Construction and Building Materials 96 (2015) 47–54

Contents lists available at ScienceDirect

Construction and Building Materials

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

Rheological and low temperature properties of asphalt composites

containing rock asphalts
Ruixia Li a,⇑, Pravat Karki b, Peiwen Hao c, Amit Bhasin b
a
School of Water Conservancy and Environment, Zhengzhou University, No. 100 Science Avenue, Zhengzhou, Henan 45001, China
b
Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, 1 University Station, C1761, Austin, TX 78712, USA
c
Highway School, Chang’an University, South Erhuan Middle Section, Xi’an, Shaanxi 710064, China

h i g h l i g h t s

 Mechanical properties of composites with rock asphalts are evaluated and compared.
 A method that does not need extraction of binder from rock asphalt is pursued.
 Rock asphalt increases stiffness but decreases low temperature performance of composites.
 Rock asphalt decelerates the aging rate of composites.

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

Article history: This paper presents the findings from a study conducted to evaluate the potential impact of different
Received 29 May 2014 types of rock asphalts on performance of asphalt composites. Fine aggregate matrix (FAM) mixtures were
Received in revised form 6 July 2015 prepared by incorporating three different types of rock asphalts – Buton, QC and UM for this study. The
Accepted 15 July 2015
methodology used in this study avoided the extraction of asphalt binder from rock asphalt and simplified
Available online 5 August 2015
the process of evaluating the potential impact of rock asphalts on mixture performance. Rheological
properties were measured using frequency–temperature sweep tests with a dynamic shear rheometer
Keywords:
and creep–relaxation tests with a bending beam rheometer. Tensile strengths of the composites at low
Rock asphalt
FAM mixtures
temperatures were also measured by applying monotonically increasing deflection. The critical cracking
Rheological properties temperatures were computed using a hypothetical cooling rate for the purposes of comparing material
Low temperature properties durability. Results from this study demonstrated that addition of rock asphalts increases material stiff-
Long-term aging ness and slightly reduces relaxation potential of asphalt composites at low temperatures. A comparison
of properties before and after long-term aging also revealed a slightly lower rate of aging for mixtures
modified with rock asphalt as compared to the control mix.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction predominant constituents of rock asphalt. Physically, rock asphalt

Rock asphalt is a naturally occurring material that can be used
as a partial substitute for conventional asphalt binder in the pro-
duction of hot-mix asphalt. In certain parts of the world, rock
asphalt has gained attention because of its benefits in terms of
reduced costs and ease of use during mixture production. Rock
asphalt is naturally produced by impregnation of petroleum or
oil into rocks such as limestone followed by the combined action
of heat, pressure, oxidation, and bacteria over millions of years.
As a result, asphalt binder and mineral fillers are the two
⇑ Corresponding author.
E-mail addresses: liruixia@utexas.edu (R. Li), pravatko@utexas.edu (P. Karki),
haopw@yahoo.com.cn (P. Hao), a-bhasin@mail.utexas.edu (A. Bhasin).
is black in color and can be easily crushed into a powder form
[1]. The mechanical properties of rock asphalts vary considerably
depending on the type and proportion of its constituents and the
type of physical and chemical alteration it undergoes over time [2].
Previous studies have reported improvement in the performance
and cracking resistance of pavement mixtures modified by incorpo-
rating different types of rock asphalts such as Gilsonite from USA,
Buton from Buton Island of the South Pacific Indonesia, QC from
China and UM from Iran. Studies by Widyatmoko et al. [2],
Anderson et al. [3], and Yilmaz et al. [4] investigated the effect of mod-
ifying asphalt mixtures with Gilsonite rock asphalt and found
improved performance at higher temperatures. However, there is
limited information in the literature that describes the characteristics
of other rock asphalts. Similar studies by Siswosoebrotho et al. [5]

http://dx.doi.org/10.1016/j.conbuildmat.2015.07.150
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
48 R. Li et al. / Construction and Building Materials 96 (2015) 47–54
and Affandi [6] about the effect of Buton rock asphalt reported
higher values of stiffness, stability, resistance to deformation, and
temperature susceptibility than the control mixtures. Another study
by Fan et al. [7] reported that QC rock asphalts increased the viscos-
ity and stiffness of asphalt binder at higher temperatures but
decreased the ductility at lower temperatures. Similarly, Lu et al.
[8] measured the properties of QC rock modified binder and reported
that there was no negative influence to the low temperature proper-
ties of the mixture as long as the concentration of the binder from
rock asphalt was within 8% by weight of the original binder content.
Also, Ameri et al. [9] found that UM rock asphalt increased the
rutting resistance of binder, but adversely affected its low tempera-
ture cracking resistance. Other forms of natural rock asphalt also
include natural asphalt binder mixed with fine mineral aggregates
such as the one from Selenice in Albania [10].
The quality and content of the binder present in rock asphalt
varies from one source to another. Therefore each source must
be evaluated individually potential use in asphalt mixture produc-
tion. One of the problems faced by researchers in evaluating bin-
ders from rock asphalts or other similar sources is the
complexity in extracting or separating the asphalt binder from
the mineral aggregate. The methodology used in this study inves-
tigates the effect of rock asphalt on asphalt mixtures without hav-
ing to extract the binder from the rock asphalt. Instead, the rock
asphalt was carefully proportioned and mixed with a fresh sample
of asphalt binder and fine aggregates to produce a fine aggregate
matrix (FAM) mixture. The proportioning was carried out by
replacing the weight and volume contribution of aggregates and
binder from the rock asphalt in the final mixture. Test specimens
from the gyratory compacted FAM samples were then fabricated
for further testing and evaluation. This is not only an efficient
approach to evaluate rock asphalt, but it is also more realistic
because it reflects the manner in which rock asphalt would be
introduced during hot-mix asphalt production. Note that the use
of FAM specimens is also more efficient in terms of the cost, time
and capital equipment required compared to testing a full asphalt
mixture, while at the same time providing meaningful insights into
expected material behavior. This same method can easily be
extended and used to evaluate the influence of other forms of nat-
ural asphalt as well as the pulverized form of reclaimed asphalt
pavement (RAP) mixed with or without rejuvenators on the
expected performance of asphalt mixtures.
In short, the main objective of this paper is to compare the influ-
ence of rock asphalt from three different sources: Buton, QC and
UM (see Fig. 1.) on the overall properties of asphalt composites.
More specifically, this study investigates (1) the influence of rock
asphalt on mechanical properties of FAM mixtures at different
loading conditions and temperatures, (2) the influence of aging
on mechanical properties of FAM mixtures with and without rock
asphalt, and (3) the influence of percentage of rock asphalt on the
Fig. 1. Three different types of rock asphalt.
aforementioned properties. The secondary objective of this paper is
to present a methodology that can be used to achieve the first
objective (and similar scenarios such as in the case of recycled
asphalt pavement materials), without having to separate the
binder from the rock.
2. Methodology
2.1. Materials
Limestone aggregates and fillers from Buda, Texas and a Superpave PG 64-22
asphalt binder obtained from a refinery in Texas were combined with rock asphalt
from three different sources – Buton, QC and UM to produce the FAM mixtures and
test specimens. Previous studies [11,12] conducted by the co-authors of this paper
recommend using 8% binder from the QC or UM rock asphalt and 20% binder from
the Buton rock asphalt by weight of total binder content in the mixture. The afore-
mentioned concentrations refer to the percentage of binder from the rock asphalt
expressed as a percentage of the binder in the original mix design by weight. For
example, 8% by weight of QC rock asphalt implies adding an amount of QC rock
asphalt such that the binder content from this rock asphalt is 8% of the optimum
binder content obtained from the original mix design. In order to compare the influ-
ence of the source of rock asphalt, 8% content was used for rock asphalts from each
of the three sources. An additional mix with 20% concentration of Buton rock
asphalt was also used to evaluate the effect of concentration of rock asphalt on
the performance of FAM mixtures. In summary, five different types of FAM mixtures
were used: a control mixture, three different types of rock asphalt mixtures at 8%
concentration and Buton rock asphalt at an additional 20% concentration.
Table 1 presents the basic properties of the three different rock asphalts used in
this study, while Fig. 2 presents the gradation of aggregates in the FAM mixture.
This gradation reflects the relative proportion of fine aggregates in a typical dense
graded asphalt mixture. The control mix was prepared using 89% of fine aggregates
and 11% of PG 64-22 binder by total weight of the mixture. As mentioned earlier,
FAM mixtures with rock asphalts were prepared by substituting an appropriate
amount of fines and binder with the rock asphalt. The substitution was done such
that the binder content from the rock asphalt was 8% of the optimum binder con-
tent for the QC and UM rock asphalts, and 8% and 20% of the optimum binder con-
tent for the Buton rock asphalt. The percentages of the PG 64-22 binder and
limestone fines were reduced for each mix to compensate for the binder and fines
added in the form of rock asphalt. In other words, the mixes with rock asphalt had
the same final gradation and binder content as the control mix.
2.2. Sample preparation
Rock asphalt, commercially obtained in the form of a powder, was first mixed
with oven dried limestone aggregates, and subsequently with the PG 64-22 asphalt
binder. The loose mixture was then placed in an oven at 135 °C for 2 more hours to
simulate short-term aging (STA) and then compacted into cylindrical samples
(150 mm diameter and 120 mm height) of maximum possible density using the
Superpave gyratory compactor. Similarly, another set of loose mixtures from each
rock asphalt source was placed in the environmental chamber at 60 °C for 30 days
prior to compaction to simulate long-term aging (LTA). The aforementioned condi-
tions produce approximately the same extent of aging as the pressure-aging vessel
used for asphalt binders [13,14]. Table 2 presents volumetric properties of the five
different short-term aged and long-term aged FAM mixtures prepared for this
study. All specimens, irrespective of aging, were compacted until refusal (maximum
density possible when using the Superpave Gyratory Compactor). The bulk specific
gravity and maximum specific gravity were measured using the compacted speci-
mens and samples of loose mixtures set aside form the same batch used for com-
paction, respectively, following procedures similar to those used with asphalt
 R. Li et al. / Construction and Building Materials 96 (2015) 47–54 49

Table 1
Properties of rock asphalts.
Property Type of rock asphalt
Buton QC
Asphalt binder content (%) 28 83
Mineral fillers content (%) 72 17
Specific gravity 1.62 1.10
Softening point (°C) 144 99
Note: These data are from rock asphalt processors, and ‘‘–’’ indicates that the data
was unavailable.
mixtures. The air voids were computed using the bulk and maximum specific grav-
ity. As seen in this table, the air void content in the test specimens varied from 1.5%
to 3% by volume of the mixture. Between replicates of any given type of mix (min-
imum of two), the coefficient of variation in the air voids varies from 0% to 12%. The
Superpave compacted specimens of the short-term and long-term aged samples
were then cut into two parts. The first part was used to extract cylindrical speci-
mens (12.25 mm diameter and 45 mm length) for the frequency–temperature
sweep tests. The second part was used to extract rectangular beam specimens
(125 mm  13 mm  7 mm) for the creep and flexural failure tests. Fig. 3 illustrates
a typical compacted specimen with the cored cylindrical test specimens as well as
the saw-cut rectangular beam specimens.
2.3. Frequency–temperature sweep tests at intermediate to high temperatures
Frequency–temperature sweep tests were conducted to evaluate and compare
the linear viscoelastic properties of FAM mixtures modified with different rock
asphalts. The cylindrical test specimens were subjected to shear strain of 0.001%
amplitude in a dynamic shear rheometer while changing the loading frequency
from 25 Hz to 0.01 Hz at 5 °C, 20 °C and 40 °C in succession and measuring the
resultant shear stress. The applied strain amplitude was low enough to avoid any
damage to the test specimen. For example, earlier studies have shown that a FAM
mixes behave linearly up to a strain amplitude of 0.0065% [15–17]. The dynamic
shear modulus, |G*| and phase angle, d were calculated at each temperature–
frequency combination. Three cylindrical specimens from each mixture (both
short-term and long-term aged for each type) were used for these tests.
The dynamic shear modulus data obtained from these tests were horizontally
shifted and superimposed to construct the master curves at a reference tempera-
ture of 20 °C by applying the time–temperature superposition principle for the cor-
responding materials. The sigmoidal model [18] was used to fit the master curves:
a creep stiffness at any time, t can be calculated directly from the applied constant
lgðjG jÞ ¼ d þ
1 þ ebþcðlg tr Þ
where, t r is the reduced time at the reference temperature, d is the minimum asymp-
totic modulus, a is the span of modulus values, and b, c are the parameters describ-
ing the shape of the sigmoidal function.
temperature cracking. Compared to a material with relatively lower stiffness and/or
higher relaxation rates, a material with relatively higher stiffness and/or reduced
relaxation rate in fact develops higher tensile stresses over time as the pavement
temperature drops. However, the tensile strength of the material ultimately dic-
UM tates whether or not the material fails during such cooling events. Therefore, in
order to assess the relative resistance of different materials to low temperature
82 cracking, it is important not only to measure their stiffness and relaxation rates
18 but also their tensile strengths at low temperatures.
1.16 In this study, the bending beam rheometer (BBR) was used to measure the creep
– stiffness of different rock asphalts at low temperatures. The BBR tests were carried
out at 6 °C, 12 °C and 18 °C by following the test procedures similar to those
used for asphalt binders [19]; however, the specimen preparation procedures used
were different for FAM mixes. Also, unlike the ASTM procedure the deformation on
the FAM specimens was recorded during loading and unloading periods for a dura-
tion of 240 s each. Five beam specimens from each mixture (both short-term and
long-term aged FAM mixture with and without rock asphalt) were used for these
tests.
To determine the failure stress of different FAM asphalt mixtures, similar beam
specimens were subjected to flexural failure at 6 °C and 12 °C by using a ten-
sion–compression dynamic mechanical analyzer. The set up for the flexural failure
test was very similar to the BBR. Each beam (125 mm  13 mm  7 mm) was
placed on two hemispherical supports with a span of 100 mm and was completely
immersed in a bath at a controlled temperature. A monotonically increasing dis-
placement was applied at a rate of 1 mm per minute. The loading assembly was first
calibrated without any specimen to correct for any buoyancy that may arise due to
the immersion of the loading block. Three beam specimens from each short-term
aged FAM mixture were used for these tests. Note that in this case, the flexural
beam geometry was being used to determine the maximum tensile stress until fail-
ure. Due to the limited scope of this paper, low temperature fracture tests were per-
formed only on short-term aged mixtures. Fig. 4 below shows the setup and typical
result for creep-recovery and flexural strength tests.
2.5. Analysis to compare low temperature cracking susceptibility of rock asphalts
The resistance of rock asphalt mixtures against low temperature cracking was
compared using two different approaches. The first approach directly compared
the stiffness and relaxation properties of different mixtures at low temperatures.
This approach, however, does not take the potential differences in the tensile
strength of different rock asphalts into account. Therefore a second approach that
compares the critical cracking temperature for different mixtures at a given rate
of cooling was also used. A brief description of both approaches is presented below.
As mentioned earlier, creep stiffness, S and the material relaxation constant,
m-value at 60 s of loading were determined as the key low-temperature properties
following ASTM-D6648. In addition, the following analysis was also carried out. The
ð1Þ
load and the measured displacement history as follows [20]:
PL3
SðtÞ ¼ 3
ð2Þ
4bh dðtÞ

where, S(t) refers to the creep stiffness at time t, P refers to the applied constant load,
2.4. Creep–recovery and flexural failure tests at low temperatures L represents the distance between supports, b and h are width and depth of the test
beam specimen, and dðtÞ represents the deflection measured at mid-span of the test
One of the concerns with using rock asphalt is its susceptibility to low temper- specimen at time t. Similarly, the creep rate, m(t) is defined as the absolute slope of
ature cracking. Typically, higher stiffness and reduced relaxation rates are consid- creep stiffness versus time data in a double logarithm coordinate system:
ered responsible for an increase in the susceptibility of a material to low
 
d lg½SðtÞ
mðtÞ ¼   ð3Þ
d lg½ðtÞ 
100 In terms of the low temperature properties, a higher value of relaxation (or
m-value) signifies the ability of the asphalt binder or its composite to relax and
90
relieve the thermal tensile stress as the temperature drops. Similarly, a lower value
80 of creep stiffness signifies that the asphalt binder or mixture develops smaller ther-
mal tensile stresses. Consequently, these two factors dictate the ability of the
70 asphalt material to resist low-temperature cracking by allowing only smaller ten-
60 sile stresses to develop with a decrease in temperature; the temperature reduction
Passing (%)

50

40 Table 2
Volumetric properties of the five different short-term aged and long-term aged FAM
30 mixtures.
20 FAM mixture type Short-term aged Long-term aged
10 Gmb Air voids (%) Gmb Air voids (%)
0 Control 2.253 1.57 2.254 1.53
0.075 0.3 0.6 1.18 2.36 QC (8%) 2.256 1.46 2.259 1.33
UM (8%) 2.248 1.79 2.247 1.81
Aggregate Size (mm)
Buton (8%) 2.245 1.91 2.248 1.81
Buton (20%) 2.231 2.52 2.220 2.99
Fig. 2. Gradation of aggregates used in the FAM mixtures.
50 R. Li et al. / Construction and Building Materials 96 (2015) 47–54

(a) (b)
Fig. 3. (a) Cored cylindrical test specimens, and (b) saw-cut rectangular test specimens.

Fig. 4. (a) Setup and typical result for creep–recovery test, and (b) setup and typical result for flexural failure test.

results in stress build up whereas the ability of the binder to relax continually coun-
ters this buildup. In a recent study, Liu et al. [21] presented an analytical method
that combines these two simultaneous processes and expresses the low tempera-
ture performance of a material in terms of a time-dependent parameter, J0 (t)
defined as:
mðtÞ 1
J0 ðtÞ   ð4Þ
SðtÞ t
Phenomenologically, this parameter signifies the changing rate of creep compli-
ance with time [21]. A higher value of J0 (t) represents greater cracking resistance of
a material at low temperatures and vice versa. This parameter can be obtained by
fitting the creep response of the material to an empirical or mechanical analog
model, such as the Burgers model [21] as used in this study.
As introduced earlier, the second approach used in this study evaluates the crit-
ical cracking temperature of FAM in order to compare the influence of different rock
asphalt types and concentration on low-temperature cracking resistance. To predict
the value of critical cracking temperature (Tcr) of FAM mixtures with different rock
asphalts, the thermal stress determined from the creep compliance test was com-
pared with the failure stress obtained from the flexural failure test. The process
of predicting, Tcr was performed by following the steps listed below:
(1) The creep compliance curves obtained from the BBR tests performed at dif-
ferent temperatures were horizontally shifted to form a single smooth mas-
ter curve at the reference temperature, 6 °C in this study. The master
curve was then fitted to a mathematical form similar to the Prony series:
DðtÞ ¼ k1  k2 ek3 t  k4 ek5 t ð5Þ
 R. Li et al. / Construction and Building Materials 96 (2015) 47–54
where, k1 , k2 , k3 , k4 , k5 are the Prony series parameters. In all cases, one
time-independent and two time-dependent terms were found to be adequate to
model the measured response. Note that, previous research studies have shown that
time (rheological) and temperature (thermal) superposition can be successfully
applied on FAM mixtures [15,17,22]. The continuity in master curves constructed
based on the same thermo-rheological principle study attests such applicability on
FAMs.
(2) The creep compliance obtained using Eq. (5) was then used to determine
the relaxation modulus for each material using appropriate inversion after
Laplace transformation.
(3) The ambient and concomitant material temperature is now assumed to
drop linearly at a rate of 5 °C/h from an initial temperature of 10 °C. It is
assumed that there are no thermal stresses or other stress histories at this
starting temperature. Also, the coefficient of thermal expansion of the
material is assumed to be 1.7  104 mm/mm/°C. This number is based
on the values reported by Anderson et al. [23] in the literature. It is possible
that the coefficient of thermal expansion of the actual materials used may
be slightly different from the aforementioned values. Also it may be argued
that this analysis is being conducted for FAM, whereas real pavement struc-
tures are made up of full asphalt mixtures. However, it must be noted that
the purpose of this exercise was not to obtain the exact failure temperature
of a pavement section but rather use the failure temperature so obtained to
compare the relative susceptibility of different materials to low tempera-
ture cracking.
(4) The thermal stress in reduced time n is given by the convolution integral:
Z stiffness of the mixtures. This increase can be attributed, at least
n
@ eðn0 Þ 0
rðnÞ ¼ Eðn  n0 Þ dn ð6Þ
0 @n0
where, n0 is the variable of integration and e(n0 ) is the thermal strain history
in reduced time. The reduced time can be obtained by using the
transformation:
Z asphalt, it is possible that the difference in the nature of the fines
t
nðtÞ ¼ a½Tðt0 Þdt
0
ð7Þ
0
where, TðtÞ is the change in temperature T with time t. For a more detailed
description of the above solution, it is recommended to refer the works of
Muki et al. [24] and Morland et al. [25]. Eqs. (6) and (7) were used to obtain
the accumulated thermal stress as the temperature dropped from 10 °C to
30 °C.
(5) The final step was to compare the calculated thermal stress to the failure
stress from the fracture test. Two fracture test temperatures were used
and linearly interpolated across the thermal stress history (the line con-
necting two points in Fig. 5). The point at which this failure stress curve
intersects the thermal stress curve was defined as the critical cracking
temperature of the FAM mixture. Fig. 5 illustrates the typical calculated
thermal stress history of the material superimposed with the tensile
strength of the material.
The authors recognize that a limitation of this study is that the tensile strength
of the FAM mixture as well as the aforementioned analysis does not take into
account the rate dependency in the tensile strength of the FAM mixture. While it
may be somewhat reasonable to assume that the tensile strength is not significantly
influenced by rate of loading at very low temperatures, the authors of this paper
plan to investigate this more rigorously in future studies.
Fig. 5. Determination of the critical cracking temperature.
51
3. Results and discussion
3.1. Frequency–temperature sweep tests
Fig. 6 illustrates master curves for the dynamic shear modulus
of short-term aged and long-term aged FAM mixtures. As expected
and seen in Fig. 6, all of the rock asphalt mixtures (STA and LTA)
show higher values of dynamic shear modulus as compared to
the control mixture, especially at lower frequencies. A notable
aspect of these results was that substituting even a small percent-
age of asphalt binder with binders from rock asphalts had a very
significant influence on the properties of the FAM. For example,
even at a reduced frequency of 10 Hz at 20 °C, mixtures modified
Buton, UM and QC rock asphalt at 8% dosage and aged for short
time (2 h in this study) had shear moduli that were 1.5, 2.1 and
2.7 times that of the control mixture, respectively. This effect
was slightly reduced after long-term aging; the mixtures modified
Buton, UM and QC rock asphalts at 8% dosage and aged for long
term (30 days in this study) exhibited shear moduli that were
1.4, 1.5 and 2.0 times that of the control mixture, respectively.
These results indicate that rock asphalts can significantly increase
partially, to the high stiffness of the rock asphalt that may be
due to high asphaltene content and high molecular weight [2].
Since the incorporation of rock asphalt also entails the substitution
of a small fraction of fines with the fines naturally contained rock
contained in rock asphalt may also contribute to the increase in
stiffness. Finally, an increase in stiffness of the mixture may be
regarded as a potential indicator of reduced resistance against
low temperature cracking. This aspect is evaluated in the later sec-
tions of this paper.
Test results also indicate that the source of the rock asphalt has
a significant influence on the overall properties of the FAM mix-
tures. For the 8% concentration of rock asphalt binder concentra-
tion, the master curves for QC and UM rock asphalt mixtures
(STA and LTA) at higher frequency were similar to each other,
and slightly higher than the ones for Buton rock asphalt mixtures.
This indicates that there is no significant difference for the magni-
tude of shift between the three rock asphalt mixtures at high fre-
quencies. However, at intermediate and lower frequencies, QC
rock mixture exhibits the highest modulus, followed by UM and
Buton rock mixtures. Moreover, an increase in the Buton rock
asphalt concentration from 8% to 20% resulted in an increase in
the modulus values for the mixtures (STA and LTA).
Since rock asphalts are naturally aged, it is expected that the
rate of aging for mixtures with rock asphalt would be slightly
lower than a conventional mixture. Table 3 presents the ratio of
the dynamic shear modulus of the long-term and short-term aged
FAM mixtures at several different reduced frequencies. As seen in
Table 3, with minor exceptions, the long-term aged rock asphalt
mixtures showed a smaller change in stiffness than the similarly
aged control mixture. This suggests that the difference in stiffness
of mixtures with blends of virgin asphalt and rock asphalts com-
pared to the stiffness of mixtures with virgin asphalt tends to
reduce over time with oxidation.
3.2. Creep tests and flexural failure tests at low temperatures
As described earlier, creep compliance tests were conducted at
three low temperatures: 6 °C, 12 °C and 18 °C. The results
from these tests were found to be qualitatively similar. Therefore
only results from the 12 °C temperature are presented in this
paper for brevity. Fig. 7 illustrates creep stiffness, S and m-value
for all FAM mixtures at 60 s and 12 °C for the mixtures subjected
52 R. Li et al. / Construction and Building Materials 96 (2015) 47–54

Dynamic Shear modulus |G*| (Pa) 1E+10 1E+11

Dynamic Shear modulus |G*| (Pa)

1E+9 1E+10

1E+8 1E+9

1E+7 1E+8
Control Control
Buton(8%) Buton(8%)
1E+6 UM(8%) 1E+7 UM(8%)
Buton(20%) Buton(20%)
QC(8%) QC(8%)
1E+5 1E+6
1E-5 1E-4 1E-3 1E-2 1E-1 1E+0 1E+1 1E+2 1E+3 1E+4 1E-5 1E-4 1E-3 1E-2 1E-1 1E+0 1E+1 1E+2 1E+3 1E+4
Reduced Frequency (Hz) Reduced Frequency (Hz)
(a) (b)
Fig. 6. Master curves for dynamic shear modulus at a reference temperature of 20 °C for (a) short-term aged and (b) long-term aged FAM mixtures.

Table 3 asphalt mixture that showed a slightly positive effect on the

Dynamic shear modulus ratio of long-term to short-term aged FAM mixtures at 20 °C. low-temperature performance in terms of m-value. A notable
Reduced frequency Fine aggregate matrix mixture type observation from Fig. 7 is that the increase in stiffness of the
(Hz)
Control Buton UM QC Buton FAM mixtures with 8% rock asphalt binder dosage after
(8%) (8%) (8%) (20%) long-term aging was only about 5–10% compared to the control
1.00E04 6.19 5.20 3.83 4.24 4.76
mixture. Similarly, the reduction in m-value was only about 10–
1.00E03 4.34 4.02 3.21 3.51 4.24 15%.
1.00E02 3.65 3.32 2.68 2.77 3.45 Under certain conditions, it is difficult to evaluate the impact of
1.00E01 2.67 2.87 2.25 2.16 2.65 material stiffness and m-value on the magnitude and relaxation of
1.00E+00 2.34 2.51 1.89 1.76 2.03
tensile stresses developed as the temperature drops. For example,
1.00E+01 2.14 2.16 1.60 1.55 1.62
1.00E+02 2.02 1.82 1.37 1.48 1.38 the short-term aged UM rock asphalt mixture has both higher stiff-
1.00E+03 1.92 1.51 1.19 1.49 1.26 ness and m-value compared to the control mixture. In such cases, it
1.00E+04 1.84 1.25 1.06 1.56 1.20 is more useful to compare the low-temperature properties using
the J0 (t) parameter as described earlier. The J0 (t) parameters for dif-
ferent FAM mixtures at 60 s were compared. Fig. 8 illustrates the
to two different aging procedures. The results are normalized with value of J0 (t) normalized with respect to the value corresponding
respect to the corresponding values for the control FAM mixture to the control mixture. As seen in Fig. 8, the short-term aged UM
for comparison purposes. It can be inferred from Fig. 7 that the rock asphalt FAM mixture had a lower value of J0 (t) compared to
creep stiffness, S of the rock asphalt mixtures was slightly higher the control mixture, indicating that the latter had better
than the stiffness of the control mixture, while m-value of the rock low-temperature properties compared to the UM rock asphalt.
asphalt mixtures was slightly lower than the m-value of the control The J0 (t) values for the other rock asphalt mixtures were also lower
mixture. The only exception was the short-term aged UM rock compared to the control mixture. This implies that the rock asphalt

120
160 STA
STA
LTA LTA
140 100
Normalized m-value (%)
Normalized stiffness (%)

120
80
100

60
80

60 40

40
20
20

0 0
Control Buton(8%) UM(8%) QC(8%) Buton(20%) Control Buton(8%) UM(8%) QC(8%) Buton(20%)

Matrix mixture type Matrix mixture type

(a) (b)
Fig. 7. (a) Creep stiffness, and (b) m-value of FAM mixtures at 12 °C normalized with respect to the control mixture.
 R. Li et al. / Construction and Building Materials 96 (2015) 47–54 53

120 that the effect of increased temperature can be offset by evaluating

STA the properties at a shorter loading time of 60 s. This is possible due
LTA to the thermo-rheologically simple behavior of asphalt binders. In
100
the context of this study, it is reasonable to compare FAM mixtures
under similar conditions (i.e., at a loading time of 60 s at 10 °C
Normalized J'(t) (%)

80
60
40
20
0 ferent FAM mixtures. The critical cracking temperature was
Control Buton(8%) UM(8%) QC(8%)
Matrix mixture type
Fig. 8. J0 (t) of FAM mixtures at 12 °C normalized with respect to the control
mixture for each aging condition respectively (note that values of control mix for
each mix are at 100% by definition of the normalizing method).
could have an adverse impact on the low-temperature properties,
which is consistent with results based on the creep stiffness and
m-value. Furthermore, it can also be noted that the three rock
asphalts at the same content, 8% exhibited similar values for the
J0 (t) parameter, suggesting that there is no significant difference
in the low-temperature properties between these rock asphalts.
Finally, for Buton rock asphalt, the J0 (t) value decreased when con-
centration of the rock asphalt was increased from 8% to 20%, indi-
cating that the concentration of the rock asphalt within this range
has some effect on the low-temperature properties.
Fig. 9 shows the J0 (t) ratio for long-term aged mixtures to those
of the short-term aged mixture. As can be seen from Fig. 9, the J0 (t)
ratio increased slightly with the addition of rock asphalts. This
indicates that rock asphalt modified mixtures are less sensitive to
long-term aging. This observation is in line with the observations
from frequency–temperature sweep test results.
The creep stiffness, m-value and J0 (t) parameters were com-
pared at 60 s in this study. The comparison is based on recommen-
dation for testing and evaluating asphalt binders using the BBR
(ASTM-D6648). These recommendations are based on the premise
that the low temperature properties can be measured at 10 °C
higher than the expected minimum pavement temperature and
higher than the expected minimum pavement temperature)
because the time–temperature shift factors for FAM mixtures and
full mixtures are expected to be very similar to that of the binder
as demonstrated by Di Benedetto et al. [26]. However, since the
limits for stiffness and m-value are completely different for binders
and their FAM mixtures, only the normalized values with respect
to the control mixture were used for comparison purposes in this
study (Figs. 7 and 8).
Fig. 10 illustrates the critical cracking temperatures, Tcr for dif-
Buton(20%)
obtained using the procedures as described earlier in this paper.
It can be inferred from Fig. 10 that the increase in Tcr of the FAM
mixtures with 8% rock asphalt was only about 2–5 °C compared
to the control mixture, indicating that rock asphalts have a slight
negative influence on the low temperature cracking resistance,
which is consistent with the creep test results. Moreover, for the
Buton rock asphalt, Tcr increased from 14.5 °C to 8.5 °C when
the concentration of rock asphalt was increased from 8% to 20%,
which means the concentration of rock asphalt adversely affected
the low temperature performance.
4. Conclusions
In this study, tests were conducted on FAM mixtures in order to
evaluate the potential impact of different rock asphalts (Buton, QC
and UM) on the properties and performance of asphalt mixtures.
Viscoelastic properties of FAM mixtures with and without the rock
asphalts were measured using the dynamic shear rheometer and
the bending beam rheometer. Tensile strength of the rock asphalt
at lower temperatures was measured by applying monotonically
increasing deflection. Similarly, the critical cracking temperatures
of different FAM mixtures at a given cooling rate were computed
and compared.
 A comparison of master curves indicates that the addition of
rock asphalts effectively increases the dynamic shear modulus
of the composite. This increase is desirable for improved rutting
resistance of asphalt mixtures.

100

-20
80
Critical cracking temerature (ºC)

-16
Ratio of J'(t) (%)

60

-12

40
-8

20
-4

0
Control Buton(8%) UM(8%) QC(8%) Buton(20%)
0
Control Buton(8%) UM(8%) QC(8%) Buton(20%)
Matrix mixture type
Matrix mixture type
Fig. 9. Ratio of J0 (t) of the long-term aged to J0 (t) of the short-term aged FAM
mixtures at 12 °C. Fig. 10. Critical cracking temperatures for FAM mixtures.
54 R. Li et al. / Construction and Building Materials 96 (2015) 47–54
 For the same concentration of rock asphalt, QC exhibits the
highest value of dynamic shear modulus, followed by UM and
then Buton rock asphalt. When the contribution of binder con-
tained in Buton rock asphalt was changed from 8% to 20%, the
dynamic shear modulus also increased.
 The rheological properties of FAM mixtures with rock asphalt
were less affected by long-term aging as compared to the con-
trol mixture. This almost passive response can be due to the
already aged nature of the rock asphalt.
 Low temperature properties of the rock asphalts were measured
using the bending beam rheometer and tension–compression
dynamic mechanical analyzer. Results show that the three rock
asphalts had similar low temperature properties. Rock asphalts
had a slight negative influence on the low temperature perfor-
mance, suggesting that rock asphalts should be carefully con-
sidered in locations that are prone to low-temperature
cracking. Moreover, the normalized ratio of J0 (t) for the
short-term and long-term aged mixtures indicated that the
mixtures with rock asphalt aged at a relatively lower rate.
This also suggests that with time and oxidative aging, stiffness
and relaxation of mixtures with rock asphalts will tend to be
similar to that of a comparable control mixture.
Overall, based on the test results conducted so far, rock asphalts
can be used to improve the stiffness and particularly high temper-
ature rutting resistance of asphalt mixtures. Increased stiffness due
to the addition of rock asphalts may adversely affect the low tem-
perature cracking resistance of an asphalt composite. Future work
by the authors will focus on the evaluation of intermediate tem-
perature cracking (fatigue cracking) resistance of mixtures with
rock asphalts as well as up-scaling the results from this study to
full asphalt mixtures.
Acknowledgements
The authors would like to acknowledge the China Scholarship
Council for providing the financial support and the University of
Texas at Austin for providing the technical resources required at
different stages for this study.
References
[1] Ziegler Chemical & Mineral Corporation. What is Gilsonite? http://www.
zieglerchemical.com/gilsonit.htm. Accessed 01.01.13.
[2] I. Widyatmoko, R. Elliot, Characteristics of elastomeric and plastomeric binders
in contact with natural asphalts, Constr. Build. Mater. 22 (2008) 239–249.
[3] D.A. Anderson, D. Maurer, T. Ramirez, D.W. Christensen, M.O. Marasteanu, Y.
Mehta, Field performance of modified asphalt binders evaluated with
superpave test: Methods I-80 Test Project, Transport Res. Rec. J. Transport
Res. Board 1661 (1999) 60–68.
[4] M. Yilmaz, M.E. Çeloglu, Effects of SBS and different natural asphalts on the
properties of bituminous binders and mixtures, Constr. Build. Mater. 44 (2013)
533–540.
[5] B.I. Siswosoebrotho, N. Kusnianti, W. Tumewu, Laboratory evaluation of lawele
Buton natural asphalt in asphalt concrete mixture, Proc. Eastern Asia Soc.
Transport. Stud. 5 (2005) 857–867.
[6] F. Affandi. The performance of bituminous mixes using Indonesia natural
asphalt, in: The 25th ARRB Conference – Shaping the future: Linking policy,
research and outcomes, Perth, Australia; 2012.
[7] Liang Fan, Yanyan Zhang, Quanjun Shen, Effect of natural rock asphalt on the
rheological properties of asphalt binder, J. China Foreign Highway 27 (2007)
160–163 (in Chinese).
[8] Zaofeng Lu, Zhaoyi He, Qin Yu, Gang Huang, Influence of rock asphalt modified
bitumen on asphalt rheological property, Journal of ChongQing JiaoTong
University (Natural Science) 28 (2009) 543–547 (in Chinese).
[9] M. Ameri, A. Mansourian, S.S. Ashani, G. Yadollahi, Technical study on the
Iranian Gilsonite as an additive for modification of asphalt binders used in
pavement construction, Constr. Build. Mater. 25 (2011) 1379–1387.
[10] E. Santagata, P.P. Riviera, D. Dalmazzo, Performance-related characterization
of bituminous binder and mixtures containing natural asphalt, Procedia – Soc.
Behav. Sci. 53 (2012) 535–545.
[11] Ruixia Li, Research on the technical characteristic of the Buton rock asphalt
and asphalt mixture (master thesis), Chang’an University, Xi’an, China, June
2010.
[12] Kai Yang. Research on the technical characteristic of asphalt mixture with the
Buton rock asphalt (master thesis), Chang’an University, Xi’an, China, June
2012.
[13] C.J. Glover, A.E. Martin, A. Chowdhury, R. Han, N. Prapaitrakul, X. Jin, J.
Lawrence. Evaluation of Binder Aging and its Influence in Aging of Hot Mix
Asphalt Concrete: Literature Review and Experimental Design. Publication
FHWA/TX-08/0-6009-1. Texas Transportation Institute, 2009.
[14] Z. Arega, A. Bhasin. Binder Rheology and Performance in Warm Mix Asphalt.
Publication FHWA/TX-12/0-6591-2. Texas Department of Transportation,
2012.
[15] Y. Kim, D.N. Little, R.L. Lytton, Fatigue and healing characterization of asphalt
mixtures, ASCE J. Mater. Civil Eng. 15 (2003) 75–83.
[16] F.T.S. Aragão, Y. Kim, P. Karki, D.N. Little, Semi-empirical, analytical, and
computational predictions of dynamic modulus of asphalt concrete mixtures,
TRB Transport. Res. Rec. 2181 (2010) 19–27.
[17] P. Karki, Y. Kim, D.N. Little. Dynamic modulus prediction of asphalt concrete
mixtures through computational micromechanics. TRB Transport. Res. Rec.
(accepted 2015).
[18] T.K. Pellinen, M.W. Witczak, R.F. Bonaquist, Asphalt mix master curve
construction using sigmoidal fitting function with non-linear least squares
optimization, in: Recent Advances in Materials Characterization and Modeling
of Pavement Systems, American Society of Civil Engineers, 2004, pp. 83–101.
[19] ASTM-D6648. Standard test method for determining the flexural creep
stiffness of asphalt binder using the bending beam rheometer (BBR), 2008.
[20] R.M. Christensen, Theory of Viscoelasticity: an Introduction, Academic Press
Inc, New York, 1982.
[21] Shutang Liu, Cao Weidong, Shujie Shang, Hui Qi, Fang Jianguo, Analysis and
application of relationships between low-temperature rheological
performance parameters of asphalt binders, Constr. Build. Mater. 24 (2010)
471–478.
[22] B.S. Underwood, Y.R. Kim, Effect of volumetric factors on the mechanical
behavior of asphalt fine aggregate matrix and the relationship to asphalt
mixture properties, Constr. Build. Mater. 49 (2013) 672–681.
[23] D.A. Anderson, M.O. Marasteanu, Y. Liu, Dilatometeric measurement of glass
transition temperatures, in: Eurobitume Workshop on Performance-Related
Properties for Bituminous Binders, Luxembourg, 1999, Paper No. 128.
[24] R. Muki, E. Sternberg, On transient thermal stress in viscoelastic material with
temperature-dependent properties, J. Appl. Mech. (1961) 193–207.
[25] L.W. Morland, E.H. Lee, Stress analysis for linear viscoelastic material with
temperature variation, Trans. Soc. Rheol. (1960) 233–263.
[26] H. Di Benedetto, C. Sauzeat, K. Bilodeau, M. Buannic, S. Mangiafico, Q.T.
Nguyen, S. Pouget, N. Tapsoba, J.V. Rompu, General overview of the time–
temperature superposition principle validity for materials containing
bituminous binder, Inter. J. Roads Airports (IJRA) 1 (2011) 35–52.