Construction and Building Materials 150 (2017) 157–166

Contents lists available at ScienceDirect

Construction and Building Materials

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

Quality verification of polymer-modified asphalt binder used in hot-mix

asphalt pavement construction
Yuhong Wang a,⇑, Dan Chong b, Yong Wen a
a
Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong, China
b
Department of Management Science and Engineering, Shanghai University, Shanghai 200444, China

h i g h l i g h t s

IDT strength surpasses ITSM in testing asphalt quality difference in cored samples.

Rolling-thin film oven does not well simulate the short-term aging of PMA.

Phase angle surpasses complex shear modulus in testing quality difference of PMA.

The ratio of FTIR peak area surpasses the ratio of peak value in testing PMA quality.

FTIR with phase angle of extracted PMA offers a promising quality control method.

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

Article history: In many regions and countries, highway agencies are interested in verifying the quality of asphalt binders
Received 25 October 2016 placed on roads for quality assurance purpose. Particularly, as PG-graded, polymer modified asphalt
Received in revised form 3 March 2017 (PMA) is increasingly used, the agencies are interested in examining if the specified PMA is fully used
Accepted 23 May 2017
in the field or if the PMA has not been severely degraded after storage, handling, and plant production.
This research investigated and compared various test methods for such purpose. Besides laboratory-
prepared asphalt binder samples, actual pavements built with neat asphalt (pen 60/70), 30% of neat
Keywords:
asphalt with 70% of PG76-16 PMA, 15% of neat asphalt with 85% of PMA, and 100% of PMA were used
Asphalt pavement
Polymer-modified asphalt
for study. The mixture production and construction processes were carefully planned and monitored
Quality control to ensure that they are consistent with real construction operations. Test methods evaluated in this study
Quality assurance include indirect tensile (IDT) strength and indirect tensile stiffness modulus (ITSM) of field cores, rheo-
Polymer content logical parameters (G⁄, d, temperatures at which G⁄/sind = 2.2 kPa) from dynamic shear rheological tests
Field test of binders, Gel Permeation Chromatography (GPC) and Fourier Transform Infrared Spectroscopy (FTIR).
The study found that samples made of different proportions of PMA can be effectively differentiated
by IDT strength, phase angle (d), and FTIR spectrum. Even the FTIR spectra of extracted asphalt binders
without being subject to strict binder extraction and evaporation procedure reveal useful information
for quality assurance purpose. The study also further proved that RTFO-treated PMA samples cannot
be used to establish quality criteria for field-extracted PMA binders because RTFO does not well simulate
field aging of PMA.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction the quality of asphalt pavement is the quality of asphalt binder.

Quality of built pavement depends on ‘‘all those planned and
systematic actions [1]” to ensure quality. Quality assurance, there-
fore, should cover all the stages in the life-cycle of pavement and
involve all the quality-related activities by agencies, contractors,
and consultants [1]. One of the fundamental factors that decide
⇑ Corresponding author.
E-mail addresses: ceyhwang@polyu.edu.hk (Y. Wang), chongdan@shu.edu.cn (D.
Chong), yong.wen@connect.polyu.hk (Y. Wen).
In recent decades, polymer modification has been increasingly
used around the world to improve asphalt pavement’s high tem-
perature performance [2–4] without degrading its low tempera-
ture performance. Polymer modified asphalt (PMA), however, is
more expensive than neat asphalt binder: It may increase binder
cost from 30% to 100% [5,6]. In addition, PMA may become unsta-
ble when being stored at high temperature for a prolonged period
of time [7–10] and may be subjected to degradation in mixture
production, transportation, and construction [11].

http://dx.doi.org/10.1016/j.conbuildmat.2017.05.196
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
158 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166
To ensure asphalt binder quality, standard procedures such as
those specified in the AASHTO standard M320 have been routinely
used by agencies to check the SuperpaveTM performance grading
(PG) of asphalt binder. In addition, various SuperpaveTM plus spec-
ifications have been made to specifically test PMA in addition to
the test procedures in AASHTO standard M320 [12]. However, for
most agencies, asphalt binder tests are limited to binder samples
provided by suppliers (or retrieved by agencies themselves) before
mixture production [13]. It is assumed that the specified binder
type will not be changed or blended with low-quality binder dur-
ing mixture production, nor be severely degraded by improper pro-
duction conditions. From the perspective of quality assurance, this
assumption may not be always safe [14], particularly considering
the possibly large price difference between PMA and neat asphalt
binder. In addition, quality loss due to improper binder storage
and mixture production cannot be adequately assessed. This cre-
ates a concern in the quality assurance procedure of asphalt
pavement.
One of the approaches to improving binder quality assurance
program is to expand asphalt binder sampling locations [13]. It is
recommended in Texas that sampling locations such as the storage
tank of contractor site, transfer line from storage tank to hot-mix
asphalt (HMA) plant, spray bar of asphalt emulsion distributor
truck may be considered [13]. This practice, however, would
require full knowledge of HMA production and construction sched-
ules. Any change in production and construction will disrupt the
sampling plan and create a logistic difficulty for agencies. Another
approach, as adopted by a number of European countries, is to
extract and recover PMA from cored field samples and conduct ver-
ification tests on the extracted binder [14]. Germany and several
other countries legally require that the extracted PMA to be tested
for softening point and elastic recovery after being recovered from
rotary evaporation extraction [14]. These two test methods, how-
ever, demand a relatively large quantity of recovered asphalt bin-
der. The binder extraction and recovery process using a rotary
evaporator is both time consuming and expensive. Questions have
also been raised on the accuracy of using these two methods to
verify binder type [14]. Therefore, more efficient and reliable
methods to verify the type and quality of asphalt binder actually
placed in the field are highly interested by many highway agencies
worldwide.
The overall goal of this study is to evaluate test methods that
can be potentially used to forensically verify the type and quality
of PMA used in asphalt mixture production. Particularly interested
Fig. 1. The analytic method of this study.
are methods to differentiate PG-graded PMA with neat non-
modified asphalt binder, binder artificially created by blending a
portion of PMA with neat asphalt, or binder that is subjected to
severe polymer degradation due to material storage and handling
issues. The reliability, costs, and amount of efforts of the potential
test methods are also part of the consideration. The detailed objec-
tives include:

To compare change in the asphalt mixture properties of a same
mixture design using PG-graded PMA blended with different
proportions of neat asphalt binder;

To compare change in the dynamic shear properties of PG-
graded PMA blended with different proportions of neat asphalt
binder;

To compare change in the chemical indicator characteristics of
PG-graded PMA blended with different proportions of neat
asphalt binder;

To assess different test methods and recommend a feasible pro-
cedure to evaluate the quality of PMA in as-constructed asphalt
mixtures.
This research is expected to have practical values for highway
agencies who are interested in verifying the specified quality
and/or the sufficiency of polymer content in PMA used in asphalt
pavement construction.
2. Research methods
Styrene-Butadiene-Styrene (SBS) modified PG76-16 (PG 76
hereafter) PMA and pen 60/70 neat asphalt binder are used in this
tudy. Both binders are commonly used in Hong Kong (HK), where
this study was conducted. The PG grade of the pen 60/70 binder is
PG 64-10 according to the SuperpaveTM grading system. The overall
research methods and procedures are shown in Fig. 1 and dis-
cussed in details as follows.
2.1. Sample preparation
2.1.1. Asphalt binder samples prepared in the laboratory
Both asphalt binder and mixture tests were conducted and
compared in this research. Besides 100% PG 76 and pen 60/70 bin-
der, special binders were created in laboratory by blending these
two binders at 165 °C into the following proportions: 70% PG 76
 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166 159

binder with 30% pen 60/70 binder and 85% PG 76 binder with 15% binders. In HK, asphalt binder used for pavement construction is
pen 60/70 binder. The special binders were blended by using a from a limited number of suppliers. Therefore, it may be possible
high-shear mixer for 15 min. The four types of binder (original to detect the relationship between asphalt mixture properties
and specially blended) were treated in a rolling-thin film oven and the types of asphalt binder used.
(RTFO) before being tested by a dynamic shear rheometer (DSR).
They were also tested for Fourier Transform Infrared Spectroscopy 2.2. Test methods
(FTIR) and Gel Permeation Chromatography (GPC) with two addi-
tional blended proportions: 80% PG 76 with 20% Pen 60/70 and Tests conducted on binder and mixture specimens in this study
90% PG 76 with10% Pen 60/70. can be categorized into three groups. The first group is to test the
mechanistic properties of sample cores. The cores used for IDT
strength tests were 100 mm in diameter and about 44 mm in
2.1.2. Asphalt binder and mixture samples obtained from the field
thickness, whereas those for ITSM tests were 150 mm in diameter
Besides tests on asphalt binders prepared in the laboratory,
and the same thickness. The IDT tests were conducted according to
another focus of this study was to test asphalt mixtures and bin-
ASTM D6931–12, except for the test temperature being raised to
ders extracted from field cores. Sample cores were taken from five
35 °C to better differentiate the PG 76 binder with the others.
road resurfacing projects, including three using PG 76 stone mastic
The ITSM tests were conducted at 35 °C by using the Nottingham
asphalt (SMA) mixtures and the other two using pen 60/70 SMA
Asphalt Tester (NAT).
mixtures.
The second group of tests focused on the rheological properties
In addition, two trial pavement sections were purposely con-
of various binders by using a DSR. The complex shear modulus (G⁄)
structed for this study, one using 30% pen 60/70 binder and 70%
and phase angle (d) of the asphalt binders at different temperature
PG 76 binder and the other using 15% pen 60/70 binder and 85%
levels were tested at 10 rad/s and the temperatures at which the
PG 76 binder. The mixture was produced in a batch plant according
rutting index (G⁄/sin d) reaches 2.2 kPa were determined. This tem-
to the same design as SMA. During mixture production, the two
perature provides an indication of the binder’s high temperature
types of binders were simply introduced into the pugmill of the
performance according to SuperpaveTM specification.
batch plant through separate binder supply systems without being
Chemical analysis techniques, including FTIR and GPC, were
pre-blended. The mixing process took about 1–3 min. Each pave-
employed for the third group of tests. Both methods can be poten-
ment section was about 4 m  10 m  0.05 m (thickness). The
tially used to test the availability and content of polymers in labo-
two special pavement mixtures, prepared by a special order at an
ratory asphalt samples [15,16]. In particular, FTIR has been
asphalt mixture plant in HK, were used to simulate the production
adopted as a standard test method (AASHTO T-302-05, ‘‘polymer
process in which the specified PG 76 binder was only partially
content of polymer-modified emulsified asphalt residue and
used. The transportation, placement, and compaction of the pave-
asphalt binders”). In the test, the presence of polymer can be deter-
ment mats were carefully controlled to match an actual construc-
mined by examining the change of the infrared spectra, but the
tion process. A picture of the placement and compaction of the
exact content of the polymer needs to be found through a cali-
special mixtures is shown in Fig. 2. Sample cores taken from the
brated plot from the known base asphalt binder and polymer.
actual and the special projects were tested for indirect tensile
The applications of both methods for quantitative characterization
(IDT) strength at 35 °C and indirect tensile stiffness modulus
of extracted PMA from field samples, especially blended asphalt
(ITSM) at 35 °C. Asphalt binders from the cored samples were
binders, have been limited.
extracted by using an extraction unit bowl according to ASTM stan-
dard D2171/D2172 M-11. The solvent used for extraction is
2.3. Statistical analysis methods
dichloromethane. To ensure that the fine particles were thoroughly
removed, the extracted bitumen solution was further clarified in a
The hypothesis of the statistical test was established as follows:
sample tube centrifuge machine according to the European Stan-
Null hypothesis (H0): The tested average property of a mixture
dard EN 12697–3:2005. Asphalt binder was recovered from the
or binder (l2) is equal to the tested average property of a mixture
clarified solution by using a rotary evaporator according to the
or binder that uses known PG76 asphalt binder (l1).
European Standard EN 12697–3:2005. Because all the mixtures
Alternative hypothesis (Ha): The tested average property of a
used in this study follow the same mix design (SMA), possible dif-
mixture or binder (l2) is not equal to the tested average property
ference in mixture properties is likely due to difference in asphalt
of a mixture or binder that uses known PG76 asphalt binder (l1).
Because the primary goal of this study is to compare the feasi-
bility and efficiency of the different test methods, the sample size
for each experiment is not large. The statistical procedure used for
making inference based on small sample size was employed to
compare sample means [17]. The statistic for testing the null
hypothesis (H0) l1 = l2 is [17]:t ¼ rðY^2 Y 1 Þ ; with t distribution hav-
Y 2 Y 1

ing a degree of freedom equal to n1 + n2  2,where,

l1 = the population mean of sample 1,
l2 = the population mean of sample 2,
Y 1 = average value of sample 1,
Y 2 = average value of sample 2,
n1 = the size of sample 1,
n2 = the size of sample 2,
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P 2 P 2
1 1 ðY 1i  Y 1 Þ þ ðY 2i  Y 2 Þ
r^ Y 2 Y 1 ^
¼r þ ; and r ^¼
n1 n2 n1 þ n2  2
Fig. 2. Construction of trial pavement sections on a concrete slab using asphalt
mixtures with two types of mixed asphalt binders. ð1Þ
160 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166

The t-value and the degree of freedom were used to identify the
probability associated with the t-value in a standard t distribution.
The test and statistical analysis results are summarized and dis-
cussed as follows.
3. Test results and discussion
3.1. Comparison of IDT strength and ITSM of sample cores using
different binders
Consequently, one may not be able to tell if the mixture uses the
specified PG 76 binder.
The ITSM test results showed high variation and hence failed to
detect significant difference between the average ITSM value of
specimens using PG 76 binder and that of the other groups (see
Fig. 4). Moreover, the average ITSM value of the specimens using
PG 76 is even less than that with pen 60/70. Therefore, ITSM test
appears not suitable for the purpose of binder type and quality
differentiation.

3.2. Comparison of DSR test results

The boxplot of IDT strengths of sample cores using the four dif-
ferent types of asphalt binders is shown in Fig. 3. The figure reveals
The complex shear modulus (G⁄) at 64 °C, phase angle (d) at
that the average IDT strength decreases with the increase of pen
64 °C, temperature at which the rutting index G⁄/sin (d) reaches
60/70 binder content, despite some variations among the test
2.2 kPa (threshold temperature) were used for comparison. Poly-
results. The IDT strengths of the specimens using 100% PG 76,
85% PG 76, and 70% PG 76 overlap with each other, but they are
quite different than the IDT strength of the specimens using pen
60/70 binder only.
The statistical test results based on the IDT strengths of the dif-
ferent groups are summarized and shown in Table 1. In the Table,
the alternative hypothesis for one side test is that the average IDT
strength of the mixture using 100% PG 76 binder is greater than the
average IDT strength of the other mixture to be compared. The
one-side test is more powerful than the two-side test because
the probability of making ‘‘significance” inference at a predefined
significance level is reduced to half. It can be seen that the IDT test
can generally differentiate the mixtures using 100% PG 76 binder
with those using a less portion. From the perspective of practical
applications, the IDT test may be a convenient method. The equip-
ment and method for this test are routine for highway agencies. It
appears possible for HK to use this method as a simple quality con-
trol test because the neat asphalt binder is typically pen 60/70 and
is usually from a limited number of suppliers. It needs to be recog-
nized, however, this study only examined the difference between Fig. 4. ITSM of specimens using different types of binders (Note: a: 100% PG 76, b:
the mixtures using PG 76 binder, pen 60/70, and the combinations 85% PG 76, c: 70% PG 76, d: 0% PG 76).
of the two. If a harder binder such as pen 35/50 had been chosen,
the IDT strength of the mixture using such binder may be high.

Fig. 3. IDT strengths of specimens using different types of binders (Note: a: 100% Fig. 5. Complex shear moduli G* (64 °C) of different types of binders after RTFO
PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76). (Note: a: 100% PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).

Table 1
Statistics for significance tests based on IDT strength.

Mixture Sample size Average Strength (kPa) Std. Dev. (kPa) Prob. by comparison with control (mix. using 100% PG 76)
100% PG 76 12 458.8 50.0 n.a.
85% PG 76 + 15% Pen 60/70 8 418.3 50.4 >0.01 but <0.05 for one side test
70% PG 76 + 30% Pen 60/70 7 398.7 49.9 >0.01 but <0.05 for two side test; <0.01 for one side test
100% Pen 60/70 7 250.0 50.1 <0.01 for two-side test
 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166 161

mer modification increases the binder’s elasticity, resulting in a

lower d and higher threshold temperature. The boxplots of G⁄, d,
and threshold temperatures of the four types of binder after being
treated in RTFO are shown in Figs. 5–7, respectively. The summary
of the statistical tests is shown in Table 2. The boxplots of all the
three parameters for the binders recovered from field samples
are shown in Figs. 8–10, respectively, and the test statistics are
shown in Table 3. Five specimens were tested for each combination
of binder type and treatment method.
Large variations in G⁄ are shown in Figs. 5 and 8 for all types of
binders except for the neat pen 60/70. As a result, the statistical
test results between the average G⁄ of the 100% PG 76 binder
and that of the blended binders are not significant. The d, however,

Fig. 8. Complex shear moduli G*(Pa, 64 °C) of different types of binders extracted
from field (Note: a: 100% PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).

Fig. 6. Phase angles d (64 °C) of different types of binders after RTFO (Note: a: 100%
PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).

Fig. 9. Phase angles d (64 °C) of different types of binders extracted from field
(Note: a: 100% PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).

shows remarkable consistency within each group of binder and

Fig. 7. Threshold temperatures (°C, G*/Sin d = 2.2 kPa) of different types of binders
after RTFO (Note: a: 100% PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).
clear difference between the groups. Consequently, the statistical
test results based on d are all significant at the 0.01 significance
level. The variations of the threshold temperatures, which are
affected by both G⁄ and d, lie between the variations of G⁄ and d.
The difference between the average threshold temperature of the
PG 76 binder and the binder blended with 85% PG 76 and 15% of
pen 60/70 is not significant, but they are significant for the other
two binders.
As mentioned previously, a few highway agencies use ‘‘Super-
paveTM Plus” specifications. One of the parameters added to the

Table 2
Statistics for significance tests based on DSR test results of RTFO-treated specimens.

Binder G* (Pa, 64 °C) d (64 °C) Threshold Temp. (°C)

Avg. St. Dev. Sig. Avg. St. Dev. Sig. Avg. St. Dev. Sig.
100% PG 76 5917 972 n.a. 61.8 0.3 n.a. 76.5 1.9 n.a.
85% PG 76 + 15% Pen 60/70 6435 410 Not sig. at 0.05 sig. level 66.5 0.1 <0.01 75.4 0.6 Not sig. at 0.05 sig. level
70% PG 76 + 30% Pen 60/70 5578 212 Not sig. at 0.05 sig. level 70.7 0.3 <0.01 73.0 0.4 <0.01
100% Pen 60/70 2301 51 <0.01 86.9 0.1 <0.01 64.3 0.1 <0.01
162 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166

Fig. 11. Comparison of phase angles (at 64 °C) of binders treated in RTFO and
extracted from field (five samples for each group).

Fig. 10. Threshold temperatures (°C, G*/Sin d = 2.2 kPa) of different types of binders the solution elute through a column packed with porous medium,
extracted from field (Note: a: 100% PG 76, b: 85% PG 76, c: 70% PG 76, d: 0% PG 76).
and determining molecular weights and concentration of chemi-
cals through the analysis of elution time. Molecules with higher
molecular weight will elute faster than those lighter ones. The
‘‘SuperpaveTM Plus” specification is d, as being used by California molecular weights of polymers in PMA are believed to be much
Department of Transportation (Caltrans) [18]. Although the greater than those of the asphalt components [11]; therefore, they
‘‘SuperpaveTM Plus” specification targets at source asphalt binder can be separated by using GPC and further analyzed for the pres-
(before mixture production and construction), this study shows ence, type, and content of polymers [11,20]. Daly et al. found that
that d can also be effectively used to differentiate PG 76 binder the average molecular weight distribution of PMA can be divided
from those using less or no polymers in the asphalt binder into very high molecular weight (VHMw, 1000-300 K Daltons,
extracted from field mixture samples. KDa), high molecular weight (HMw, 300-45KDa), medium molecu-
Since d can be used to effectively distinguish binders containing lar weight (MMw, 45-19KDa), asphaltenes (19-3KDa and maltenes
different polymer content, a question that may be interested by (3–0.2KDa) [20]. The sum of VHMw, HMw, and MMw (SVHM) is
highway agencies is if the d value of RTFO-aged binder can be used predominantly attributed to polymer additives [20]. Therefore, it
as a quality acceptance criterion for field-extracted binder. This is possible to approximately calculate the amount of polymer
study, however, further found that RTFO does not accurately sim- products added to asphalt. They further observed that 1 wt% up
ulate the short-term aging of PMA in the field. The d values (at
64 °C) of the different types of binders that vary in polymer content
are shown in Fig. 11, which reveals a clear difference of d for 100%
PG 76 binders treated in RTFO and that recovered from the field.
Moreover, as the polymer content decreases, the gap between
the d is narrowed, until it diminishes for the neat pen 60/70 bin-
ders. Essentially, the more polymer is contained in the binder,
the larger difference it causes between the RTFO-aged binders
and those binders aged in plant production, transportation, and
construction. Therefore, it may not be appropriate to use DSR
parameters from RTFO-treated PMA to establish a quality accep-
tance criterion for binders extracted from field. The inadequacy
of RTFO to simulate short-term aging of PMA has also been
reported in other studies (e.g., [19]).

3.3. Trial tests using GPC and FTIR

GPC and FTIR are two commonly used chemical analysis meth-
ods to identify polymer content in PMA (e.g., [11,20,21]). GPC, also
known as Size Exclusion Chromatography (SEC), can be used to
separate particles or chemical compounds of different sizes. The
process involves dissolving asphalt in a standard solvent, making Fig. 12. GPC chromatograms of binders with different polymer content.

Table 3
Statistics for significance tests based on DSR test results of field-extracted asphalt specimens.

Binder G*(Pa, 64 °C) d (64 °C) Threshold Temp. (°C)

Avg. St. Dev. Sig. Avg. St. Dev. Sig. Avg. St. Dev. Sig.
100% PG 76 6563 199 n.a. 65.4 1.0 n.a. 76.6 0.6 n.a.
85% PG 76 + 15% Pen 60/70 6993 791 Not sig. at 0.05 sig. level 68.0 0.3 <0.01 76.1 1.2 Not sig. at 0.05 sig. level
70% PG 76 + 30% Pen 60/70 5705 737 Not sig. at 0.05 sig. level 71.3 0.4 <0.01 73.6 1.2 <0.01
100% Pen 60/70 2114 370 <0.01 86.8 0.3 <0.01 63.7 1.4 <0.01
 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166 163

to 1.98 wt% polymer is needed to achieve PG 70-22 and a mini-
mum of 2 wt% polymer is needed to achieve PG 76-22 [20].
In this study, PG 76 and pen 60/70 binders were also blended
into different proportions and examined for polymer content by
using GPC. The chromatograms of blended binders with different
proportions of PG 76 and an extracted neat Pen 60/70 binder are
shown in Fig. 12. The figure reveals clear peaks at about 12 min
for binders with PG 76, but the neat pen 60/70 binder lacks the
peak. Therefore, the presence of polymer can be clearly checked.
Although it appears possible to use the SVHM of asphalt binders
extracted from cored mixture samples as a supplementary indica-
tor for polymer content, several questions remain to be answered.
Firstly, several studies found that polymers are subjected to degra-
dation during the mixing and construction process [9,14]. The
extent of polymer content change in field construction needs to
be considered if the polymer content is to be used as a quality
assurance parameter. Secondly, there are some large molecules
or microstructures in neat asphalt binders [22], which may bias
the SVHM results. Thirdly, GPC test on binder recovered from field
is a complex process that demands skilled technician and testing
equipment, which pose a challenge for highway agencies to adopt
it as a routine quality assurance test method. Therefore, GPC was
not further pursued in this study.
FTIR spectroscopy has been widely used to characterize the
functional groups in asphalt [e.g., 23, 24]. The absorbance (or trans-
mittance) peak value around the wavenumber 965–966 cm1 is an
indicator of Styrene-Butadiene-Rubber (SBR), Styrene-Butadiene
(SB), or SBS polymers, while the peak value at the wavenumber
1375 cm1 is an indicator of base bitumen [23,24]. The ratios of
the two peak values or peak areas [23,24] indirectly indicate the
polymer content in the binder. However, to quantify the polymer
content, base bitumen and polymer of the same source as those
used in the tested PMA need to be used to develop a calibration
plot. For agencies that import PMA from international markets,
the base bitumen and polymer modifiers may not be easily
obtained from the source refinery plants for conducting such test.

Fig. 13. FTIR spectrum of 100% Pen 60/70 binder.

Fig. 14. FTIR spectrum of 100% PG 76 binder.

164 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166

Therefore, this study intends to find: (1) if it is possible to identify

the relative content of PG 76 binder instead of the exact polymer
content, without knowing the base bitumen and polymer informa-
tion; (2) if it is possible to develop a simple bitumen recovery pro-
cedure that can be used for subsequent FTIR test without following
the lengthy process of binder extraction and rotary evaporation.
The Attenuated Total Refraction (ATR) FTIR spectroscopy was
used to characterize the various binders in this study. The ATR
mode examines the surface of asphalt sample by penetrating the
sample only by a few microns. Initially, only laboratory asphalt
specimens were used in this study. Besides 100% PG 76 binder
and pen 60/70 binder, the following blended binders were pre-
pared: 70% PG 76 + 30% pen 60/70, 80% PG 76 + 20% pen 60/70,
85% PG 76 + 15% pen 60/70, 90% PG 76 + 10% pen 60/70. Duplicate
tests were conducted for each type of binder. The FTIR spectra of
the 100% pen 60/70 specimens and 100% PG76 specimens are
shown in Figs. 13 and 14, respectively. It can be clearly seen that Fig. 15. Change of the absorbance peak ratio with the proportion of PG 76 binder
the pen 60/70 binder lacks a peak at the wavenumber 965 cm1. (duplicate samples).
Therefore, FTIR test can be conveniently used to detect the pres-
ence of polymer.
The FTIR spectra reveal that as the proportion of PG 76 binder
increases, the absorbance peak value around 965 cm1 increases,
too. Because the exact polymer content in the 100% PG 76 is
unknown, the relationship between the absorbance peak value
and polymer content cannot be directly linked. However, if we
assume that the polymer content in the PG 76 binder is k and
the polymer content in the pen 60/70 binder is 0, the proportion
of PG 76 binder in the blended binder is p, and the total mass of
the mixed binder is m, the polymer content of a blended binder
can be estimated as:

Cp ¼ ½k  p  m þ 0  ð1  pÞm=m ¼ k  p
Therefore, the polymer content in the blended binder is propor-
tional to the PG 76 binder content. From quality assurance per-
spective, ensuring that 100% PG 76 binder is used or not be
severely degraded may be more relevant than knowing the exact
polymer content. The ratios of the absorbance peak values between
the wavenumber 965 cm1 and 1375 cm1 for all the tested bin-
ders are shown in Table 4.
Without considering the actual value of k, the change of absor-
bance peak ratios with the proportion of PG 76 binder is shown in
Fig. 15. The high R-square indicates that the model fits the data
very well, and the repeatability of the test at each proportion is
good, too.
Another method to assess the relative polymer content in
asphalt binder is to use the ratios of the integrated areas around
the wavenumber 965 cm1 and 1375 cm1, as shown in Fig. 16.
The use of the integrated area to evaluate the change of chemical
components in asphalt binder is also widely used, especially in
the evaluation of oxidative aging of asphalt binder [25] as well as
in the determination of the additive percentages in polymer mod-
ified asphalt cements [23,24].
ð2Þ
Fig. 16. Peak areas around the wavenumber 965 cm1 and 1375 cm1.
The ratios of the calculated peak areas around the wavenumber
965 cm1 and 1375 cm1 for all the tested binders are shown in
Table 5, and the change of the peak area ratios with the proportion
of PG 76 binder is shown in Fig. 17. It can be seen that the use of
this ratio also fits the data very well. Therefore, FTIR is a promising
method to verify the availability of polymer modifier as well as the
proportion of PG 76 binder used in a blended binder. The use of
both peak height ratio and peak area ratio appear to be equally
acceptable.
To further assess the feasibility of using FTIR to identify the
availability and content of polymer in asphalt binder from the field,
FTIR tests were conducted on extracted binder from the cored
specimens. The same batches of cored samples used for DSR tests
were also used for FTIR tests. In performing the tests, less than
50 grams of bituminous mixtures from the field cores were used.

Table 4
Change of absorbance peak value ratio with PG 76 binder content (Note: k is the Table 5
unknown polymer content in the PG 76 binder). Change of absorbance peak area ratio with PG 76 binder content (Note: k is the
unknown polymer content in the PG 76 binder).
Proportion of Polymer content Ratios of absorbance peak
PG76 binder values Proportion of Polymer content Ratios of absorbance peak areas
PG76 binder
Specimen 1 Specimen 2 Specimen 1 Specimen 2
0 0 0.0377 0.0667 0 0 0 0
0.7 0.7 k 0.2829 0.3077 0.7 0.7 k 0.2733 0.2725
0.8 0.8 k 0.3830 0.3789 0.8 0.8 k 0.3702 0.3896
0.85 0.85 k 0.3882 0.3767 0.85 0.85 k 0.3745 0.3997
0.9 0.9 k 0.3958 0.4043 0.9 0.9 k 0.414 0.4271
1 k 0.4526 0.4505 1 k 0.4581 0.431
 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166 165

Fig. 17. Change of the absorbance peak area ratios with the proportion of PG 76 Fig. 19. Change of the ratios of absorbance peak areas with the proportion of PG 76
binder (duplicate samples). binder in field samples (duplicate samples).

of absorption peaks. Conversely, the absorption area may be less

affected by the global shift of the spectrum and hence provides
more consistent results. The change of absorption peak area ratios
with the proportion of PG 76 is shown in Fig. 19, which shows that
absorption area ratios approximately grow linearly with the PG 76
proportion.

4. Summary, Conclusion and recommendation

Ensuring that specified PMA is fully used and not severely

degraded in asphalt mixture production is a critical link in the
quality assurance process of asphalt pavement construction. Using
SBS modified PG 76, neat pen 60/70, and blended binders created
from these two, this study tested a variety of methods to differen-
tiate PG 76 binder from those containing none or only a portion of
the PG 76 binder. The test methods involved mechanical tests on
asphalt mixtures, rheological tests on asphalt binders, and chemi-
Fig. 18. High speed centrifuge apparatus for precipitating fine aggregate and filler.
cal analysis on asphalt binders. The following conclusions are
drawn from the study.
Asphalt binders from the bituminous mixtures were not recovered
from rotary evaporation. Instead, the recovery only involved add- (1) The indirect tensile (IDT) strength test can be potentially
ing solvent (dichloromethane) to the broken mixture, screening used to differentiate asphalt mixture using PG 76 binder
the aggregates, spinning the solution in a high-speed centrifuge from those pen 60/70 binder using none or only a portion
apparatus (Fig. 18) to precipitate fine aggregate and filler, evapo- (85% and 70%) of PG 76 binder, but the stiffness modulus
rating the small amount of solution, and placing the evaporated cannot reveal the difference.
residual in an oven at 160 °C until it approaches to a constant mass. (2) The phase angle of recovered asphalt binder can be effec-
The IR spectra of bitumen extracted from field reveal that, if the tively used to differentiate field mixture using PG 76 binder
binder contains a portion of PG76 binder, a peak at wavenumber from those using neat or partial PG 76 binders, but the com-
965 cm1 will appear. To identify the portion of PG76 binder used plex shear modulus or threshold temperature based on rut-
in the extracted binder, the ratios of the absorbance peak values ting index (G⁄/sind) cannot effectively serve this purpose.
and absorbance peak areas are summarized and shown in Table 6. (3) The phase angle of PMA subjected to RTFO aging cannot be
Table 6 indicates that the ratios generally increase with the reliably used as a quality acceptance criterion for PMA
increase of PG 76 binder proportion, and the ratios of absorbance extracted from field mixtures, due to the limit of RTFO sim-
peak areas are more consistent. It is suspected that additional ulating PMA aging during mixture production, transporta-
aging and residual solvent may cause inconsistent measurements tion, and construction.

Table 6
Change of absorbance peak value ratios and peak area ratios with PG 76 binder content in field samples (Note: k is the unknown polymer content in the PG 76 binder).

Proportion of PG76 binder Polymer content Ratios of absorbance peak values Ratios of absorbance peak areas
Specimen 1 Specimen 2 Specimen 1 Specimen 2
0 0 0 0 0 0
0.7 0.7 k 0.18 0.21 0.1608 0.1571
0.85 0.85 k 0.16 0.19 0.2489 0.2222
1 k 0.31 0.30 0.3167 0.3067
166 Y. Wang et al. / Construction and Building Materials 150 (2017) 157–166
(4) Chemical analysis methods such as GPC and FTIR can be
potentially used to detect the availability of polymer and
estimate the polymer content or relative polymer content
in recovered asphalt binder. The sample preparation and
testing procedures using FTIR is more convenient than those
using GPC, and FTIR is less affected by large molecules or
microstructures in base asphalt binder.
(5) The ratios of absorbance peak areas of FTIR spectra are more
effective in differentiating compliant PMA binders from
those non-compliant ones than the ratios of absorbance
peak values.
The findings suggest that IDT strength, phase angle of recovered
asphalt, and FTIR may be potentially used for quality compliance
test of PMA binder. IDT strength is a conventional test that may
be used together with other quality acceptance tests (such as den-
sity) after construction; therefore, the additional effort for an
agency may be minimal. However, large variation shown in the
field samples suggests a relatively large sample size is needed. In
addition, IDT strength may not be able to serve as a quality indica-
tor for PMA if a harder neat asphalt binder is blended into the mix-
ture. Phase angle of recovered asphalt binder demands a rigorous
binder extraction and recovery procedure, which is a barrier for
some highway agencies to use this method as a routine test. In
addition, because RTFO does not well simulate the short-term
aging of PMA, extracted asphalt binder samples from the field,
instead of RTFO-treated samples, may need to be used to establish
quality compliance criteria. FTIR is a relatively simple technique to
use, especially by following the sample preparation and analysis
procedure developed in this study. However, due to polymer
degradation in mixture production and placement, more field spec-
imens need to be used to establish a compliance database or crite-
ria. Moreover, FTIR only provides an indication of polymer in the
PMA, not its mechanical properties.
Ideally, two methods may be combined to test the compliance
of PMA. FTIR may serve as a convenient screening test to assess
the adequacy of polymer content, by comparing the FTIR absorp-
tion peak area ratios of field-obtained samples with a threshold
created from historical data. If abnormality is suspected, validation
tests may be further performed on extracted and recovered asphalt
binders. This study found that phase angle from DSR test is suffi-
cient for such purpose, however, one may also attempt the rela-
tively new method—the Multiple Stress Creep Recovery (MSCR)
test (AASHTO TP70) [26], which is not investigated in this study.
Acknowledgement
The research is funded by the Highways Department in Hong
Kong and an internal research grant (PolyU 5277/12E).
References
[1] Transportation Research Board, 2005. Glossary of highway quality assurance
terms. Transportation research circular e-c074, Transp Res Board, Washington
D.C.
[2] M.G. Bouldin J.H. Collins Influence of binder rheology on rut resistance of
polymer modified and unmodified hot mix asphalt 1992 Polymer Modified
Asphalt Binders ASTM International
[3] C.J. Glover, R.R. Davison, C.H. Domke, Y. Ruan, P. Juristyarini, D.B. Knorr, S.H.
Jung, Development of a new method for assessing asphalt binder durability
with field validation, (No. FHWA/TX-05/1872-2), 2005.
[4] X. Lu, U. Isacsson, Chemical and rheological characteristics of styrene-
butadiene-styrene polymer-modified bitumens, Trans. Res. Rec. J. Trans. Res.
Board 1661 (1999) 83–92.
[5] Argus Media Ltd. 2011, Argus Asphalt Report (Nov. 11, 2011): Weekly
International Pricing and Analysis. 11M-45.
[6] J. D’Angelo, Modified binders and superpave plus specifications, Superpave
Technical Issues, Asphalt Institute, 2004.
[7] H. Fu, L. Xie, D. Dou, L. Li, M. Yu, S. Yao, Storage stability and compatibility of
asphalt binder modified by SBS graft copolymer, Constr. Build. Mater. 21 (7)
(2007) 1528–1533.
[8] B. Sengoz, G. Isikyakar, Analysis of styrene-butadiene-styrene polymer
modified bitumen using fluorescent microscopy and conventional test
methods, J. Hazard. Mater. 150 (2) (2008) 424–432.
[9] D.O. Larsen, J.L. Alessandrini, A. Bosch, M.S. Cortizo, Micro-structural and
rheological characteristics of SBS-asphalt blends during their manufacturing,
Constr. Build. Mater. 23 (8) (2009) 2769–2774.
[10] Í.A. de Carcer, R.M. Masegosa, M.T. Viñas, M. Sanchez-Cabezudo, C. Salom, M.G.
Prolongo, V. Contreras, F. Barceló, A. Páez, Storage stability of SBS/sulfur
modified bitumens at high temperature: influence of bitumen composition
and structure, Constr. Build. Mater. 52 (2014) 245–252.
[11] I. Negulescu, L. Mohammad, W. Daly, C. Abadie, R. Cueto, C. Daranga, I. Glover,
Chemical and rheological characterization of wet and dry aging of SBS
copolymer modified asphalt cements: Laboratory and field evaluation (with
discussion), J Assoc Aspha Pav Tech 75 (2006).
[12] J.A. D’Angelo, 2002. Superpave Plus Binder Specifications Why do We Have
them, <http://amap.ctcandassociates.com/wp/wp-content/uploads/
SuperPaveBinder.pdf> (accessed on October 21, 2016).
[13] A. Epps Martin, E.S. Park, E. Arambula, C. Spiegelman, Assessment of the TxDOT
binder quality assurance program, Report: 0-4047-2, <https://tti.tamu.edu/
publications/catalog/record/?id=6208> (accessed on Sep. 1, 2016), 2003.
[14] I. Nösler, T. Tanghe, H. Soenen, 2008, Evaluation of binder recovery methods
and the influence on the propertiesof polymer modified bitumen, in:
Proceedings of the 4th Eurasphalt and Eurobitume Congress held May 2008,
Copenhagen, Denmark.
[15] J.B. Wei, J.C. Shull, Y.J. Lee, M.C. Hawley, Characterization of asphalt binders
based on chemical and physical properties, Int J Polym Analy Char 3 (1) (1996)
33–58.
[16] S. Diefenderfer, 2006, Detection of polymer modifiers in asphalt binder, Report
no.: FHWA/VTRC 06-R18, < http://www.virginiadot.org/vtrc/main/
online_reports/pdf/06-r18.pdf> (accessed on Aug. 1, 2016).
[17] A. Agresti, B. Finlay, Statistical Methods for the Social Sciences, Dellen, San
Francisco, 1997.
[18] C.C. Gerber, R. Schlierkamp, 2012, California vs. Nevada binder specs: What’s
the difference. <http://www.qualityincalifornia.com/2012_12_01_archive.
html>, (accessed October 21.2016).
[19] D.A. Anderson, R.F. Bonaquist, 2012, Investigation of short-term laboratory
aging of neat and modified asphalt binders, NCHRP Report 709, Transp Res
Board, Washington D.C.
[20] W.H. Daly, I. Negulescu, I. Glover, 2010, A comparative analysis of modified
binders: original asphalts and materials extracted from existing pavements,
Federal Highway Administration Report No. FHWA/LA 10 (2010) 462.
[21] J.C. Petersen, 1986, Quantitative functional group analysis of asphalts using
differential infrared spectrometry and selective chemical reactions–theory and
application, Transportation Research Record, (1096).
[22] Y. Wang, K. Zhao, Different forms of asphaltene microstructures discovered in
transmission electron microscopy, J. Mater. Civ. Eng. 28 (11) (2016) 04016137.
[23] AASHTO, 2005, Standard method of test for determination of polymer modifier
in asphalt, Temporary standard. http://www.arc.unr.edu/Deliverables/
Determination_of_Polymer_Modifier_in_Asphalt-Darft-1-Apr2011.pdf
(accessed October 21.2016).
[24] Txdot, 1999, Determining polymer additive percentages in polymer modified
asphalt cements, Designation: Tex-533-c. http://ftp.dot.state.tx.us/pub/txdot-
info/cst/TMS/500-C_series/pdfs/aph533.pdf (accessed October 21.2016).
[25] Y. Wang, Y. Wen, K. Zhao, D. Chong, J. Wei, Connections between the
rheological and chemical properties of long-term aged asphalt binders, J.
Mater. Civ. Eng. 27 (9) (2014) 04014248.
[26] Y. Wen, Y. Wang, K. Zhao, A. Sumalee, The use of natural rubber latex as a
renewable and sustainable modifier of asphalt binder, Int. Journal of Pave. Eng.
(2015) 1–13.