost more up front to the agency, and are significantly stiffer than intended.
From the data shown, it can be reasonably expected that current formulations will likely change over time
to more closely meet the M332 specification, and associated costs will likely drop accordingly.

Table 3. M332 Grading Information for Example CSBG Data

Binder PG 58-34* PG 64-28 PG 64-34 PG 70-22 PG 70-28 PG 70-34 PG 76-28

MSCR %R Limit
30% 30% 55% 55% 55% 75% 75%
@ 58C
No. Tests 109 374 274 52 176 31 2
No. Passing 92 290 259 47 173 31 2
% Passing 84.4% 77.5% 94.5% 90.4% 98.3% 100.0% 100.0%
MSCR Grade Des. H H V V V E E
Jnr @ 3.2 kPa
2.0 2.0 1.0 1.0 1.0 0.5 0.5
Limit (kPa)
No. Tests 109 374 274 52 176 31 2
No. Passing 107 373 270 52 176 30 2
% Passing 98.2% 99.7% 98.5% 100.0% 100.0% 96.8% 100.0%

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 217

4.4 Relationship of MSCR Criteria to Performance

From a pavement engineering point of view, elastic materials are desirable since they show complete strain
recovery after deformation, which implies resistance to rutting. However, such pavements could be a
problem for thermal cracking, due to the possible reduction in stresses relaxation ability by viscous flow.
The direct relationship between elastic behaviour and overall performance is therefore not clear, and could
be highly dependent on climate, loading, and type of distress. In addition, the conventional ways used to
measure elasticity in asphalt binders may not be representative of field conditions or loading, and are not
actually measuring a true material property, per se. For example, the AASHTO T301 elastic recovery test
strains the binder significantly more than would be expected in the field, potentially damaging the polymer
structure. Nevertheless, these tests continue to be used to measure relative elasticity.

In this study, WCTG binders were correlated to MSCR test parameters and mixture flow number results.
Each asphalt mixture selected for this comparison has different aggregate sources, design considerations,
and design traffic levels. The findings (as shown in Figure 7) indicate that if the MSCR test is performed at
the mixture performance test temperature, a logical trend exists between flow number and J nr at 3.2 kPa,
with a reasonably high explained variance (R2) considering the variation in mixtures tested. A power law
was used since Jnr cannot fall below zero, yet flow number can continue to increase indefinitely. This trend
presents an important observation: there is a diminishing return on very low Jnr values (i.e., agencies could
be paying more for a limited and marginal increase in improvement to performance). Another important
observation is that, in terms of similar levels of performance, binders can show large changes in the J nr value
(note the Jnr is on a logarithmic scale). Clearly this indicates that while a general trend exists, and it appears
a lower Jnr improves rut resistance, the other components/design factors in the mixture could have a large
impact on performance.

In terms of the %R parameter, although the power fit is similar, the general scatter in the data suggests that
the %R is much less important for high temperature performance. For changes in recovery of less than 5
percent, nearly two orders of magnitude difference in flow number exists for the data. For softer binders,
which by association would allow more deformation for each load pulse, it is plausible that %R has a larger
impact since more recovery could take place during each cycle. %R is not normalized to the load as is Jnr,
so it is a parameter derived from the deformation that takes place during each cycle. Therefore, if a binder
had a very large %R and a relatively high J nr, the %R parameter might impact performance more. This is
the converse of how the AASHTO M332 specification is written, however, in that high J nr values require
lower minimum %R values.

4.5 Percent Jnr Difference and Stress Dependence of Modified Binders

In the MSCR procedure, the 0.1 and 3.2 kPa stress levels are used to estimate the %Jnr Difference parameter
(Equation 3). Since almost all asphalt binders are stress-softening, the 3.2 kPa Jnr should always be equal to
or larger than the 0.1 kPa Jnr value. Depending on the base asphalt and modification type, the %Jnr Difference
can range widely for different asphalt binders. To ensure that asphalt binders do not exhibit a high stress
sensitivity, a maximum %Jnr Difference of 75 percent is placed on the J nr when stress is increased from 0.1
to 3.2 kPa in AASHTO M332 specification. To understand how this requirement applies to the binders
provided by Pooled Fund members [11], binders were tested at three stress levels: 0.1, 3.2 and 10 kPa. The
%Jnr Difference was calculated between each subsequent stress level (i.e. %J nr difference between 0.1 and
3.2 kPa, and %Jnr difference between 3.2 and 10 kPa). Each Pooled Fund binder %Jnr Difference values are
compared in Figure 8. Each binder was labeled with the first letter of the respective partner state and the PG
grade for identification. The MSCR testing was completed at the high grade temperature of the binder.

© Canadian Technical Asphalt Association 2017

218 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

1.0000
Jnr, 3.2kPa, kPa @ 46 °C

0.1000

0.0100
R² = 0.60

0.0010
1 10 100 1,000 10,000 100,000

Flow Number (FN) @ 46 °C

(a)
MSCR %R, 3.2 kPa @ 46 °C

100.0

90.0

80.0
R² = 0.50
70.0

60.0

50.0

40.0
1 10 100 1,000 10,000 100,000

Flow Number (FN) @ 46 °C

(b)

Figure 7. Flow Number testing relative to Multiple Stress Creep Recovery (MSCR) data on WCTG
mixtures [15].

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 219

Figure 8. % Jnr Difference values for each Pooled Fund binder. 0.1-3.2 kPa data represents the %
Jnr Difference between stress levels of 0.1 and 3.2; and 3.2-10 kPa data represents the % Jnr
Difference between stress levels of 3.2 and 10 kPa.

As it can be seen in Figure 8, there is a wide range of potential outcomes depending on the binder type.
Three of the 0.1-3.2 kPa %Jnr Difference values fail the %Jnr Difference parameter and have a 0.1-3.2 kPa
%Jnr Difference that is significantly greater than the 3.2-10 kPa %Jnr Difference. For each of these binders,
one was graded to withstand Very heavy traffic (V) and two were graded for Extremely heavy traffic (E)
according to AASHTO M 332. Therefore, the asphalt binders meet the J nr requirements for large volumes
of traffic, but fail the specification due to high stress sensitivity. These results are not logical, given that the
same three binders have 0.1-3.2 kPa %Jnr Difference values greater than the corresponding 3.2-10 kPa %Jnr
Difference values. As previously stated, as the stress level increases, the J nr is also expected to increase.
Therefore, the 0.1-3.2 kPa %Jnr Difference would be expected to be lower than the 3.2-10 kPa %Jnr
Difference. A potential cause of the stress sensitivity failures could be found in the test procedure itself. For
E and V binders, the Jnr values at 0.1 kPa are extremely small, even approaching the resolution limits of the
DSR, so when the %Jnr Difference is calculated, a very large number usually results.

To supplement the Pooled Fund data [11], the Ohio binder database was also used to determine if a trend
exists between the Jnr parameter and the %Jnr Difference. Thus, the %Jnr Difference as reported in the Ohio
database was compared against the measured Jnr values at 0.1 kPa and 3.2 kPa. All testing was conducted at
64 °C, and results are presented in Figure 9.

© Canadian Technical Asphalt Association 2017

220 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

1000

100
64-22
%Jnr Difference (%)

70-22
10
76-22

88-22
1

0.1
0 0.5 1 1.5 2 2.5
Jnr, 0.1 kPa

(a)
1000

100
%Jnr Difference (%)

64-22
10 70-22
76-22
88-22
1
Average across all grades = 45.7%

0.1
0 0.5 1 1.5 2

Jnr, 3.2 kPa

(b)

Figure 9. ODOT Multiple Stress Creep Recovery (MSCR) %Jnr Difference vs. Jnr at two stress
levels: (a) 0.1 kPa, and (B) 3.2 kPa.

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 221

As shown in Figure 9, there is no direct correlation between the measured Jnr value at either stress level and
%Jnr Difference. However, as the Jnr becomes increasingly small, the %Jnr Difference parameter appears to
show much higher variation for a given Jnr value (but not necessarily a high value). It is shown that binders
that fail the maximum limit of 75 percent are mostly of the PG 88-22 and the PG 72-22, which are currently
used for only the very heavy traffic conditions. In addition, none of the binders that would effectively be
graded as H of V by Jnr limits of 0.5 and 1.0 fail the specification for this data set. This could be due to
several factors including the test procedure itself, as alluded to when describing the findings in Figure 8. It
may also be due to the concentration of polymer, since to achieve an E grade from the same base binder,
more polymer would inherently be needed. The polymer concentrations and formulation processes for the
PG 76 and PG 88 binders are unknown, however the trend indicates that using the Jnr Difference limit of
75% could lead to change in formulations of such grades to meet the requirement.

Clearly the analysis of the combined data indicates that more work is needed to better understand this
parameter, particularly for high performing grade binders, before agencies should consider implementing it
in their specifications. Consideration should be made to list the parameter ‘Report Only’ until a clearer
understanding of the parameter is achieved. This is recommended since there is no clear evidence that the
parameter is related to performance in the lab or in the field, or that the limit is selected to reduce risk of
premature rutting.

5.0 SUMMARY OF FINDINSG AND CONCLUSIONS

Based on the data presented in this study, the implementation of the MSCR test can target more than one
objective. The following recommendations are offered based on the objective of the specifying agency and
the feasibility of using one or more of the MSCR test parameters.

 The correlation between the MSCR %R parameter and elastic recovery AASHTO T301 testing
procedure is relatively poor and shows a lot of scatter. However, if the objective is to maintain the
same modified binder formulations as controlled today by the AASHTO T301/ASTM D6084 Elastic
Recovery (ER), the %R parameter is a possible candidate that can be used to detect the presence of
and potentially quality/quantity of elastomeric modification. It is shown to directly correlate with the
current T301/D6084 procedures and the phase angle on a state by state basis, but not on a universal
basis.

 Recommending universal limits for the MSCR %R parameter, as stipulated in the AASHTO M332,
that are dependent on Jnr value is not practical nor useful from pavement performance point of view,
since %R is not clearly related to rutting performance.

 If the objective is to replace the G*/sin(δ) parameter and ensure good contribution of binders to rutting
resistance, the Jnr parameter measured at 3.2 kPa is a good candidate, irrespective of the %R. Rutting
resistance is highly correlated with mixture Flow Number results and literature clearly show it is a
better choice than the G*/sin (δ).

 The conversion of the grades determined based on the AASHTO M320 with grade bumping (e.g., PG
70, PG 76, and PG 82) to the traffic grades (PG 58 or PG 64 S, H, V and E) is not simple because the
G*/sin(δ) used in the AASHTO M320 does not correlate well with the Jnr at 3.2kPa. Therefore, if
agencies wish to keep the same binder formulations (or minimize change), but wish to implement the
MSCR Jnr parameter, the limits for Jnr at 3.2 kPa will likely be different between regions. However,

© Canadian Technical Asphalt Association 2017

222 NEW TEST METHODS AND SPECIFICATIONS FOR MODIFIED BINDERS

since the Jnr parameter is well-related to rutting performance of mixtures, agencies should not try to
maintain formulations and focus on using universal values of Jnr as related to traffic and climate. This
is the part that AASHTO M332 that is ready for implementation. There is more work to be done to
verify that the Jnr values of 4.5, 2, 1.0 and 0.5 1/kPa correctly correspond to the traffic speed and
volume designated in the AASHTO M332.

 The %Jnr Difference parameter is highly dependent on the binder formulations but lacks a clear
relationship to rutting performance. It is claimed in the literature to be an indicator of modification
quality, unfortunately no clear evidence is provided. In addition, since it is measured relative to the
Jnr tested at 0.1 kPa, its reliability in terms of variability and in terms of actual condition in typical
asphalt mixtures is questionable. If a universal value of 75 percent, as currently listed in the M332 is
used, changes in formulations of binders to meet this limit are expected as shown in the data collected
in this study.

REFERENCES

[1] American Association of State Highway and Transportation Officials (AASHTO) M320-16,
“Standard Method of Test for Performance-Graded Asphalt Binder”, Standard Specifications for
Transportation Materials and Methods of Sampling and Testing, Part 2B, 36th Edition, Washington,
DC (2016).

[2] American Association of State Highway and Transportation Officials (AASHTO) M332-14,
“Standard Specification for Performance-Graded Asphalt Binder Using Multiple Stress Creep
Recovery (MSCR) Test”, Standard Specifications for Transportation Materials and Methods of
Sampling and Testing, Part 2B, 34th Edition, Washington, DC (2014).

[3] American Association of State Highway and Transportation Officials (AASHTO) T301-13,
“Standard Method of Test for Elastic Recovery Test of Asphalt Materials by Means of a
Ductilometer”, Standard Specifications for Transportation Materials and Methods of Sampling and
Testing, Part 2B, 33th Edition, Washington, DC (2013).

[4] American Association of State Highway and Transportation Officials (AASHTO) T315-16,
“Standard Method of Test for Determining the Rheological Properties of Asphalt Binder Using a
Dynamic Shear Rheometer (DSR)”, Standard Specifications for Transportation Materials and
Methods of Sampling and Testing, Part 2B, 36th Edition, Washington, DC (2016).

[5] American Association of State Highway and Transportation Officials (AASHTO) T51-13, “Standard
Method of Test for Ductility of Bituminous Materials”, Standard Specifications for Transportation
Materials and Methods of Sampling and Testing, Part 2A, 33th Edition, Washington, DC (2013).

[6] Bahia H, Hanson DI, Zeng M, Zhai H, Khatri MA. Characterization of Modified Asphalt Binders in
Superpave Mix Design, National Cooperative Highway Research Program (NCHRP) Report 459,
Transportation Research Board, National Research Council, National Academies, Washington, D.C.
(2001).

[7] D’Angelo J. New High-Temperature Binder Specification Using Multistress Creep and Recovery,
Transportation Research Circular EC-147, Transportation Research Board, National Research
Council, National Academies, Washington, D.C., 1-13 (2010).

© Canadian Technical Asphalt Association 2017

 MORAES, SWIERTZ & BAHIA 223

[8] ASTM International (ASTM) D6084, “Standard Test Method for Elastic Recovery of Asphalt
Materials by Ductilometer”, Annual Book of ASTM Standards, Road and Paving Materials; Vehicle-
Pavement Systems, 04-03, West Conshohocken, Pennsylvania (2011).

[9] The Multiple Stress Creep Recovery (MSCR) Procedure. Federal Highway Administration’s
(FHWA’s) Asphalt Pavement Technology Program, TechBrief, Washington, D.C. (2011).

[10] American Association of State Highway and Transportation Officials (AASHTO) T350-14,
“Standard Method of Test for Multiple Stress Creep Recovery (MSCR) Test of Asphalt Binder Using
a Dynamic Shear Rheometer (DSR)”, Standard Specifications for Transportation Materials and
Methods of Sampling and Testing, Part 2B, 34th Edition, Washington, DC (2014).

[11] Bahia H, Lyngdal E, Varma R, Sweirtz D, Teymourpour P. “Modified Binder (PG+) Specifications
and Quality Control Criteria”, Wisconsin Highway Research Program (WHRP) 0092-14-22,
Transportation Pooled Fund Program (TPF-5/302) (2016).

[12] American Association of State Highway and Transportation Officials (AASHTO) T240-13,
“Standard Method of Test for Effect of Heat and Air on a Moving Film of Asphalt Binder (Rolling
Thin-Film Oven Test)”, Standard Specifications for Transportation Materials and Methods of
Sampling and Testing, Part 2B, 33th Edition, Washington, DC (2013).

[13] American Association of State Highway and Transportation Officials (AASHTO) TP79-16,
“Standard Method of Test for Determining the Dynamic Modulus and Flow Number for Hot Mix
Asphalt (HMA) Using the Asphalt Mixture Performance Tester (AMPT)”, Standard Specifications
for Transportation Materials and Methods of Sampling and Testing, Part 2A, 36th Edition,
Washington, DC (2016).

[14] Dongré R, D’Angelo J, Copeland A. “Refinement of Flow Number as Determined by Asphalt Mixture
Performance Tester”, Transportation Research Record No. 2127, Recycled Rubber, Aggregate, and
Filler in Asphalt Paving Mixtures, Transportation Research Board, National Research Council,
National Academies, Washington, D.C., 127-136 (2009).

[15] Lyngdal E. “Critical Analysis of PG and PG+ Asphalt Binder Test Methods”, Master Dissertation,
University of Wisconsin-Madison (2015).

© Canadian Technical Asphalt Association 2017

View publication stats