Construction and Building Materials 73 (2014) 479–489

Contents lists available at ScienceDirect

Construction and Building Materials

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

Evaluation of moisture susceptibility of asphalt mixes containing RAP

and different types of aggregates and asphalt binders using the surface
free energy method
Rouzbeh Ghabchi a,⇑, Dharamveer Singh b,1, Musharraf Zaman c,2
a
School of Civil Engineering and Environmental Science, University of Oklahoma, 202 W. Boyd St., CEC 334, Norman, OK 73019, USA
b
Department of Civil Engineering, Indian Institute of Technology (IIT), Bombay 400076, India
c
College of Engineering, University of Oklahoma, 202 W. Boyd St., CEC 450(C), Norman, OK 73019, USA

h i g h l i g h t s

 The binder–aggregate work of adhesion increased with an increase in RAP content.

 The binder–aggregate work of debonding decreased with an increase in RAP content.
 Use of RAP in an asphalt mix may reduce the moisture-induced damage potential.
 A high total SFE component of aggregate may increase moisture damage potential.

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

Article history: The surface free energy (SFE) measurement of asphalt binder and aggregate is considered a reliable mech-
Received 22 February 2014 anistic framework for evaluating the moisture-induced damage potential of asphalt mixes. In the present
Received in revised form 10 September 2014 study, the SFE method was used to evaluate the effects of asphalt binder type, Reclaimed Asphalt Pave-
Accepted 17 September 2014
ment (RAP) and its amount, and aggregate type on the moisture-induced damage potential of asphalt
mixes. The SFE components (non-polar, acid and base) of a PG 64-22 and a PG 76-28 (polymer-modified)
asphalt binders, blended with different amounts of RAP binder (0%, 10%, 25% and 40%) were measured in
Keywords:
the laboratory using a Dynamic Contact Angle (DCA) analyzer. Also, the SFE components of six types of
Surface free energy
Asphalt mix
aggregates, namely limestone, rhyolite, sandstone, granite, gravel, and basalt were used in this study.
Reclaimed Asphalt Pavement The SFE components of limestone and rhyolite aggregates were measured using a Universal Sorption
Moisture-induced damage Device (USD), while those of the sandstone, granite, gravel, and basalt aggregates were obtained from
Work of adhesion the literature. The energy ratio parameters estimated based on the spreading coefficient, the work of
Work of debonding adhesion, and the work of debonding were used to assess the moisture-induced damage potential of dif-
Wettability ferent combinations of asphalt binders and different RAP binder contents and aggregates. The SFE test
results indicated that the acid SFE component of PG 64-22 and PG 76-28 asphalt binders increase with
the addition of RAP binder, while the base SFE component remains almost unchanged. Also, the wettabil-
ity and the work of adhesion of both PG 64-22 and PG 76-28 asphalt binders over different types of aggre-
gates increased with an increase in RAP content (by 25% and more). Based on the energy ratio parameters,
it was found that the resistance to moisture-induced damage increased with an increase in RAP content
for both PG 64-22 and PG 76-28 asphalt binders and all types of aggregates, specifically when higher RAP
contents were used. Moreover, it was found that the higher the total SFE of the aggregates, the lower the
energy ratio parameter values. Therefore, a high total SFE component of aggregate may result in a high
moisture-induced damage potential of the mix. The results presented herein are expected to be helpful
in mechanistically assessing the moisture-induced damage potential of asphalt mixes, produced with
polymer-modified and non-polymer-modified asphalt binders, containing RAP.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
⇑ Corresponding author. Tel.: +1 (405) 818 2845.
E-mail addresses: [email protected] (R. Ghabchi), [email protected]
(D. Singh), [email protected] (M. Zaman).
Hot Mix Asphalt (HMA) is the most widely used paving material
1
Tel.: +91 (976) 901 7304. in the U.S. Nationwide, more than 90% of pavements are paved
2
Tel.: +1 (405) 325 2626. with asphalt [1]. Each year, over 550 million tons of HMA are

http://dx.doi.org/10.1016/j.conbuildmat.2014.09.042
0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.
480 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489

produced and used for construction of flexible pavements. Increas- aggregates were measured using a USD device, while the SFE com-
ing environmental awareness, rising oil prices and scarcity of high ponents of sandstone, granite, gravel, and basalt aggregates were
quality aggregates in many areas spur development of methods obtained from the literature. Consequently, a wide range of binder
and technologies for reduced use of virgin asphalt binder and types, RAP amounts, and aggregates were covered in this study.
aggregates and increased use of recycled or reclaimed materials. Finally, wettability, work of adhesion, work of debonding, and
Therefore, over the past two decades, many transportation agen- energy ratio parameters were estimated to mechanistically discuss
cies, asphalt producers and pavement construction companies the effect of binder type, RAP amount and aggregate types on the
have taken major initiatives to implement green paving technologies moisture-induced damage potential of asphalt mixes containing
[2]. Use of Reclaimed Asphalt Pavement (RAP) in asphalt produc- RAP. The results presented herein are expected to be helpful in
tion is an important element of such initiatives. Based on a report evaluation of aggregates–asphalt binder–RAP combinations, dur-
published by the National Asphalt Pavement Association (NAPA), ing the material selection for asphalt mixes, in order to minimize
asphalt is being recycled and reused at a rate of over 99% [2]. Recy- the possibility of moisture damage in pavement.
cling of asphalt pavements and asphalt shingles conserved 20.5
million barrels of asphalt binder in 2010 [2]. Using RAP in HMA 2. Objectives
not only benefits the economy and environment by cutting the cost
and preserving the natural resources, but also increases the rut This study aims to evaluate the effect of asphalt binder type,
resistance of the mix [3–6]. Concerns over premature pavement RAP binder, and aggregate type on the moisture-induced damage
distresses due to increased RAP amounts in HMA, limits the use potential of mixes using the SFE approach. The specific objectives
of RAP. For example McDaniel et al. [7] reported that incorporating of this study were to:
more than 20% RAP in asphalt mixes resulted in reduced fatigue
life, when compared with that of the virgin mix. In a similar study, 1. Determine the SFE components of PG 64-22 and PG 76-28
Shu et al. [8] concluded that using RAP in HMA reduces the fatigue asphalt binders with and without addition of different
life of asphalt mixes. In addition to rut and fatigue, moisture- amounts of RAP binder (i.e., 0%, 10%, 25%, and 40%), using a
induced damage is another important distress in HMA pavements, Dynamic Contact Angle (DCA) analyzer.
including those containing RAP. 2. Determine the SFE components of different types of aggre-
Moisture-induced damage is defined as the loss of asphalt bin- gates, using a Universal Sorption Device (USD).
der’s tensile strength (cohesive failure) or bonding failure at the 3. Evaluate the asphalt binder’s coating quality with and with-
asphalt binder–aggregate interface (adhesive failure), due to the out addition of RAP binder on different types of aggregates
presence of moisture [9]. Retained indirect Tensile Strength Ratio using wettability parameter (spreading coefficient).
(TSR) and Hamburg Wheel Tracking (HWT) tests, in accordance 4. Evaluate moisture-induced damage potential of mixes con-
with AASHTO T 283 [10] and AASHTO T 324 [11], respectively, taining RAP with different types of aggregates and asphalt
are used to evaluate moisture-induced damage potential of asphalt binders based on the energy ratio parameters estimated
mixes. However, it is reported that a TSR test may fail to correlate based on wettability, work of adhesion, and work of
with the field observations, due to its empirical nature and lacking debonding.
a mechanistic base [12]. Similarly, HWT test does not directly
address the ‘‘failure mechanism’’ that governs the stripping in 3. Background on surface free energy
asphalt pavements. Therefore, using a mechanistic approach to
characterize the moisture-induced damage of asphalt mixes con- 3.1. Surface free energy components
taining RAP is important to stripping evaluation.
Recently, the surface free energy (SFE) method has been used The SFE of a solid (or liquid) is defined as the work required for
successfully, to evaluate the moisture-induced damage potential increasing its surface by a unit area under vacuum [20]. Van Oss
of asphalt binder–aggregate systems [9,12–19]. The SFE method et al. [20] proposed a theory (Good-van Oss-Chaudhury theory)
is a mechanistic approach which directly addresses the adhesion which suggests three components of SFE. Based on the Good-van
and debonding of the asphalt binder and aggregates, in presence Oss-Chaudhury theory, the total SFE can be expressed in the form
of moisture. Many studies have applied the SFE approach to assess of: (i) a monopolar acidic component (C+), (ii) a monopolar basic
moisture susceptibility of different asphalt binders, Warm Mix component (C), and (ii) an apolar or Lifshitz–van der Waals com-
Asphalt (WMA) additives, and anti-stripping agents in contact with ponent (CLW). Also, the total SFE component (Ctotal) can be
different types of aggregates [9,12–19]. However, to the authors’ expressed in terms of a Lifshitz–van der Waals (CLW) and an
knowledge, no study was found in the literature to evaluate the acid-base (CAB) component, shown in Eqs. (1) and (2), respectively.
moisture-induced damage potential of mixes containing RAP, using
the SFE method. RAP contains aged and stiff binder, which may CTotal ¼ CLW þ CAB ð1Þ
alter the chemical and surface properties of a virgin binder when
where,
mixed in different proportions. Hence, the SFE components of an pffiffiffiffiffiffiffiffiffiffiffiffiffi
asphalt mix prepared with different amounts of RAP are expected CAB ¼ 2 Cþ C ð2Þ
to be different than those for a virgin mix. This, in turn, may affect
the adhesion potential and bond strength of the asphalt binder– 3.1.1. Surface free energy of asphalt binder
aggregate system. Therefore, the current study was undertaken Thomas Young in 1805, described the occurrence of the wetting
to evaluate the effect of virgin asphalt binders with different Per- and spreading of a liquid over a surface to be directly related to the
formance Grades (PG), different amounts of RAP, and different interaction between the cohesive and adhesive forces. Therefore,
types of aggregates, on the moisture-induced damage potential contact angle of a liquid over a solid surface determines the wetta-
of asphalt mixes. The SFE measurements consisted of testing two bility of the liquid over a surface. This contact angle is known to be
types of asphalt binders, namely PG 64-22 and PG 76-28, each a function of the surface energies of the system. Thus, the SFE com-
mixed with different amounts of RAP binder, namely 0%, 10%, ponents of an asphalt binder can be determined by measuring the
25%, and 40%, using a DCA analyzer. The aggregates included in this contact angles (h) between the asphalt binder and three different
study consisted of limestone, rhyolite, sandstone, granite, gravel, solvents (one apolar, one monopolar and one bipolar solvent),
and basalt. The SFE components of the limestone and rhyolite using the Wilhelmy plate test method [9,14–17,21–23]. Then
 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489 481

Eq. (3) was solved for three contact angles measured with three where WAS = the work of adhesion between an asphalt binder (sub-
solvents [24] to determine the three unknown SFE components script A) and aggregate or stone (subscript S). Therefore, the higher
of the asphalt binder. the bond strength between the asphalt binder and aggregate in dry
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi condition, the higher the WAS [12].
CL ð1 þ cos hÞ ¼ 2 CLW LW
A CL þ CþA CL þ CA CþL ð3Þ
3.2.3. Work of debonding
where A and L subscripts represent the energy parameters associ- The energy released as a result of spontaneous separation of the
ated with asphalt binder and probe liquid, respectively. A descrip- asphalt binder from aggregate surface in the presence of water is
tion of the method used for conducting the Wilhelmy plate test is called the work of debonding. The work of debonding is deter-
provided, later in this paper. mined from Eq. (8).

3.1.2. Surface free energy of aggregates W wet

ASW ¼ CAW þ CSW  CAS ð8Þ
Adsorption isotherms developed using three different probe where W wet
= the work of debonding of asphalt binder from aggre-
ASW
vapors are used to determine the SFE components of aggregates gate in presence of water; CAW = interfacial energy between asphalt
using a USD, applying a method suggested by Bhasin et al. [12]. binder and water (subscript W); CSW = interfacial energy between
For this purpose Eqs. (4) and (5) were used to calculate the spread- aggregate and water; and CAS = interfacial energy between asphalt
ing pressure and the work of adhesion between probe vapor and binder and aggregate. Interfacial energy is defined as the energy
aggregates, respectively. equal to the surface tension at an interface. Interfacial energy
Z pn
RT n between two materials is determined from Eq. (9) [12].
pe ¼ dp ð4Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi
MA p
0
Cij ¼ Ci þ Cj  2 CLW LW
i Cj  2 Cþi Cj  2 Ci Cþj ð9Þ

W SV ¼ pe þ 2CTotal
V ð5Þ where Cij = interfacial energy between materials i and j; Ci and Cj =
total surface free energies of materials i and j, respectively. Since the
where pe = equilibrium spreading pressure of the probe vapor on
spontaneous debonding of the asphalt binder from aggregate due to
aggregate surface; R = universal gas constant; T = test temperature;
the presence of water releases energy, the W wet ASW is a negative value.
M = molecular weight of probe vapor; n = adsorbed mass of probe
This will cause the total energy level of the system to reduce, a ther-
vapor per unit mass of the aggregate at probe vapor pressure of p;
modynamically favorable mechanism. It can be concluded that the
A = specific surface area of aggregate; WSV = work of adhesion
greater the jW wet
ASW j, the higher the debonding potential of the asphalt
between aggregate surface and probe vapor (subscript V); and
binder from the aggregate, in presence of moisture [12].
CTotal
V = total surface free energy of probe vapor. After calculation
of WSV for each probe vapor, by Eq. (5), the SFE components of the
3.2.4. Energy ratio parameters
aggregate are determined by solving the adhesion equation the
To this end, based on the wettability, the work of adhesion and
from adsorption isotherm results for three solvents. A detailed dis-
the work of debonding determined from Eqs. (6)–(8), respectively,
cussion on the measurement of the SFE components of aggregates,
it is evident that stripping of the asphalt binder from the aggregate
using a USD, is presented subsequently in this paper.
depends on these parameters. The resistance to moisture-induced
damage increases with an increase in the wettability (SA/S) and
3.2. Performance parameters estimated using SFE
work of adhesion (WAS), and decreases with an increase in the mag-
nitude of the work of debonding (jW wet ASW j). Therefore, Bhasin et al.
3.2.1. Wettability
[12] suggested the use of WAS and jW wet ASW j into a single parameter,
Wetting is a liquid’s spreading and maintaining its contact over
namely energy ratio (ER1), shown in Eq. (10).
a surface, which may include penetration into porous medium
[25]. Therefore, the potential of an asphalt binder to coat an aggre- W AS
ER1 ¼ j j ð10Þ
gate can be studied by evaluation of wettability. Wettability is W wet
ASW
determined by the liquid contact angle measured on the surface From Eq. (10), it is evident that ER1 increases with an increase in
of the solid phase. The contact angle is formed based on a balance WAS and decreases with an increase in jW wet ASW j. However, Eq. (10)
between the adhesive and cohesive forces [25]. The spreading coef- does not include the role of wettability, as a measure of coating
ficient, as a quantitative measure of wettability, is defined as the quality and forming bond between aggregate and asphalt binder.
reduction in SFE during losing the bare solid surface and forming Therefore, a second parameter, ER2, which considers the wettability
a new solid–liquid and liquid–vapor interface [26]. The spreading parameter, was suggested by Bhasin et al. [12], as shown in Eq. (11).
coefficient is given in Eq. (6) [17].
SA=S
SA=S ¼ CS  CAS  CA ð6Þ ER2 ¼ j j ð11Þ
W wet
ASW
where SA/S = spreading coefficient of asphalt binder (subscript A)
The wettability parameter, namely spreading coefficient (SA/S)
over the aggregate or stone (subscript S); CS = total surface free
can be estimated from Eq. (6). The ER1 and ER2 were used in this
energy of aggregate, CAS = interfacial energy between asphalt bin-
study to evaluate moisture-induced damage potential of different
der and aggregate, and CA = total surface free energy of asphalt
mixes containing RAP with varieties of aggregates and asphalt
binder.
binders.

3.2.2. Work of adhesion

The free energy required to create two interfaces from one 4. Materials
interface, consisting of two different phases in contact, is defined
as the work of adhesion. The work of adhesion between an asphalt 4.1. Asphalt binders and aggregates
binder and aggregate can be determined from Eq. (7).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffi The PG 64-22 and PG 76-28 asphalt binders, used in this study,
W AS ¼ 2 CLWA CS þ 2
LW
CþA CS þ 2 CA CþS ð7Þ were obtained from the Valero refinery located in Muskogee, OK.
Both of these asphalt binders are commonly used in Oklahoma
482 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489

for construction of pavement. Different types of aggregates, tested (ii) 10% RAP binder +90% virgin binder, (iii) 25% RAP binder +75% vir-
in this study, were collected from different quarries in Oklahoma. gin binder, and (iv) 40% RAP binder +60% virgin binder. The selection
The collected aggregates are among the common aggregates used of these proportions was based on the RAP contents commonly used
in Oklahoma for production of asphalt mixes. in different types of pavements in Oklahoma. Thus, a total of 8 differ-
ent asphalt binders and RAP combinations (four for PG 64-22 and
4.2. RAP binder four for PG 76-28), were prepared. A summary of the asphalt binder
mixes, prepared for this study is shown in Table 1.
The chemicals used in binder extraction methods from RAP may
significantly alter the asphalt binders’ chemical and surface prop-
erties and may introduce error in measured SFE parameters of 5.2. Measurement of surface free energy components of asphalt
the extracted RAP binder. Therefore, in order to produce RAP bin- binders
der, the Rolling Thin-Film Oven (RTFO) and Pressure Aging Vessel
(PAV) methods, in accordance with AASHTO T 240 [27] and The SFE components of each asphalt binder were determined by
AASHTO R 28 [28] were used, respectively, to long-term age the measuring its contact angles with different solvents, using the
PG 64-22 binder. The binder prepared using this method was DCA. For this purpose, three different solvents with known SFE
stored in small canisters for further testing. components were used, namely water (bi-polar solvent), glycerin
The asphalt binder aged according to this procedure represents (apolar solvent) and formamide (mono-polar solvent). To prepare
a long-term aging equivalent to seven to ten years of in-service DCA samples, standard cover glasses having 25 mm width by
aging [29]. This method is used by many researchers to produce 50 mm length were coated with asphalt binder. For this purpose,
simulated RAP binder in the laboratory. It has been reported that approximately 100 g of asphalt binder mix in a canister was kept
the PAV method can simulate both chemical and physical changes in an oven at 165 °C, for 2 h to liquefy the binder and gently mixed
of asphalt binders during its service life [30]. Since PG 64-22 bind- couple of times to ensure the consistency and proper mixing of RAP
ers are used in a majority of pavements in Oklahoma, a PG 64-22 binder and virgin binder. Then the glass plate surface was ‘‘flamed’’
binder was selected for aging and was used for RAP binder by passing it through an oxygen flame for at least three times in
production. less than 5 s, in order to obtain a moisture-free and clean surface.
Thereafter, the glass plate was carefully dipped at least 20 mm in
the asphalt binder mix in the oven and moved back and forth three
5. Methodology times in 5 s, to ensure proper asphalt coating on the glass plate.
The coated plate was then kept in the oven on a stand in vertical
A summary of the work flow and techniques used for evaluation position for 2 min to drip down the excess binder and to obtain a
of the moisture-induced damage potential of asphalt binders smooth surface. Finally, the sample was cured for 24 h in a desic-
mixed with RAP and aggregates is presented in Fig. 1. cator, before conducting the test. Table 1 presents the test matrix
for the DCA tests conducted on asphalt binder mixes, in this study.
5.1. Preparation of asphalt binder for SFE testing As evident from Table 1, a total of 120 (2 binders  4 RAP percent-
ages  5 replicates  3 solvents) asphalt binder samples were
The asphalt binder mixes, used for the SFE testing, consisted of tested using the DCA analyzer. After measuring the contact angles,
two sets. The following proportions (by weight) were used for each the SFE components of asphalt binder mixes were determined by
type of virgin binder (PG 64-28 and PG 76-28): (i) 100% virgin binder, using Eq. (3) for each asphalt binder and solvent.

Fig. 1. Work flow for evaluation of moisture damage potential using SFE method.
 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489 483

Table 1
Matrix of the dynamic Wilhelmy plate tests conducted on binder mixes.

Asphalt binder mix No. of samples tested with each solvent Total samples
Virgin binder type and amount (%) RAP binder amount (%) Water Glycerin Formamide
PG 64-22 100 0 5 5 5 15
90 10 5 5 5 15
75 25 5 5 5 15
60 40 5 5 5 15
PG 76-28 100 0 5 5 5 15
90 10 5 5 5 15
75 25 5 5 5 15
60 40 5 5 5 15
Total asphalt binder samples tested using DWP test 120

5.3. Measurement of surface free energy components of aggregates 6. Results and discussion

The SFE components of different aggregates, collected from 6.1. SFE components of asphalt binders
quarries in Oklahoma, were determined by USD testing. Further-
more, the SFE components of a number of other aggregates were The SFE components of asphalt binder play an important role in
adopted from the literature [31,32]. Surface free energies of aggre- defining its ability to adhere to aggregates. The SFE components of
gates were determined from vapor sorption isotherms, i.e. the asphalt binders, along with those obtained from aggregates are
amount of vapor adsorbed or desorbed on the solid surface at a needed for determination of the energy ratio parameters. These
fixed temperature and partial pressure [31]. These aggregates con- parameters are used to evaluate the moisture-induced damage
sisted of limestone and rhyolite tested in this study, and sandstone, potential of the asphalt binder–aggregate systems. The SFE compo-
granite, gravel, and basalt, adopted from the literature [31,32]. The nents of PG 64-22 and PG 76-28 asphalt binders mixed with differ-
methodology used by Bhasin and Little [31] was applied for deter- ent amounts of RAP binder are presented in Fig. 2a and b,
mination of the SFE components. In this study, three probe vapors, respectively. The non-polar SFE component (CLW) values of PG
namely, water (bi-polar vapor), methyl propyl ketone or MPK 76-28 binder were found to be approximately 2 mJ/m2 higher than
(mono-polar vapor) and n-hexane (apolar vapor) were used to those of PG 64-22 binder for different percentages of RAP binder
determine adsorption isotherms. For aggregate sample prepara- (Fig. 2). It is important to note that the PG 76-28 is a polymer-mod-
tion, selected aggregates were oven dried at 60 °C for 24 h and ified asphalt binder; higher CLW values for this type of asphalt bin-
cooled down to room temperature. Then, they were crushed in lab- der may be attributed to polymer modification. The CLW component
oratory and the size fraction with particle larger than 2.36 mm of both PG 64-22 and PG 76-28 asphalt binders reduced with the
(retaining on a No. 8 sieve) and smaller than 4.75 mm (passing addition of 10% and 25% RAP binder. This component increased
to a No. 4 sieve) were selected. Thereafter, the selected fraction with addition of higher amounts of RAP (i.e., 40%). Nonpolar mole-
of aggregates was washed several times with distilled water to cules in an asphalt binder are known to work as a matrix for polar
obtain dust-free clean surfaces. Then, they were oven-dried at molecules. This matrix arrangement is responsible for asphalt’s
120 °C for 12 h and allowed to cool to room temperature in a des- elastic properties of asphalt binders [33]. Thus, an increase in the
iccator sealed with silica gel, before testing. About 20 g of the pre- CLW component may result in an increase in work of adhesion,
pared aggregate sample was subjected to probe vapors (water, which is an indicator of a better binder–aggregate bond in dry
MPK, and n-hexane), in a USD. The recorded changes in sample condition.
weight at different relative humidity or pressures were measured Furthermore, an increase in RAP content, increased the acid SFE
using a built-in Cahn D-101 microbalance. A data acquisition component (C+) and, in general, decreased the base SFE component
system recorded the sample weight, temperature, and relative (C) of both PG 64-22 and PG 76-28 asphalt binders (Fig. 2). This
humidity or pressures in a data file, at user-defined intervals. The increase in C+ and reduction in C values with increasing RAP
collected data was then used for developing the sorption isotherms amounts are more pronounced for the PG 76-28 binder. Highly
for each aggregate tested with each probe vapor. Then, Eqs. (4), (5) acidic binders are known not to produce a strong bond with acidic
and (7) in conjunction with the developed sorption isotherms were aggregates. This is due to surface chemistry of Lewis acid and base
used to determine the SFE components of each aggregate [31]. which does not favor adhesion in this case [23]. It is known that sig-
Table 2 presents the test matrix of the USD tests conducted on nificantly high polar components (C+ and C) in asphalt binder may
aggregates tested in this study and those adopted from the litera- result in moisture-induced damage potential, fatigue cracking in
ture. According to Table 2, a total of 18 aggregate (2 aggregates  3 thick pavement layers and rutting [33]. However, the moisture sus-
replicates  3 solvents) samples were tested in the USD device. ceptibility of a mix should evaluated based on parameters which

Table 2
Matrix of the sorption tests conducted on aggregates and those from literature.

Type of aggregate Aggregate code No. of samples tested with each solvent Total samples
Water Glycerin Formamide
Limestone LS 3 3 3 9
Rhyolite RH 3 3 3 9
Sandstone SS Adopted from Bhasin and Little [31]
Granite GN Adopted from Buddhala et al. [32]
Gravel GV Adopted from Buddhala et al. [32]
Basalt BS Adopted from Buddhala et al. [32]
Total aggregate samples tested using USD 18
484 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489

higher wettability of asphalt binder over aggregate surface will

help the asphalt binder to easily coat the aggregate. The wettability
can be expressed by the released energy as the asphalt binder
flows over the aggregate to coat it [16,26]. The released energy,
namely, spreading coefficient of the asphalt binder over the aggre-
gates (SA/S), can be used to quantify the wettability. In this study,
the spreading coefficients of PG 64-22 and PG 76-28 binders mixed
with different amounts of RAP binder (0%, 10%, 25% and 40%) over
different types of aggregates (limestone, rhyolite, sandstone, gran-
ite, gravel and basalt) were determined using Eq. (6).

6.3.1. Effect of asphalt binder type and RAP content on wettability

The spreading coefficients of PG 64-22 and PG 76-28 asphalt
binders with different percentages of RAP binder and aggregates
are presented in Fig. 3. It can be seen from Fig. 3 that the spreading
coefficient of PG 64-22 and PG 76-28 asphalt binders with different
Fig. 2. SFE components of (a) PG 64-22 and (b) PG 76-28 binders mixed with RAP types of aggregates did not change significantly up to 10% RAP con-
binder. tent. However, the spreading coefficients for both PG 64-22 and PG
76-28 asphalt binders increased at higher RAP contents (25% and
include wettability, adhesion and debonding properties of an 40%), for all types of aggregates, except for granite. Therefore, use
asphalt binder–aggregate system, which is introduced in this paper. of higher amounts of RAP will be beneficial for a better aggregate
coating with binder, contributing to a better bond. Furthermore,
6.2. SFE components of aggregates it was observed that the PG 64-22 asphalt binder without RAP
and 10% RAP had higher spreading coefficients than those of the
The SFE components of the aggregates tested (limestone and rhy- PG 76-28 asphalt binder for almost all types of aggregates. For
olite) and those adopted from literature (sandstone, granite, gravel, example, spreading coefficients of the PG 64-22 asphalt binder
and basalt) are presented in Table 3. The SFE components of the with 0% and 10% RAP was found to be 104.4 and 104.7 mJ/m2 on
aggregates, tested herein (limestone and rhyolite), were found to gravel aggregate, while they were 92.9 and 92.1 mJ/m2 for the PG
be in the range of those reported in the literature [9]. The gravel 76-28 asphalt binder (approximately 11% reduction). This means
and basalt aggregates have the highest and the lowest total SFE com- that, when mixed with low amounts of RAP (0% and 10%), a PG
ponents (356.7 and 72.1 mJ/m2), respectively. Comparatively, gravel 64-22 binder may have a higher tendency to coat the aggregates
and granite were found to have the highest acidic SFE components than a PG 76-28 binder. However, when the amount of RAP
(24.1 and 23.0 mJ/m2, respectively), among the other aggregates. A increased (25% and 40%), both PG 64-22 and PG 76-28 binders
comparison between the asphalt binder and aggregate SFE compo- showed similar values of the spreading coefficients with different
nents (Table 3 and Fig. 2) reveals that the aggregates have relatively aggregates, as shown in Fig. 3. Thus, it can be concluded that the
higher acid, base and acid-base components, compared to those of wettability improves and becomes independent of binder type at
asphalt binders. For example, the acid-base SFE components of PG higher RAP contents (25% and 40%).
64-22 and PG 76-28 binders, mixed with different amounts of RAP,
are in the range of 1.09–1.84 mJ/m2. These ranges are considerably 6.4. Work of adhesion
lower those measured for aggregates (19.8–299.2 mJ/m2). There-
fore, water, with a polar molecule, has a higher wetting potential The work of adhesion (WAS) under dry condition (without effect
on aggregates than that of asphalt binders [34]. of moisture) was determined to evaluate the bond strength
between aggregate and asphalt binder [16]. The work of adhesion
6.3. Wettability is the energy required for separation of an asphalt binder from the
aggregate–binder interface [12]. Therefore, a higher WAS value is
The wettability is defined as the tendency of a liquid to wet a desirable for a stronger bond between asphalt binder and aggre-
solid surface. Asphalt binder and aggregates are hydrophobic and gate under dry condition. However, the work of adhesion alone
hydrophilic materials, respectively [34]. Hence, wetting a cannot rank asphalt mixes based on their moisture-induced dam-
hydrophilic surface (aggregate) with a hydrophobic material age potentials. This parameter is required to estimate the energy
(asphalt binder) is not easy. Therefore, it is important to study ratio parameters, by which the moisture susceptibility is
the wettability of asphalt binder over the aggregate surface. A evaluated.

Table 3
SFE components of aggregates.

Type of aggregate Aggregate code Aggregate surface free energy components (mJ/m2)
CLW (non-polar) C (base) C+ (acid) CAB (acid-base) Ctotal
Limestone LS 51.4 741.4 17.5 227.9 279.3
Rhyolite RH 48.9 877.9 7.5 161.9 210.8
Sandstonea SS 58.3 855.0 14.6 223.5 281.8
Graniteb GN 133.2 96.0 24.1 96.2 229.4
Gravelb GV 57.5 973.0 23.0 299.2 356.7
Basaltb BS 52.3 164.0 0.6 19.8 72.1
1
Adopted from Bhasin and Little [31].
2
Adopted from Buddhala et al. [32].
 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489 485

Fig. 3. Spreading coefficients of different aggregates and (a) PG 64-22 and (b) PG Fig. 4. Work of adhesion of different aggregates and (a) PG 64-22 and (b) PG 76-28
76-28 asphalt binders mixed with different RAP amounts. asphalt binders mixed with different RAP amounts.

highest and the lowest work of adhesion with both binder types
6.4.1. Effect of asphalt binder type and RAP content on work of (PG 64-22 and PG 76-28, respectively. It means that the gravel and
adhesion basalt aggregates have the highest and the lowest dry bond strength
Fig. 4a and b presents the work of adhesion between the aggre- by asphalt binder, respectively. Furthermore, it was observed that
gates and the PG 64-22 and PG 76-28 asphalt binders, respectively, using RAP with PG 64-22 and PG 76-28 binders on the granite aggre-
with different amounts of RAP. In general, it can be seen that the gate may not significantly improve the work of adhesion (Fig. 4).
work of adhesion for the PG 64-22 and PG 76-28 asphalt binders
increases significantly at higher RAP contents (25% and 40%), for
6.5. Work of debonding
all types of aggregates, except for granite, when the PG 64-22 binder
was used. However, the increase in the work of adhesion was not sig-
The work of debonding (W wet ASW ) is a measure of aggregate–
nificant at a lower percentage of RAP (i.e., 10%). Therefore, it can be
asphalt binder separation potential in presence of moisture. As a
concluded that addition of low amounts of RAP (10%) may not affect
result of stripping, a reduction in the free energy of the system
the work of adhesion, while use of higher RAP amounts (25% and
occurs when asphalt binder, in presence of water, separates from
more) was found to be beneficial to improve aggregate–asphalt bin-
the aggregate–asphalt binder interface. This reduction in free
der adhesion. These observations are consistent with the results
energy of the system is called the work of debonding and is deter-
reported based on the wettability of asphalt binders for different
mined by using Eq. (8). Therefore, a higher magnitude of the work
amounts of RAP binder, discussed in the previous section.
of debonding (negative number) implies that the occurrence of
While comparing work of adhesion for both the asphalt binders,
stripping is thermodynamically more favorable [31]. Therefore, it
it was found that the PG 64-22 asphalt binder without RAP and 10%
is essential to evaluate the work of debonding for different combi-
RAP showed a higher work of adhesion than those of PG 76-28
nations of virgin asphalt binder, RAP binder, and aggregates to fully
asphalt binder, for all types of aggregates, except for granite. For
characterize the moisture damage mechanism. However, it should
example, the work of adhesion for the PG 64-22 asphalt binder for
be noted that the work of debonding alone cannot rank moisture
gravel aggregates, with 0% and 10% RAP was found to be 128.5 and
damage potentials of asphalt mixture. This parameter is required
128.4 mJ/m2, while it was found to be 121.0 and 118.0 mJ/m2 for
for estimation of energy ratio parameter, which is a crucial to
PG 76-28 asphalt binder on the same aggregate (approximately 6%
evaluate the moisture damage potential.
and 8% reduction, respectively). This means that at low amounts of
RAP (0% and 10%), a PG 64-22 binder has higher bonding strength
with aggregates than that of the PG 76-28 binder. A similar trend 6.5.1. Effect of asphalt binder type and RAP content on work of
was found for wettability as well. However, it was observed that, debonding
with further increasing in RAP amounts (25% and 40%), both the Fig. 5a and b presents the work of debonding between the
PG 64-22 and PG 76-28 binders showed close work of adhesion val- aggregates and the PG 64-22 and PG 76-28 asphalt binders, respec-
ues with different aggregates, indicating that adhesion improves at tively, with different amounts of RAP. It is evident that the magni-
higher RAP contents (25% and 40%), regardless of the type of virgin tude of the work of debonding for both PG 64-22 and PG 76-28
asphalt binder used. The gravel and basalt aggregates had the asphalt binders with different aggregates (except for granite)
486 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489

decrease with an increase in the amount of RAP binder. The reduc- for each binder type (PG 64-22 and PG 76-28). Therefore, based
tion in the work of debonding is more pronounced at higher RAP on the work of debonding values, it can be concluded that when
amounts (25% and 40%), while it is not as significant as that for using granite and basalt aggregates, it is not expected to gain
lower RAP content (10%). Therefore, it can be concluded that the benefits from using RAP in terms of reducing the work of debond-
use of higher RAP amounts (25% and more) may lead to less strip- ing. However, using gravel, sandstone, rhyolite and limestone
ping. Furthermore, according Fig. 5a and b, comparing the work of aggregates was found to maximize the reduction in the work of
debonding of the virgin PG 64-22 and PG 76-28 asphalt binders (0% debonding, as a result of adding RAP. It is important to note that
RAP) reveals that the PG 64-22 has a lower magnitude of the work this parameter (work of debonding) should be considered in con-
of debonding for all types of aggregates than that of PG 76-28 junction with the wettability and work of adhesion, to evaluate
asphalt binder. For example, the work of debonding of rhyolite moisture-induced damage potential, as discussed next.
aggregate with virgin PG 64-22 asphalt binder was 176.2 mJ/
m2, which is approximately 7% lower than that of PG 76-28 6.6. Energy ratio parameters
(187.6 mJ/m2). A similar trend also exists for other types of aggre-
gates. As discussed earlier, the PG 76-28 is a polymer-modified 6.6.1. Effect of asphalt binder type and RAP content on ER1 and ER2
binder; and the observations based on the work of adhesion and The ER1 and ER2 values were determined for different combina-
the work of debonding, suggest moisture susceptibility concerns tions of the asphalt binder types (PG 64-22 and PG 76-28), RAP
for this polymer-modified asphalt binder when it is compared with contents (0%, 10%, 25%, and 40%), and different aggregate types
non-polymer-modified type. Therefore, more study on the effect of (limestone, rhyolite, sandstone, granite, gravel, and basalt) and
using polymer-modified asphalt binder in asphalt mix on its mois- are presented in Figs. 6 and 7. It is evident from Figs. 6 and 7 that,
ture susceptibility is recommended. for both PG 64-22 and PG 76-28 asphalt binders, the ER1 and ER2
Furthermore, it is evident that the work of debonding changes values increase with an increase in RAP content for all types of
significantly with the change in aggregate type for both PG 64-22 aggregates. The addition of low amounts of RAP (10%) does not
and PG 76-28 asphalt binders (Fig. 5). However, the trends of var- seem to influence the ER1 and ER2 values, while addition of higher
iation in the work of debonding with aggregate type and RAP con- amounts (25% and 40%) of RAP significantly improved the ER1 and
tent when the PG 64-22 and PG 76-28 asphalt binders were used, ER2 values for both types of asphalt binders (Figs. 6 and 7). It means
were found to be similar. Furthermore, it was observed that the that, in general, addition of higher percentages of RAP improves the
work of debonding for granite aggregate with PG 64-22 and PG resistance to moisture-induced damage for both PG 64-22 and PG
76-28 asphalt binders remain unchanged at different RAP contents. 76-28 asphalt binders.
In a similar way, basalt aggregate shows very low reduction in Furthermore, it was observed that the PG 64-22 asphalt binder
work of debonding with an increase in RAP content, when PG 64- without RAP and 10% RAP generally showed higher ER1 and ER2 val-
22 and PG 76-28 asphalt binders were used. However, the work ues than those obtained for the PG 76-28 asphalt binder for all types
of debonding found for gravel, sandstone, rhyolite and limestone of aggregates, except for granite. For example, the ER1 values of a PG
aggregates show similar sensitivities to change in RAP content 64-22 asphalt binder and basalt aggregate with 0% and 10% RAP
were found to be 2.70 and 2.78, respectively, while they were 2.27

Fig. 5. Work of debonding of different aggregates and (a) PG 64-22 and (b) PG 76- Fig. 6. ER1 values determined for different aggregates and (a) PG 64-22 and (b) PG
28 asphalt binders mixed with different RAP amounts. 76-28 asphalt binders mixed with different RAP amounts.
 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489 487

for granite aggregate range between 1.91 and 1.97 with the PG 64-
22 binder and between 1.96 and 2.25, for the PG 76-28 binder. A
similar trend in variation was observed for ER2. From the above dis-
cussion, it may be concluded that, ER1 and ER2 exhibited similar
and consistent trend of variation with change in asphalt binder
type, RAP content and aggregate type. Therefore, both of these
parameters may be used for the evaluation of moisture-induced
damage potential of mixes. Based on the ER1 and ER2 values, the
aggregates combined with PG 64-22 and PG 76-28 asphalt binders
used in this study were ranked based on their resistance to mois-
ture-induced damage, from the highest to the lowest, as: basalt,
granite, rhyolite, limestone, sandstone, and gravel. It is interesting
to note that the following order was found when the aggregates
were ranked with respect to their total SFE from the lowest to
the highest, as: basalt, rhyolite, granite, limestone, sandstone,
and gravel. This ranking is almost the same as that found from
ER1 and ER2 values. This is expected since the non-polar compo-
nents of the aggregates are very low and the most effective SFE
component contributing to the total SFE of aggregates is the acid
and base components. Therefore, an increase in acid and base com-
ponents of the aggregates will lead to a high total SFE. As discussed
earlier, high acid and base component in aggregates are known to
increase moisture-induced damage potential. However, use of ER1
and ER2 will produce more accurate evaluation of moisture damage
because of considering the effects of both aggregate and asphalt
binder. The ranking provided herein is expected to be helpful for
the pavement engineers in selecting aggregates for asphalt mixes
containing RAP, so as to minimize the moisture-induced damage
potential of the resulting mix.
Fig. 7. ER2 values determined for different aggregates and (a) PG 64-22 and (b) PG
76-28 asphalt binders mixed with different RAP amounts.
7. Conclusions and recommendations
and 2.28 mJ/m2, respectively, for the PG 76-28 asphalt binder
(approximately 16% and 18% reduction with respect to the PG 64- The present study evaluated the effect of two types of asphalt
22 binder). Similarly, the ER2 values of a PG 64-22 asphalt binder binders, namely PG 64-22 and PG 76-28, four RAP contents,
and basalt aggregate with 0% and 10% RAP were found to be 1.87 namely, 0%, 10%, 25% and 40%, and six different types of aggregates,
and 1.94, while they were found to be 1.43 and 1.47 mJ/m2, respec- namely limestone, rhyolite, sandstone, granite, gravel, and basalt,
tively, for the PG 76-28 asphalt binder (approximately 25% and 24% on the wettability and moisture-induced damage potential of asso-
reduction). This means that at low amounts of RAP (0% and 10%), the ciated asphalt mixes, applying the surface free energy (SFE)
PG 64-22 binder exhibits higher ER1 and ER2 values for different approach. For this purpose, the SFE components of the abovemen-
types of aggregates than those of the PG 76-28 binders. This can tioned asphalt binders and aggregates were used to estimate the
be attributed to the effect of using a polymer-modified asphalt bin- wettability, work of adhesion and work of debonding for different
der (PG 76-28), which increases the moisture induced damage combinations of asphalt binders and aggregates. Thereafter, two
potential, compared to non-modified asphalt binder (PG 64-22). different energy ratio parameters, namely ER1 and ER2, calculated
However, it was observed that with further increasing the RAP based on wettability, work of adhesion and work of debonding
amounts (25% and 40%), both PG 64-22 and PG 76-28 asphalt bind- were used to evaluate moisture-induced damage potential of the
ers with different aggregates, showed similar ER1 and ER2 values. mixes. The following conclusions may be drawn from the results
Thus, at higher RAP contents type of binder does not have any signif- and discussions presented in the preceding sections.
icant effect on the moisture-induced damage potential.
1. The non-polar SFE component of PG 76-28 asphalt binder
was found to be higher than that of PG 64-22, for all RAP con-
6.6.2. Effect of aggregate type on ER1 and ER1
tents. This SFE component increases with addition of RAP in
Figs. 6 and 7 show that the ER1 and ER2 values highly depend on
higher amounts (i.e., 25% and 40%) for both PG 64-22 and PG
the aggregate type. It was observed that the basalt aggregate exhib-
76-28 asphalt binders. The acid SFE component (polar com-
ited the highest ER1 and ER2 values, ranging from 2.27 to 3.31, and
ponent) of both PG 64-22 and PG 76-28 asphalt binders were
1.43 to 2.27, respectively, for different RAP amounts for both PG
found to be higher than their base SFE component, and
64-22 and PG 76-28 asphalt binders. This is due to very low work
increased with an increase in RAP content; but, the range
of debonding found for basalt, compared to its work of adhesion
of variations for acid SFE component was similar for both
and wettability. Comparatively, the gravel aggregate showed the
types of asphalt binders. However, the base SFE component
lowest ER1 and ER2 values ranging from 0.55 to 0.73 and 0.41 to
of both PG 64-22 and PG 76-28 asphalt binders did not
0.59, respectively, for different RAP amounts and asphalt binders.
change significantly, with increasing RAP content.
This shows that the mixes with different amounts of RAP, in which
2. Based on the wettability parameter estimated for different
the basalt and gravel aggregates are used, have the highest and the
combinations of asphalt binder type, RAP contents, and
lowest resistance to moisture-induced damage, respectively.
aggregates, it was found that the coating quality of both PG
It is worth noting that the ER1 and ER2 values were less sensitive
64-22 and PG 76-22 asphalt binders over the different types
to the change in RAP amount for granite aggregates with both PG
of aggregates increased with an increase in RAP content (25%
64-22 and PG 76-28 asphalt binders. For examples, the ER1 values
488 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489

and more). The gravel and basalt aggregates showed the 2. Additional studies on the effect of polymer-modified asphalt
highest and the lowest coating qualities among the tested binders with different PG-plus grades on moisture-induced
aggregates, respectively. damage potential, using the SFE method, is recommended.
3. The bond strength between aggregates and asphalt binder 3. Performance tests on the asphalt mixes, using the aggregates
systems under dry condition was evaluated based on the and asphalt binders tested herein, are recommended to
work of adhesion. It was found that the work of adhesion of cross-check the results from the SFE method with those
the PG 64-22 and PG 76-22 asphalt binders with different obtained from laboratory mix testing.
types of aggregates, increases with an increase in RAP content
(25% and more). The improvement in the work of adhesion
(under dry condition) was found very low for the granite Acknowledgments
aggregate. The gravel and basalt aggregates showed the high-
est and the lowest work of adhesion (in dry condition) among The authors would like to acknowledge the financial support
the tested aggregates, respectively. This finding is consistent from the Oklahoma Transportation Center (OkTC) and the
with the results obtained for wettability. Use of PG 64-22 Oklahoma Department of Transportation. Assistance of Dr. Edgar
asphalt binder resulted in a higher work of adhesion (under O’Rear, Dr. Nazimuddin Wasiuddin, Ms. Aravinda Buddhala, Mr.
dry condition) with different aggregates and RAP amounts. Jackson Autrey and Mr. Michael Hendrick is highly appreciated.
A higher work of adhesion is expected to improve the aggre- Also, assistance received from Schwarz Paving in the collection of
gate–asphalt binder bond strength, under dry condition. materials is gratefully acknowledged.
4. The debonding potential of aggregate from asphalt binder
under wet condition was evaluated based on the work of References
debonding. It was found that the work of debonding for both
PG 64-22 and PG 76-22 asphalt binders with different types [1] Northeast center of excellence for pavement technology (NECEPT). Superpave
system; 2010. <http://www.superpave.psu.edu/superpave/system.html>
of aggregates decreased with an increase in RAP content (accessed 21.02.14).
(25% and more). The reduction in the work of debonding [2] National asphalt pavement association (NAPA). Recycling; 2011. <https://
with increasing RAP content was found to be insignificant www.asphaltpavement.org/index.php?option=com_content&task=view&id=25
&Itemid=45> (accessed 21.02.14).
when granite and basalt aggregates were used. Gravel and [3] Al-Qadi IL, Aurangzeb Q, Carpenter SH, Pine WJ, Trepanier J. Impact of high RAP
basalt aggregates showed the highest and the lowest work content on structural and performance properties of asphalt mixtures.
of debonding among the other aggregates, respectively. A Research report FHWA-ICT-12-002 illinois center for transportation; 2012. p.
53.
higher work of debonding is expected to increase the separa-
[4] Huang B, Zhang Z, Kinger W. Fatigue crack characteristics of HMA mixtures
tion potential of asphalt binder from aggregate, in presence containing RAP. In: Proceedings, 5th international RILEM conference on
of water. However, it should be discussed in conjunction cracking in pavements, Limoges, France; 2004.
with the wettability and work of adhesion to make a sound [5] Mohammad LN, Negulescu II, Wu Z, Daranga C, Daly WH, Abadie C.
Investigation of the use of recycled polymer modified asphalt binder in
judgment on stripping potential. asphalt concrete pavements (with discussion and closure). J Assoc Asphalt
5. Overall, the energy ratio parameters (ER1 and ER2), as mech- Paving Technol 2003;72:551–94.
anistic indicators of resistance of the asphalt mix to mois- [6] McDaniel R, Shah A. Asphalt additives to control rutting and cracking.
Publication FHWA/IN/JTRP-2002/29, Joint Transportation Research Program.
ture-induced damage, consistently increased with an West Lafayette, Indiana: Indiana Department of Transportation and Purdue
increase in RAP content for both PG 64-22 and PG 76-28 University; 2003.
asphalt binders and all types of aggregates. Based on the [7] McDaniel RS, Soleymani H, Anderson RM, Turner P, Peterson R. Recommended
use of reclaimed asphalt pavement in the superpave mixture design method.
ER1 and ER2 values, use of polymer-modified asphalt binder NCHRP final report (9–12). Washington, DC: Transportation Research Board;
(PG 76-28) was found to increase the moisture-induced 2000.
damage potential at lower RAP contents (0% and 10%) com- [8] Shu X, Huang B, Vukosavljevic D. Laboratory evaluation of fatigue
characteristics of recycled asphalt mixture. J Construct Build Mater
pared to non-polymer-modified asphalt binder (PG 64-22). 2008;22:1323–30.
At higher RAP contents (25% and 40%), the improvement in [9] Howson J, Bhasin A, Masad E, Lytton R, Little D. Development of a database for
resistance to moisture-induced damage was found to be surface energy of aggregates and asphalt binder. Report No. FHWA/TX-5-4524-
01-1. Texas A&M University; 2009.
similar for both types of asphalt binders (PG 64-22 and PG
[10] AASHTO. Standard method of test for resistance of compacted asphalt
76-28). For different amounts of RAP and different asphalt mixtures to moisture-induced damage. T 283–07, Washington, DC:
binder types (PG 64-22 and PG 76-28), basalt and gravel American Association of State and Highway Transportation Officials; 2007.
aggregates showed the highest and the lowest resistance to [11] AASHTO. Standard method of test for hamburg wheel-track testing of
compacted hot-mix asphalt (HMA). T 324–11. Washington, DC: American
moisture-induced damage, respectively. Association of State and Highway Transportation Officials; 2011.
6. It was found that the higher the total SFE of the aggregates, [12] Bhasin A, Little DN, Vasconcelos KL, Masad E. Surface free energy to identify
the lower the ER1 and ER2 values. Therefore, a high total moisture sensitivity of materials for asphalt mixes. Transp Res Rec
2007;2001:37–45.
SFE component of aggregate may result in a high moisture- [13] Ghabchi R, Zaman M, Bulut R, Koc M, Singh D. WMA pavements in Oklahoma:
induced damage potential in the mix. moisture damage and performance issues. Final report. Oklahoma
7. Based on the outcomes of this study, the recommended prac- Transportation Center (OkTC), 2601 Liberty Parkway, Suite 110, Midwest
City, Oklahoma 73110; 2013.
tice for evaluation of the moisture-induced damage potential [14] Arabani M, Roshani H, Hamedi GhH. Estimating moisture sensitivity of warm
of asphalt mixes is a combined use of SFE approach and mix asphalt modified with zycosoil as an antistrip agent using surface free
traditional testing methods (e.g., HWT and TSR). energy method. J Mater Civ Eng 2012;24(7):889–97.
[15] Hossain Z, Bhudhala A, Zaman M, O’Rear E, Cross S, Lewis S. Evaluation of the
use of warm mix asphalt as a viable paving material in the United States.
Based on the methodology and the materials used in this study, Research report, Virginia: Federal Highway Administration, Turner-Fairbank
the following recommendations are suggested. Highway Research Center; 2009.
[16] Wasiuddin NM, Zaman MM, O’Rear EA. Effect of sasobit and aspha-min on
wettability and adhesion between asphalt binders and aggregates. Transp Res
1. SFE components of the asphalt binder are expected to change Rec 2008;2051:80–9.
with changing its source, due to variability in chemical com- [17] Wasiuddin NM, Fogle CM, Zaman MM, O’Rear EA. Effect of anti-strip additives
position of crude oil. Therefore, the use of asphalt binders on surface free energy characteristics of asphalt binders for moisture-induced
damage potential. J Test Eval 2007;35(1):36–44.
from different sources and adding them to the test matrix [18] Lytton RL, Masad E, Zollinger C, Bulut R, Little DN. Measurement of surface
is recommended. energy and its relationship to moisture damage. TxDOT report number 0-
 R. Ghabchi et al. / Construction and Building Materials 73 (2014) 479–489 489

4524-2. College Station, Texas: FHWA, Texas Transportation Institute, Texas Washington, DC: American Association of State and Highway Transportation
A&M University; 2005. Officials; 2013.
[19] Cheng D, Little DN, Lytton RL, Holste JC. Use of surface free energy of asphalt- [28] AASHTO. Standard practice for accelerated aging of asphalt binder using a
aggregate system to predict moisture damage potential. J Assoc Asphalt Paving pressurized aging vessel (PAV). R 28-12. Washington, DC: American
Technol 2002;71:59–88. Association of State and Highway Transportation Officials; 2012.
[20] Van Oss CJ, Chaudhury MK, Good RJ. Interfacial Lifshitz–van der Waals and [29] Bahia HU, Anderson DA. The pressure aging vessel (PAV): a test to simulate
polar interactions in macroscopic systems. Chem Rev 1988;88(6):927–41. rheological changes due to field aging. In: Hardin JC, editor. American society
[21] Ghabchi R, Singh D, Zaman M, Tian Q. Mechanistic evaluation of the effect of for testing and materials. West Conshohocken, PA: ASTM Special Technical
WMA additives on wettability and moisture susceptibility properties of Publication 1241; 1995.
asphalt mixes. ASTM J Test Eval 2013;41(6):933–42. [30] Galal KA, White TD, Hand AJ. Second phase study of changes in in-service
[22] Ghabchi R, Singh D, Zaman M, Tian Q. A laboratory study of warm mix asphalt asphalt. In: Joint transportation research program. Indiana: Purdue University;
for moisture damage potential using surface free energy method. In: 2000.
Proceedings of the 2013 airfield and highway pavement conference: [31] Bhasin A, Little DN. Characterization of aggregate surface energy using the
sustainable and efficient pavements, Los Angeles, California: Transportation universal sorption device. ASCE J Mater Civ Eng 2007;19(8):634–41.
& Development Institute of ASCE, June 9–12; 2013. p. 54–63. [32] Buddhala A, Hossain Z, Wasiuddin NM, Zaman M. Effects of an amine anti-
[23] Arabani M, Hamedi GhH. Using the surface free energy method to evaluate the stripping agent on moisture susceptibility of sasobit and aspha-min mixes by
effects of polymeric aggregate treatment on moisture damage in hot mix surface free energy analysis. ASTM J Test Eval 2011;40(1):1–9.
asphalt. J Mater Civ Eng 2011;23(6):802–18. [33] Jones DR, Kennedy TW. The asphalt model. Results of the SHRP asphalt
[24] Good RJ, van Oss CJ. The modern theory of contact angles and the hydrogen research program. SHRP A-1001 contract. Austin, TX: Center for
bond components of surface energies. New York, NY: Plenum Press; 1991. Transportation Research, University of Texas; 1991.
[25] Berg JC. Wettability. New York: Marcel Dekker Inc.; 1993. [34] Tarrer AR, Wagh V. The effect of the physical and chemical characteristics of
[26] Zettlemoyer AC. Hydrophobic surfaces, vol. 8. New York and the aggregate on bonding. Strategic highway research program. Report: SHRP-
London: Academic Press; 1969. A/UIR-91-507; 1991.
[27] AASHTO. Standard method of test for effect of heat and air on a moving film of
asphalt (rolling thin-film oven test) (ASTM designation: D 2872-04). T 240-13.