Construction and Building Materials 153 (2017) 695–703

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Development of paving performance index system for selection of

modified asphalt binder
Chao Wang a,b, Lidong Zhao c, Wei Cao d,⇑, Dongwei Cao c, Bo Tian c
a
Key Laboratory of Road Structure & Material Ministry of Transport, Beijing, PR China
b
Department of Road and Railway Engineering, Beijing University of Technology, Beijing 100124, PR China
c
Highway Engineering Research Center, Research Institute of Highway Ministry of Transport, Beijing 100088, PR China
d
Louisiana State University, Louisiana Transportation Research Center, 4101 Gourrier Ave., Baton Rouge, LA 70808, USA

h i g h l i g h t s

A comprehensive paving performance index (PPI) is established for modified asphalt.

PPI considers various mechanical performances of asphalt binder.

An adequate linear relationship is observed between PPI and binder cost.

The performance per unit price is utilized to assess the cost-effectiveness.

a r t i c l e i n f o a b s t r a c t

Article history: The objective is to establish a comprehensive paving performance index (PPI) for quantifying the cost-
Received 27 March 2017 effectiveness of modified asphalt binders. Eight modified binders were characterized using a suite of
Received in revised form 18 June 2017 mechanical tests for the rutting, fatigue, yield, and recovery properties. The current specification param-
Accepted 19 July 2017
eter for each test was evaluated to establish the individual performance index (IPI). The developed PPI
considered contributions from all IPIs and was found to correlate reasonably well with cost via a linear
function, except for the crumb rubber (CR) modified binder. By evaluating the binder performance per
Keywords:
unit price, CR was identified as the most cost-effective modifier.
Modified asphalt
Rutting
Ó 2017 Elsevier Ltd. All rights reserved.
Fatigue
Yield resistance
Elastic recovery
Cost effectiveness

1. Introduction raise economic concerns and perhaps compatibility and stability

issues [1,2].
In the past two decades, modified asphalts have been widely Crumb rubber (CR) is another popular modifier that refers to the
utilized in pavement engineering and demonstrated better perfor- application in which ground recycled rubber and paving asphalt
mance than conventional neat asphalts in many aspects such as are combined [3]. Additional economic advantages can be obtained
rutting, thermal cracking, fatigue cracking, stripping, and tempera- from the CR modified asphalt when the rubber is recycled from
ture susceptibility. Polymer additives, such as styrene- automotive and truck tires. The United States Federal Highway
butadienestyrene (SBS), styrene-butadiene rubber (SBR), and ethy- Administration (FHWA) conducted a life cycle cost analysis (LCCA)
lene vinyl acetate (EVA), have seen success when used in heavily for modified asphalt mixtures, which indicated that CR modified
trafficked pavements. Among various polymer modifiers, SBS is asphalt is the most cost effective among all evaluated materials
probably the most widely used, although the addition of SBS may [4]. Furthermore, high modulus asphalt binders (HMABs) which
are normally manufactured from hard-grade asphalt, rock asphalt
modification, and polyolefin modification, are widely adopted in
several European countries as well as in South Africa, China, and
⇑ Corresponding author. Korea for enhancing the rutting resistance of pavement structure
E-mail addresses: wangchao@bjut.edu.cn (C. Wang), 472110053zld@163.com (L. [5–8]. High viscosity (HV) modified asphalt, polyphosphoric acid
Zhao), wcao@lsu.edu (W. Cao), caodongwei@vip.126.com (D. Cao), tbb73@yahoo.
(PPA) modified asphalt, and also various alternative binders from
com (B. Tian).

http://dx.doi.org/10.1016/j.conbuildmat.2017.07.151
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
696 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703

biomass have been recently investigated for sustainable asphalt 2. Materials and testing
pavement technology [9–11]. In summary, various modified
2.1. Materials
asphalt binders have been verified for improved paving perfor-
mance in both laboratory and field. However, it should be noted In this study, a total of eight modified asphalt binders were evaluated, which
that the asphalt pavement infrastructure is becoming increasingly covered two typical SBS binders, HV binder, SBS + HV compound modified binders,
costly in terms of construction and maintenance during service life CR modified binder, and two different HMABs. Details of the various modifiers are
summarized in Table 1. All these modified binder materials are frequently applied
due to asphalt modification. Therefore, in practice there is an
to heavily trafficked pavement structures and/or extreme climate conditions in
urgent need to comprehensively quantify the performance China. The material cost of each binder is also given in Table 1. The prices of SBS
improvement achieved by asphalt modification and to optimize binders and CR binder are directly provided by the material producer. The costs
material selection by incorporating the cost-effectiveness analysis. of HV, two SBS + HV and HMABs binders are calculated using the unit prices of
A significant amount of research has been conducted to develop the base binders and modifiers. The final costs of all modified binders are utilized
to analyze the cost-effectiveness of binders later.
a performance related purchase specification for asphalt binder.
During the United States Strategic Highway Research Program
(SHRP), a series of rheological test procedures were proposed to 2.2. Testing methods
determine the performance grade (PG) of asphalt binder with the
The modified asphalt binders were subjected to a short-term aging condition in
use of dynamic shear rheometer (DSR) and bending beam rheome-
the conventional rolling thin-film oven (RTFO) test (AASHTO T 240) to simulate the
ter (BBR) [12]. However, the SHRP specification parameters are asphalt aging process during mixture production and pavement construction [30].
measured within the linear viscoelastic (LVE) domain of asphalt All RTFO-aged binders were tested using an Anton Paar MCR 302 Rheometer, which
binder, and thus no damage related material characteristics were is capable of all the rheological tests conducted in rotational and oscillatory modes.
accounted for in distinguishing the binder performance. Besides, The modularity of the system allows the integration of a wide range of temperature
devices and application-specific accessories. The standard 25-mm parallel plates
the SHRP study was mainly conducted with neat (unmodified) with 1-mm gap and 8-mm parallel plates with 2-mm gap configurations were
asphalt materials. These existing limitations were addressed in employed respectively for the high and intermediate temperature performance
the subsequent NCHRP Project 9–11 which focused on modifica- testing.
tions to the SHRP binder specification in order to accommodate
modified binders. Several new damage-based testing approaches
2.2.1. Multiple stress creep recovery (MSCR) test
for rutting and fatigue resistance were explored and developed, Each cycle in the MSCR test consists of a creep load of 1-s duration followed by
based on which the role of modified binder in rheology and dam- 9-s recovery with zero load. The load profile consists of ten creep-recovery cycles
age resistance behavior of asphalt mixtures were more clearly under the creep stress of 0.1 kPa immediately followed by another ten creep-
recovery cycles with an increased creep stress level of 3.2 kPa [16]. The total testing
observed [13,14]. However, these new test procedures from the
time is 200 s. A typical strain history from the MSCR test with the SBS-R binder is
NCHRP 9–11 study were time-consuming and experimentally given in Fig. 1.
intense, and thus finally were not implemented for binder The performance indicators of the MSCR test are percent recovery (R) and non-
specification. recoverable compliance (Jnr). For a given creep-recovery cycle, R and Jnr are calcu-
In recent years, a continuing effort is being made to improve the lated according to Eqs. (1) and (2), respectively:

PG specification system of asphalt binder by incorporating new cp  cn

damage-based performance tests [15]. The multiple stress creep R¼ ð1Þ
cp  c0
recovery (MSCR) test (AASHTO TP 70) was developed as a specifi-
cation procedure for evaluating the rutting potential of asphalt bin-
cn  c0
ders at high temperature. This methodology was successfully Jnr ¼ ð2Þ
s
verified by correlating the obtained material parameter to rutting
performance of asphalt mixtures and field pavements [16–18]. where, c0 represents the shear strain at the beginning of the cycle, cp is the peak
strain after 1 s creep loading, cn represents the non-recoverable strain at the end
On the other hand, to develop a new fatigue evaluation approach
of this cycle after 9 s of recovery, and s represents the creep stress in each cycle.
for asphalt binder is challenging, as the test procedure should be For each stress level, the R and Jnr values were averaged from the 10 creep-
practical to implement, without the time-consuming efforts typi- recovery cycles. Thus, totally four parameters, R0.1, Jnr0.1, R3.2, and Jnr3.2 can be deter-
cally associated with conventional fatigue testing. Johnson et al. mined, and Jnr3.2 is currently the specification parameter for distinguishing the rut-
[19–21] developed the linear amplitude sweep (LAS) test (AASHTO ting resistance of asphalt binders.

TP 101) as an accelerated fatigue procedure to predict the binder

fatigue life under cyclic loading. Recently, Wang et al. [22] 2.2.2. Linear amplitude sweep (LAS) test
improved the prediction accuracy of the LAS-based binder fatigue The standardized LAS procedure (AASHTO TP101) consists of two steps [19].
life by establishing a unified energy-based failure criterion. Mean- First, a nondestructive frequency sweep test is conducted to determine the undam-
aged material response. Second, a linear oscillatory strain sweep with strain ampli-
while, the newly released AASHTO TP 123 [23] procedure specifies
tudes ranging from 0.1% to 30% within 5 min is employed to assess the asphalt
binder testing at intermediate temperature using a monotonic con- binder damage tolerance. The LAS test data interpretation is built upon the
stant shear rate loading, which consists of two performance tests, simplified-viscoelastic continuum damage (S-VECD) theory of asphalt concrete fati-
namely the binder yield energy test (BYET) and DSR-based elastic gue modeling [31–33].
recovery (DSR-ER) test. Measuring the binder yield properties from Recently, an energy-based failure analysis approach is proposed for the LAS test
to more accurately simulate the control-strain cyclic fatigue life (Nf) of asphalt bin-
BYET has been shown promising for predicting fatigue cracking der, in which the LAS loading duration of 5 min were respectively extended to
and even thermal cracking at low temperature [24–26]. Besides, 10 min and 15 min [22]. This improved LAS-based binder fatigue evaluation
the ER test of asphalt binder also provides a reliable means to char- approach is able to produce three material characteristic functions, i.e., dynamic
acterize asphalt modification with a traditional ductilometer [27]. shear modulus mastercurve, damage characteristic curve, and failure criterion.
Fig. 2 (a)–(c) respectively presents the three material functions of the SBS-R binder,
Efforts has been successfully conducted to measure binder elastic
followed by the final Nf prediction result in Fig. 2 (d). The damage characteristic
recovery on DSR, which is the main device in the current PG spec- curve, which is independent of loading history, presents a unique relationship
ification [28,29]. between the material integrity indicated by pseudo stiffness (C) and the internal
This paper presents a framework to establish a paving perfor- state variable of damage intensity (S). The failure criterion gives the characteristic
mance index (PPI) for modified asphalt binders based on both function between the releasing rate of pseudo strain energy (GR) and the fatigue life
Nf, which is also unique for any loading histories. The details regarding the estab-
the advanced performance tests and material costs. It is aimed to lishments of the LAS-based binder fatigue modeling approach are provided else-
provide an effective approach to the selection of optimum binder where [22]. The predicted fatigue life for 3% strain, denoted as 3%Nf as shown in
modification for asphalt materials in practice. Fig.2(d), is utilized for fatigue performance comparison in this study.
 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703 697

Table 1
Summary of tested asphalt binders.

Asphalt Binder Designation Modifier Unit Price (USD/ton)

Base Neat Binder Modifiers Modified Binder
SBS Binders SBS-R 4% SBS radial polymer N/A N/A 538
SBS-L 4% SBS linear polymer N/A N/A 538
High Viscosity Binder HV 4% High viscosity additives 308 6154 554
SBS + HV Binders SBS-R + 5%HV 4% SBS radial polymer + 5% high viscosity additives N/A 6154 846
SBS-R + 8%HV 4% SBS radial polymer + 8% high viscosity additives N/A 6154 1031
Crumb Rubber Binder CR 20% Crumb rubber additives N/A N/A 385
High Modulus Asphalt Binders RA-HMAB 0.3% Rock asphalt additives 308 2000 314
PR-HMAB 0.3% Polyolefin additives 308 2462 315

Note: N/A = not available.

3. Individual performance evaluation

In this section, the performance test results of the eight modi-

fied binders are presented and compared, followed by the estab-
lishment of a centesimal score system for each individual
performance index (IPI). The developed performance ranking
approach utilizes the same concept that is currently applied in
China for pavement condition assessments in terms of rutting,
cracking, roughness, and skid resistance.
A two-step analysis approach was developed in this study to
obtain the centesimal score for each performance indicator. For
this purpose, a reference value, denoted as pref, should be first
selected and was set as the maximum level of the performance
indicator under consideration. The IPI was then determined as
pi
IPI ¼  100 ð4Þ
Fig. 1. Typical MSCR test data of the SBS-R binder. pref

where, pi is the value of the performance indicator of each binder.

2.2.3. Binder yield energy test (BYET)
The BYET is designed for measuring the binder yield characteristics at interme- 3.1. Rutting resistance index (RRI)
diate temperature. The prepared binder sample is tested at the desired test temper-
ature at which a monotonic constant shear strain rate of 2.315% s–1 is applied to the
sample. Both stress and strain are recorded at a sampling rate of one data point per The time histories of strain responses from the MSCR test of all
two seconds. The test is concluded once an obvious peak on the stress-strain curve binders are shown in Fig.5(a). It can be observed that the PR-HMAB
is achieved. The BYET-based performance parameter is named yield energy (Wy), binder exhibited the best permanent deformation resistance. Fig. 5
which is quantified as the area under the curve up to the peak shear stress. A typical
(b) and (c) present the results of R and Jnr under the two creep
BYET result of the SBS-R binder is presented in Fig. 3.
stress levels, respectively. Trends in Fig.5(b) demonstrate that the
strain recovery of the two HMABs are much lower than the other
2.2.4. DSR-Elastic recovery (DSR-ER) test modified binders. Meanwhile, the compound modified SBS + HV
The DSR-ER test is developed as a surrogate test procedure for measuring the binders show slightly higher R values than the SBS and HV binders.
elasticity of modified asphalt binders at intermediate temperature. First, a constant
strain rate loading of 2.315% s–1 is applied to the binder sample until a strain of
Similar trends can be found for the Jnr results as compared in Fig.5
277.78% is achieved (within approximately 2 min). Afterwards, a recovery step is (c), in which Jnr of HMABs and CR binder were lower than those for
carried out by applying zero shear stress to the sample for 30 min. Both stress the SBS and HV binders. In addition, increase of the HV additive in
and strain are recorded at a sampling rate of one data point per two seconds the SBS binder reduced Jnr to a level close to those of HMABs and CR
throughout the loading and recovery periods. A typical DSR-ER test result of the
binder. The better rutting resistance of CR binder than SBS binders
SBS-R binder is given in Fig. 4. The performance indicator of ER is defined in Eq. (3):
has also been reported in other studies based on FHWA-
c1  c2 Accelerated Loading Facility (ALF) both in laboratory MSCR testing
ER ¼ ð3Þ
c1 and field rutting measurements [17].
In order to further investigate the elastic response of the mod-
where, c1 is the strain at the end of the loading period, and c2 is the residual strain at ifiers used in this study, the relations of Jnr3.2 versus R3.2 are plotted
the end of the recovery period.
in Fig. 5(d) and compared with the reference curve provided in
AASHTO TP 70 to identify the presence of elastomeric polymer
2.3. Experimental plan [16]. It is found that only the two HMABs fell below the reference
curve, which confirmed that the SBS, HV and CR binders used in
The test plan of this study is summarized in Table 2. The MSCR test of all binders
this study were modified with acceptable elastomeric polymers.
were conducted at a single temperature of 60 °C rather than the corresponding PG
temperatures to consistently compare the rutting resistance. A typical intermediate Since the MSCR-based performance indicator, Jnr3.2, is inversely
temperature of 20 °C in Beijing area is selected for LAS, BYET, and DSR-ER tests [34]. proportional to the rutting resistance, 1/Jnr3.2 was utilized for cal-
Regarding the performance indicators, the predicted 3%Nf is utilized for the LAS test culating the centesimal score of RRI. According to the previously
and the specification parameters are employed for MSCR, BYET, and DSR-ER tests. introduced two-step analysis approach, the maximum value of 1/
For most modified binders, normally two replicates were conducted for all perfor-
mance tests to reduce experimental variability. A third or even more replicates
Jnr3.2 (from the PR-HMAB binder) was selected as the reference,
were tested for the SBS + HV compound modified binders in certain cases to ensure and the RRI values for all binders were then determined as given
the coefficient of variation was within 10 percent. in Table 3.
698 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703

Fig. 2. Typical LAS test results of the SBS-R binder (a) dynamic modulus mastercurve (b) damage characteristic curve (c) failure criterion (d) fatigue life prediction.

Fig. 3. Typical BYET result of the SBS-R binder. Fig. 4. Typical DSR-ER test result of the SBS-R binder.

3.2. Fatigue resistance index (FRI) in this study demonstrated better fatigue resistance than the linear
SBS modifier, which is probably due to the better cross-linked net-
Applying the LAS-based fatigue modeling approach, the control- work of polymer structure for the radial modifier [15]. Addition-
strain fatigue life of all modified binders were simulated and the ally, Wang et al. [22] previously demonstrated better fatigue
results are compared in Fig. 6. It can be observed that CR binder resistance of CR binder than SBS binder from the FHWA-ALF mate-
exhibited the best fatigue resistance, followed by compound SBS rials using the LAS testing approach, which is also consistent with
+ HV binders, SBS binders, HV and RA-HMAB binder. The PR- the LAS analysis results in this study.
HMAB binder showed the lowest fatigue resistance. It is also noted The predicted 3%Nf was employed as the performance indicator
that increasing the HV content from 5% to 8% into the SBS-R binder for fatigue resistance. As given in Table 4, the 3%Nf of the CR binder
improved the fatigue performance. Besides, the radial SBS modifier
 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703 699

Table 2 exhibited an entirely different yield process featured with signifi-

Test plan. cantly high yield stresses and low yield strains. Additionally, the
Performance Temperature Test Condition Performance two compound SBS + HV binders presented a multiple-peak phe-
Test Indicator nomenon which was further verified for repeatability using addi-
MSCR Test 60 °C 1 s creep + 9 s recovery and Jnr3.2 tional replicates. This multiple-peak behavior was confirmed in
10 repeated cycles for each existing literature when highly modified binders were tested
creep stress; [24]. In this paper, the first stress peak was identified for the SBS
Creep stress levels: 0.1 kPa
and 3.2 kPa.
+ HV binders and employed for the yield energy (Wy) calculation.
LAS Test 20 °C Frequency sweep at 5 °C, 3%Nf The SBS-L binder yielded the maximum Wy value and was chosen
20 °C and 35 °C; as the reference for YRI quantification for all binders as given in
LAS tests conducted Table 5.
respectively within 5 min,
10 min and 15 min at 20 °C.
BYET 20 °C Constant strain rate: 2.315% Wy
s1,
Continuous loading 3.4. Elastic recovery index (ERI)
DSR-ER Test 20 °C Constant strain rate: 2.315% ER
s1,
2 min loading + 30 min The strain responses of all binders from the DSR-ER tests are
recovery given in Fig. 8. The results again demonstrate a better deformation
recovery potential of SBS, HV, and CR modified binders than that of
the two HMABs. In addition, incorporation of more HV additives
was set as the reference and then the centesimal score for each into the SBS-R binder produced an increased ER level. Recall that
binder was obtained according to Equation (4). a similar phenomenon has been observed in percent recovery (R)
in Fig. 5(b) from the MSCR test. Fig. 9 shows a reasonable correla-
tion between ER, obtained at a constant strain rate at intermediate
3.3. Yield resistance index (YRI) temperature for identifying recovery, and R3.2 that was evaluated
using a constant creep stress for recovery at high temperature. It
The stress-strain responses of all modified binders in the BYET is therefore indicated that deformation recovery can be deemed
test are illustrated in Fig. 7. Compared with the modifiers that were as a characteristic material property that is independent of loading
previously identified as the elastomeric polymers, the two HMABs mode and temperature.

Fig. 5. MSCR test results: (a) time histories of strain responses, (b) percent recovery (R), (c) non-recoverable compliance (Jnr), and (d) relationship between Jnr3.2 and R3.2.
700 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703

Table 3
MSCR test results and RRI.

Asphalt Binder Jnr3.2 (kPa1) 1/Jnr3.2 (kPa) Reference 1/Jnr3.2 (kPa) RRI
SBS-R 0.134 7.46 26.84 27.8
SBS-L 0.114 8.79 32.7
HV 0.136 7.34 27.3
SBS-R + 5%HV 0.090 11.06 41.2
SBS-R + 8%HV 0.060 16.77 62.5
CR 0.046 21.79 81.2
RA-HMAB 0.054 18.66 69.5
PR-HMAB 0.037 26.84 100.0

Table 5
BYET test results and YRI.

Asphalt Binder Wy Reference Wy YRI

SBS-R 1723048 1975136 87.2
SBS-L 1975136 100.0
HV 1882915 95.3
SBS-R + 5%HV 1079635 54.7
SBS-R + 8%HV 1666455 84.4
CR 525040 26.6
RA-HMAB 208765 10.6
PR-HMAB 188808 9.6

Fig. 6. LAS-based control-strain fatigue predictions.

Table 4
LAS-based fatigue prediction results and FRI.

Asphalt Binder 3%Nf Reference 3%Nf FRI

SBS-R 21772 43416 50.1
SBS-L 15869 36.6
HV 9166 21.1
SBS-R + 5%HV 25642 59.1
SBS-R + 8%HV 37100 85.5 Fig. 8. DSR-ER test results.
CR 43416 100.0
RA-HMAB 9360 21.6
PR-HMAB 5026 11.6

Fig. 9. Relationship between ER and R3.2.

Fig. 7. BYET stress-strain curves.

4. Development of paving performance index (PPI)

As previously mentioned, the increased material cost due to

Table 6 presents the centesimal scores of ERI for all tested bin- asphalt modification is a continuous concern in practice, and a
ders, which were calculated using the maximum ER value from the relationship between binder price and the expected performance
SBS-R + 8%HV binder as reference. is thus of tremendous benefit for the purchase specification of
 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703 701

Table 6 ever, in practice weight assignment is associated with project needs

DSR-ER test results and ERI. which include typical distress types under the in-situ traffic and cli-
Asphalt Binder ER Reference ER ERI mate conditions.
SBS-R 67% 87% 76
Fig. 11 illustrates the relationship between binder cost and PPI
SBS-L 49% 57 in four different cases using various weight percentages. In Case A,
HV 41% 47 an equal weight of 0.25 was applied to each IPI. Different weights
SBS-R + 5%HV 82% 94 were assigned in other cases. The current pavement condition
SBS-R + 8%HV 87% 100
assessments in China consider four main performances, i.e., rut-
CR 63% 72
RA-HMAB 14% 16 ting, cracking, skid resistance and roughness, in which the weights
PR-HMAB 13% 15 assignments are respectively 0.15, 0.35, 0.1 and 0.4. For Case D, the
RRI and FRI weight percentages are respectively the same with the
weights for rutting and cracking (i.e., 0.15 and 0.35), while the
modified binders. In this study, the current material cost for the weights of YRI and ERI correspond to those for skid resistance
tested modified binders are calculated based on the prices of the and roughness, respectively (i.e., 0.1 and 0.4).
neat binder and various modifiers. The modified binder cost was As shown in Fig. 11, reasonable correlations can be observed
then plotted with respect to the identified IPI including RRI, FRI, between PPI and binder cost in all cases when the CR binder was
YRI, and ERI, as shown in Fig. 10 (a)–(d). It can be seen that no clear excluded. It is indicated that multiple performance indices are
correlation exists between binder cost and each IPI, which suggests indeed necessary to assess the cost-effectiveness of asphalt
that it may be more appropriate to evaluate the cost effectiveness modification.
of asphalt modification based on a combined index instead of indi- In order for a better comparison, the binder performance per
vidual ones. unit price, denoted as P/C, was obtained, which was defined as
Given the above discussion, a comprehensive paving perfor- the ratio of PPI to the corresponding binder cost. For all tested
mance index (PPI) of modified asphalt binder is proposed in this modified binders, the P/C values were summarized in Fig. 12 for
paper. By applying a weight percentage to each IPI, all IPIs are com- various cases as described in Fig. 11. It was found that P/C was rel-
bined and the resulting weighted average is defined as PPI: atively stable for each binder in different cases in this study. For
most modified binders, except the CR binder, an averaged P/C level
PPI ¼ xRRI RRI þ xFRI FRI þ xYRI YRI þ xERI ERI ð5Þ of 0.09 was identified and shown as the dashed horizontal line in
where, xRRI, xFRI, xYRI and xERI represent the weight percentage for Fig. 12, while the averaged P/C of the CR binder is approximately
RRI, FRI, YRI, and ERI, respectively. The final PPI value is certainly 0.19 as indicated by the solid line. The observed high cost-
dependent upon the weight percentage assigned to each IPI. How- effectiveness of the CR binder is consistent with the findings in

Fig. 10. Relationship between binder costs and IPI (a) RRI (b) FRI (c) YRI (d) ERI.
702 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703

Fig. 11. Relationship between binder cost and PPI using various weight percentages: (a) Case A, (b) Case B, (c) Case C, and (d) Case D.

and DSR-ER tests. A centesimal score system was developed for

the PPI parameter that is quantified using the IPIs representing rut-
ting potential, fatigue resistance, yield property, and elastic recov-
ery. Specific findings of this study are:

(1) The two HMABs presented the best rutting resistance but
relatively poor performance in the other respects. The CR
modified binder demonstrated the highest fatigue resistance
among all binders.
(2) The modifiers used in this study for SBS, HV and CR binders
were identified as acceptable elastomeric polymers, and
increase in the HV content in SBS binder increased the rut-
ting, fatigue, and elastic recovery performance.
(3) The laboratory measured performance indicators were con-
verted into a centesimal score of IPI based on a two-step
analysis approach. These IPIs failed to correlate with the
Fig. 12. P/C comparison for the tested modified asphalt binders under various prices of the modified binders.
percent weights. (4) Under various assignments of weight percentages, a reason-
able linear relationship was observed between PPI and bin-
der costs when the CR binder was excluded, which
the LCCA study conducted by U.S. FHWA [4], and is due to the fact suggests that a comprehensive consideration of multiple
that rubber is normally recycled from used automotive tires. performance indices is required to assess the cost-
effectiveness of asphalt modification.
5. Summary and conclusions (5) The ratio of binder PPI to cost, P/C, represents the perfor-
mance achieved for a unit price. It was found that the aver-
This paper presented the development of a comprehensive PPI aged P/C value of the seven modified binders other than the
system for evaluating the cost-effectiveness of modified asphalt CR binder was fairly consistent, approximately 0.09,
binders. Materials involved in the experimental study included whereas the CR binder exhibited a higher averaged P/C of
SBS, HV, compound SBS + HV, CR binders and HMABs. A suite of
performance tests were performed, including MSCR, LAS, BYET,
 C. Wang et al. / Construction and Building Materials 153 (2017) 695–703 703

0.19. Therefore, the CR modified binder is identified as the [14] H.R. Soleymani, H. Zhai, H. Bahia, Role of modified binders in rheology and
damage resistance behavior of asphalt mixtures, Transp. Res. Rec. 1875 (2004)
most cost-effective among the binders evaluated in this
70–79.
study. [15] Transportation Research Circular E-C147. Development in asphalt binder
specifications. Papers from a Workshop. Papers from a Workshop, January,
Future work is suggested to explore and provide a guideline on 2008, Washington, D.C.
[16] AASHTO. Standard method of test for multiple stress creep recovery (MSCR)
the selection of weight assignment that addresses the design test of asphalt binder using a dynamic shear rheometer (DSR). AASHTO TP 70,
needs, and local traffic and climate characteristics. 2010, Washington D.C.
[17] J. D’Angelo, The relationship of the MSCR test to rutting, Road Mater. Pavement
Design 10 (S1) (2009) 61–80.
Acknowledgements [18] J. D’Angelo, R. Dongré, Practical use of multiple stress creep and recovery test,
Transp. Res. Rec. 2126 (2009) 73–82.
The authors would like to acknowledge the opening funding [19] AASHTO. Standard method of test for estimating damage tolerance of asphalt
binders using the linear amplitude sweep. AASHTP TP 101, 2014, Washington
supported by the key Laboratory of Road Structure & Material, D.C.
Ministry of Transport, Beijing, P.R.C. The supports from National [20] C. M. Johnson. Estimating asphalt binder fatigue resistance using an
Natural Science Foundation of China (Grant No. 51608018) and accelerated test method. 2010, Ph.D. dissertation, University of Wisconsin-
Madison, Madison, WI.
Beijing Natural Science Foundation (Grant No. 8174059) are also [21] C. Hintz, H. Bahia, Simplification of linear amplitude sweep test and
gratefully acknowledged. specification parameter, Transp. Res. Record: J. Transp. Res. Board 2370
(2013) 10–16.
[22] C. Wang, C. Castorena, J. Zhang, Y.R. Kim, Unified failure criterion for asphalt
7. References
binder under cyclic fatigue loading, J. Assoc. Asphalt Paving Technol. 84 (2015)
269–299.
[1] Y. Becker, M.P. Mendez, Y. Rodriguez, A polymer modified asphalt, Vision [23] AASHTO. Measuring asphalt binder yield energy and elastic recovery using the
Technol. 9 (2001) 39–50. dynamic shear rheometer. AASHTO TP 123, 2016, Washington D.C.
[2] Y. Yildirim, Polymer modified asphalt binders, Constr. Build. Mater. 21 (2007) [24] C.M. Johnson, H. Wen, H.U. Bahia, Practical application of viscoelastic
66–72. continuum damage theory to asphalt binder fatigue characterization, J.
[3] G. King, H. King, R.D. Pavlovitch, A.L. Epps, P. Kandahl, Additives in Asphalt, J. Asphalt Paving Technol. 78 (2009) 597–638.
Assoc. Asphalt Paving Technol. 68 (1999) 32–69. [25] H. Wen, S. Shen, Z. Ma, J. Wang, Modeling the effects of temperature and
[4] R.G. Hicks, J.A. Epps. Life cycle cost analysis of asphalt-rubber paving materials, loading rate on fatigue property of asphalt binder, J. Test. Eval. 38 (6) (2010)
in Proceedings of the First International Conference on the World of Asphalt 647–652.
Pavements, Sydney, Australia, 2000, 69–88. [26] H. Wen, S. Bhusal, Toward development of a new thermal cracking test using
[5] J.-F. Corte, Development and uses of hard-grade asphalt and of high-modulus the dynamic shear rheometer, J. Test. Eval. 41 (3) (2013) 1–8.
asphalt mixes in France, Transp. Res. Circular 503 (2001) 12–31. [27] M. Walaa, A. Alexander, E.K. Muhammed, F. Zhou, Evaluation of binder elastic
[6] J. Komba, J. Anochie-Boateng, J. O’Connell, B. Verhaeghe, Long-term pavement recovery on HMA fatigue cracking using continuum damage and overlay test
performance monitoring and the revision of performance criteria for high based analyses, Road Mater. Pavement Des. 12 (S2) (2011) 345–376.
modulus asphalt in South Africa, Proc. Roles Accelerated Pavement Testing [28] C.S. Clopotel, H. Bahia, Importance of elastic recovery in the DSR for binders
Pavement Sustainability (2016) 177–194, http://dx.doi.org/10.1007/978-3- and mastics, Eng. J. 16 (4) (2012) 99–106.
319-42797-3_12. [29] H.A. Tabatabaee, C. Clopotel, A. Arshadi, H. Bahia, Critical problems with using
[7] X. Zou, A. Sha, W. Jiang, X. Huang, Modification mechanism of high modulus the asphalt ductility test as a performance index for modified binders, Transp.
asphalt binders and mixtures performance evaluation, Constr. Build. Mater. 90 Res. Rec. 2370 (2013) 84–91.
(2015) 53–58. [30] AASHTO. Standard method of test for effect of heat and air on a moving film of
[8] H.J. Lee, J.H. Lee, H.M. Park, Performance evaluation of high modulus asphalt asphalt binder (rolling thin-film oven test). AASHTO T 240, 1997, Washington
mixtures for long life asphalt pavements, Constr. Build. Mater. 21 (2007) 1079– D.C.
1087. [31] Y.R. Kim, H.J. Lee, D.N. Little, Fatigue characterization of asphalt concrete using
[9] L. Li, H. Geng, Y. Sun, Simplified viscosity evaluating method of high viscosity viscoelasticity and continuum damage theory, J. Assoc. Asphalt Paving
asphalt binders, Mater. Struct. 48 (2015) 2147–2156. Technol. 66 (1997) 520–569.
[10] Transportation Research Circular E-C160. Polyphosphoric acid modification of [32] J.S. Daniel, Y.R. Kim, Development of a simplified fatigue test and analysis
asphalt binders. Papers from a Workshop, April 7–8, 2009, Minneapolis, procedure using a viscoelastic damage model, J. Assoc. Asphalt Paving Technol.
Minnesota. 71 (2002) 619–650.
[11] Transportation Research Circular E-C165. Alternative binder for sustainable [33] B. S. Underwood. Multiscale constitutive modeling of asphalt concrete. 2011,
asphalt pavements. Papers from a Workshop, January 22, 2012, Washington, Ph.D. Dissertation, North Carolina State University, Raleigh, N.C.
D.C. [34] J. Zhang, C. Jing, D. Han, and C. Wang. Study on working conditions of
[12] D.A. Anderson, T. Kennedy, Development of SHRP binder specification, J. expressway asphalt pavement by monitoring test road. Proceedings of 4th
Asphalt Paving Technol. 62 (1993) 481–507. International Chinese European Workshop on Functional Pavement Design
[13] H. Bahia, D.I. Hanson, M. Zeng, H. Zhai, M.A. Khatri, R.M. Anderson. (CEW 2016), Delft, Netherlands, 29 June–1 July, 2016, 1585–1594.
Characterization of modified asphalt binders in superpave mix design.
NCHRP Report 459, Transportation Research Board, National Research
Council, Washington D.C., 2001.