Experimental Investigation into the Multiscale

Performance of Asphalt Mixtures with High

Contents of Reclaimed Asphalt Pavement
Jian Li, Ph.D. 1; Fujian Ni, Ph.D. 2; and Qingqing Lu 3
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Hot-mix asphalt (HMA) mixtures with high content of reclaimed asphalt pavement (RAP) have been investigated and applied in
the laboratory and in plants. An asphalt mixture can be considered as a four-scale assemblage of components with different characteristic
length scale, binder, mastic, mortar, and finally mixture. This paper investigates the mechanical properties of HMA mixtures with high RAP
contents through the multiscale of binder, mortar, and mixture, and analyzes the property correlations between them. RAP contents of 30, 50,
and 85% were adopted for tests; meanwhile fresh asphalt (0%RAP) and 100%RAP were also prepared for comprehensive comparison.
Frequency sweep test, repeated creep test, and bending beam test were used to measure the properties of mortars, meanwhile, dynamic
modulus test, flow number test, and semicircular bending (SCB) test were used to measure the properties of mixtures. It is found that with
the increase of the RAP content, binder, mortar, and mixture show similar changes of mechanical properties, behaving as complex modulus
increases with viscoelasticity reduce, and rutting resistance improves, whereas thermal cracking resistance declines. The mechanical
properties of binder, mortar, and mixture at high and low temperature are well related. Fine aggregates have significant influence on
the high-temperature properties of mortars, and the mixture modulus is highly affected by coarse aggregates. DOI: 10.1061/(ASCE)
MT.1943-5533.0002269. © 2018 American Society of Civil Engineers.
Author keywords: Asphalt mortar; Asphalt mixture; Multiscale; High reclaimed asphalt pavement (RAP); Mechanical property; Correlation.

Introduction acceptable because RAP in situ has homogenous components that

guarantee quality stability of HIR pavement (Ma 2010).
In recent years, the shortage of petroleum and aggregate supply and Plenty of studies have been carried out on the characteristics of
the trend of sustainable development are calling for high-quality, asphalt mixture with high RAP content, and most of them focused
cost-effective alternatives to virgin paving mixtures. Under this cir- on the asphalt binder and mixture. However, asphalt mortar which
cumstance, pavement recycling started to attract attention, and re- is composed of binder, fine aggregates, and mineral filler, as an
claimed asphalt pavement (RAP) are utilized in larger and larger important part of the mixture, should also be considered. Nicholls
quantities each year in new construction and maintenance of pave- (1976) suggested that asphalt concrete microstructure could be rep-
ment. Although recycling of RAP benefits both the economy and resented by a two-phase model, which consisted of a large spherical
environment, a major concern is that because aged asphalt binder is aggregate particle surrounded by a spherical shell of fine aggregate-
applied in RAP, mixtures with RAP may not perform as well as filler-binder mixture as the matrix. And the fine aggregate-filler-
mixtures with virgin binder (Copeland 2011). Therefore, the con- binder mixture was further represented by a two-phase model,
tents of RAP are strictly limited in plant-produced hot mix asphalt which treated fine aggregate as a spherical inclusion and the
(HMA) mixtures (usually less than 25%), especially for using in mixture of filler and binder as the matrix. Nicholls’s model for
upper layers (Loria et al. 2011; Diefenderfer and Harikrishnan asphalt concrete was further supported by experimental results.
2014). However, many attempts have been made to increase the Buttlar et al. (1999) used a simplified Christensen and Lo’s spheri-
content of RAP and several studies show that mixtures with cal model to predict the stiffness of the filler-binder mixture, with
30% or more RAP can perform as well as conventional HMA mix- satisfactory results. Zhu and Nodes (2000) proposed a similar con-
tures when properly designed (Nash et al. 2012; Hong et al. 2010; cept for the microstructure of asphalt concrete—asphalt concrete
Bennert et al. 2014). Meanwhile, RAP contents in hot in-place was a coarse aggregate assembly coated by an asphalt binder
recycling (HIR) pavement are usually more than 80%, which is mortar, which itself consists of fine aggregates and asphalt binder.
Furthermore, a multiscale model comprised of five scales of
observation—bitumen, mastic, mortar, mixture, and pavement, is
1
Engineer, Shanghai Road and Bridge (Group) Co., Ltd., Shanghai generally accepted for asphalt (Lackner et al. 2005).
200433, People’s Republic of China (corresponding author). E-mail: The characteristic scale of mortar is immediately smaller than
[email protected] the mixture and immediately larger than mastic, and it could help
2
Professor, School of Transportation, Southeast Univ., Nanjing 210096, better recognize the characteristics of asphalt concrete. Plenty of
People’s Republic of China. E-mail: [email protected] research has been performed on experimental investigations of
3
Senior Engineer, Shanghai Road and Bridge (Group) Co., Ltd., Shanghai
mortar and multiscale models of asphalt concrete. Kim (2003)
200433, People’s Republic of China. E-mail: [email protected]
Note. This manuscript was submitted on June 27, 2017; approved on studied fatigue characteristics of asphalt concrete on sand asphalt
November 8, 2017; published online on April 6, 2018. Discussion period (a type of mortar) under the assumption that crack initiation starts at
open until September 6, 2018; separate discussions must be submitted for a material level that exists between coarse aggregate particles. In his
individual papers. This paper is part of the Journal of Materials in Civil study, mortar samples were fabricated with a bitumen content of
Engineering, © ASCE, ISSN 0899-1561. 8% by total mass which was chosen based on the assumptions

© ASCE 04018105-1 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 of an average value of 10 μm asphalt film thickness within asphalt The aim of this paper is to investigate the viscoelastic behavior
concrete. Zollinger (2005) followed the work of Kim (2003) and of asphalt mixtures with high RAP contents and to identify the sen-
evaluated the moisture susceptibility of mortar which was created sitivity of material properties with changes in RAP content. The
from actual mixture gradations by taking only materials finer than investigation was implemented through three asphalt scales of
1.18 mm. Li and Metcalf (2005) proposed a two-step approach binder, mortar, and mixture, and the building of property correla-
from the existing micromechanical models to predict the modulus tions between them were attempted. The RAP contents of asphalt
of asphalt concrete utilizing the volumetric fractions, Poisson’s ra- mixtures adopted in this study were chosen based on the practices
tios, moduli of the aggregate, and filler-binder mixture. Aigner et al. of hot in-plant recycling and hot in-place recycling pavement in
(2009) proposed a multiscale model to determine the complex vis- Jiangsu province.
coelastic properties of mastic, mortar, and asphalt concrete from
the rheological properties of bitumen, and validated the model
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

by means of cyclic tests and static uniaxial creep tests. Underwood Experimental Plan
and Kim (2011, 2013a) have done significant work to understand
the fundamental properties of different asphalt material scales.
Materials
They performed the temperature and frequency sweep test on
different material scales (from binder to mixture) to identify the In this study, two groups of asphalt material, mortars, and coarse
sensitivity of the dynamic shear modulus (G
) and phase angle graded mixtures were prepared for property tests. RAP contents
(δ) to changes in the volumetric composition (air void and binder of 30, 50, and 85% were selected for production of mortars and
content). Then, they made quantitative comparisons between mixtures. Meanwhile, fresh HMA mixture (0%RAP) and RAP mix-
mortar and mixture properties with compositional changes, and ture (100%RAP) were prepared for comprehensive comparison. The
proposed a micromechanical model and an empirical correlation, RAP utilized in the study were collected in the hot in-place recycling
respectively, to study the linkages between the moduli and fatigue process from the surface layer of expressway asphalt pavement in
properties of mortar and mixture. Cucalon et al. (2016) utilized a Jiangsu province. Styrene-butadiene-styrene (SBS) polymer modi-
four-level multiscale computational micromechanics methodology fied bitumen was used as the virgin binder for fresh mixtures,
to predict the effect of mineral filler, fine aggregate, and large and the virgin basalt coarse aggregates and limestone fine aggregates
aggregate on the dynamic modulus of mastic, mortar, and asphalt were both from the Zhenjiang, Jiangsu province.
concrete, and pointed out that chemical interactions should be The HMA mixtures with 13.2 mm NMAS dense aggregate
considered in the multiscale approach. gradation were designed based on the Standard Specification for
Even though mortar is defined as the materials exists between Construction of Jiangsu Province Expressway Asphalt Pavements
the coarse aggregates, it is hard to determine the proper size ranges [DB32/T1246-2008 (Jiangsu Bureau of Quality and Technical
for mortar that could truly represent the material. In the study of Supervision 2008)]. For the mortars, 4.75 mm was selected as
Zhu and Nodes (2000), the cutoff size of coarse and fine aggregates the maximum fine aggregate size based on the following consid-
was defined as 1.18 mm. Kim (2003) and Zollinger (2005) also erations: First, the 4.75 mm size sieve was utilized as the key sieve
used 1.18 mm as the maximum size of fine aggregates for mortar in the design of the 13.2 mm dense graded asphalt mixture accord-
studies. Other researchers used the aggregate sizes smaller than ing to the specification of Jiangsu; Second, a larger aggregate size
2.36 mm (Dai and You 2007; Underwood and Kim 2011), could help fabricate the mortar specimens in small volume to re-
2.26 mm (Valenta et al. 2010), or 2 mm (Lackner et al. 2005; duce the negative effect on test results. Underwood and Kim
Aigner et al. 2009; Cucalon et al. 2016), and these choices of size (2013b) suggested that the total asphalt content and filler content
have been determined by computational modeling and imaging of the mortar used in experiments should not include the mastic that
limitations. Underwood and Kim (2013b) indicate that mortar ag- coats particles larger than mortar-sized particles. Therefore, mortars
gregate particles have a maximum size that is dependent upon the in this study were obtained by passing the 4.75 mm sieve with
nominal maximum aggregate size (NMAS) for the mixture and the HMA mixtures blended in the laboratory. The actual aggregate
could be determined by using the fine aggregate initial break gradations and binder contents of mortars can be measured using
(FAIB) sieve as taken from the Bailey method of gradation evalu- a bitumen extractor [JTG E20-2011 (Ministry of Transport of the
ation. It is suggested that for 9.5 and 12.5 mm mixtures, the aggre- People’s Republic of China 2011c), T0725 (Ministry of Transport
gate sizes of mortar are smaller than 2.36 mm, and for 19.0 and of the People’s Republic of China 2011b) and T0722 (Ministry of
25.0 mm mixtures, the aggregates sizes are smaller than 4.75 mm. Transport of the People’s Republic of China 2011a)]. As a result,

Table 1. Gradations and Binder Contents of Mixtures and Mortars

Percent passing (%) of each size sieve (mm)
RAP Binder
Material content (%) content (%) 13.2 9.5 4.75 2.36 1.18 0.6 0.3 0.15 0.075
Asphalt mixture 0 4.67 92.6 67.3 42.6 29.5 21.8 15.6 12.6 9.5 6.7
30 4.67 92.8 68.8 41.7 29.1 21.8 15.7 11.9 9.0 6.8
50 4.68 93.0 69.8 41.2 29.0 21.7 15.7 11.4 8.7 6.5
85 4.72 93.3 71.6 40.1 28.4 21.6 15.8 10.6 8.1 6.0
100 4.76 93.4 72.3 39.7 28.2 21.6 15.8 10.3 7.9 5.8
Upper limit 100.0 73.0 48.0 34.0 26.0 18.0 16.0 14.0 8.0
Lower limit 80.0 62.0 37.0 28.0 19.0 12.0 10.0 7.0 4.0
Asphalt mortar 0 7.10 100.0 100.0 100.0 56.7 35.8 28.0 18.6 15.4 10.1
30 7.09 100.0 100.0 100.0 56.2 36.1 27.8 18.3 14.9 10.2
50 7.10 100.0 100.0 100.0 56.8 36.3 28.2 17.8 15.0 10.2
85 7.13 100.0 100.0 100.0 57.0 35.9 27.9 18.0 15.2 10.6
100 7.14 100.0 100.0 100.0 57.3 35.5 27.5 18.4 15.4 10.9

© ASCE 04018105-2 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 the gradations and binder contents of mortars and mixtures pre- α
lg jE
j ¼ δ þ ð1Þ
pared for study are listed in Table 1. 1 þ eβþγ
lgðfr Þ
where jE
j = dynamic modulus; f r = reduced frequency, Hz; δ =
Test Methods for Mixture minimum value of jE
j; δ þ α = maximum value of jE
j; and β,
γ = parameters describing the shape of the sigmoidal function.
It is generally accepted that the dynamic modulus test, the flow The following equation defines the reduced frequency fr as the
number test, and the semicircular bending (SCB) test are the proper actual loading frequency multiplied by the time-temperature shift
test methods to evaluate the mechanical properties of the HMA factor, aðTÞ
mixture, and the evaluation parameters of these tests are highly re-
lated to the behavior of asphalt pavement in the field. The dynamic lgðf r Þ ¼ lgðfÞ þ lgðaðTÞÞ ð2Þ
modulus test was used to measure the dynamic modulus and phase
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

angle of the mixtures. The flow number test was employed to ob- where f = loading frequency, Hz; aðTÞ = shift factor as a function
tain the flow number of the mixtures to evaluate the rutting resis- of temperature; and T = temperature.
tance at high temperature, and the SCB test was utilized to measure
the thermal cracking property of the mixtures at low temperature. Flow Number Test
A UTM-25 test machine was utilized to conduct the above tests. The flow number test is used as a laboratory test to evaluate the
rutting resistance of asphalt mixtures (Witzcak 2002). Specimens
Dynamic Modulus Test for the test were 100 mm in diameter and 100 mm in height. A
The dynamic modulus test, in accordance with the standard method repeated dynamic load was applied in a haversine waveform with
of testing to determine the dynamic modulus and flow number for a wavelength of 0.1 s and followed by a dwell or rest period of 0.9 s.
hot mix asphalt (HMA) using the asphalt mixture performance tester The specimens were loaded in unconfined condition with a devia-
[AASHTO TP79 (AASHTO 2009)], is used to determine the visco- toric stress of 0.7 MPa, and continued until the permanent strain of
elastic characteristics of hot mix asphalt (HMA) mix. Cylindrical 5% or 10,000 loading cycles. The test temperature was set as 60°C.
specimens measuring 150 mm (in diameter) by 178 mm (in height) The typical strain history from a flow number test contains three
were prepared using a Superpave gyratory compactor and were cored zones, i.e., primary, secondary, and tertiary. The permanent strain in
in the center and cut at the ends to obtain 100 mm (in diameter) the secondary zone can be expressed as a linear function of cycle
by 150 mm (in height) specimens with air voids of 5  0.5%. number as follows (Zhou et al. 2004):
The dynamic modulus jE
j and phase angle θ were measured at five εp ¼ aN þ b ð3Þ
temperatures (−10, 4.4, 21.1, 37.8, and 54.4°C) and six loading
frequencies (0.1, 0.5, 1, 5, 10, and 25 Hz) with 138 kPa as the con- where εp = permanent strain; N = number of loading cycles; a,
fining pressure. For each mix, three replicates were used. b = slope and intercept of the linear function, respectively.
In the proposed “2002 Guide for the Design of Pavement During the test, permanent strain versus the number of loading
Systems,” currently under further development in NCHRP Project cycles is recorded automatically, and the results are used to estimate
1-37A, the modulus of the asphalt concrete at all analysis levels of the flow number (FN). The flow number was determined numeri-
temperature and time rate of load is determined from a master curve cally as the cycle number at which the strain rate is at a minimum
constructed at a reference temperature. Master curves are con- based on the Francken model (Bhasin et al. 2004).
structed using the principle of time-temperature superposition
(Bonaquist 2008). The data at various temperatures should be Semicircular Bending Test
shifted with respect to the logarithm of time until the curves merge The SCB test is performed to characterize the fracture resistance of
into a single smooth one. The resulting dynamic modulus master asphalt mixtures (Krans et al. 1996). The semicircular specimens
curve, as a function of time/frequency, describes the time/frequency for this test, 150 mm in diameter by 50 mm in thickness without a
dependency of the material, whereas the amount of shift required at notch in the middle bottom, were cut from cylindrical specimens
each temperature indicates the temperature dependency. In general, with 150 mm in diameter and 120 mm in height prepared using a
the master curve can be mathematically expressed by a sigmoidal SGC. The test set up and specimen configuration are shown in
function (Pellinen and Witczak 2002): Fig. 1. The distance S between the two supporting points is equal

Fig. 1. SCB test setup and specimen configuration

© ASCE 04018105-3 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 to 0.8 times the specimen diameter. Tests were conducted at −10°C void content. Finally, the beam specimens with 50 mm in length,
and under 50 mm=min loading rate. Load and displacement data 10 mm in width, and 10 mm in height for testing were fabricated
were recorded in the test. with a high precision double-side saw. The test specimens had an
The tension stress in the middle bottom of the specimen and air void content of 5  0.5%. The DSR test setup and mortar speci-
stiffness modulus can be calculated using the equation as follows men are shown in Fig. 3(a).
(Molenaar et al. 2002): The bending beam rheometer (BBR) test was used to measure
the low-temperature properties of asphalt mortars. This test method
4.8F
σt ¼ ð4Þ follows the method developed at the University of Minnesota under
BD an NCHRP IDEA project to determine the creep stiffness of thin
1.84F mixture beams with the BBR equipment commonly used to deter-
Mr ¼ ð5Þ mine low-temperature properties of asphalt binder (Marasteanu
BV
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

et al. 2009). Specimens for the test, with 127 mm in length,

where σt = tension stress in the middle bottom of the specimen, 6.35 mm in width and 12.7 mm in height, and an air void content
MPa; M r = stiffness modulus, MPa; F = loading force, N; V = of 5  0.5%, were fabricated in the same way that the first ones
vertical displacement in the middle bottom of specimen, mm; were cored and then cut from cylindrical specimens of 150 mm
B = specimen thickness, mm; and D = specimen diameter, mm. in diameter and 150 mm in height which were compacted by
Asphalt mixtures are considered as elastic material at low tem- SGC. Fig. 3(b) shows the BBR test setup and the specimen.
perature, and the elastic material cracking process can be illustrated
by dissipated energy. According to the material damage criterion, Frequency Sweep Test
the fracture of material can be represented by the function of strain The frequency sweep test according to AASHTO TP7-94
energy density dW=dV, and its calculation equation is as follows: Procedure E is used to measure the complex modulus (G
) and
Z ε phase angle (δ) at different temperatures and a wide range of
dW 0 frequencies. In this test, mortar specimens were tested under five
¼ σij dεij ð6Þ
dV 0 temperatures (30, 40, 50, 60, and 70°C) and 25 loading frequencies
(1, 2, 3, : : : . 24, 25 Hz) with 0.01% of control strain. A aster curve
where σij , εij = stress and strain; and ε0 = critical strain (strain at the of complex modulus at a reference temperature can be built by
peak stress). shifting the data of different temperatures according to the principle
The critical value of dW=dV is the area under the stress-strain of time-temperature superposition. Zeng et al. (2001) developed
curve before the peak stress, which is shown in the shaded area of the CAM model to fit the master curve:
Fig. 2. The larger critical value of the strain energy density indicates
greater energy the material requires for fracture, and thus indicates G
g − G
e
G
¼ G
e þ ð7Þ
better cracking resistance. ½1 þ ðfc =f 0Þk me =k

where G
e = G
(f → 0), equilibrium complex modulus, G
e ¼ 0 for

Test Methods of Mortar
binders and G
e > 0 for mixtures in shear; G
g ¼ G
(f → ∞), glass
A creep performance test for mixtures using a dynamic shear rhe- complex modulus; fc = location parameter with dimensions of
ometer (DSR) was developed in the Transportation Research frequency; f 0 = reduced frequency, function of both temperature
E-circular “New simple performance tests for asphalt mixes” and strain; k, and me = shape parameters, dimensionless.
(Button 2004). In this study, the frequency sweep test and repeated The distance between G
(f c ) and G
g is given by
creep test following this method were conducted in the DSR to
characterize the viscoelasticity and rutting resistance of asphalt 2me =k
R ¼ log ð8Þ
mortars. Izadi et al. (2011) confirmed that using SGC produced 1 þ ð2me =k − 1ÞG
e =G
g
homogeneous air void samples with microstructure that most
closely resembles the mortar of the asphalt mixture microstructure. R is an indicator of the width of the relaxation spectrum. A
In this study, cylindrical specimens of mortar were compacted at higher value is an indication of more gradual transition from the
150°C in the Superpave Gyratory Compactor to a diameter of elastic behavior to the viscous behavior. It indicates less sensitivity
150 mm and a height of 70 mm. First, the gyratory specimens were to frequency changes, generally lower G
values, and higher phase
cored in the center and then cut in the ends because of the higher air angles within the intermediate range of frequency.

Repeated Creep Test

A repeated creep test was developed under NCHRP 9-10 project
“Superpave Protocols for Modified Asphalt Binders” to determine
the high-temperature property for asphalt binders (Bahia et al.
2001). A creep–recovery stress control loading mode is adopted
in this test which considers the high delayed elasticity of modified
asphalt. The creep behavior of asphalt binders can be represented
by the four-element Burgers model as follows:
τ τ  G 
−t 1 τ
γðtÞ ¼ 0 þ 0 1 − e η1 þ 0 t ð9Þ
G0 G1 η0

where γðtÞ = shear strain; τ 0 = constant shear stress; G0 = spring

constant of Maxwell model; G1 = spring constant of Kelvin model;
η0 = dashpot constant of Maxwell model; η1 = dashpot constant of
Fig. 2. Stress versus strain curve of SCB test
Kelvin model; and t = creep time.

© ASCE 04018105-4 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Test setup of asphalt mortar: (a) DSR test; (b) BBR test

The viscous component of the creep stiffness (Gv ) is proposed ductility, softening point, and viscosity. Moreover, the dynamic
as the indicator of high-temperature property for asphalt binders. shear rheometer test according to AASHTO T315-09 (AASHTO
Gv can be calculated using the parameters of the Burgers model 2006b) was conducted at 76°C to measure the rutting factor
according to the follow equation: (G
/sin δ) of the binder, and the bending beam rheometer test
Gv ¼ η0 =t ð10Þ according to AASHTO T313 was conducted at −12°C to measure
the creep stiffness and creep rate (m-value) of the binder. The test
In this test, the mortar specimens were tested by 1 s loading results are listed in Table 2.
and 9 s unloading for 200 cycles at 60°C. The constant shear stress As shown in Table 2, comparing with the virgin bitumen, the
was set as 15 kPa. Gv was obtained by fitting the 100th and 101th RAP binder has much smaller penetration and ductility, and much
creep-recovery curves using Burgers model. bigger viscosity indicates that the RAP binder became stiffer by the
aging effect. However, the softening point of the RAP binder is
BBR Test
lower than that of virgin bitumen. The softening point of the
The BBR test of asphalt mortar were conducted at −12 and 0°C.
The load applied to asphalt mortar at 0°C was 1,960 mN, and the SBS modified bitumen is influenced by both the basis bitumen
load at −12°C was 4,410 mN. Beam deflection during the entire and SBS modifier; therefore, its changes in the aging process would
duration of the creep test was between 30 and 300 m. The creep be more complicated. Previous work (Li 2016) shows that the soft-
stiffness, SðtÞ, and the m-value, mðtÞ, were obtained following ening point of the SBS modified bitumen increases under short-
the same equations described in the binder BBR test method term aging, but decreases when it reaches a certain value under
[AASHTO T313-06 (AASHTO 2006a)]. long-term aging. The DSR test shows that the RAP binder has big-
ger G
/sin δ at 76°C, indicating a higher dynamic shear modulus
jG
j and better rutting resistance at high temperature. The BBR
Results Analysis test shows the RAP binder has bigger creep stiffness and smaller
m-value at −12°C, indicating poorer deformation ability and crack-
Asphalt Binder ing resistance at low temperature. With the increase of the RAP
Asphalt binders of the mixtures were tested to characterize the content, the penetration, ductility, softening point, m-value of
properties of virgin binder, mixed binder, and RAP binder. Binders mixed binder decrease, whereas viscosity, G
/sin δ, and creep stiff-
were first extracted from mixtures by a centrifuge and then recov- ness increase, indicating that the properties of the binder are very
ered using a rotary evaporator. The test item includes penetration, sensitive to the RAP content. The mixed binder could obtain better

© ASCE 04018105-5 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 Table 2. Parameter Values of Binder for a Given RAP Content 30000

RAP content (%) 25000 0%RAP

Parameters 0 30 50 85 100 25%RAP
20000

|E*| (MPa)
Penetration 62 44 37 28 24 55%RAP
(25°C) (0.1 mm) 15000
Ductility (5°C) (cm) 43 26.5 17.3 6.9 3.2 85%RAP
Softening point (°C) 86 80.3 75.8 67.1 62.5 10000 100%RAP
Brookfield viscosity 2.105 2.351 2.539 2.703 2.747
(135°C) (Pa · s) 5000
G
/sin δ (76°C) 1.338 1.824 2.205 2.829 2.985
(kPa) 0
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

Creep stiffness 233 307 373 412 431 1.E-06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06
(−12°C) (MPa) (a) fr (Hz)
m-value (−12°C) 0.382 0.315 0.273 0.242 0.227
1.E+05

rutting resistance with higher RAP content; however, its thermal 1.E+04
cracking potential also increases. 0%RAP

|E*| (MPa)
1.E+03 25%RAP
Asphalt Mixture 55%RAP
1.E+02 85%RAP
Dynamic Modulus Test
100%RAP
The dynamic modulus and phase angle master curves for mixtures
at reference temperature of 21.1°C are plotted in Fig. 4. 1.E+01
1.E-06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06
The discrepancies of dynamic modulus among mixtures with
(b) fr (Hz)
different RAP contents at high frequencies (low temperatures)
are shown in Fig. 4(a), whereas the log-log scale is utilized in 40 0%RAP
Fig. 4(b) to exhibit the discrepancies of dynamic modulus at 25%RAP
35
low frequencies (high temperatures). Test results show that the
55%RAP
fresh mixture (0%RAP) has the smallest jE
j and biggest θ, 30
85%RAP
whereas the 85%RAP or 100%RAP mixtures have the biggest 25
(deg)

jE
j and smallest θ. The minor differences between the test data 20
100%RAP
of 85%RAP and 100%RAP mixtures may be attributed to the test- 15
ing accuracy. The dynamic modulus and phase angle of mixtures
10
are quite sensitive to the RAP content in which the mixtures with
higher RAP have bigger jE
j and smaller θ. Because the phase an- 5
gle represents the ratio of elastic component versus the viscous 0
component of the asphalt material, the fact that it decreases with 1.E-06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06
(c) fr (Hz)
higher RAP content indicates the viscous component decreases
and elastic component increases for asphalt mixtures. This change
Fig. 4. Master curves: (a) dynamic modulus; (b) dynamic modulus in
of viscoelasticity is beneficial for asphalt mixtures to resist perma- log-log scale; (c) phase angle
nent deformation at high temperature, whereas changing viscoelas-
ticity introduces the negative effect on thermal cracking at low
temperature.
Table 3. Flow Number Test Results
Flow Number Test
Secondary zone
The flow number test results of the mixtures are shown in Table 3, Asphalt Air εp ðFNÞ
and permanent strain curves are depicted in Fig. 5. The comparison mixture Number voids (%) a b FN (με)
diagram of FN for the asphalt mixtures is shown in Fig. 6. 0%RAP 1 4.9 40.412 20,980 660 47,810
It is generally accepted that a higher FN indicates a lower rutting 2 5.2 27.128 16,162 620 33,074
susceptibility. As shown in Fig. 6, the FN of 100%RAP is much Mean 5.0 33.77 18,571 640 40,442
higher than the FN of fresh mixture, indicating that the RAP mix- 30%RAP 1 4.8 26.001 18,139 900 41,699
ture has better rutting resistance than the fresh HMA mixture. 2 4.5 23.729 21,385 1,150 49,015
Mean 4.7 24.87 19,762 1,025 45,357
Because the two mixtures have the same aggregate gradations 50%RAP 1 5.3 17.009 15,362 950 31,769
and air voids, the difference in rutting resistance primarily depends 2 4.8 11.4 18,502 1,650 37,464
on the binder. DSR test results of bitumen in Table 2 show that Mean 5.1 14.20 16,932 1,300 34,617
comparing with the virgin bitumen, the RAP binder becomes stiffer 85%RAP 1 4.7 12.427 18,953 2,000 44,035
and its high-temperature property enhances, which resulted in the 2 4.9 5.9474 20,245 3,400 40,674
improvement of rutting resistance of the RAP mixture. It is ob- Mean 4.8 9.19 19,599 2,700 42,355
served that with the increase of RAP contents, the FN of asphalt 100%RAP 1 4.9 8.1552 20,509 2,800 43,564
mixtures increases, indicating better rutting resistance. Conclusions 2 4.9 7.7316 19,116 3,100 43,429
Mean 4.9 7.94 19,812.5 2,950 43,497
could be made that with proper design of gradation and volume,

© ASCE 04018105-6 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 asphalt mixtures with high RAP content exhibit better rutting As shown in Fig. 7, comparing with 0%RAP mixture, the 100%
resistance than fresh HMA mixtures. RAP mixture has a lower critical value of strain energy density (J c )
and tension stress (σt ); especially J c is not even half that of 0%RAP
SCB Test mixture. It indicates that the low-temperature cracking resistance of
The critical value of strain energy density (J c ) and tension stress a RAP mixture is far worse than a fresh HMA mixture. The low-
(σt ) of the mixtures measured by SCB tests are shown in Fig. 7. temperature property of an asphalt mixture primarily depends on
the asphalt binder, and the low-temperature property of aged
RAP binder declines severely compared with the virgin bitumen,
50000 which can be concluded from the BBR test results of binders shown
45000 in Table 2. Also, with the increase of RAP content, the J c of the
40000 asphalt mixture decreases, meaning that the fracture energy of the
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

35000 mixture reduces, and the potential susceptibility to thermal crack-

30000 ing increases. As a result, applying the RAP in the HMA mixture
0%RAP-1 0%RAP-2
would lead to the decline of cracking resistance at low temperature,
25000
and it needs to be seriously considered in the design of mixtures
p

20000 30%RAP-1 30%RAP-2

with high RAP.
15000 50%RAP-1 50%RAP-2
10000 85%RAP-1 85%RAP-2
Asphalt Mortar
5000 100%RAP-1 100%RAP-2
0
0 1000 2000 3000 4000 5000 Frequency Sweep Test
Loading Cycles The complex modulus master curves of various asphalt mortars at
referenced temperature of 50°C are shown in Fig. 8. The parameter
Fig. 5. Permanent strains versus number of loading cycle curves values of the CAM model fitted to the master curves are summa-
rized in Table 4.
As the test results show, 0%RAP mortar has the lowest complex
4000
modulus (jG
j), whereas 100%RAP mortar has the highest jG
j. G
e
represents the modulus at very low frequencies or high tempera-
3500 tures, and it increases with higher RAP contents. G
g represents
the modulus at very high frequencies or low temperatures, and
3000
it also increases with higher RAP contents with acceptable fitting
2500 error at 30%RAP mortar. R is the indicator of rheological property,
and higher R indicates a wider relaxation spectrum. Table 4 shows
FN

2000
that 0%RAP mortar has the largest R value, which indicates the
1500 fresh mortar has the best rheological property. Mortars with higher
content of RAP have lower R values, and when the RAP content
1000 increases to more than 50%, the R value decreases sharply. The
500 results indicate that mortar with higher RAP has lower sensitivity
640 1025 1300 2700 2950 to frequency change or temperature change, and the rheological
0 property declines.
0%RAP 30%RAP 50%RAP 85%RAP 100%RAP
Repeated Creep Test
Fig. 6. Flow number of asphalt mixtures with different contents of
Creep-Recovery strain data for asphalt mortars in each cycle are
RAP
recorded in this test, and Fig. 9 shows the strain curves of the

160.0 13.1 13.5

140.0 12.6 13.0

12.3
120.0 12.5
12.0
100.0 11.5
Jc (kJ.m-3 )

(MPa)

11.2 11.5
80.0
11.0
t

60.0
10.5
40.0 10.0
20.0 9.5
126.1 93.6 84.4 60.3 55.6
0.0 9.0
0%RAP 30%RAP 50%RAP 85%RAP 100%RAP
Fig. 7. Experimental results for Jc and σt of asphalt mixtures

© ASCE 04018105-7 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 1.0E+10 Table 5. Burgers Model Fitting Results of Creep and Recovery
0%RAP
30%RAP Asphalt
50%RAP
mortar G0 (Pa) η0 (Pa · s) G1 (Pa) η1 (Pa · s) Gv (Pa)
1.0E+09
85%RAP 0%RAP 6.18 × 108 4.59 × 109 1.73 ×109 8.91 × 107 4.59 × 109
6.56 × 108 4.94 × 109 2.18 × 109 1.25 × 108 4.94 × 109
|G*| (Pa)

30%RAP
100%RAP
50%RAP 7.98 × 108 5.55 × 109 4.38 × 109 4.45 × 108 5.55 × 109
85%RAP 1.11 × 109 1.58 × 1010 5.11 × 109 2.50 × 108 1.58 × 1010
1.0E+08 100%RAP 1.36 × 109 2.30 × 1010 9.87 × 109 6.07 × 108 2.30 × 101

4000
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

1.0E+07 3383
1.0E-05 1.0E-03 1.0E-01 1.0E+01 1.0E+03 1.0E+05 3500 3160
fr (Hz) 2936
3000
2380 2523
2500

S (60s) (MPa)
Fig. 8. Complex modulus master curves of asphalt mortars
2000
0°C
1500
-12°C
Table 4. CAM Model Fitting Results of Master Curves 1000 804 867
500 596
396
Asphalt mortar 500
Parameters 0%RAP 30%RAP 50%RAP 85%RAP 100%RAP 0
0 30% 50% 85% 100%
G
e (Pa) 2.04 × 10 3.01 × 10 3.60 × 10 4.88 × 107 5.14 × 107
7 7 7
(a)
G
g (Pa) 1.158 × 109 2.1E × 109 1.3 × 109 1.69 × 109 2.84 × 109 RAP Content
f c (Hz) 14,136.7 13,159.4 11,010.6 13,598.7 14,253.3
k 2.76 5.56 10.11 8.18 8.63 0.5
me 0.46 0.54 0.38 0.41 0.36 0.432
0.403
R 0.049 0.029 0.011 0.014 0.012 0.4 0.372
0.321 0.311
m-value (60s)

0.3
0.247
0.226 0.215
4.0E-05 0.182 0°C
0.2 0.174
-12°C
0%RAP
3.0E-05 30%RAP 0.1

50%RAP
Strain

85%RAP 0
2.0E-05 0 30% 50% 85% 100%
100%RAP (b) RAP Content

1.0E-05 Fig. 10. Experimental results for asphalt mortars: (a) creep stiffness;
(b) m-value
0.0E+00
990 995 1000 1005 1010 1015
Time (S) of Gv indicates a better rutting resistance of mortar with the
increase of RAP content.
Fig. 9. Creep and recovery of the 100th and 101th cycles for asphalt
mortars BBR Test
The BBR system gave the force, deflection, creep stiffness (S), and
m-value at 8, 15, 30, 60, 120 and 240 s. The creep stiffness and
m-value at 60 s were utilized to evaluate the low-temperature prop-
100th and 101th cycles. The parameter values of the Burgers model erty in this test. The creep rate (m-value) represents the ability to
fitted the curves and are summarized in Table 5. relax under stress. A lower stiffness and a higher m-value indicate
1. The creep-recovery curves show excellent delayed elasticity re- better resistance to thermal cracking. The creep stiffness and
covery ability of all the mortars for only a small strain residual m-value obtained in the tests are shown in Fig. 10.
after the recovery period. It is obvious that both the elastic strain As shown in Fig. 10, asphalt mortar has higher creep stiffness
and accumulated permanent strain in the loading period de- and lower m-value at −12°C than 0°C, indicating higher thermal
crease with higher RAP content, which indicates the mortar with cracking susceptibility at lower temperature. With the increase
higher RAP has better resistance for permanent deformation at of RAP content, the mortar’s stiffness increases, and the m-value
high temperature. The Burgers model fits the creep-recovery decreases, 0%RAP mortar has the smallest stiffness and biggest
curves well, and the fitting results listed in Table 5 show that m-value, whereas 100%RAP mortar has the biggest stiffness and
the four parameters of the Burgers model regarding the mortar smallest m-value. The authors conclude that the low-temperature
viscoelasticity, namely elasticity modulus G0 , G1 and viscosity cracking resistance of RAP mortar declines compared with
η0 , η1 , increase with higher RAP. According to the meaning of the fresh mortar, which is primarily caused by the inferior low-
the viscous component of the creep stiffness, a higher value temperature property of aged binder. Asphalt mortar’s ability

© ASCE 04018105-8 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 to resist low-temperature cracking decreases with higher RAP Table 6 shows that mechanical properties of the binder, mortar,
content. and mixture at high and low temperatures are highly related with
correlation coefficients larger than 0.9. The correlations of high-
temperature modulus between binder, mortar, and mixture are
Correlation Analysis not as significant as those of low-temperature modulus. It signifies
According to the test results for asphalt binders, mortar, and mix- that the binder has greater effect on asphalt mixture modulus at low
ture, the complex modulus (jG
j) and high-temperature rutting temperature than at high temperature. A similar conclusion could
resistance improve with higher RAP content, whereas the low- be reached by comparing the correlations of rutting resistance and
temperature cracking resistance declines. To quantitatively investi- thermal cracking resistance. The reason for the above phenomenon
gate the performance correlations of asphalt in different scales, is that asphalt binder behaves much more viscous than elastic at
three sets of data measured for binder, mortar, and mixture, respec- high temperature which leads to the coarse aggregates dominating
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

tively, were correlated pairwise by calculating the Pearson correla- the mechanical properties of the asphalt mixture.
tion coefficient (r) using the following equation: The lower correlation of rutting resistance between binder and
mortar indicates the significant influence of fine aggregates on the
covðx; yÞ rutting resistance of mortar material scale, and the lower correlation
r ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð11Þ
DðxÞ þ DðyÞ of high-temperature modulus between the binder and mixture also
shows the importance of fine aggregates in the high-temperature
For high-temperature modulus analysis, the complex shear mechanics of mortar. Mortar produced in this study could be seen
modulus jG
j (76°C, 10 Hz) of binder, complex shear modulus as a 4.75 mm NMAS dense graded mixture, and the fine aggregates
jG
j (60°C, 10 Hz) of mortar and dynamic modulus jE
j (54°C, for the mixtures scale converts to the coarse aggregates in the mor-
10 Hz) of mixture are used, whereas the creep stiffness S tar scale. Therefore, it is not difficult to understand the significant
(−12°C) of binder and mortar, stiffness modulus Mr (−10°C) of influence of fine aggregates on the mechanical properties of mortar
mixture are used for low-temperature modulus analysis. For rutting at high temperature.
resistance analysis, rutting factor G
/sin δ (76°C) of binder, the vis- For high-temperature and low-temperature modulus, the corre-
cous component of creep stiffness Gv (60°C) of mortar, and lation between binder and mortar are better than that between mor-
flow number FN (60°C) of mixture are selected. For thermal crack- tar and mixture, which indicates that the effect of coarse aggregates
ing resistance, the m-value (−12°C) of binder and mortar, the criti- on the modulus of the mixture is higher than that of fine aggregates.
cal value of strain energy density J c (−10°C) of mixture are Better correlations of modulus are observed between the binder and
adopted for analysis. Linear fitting was then applied to show the mixture than between the mortar and mixture, meaning that the as-
correlations among the parameters of asphalt materials, as depicted phalt binder could be a more reliable material scale to predict the
in Figs. 11–13, and the goodness of fit (R2 ) values are marked in the modulus of the mixture. The lower correlations of modulus be-
figures. The correlation coefficient (r) values for different data sets tween the mortar and mixture may be caused by the fundamental
are provided in Table 6. differences of the indexes (the creep stiffness of mortar and the

250 3500
|G*| of Mortar (MPa)

200
S of Mortar (MPa)

R² = 0.7647 3000 R² = 0.9426

150

100
2500
50

0 2000
1.0 1.5 2.0 2.5 3.0 200 250 300 350 400 450
(a) |G*| of Binder (kPa) (b) S of Binder (MPa)

2.5E+07 0.270

2.0E+07
Gv of Mortar (kPa)

m-value of Mortar

R² = 0.9205
R² = 0.7916 0.230
1.5E+07

1.0E+07
0.190
5.0E+06

0.0E+00 0.150
1.0 1.5 2.0 2.5 3.0 0.2 0.3 0.3 0.4 0.4
(c) G*/sin of Binder (kPa) (d) m-value of Binder

Fig. 11. Linear fitting curves of binder versus mortar for (a) high-temperature modulus; (b) low-temperature modulus; (c) rutting resistance;
(d) thermal cracking resistance

© ASCE 04018105-9 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 3000 800

|E*| of Mixture (MPa)

Mr of Mixture (MPa)
2500
700
R² = 0.9561
2000 R² = 0.898
600
1500
500
1000

500 400
1.0 1.5 2.0 2.5 3.0 200 250 300 350 400 450
|G*| of Binder (kPa) S of Binder (MPa)
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

(a) (b)

4000 140

Jc of Mixture (MPa)
120
3000
FN of Mixture

100 R² = 0.9791
2000 R² = 0.9455
80
1000
60

0 40
1.0 1.5 2.0 2.5 3.0 0.2 0.3 0.3 0.4 0.4
(c) G*/sin of Binder (kPa) (d) m-value of Binder

Fig. 12. Linear fitting curves of binder versus mixture: (a) high-temperature modulus; (b) low-temperature modulus; (c) rutting resistance; (d) thermal
cracking resistance

3000 800
|E*| of Mixture (MPa)

Mr of Mixture (MPa)

2500
700
2000
600
1500 R² = 0.6528 R² = 0.8174

500
1000

500 400
0.0 100.0 200.0 300.0 2000 2500 3000 3500
(a) |G*| of Mortar (MPa) (b) S of Mortar (MPa)

4000 140
Jc of Mixture (MPa)

120
3000
FN of Mixture

100
2000
R² = 0.9174 80
R² = 0.9601
1000
60

0 40
0.0E+00 1.0E+07 2.0E+07 3.0E+07 0.150 0.190 0.230 0.270
(c) Gv of Mortar (kPa) (d) m-value of Mortar

Fig. 13. Linear fitting curves of mortar versus mixture: (a) high-temperature modulus; (b) low-temperature modulus; (c) rutting resistance; (d) thermal
cracking resistance

© ASCE 04018105-10 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 Table 6. Correlation Coefficient r of Statistical Analysis However, it is noted that the lack of study on mastic (a scale be-
Binder versus Binder versus Mortar versus tween binder and mortar), which is composed of asphalt binder and
Property mortar mixture mixture filler, leads to difficulties in distinguishing the effect of filler and
fine aggregate on properties of asphalt mixtures. Therefore, inves-
High-temperature modulus 0.874 0.948 0.808
Low-temperature modulus 0.971 0.978 0.904
tigation into the mechanical properties of mastic and correlation
Rutting resistance 0.890 0.972 0.958 analysis with other asphalt scales would be the emphasis in a fur-
Thermal cracking resistance 0.959 0.990 0.980 ther study. Besides, to keep consistency of the volumetric character-
istics of mortars and mixtures, cylinder specimens of the asphalt
mortar would be utilized in the DSR tests to further characterize
the mortars.
stiffness modulus of mixture) selected to represent the material
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

mechanical stiffness, or the inconsistency of test methods (i.e., the

volumetric differences between cuboid specimens for mortar tests Acknowledgments
and cylinder specimens for mixture tests) used to measure the
modulus of the mortar and mixture. Based on the above consider- The authors acknowledge Mr. Kai Xu and Mr. Zhou Zhou for their
ations, consistent cylinder specimen tests applied for the mortar help on experiments of the study.
study and mechanical parameters obtained from the tests should
be utilized and analyzed in further work.
References

Summary and Conclusions AASHTO. (2006a). “Standard method of test for determining the flexural
creep stiffness of asphalt binder using the bending beam rheometer
The mechanical properties of asphalt mixtures with high RAP were (BBR).” AASHTO T313-06, Washington, DC.
AASHTO. (2006b). “Standard method of test for determining the rheologi-
investigated in this paper through three asphalt scales of binder,
cal properties of asphalt binder using a dynamic shear rheometer
mortar, and mixture. RAP contents of 30, 50, and 85% were (DSR).” AASHTO T315-09, Washington, DC.
selected for the tests; meanwhile, fresh asphalt (0%RAP) and AASHTO. (2009). “Standard method of test for determining the dynamic
100%RAP were prepared for comprehensive comparison. The cor- modulus and flow number for hot mix asphalt (HMA) using the asphalt
relations of mechanical properties between binder, mortar, and mixture performance tester (AMPT).” AASHTO TP79, Washington,
mixture were quantitatively analyzed by calculating the Pearson DC.
correlation coefficient (r). Important observations and conclusions Aigner, E., Lackner, R., and Pichler, C. (2009). “Multiscale prediction
from this study could be summarized as follows: of viscoelastic properties of asphalt concrete.” J. Mater. Civ. Eng.,
• For the mechanical properties of asphalt mortars, with higher 10.1061/(ASCE)0899-1561(2009)21:12(771), 771–780.
RAP contents, the complex modulus jG
j increases, and the Bahia, H. U., Hanson, D. I., Zeng, M., Zhai, H., Khatri, M. A., and
Anderson, R. M. (2001). “Characterization of modified asphalt binders
rheological indicator R value decreases, indicating lower sensi-
in superpave mix design.” NCHRP Rep. No. 459, National Academy
tivity to frequency change or temperature change. With higher Press, Washington, DC, 46–49.
RAP content, lower accumulated permanent strains and higher Bennert, T., Daniel, J., and Mogawer, W. (2014). “Strategies for incorpo-
viscous components of the creep stiffness (Gv ) are observed rating higher RAP percentages: Review of northeast states implemen-
from repeated creep tests, signifying better rutting resistance tation trials.” Transp. Res. Rec., 2445, 83–93.
of mortars; meanwhile, the increasing creep stiffness and de- Bhasin, A., Button, J., and Chowdhury, A. (2004). “Evaluation of simple
creasing m-value observed from the BBR test indicate declining performance tests on hot-mix asphalt mixtures from south central
low-temperature cracking property of mortar. United States.” Transp. Res. Rec., 1891, 174–181.
• For the mechanical properties of asphalt mixtures, with higher Bonaquist, R. F. (2008). Refining the simple performance tester for use in
RAP contents, the dynamic modulus increases and phase angle routine practice, Vol. 614, Transportation Research Board, Washington,
decreases, indicating that the viscous component decreases and DC.
the elastic component increases. The increasing FN obtained Buttlar, W., Bozkurt, D., Al-Khateeb, G., and Waldhoff, A. (1999).
“Understanding asphalt mastic behavior through micromechanics.”
from the flow number test and the decreasing critical value
Transp. Res. Rec., 1681, 157–169.
of strain energy density (J c ) obtained from the SCB test indicate
Button, J. W. (2004). New simple performance test for asphalt mixes,
better rutting resistance and lower thermal cracking property of Transportation Research Board, Washington, DC, 1–44.
mixtures with higher RAP content. Copeland, A. (2011). “Reclaimed asphalt pavement in asphalt mixtures:
• The correlations of high-temperature properties between binder, State of the practice.” Rep. No. FHWA-HRT-11-021, Transportation Re-
mortar, and mixture are not as significant as those of the low- search Board, Washington, DC.
temperature properties, signifying that the binder has greater ef- Cucalon, L. G., Rahmani, E., Little, D. N., and Allen, D. H. (2016). “A
fect on asphalt mixture properties at low temperature than at multiscale model for predicting the viscoelastic properties of asphalt
high temperature. The lower correlations of high-temperature concrete.” Mech. Time-Dependent Mater., 20(3), 325–342.
modulus and rutting resistance between the binder and the mor- Dai, Q., and You, Z. (2007). “Prediction of creep stiffness of asphalt mix-
tar indicate the significant influence of fine aggregates on high- ture with micromechanical finite-element and discrete-element models.”
temperature properties of the mortars. Better correlations of J. Eng. Mech., 10.1061/(ASCE)0733-9399(2007)133:2(163), 163–173.
modulus between the binder and mortar than that between Diefenderfer, S., and Harikrishnan, N. (2014). “Evaluation of high RAP
mixture production, construction, and properties.” Proc., 93rd Annual
the mortar and mixture indicates that the effect of coarse aggre-
Meeting of the Transportation Research Board, Transportation Re-
gates on the modulus of the mixture is higher than that of fine search Board, Washington, DC.
aggregates. Hong, F., Chen, D. H., and Mikhail, M. (2010). “Long-term performance
The outcomes of this study regarding the correlations between evaluation of recycled asphalt pavement results from Texas: Pavement
the mortar, mixture, and binder is meaningful for providing insight studies category 5 sections from the long-term pavement performance
to the fundamental properties of asphalt mixtures with high RAP. program.” Transp. Res. Rec., 2180, 58–66.

© ASCE 04018105-11 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105

 Izadi, A., Bhasin, A., and Motamed, A. (2011). “Designing fine aggregate Molenaar, A. A. A., Scarpas, A., Liu, X., and Erkens, S. M. J. G. (2002).
mixtures to evaluate fatigue crack growth in asphalt mixtures.” Rep. No. “Semi-circular bending test; simple but useful?” J. Assoc. Asphalt
SWUTC/11/161022-1, Univ. of Texas at Austin, TX. Paving Technol., 71(1), 794–815.
Jiangsu Bureau of Quality and Technical Supervision. (2008). “Standard Nash, T., Sholar, G., Page, G., and Musselman, J. (2012). “Evaluation of
specification for construction of Jiangsu province expressway asphalt asphalt mixture with high percentage of reclaimed asphalt pavement in
pavements.” DB32/T1246-2008, Jiangsu, China. Florida.” Transp. Res. Rec., 2294, 16–25.
Kim, Y. R. (2003). “Mechanistic fatigue characterization and damage Nicholls, R. L. (1976). Composite construction material handbook,
modeling of asphalt mixtures.” Ph.D. dissertation, Texas A&M Univ., Prentice Hall, Englewood Cliffs, NJ.
College Station, TX. Pellinen, T. K., and Witczak, M. W. (2002). “Stress dependent master curve
Krans, R. L., Tolman, F., and Van de Ven, M. F. C. (1996). “Semi-circular construction for dynamic (complex) modulus (with discussion).” J. As-
bending test: A practical crack growth test using asphalt concrete soc. Asphalt Paving Technol., 71(1), 281–309.
cores.” RILEM Proc., Chapman & Hall, London, 123–132. Underwood, B. S., and Kim, Y. R. (2011). “Experimental investigation into
Lackner, R., Spiegl, M., Blab, R., and Eberhardsteiner, J. (2005). “Is
Downloaded from ascelibrary.org by S.V. National Institute Of Technology on 09/23/24. Copyright ASCE. For personal use only; all rights reserved.

the multiscale behavior of asphalt concrete.” Int. J. Pavement Eng.,

low-temperature creep of asphalt mastic independent of filler shape 12(4), 357–370.
and mineralogy? Arguments from multiscale analysis.” J. Mater. Civ. Underwood, B. S., and Kim, Y. R. (2013a). “Effect of volumetric factors
Eng., 10.1061/(ASCE)0899-1561(2005)17:5(485), 485–491. on the mechanical behavior of asphalt fine aggregate matrix and the
Li, J. (2016). “Research on hot in-place recycling technology for modified relationship to asphalt mixture properties.” Constr. Build. Mater.,
asphalt pavement.” Ph.D. dissertation, Southeast Univ., Nanjing, China.
49(12), 672–681.
Li, Y., and Metcalf, J. B. (2005). “Two-step approach to prediction of
Underwood, B. S., and Kim, Y. R. (2013b). “Microstructural investigation
asphalt concrete modulus from two-phase micromechanical models.”
of asphalt concrete for performing multiscale experimental studies.”
J. Mater. Civ. Eng., 10.1061/(ASCE)0899-1561(2005)17:4(407),
Int. J. Pavement Eng., 14(5), 498–516.
407–415.
Valenta, R., Sejnoha, M., and Zeman, J. (2010). “Macroscopic constitutive
Loria, L., Hajj, E., Sebaaly, P., Barton, M., Kass, S., and Liske, T. (2011).
“Performance evaluation of asphalt mixtures with high recycled asphalt law for mastic asphalt mixtures from multiscale modeling.” Int. J. Multi-
pavement content.” Transp. Res. Rec., 2208, 72–81. scale Comput. Eng., 8(1), 131–149.
Ma, T. (2010). “Research on hot in-place recycling technology of SMA.” Witzcak, M. W. (2002). Simple performance test for superpave mix
Ph.D. dissertation, Southeast Univ., Nanjing, China. design, Vol. 465, Transportation Research Board, Washington, DC,
Marasteanu, M., Velasquez, R., Cannone Falchetto, A., and Zofka, A. 73–79.
(2009). “Development of a simple test to determine the low temperature Zeng, M., Bahia, H. U., Zhai, H., Anderson, M. R., and Turner, P. (2001).
creep compliance of asphalt mixtures.” IDEA Program Final Rep. “Rheological modeling of modified asphalt binders and mixtures (with
NCHRP, Transportation Research Board, Washington, DC, 133. discussion).” J. Assoc. Asphalt Paving Technol., 70(1), 403–441.
Ministry of Transport of the People’s Republic of China. (2011a). “Stan- Zhou, F., Scullion, T., and Sun, L. (2004). “Verification and modeling
dard method of test for determining the binder content of asphalt of three-stage permanent deformation behavior of asphalt mixes.” J.
mixture using a centrifugal extractor.” JTG E20-2011, T0722, Beijing. Transp. Eng., 10.1061/(ASCE)0733-947X(2004)130:4(486), 486–494.
Ministry of Transport of the People’s Republic of China. (2011b). “Stan- Zhu, H., and Nodes, J. E. (2000). “Contact based analysis of asphalt
dard method of test for determining the gradation of aggregates of pavement with the effect of aggregate angularity.” Mech. Mater., 32(3),
asphalt mixture.” JTG E20-2011, T0725, Beijing. 193–202.
Ministry of Transport of the People’s Republic of China. (2011c). “Stan- Zollinger, C. J. (2005). “Application of surface energy measurements to
dard test methods of bitumen and bituminous mixtures for highway evaluate moisture susceptibility of asphalt and aggregates.” Masters
engineering.” JTG E20-2011, Beijing. thesis, Texas A&M Univ., College Station, TX.

© ASCE 04018105-12 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2018, 30(6): 04018105