Construction and Building Materials 38 (2013) 423–442

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

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

Review

A review of state of the art on stripping phenomenon in asphalt concrete

Amir Mehrara, Ali Khodaii ⇑
Department of Civil and Environmental Engineering, Highway and Transportation Div., Amirkabir University of Technology, Tehran, Iran

h i g h l i g h t s

" Providing a comprehensive review of researches on stripping potential of HMA.

" Classifying stripping tests on HMA according to the nature of the tests.
" Extracting some applicable stripping indicators based on previous researches.
" Highlighting recent numerical research areas on stripping potential.

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

Article history: Stripping is an important type of distress about which gaining thorough information can improve the
Received 5 April 2012 pavements design methods. A brief review of researches on the stripping potential of asphalt mixtures
Received in revised form 28 July 2012 over the past 40 years has been presented in this paper. The contents of the article include mechanisms
Accepted 11 August 2012
of moisture damage occurrence and contributing factors, as well as an introduction to the variety of steps
Available online 29 September 2012
taken in an effort to experimentally and numerically model the moisture damage propagation.
Researches on stripping have been divided into three categories of field studies, laboratory investigations,
Keywords:
and numerical and computational analysis. Laboratory experiments are also categorized into five groups.
Asphalt concrete
Stripping
Tests on loose mixtures, destructive mechanical tests on asphalt concrete, nondestructive mechanical
Moisture susceptibility tests on asphalt concrete, mechanical tests which provide a measure of cohesion in asphalt and adhesion
Moisture damage between asphalt and aggregate, non-mechanical tests which provide a measure of cohesion and adhesion
Computational modeling based on surface energy theories, and non-destructive non-mechanical tests. In different sections of the
article, achievements including the shortcomings and blessings in each branch of researches have been
discussed and finally suggestions for future investigations have been presented.
Ó 2012 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
2. Moisture damage (definition, mechanisms, controlling factors) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
2.1. Moisture damage mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
2.1.1. Moisture damage micro-mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
2.1.2. Moisture damage macro-mechanisms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
2.2. Moisture damage controlling factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
3. Categorizing the conducted researches on moisture damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
3.1. Field studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
3.2. Laboratory studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
3.2.1. Tests on loose mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
3.2.2. Destructive mechanical tests on compacted mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

⇑ Corresponding author. Address: Department of Civil Engineering, Amirkabir

University of Technology, No. 424, Hafez Ave., Tehran 15914, Iran. Tel.: +98 021
64543000.
E-mail addresses: amir.mehrara@aut.ac.ir (A. Mehrara), khodaii@aut.ac.ir (A.
Khodaii).

0950-0618/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2012.08.033
424 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442

3.2.2.1. Stripping evaluation based on indirect tensile strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

3.2.2.2. Stripping evaluation based on permanent deformation index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
3.2.2.3. Stripping evaluation based on fatigue life index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
3.2.3. Nondestructive mechanical test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.2.4. Stripping evaluation using energy base theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.2.4.1. Stripping evaluation using mechanical tests that measure adhesion and cohesion potential in the mixture . . . . . . . . . . . . . . 431
3.2.4.2. Stripping evaluation using non-mechanical tests on mixture constitutive parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431
3.2.4.3. Stripping evaluation based on fracture mechanics and energy based theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432
3.2.5. Non-destructive non-mechanical tests on compacted mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
3.2.5.1. Stripping potential evaluation based on permeability and diffusion potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434
3.2.5.2. Stripping evaluation based on emission speed of surface waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
3.3. Numerical analysis and modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
4. Conclusions and suggestions for later research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440

1. Introduction
Moisture sensitivity of an asphalt mixture, generally called
stripping potential, is among the most important distresses of
asphaltic pavements. Although the effect of moisture on asphalt
mixtures, unlike the traffic loading and thermal stress, is not con-
sidered as a major solicitation, penetration of moisture through the
asphalt mixtures can increase the pavements vulnerability to the
other two solicitations [1–3]. Perhaps it has been near a century
that this type of distress has been discovered. Since then, lots of re-
searches have been conducted to understand the mechanisms of its
initiation and to find remedies to control its formation in pave-
ments. A brief look at the history of researches reveals that field
investigation, laboratory experiments and analytical studies are
the three major research areas around moisture damage. Before
the fifties, most researches were limited to field observations, but
afterwards, with the advent of laboratory test methods, researches
were concentrated on laboratory studies to find a correlation be-
tween test results and field observations. From about the turn of
the century, another branch of investigation has been gaining
attention, in which the stripping phenomenon has been studied
from a theoretical point of view, and an attempt has been made
to model the coupling effects of moisture damage and traffic
loading.
What is being put forth in this article is a summary of extended
researches having been conducted on the reasons and causes of
stripping from different researchers viewpoints during the past
40 years. Also, it is intended to take a look at studies which have
been done with the aim of evaluation, estimation and preservation
of stripping in asphalt concrete and asphaltic pavements consider-
ing the achievements and shortcomings of each. Finally, the new
areas of research based on the improved theoretical bases of eval-
uation and recognition of stripping are introduced.
2. Moisture damage (definition, mechanisms, controlling
factors)
Different varieties of definitions having the concept of moisture
distress can be seen in the literature. According to the definitions, a
separation of asphalt from aggregate or the rupture of asphalt tex-
ture in asphalt mixtures under the action of cyclic traffic load and
presence of water or water vapor at the same time is named strip-
ping. Kiggundu and Roberts gathered various definitions from a
number of researchers’ viewpoints such as, Petersen, Kennedy and
Tunicliff [4]. According to the named researchers, stripping takes
place due to loss of adhesion between aggregates and asphalt in
the presence of water; however, Kiggundu and Roberts provided
a more comprehensive definition for stripping in which both cohe-
sive and adhesive failures were considered as the reasons of mois-
ture damage. Although in some researches [5–7] the exertion of
cyclic traffic load is taken as another cause besides moisture, some
other scientists such as Birgisson et al. [8] believe that to evaluate
the effect of moisture on asphalt concrete, it is better to consider
the stripping affect of moisture due to its long term presence in
voids of asphalt mixture without encountering the complexity of
coupling of cyclic loading and moisture damage.
Apart from the definition of moisture damage and the mecha-
nisms of its formation in asphalt concrete there are some factors
related to the constitutive parts of the mixture and to the prevail-
ing environmental conditions that can have a profound effect on
expediting or impedance of moisture damage that will be dis-
cussed throughout the section.
2.1. Moisture damage mechanisms
Two mainstreams can be found in the researches that are con-
ducted on mechanisms of stripping. These two mainstreams are
studying the micro-mechanisms and macro-mechanisms of mois-
ture damage. In other words, there are some theories that explain
adhesion between asphalt and aggregate, cohesion in asphalt and
the failure of these two bonds on a molecular scale, yet some other
theories explain the adhesive and cohesive failure using mechani-
cal theories on a macro-scale. However, both approaches can be
seen in most of the recent researches.
2.1.1. Moisture damage micro-mechanisms
The micro-mechanism of adhesion, probably known as adhe-
sion theories can be categorized as below according to several
researchers [4,9–12].
 Mechanical theory: Based on this theory, the penetration of
asphalt to the irregularities of the aggregate surface can pro-
duce mechanical interlocking.
 Chemical reaction theory: This theory attributes the adhesion
between asphalt and aggregate to the chemical reactions that
may exist between the aggregate particles and the adsorbed
asphalt.
 Molecular orientation theory: According to this theory, asphalt
molecules adjacent to the aggregate surface, depending on the
amount of their polarity, align themselves towards the aggre-
gate surface, so that they can satisfy the energy demand of
the aggregate.
 Surface energy theory: This theory, which is developed based on
the thermodynamic principles, assumes that adhesion has a
thermodynamic nature and depends on the surface energy of
the materials.
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442 425

 Weak boundary theory: Adhesion failure can be assumed to hap-
pen as a cohesion failure in a weak layer of binder or the mate-
rial underneath (here aggregate).
As well as the five microscopic theories of adhesion, microscopic
effects of the penetration of water into asphalt mixtures were stud-
ied and six theories, named stripping mechanisms in some re-
searches [4,5,13–17] and asphalt–aggregate system response to
moisture in others [12,18] were developed. Caro and her coworkers
have presented a more comprehensive gathering of these theories
and a more complete explanation of each as below [12]:
1. Detachment (debonding): The separation of an asphalt film from
an aggregate surface by a thin film of water without an obvious
break in the binder layer (with a chemical or thermodynamic
nature) [12].
2. Displacement: Loss of material from the aggregate surface
through a break in the asphalt film and/or possible separation
of the aggregate/mastic (with a mechanical nature) [12].
3. Dispersion of the mastic: Weakening of the cohesion in the
asphalt binder or mastic due to long-term diffusion periods
and loss of material due to the presence of water flow (with a
mechanical nature) [12].
4. Film rupture and micro-crack: Ruptures in the mastic or aggre-
gates (with either a mechanical or thermodynamic nature) [12].
5. Desorption: Washing away of the outer layers of mastics due to
the presence of flow (a mechanical process after other types of
responses) [12], ‘‘According to Kringos this type of response is
an advective transport of mastic and happens following the dif-
fusion process’’ [17].
6. Spontaneous emulsification: Inverted emulsion of water droplets
in binders (with a chemical nature) [12].
Perhaps Fromm and Lottman were among the first researchers
who unlike their contemporary researchers in the 1960s and
1970s, evaluated the moisture damage on a micro-scale [13].
Fromm took two mechanisms of moisture damage into consider-
ation in 1974. The first mechanism, upon which the principle of
diffusion theory was established, was spontaneous emulsification
and movement of water droplets in asphalt until they got to the
aggregate and asphalt interface. The second mechanism was as-
phalt film rupture due to interfacial stress between three phases
of air, asphalt and water [13]. Fromm did not put forth any idea
about the reason of diffusion occurrence; however, he cited other
researches in which diffusion had been attributed to the osmosis
phenomenon [13]. This mechanism assumes that there is either a
considerable film of adsorbed water or pockets of moisture in the
pores which contain dissolved salts that cause water to pass
through the film by osmotic pressure. The effect of this mechanism
(osmosis) has been considered negligible by Thelen in the same
Ref. [13]. Kiggundu and Roberts also presented both opinions in
rejecting and supporting the osmosis action [4].
Ishai and Craus, focusing on the adhesion theory, utilized three
mechanisms of the chemical, mechanical and surface energy the-
ory to explain adhesion [9]. They also introduced the third theory
as the most powerful method of assessing adhesion between as-
phalt and aggregate compared to the other two. The chemical
adhesion theory was later used in other researches such as
[19,20]. Dalter and his coworkers [21] believed that all types of
aggregates, because of their hydrophilic nature, maintained a

Table 1
Characteristics of asphalt mixture and its components affecting the moisture susceptibility.

Main contributor Determining characteristics Favorable properties Supporting researches

Asphalt properties Viscosity High viscosity Thelen, Schmit, and Graf, cited from [27]
Chemical structure Existence of phenol and nitrogen Curtis et al. cited from [27]
Asphalt film thicknessb High asphalt film thickness Hicks, Tunnicliff, and Root, cited from
[27,8,10]
Permeation water holding capacity Low diffusivity and low water holding [17,28,29]
capacity
Aggregate properties Surface texture Coarse texture Hicks, Majidzadeh, and Brovold cited from
[27,30]
Porosity Dependant on the size of the pores Hicks and Thelen cited from [27]
Mineralogy Using basic aggregates – absence of clay Rice, Majidzadeh, and Brovold cited from
mineral [27,31]
Surface coating No coating Majidzadeh and Brovold, Tunnicliff and
Root, cited from [27], Kendhal [23]
Surface moisture Being dry Majidzadeh and Brovold, Kim et al. cited
from [27,23]
Surface chemical characteristics Ability to share electrons or to form a Hicks cited from [27]
hydrogen bond
Mineral fillers Existence of fillers in order to increase the Hicks cited from [27]
asphalt viscosity
Asphalt mixtures properties Void ratio Very low or very high (the critical range of Terrel and Shute cited [27,6,32]
void ratio is between 6% and 14%)
Average void size and distribution More or less that 1.3 for granite and more Masad et al. 2006 cited from [12]
of air voids in mixture or less that 0.9 for lime stone
Permeability Less than 103 cm/s [33]
Gradationa Very dense or open graded Hicks et al. (1991) and Kandhal (1994)
cited from [27]
Asphalt contentb High asphalt content Hickes [27]
Asphalt age Less aged asphaltc Cheng et al. [7,27]
Type of fine aggregates in mixture Using less natural sand as possible [34,35]
Additives Using anti-stripping ore preferably Tunnicliff and Root cited from [27,6,36]
polymeric additives
a
Birgisson and his coworkers in a paper that was published a few years ago claimed that dense fine graded mixtures and open coarse graded mixtures are more prone to
stripping [8]. Graf also put the same idea forward about open graded mixtures [30].
b
According to the research, as long as the asphalt film thickness around the aggregate is less than 60 lm, strength to adhesive failure will be less than strength to cohesive
failure. However, in a film thickness of more than 60 lm, the reverse trend is expected [16]. Arambula quoted the same implication from Lytton [37].
c
Aging can increase the stripping and decrease the healing potential; however, in the case that stripping is due to the oxidation process, a reverse trend may be seen [2].
426 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
Table 2
Environmental conditions during and after construction and the construction procedure affecting the moisture susceptibility.
Main contributor Determining characteristics
Prevailing conditions during Environmental temperature
construction of pavement
Precipitation
Compaction
Time period between construction of a friction
layer on a newly constructed HMA layer
Prevailing conditions after Precipitation
construction of pavement
Freeze–thaw cycles
Temperature
Wet-dry cycles
Sub grade water content
Drainage condition
Traffic load
Micro-organisms activity
PH of water that flows through the pavement
single-molecular layer of water which was chemically attracted to
their surface even after being hydrated at very high temperatures.
They assumed that the layer was one of the triggers of moisture
damage [21]. In 1991, Curtis and his coworkers, using a physio-
chemical mechanism, considered both adhesion and cohesion con-
tributions to chemical bonds between different types of aggregates
and asphalts by taking the effects of aging and oxidation into ac-
count [22].
2.1.2. Moisture damage macro-mechanisms
As well as the recognition of micro-mechanisms contributing to
moisture damage, some other macro-mechanisms were also devel-
oped. Although the micro-mechanisms finally lead the asphalt con-
crete towards failure due to microscopic activities, taking the
macro-damage point of view can be a better approach in order to
devise more realistic experimental methods of evaluating moisture
damage. To this end, the macro-mechanism of moisture damage
can be introduced as follows, according to the literature:
1. Formation of excess pore pressure in saturated pavement due to
traffic loads or thermal shocks [4,5], and formation of tensile
stress reaching as high as 20 psi, due to water evaporation in
asphalt mixture voids [23].
2. Hydraulic scouring due to the moving traffic load on saturated
pavement [4,24] which is addressed as pumping action in some
researches [17,26].
3. Physical erosion of asphalt due to high velocity hydraulic flows
in mixture voids following the diffusion process [23,25] which
is addressed as advective movement in recent researches [17].
It might be concluded that macro-mechanisms of moisture
damage are in fact a combination of effects of physical solicitations
such as stress caused by traffic and thermal loading and water en-
trance to pavement structure.
2.2. Moisture damage controlling factors
In a swift review of published papers about stripping, various
factors influencing the moisture damage potential can be seen. Per-
haps, the most comprehensive review of all probable factors that
affect the stripping potential has been provided by Bahia and See-
mab [27]. The same gathering can also be seen in some other works
Favorable properties Supporting researches
Warm weather Hicks, Majidzadeh, and
Brovold cited from [27]
No snow or rainfall Hicks cited from [27]
Sufficient Hicks cited from [27, 43]
About two summers [6]
No snow or rainfall-occurrence of dry periods after Hicks cited from [27,30]
precipitation
No freeze–thaw cycles Lottman, Taylor, and
Khosla cited from
[27,29,38]
Mild temperature-low difference between day and night [6]
temperature
No wet-dry cycles at high temperatures [22]
Low water content [6]
Good performance of drains [6,23]
Low traffic Fromm, Gzemski et al.,
cited from [27]
Utilizing additives that can impede the activity of micro- [39,40]
organisms in the asphalt and aggregate interface
Maintaining PH at low levels [4]
[7]. In this part of the paper, moisture damage controlling factors
have been presented in a similar format as in Bahia and Seemab’s
work [27] after adding some supplementary data to it.
Generally, the quality of adhesion between asphalt and aggregate
and cohesion in asphalt can be influenced by a variety of factors. A
part of these factors are related to the characteristics of the asphalt
mixture and its constitutive parts as can be seen in Table 1. Another
part of the factors are dictated by the construction and prevailing
environmental conditions which are introduced in Table 2:
3. Categorizing the conducted researches on moisture damage
The conducted research may be categorized in three groups of
field, laboratory and numerical analysis and modeling.
3.1. Field studies
Although scientists of other fields like Nicholson and Thelen have
studied the microscopic effects of the presence of water in asphalt
mixtures since before the 1950s [41,42], pavement experts such as
Fromm and Lottman began their continuous researches from the
early 1970s [13]. Kandhal is also among other contemporary
researchers who extensively evaluated the reasons of the stripping
based on field observations. He introduced the mechanisms of initi-
ation and propagation of detrimental effects of moisture in asphaltic
pavements lacking appropriate drains using field data and suggested
some diagnostic symptoms through which the moisture damage
could be pre-diagnosed in his 1989 and 1992 papers [23,25]. Gener-
ally, in the researches similar to those mentioned, the field perfor-
mance of pavements exposed to moisture damage has been
observed and used as a criterion to assess conventional laboratory
methods in estimating moisture damage potential or the effective-
ness of different additives in decreasing stripping potential.
3.2. Laboratory studies
It can be implied that most of the studies about stripping poten-
tial were experimental and based on laboratory investigations.
These researches were mainly performed with one or all of three
aims introduced here in.
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442 427

1. Evaluating the effectiveness of anti-stripping additives in the
impedance of moisture damage using laboratory tests.
2. Developing new experimental methods and comparing their
ability to evaluate stripping potential with conventional test
methods and field experiments.
3. Extracting strength and behavioral parameters in order to esti-
mate the performance of asphaltic pavements in presence of
moisture detrimental effects.
Different methods have been utilized by researchers to evaluate
stripping potential. This diversity may be attributed to the enor-
mous factors that contribute to stripping and their cross interac-
tions. A summary of common test methods can be found in
many researches such as [7,23,43,44]. Solaimanian has presented
the most comprehensive collection of test methods in this area
[45]. Tables 3 and 4 exhibit moisture damage evaluation tests
based on Solaimanian’s [45] work which has been completed by
adding some other types of tests from other references and reviews
of different researchers about the tests.
As mentioned earlier, the tests introduced above have been
developed approaching one of the three aims. For instance, in
many researches [14,38,55–64,64,66], the main aim was to assess
the applicability of tests introduced in Tables 3 and 4 in measuring
the effectiveness of using additives such as lime, amino-acid, and
polymers in controlling stripping occurrence.
A lot of other researchers have been aiming to compare new
developed experimental methods with conventional tests and to
introduce new parameters and criteria as measures of stripping
susceptibility. Caro and her coworkers have published two papers
about stripping mechanisms and behavioral parameters developed
by different researchers to evaluate stripping potential [12,66]. In
this work, the introduced indices were categorized based on the
number of behavioral parameters each one includes. Caro has used
a sophisticated generic method of grouping the experimental tests
[12,66]. In this method, the stripping tests are mainly divided into
single parameter and multiparameter tests. In single parameter
tests, the output merely includes a single behavioral parameter,
but multi-parameter tests provide criteria with several behavioral
and mechanical parameters [12,66]. In this article, an attempt has
been made to categorize the tests according to the performance of
the test and the philosophy upon which the tests were invented. To
this end, five categories can be introduced as below:
1. Tests on loose mixtures.
2. Destructive mechanical tests on asphalt concrete.
 Evaluation of stripping potential using indirect tensile
strength test.
 Evaluation of stripping potential using permanent deforma-
tion index.
 Evaluation of stripping potential using fatigue index.

Table 3
Variety of test methods on loose asphalt mixture or on components of mixture to evaluate stripping potential.

Test name Measured parameter in test
Methylene Blue The amount of harmful clays of the smectite
(montmorillinite) group, organic matter and
iron hydroxides present in fine aggregates
Static immersion Percent of aggregates surface that have
maintained their asphalt coatings after static
immersion in water
Dynamic immersion Percent of aggregates surface that have
maintained their asphalt coatings after being
agitated in water
Chemical immersion A quantitative index based on the
concentration of a chemical material for the
initiation of moisture damage
Boiling Percent of stripped aggregates after immersion
in boiling water
Surface reaction A quantitative index based on the pressure of
produced gas due to reaction of a chemical
with the stripped surface of aggregates
Net adsorption A quantitative index based on the difference of
the adsorbed asphalt to aggregate surface in
presence and in absence of moisture
Bottle test Measuring the adhesion capability of asphalt to
Ottawa sand
PATTI Measuring the required force to cause adhesion
failure between asphalt and aggregate in
presence and absence of moisture
TTS Measuring the required force to cause cohesion
failure in asphalt
Wilhelmy plate Measuring the surface energy of asphalt
USD Measuring the surface energy of aggregate
Approach of the test
Concentrating on the effects of mineralogy
of the surface coating of aggregates on
adhesion bond
Concentrating on adhesion bond failure
Similar to previous test
Similar to previous test
Similar to previous test
Similar to previous test
Concentrating on the adsorption potential
of aggregate based on surface energy
theory
Concentration on effect of anti-stripping
additives on asphalt
Evaluation of adhesion potential and
providing the possibility of cohesion
failure
Evaluation of cohesion failure potential
Determining the cohesive energy of
asphalt using surface energy theory
Determining the adhesive energy between
asphalt and aggregate using surface
energy theory
Evaluation of test performance
A suitable method of evaluating the effect
of fine aggregates on mixtures stripping
potential and showing a good correlation
with TRS index [25]
Focusing on performance of coarse
aggregates without taking the effects of
fines on stripping into account-using a
method that can hardy simulate adhesion
bond failure [43]
Similar to previous test [46]
Similar to previous test [46]
Estimation of 58% success in predicting
moisture damage [4]
[45,46]
Poor indicator of the moisture sensitivity
of binder [47]
[43,46]
Good correlation with TSR and wheel
tracking test results in estimating
resistance to stripping and the failure
mode (adhesion or cohesion failure) [36]
Good correlation between the amount of
stripped surface of aggregates, which is
calculated using the surface energy
parameters, and ratio of change in
stiffness in immersion compression test
[28,29]
428 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
Table 4
Variety of test methods on compacted asphalt mixtures to evaluate stripping potential.
Test name Measured parameter in test
Moisture vapor Measuring the ratio of Heveem stability
susceptibility after moisture conditioning to that before
conditioning
Immersion Measuring the ratio of compressive
compression strength after moisture conditioning to
that before moisture conditioning
Marshall immersion Measuring the ratio of Marshall strength
after moisture conditioning to that before
moisture conditioning
Freeze–thaw pedestal The number of freeze–thaw cycles to crack
initiation in sample before and after
moisture conditioning
Lottman Measuring the ratio of diametric strength
after moisture conditioning to that before
moisture conditioning (with freeze and
thaw cycles)
Root–Tunicliff Measuring the ratio of diametric strength
after moisture conditioning to that before
moisture conditioning (without freeze and
thaw cycles)
Double punch Measuring the ratio of punch shear
strength after moisture conditioning to
that before moisture conditioning
ECS Measuring the permeability of compacted
asphalt mixtures and changes of their
resilient modulus during application of
thermal cycles and cyclic loading
Wheel tracking Number of cycles corresponding to the
intersection point of the slop of second
and third part of creep curve (known as
stripping turning point)
Beam fatigue Fatigue life before and after moisture
conditioning
RSST-CH Measuring the ratio of permanent shear
strain of samples after moisture
conditioning to that before moisture
conditioning
SATS Measuring the ratio of resilient modulus
after moisture conditioning to that before
moisture conditioning
SPT Measuring dynamic modulus and drawing
master curve before and after moisture
conditioning
DMA Number of loading cycles corresponding
to the turning point of shear modulus
curve versus loading cycles
MASW Measuring modulus, Poisons’ ratio and
density of samples before and after
moisture conditioning
3. Nondestructive mechanical tests on asphalt concrete.
4. Energy based methods.
 Mechanical tests that measure adhesion and cohesion
potential in the mixture.
 Non-mechanical tests on mixture constitutive parts.
 Fracture mechanics and energy based indexes.
5. Non-destructive non-mechanical tests.
Approach of the test
Concentration on stripping due to water
vapor dispersion into mixtures void areas
Concentration on effect of moisture
damage on permanent deformation
resistance
Concentration on effect of moisture
damage on Marshall resistance
Concentration on effect of moisture
damage on thermal cracking
Concentration on effect of moisture
damage on diametric strength which is a
measure of fatigue and thermal cracking
damage
Similar to previous test
Similar to previous test
Concentration on effect of moisture
damage on samples resilient modulus (an
important parameter in flexible
pavements design during last 2 decades)
Concentration on effect of moisture
damage on permanent deformation
resistance
Concentration on effect of moisture
damage on fatigue cracking
Concentration on effect of moisture
damage on permanent deformation
resistance
Concentration on effect of moisture
damage on resilient modulus (with taking
the effects of aging into consideration)
Concentration on effect of moisture
damage on samples complex modulus (an
important parameter in flexible
pavements design in recent years)
Concentration on effect of moisture
damage on sand-asphalt samples shear
modulus and fatigue life
Providing a complex parameter that takes
mixtures behavior under compression and
shear loads using emission of mechanical
waves in both directions
 Using permeability as a measure of stripping potential.
 Using mechanical wave moving velocity as a measure of
stripping potential.
3.2.1. Tests on loose mixtures
These types of tests, which have been taken favorably, even un-
til the early 1990s, generally include the tests introduced in Table
3. A part of the tests such as static immersion, dynamic immersion
Evaluation of test performance
[23,46]
Estimation of 47% success in predicting
moisture damage according to field
experience [23]
[7,23]
Good ability in determining the
effectiveness of anti-stripping additives
[23]
Existence of difference in moisture
damage estimation capability between
this test and ITS due to difference in the
rate of strain application [30]
Estimation of 67% and 76% success in
predicting moisture damage according to
Original and modified Lottman’s method
[23]
Estimation of 60% and 67% success in
predicting moisture damage in the case of
using 70% and 80% TSR criteria
respectively [23]
Higher repeatability of results (low
standard deviation) in comparison to ITS
test [48]
Having the capability to be used in
mixture design procedure (cited from
[49])
Higher repeatability of results (low
standard deviation) in comparison to ITS
test [50]
Better and more similarity in performance
to factors causing stripping and
permanent deformation in field compared
to other types of tests such as TSR [35]
Underestimation of stripping potential
mixtures made of softer asphalt in
comparison to flexural beam test [51,52]
Good capability of estimating moisture
damage potential and agreement with ITS
test results [51,52]
Incapability in evaluating moisture
damage potential- need to be studied
more [51,52]
Yielding the same results as AASHTO ITS
test [44]
Having a better performance in estimating
stripping potential compared to Wheel
tracking and ITS tests [53]
Having a good capability in measuring
stripping potential [1,16]
Being superior to ITS or dynamic modulus
because of taking directional differences
in mixtures mechanical traits [53]
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
and boiling tests provide a qualitative measure of stripping. The
other part of the tests like net-adsorption and surface reaction,
introduce indirect quantitative measures through which the
moisture damage is evaluated merely according to the final steps
of adhesion failure regardless of the process of water permeation
and diffusion. The other deficiency that can be attributed to tests
on loose mixtures is that the tests mostly focus on a narrow range
of aggregate size and do not consider the stripping potential of the
whole grading in a mixture. For example, in a Methylene Blue test,
the effect of fine aggregates is considered on stripping potential.
Although the results of the researches such as [23,34,68] testifies
that the asphalt mixtures stripping potential is mainly affected
by fine rather than coarse aggregates, tests on loose mixtures,
which are not developed on the ground of a fundamental mechan-
ical theory, cannot explain the stripping potential comprehen-
sively. These tests may not allow the study of probable changes
in the behavior of asphalt concrete under the coupling effect of
moisture and load. The only reason for the long-term durability
of these tests was their simple, fast and cheap approach on the
one hand and lack of appropriate equipment on the other hand.
3.2.2. Destructive mechanical tests on compacted mixtures
These types of tests are rather versatile. ITS (Indirect Tensile
Strength), Heveem, Double punch, Beam fatigue, Marshall and
Compression tests are the common tests of this type. During the
last few decades, a number of valuable extensive researches have
been performed on improving the test methods and extracting
stripping parameters. After evaluating the suitability of three types
of tests including static and dynamic immersion, color indicator
and bottle test and mechanical tests (Marshall, ITS, resilient mod-
ulus, compression and Heveem tests) Tunnicliff and Root con-
cluded that mechanical tests can simulate the mixtures behavior
more realistically and can be effectively used to assess the perfor-
mance of anti-stripping additives [43]. They also believed that an
ITS test can better explain the adhesion mechanism. They added
that in the case of the main idea being the assessment of the struc-
tural adequacy of the pavement, using the resilient modulus as a
stripping index would be more logical [43]. In later papers pub-
lished by the writers in 1983, a more critical review was allocated
to the other types of tests and the ITS test was strongly recom-
mended [69].
3.2.2.1. Stripping evaluation based on indirect tensile strength. In
most researches that have been conducted before the 1990s and
especially the late 1970s and early 1980s, the prevailing method
of stripping evaluation was Lottman’s indirect tensile strength
method or similar modified versions of it. According to Kennedy,
the different methods have been utilized by researchers to evalu-
ate stripping potential. ITS test, while simple, has the capability
of measuring the potential of asphalt concrete to different dis-
tresses such as thermal and fatigue cracking and permanent defor-
mation as well as stripping [70].
Gilmor and his coworkers in a paper published in 1984 defined
the concept of adhesion and cohesion and their relation with indi-
rect tensile strength as follows [71]:
DTS10 ¼ Cl0 ðDS01 Þ þ DCl10 ð1  S1 Þ
In Eq. (1), the subscript ‘‘0’’ refers to sample properties before water
conditioning and subscript ‘‘1’’ refers to the same after moisture
damage (water conditioning). Having this in mind, the rest of the
parameters can be defined as follows:
 TS: Tensile strength of asphalt concrete in the indirect tension
mode.
429
 DTS1-0: Difference between indirect tensile strength of samples
that have gone through water conditioning and that of dry
samples.
 S0: Stripped surface of aggregates in the fractured sample before
moisture conditioning.
 S1: Stripped surface of aggregates in the fractured sample after
moisture conditioning.
 Cl0: Indirect tensile strength per surface of non-stripped aggre-
gates before moisture conditioning. This strength is an indica-
tion of cohesion strength ðCl0 ¼ TS0 =ð1  S0 ÞÞ.
 Cl1: Indirect tensile strength per surface of non-stripped aggre-
gates after moisture conditioning. This strength is an indication
of adhesive strength ðCl1 ¼ TS1 =ð1  S1 ÞÞ.
 DS1-0: Change in the stripped surface of aggregates in the frac-
tured sample before and after moisture conditioning.
On the right hand side of Eq. (1), the first term is an index of
adhesion failure and the second term is an index of cohesion failure
in the mixture. It should be noted that in developing Eq. (1), the
authors hypothesized that indirect tensile strength is attributed
to the cohesion bond of the part of asphalt that has maintained
its adhesion to aggregate, so the parts of the asphalt that have gone
through adhesion failure do not contribute to tensile strength. Gil-
mor also concluded that adhesion failure happens due to environ-
mental solicitations rather that traffic loading [71].
Lottman and Leonard presented a pavement management soft-
ware named ACMODAS (Asphalt Concrete Moisture Damage Anal-
ysis System) (quoted from [71]). ACMODAS had two subprograms.
One of the subprograms predicted the remaining life of the pave-
ment using ITS and resilient modulus before and after stripping,
while the other subprogram was making a decision whether to
use anti-stripping from an economical point of view. The ACMO-
DAS program predicts life of the pavement under wet conditions
based on the magnitude of tensile strength and retention of
strength following moisture conditioning in accordance with Lott-
man’s method in a laboratory [14]. Busching and his coworkers
used the software in their research later [14]. In the same year,
Nesichi and his coworkers from a critical point of view to conven-
tional stripping criteria stipulated that even though there were
various methods to evaluate stripping potential, no logical relation
existed to relate the laboratory experiences to the reduced life of
pavement in the field [72].
According to these researchers, even though Lottman tried to
reduce the deficiencies of ACMODAS in his later works, there were
still some limitations in regard to presumptions made to develop
the life estimation algorithm of the program, the most important
of which are presented herein [72]:
 Calculating the life of pavement by assuming that the mechan-
ical characteristics of other layers are constant and prefixed.
 Taking only two fixed stages of field moisture damage into
account in calculating the retained life of the moisture damaged
pavement.
 Assuming that the moisture damage merely affects the fatigue
life and its influence on other modes of damage was ignored.
 Inclusion of a qualitative traffic factor in calculations. This factor
ð1Þ
could be representative of four traffic levels of moderate, inter-
mediate, heavy and very heavy.
Nesichi and Ishai presented a well developed model for predict-
ing the effect of moisture damage based on ITS as below [72]:
TSRT
TSR ¼ ¼ ð1  eK=T Þ ð2Þ
TSR0
430 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
TSR2: diametric tensile strength ratio of samples with acceler-
ated conditioning (freeze plus warm water soak) [72]. This param-
eter is related to the long term performance of pavements which
have experienced full moisture damage.
In the work conducted by Nesichi and Ishai, a continuous math-
ematical equation was developed in which the actual amount of
equivalent axle load could be taken into consideration [72]. Based
on the results, good correlation was found between the TSR (tensile
strength ratio) values predicted for the pavement and their actual
performance after the moisture damage period [72].
In 1999, Bahia and Seemab, after gathering field data in regard
to pavement performance in sustaining moisture damage, devel-
oped an equation through which the ITS of a saturated sample is
estimated using its dry ITS. In Eq. (3), despite the complexity of
the model due to the inclusion of a higher number of variables,
the indirect strength of wet samples can be precisely estimated
with the R2 = 83.4%. In this equation, parameters such as air void,
anti-stripping usage and the saturation level of the mixture have
been integrated. It has been concluded that the TSR index should
adopt an amount between 0.602 and 0.805 in 95% reliability level
in order for the mixture to be acceptable [27].
Wet ¼ 133:74 þ C 7 þ C 8 þ 1:12Dry  0:422DeltaSat
2
þ 44:84Void  0:003Dry2  3:34Void
In Eq. (3):
 Wet: Indirect strength of saturated samples.
 Dry: Indirect strength of dry samples.
 DeltaSat: Saturation level of samples.
 Void: Void ratio of sample.
 C7, C8: Regression constants.
Although a lot of useful work has been done based on ITS tests
and TSR parameters, there are some shortcomings with regard to
determining the stripping potential using the ITS test, the most
important of which are given here:
 Using short-period freeze and thaw cycles or sudden saturation
which are not necessarily field like, can cause detachment
between asphalt and aggregate and this detachment differs
from real stripping [43]. The wetting process should be in a
way that the entrance of water to the voids of the mixture does
not cause immediate and sudden damage, but should result in
spontaneous and perpetual damage.
 The repeatability of test results using the ITS method is low
[27]. For example, small changes in the samples air void can
remarkably change the TSR index.
 Limiting the TSR index to a certain amount, as the only measure
to control stripping, particularly in the case of dealing with anti
stripping-modified samples may not guarantee their stripping
resistance performance. To overcome this inadequacy a limita-
tion should be set for the minimum amount of ITS for the
stripped sample as well as the limitation for TSR [23]. Dukatz
and Phillips suggested the minimum amount of 46–60 psi for
ITS [74].
 There does not seem to be an acceptable correlation between
laboratory test results and field experiments [23].
 An application of mechanical and moisture induced damage in
two different phases seems unrealistic [7].
 Relying on a single parameter to evaluate stripping potential
does not suffice. This single parameter may not necessarily
show the behavioral characteristics of the mixture and the
mechanisms contributing to moisture damage.
3.2.2.2. Stripping evaluation based on permanent deformation
index. Following the numerated shortcomings for the ITS test,
other types of tests were developed to evaluate the stripping po-
tential. Swailmi and Terrel invented a machine named ECS (Envi-
ronmental Conditioning System) in 1990 [50]. The advent of the
ECS machine brought about some drastic changes in measuring
the stripping potential of asphalt mixtures. The loading mechanism
in this machine was designed so that compression forces (causing
permanent deformation) could be exerted in two forms of static
and cyclic to asphalt samples while they were facing the condition-
ing process. The conditioning included wetting and applying ther-
mal cycles. Meanwhile, the changes in resilient modulus of the
samples were recorded as a measure of moisture damage [50].
Therefore, this method was established based on the integration
of permanent deformation and stripping potential. A comparison
between ECS and ITS tests indicated that the former yields more
repeatable results with lower standard deviation [75]. During later
years, the introduced method was improved and the modifications
were applied to the method of measuring resilient modulus, per-
manent strains and the conditioning and load application systems
by different researchers [3,76,77].
In a report by Chen and Huang [78], three experimental tests of
IDT for creep (Indirect Tensile Test), ITS and dynamic modulus
have been compared and an energy based index has been intro-
ð3Þ duced as seen in Birgisson’s work [8]. Chen and Huang also criti-
cized the ECS test due to its complicated mechanism and costly
equipment.
It should be noted that ECS methods had met lots of the imper-
fections of the conventional tests. For example, in this test, the
samples saturation was controlled by the level of vacuum and
the accessibility of water despite controlling the level of saturation
[67]. Furthermore, the potential of stripping was measured not as
an instant single index, but as a parameter (say resilient modulus)
change of which can be traced and measured during the test. How-
ever, the only question that remains is whether this resilient mod-
ulus can be indicative of a behavioral trait of the asphalt mixture?
Or whether it could provide a correct understanding of the mech-
anisms involved in moisture damage.
In the late 1990s, following a tendency to use wheel track tests,
utilizing the same test to extract an index for moisture damage po-
tential of mixtures became a common approach. In this method that
has been used in several studies [35,79], the main idea was defining
a stripping index based on the interaction of permanent deforma-
tion and moisture effect. This index can be obtained from the turn-
ing point of permanent deformation versus cycles of the load
application curve. One of the most important advantages of this in-
dex is its dependence on permanent deformation which is not only
traceable but also explains more complicated mechanical behavior
than changes of resilient modulus during damage evolution. In
other words, although the method is not sophisticated enough to
discriminate the mechanisms involved in stripping, the performer
can be given the assurance that a more comprehensive parameter
than resilient modulus has been obtained and the test can better
simulate the macroscopic mechanism such as pumping effects.
With the improvement of laboratory equipment, a third gener-
ation of test methods became the center of attention in which a
cyclic compressive load exerted to samples and changes in dy-
namic and creep modulus were recorded during the tests. In re-
searches like [7,79,80], this new method has been studied. In the
former paper, the authors have compared three indices obtained
from the dynamic creep test according to Kaloush’s method [81].
The indices are cohesion failure index, aggregates friction failure
index, and DPSE (Dissipated Permanent Strain Energy) index [7].
In the second research [80], the authors attempted to provide a
formulation, through which the stripping potential of asphalt
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
concrete could be monitored during compressive cyclic loading.
They proposed their formulation based on the three-stage creep
model developed by Zhou and his coworkers [82]. In the third re-
search in this category, the NAT (Nottingham Asphalt Tester) ma-
chine was used [46]. Results of the dynamic creep test using NAT
showed that between the flow number and the slope of linear part
of the creep curve, the latter has more repeatability in estimating
stripping potential [46].
3.2.2.3. Stripping evaluation based on fatigue life index. This type of
test can be found in three different forms in the literature. The first
form, which may be the most common one, is using the cyclic indi-
rect tension test as IDT. The second form is using the DMA (Dy-
namic Mechanical Analysis) test which is used for fine aggregate
mixtures. The flexural beam test is the last form of using a fatigue
test for stripping detection.
The cyclic indirect test has been used in a number of researches
[8,71]. Gilmore and his coworkers studied the effect of lime in con-
trolling the detrimental effects of moisture on asphalt concrete
[71]. Birggisson and his coworkers have also extracted a well-
developed energy base parameter named DCSE (dissipated creep
strain energy) using IDT [8]. In the DMA test method, a cylindrical
fine aggregate mixture is subjected to cyclic torsion [1]. Kim and
his coworkers first used this test to evaluate the stripping potential
of sand asphalt mixtures [1]. The details of the research will be gi-
ven in the following section. The four point bending test is a well
known method of testing the fatigue life of asphalt concrete; how-
ever its use for the purpose of measuring moisture damage in com-
pacted mixtures dates back to 1995 [83]. Two parameters of initial
stiffness and fatigue life were used by Shatnawi and his coworkers
as stripping indexes [83]. Studies conducted by Lu and Harvey
were other examples of using the beam bending test to measure
moisture damage [51,52]. Lu, in his thesis [84], compared this
method with ITS and the wheel track tests and recommended
using the bending beam as a comparable test to ITS and a better
test than wheel track. According to [51,52], fatigue life can better
distinguish the mixtures with different stripping potential in com-
parison to initial stiffness.
3.2.3. Nondestructive mechanical test
In most studies having been carried out since 1980, resilient
modulus tests were conducted besides destructive tests as second-
ary outputs, or the trend of the modulus changes have been re-
corded while the samples were experiencing permanent distress
such as permanent strains and fatigue damage. Samples of these
researches are [22,43,44,50,70,75]. Development of more sophisti-
cated behavioral models on one hand and the advent of new ap-
proaches in design methods in which a complex modulus was
substituted with a resilient modulus on the other hand [90], set
the foundation for the new non-destructive test methods to gain
interest. In these types of tests, the stripping potential of the intact
sample is measured merely based on the master curve of dynamic
modulus rather than performing destructive permanent deforma-
tion or fatigue tests. In researches such as [28,53] a comparison
has been made between the non-destructive dynamic modulus
test and other tests like wheel tracking or ITS. As a result, dynamic
modulus has been introduced as a valuable parameter for measur-
ing moisture damage [28,53].
As cited from Schwartz by Solaimanian, ‘‘in the case that initial
modulus of asphalt concrete decreases for about 25% due to mois-
ture damage, a 60% increase in rutting and a 30% increase in fatigue
damage will be expected’’ [53]. Based on this citation, Solaimanian
believed that developing a model for the change of the complex
modulus during the life time of asphalt concrete could have a
key role in the mechanistic-empirical design method of pavements
[53]; however, he did not provide any model in his research. In
431
2009, Jahromi used a dynamic modulus test to evaluate the anti-
stripping effects of lime on asphalt mixture [65].
3.2.4. Stripping evaluation using energy base theories
3.2.4.1. Stripping evaluation using mechanical tests that measure
adhesion and cohesion potential in the mixture. In these tests, the
stripping potential is evaluated directly based on the probability
of damage in the cohesion bond in asphalt or adhesion bond be-
tween asphalt and aggregate. Kanitpong and Bahia’s research
[10] is among the research carried out in this field. According to
Kanitpong and Bahia, stripping in asphalt is usually defined as
the damage in adhesion bond between asphalt and aggregate;
however, there are other physical and chemical mechanisms that
cause cohesion failure in asphalt by softening it [10].
Even though cohesive failure, in accordance with its effect on
moisture damage, is less important than adhesive failure, some
researchers believe that this type of failure is more likely to happen
[7]. The results of tests conducted by Kanitpong and Bahia on the
fractured faces of samples in the TSR test confirm this idea. In their
research, the fractured faces of ITS samples showed that the sam-
ples failure was more due to cohesive failure than adhesive failure
in dry samples; however, when samples were affected by moisture,
the predominant failure mode changed from cohesive to a mixture
of adhesive and cohesive [10]. Khattak and his coworkers have
come to a different conclusion in their research [93]. According
to them, the occurrence of cohesive or adhesive failure is depen-
dent on prevailing temperature [93].
Kanitpong and Bahia also added that the cohesion failure could
be divided into two parts, asphalt softening failure and adhesion
failure. They believed that in previous researches, fewer attempts
had been made to determine the contribution of any of the men-
tioned parts in stripping [10]. In this research, they used test meth-
ods named PATTI (Pneumatic Adhesion Tensile Testing Instrument)
and TTS (Tact Test System) to measure the adhesion between as-
phalt and aggregate and cohesion in asphalt, respectively [10]. In
this research, an attempt has been made to find a relation between
the results of the mentioned tests and the ITS test to see whether
adhesion and cohesion tests can describe the decrease in ITS after
moisture conditioning. According to the results, the main reason
for failure in dry asphalt samples was cohesion failure, but wet
samples mostly showed adhesion failure. They also found that
there was a good relation between the ITS of dry samples and
the average energy required for asphalt cohesion failure according
to the TTS test, and a good relation between ITS of wet samples and
the average of detachment strength from the PATTI test and mean
energy required for cohesion failure according to the TTS test. In
another study performed by Kanitpong and Bahia, the results of
introduced tests were compared with the results of ITS, axial com-
pression and wheel tracking test in evaluating the stripping poten-
tial of PMA mixtures. The results showed that PATTI and TTS
methods were capable of illustrating mixture behavior in the pres-
ence of moisture and their resistance to rutting [36].
3.2.4.2. Stripping evaluation using non-mechanical tests on mixture
constitutive parts. Unlike the tests introduced in the previous part,
in these tests, an attempt is made to measure the surface free en-
ergy of mixture parts (aggregates and asphalt) separately and to
calculate the strength of cohesion and adhesion bonds based on
the energy that is required to overcome the bonding. This method
has its roots in physio-chemical principles and has been studied for
the past several decades ago; however, its usage as an index for
moisture damage potential in asphalt mixtures dates back to the
late 1990s and early 2000. During this time span, using surface en-
ergy concepts in determining cohesion and adhesion strength in
asphalt mixtures gained popularity. Researchers such as Elphing-
ston, Li, Cheng, Little, Lytton, Hefer and King have measured the
432 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
physical and chemical forces between aggregate and asphalt using
the acid-base and Lifshitz-Van Der Waals components of surface
energy [28,29,61,91]. In this regard, Wilhelmy and USD (Universal
Sorption Device) methods can be named as the most-used tests. In
the Wilhelmy method, the surface energy of different types of
binders can be calculated based on the advancing or receding an-
gles [89] through which the potential of healing or cracking can
be evaluated. The USD method is used to measure the surface en-
ergy and adsorption capability of solid or semi-solid materials [28].
Utilizing these methods to calculate the surface energy of asphalt
and aggregate allows the most compatible compound of asphalt
and aggregate to have the strongest adhesion bond found and
the probability of moisture damage in the mixture to decrease
[28]. Cheng and his coworkers compared the results of the cyclic
compression test on compacted mixtures and the surface energy
tests in 2002b [28]. According to their conclusion, mechanical tests
validated the results of Wilhelmy and USD tests. They suggested
the equation below for calculating the surface percentage of the
aggregates that had been stripped [28].
Ewet DG12  ð1  PÞ þ DG132 P
¼ ð4Þ
Edry DG12
In the above equation:
 Ewet: Stiffness of saturated sample.
 Edry: Stiffness of dry sample.
 DG12: Adhesion surface energy between asphalt and aggregate
in absence of water.
 DG132: Adhesion surface energy between asphalt and aggregate
in presence of water.
 P: Percentage of surface area of the aggregate that is replaced by
water in the mixture.
Cheng and his coworkers published another paper in 2003 in
which the process of water diffusion in asphalt was chiefly ad-
dressed and apart from the factors that may be involved in this
phenomenon (as described by Fromm [13]), a diffusion model for
absorption and adsorption was developed using the USD test
method [29]. The determinant effect of the amount of absorbed
water to asphalt, in moisture reaching the aggregate surface and
in initiation of adhesive failure was set forth as significant in
achieving a diffusion model by the researchers [29].
Similar to what was seen in Kim’s work [1], Masad and his
coworkers developed an equation similar to Cheng and his cowork-
ers between complex modulus parameters from DMA test results
and the amount of stripped surface of aggregates [16].
 0 
G
aPW G DGaD  ð1  PÞ þ DGaW P
DG ¼  0 W ¼
G DGaD
G
D energies into account has the advantage of considering the real
In Eq. (14):
 DGaD: Adhesion energy bond between asphalt and aggregate in
dry conditions.
 DGaW: Adhesion energy bond between asphalt and aggregate in
wet conditions.
 DGaPW: Adhesion energy bond between asphalt and aggregate
in semi-wet conditions.
 P: Percentage of surface area of the aggregate that is replaced by
water in the mixture.
 aD 
 
The results obtained in the research DDGGaW  can be used as a good
indicator of moisture-susceptible combination of asphalt and
 aD 
 
aggregates. Values of DDGGaW  lower than 0.8, may show the possibil-
ity of moisture damage [16].
According to the literature, there are different varieties of theo-
ries that explain the adhesive and cohesive bond in asphalt mix-
tures. Most of the time a combination of two or three theories
might be used to provide a realistic perception of the mixtures
integrity or mechanisms involved in stripping. However, among
all those theories, the surface energy theory has a number of
advantages over other theories that distinguishes it from them.
First of all, being on the basis of thermodynamics, the surface en-
ergy theory can explain adhesion and cohesion bonds according
to both aggregate and asphalt physio-chemical properties. Sec-
ondly, as the surface texture of aggregates is taken into account
in the USD method, it can be realized that mechanical and physical
based adhesions can be effectively reflected in parameters used in
the surface energy theory. Thirdly, having the ability to separately
calculate the adhesion and cohesion bonds, the surface energy
method can give an idea about the contributions of any of these
bonds in the mixtures integrity. Last but not least, surface energy
provides a means to bridge the gap between physio-chemical
properties of asphalt and aggregate and the fracture mechanics
of mixtures. Using this capability, an explanation of the phenom-
ena such as healing or aging would be easier and more understand-
able. A sample of the latter can be seen in Masad’s work [16].
3.2.4.3. Stripping evaluation based on fracture mechanics and energy
based theories. An energy based index is more like a fabricated
parameter used for analyzing other test outputs rather than being
an experimental testing method. This technique is rooted from
fracture mechanics. By integrating different parameters obtained
from an experimental test using this technique, multi-factor
parameters for evaluating the behavior of mixtures can be at-
tained. As mentioned earlier, Birgisson and his coworkers have
used IDT in their research to measure moisture damage [8].
According to Birgisson, water can deteriorate the pavement in
two ways. The most well-known method which has been the ob-
ject of most recent research is the generation of cyclic hydraulic
pressure inside the voids of asphalt concrete under traffic loading
[8]. The second method is the prolonged presence of still water
in the mixture pores. The latter form has been addressed with
the concept of diffusion in some studies [28].
Birggisson has focused on the second form of moisture damage
without taking the baffling effects of hydraulic flow and high pres-
sure fields inside asphalt voids into account. Using the concept of
fracture mechanics, Birggison has attempted to find the effect of
moisture damage on certain key parameters of asphalt concrete.
To this end, he defined the DCSE threshold and measured the num-
ber of loading cycles after which the DCSE reaches its threshold,
and a preexisting crack propagates for a length of 1 in. According
ð5Þ
to Birggisson and some other researchers, taking dissipated creep
behavior of the mixture over a wide range of stresses including
destructive and nondestructive ones [8,85]. On the contrary, using
a single parameter criterion such as resilient modulus or tensile
strength, may not allow this comprehensive perception to be
achieved [8,85]. In Birggisson’s work, resilient modulus and stress
relaxation and ITS tests were conducted to gain the principle
behavioral parameters of the mixtures like FE (fracture energy)
and DSCE which, according to the definitions presented in previous
studies, show the sudden fracture and growth of the micro-cracks
threshold respectively [85–87].
Birggison believed that the parameters introduced above, can
not only control the initiation and propagation of crack in asphalt
concrete, but can also be effectively used to evaluate the detrimen-
tal effects of water [8]. Assuming that the wave form of the cyclic
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442 433

load was haversin with the loading and resting time of 0.1 and
0.9 s, Birggison defined the DCSE per cycle as follows [8]:
1 2
DCSE=cycle ¼ r D1 mð100Þm1 ð6Þ
20 AVE
In Eq. (6), rAVE is the average stress in the crack process zone, D1 and
m are regression parameters from the stress relaxation test. Birgis-
son, in his later researches, defined the energy ratio concept as fol-
lows [88]:
a  DCSE
ER ¼ ð7Þ
m2:98 D1

a ¼ 0:0299r3:1 ð6:36  St Þ þ 2:46  108 ð8Þ

Fig. 1. Pseudo stiffness versus number of loading cycles for two different strain
 ER: Energy ratio.
levels of 0.28% and 0.40% (adapted from [1]).
 St: Indirect tensile strength.
 r: Tensile stress of asphalt layer.

A year later in 2004, Kim and his coworkers tried to find a log-
ical relation between the damage index presented by Cheng
(according to Eq. (4)) and the ratio of pseudo stiffness of wet to
dry samples [1]. They drew these two parameters versus the num-
ber of load application and found a good agreement between them
(increase in stripped surface index developed by Cheng was equiv-
alent to decrease in pseudo stiffness ratio presented by Kim) [1].
Kim and his coworkers have used a similar concept as Birggisson
and defined a fracture mechanics base parameter using the outputs
of the DMA test as mentioned earlier [1]. In this regard, three
parameters of DPSE (Dissipated Pseudo Strain Energy), PS (pseudo
stiffness) and the number of loading cycles until failure were de-
fined as criteria for the stripping potential. Fig. 1 shows a sample
of results in Kim’s work [1].
According to Kim and his coworkers, DPSE and PS for wet sam-
ples were lower than that for dry samples, and increasing the
Fig. 2. R(N) versus N for dry samples (adapted from [16]).
strain level or moisture conditioning, the sample can result in high-
er rates of DPSE reduction. Kim believes if changes of shear modu-
lus are drawn versus the number of cycles, a transient point will be
observed. This transient point is an appropriate criterion for strip-
ping evaluation [1].
A couple of years later, Masad and his coworkers applied some
modifications to the DMA test method and developed a criterion
based on the crack growth rate. Taking advantage of the principle
law of viscoelastic fracture introduced by Schapery in 1984 [92], a
relation was set between the surface energy of crack faces and the
dissipated energy due to formation of cracks in each cycle of load-
ing as introduced in the equation below [16]:
  n 1=2nþ1
r ðNÞ D1 GR b
RðNÞ ¼ ¼ ð2n þ 1Þnþ1 a N ð9Þ
K 1=2nþ1
4pG1 DG

@W R
b¼ ð10Þ
@ lnðNÞ

DPSE Fig. 3. R(N) versus N for wet samples (adapted from [16]).
WR ¼ 0 ð11Þ
G

G
 K: Material constant that is reversely related to the square of
In the above equations: tensile strength.

 DGa: Adhesive bond energy. According to Masad, ‘‘Eq. (9) combines a mixture of physical,
 N: Number of load application cycles. chemical, and mechanical properties to calculate crack growth in
 GR: Reference modulus. dry and wet specimens’’ [16]. Therefore, in the case that samples
0
 G : Complex modulus of moisture damaged sample. have gone through moisture conditioning, the crack growth rate

 G : Complex modulus of intact sample. based on Eq. (7) would be higher than the same parameter in dry
 D1: Regression coefficient from stress relaxation test. samples. Details about how to obtain the crack growth law are gi-
434 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
ven in the related Ref. [16]. Figs. 2 and 3 show the crack growth
rate in dry and moisture conditioned samples based on Masad’s
work [16].
Research conducted by Khalid and Monney [89] is among other
studies having been conducted recently based on the concepts of
fracture mechanics. In this research, fracture energy and fracture
toughness of emulsified asphalt mixture have been used as criteria
for mixture moisture susceptibility.
Using the fracture mechanics theory and concepts such as frac-
ture and creep energy can be taken as a turning point in moisture
damage evaluation because these concepts allow the elimination
of viscoelastic strains from the total strain and reach an index
which shows real damage in the asphalt mixture, as well as the cri-
terion based on fracture mechanics having a multi-parameter nat-
ure. Caro believes that a multi-parameter criterion can allow a
more comprehensive evaluation of the moisture damage [67].
3.2.5. Non-destructive non-mechanical tests on compacted mixtures
In this section some properties of the asphalt mixture may not
have a direct role in the stripping phenomenon, but have a high
capability in realizing the factors that potentially affect the mois-
ture susceptibility mixtures without going through expensive
and time consuming experimental methods are presented.
3.2.5.1. Stripping potential evaluation based on permeability and
diffusion potential. Apart from the predominant failure mode in dry
or saturated samples, permeation of water into the asphalt–aggre-
gate system and water diffusion into asphalt is the first step in
moisture damage. A glance over the literature shows three proba-
ble ways that water penetrates into asphalt mixture: permeation
due to participation, vapor permeation from subgrade soil and cap-
illary rise. Therefore, in some researches, permeation or diffusion
coefficients are taken as moisture susceptibility indexes. Kutay
has developed a permeation model in which the interconnectivity
and tortuosity of voids as well as the porosity of the mixture have
been taken into account to provide a more realistic hydraulic flow
inside the material. Furthermore, he improvised the experimental
method of measuring hydraulic conductivity in order to obtain dy-
namic hydraulic conductivity in 2005 [94]. Although Kutay did not
relate his improved model to moisture degradation in asphalt con-
crete, his model may lay the foundation for developing a more pre-
cise permeation and diffusion model.
Choubane and his colleagues tended to determine the extension
of stripping based on the permeability of asphalt concrete and to
set a maximum limitation for the coefficient of permeability in vol-
umetric design methods. According to their results, amounts lower
than 103 for the permeability coefficient can greatly control strip-
ping risk [33].
In 2007, Masad and his coworkers presented a method in
which capillary movements of water in asphalt concrete voids
could be traced using the CT (computed tomography) scan meth-
od. Using the method and an integrated program, they found a
relation between the amount and distribution of voids and the
capillarity of the mixture. Considering the fact that the rate of
water penetration and the amount of water that has entered the
mixture is one of the determinant factors in moisture damage, Ma-
sad’s method can provide a good estimation of the amount of
water that may enter the pavement in sites where capillary rise
and diffusion is known to be the main source of moisture exis-
tence in the asphalt mixture [95]. Based on the results in Masad’s
work [95], moisture conditioning in some cases could cause more
voids than the connected voids already existing (in the dry state)
filled with water. He attributed this phenomenon to the damage
caused by moisture conditioning forming more connected voids
[95]. He also concluded that the smaller the voids, the higher
the capillary rise under static equilibrium with water weight,
and accordingly, the more time for water rise in the smaller voids.
He believed that his findings can be effectively used in the adop-
tion of an optimum air void structure (including the air void ratio
and the voids size distribution) which makes the mixture more
resistant to moisture damage [95].
In the same year, Arambula and her colleagues tried to make a
connection between air void distribution in mixtures and their
mechanical traits [37]. According to the obtained results, the smal-
ler the average radii of air void the higher the ratio of crack growth
rate of wet samples to dry samples. On the contrary, the larger the
voids’ radii, the lower the stripping potential [37].
In some cases the researchers have intended to develop an
experimental method to quantify the diffused water as well as
developing a diffusion model. In this regard, an experimental equa-
tion was introduced by Arambula in 2010, and an effective diffu-
sion coefficient of material can be found by means of the
equation as follows [96]:
 
1 dW H2 O RT L
Deff ¼  ð12Þ
A dt M H2 O P1  P0
P ¼ RHðPSat Þ ð13Þ
In the equations above:
 A: The part of the sample cross section through which water
vapor can escape (Fig. 4).
dW H
2O
 dt
: The rate of water that escapes from the container (can be
calculated by weighing the sample).
 L: Thickness of the sample (Fig. 4).
 P1, P0: Vapor pressure on two sides of the sample.
 R: Universal gas constant.
 T: Temperature in Kelvin.
 RH: Relative humidity.
 M H2 O : Molecular mass of water.
Using this effective diffusion coefficient and numerical model-
ing techniques allowed the water concentration profiles to be ob-
tained. Fig. 5 presents a sample of diffused moisture into a
micro-model developed by Arambula and her coworkers [96]:
The technique presented by Arambula and her coworkers may
not directly be related to moisture damage, but having a profile
of water concentration inside a micro-model of the mixture can
introduce the most vulnerable parts of the mixtures for different
combinations of aggregates and binders. In other words, any deg-
radation in cohesive and adhesive bonds in different parts of the
mixture can be effectively related to the water concentration in
those parts. As it has been stated by Arambula, ‘‘The numerical sim-
ulations demonstrate that moisture-related material properties of the
components of asphalt mixtures can be efficiently used to track mass
transport processes and their related structural degradation processes
without the need of conducting complex and expensive experimental
work. Therefore, they can be used as input in numerical models de-
voted to study the role of diverse environmental conditions and/or
characteristics of the asphalt mixtures (e.g., different gradation, air
voids content, asphalt binder content, etc.) to the initiation and evo-
lution of moisture induced damage’’ [96].
Arambula et al. [96] utilized the CT scan technique to extract
a numerical model of water diffusion [97]. They also made esti-
mations about the diffusion coefficient and compared the results
with their experimental measurements in their previous re-
search. The obtained pictures from computed tomography in this
work were analyzed using FE (finite element) and FD (finite dif-
ference) methods. Their results testified that the numerical
method in both FE and FD approaches overestimated the
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442 435

Fig. 4. Specimen setup illustrating the dimensions and experimental boundary conditions; the curved arrows represent hJi the molar flux from inside the container with
saturated water vapor pressure to the environmental chamber at 15% RH (adapted from [96]).

Fig. 5. Moisture diffusion within the FAM (fine aggregate mixture) for a sandstone mixture (in g/mm2, per unit of thickness) (adapted from Ref. [96]).

diffusion coefficients in comparison to the experimental method. In this equation:

Limitations in picturing techniques were the first reason that the
observed difference could be attributed to [97]. The second rea-
son was that in the experimental method, the diffusion coeffi-
cient was obtained for part of the mixture, yet in the
numerical analysis, the whole mixture was taken into account.
Finally, in the numerical method, it was assumed that constant
flow conditions prevailed [97]. They also concluded that in the
case of limited picture resolution, the FD method yields more
reasonable results compared to the FE method [97].
In another research, an experimental method in which the dif-
fusion coefficients were calculated by weighing the mastic samples
that were kept inside water was innovated [98]. In this research,
the amount of absorbed water into the mastic was obtained by
the gravimetric method and the effect of temperature on the diffu-
sion rate was also investigated. According to the results, asphalt
mixtures, mastics or asphalts that have the same diffusion coeffi-
cient in a given temperature might show a remarkable difference
in their diffusion coefficient at higher temperatures [98]. It is
worth mentioning that the results of the experimental test in this
research were compared with the diffusion model that had been
introduced by Pereira and Yarwood in 1996 [98]. The equation
below was suggested by Vasconcelos to consider the temperature
dependency of diffusion [98]. The diffusion coefficient calculated
for fine aggregate mixture in this research was at the same order
as what was reported by Kringos and her coworkers [99].
Ea
Deff ðTÞ ¼ D0 expðRT Þ
 Deff: Efficient diffusion coefficient obtained by combining two
modes of diffusion.
 D0: Diffusion coefficient when temperature goes to infinity
(m2/s).
 Ea: Activation energy which is the energy that a system must
take until a process can occur in it (J/mol).
 R: Universal gas constant.
 T: Absolute temperature in Kelvin.
3.2.5.2. Stripping evaluation based on emission speed of surface
waves. The MASW (Multichannel Analysis Surface Wave) method
is another experimental test which is used to measure moisture
damage and has newly been introduced by Barnes and Trottier
[54]. The method, which is a nondestructive method, has been
compared with dynamic modulus, ITS and visual inspection tests.
In this research, MASW has been developed based on the concept
that the speed of travel of mechanical waves produced by an im-
pact in an elastic or visco-elastic material is a function of material
density, modulus and Poisson coefficient [54]. Barnes and Trottier
believed that in the dynamic modulus test, the behavior of mix-
tures that were tested in the compression mode were affected by
stone to stone contact. On the other hand, in the ITS test, strength
criteria are defined not by stone to stone contact but by cohesive
and adhesive bonds under tensile and shear stresses. Putting these
ð14Þ two facts next to each other led writers to conclude that neither of
436 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
the mentioned tests could provide a comprehensive means to eval-
uate moisture damage [54]. However, in the MASW method, be-
cause due to part of the waves traveling vertically and the other
part traveling horizontally, the mechanical capability of the mix-
ture can be evaluated in both directions at the same time, so a
more general evaluation may be performed [54]. In this research,
Achenbach’s equation was used to transform shear and radial
waves travel speed to each other and to the dynamic modulus as
below [54]. In addition to that, it was attempted to define a mini-
mum acceptable amount of 0.85 and 0.9 for TSR and modulus ratio,
respectively, according to the comparison between results of the
new method and visual inspections of the stripped mixtures. The
introduced method in this research has been found to have high
sensitivity in detecting moisture susceptible mixtures with air
voids lower than 2%.
V S ð0:862 þ 1:14tÞ
VR ¼
1þt
E ¼ 2q V 2S ð1  tÞ
In Eqs. (15) and (16):
 VR: Surface wave velocity.
 VR: Shear wave velocity.
 t: Poisson’s ratio.
 q: Material density.
 E: Elastic modulus.
3.3. Numerical analysis and modeling
Researchers in the field of pavement engineering have been try-
ing to develop a test method which can measure state parameters
of materials in a way that the effects of all prevailing environmen-
tal and traffic conditions are reflected in it. In addition, the param-
eters are representative of the behavior of the material
independent of the size effect. They were also aiming to extract
the constitutive relation between stress and strain and to apply it
in predicting the material behavior in different conditions. More-
over, they are interested in integrating the characteristic behavior
into computerized numerical and analytical methods to provide a
mechanistic design procedure as may be seen in designing steel
Fig. 6. Displacement u (deformed mesh), water velocity w (arrows) and pressure p (color) plots at time t = 0.005 s for Case (a) hard asphalt (left) and Case (b) soft asphalt
(right) (adapted from [26]).
structures within the boundary of elastic behavior. Bearing this
in mind, the outstanding improvements in analytical softwares
have caused the first steps toward numerical analysis of a compli-
cated phenomenon such as moisture damage under traffic loading
to be taken from about 6 years ago.
Probably, the first numerical research in which the coupling ef-
fects of moisture and loading were analyzed might be conducted
by Kettil et al. [26]. In this research, a program based on the FE
analysis method was written. In the program, three principle equa-
tions on the basis of mass and momentum conservation were
established (Eq. (17)) and solved to obtain the amounts of defor-
mation, water speed and pressure inside the pavement [26].
8
> ~ T r  qu
r €  q f ðw _ þ wðrT wÞÞ þ qb ¼ 0
>
<
rp  R  qf u€  qnf ðw _ þ wðrT wÞÞ þ qf b ¼ 0 ð17Þ
>
>
: T _
p
r w þ amT e_ þ Q ¼ 0
ð15Þ
In this equation:
 u: Displacement vector.
ð16Þ € : Second derivative of displacement vector.
u
 p: Pore pressure vector.
_ First derivative of pore pressure vector.
 p:
 w: Darcy velocity of hydraulic flow vector.
 w:_ First derivative of hydraulic flow vector.
 q: Total mixture density.
 qf: Fluid density (water here).
 R: Arisen contact force due to the Darcy velocity.
 b: Body force.
 e_ : First derivative of strain vector.
 r: Total stress.
 Q: Effective bulk modulus.
 a: Factor that is determined according to the bulk modulus of
total mixture and the solid skeleton.
m ¼ ½1 1 1 0 0 0
In Kettil’s work, the pavement structure with all its layers has
been modeled in a real scale [26]. The hydraulic scouring and
pumping effects have been investigated by applying a pseudo-sta-
tic load to a saturated 2-D and 3-D FE model. Stress and deforma-
tion counters, water velocity vectors and pressure fields inside the
pavement were obtained as output in the cross section and longi-
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
tude section of the pavement in the 2-D model; however, 3-D anal-
ysis encounters the singularity problem and this did not converge.
Fig. 6 shows a sample of finite element analysis adapted from
Kettil’s work [26].
According to the results presented in this paper, the application
of moving load on pavement could create considerably high pres-
sure inside the pavement. On the contrary, water velocity was so
negligible that its ability to cause stripping was somehow under
question and was needed to be further studied. The behavior of
two types of flexible and semi-solid pavements were also investi-
gated and some differences were seen in them. The results showed
that in the hard pavement (flexible pavement with stiffer layers),
the pressurized water under the wheel load was pumped out of
the pavement and after the wheel passed over the section, it was
sucked back into the pavement. In the soft pavement (flexible
pavement with less stiffness) though, the direction of water flow
was out of the pavement in front of the wheel and towards the in-
side of the pavement in its back. The other difference was that in
the hard pavement the maximum pressure of 2 MPa was formed
under the wheel load and the maximum recorded velocity was
7E7 m/s, whereas the two mentioned parameters were
0.02 MPa and 2E2 m/s, respectively for flexible pavement [26].
Fig. 7. Two aggregate micro-model schematic (adapted from [99]).
Fig. 8. Comparison of damage development in different nodes of asphalt–aggregate micro-model for (a) dry condition (left), (b) 42 days of diffusion (right) (adapted from
[99]).
Fig. 9. Comparison of damage development in different nodes of asphalt–aggregate micro-model in case of (a) pumping action (left), (b) pumping action and erosion (right)
(adapted from [99]).
437
It should be noted, despite the fact that the theme of Kettil’s
work is innovative and can lay the foundation for later studies;
moisture damage has been referred implicitly in it. In other words,
Kettil has not provided any type of relation between the intro-
duced detrimental parameters (pressure and velocity) and damage
propagation in his work.
The second instructive study with the aim of coupling moisture
and loading effects has been conducted as a part of an extensive re-
search program, with the aim of completion of a finite element
modeling software named CAPA by Kringos in TU Delft University
in 2007 [17]. In the mentioned research that was performed as a
PHD dissertation, stripping in asphalt concrete was modeled on
the basis of the fact that involving factors in the phenomenon fall
into two groups: physical factors and mechanical factors. Accord-
ing to Kringos, physical factors cause damage through water diffu-
sion and advective transport or the washing away process which
can soften the mastic and erode it by hydraulic flows respectively,
and mechanical factors can produce damage through the pumping
effect of water following the formation of high pressure fields in-
side the mixture under the wheel load [17]. The equations used
by Kringos to consider the coupling effects of moisture and load
have been categorized into four parts as: developing the prevailing
equation of diffusion based on conservation of mass, finding the
equation of mass erosion, defining damage indexes for mastic
cohesion and adhesion bond between asphalt and aggregate, and
finally, developing the elasto-visco-plastic damage equation. For
the sake of brevity, the details of the mentioned equations have
not been presented here. For more information about the details,
readers are recommended to refer to [17,99,100].
After adding the supplementary subprogram to CAPA in order to
include moisture damage effects on the elasto-visco-plastic consti-
tutive equation, Kringos studied the mechanical behavior of a mi-
cro-model made of two aggregates (Fig. 7). Different cases were
studied in Kringos’ work. In Figs. 8 and 9, examples of plastic dam-
age formation under cyclic loading in the four cases of analysis
have been depicted [99].
438 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
Even though Kringos’ researches, unlike experimental tests, do
not provide an engineering approach to pavement design, involv-
ing parameters in diffusion and erosion have been effectively intro-
duced and discussed therein. According to Kringos, in the case of
encountering stripping damage, designers should look for the main
parameters having a contribution in damage rather than just using
experimental tests that merely control the cohesion and adhesion
bond adequacy. Involving parameters in diffusion and erosion are
suggested by Kringos as a diffusion coefficient of moisture, maxi-
mum capacity of mastic moisture intake, mastic desorption coeffi-
cient, diffusion coefficient of mastic and dispersion coefficient of
mastic [99]. Viewpoints similar to Kringos’ can be seen in Cheng’s
works in which a model for water absorption and adsorption has
been developed [29]. Kringos put forth that the asphalt binder
should have a low diffusion coefficient and the amount of filler
should also be determined in a way that the diffusion coefficient
of mastic and its erosion potential remains as low as possible [99].
In 2008, a paper was published by Mo and his coworkers in TU
Delft University in connection with analyzing the raveling poten-
tial of pavement using CAPA software [101]. In this paper, a
meso-scale model was developed to simulate the aggregates ravel-
ing off the pavement surface due to passing wheels. For more infor-
mation the readers might want to refer to [101].
In recent years, Texas A&M University has also joined the re-
search stream around asphalt mixtures stripping. Caro and her
coworkers have developed and presented a model to couple the ef-
fects of moisture damage and load application. However, unlike
Kringos’ approach, in which the thermodynamic theory and plas-
ticity law were used to form the required equations, they tried to
utilize fracture mechanics concepts in their model. In this work,
CZM (cohesive zone modeling) has been used as one of the fracture
mechanics models based on the traction separation law [18]. This
model is one of the predefined models in the Abaqus commercial
software. The presented finite element model in Caro’s work can
simulate the coupling effects of moisture diffusion and loading in
four steps. In the first step, the FE model is subjected to hydraulic
gradient as a boundary condition and through this period the dif-
fused water profile inside the model is obtained and the moisture
concentration can be extracted, then the visco-elastic properties of
FAM (fine aggregate mixture) and cohesive elements (interface be-
tween FAM and aggregates) can be calculated using the moisture
content data and the traction separation law. In step two, a
mechanical load is applied to the model and the deformation and
stress counters are recorded. During steps three and four, the
two previous steps are repeated. This replication can highlight
the most vulnerable parts of the model based on the maximum re-
corded amounts of stress, and also determine the probability of
adhesive or cohesive failure [18].
Fig. 10. Force versus time curves from a mixture comprised of FAM type 2 and (a) aggregate-1 and (b) aggregate 2, for the four possible moisture profiles generated using an
external moisture gradient type D (adapted from [102]).
In Caro’s researches, stripping of the asphalt mixture was mod-
eled merely due to diffusion based on Fick’s law and in order to
make the analysis simple, other damage mechanisms introduced
by Kringos (advective transport and water pumping) were ignored.
The effect of diffusion on the mechanical properties of the mastic
was considered through the two initiatives presented below and
the aggregates behavior was modeled elastically and independent
of moisture content [102]:
1. Visco-elastic properties of mastic are degraded as a function of
moisture content having been diffused into the mastic. This can
indicate the cohesive damage.
2. Visco-elastic properties on the interface material (between
mastic and aggregate) in the finite element model are degraded
as a function of moisture content having been diffused into the
mastic. This can indicate the adhesive damage.
Assessing the role of the moisture diffusion coefficient of FAM
and aggregates in damage development was one of the major
activities conducted by Caro and her coworkers [102]. Fig. 10
shows the force versus time and force versus displacement curves
for mixtures made of FAMs and aggregates with different moisture
diffusion coefficients in Caro’s work [102]. According to the ob-
tained results, samples that were subjected to more aggressive
moisture profiles went through more damage. A difference of one
order of magnitude in the diffusion coefficient of FAM may cause
a reduction up to 3.7 times in maximum sustainable force by the
mixture at a given displacement compared to a similar dry mix-
ture, and the two order of magnitude escalation in the aggregates
diffusion reduces the maximum force of resistance of the mixtures
from 1.2 to almost 2 times in comparison with similar dry mixes
[102].
The distribution effect of air voids inside the mixture is another
valuable work that has been conducted by Caro and her coworkers
[103,104]. According to Caro, there are two common techniques to
take the effect of air voids into consideration in numerical analysis
methods. In the first method, which is far from reality, the visco-
elastic properties of the FAM part are defined in a way that in-
cludes the effect of voids. In the second method, which despite
the higher accuracy makes the analysis more complicated and time
consuming, the actual voids are geometrically integrated into the
model [104]. In a part of her research program, Caro applied the
second method to find out the extent of the effect of random dis-
tribution of air void on the moisture related mechanical behavior
of a micro-model [103].
To achieve a model of this characteristic, she applied a probabi-
listic function of air void size having been obtained using the CT
scan method [103]. In her developed model, two methods were
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442 439

Fig. 11. Mean values of the 95% confidence intervals of (a) dissipated energy and (b) stiffness as a function of total air void content of asphalt mixtures subjected to 10 days
moisture diffusion and mechanical loading (adapted from [103]).

taken to allocate air voids to the model structure. In the first case,
the voids were substituted for a part of FAM and in the second case
voids were substituted for a part of aggregates [103]. The behav-
ioral parameters chosen to monitor the effect of air void distribu-
tion were DE (dissipated energy) and stiffness (area under the
stress–displacement curve after crack initiation). Fig. 11 shows a
sample of obtained results in which the 95% confidence intervals
of the two mentioned parameters using the bootstrapping tech-
nique have been presented from Caro’s work [103]:
As observed from Fig. 11, 95% confidence intervals can effec-
tively highlight an increase in the amount and variability of air
voids during the construction and can have a considerable effect
on the mixture performance not only in a moist environment but
even in dry conditions. According to Caro, ‘‘the change of the total
dissipated energy or stiffness with respect to the increase of one unit
of air voids content (for both dry and moisture-conditioned cases) is
a useful parameter to evaluate the overall damage’’ [103]. However,
it can be realized from the results of her work that the stiffness
of mixtures was less sensitive to the stated cases of void allocation
in comparison to dissipated energy. It was also seen that the vari-
ability of these two mechanical parameters were higher in the 7%
level of air void. On parts of the stripping potential evaluation, Caro
and her colleagues came to the conclusion that the stiffness param-
eter was almost 40% lower for wet samples than dry samples. They
also added that in dry samples, the method of void allocation and
the void amount could remarkably alter the stiffness parameter,
yet in wet samples, the stiffness was found almost insensitive to
the volumetric parameters. Therefore, they concluded that dissi-
pated energy might be a better parameter in order to evaluate
the mixture’s moisture sensitivity [103].
In another research conducted by Caro, a third method of consid-
ering the effect of randomly distributed air voids has been intro-
duced. In her suggested method, a stochastic modeling technique
based on random field theory was utilized according to which the
micro-model was divided into several sections and the air voids
were randomly distributed in each section. In this method, the allo-
cation of other parameters such as mechanical and hydraulic prop-
erties with the same distribution would be much easier [104].
In Caro’s research, a simple yet sophisticated cohesive zone mod-
eling was used to evaluate the stripping potential of mixtures. By
relying on fracture mechanics concepts, this model can effectively
allow the consideration of phenomena such as aging and healing
in the mixtures behavior. However, the assumptions that were made
for simplification need to be investigated more in the experiment
modulus. The connectivity and tortuosity of air voids which are
important factors in moisture damage are also a missing part in
Caro’s researches. Because of this, the hydraulic scouring and water
pumping effect as considered in Kringos’ and Kettil’s works and as
occurring in experimental tests and in field could not be modeled.
Fig. 12. Stress–strain diagrams for constant uniaxial strain rate simulations for
several moisture conditioning levels (adapted from [105,107]).
Following a research stream in an attempt to develop a realistic
continuum damage model for asphalt concrete in Texas A&M, Gra-
ham modified a nonlinear viscoelastic–viscoplastic model that had
been developed by Huang [106] in order to computerize the cou-
pling effect of moisture and loading in 2008 [105]. According to
Graham [105], in the models that were developed by Kringos
[17] and Caro [18], despite the pursuit of different principles and
methods in the simulation of moisture induced damage, an impor-
tant common deficiency can be seen [105], and that is the igno-
rance of irreversible damages induced by moisture. In other
words, in either of the mentioned works [17,18] the damage
caused by moisture is dependant only on the current amount of
diffused moisture and not on the history of changes in the moisture
content. Therefore, a decrease in moisture content would result in
a full recovery of mechanical properties in the material [105].
The model that has been modified by Graham suits plasticity
damage rules, but unlike the visco-elsto-plastic model developed
by Kringos [17] it takes the viscoplastic deformations into account
[105]. In this case, the damage occurrence in asphalt mixture
would also be rate dependant (dependant on loading rate) [105].
In Graham’s model, the nonlinear reversible deformations are ob-
tained using Schapery’s theory, and the rate dependant irreversible
deformations are calculated using Perzyna viscoplasticity with
modified Drucker–Prager yield surface [105]. Moisture damage in
Graham’s work has been modeled in accordance with the degrada-
tion in adhesive and cohesive bonds in asphalt concrete using inde-
pendent damage laws [105]. An extended parametric study has
been conducted by Graham in his thesis. Fig. 12 shows a sample
440 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
of the obtained results in which the normalized moisture content
(h) effect has been shown on the stress–strain relation of asphalt
concrete [105,107].
The model for coupling mechanical moisture damage presented
by Graham is probably the most comprehensive continuum dam-
age model up until this moment that has been developed to con-
sider the effects of moisture damage in asphalt concrete. It
should be noted that according to laboratory and field experi-
ments, the damage that has been caused by moisture in asphalt
concrete, may partly be healed in the absence of water (during
dry periods) [13,30], so the hypothesis of total irreversibility in
moisture damage has been proposed by Graham, although more
precise than a full recovery assumption, it cannot be generally ac-
cepted. Therefore, a parameter in regard with the healing rate
should probably be integrated into the modeling algorithm.
4. Conclusions and suggestions for later research
According to the reviewed researches about asphalt concrete
moisture damage, the most effort in this field has been made on
developing a test method that can not only simulate the process
of damage occurrence in the field but also provide an assessment
method through which the suitability of mixtures would be esti-
mated in designing steps and would be guaranteed during the
pavements service life. In some studies, as it was observed, taking
advantage of improved laboratory testing systems (ECS, UTM, NAT,
Wheel Tracking, DMA, and . . .) have made researchers capable of
providing the most similar loading and environmental conditions
to the field. In the literature, some investigators, including Lott-
man, Nesichi and Ishai [5,72,73], have tried to extract an estima-
tion model for a mechanical parameter (say ITS) to consider the
probable effects of moisture damage. However, an estimation
model based on a simple mechanical parameter cannot include
all the behavioral complexities of asphalt concrete.
Meanwhile, researches about theoretical bases of the stripping
have also played an important role in extracting and introducing
the most important factors involved in moisture damage and in
defining multi-parameter criteria to evaluate the stripping poten-
tial of asphalt. Studies conducted by Birgisson, Kim and Masad
are among those researches [1,16,88]. In these researches, a variety
of mechanical parameters have been put together to derive a com-
prehensive multi-parameter criteria such as dissipated creep strain
energy, fracture energy and crack growth rate. Having known the
mechanisms of stripping and involving factors from experimental
studies and putting them together with the capability of computa-
tional methods may provide a rigorous mechanistic-empirical de-
sign method through which the changes in mechanical traits of
pavements would be predictable.
As it can be understood from studies during the last decade, and
especially the past 5 years, researchers have been more inclined to
develop finite element mechanical models and use computational
methods to analyze the stresses and strains that might be formed
under the coupling effects of traffic and moisture permeation and
diffusion in the pavement. Kringos, Graham and Caro have taken
some essential steps to this end. Involving parameters in diffusion
and erosion are suggested by Kringos as the diffusion coefficient of
moisture, maximum capacity of mastic moisture intake, mastic
desorption coefficient, diffusion coefficient of mastic and disper-
sion coefficient of mastic [99]. Caro has also performed valuable
work in assessing the effect of random distribution of air voids in-
side a micro-model structure on the response of micro-models to
loading and moisture diffusion [12,103,104]. Graham, in his con-
tinuum damage model, has addressed most of the shortcomings
in previous coupled models as well [105,106].
In spite of all the precious steps taken toward scrutinizing mois-
ture damage in asphalt concrete, there is still a lot remaining. First
of all, despite the huge progresses that have been made in gaining
knowledge about theories of adhesion and cohesion and the mech-
anisms of stripping, the contribution of each of the theories in
explaining cohesive and adhesive bonds and each of the mecha-
nisms in causing stripping has not been determined for sure. Sec-
ondly, absolute agreements have not been observed between
laboratory test methods (especially newly developed methods)
and field experience in recognizing the susceptibility of asphalt
mixtures. According to the reports of Kandhal, a correlation of
47–76% has been recorded between experimental tests such as
ITS or immersion compression and field observations [23].
Thirdly, the FE models (Mostly 2-D) have been developed with
lots of simplifying assumptions about hydraulic conductivity, air
void distribution, etc. In addition, the behavior of the models has
not been verified by extensive experimental data. Finally, there
are some other important factors that can influence the moisture
sensitivity during service life. The aging process for example, can
make changes in the surface energy properties of asphalt and
aggregates and accordingly result in different cohesive and adhe-
sive bond energies as seen in Cheng’s works [28,29]. To sum up,
some research fields that the authors believe might shed light upon
stated uncertainties are suggested herein:
1. Trying to find a more realistic way to consider the effects of air
voids in numerical 3-D models in which tortuosity and inter
connectivity of air voids are taken into account.
2. Moisture can penetrate asphalt concrete either through perme-
ation or diffusion. Considering both ways in developing mois-
ture damage models is necessary in later researches.
3. Since promising results have been achieved using DEM in other
researches related to asphalt concrete mechanical behavior
[108], using the discrete element method in modeling the adhe-
sion and cohesion bonds might also be worth trying.
4. Extending the current micro-models to meso-scale 2-D models
and trying to verify the analysis results (stresses, strains, dissi-
pated energies) with laboratory tests.
5. Since the dynamic modulus is a critical parameter in designing
asphaltic pavements, achieving an estimation model for its
change during the service life of the pavement and applying this
model to FE simulations can provide better predictions of dam-
age propagation.
6. Using the concepts of extended finite element in order to model
the crack initiation and propagation in CZM and comparing it
with experimental results.
7. Integrating a healing parameter into the current micro-models
developed by Caro and Kringos and also the continuum damage
model introduced by Graham Fig. 5.
References
[1] Kim YR, Little DN, Lytton RL. Effect of moisture damage on material properties
and fatigue resistance of asphalt mixtures. J Transport Res Rec (TRB)
2004;1891:48–54.
[2] Kim YR. Modeling of asphalt concrete. USA: ACSE Press McGraw-Hill
Construction; 2009.
[3] Alam MM, Tandon V, Nazarian S, Tahmoressi M. Identification of moisture-
susceptible asphalt concrete mixes using modified environmental
conditioning system. J Transport Res Rec (TRB) 1998;1630:106–16.
[4] Kiggundu BM, Robberts FL. Stripping in HMA mixtures: state-of-the art and
critical review of test methods. National Center for Asphalt Technology, NCAT
report no. 88-2. Auburn University, Alabama; 1988.
[5] Lottman PR. Laboratory test method for predicting moisture – induce damage
to asphalt concrete. J Transport Res Rec TRB 1982:843.
[6] Kandhal PS, Rickards IJ. Premature failure of asphalt overlays from stripping:
case histories. J Assoc Asphalt Pav Technol (AAPT) 2001;703:47–406.
[7] Kim S, Coree BJ. Evaluation of hot mix asphalt moisture sensitivity using the
Nottingham asphalt test equipment. Department of Civil, Construction, and
 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
Environmental Engineering. Report no. IHRB Project TR-483. Iowa State
University; 2005.
[8] Birgisson B, Roque R, Page GC. Evaluation of water damage using hot mix
asphalt fracture mechanics. J Assoc Asphalt Pav Technol (AAPT)
2003;72:516–62.
[9] Ishai I, Carus J. Effect of the filler on aggregate-bitumen adhesion properties in
bituminous mixtures. J Assoc Asphalt Pav Technol (AAPT) 1977;46:228–59.
[10] Kanitpong K, Bahia HU. Role of adhesion and thin film tackiness of asphalt
binders in moisture damage of HMA. J Assoc Asphalt Pav Technol (AAPT)
2003;72:611–42.
[11] Hefer AW, Little DN, Lytton RL. A synthesis of theories and mechanisms of
Bitumen-aggregate adhesion including recent advances in quantifying the
effects of water. J Assoc Asphalt Pav Technol (AAPT) 2005;46.
[12] Caro S, Masad E, Bhasin A, Little DN. Moisture susceptibility of asphalt
mixtures, Part 1: mechanisms. Int J Pave Eng 2008;9(2):81–98.
[13] Fromm HJ. The mechanisms of asphalt stripping from aggregate surfaces. J
Assoc Asphalt Pav Technol (AAPT) 1974;43:191–223.
[14] Bushing HW, Amirkhanian SN, Burati JL, Alewine JM, Fletcher MO. Effects of
selected asphalts and antristrip additives on tensile strength of laboratory-
compacted Mershall specimens – a moisture susceptibility study. J Assoc
Asphalt Pav Technol (AAPT) 1986;55:120–49.
[15] Tarrer AR, Wagh V. The effect of physical and chemical characteristics of the
aggregate on bonding. Washington (DC): National Research Council; 1991.
[16] Masad E, Zollinger C, Bulut R, Little D, Lytton R. Characterization of HMA
moisture damage using surface energy and fracture properties. J Assoc
Asphalt Pav Technol (AAPT) 2006;46:919–53.
[17] Kringos N. Modeling of combined physical–mechanical moisture induced
damage in asphaltic mixes. PHD dissertation. TU Delft University, The
Netherlands; 2007.
[18] Caro S. A Coupled micromechanical model of moisture induced damage in
asphalt mixtures: formulation and applications. PHD dissertation. Texas A&M
University, USA; 2009.
[19] Anderson DA, Dukatz EL, Petersen C. The effect of antistrip additives on the
properties of asphalt cement. J Assoc Asphalt Pav Technol (AAPT)
1982;51:298–318.
[20] Scott NR, Ritter JB. Field and laboratory evaluation of debonding test
procedures. Report no. FHWA-AZHPR-A-15(149). Arizona Department of
Transportation; 1980.
[21] Dalter RS, Gilmore DW. The mechanisms of asphalt stripping from aggregate
surfaces. J Assoc Asphalt Pav Technol (AAPT) 1974;43:191–224.
[22] Curtis CW, Terrel RL, Perry LM, AL-Swailm S, Braanan CJ. Importance of
asphalt–aggregate interactions in adhesion. J Assoc Asphalt Pav Technol
(AAPT) 1991;60:476–533.
[23] Kandhal PS. Moisture susceptibility of HMA mixes: identification of problem
and recommended solutions, national center for asphalt technology. NCAT
report no. 92-1. Auburn University, Alabama; 1992.
[24] Taylor AM, Khosla NP. Stripping of asphalt pavements: state of the art. J
Transport Res Rec (TRB) 1983:911.
[25] Kandhal PS, Lubold CW, Roberts FL. Water damage to asphalt overlays: case
histories. J Assoc Asphalt Pav Technol (AAPT) 1989;58:40–77.
[26] Kettil P, Engstrom G, Wiberg NE. Coupled hydro-mechanical wave
propagation in road structures. Comput Struct 2005;83:1719–29.
[27] Bahia H, Seemab A. Evaluation and correlation of lab and field tensile strength
ratio (TSR) procedures and values in assessing the stripping potential of
asphalt mixes. Report no. WIISPR-10-99. University of Wisconsin – Madison.
Department of Civil and Environmental Engineering; 1999.
[28] Cheng DX, Little DN, Lytton RL, Holste JC. Use of surface free energy properties
of the asphalt–aggregate system to predict damage potential. J Assoc Asphalt
Pav Technol 2002;71:59–88.
[29] Cheng DX, Little DN, Lytton RL, Holste JC. Moisture damage evaluation of
asphalt mixtures by considering both moisture diffusion and repeated-load
conditions. J Transport Res Rec (TRB) 2003;1832:42–9.
[30] Graf PE. Factors affecting moisture susceptibility of asphalt concrete mixes. J
Assoc Asphalt Pav Technol (AAPT) 1986;55:175–213.
[31] Terrel RL, Scholz TV, Al-Joaib A, Al-Swailmi S. Validation of binder properties
used to predict water sensitivity of asphalt mixtures. J Assoc Asphalt Pav
Technol (AAPT) 1993;62:172–223.
[32] Terrel RL, Al-Swailmi S. Water sensitivity of asphalt–aggregate mixes: test
selection. National Research Council, SHRP-A-403. Oregon State University,
Carvallis; 1994.
[33] Choubane B, Page GC, Musselman JA. Investigation of water permeability of
coarse graded superpave pavements. J Assoc Asphalt Pav Technol (AAPT)
1998;67:254–71.
[34] Maupin GW. The use of antistripping additives in Virginia. J Assoc Asphalt Pav
Technol (AAPT) 1982;51:342–63.
[35] Pan C, White TD. Evaluation of stripping for asphalt concrete mixtures using
accelerated testing methods. J Transport Res Rec (TRB) 1998;1630:98–105.
[36] Kanitpong K, Bahia HU. Relating adhesion and cohesion of asphalts to the
effect of moisture on laboratory performance of asphalt mixtures. J Transport
Res Rec (TRB) 2005;1901:33–43.
[37] Arambula E, Masad E, Martin AE. Influence of air void distribution on the
moisture susceptibility of asphalt mixes. Mater Civil Eng (ASCE)
2007;19(8):655–64.
[38] Scherocman JA, Mesch KA, Proctor JJ. The effect of multiple freeze–thaw cycle
conditioning on the moisture damage in asphalt concrete mixtures. J Assoc
Asphalt Pav Technol (AAPT) 1986;55:213–37.
441
[39] Brown LR, Pabst GS, Marcev JR. The contribution of microorganisms to
stripping and the ability of an organofunctional silane to prevent stripping. J
Assoc Asphalt Pav Technol, (AAPT) 1990;59:360–82.
[40] Parker F, Benefield L. Effects of microbial activity on asphalt–aggregate bond.
J Assoc Asphalt Pav Technol (AAPT) 1991;60:188–207.
[41] Thelen E. Surface energy and adhesion properties in asphalt–aggregate
systems. Highway Res Board 1958;192:63–74.
[42] Nicholson V. Adhesion tension in asphalt pavements, its significance and
methods applicable in its determination. J Assoc Asphalt Pav Technol (AAPT)
1932;3:29–49.
[43] Tunnicliff DG, Root RE. Antistripping additives in asphalt concrete state-of-
the-art. J Assoc Asphalt Pav Technol 1981;51(1982):265–94.
[44] Airey GD, Choi YK, Collop AC, Moore AJV, Elliott RC. Combined laboratory
ageing/moisture sensitivity assessment of high modulus base asphalt
mixtures. J Assoc Asphalt Pav Technol 2005.
[45] Solaimanian M. Moisture damage – the research for a test we can trust, rocky
mountain asphalt conference and equipment show, Denver-Clorado; 2003.
[46] Williams C, Breakah TM. Evaluation of hot mix asphalt moisture sensitivity
using the nottingham asphalt test equipment. Report no. IHRB Project TR-
555. Institute for Transportation-Iowa State University; 2010.
[47] Scholz TV, Terrel RL, Al-Joaib A, Bea J. Water sensitivity: binder validation
national research council. SHRP-A-402. Washington (DC); 1994.
[48] Jimenez RA. Methods and treatments to control debonding. J Assoc Asphalt
Pav Technol (AAPT) 1990;59:93–138.
[49] Scott JAN. Adhesion and disbonding mechanisms of asphalt used in highway
construction and maintenance. J Assoc Asphalt Pav Technol (AAPT)
1978;47:19–49.
[50] Swailmi S, Terrel RL. Evaluation of water damage of asphalt concrete mixtures
using the environmental conditioning system (ECS). J Assoc Asphalt Pav
Technol (AAPT) 1992;61:405–46.
[51] Lu Q, Harvey JT. Investigation of conditions for moisture damage in asphalt
concrete and appropriate laboratory test methods. RR No: UCPRC-RR-2005-
15. Caltrans Division of Research and Innovation; 2005.
[52] Lu Q, Harvey JT. Evaluation of moisture sensitivity of hot mix asphalt by
flexural beam fatigue test. In: Proceedings of the R. Lytton symposium on
mechanics of flexible pavements (ASCE); 2006.
[53] Solaimanian M. Simple performance test for moisture damage prediction in
asphalt concrete. J Assoc Asphalt Pav Technol (AAPT) 2006;75:346–82.
[54] Barnes CL, Trottier J-F. Evaluating laboratory-induced asphalt concrete
moisture damage using surface waves. Int J Pave Eng 2010;11(6):489–97.
[55] Kennedy TW, Turnham N, EPPS JA, Smoot CW, Young FM, Button JW, et al.
Evaluation of methods for field applications of lime to asphalt concrete
mixtures. J Assoc Asphalt Pav Technol (AAPT) 1983;52:508–35.
[56] Suntucci LE, Allen DD, Coats RL. The effects of moisture and compaction on
the quality of asphalt pavements. J Assoc Asphalt Pav Technol (AAPT)
1985;54:168–209.
[57] Connor DL. Action taken by the texas department of highways and public
transportation to identify and reduce the stripping potential of asphalt mixes.
J Assoc Asphalt Pav Technol (AAPT) 1984;53:631–9.
[58] Christensen DW, Anderson DA. Effect of amine additives on the properties of
asphalt cement. J Assoc Asphalt Pav Technol (AAPT) 1985;54:593–619.
[59] Kennedy TW, Ping WV. An evaluation of effectiveness of antistripping
additives in protecting asphalt mixtures from moisture damage. J Assoc
Asphalt Pav Technol (AAPT) 1991;60:230–64.
[60] Dukatz EL, Phillips RS. The effect of air voids on the tensile strength ratio. J
Assoc Asphalt Pav Technol (AAPT) 1987;56:517–55.
[61] King G, Bishara SW, Fager G. Acid/base chemistry for asphalt modification. J
Assoc Asphalt Pav Technol (AAPT) 2002;71:171–99.
[62] Aksoy A, Samlioglu K, Tayfur S, Ozen H. Effects of various additives on the
moisture damage sensitivity of asphalt mixtures. Constr Build Mater
2005;19(1):11–8.
[63] Tunnicliff DG. Performance of antistripping additives. J Assoc Asphalt Pav
Technol (AAPT) 1997;66:344–69.
[64] Sebaaly PE. Comparison of lime and liquid additives on the moisture damage
of hot mix asphalt mixtures. National Lime Association; 2007.
[65] Jahromi SG. Estimation of resistance to moisture destruction in asphalt
mixtures. Constr Build Mater 2009;23(6):2324–31.
[66] Reinke G, Glidden S, Herlitzka D, Veglahn S. PPA modified binders and
mixtures: aggregate and binder interactions, rutting and moisture sensitivity
of mixtures. J Assoc Asphalt Pav Technol (AAPT) 2010;79:719–42.
[67] Caro S, Masad E, Bhasin A, Little DN. Moisture susceptibility of asphalt
mixtures, Part 2: characterization and modeling. Int J Pave Eng
2008;9(2):99–114.
[68] Kennedy TW, Roberts FL, Lee KW. Evaluation of moisture susceptibility of
asphalt mixtures using the texas freeze–thaw pedestal test. J Assoc Asphalt
Pav Technol 1982;51:327–42.
[69] Tunnicliff DG, Root RE. Testing asphalt concrete for effectiveness of
antistripping additives. J Assoc Asphalt Pav Technol (AAPT) 1983;52:535–61.
[70] Kennedy TW. Characterization of asphalt pavement materials using the
indirect tensile test. J Assoc Asphalt Pav Technol (AAPT) 1977;46:132–50.
[71] Gilmore DW, Lottman RP, Scherocman A. Use of indirect tension
measurements to examine the effect of additives on asphalt concrete
durability. J Assoc Asphalt Pav Technol (AAPT) 1984;53:495–525.
[72] Nesichi S, Ishai I. A modified method for predicting reduced asphaltic
pavement life from moisture damage. J Assoc Asphalt Pav Technol (AAPT)
1986;55:149–75.
442 A. Mehrara, A. Khodaii / Construction and Building Materials 38 (2013) 423–442
[73] Lottman RP. Predicting moisture-induced damage to asphaltic concrete-field
evaluation phase. Interim report no. ITD-RP081. Transportation Research
Board. National Cooperative Highway Research Program; 1982.
[74] Dukatz EL, Phillips RS. Control of stripping with polymer treatment of
aggregates. J Assoc Asphalt Pav Technol (AAPT) 1993;62:223–47.
[75] Swailmi S, Terrel RL. Evaluation of the environmental conditioning system
(ECS) and comparison with AASHTO T-283. J Assoc Asphalt Pav Technol
(AAPT) 1993;62:150–72.
[76] Tandon V, Vemuri N, Nazarian S, Tahmoressi M. A comprehensive evaluation
of environmental conditioning system. J Assoc Asphalt Pav Technol (AAPT)
1997;66:187–206.
[77] Tandon V, Alam MM, Nazarian S, Vemuri N. Significance of conditioning
parameters affecting distinction of moisture susceptible asphalt concrete
mixtures in the laboratory. J Assoc Asphalt Pav Technol (AAPT)
1998;67:334–49.
[78] Chen X, Haung B. Evaluation of moisture damage in hot mix asphalt using
simple performance and superpave indirect tensile tests. Constr Build Mater
2008;22:1950–62.
[79] Kandhal PS, Cynthia YL, Frazier P. Tests for plastic fines in aggregates related
to stripping in asphalt paving mixtures. J Assoc Asphalt Pav Technol (AAPT)
1998;67:223–37.
[80] Mehrara A, Khodaii A. Evaluation of asphalt mixtures’ moisture sensitivity by
dynamic creep test. J Mater Civil Eng (ASCE) 2011;23(2):212–9.
[81] Kaloush KE, Witczak MW. Tertiary flow characteristics of asphalt mixtures. J
Assoc Asphalt Pav Technol (AAPT) 2002;71:278–306.
[82] Zhou F, Scullion T, Sun L. Verification and modeling of three-stage permanent
deformation behavior of asphalt mixes. J Transport Eng (ASCE)
2004;130(4):486–94.
[83] Shatnawi S, Nagarajaiah M, Harvey J. Moisture sensitivity evaluation of
binder–aggregate mixtures. J Transport Res Rec (TRB) 1995(1492):71–84.
[84] Lu Q. Investigation of conditions for moisture damage in asphalt concrete and
appropriate laboratory test methods. PHD dissertation. University of
California, Berkeley; 2005.
[85] Roque R, Birgisson B, Sangpetngam B, Zhang Z. Hot mix asphalt fracture
mechanics: a fundamental crack growth law for asphalt mixtures. J Assoc
Asphalt Pav Technol (AAPT) 2002;71:957–1064.
[86] Zhang Z, Roque R, Birgisson B, Sangpetngam B. Identification and verification
of a suitable crack growth law. J Assoc Asphalt Pav Technol (AAPT)
2001;51:239–79.
[87] Roque R, Birgisson R, Zhang Z, Sangpetngam B, Grant T. Implementation of
SHRP indirect tension tester to mitigation cracking in asphalt pavements and
overlays University of Florida. UF project no. 4910450457912. Florida
department of Transportation; 2002.
[88] Birgisson B, Roque R, Page G. The use of a performance based fracture
criterion for the evaluation of moisture susceptibility in hot mix asphalt. J
Transport Res Rec (TRB) 2004(1891):51–61.
[89] Khalid HA, Monney OK. Moisture damage potential of cold asphalt. Int J Pave
Eng 2009;10(5):311–8.
[90] Clyne TR, Li X, Marasteanu MO, Skok EL. Dynamic and resilient modulus of
Mn/DOT asphalt mixtures. Report No. MN/RC – 2003-09. Department of Civil
Engineering. University of Minnesota; 2003.
[91] Hefer AW, Bhasin A, Little DN. Bitumen surface energy characterization using
a contact angle approach. Mater Civil Eng (ASCE) 2006;18(6):759–67.
[92] Schapery RA. Correspondence principles and a generalized J integral for large
deformation and fracture analysis of viscoelastic media. Int J Fract
1984;25:195–223.
[93] Khattak MJ, Baladi GY, Drzal LT. Low temperature binder–aggregate adhesion
and mechanistic characteristics of polymer modified asphalt mixtures. J
Mater Civil Eng (ASCE) 2007;19(5):411–22.
[94] Kutey MN. Modeling moisture transportation in asphalt pavements. PHD
dissertation. University of Maryland; 2005.
[95] Masad E, Arambula E, Ketcham RA, Abbas AR, Martin AE. Nondestructive
measurements of moisture transport in asphalt mixtures. J Assoc Asphalt Pav
Technol (AAPT) 2007;46:713–55.
[96] Arambula E, Caro S, Masad E. Experimental measurement and numerical
simulation of water vapor diffusion through asphalt pavement materials.
Mater Civil Eng (ASCE) 2010;22(6):588–98.
[97] Arambula E, Garbocz EJ, Masad E, Kassem E. Numerical analysis of moisture
vapor diffusion in asphalt mixtures using digital images. Mater Struct
2010;43:897–911.
[98] Vasconselos KL, Bhasin A, Little DN, Lytton RL. Experimental measurement of
water diffusion through fine aggregate mixtures. J Mater Civil Eng (ASCE)
2011;23(4):445–52.
[99] Kringos N, Scarpasa A, Copeland A, Youtcheff J. Modeling of combined
physical-mechanical moisture-induced damage in asphaltic mixes Part 2:
moisture susceptibility parameters. Int J Pave Eng 2008;9(2):129–51.
[100] Kringos N, Scarpas A, Kasbergen1 C. Three dimensional elasto-visco-plastic
finite element model for combined physical-mechanical moisture induced
damage in asphaltic mixes. J Assoc Asphalt Pav Technol (AAPT)
2007;76:495–524.
[101] Mo LT, Huurmana M, Wub SP, Molenaara AAA. 2D and 3D meso-scale finite
element models for raveling analysis of porous asphalt concrete. Finite Elem
Anal Des 2005;44:186–96.
[102] Caro S, Masad E, Bhasin A, Little D. Micromechanical modeling of the
influence of material properties on moisture-induced damage in asphalt
mixtures. Construct Build Mater J 2010;24:1184–92.
[103] Caro S, Masad E, Bhasin A, Little D, Sanchez-Silva M. Probabilistic modeling of
the effect of air voids on the mechanical performance of asphalt mixtures
subjected to moisture diffusion. J Assoc Asphalt Pav Technol (AAPT)
2010;79:221–52.
[104] Caro S, Masad E, Sanchez-Silva M, Little D. Stochastic micromechanical model
of the deterioration of asphalt mixtures subject to moisture diffusion
processes. Int J Numer Anal Meth Geomech; 2010.
[105] Graham MA. Viscoelastic–viscoplastic damage model for asphalt concrete.
MSc thesis. Texas A&M University, USA; 2008.
[106] Huang CW. Development and numerical implementation of nonlinear
viscoelastic–viscoplastic model for asphalt materials. PhD dissertation.
Texas A&M University, USA; 2008.
[107] Abu Al-Rub RK, Masad EA, Graham MA. Physically based model for predicting
the susceptibility of asphalt pavements to moisture induced damage. Report
no. SWUTC/10/476660-00012-1. Texas Transportation Institute, USA; 2010.
[108] Dai Q, You Z. Prediction of creep stiffness of asphalt mixture with
micromechanical finite-element and discrete-element models. J Eng Mech
(ASCE) 2007;133(2):163–73.