New Parameter to Evaluate Moisture Damage of

Asphalt-Aggregate Bond in Using Dynamic

Shear Rheometer
Dong-Woo Cho1 and Hussain U. Bahia2
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Most of the techniques used today to evaluate practically moisture damage in asphalt pavements are involved in compacted
asphalt mixtures. These tests generally simulate field conditions, and provide a performance-related parameter of moisture effects by
applying mechanical techniques. No test of the compacted mixtures, however, can isolate the fundamental and rheological properties of
the asphalt-aggregate bond, which hinder proper understanding and quantification of the moisture damage phenomenon. To avoid these
complications and improve the evaluation of asphalt-aggregate interfacial properties in moist situations, this study proposes a new
experimental method. In this method, dynamic shear rheometer was used to measure asphalt-aggregate interfacial properties. A degree of
moisture damage relating to the interfacial properties was then quantified using linear viscoelastic concepts to propose a meaningful
parameter, called wet to dry yield shear stress ratio. Moreover, possible factors such as film thicknesses, conditioning times, and chemical
effects were studied in terms of the W/D YSS ratio.
DOI: 10.1061/共ASCE兲0899-1561共2010兲22:3共267兲
CE Database subject headings: Asphalt pavements; Damage; Viscoelasticity; Rheology; Parameters; Moisture.
Author keywords: Pavement; Asphalt; Moisture damage; Rheology; Viscoelasticity.

Introduction balance between analyticity and simplicity in moisture damage

tests that use either loose or compacted asphalt mixtures. The
Moisture damage has long been a major concern for pavement motivation of this research, therefore, is to develop a simple and
durability, which is one of the principal factors accelerating major analytical asphalt-aggregate test method using a dynamic shear
distresses such as rutting, fatigue, and thermal cracking 共Asphalt rheometer 共DSR兲 for evaluating moisture effects. The DSR is
Institute 1987; Majidzadeh and Brovold 1968兲, and drives the widely used for characterizing asphalt binders and is required for
high cost of pavement maintenance and rehabilitation. Conse- specifications in many countries. Under well-controlled condi-
quently, testing methods to measure moisture damage or index a tions of temperature, modes of loading, time, and moisture, the
degree of the damage have been interested in many research DSR accurately measures the rheological responses of asphalt
areas. These testing methods typically consist of loose and com- films adhering to aggregate’s surfaces. The measured responses
pacted asphalt mixture tests in accordance with the type of include linear viscoelastic 共LVE兲 properties, which are developed
samples used 共Solaimanian et al. 2002兲. The tests using loose in the Strategic highway research program 共Anderson et al. 1994;
asphalt mixtures are based on the visual inspection of aggregates
Petersen et al. 1994兲, and the rheological responses within non-
coated with an asphalt binder. Unlike the loose asphalt mixture
linear viscoelastic 共NLVE兲 regions by applying a stress sweep
tests, the tests using compacted asphalt mixtures usually provide a
protocol. Fig. 1 shows the characteristics of the LVE and NLVE
mechanical index which is obtained by measuring and comparing
regions based on the stress sweep. The stress sweep has been used
mechanical properties before and after moisture exposure. These
mechanical tests developed can generally simulate environmental recently by a few researchers and claimed to be a better indicator
conditions, and provide a performance-related parameter of mois- of binder resistance to damage 共Delgadillo et al. 2006兲. The tests
ture effects 共Bausano and Williams 2006兲. conducted in this research follow the basic protocols and analysis
Despite their advantages, it is difficult to achieve a proper methods of the SHRP binder characterization. A major modifica-
tion introduced is using rock disks 25 mm in diameter by 5 mm
1
Senior Researcher, Highway Research Div., Korea Institute of Con- thick, which are bonded on the metal spindle and bottom plate of
struction Technology 共KICT兲, Ilsanseo-Gu, Goyang-Si, Gyeonggie-Do, the DSR. This modification was applied to create easy water ac-
Republic of Korea 共corresponding author兲. E-mail: dongwoocho@kict. cess to the asphalt-aggregate interface. The preliminary results of
re.kr; [email protected] this experiment were reported and a provisional parameter of
2
Professor, Dept. of Civil and Environmental Engineering, Univ. of
LVE, wet to dry 共W/D兲 shear stress ratio 共the LVE limit, in stress
Wisconsin-Madison, Madison, WI. E-mail: [email protected]
Note. This manuscript was submitted on April 30, 2008; approved on values, of wet condition divided by that of dry condition兲 was
October 27, 2009; published online on February 12, 2010. Discussion proposed for indexing moisture susceptibility of the asphalt-
period open until August 1, 2010; separate discussions must be submitted aggregate bond, or interface 共Cho and Bahia 2007; Cho 2008兲.
for individual papers. This paper is part of the Journal of Materials in This paper is upon the continuous work to verifying and convinc-
Civil Engineering, Vol. 22, No. 3, March 1, 2010. ©ASCE, ISSN 0899-
ing the proposed parameter.
1561/2010/3-267–276/$25.00.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010 / 267

J. Mater. Civ. Eng. 2010.22:267-276.

 Stress Strain loading rate, film thickness, and temperature controls. Most im-
Period (Cycle) portantly, it allows a fundamental study of moisture effects where
moisture effects are isolated and not confounded by mixture
Stress or Time Scale
Strain Scale voids, variable film thicknesses, and different aggregate sizes.
Super Glue gel epoxy 共Super Glue Corporation, Rancho Cu-
camonga, CA.兲 was used to bond cored rock disks to the upper
Delayed Time
Stress Amplitude
Strain Amplitude
spindle and the bottom plate of DSR. The epoxy’s temperature
range is −36 to 149° C, and its maximum strength is 10 MPa. For
(a) Oscillatory Stress Input and Strain Output water conditioning of the interface between a rock disk and as-
phalt binder, a copper water cup was fabricated whose dimension
is 50 mm in inner diameter and 60 mm outer diameter, and 27
Nth Period Final Period
at LVE Stress Amp.
mm outer in height and 20 mm inner. To clean easily the used
Stress Amp. This Slope is epoxy with heat in an oven, an old spindle was used, whose
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

Non-Linear G*
cylinder part can easily be screwed in and out of the shaft part but
This Slope is Linear G*
is secure enough screwed in. Fig. 2 shows the setting and cleaning
(Complex Modulus) of materials and equipment.
st
1 Period Nth Period

Strain Amp. Strain Amp

LVE Region (Stress Independence) NLVE Region (Stress Dependence) Materials and Equipment
(b) Linear and Non-Linear G*(Complex Modulus) at Stress Sweep
Two aggregate types were selected: 共1兲 limestone 共LS兲—13% po-
rosity, 97% carbonate lime and 共2兲 sandstone 共SS兲—16% poros-
LVE Limit: The Maximum Limit to Keep the ity, 96% quartz, and 2% feldspar. The LS and SS were used
Structural Durability, Measured at % of LVE G*
because they have similar texture and porosity although their min-
G* Scale: eralogy is different. The two aggregates were selected because
there is consensus in the literature that LS has better moisture
= Stress Amp.
Strain Amp. LVE Region resistance than SS 共Ishai and Yuval 2002; Huang et al. 2005; Kim
et al. 2008兲. The porosities of these LS and SS were estimated
LVE G*: Measured at
100 Pa Shear Stress
from measuring water saturation. In addition, three different as-
phalt binders were selected for testing with similar performance
NLVE Region Stress Scale grades 共PG兲: 共1兲 M58-28: an unmodified asphalt binder; 共2兲
(c) LVE and NLVE Region at Stress Sweep M58-28+ EVA: modified with ethyl vinyl acetate 共EVA兲; and 共3兲
SBS PG64-28: mixed from a different source with styrene buta-
Fig. 1. Linear region, nonlinear region, and LVE limit diene styrene 共SBS兲.

Experimental Design Experimental Plan

The experimental design is based on modification of the standard
Table 1 shows the overall testing plan used for conducting the
DSR test. Cored rock disks 25 mm in diameter and 5 mm thick
DSR moisture damage test in this research. The variables for this
are used as the substrates for adhering asphalts. The disks are
research are classified into environmental and material variables.
glued on the DSR metal spindle and the base metal plate as
A detailed description of preparation and procedures for this ex-
shown in Fig. 2. The disks and an asphalt binder simulate the
periment is described in the following sections.
asphalt-aggregate interfaces in asphalt mixtures. The water cup,
fabricated specially for the DSR 共Paar Physica MCR 300兲, is used
to allow continuous water access to the interfaces. A stress sweep
facilitates measuring changes in the rheological properties of the Experimental Conditions and Testing Protocol
asphalt binder as it responds to gradually increasing stress condi-
tions. Using the DSR in this study has many advantages, such as Basic Testing Conditions Used in the Stress Sweep
Test
1. Sample setting and trimming at 85° C.
2. Wet conditioning temperature at 40° C.
3. Oscillation load for conditioning time at 1% shear strain, 1.6
Hz 共or 10 rad/s兲 frequency, and 9-min interval.
4. Stress sweep test at 200 points from 10 to 50,000 Pa with
lamp-log profile, and 10 Hz frequency.

Description of Dynamic Shear Rheometer Moisture

Damage Testing Protocol
Fig. 2. Setup and cleaning for newly developed DSR moisture dam- 1. Asphalt sample preparation: three different asphalt binders—
age test M58-28 共PG58-28兲, M58-28+ EVA 共PG64-22兲, and SBS

268 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 1. Experimental Plan for This Research
Environmental variables Material variables
Testing Film thickness Conditioning time Contact
task 共mm兲 共h兲 Solution material Asphalt binder Replicate
T1 1 6 Tap water Sandstone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
Limestone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
T2 1 12 Tap water Sandstone M58-28 3
M58-28+ EVA 3
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

SBS PG64-28 3
Limestone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
T3 0.1 6 Tap water Sandstone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
Limestone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
T4 1 6 0.5-M sodium chloride solution with tap water Sandstone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
Limestone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
T5 1 6 0.5-M calcium chloride solution with tap water Sandstone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3
Limestone M58-28 3
M58-28+ EVA 3
SBS PG64-28 3

PG64-28—were prepared for each testing condition 共T1, T2, g. The combinations of samples prepared are summarized
T3, T4, and T5兲. in Table 2.
a. For each testing condition, twelve molded samples 2. Rock disk sample preparation:
were prepared from each asphalt binder. Therefore, the a. Two aggregate types 共SS and LS兲 were selected.
test consists of 5 testing conditions ⫻3 asphalt binders b. Rock cylinders 25 mm in diameter were first produced
⫻12 molded samples= total 180 samples. by coring rock blocks using a coring machine, and then
b. All 12 molded samples of an asphalt binder were pre- the rock cylinders were cut using a sawing machine
pared from either Can1 or Can2, which is saved from into 5-mm-thick disks.
one asphalt source 共a quart size兲 at 163° C. c. The cut disks were flattened using a lapping machine
c. To make the 12 samples, M58-28, an unmodified as- with 70-␮m diamond surfaces.
phalt binder, was heated for 30 min at 163° C, and 3. Equipment and testing preparation:
M58-28+ EVA and SBS PG64-28, modified asphalt a. On the measuring station of the DSR, the water cup is
binders, for 1 h at 163° C. installed, and a bottom plate is firmly screwed into the
d. The samples for 1-mm film thickness and 6-h condi- water cup using a wrench.
tioning time 共T1, T4, and T5兲 were picked from Can1, b. To epoxy the rock disks and DSR:
and the samples for 0.1-mm film and 12 h 共T2 and T3兲 • First, a zero gap is set onto the bottom plate-installed
from Can2. water cup. The zero gap setting is used for epoxying
e. Among the 12 samples, six samples were used for the and testing because the loading speed and force can be
SS combination and another six samples for the LS better controlled with gap-setting parameters. The gap-
combination. setting parameters selected for this research are shown
f. Among the six samples of each SS and LS combina- in Fig. 3.
tion, three samples were used as three replicates for dry • Second, epoxy is put on a rock disk and then the
stress sweep tests and another three samples as three epoxyed-rock disk is slightly pressed down between the
replicates for wet stress sweep tests. bottom plate and DSR spindle by the zero gap work.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010 / 269

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 2. Summary of Combinations of Sample Preparation
Heating cycle Prepared sample
Sandstone combination Limestone combination
Testing task Asphalt binder Can 1 Can 2 Total sample Dry Wet Dry Wet
T1 M58-28 1 12 3 3 3 3
M58-28+ EVA 1 12 3 3 3 3
SBS PG64-28 1 12 3 3 3 3
T2 M58-28 1 12 3 3 3 3
M58-28+ EVA 1 12 3 3 3 3
SBS PG64-28 1 12 3 3 3 3
T3 M58-28 2 12 3 3 3 3
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

M58-28+ EVA 2 12 3 3 3 3
SBS PG64-28 2 12 3 3 3 3
T4 M58-28 2 12 3 3 3 3
M58-28+ EVA 2 12 3 3 3 3
SBS PG64-28 2 12 3 3 3 3
T5 M58-28 3 12 3 3 3 3
M58-28+ EVA 3 12 3 3 3 3
SBS PG64-28 3 12 3 3 3 3
Note: Heating cycle: 1⫽first heating with a can and 2⫽second heating with the same can.

Then, the rock disk should be centered as precisely as parallel along the flat bottom of the plate and spindle,
possible using a DSR spatula. This epoxyed bottom the zero gap-setting parameters for this research are
composite, thereafter, is statically stationed for provid- shown in Fig. 3.
ing enough partial curing time 共at least 30 min兲 but not c. After finishing the equipment preparation, the epoxyed
too much. composites are heated at 85° C for 30 min for sample
• Third, after the curing is finished, another zero gap preparing and trimming. As a tip, before sample prepar-
should be carried out to adjust the position against any ing and trimming, it is advised to wind Teflon tape
changes from the first epoxying work. around the sides of rock disks to avoid asphalt coating
• Finally, the second step is repeated for the DSR spindle the sides of the rocks.
and another rock disk. This second rock disk should be d. After preparing and trimming the synthesized asphalt-
centered parallel to the bottom plate epoxyed-rock disk. rock sample using the same standard procedures, it is
Unlike the bottom work, this step requires full curing heated at 85° C for 15 min to ensure adhesion.
time 共at least 6 h兲 for both top and bottom composites. e. The fixed heating time, 30 min for preparing and trim-
The first picture in Fig. 4, the testing setup, shows how ming and 15 min for adhesion, is applied to all tests to
the top and bottom composites should look when they have the same effects of thermal condition during
are complete. preparation.
• As a precaution, when the rock disks are epoxyed with 4. Stress sweep in dry and wet conditions 共dry stress sweep and
the bottom plate or the spindle, the normal force control wet stress sweep兲:
option of the zero gap setting should be used and set at
a. After finishing the testing preparation, 40° C is set as
a very slow speed. The slow movement and normal
the testing temperature and then, in dry conditions, the
force control setting results in maintaining the disks
stress sweep is run after 30-min waiting time to ensure
a temperature equilibrium state. The fixed 30-min wait-
ing time is applied to all tests to have the same thermal
and stabilizing effects.
b. After finishing the dry test, another sample is prepared
and trimmed on the same rock disk. To reduce super-
fluous asphalt near the perimeter of the disk, it is rec-
ommended to use small samples, though they should be
properly quantified for testing geometry. Here, the Te-
flon tape is removed to provide water or a chemical
solution access before it is poured.
c. Tap water or a chemical solution is poured into the
water cup in which the testing sample is installed, and
the testing sample is conditioned for 6 or 12 h at 40° C
before running a wet stress sweep.
d. Although distilled or deionized water is generally best
suited for chemical experiments, tap water is used for
practical purposes in this experiment. This is because
Fig. 3. Setting parameters for epoxying and testing tap water is a chemical mixture and therefore provides

270 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010

J. Mater. Civ. Eng. 2010.22:267-276.

 (a) Testing Setup (b) Dry-Condition (c) Wet-Condition (d) Failed Sample

Fig. 4. DSR moisture damage test procedures

enough chemical effect to the sample to save testing a potential LVE limit to separate between LVE and NLVE regions
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

time. The quantity of tap water required for all tests is 共Cho and Bahia 2007; Cho 2008兲. The shear stresses met at each
decided upon and taken at a fixed day and place for percentage were selected for dry-conditioned and wet-conditioned
quality control. stress sweeps, and the W/D ratio of shear stresses at the same
e. During the conditioning procedure, the load should percentage was calculated. A statistical comparison for the shear
continuously oscillate to gently stir the sample with 1% stress responses was then carried out on all the data in T1, T2, T3,
strain, 1.6-Hz frequency, and at 9-min intervals for ob- T4, and T5 tasks. As a statistical measure to compare shear
serving changes in the rheological properties. stresses at the three different percentages, coefficient of variation
f. After reaching the given conditioning time, a wet stress 共COV兲 关SD divided by average 共Avg.兲兴 was chosen because the
sweep is run at the same temperature as the condition- dispersion relative to the selected shear stresses is important for
ing temperature, 40° C. this analysis 共Ang and Tang 2007兲. In the statistical analysis, the
g. Fig. 4 shows the testing procedures of the DSR mois- lowest COV value regarded as having relatively less dispersion
ture damage test. was selected among the W/D ratios at 95, 90, and 85%. Table 3
tabulates the COV. values at 85, 90, and 95% for each task; the
bold numbers indicate the lowest COV values among these per-
Results and Analysis centages.
Table 4 highlights the percentage LVE Gⴱ at which each task
Having a linear relation between stress and strain provides con- has the lowest COV. As 85% of LVE Gⴱ had the lowest COV 19th
venience, and is adequate for practical engineering problems, be- times, more than the other percentages, 85% LVE Gⴱ was selected
cause a constant modulus 共in this case Gⴱ = complex modulus兲 in as the best choice for the shear stress comparison. This shear
the relation can be used to estimate engineering performance rel- stress at 85% LVE Gⴱ will be hereafter called “yield shear stress”
evant to stress and strain. In contrast, characterizing and analyz- 共YSS兲 by considering NLVE condition as a yielding state.
ing a nonlinear relation between stress and strain is difficult In this research, to more generalize W/D YSS ratio at 85%
because of the stress or strain dependence and the fact that every LVE Gⴱ, the W/D YSS ratio was compared with the W/D LVE Gⴱ
material follows a different trend 共Anderson et al. 1994兲. For the ratio that may be a potential parameter to index moisture damage;
initial analysis using the DSR, the LVE limit was defined as the Tables 5 and 6 list the values calculated from all tasks. While the
stress or strain at which the Gⴱ measured decreases to 95% of the W/D LVE Gⴱ ratios, LVE Gⴱ 共at 100-Pa shear stress兲 of wet
initial value 共Cho and Bahia 2007兲; later, an optimized percentage condition divided by that of dry condition, stay near the 1.0 value,
of 85% was chosen through the selection process described in the the W/D YSS ratio shows more sensitive results between minerals
following subsection. This LVE limit is based on the same defi- and between tasks. Therefore, with the proven sensitivity, the
nition used by researchers as the upper limit of the LVE region W/D YSS ratio is affirmed for the analytical parameter to evaluate
共Anderson et al. 1994兲. After affirming the effects of the LVE moisture effects and used in the data analysis of the planned ex-
limit relating to moisture damage, a new parameter of the LVE
periment. The analysis of the W/D YSS ratio in each task follows
limit was developed. Using this new parameter, the results of
in the next sections.
planned tasks were analyzed quantitatively and qualitatively. In
the wet testing conditions, the failure shapes existing within the
interface between an asphalt binder and its sandstone disks are all Study of Conditioning Time „Tasks 1 and 2…
very similar as a clear debonding shape of the asphalt binder from
the sandstone disks. However, the interface between an asphalt Before starting the analysis of conditioning time, some back-
binder and its limestone disks shows mostly necked shapes of the ground on the used tap water is needed. The tap water used was
asphalt films when the stresses are in NLVE. not based on rigorous scientific experiment. If distilled water had
been used for this research, 6 or 12 h may not have been enough
to see visible or measurable results. This is because pure water
Selection of Shear Stress at 85% Linear Viscoelastic does not react as quickly with asphalt and mineral. Therefore, tap
Gⴱ and W/D YSS Ratio water, which is chemically mixed, was used to save effort and
At a fixed percentage of LVE Gⴱ, the rheological responses of a time. Thus, if this experiment is repeated it is important to note
material can be compared with the responses of another material that results may vary due to locally and seasonally picked tap
at the same percentage of LVE Gⴱ. With this basic idea in mind, waters where the ratio of water to chemicals is different. With this
this research selected three fixed percentages, 95, 90, and 85% of understanding, the analysis of conditioning time begins here.
LVE Gⴱ, as the base for comparing shear stresses. In addition, Table 7 and Fig. 5 summarize the measured and calculated results
these potential points, selected by the fixed percentages, could be of Tasks 1 and 2; the analysis of this data follows:

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010 / 271

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 3. COV Values for Each Task
COV
Percent
Rock disk Asphalt LVE Gⴱ T1 T2 T3 T4 T5
type binder type 共%兲 共%兲 共%兲 共%兲 共%兲 共%兲
Sandstone M58-28 95 34 19 9 57 39
90 27 22 23 50 15
85 25 25 30 50 19
M58-28+ EVA 95 27 50 35 24 57
90 28 26 21 8 36
85 30 21 18 2 29
SBS PG64-28 95 27 28 27 20 15
90 23 29 23 9 7
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

85 21 31 25 8 13
Limestone M58-28 95 9 13 17 61 18
90 2 10 24 47 12
85 2 17 27 44 11
M58-28+ EVA 95 14 0 7 39 33
90 4 2 9 35 7
85 2 4 17 34 6
SBS PG64-28 95 6 35 47 24 36
90 5 17 34 12 17
85 4 10 21 10 9
Note: Bold numbers indicate the lowest COV value among the COV values of 95, 90, and 85% LVE Gⴱs in the asphalt binder type.

1. Phase-angle is insensitive between 6- and 12-h wet condi- creasing or decreasing according to the conditioning times.
tionings. 4. W/D YSS ratio is sensitive in mineral types of SS and LS
2. Gⴱ cannot be the indexing parameter to compare moisture because the W/D YSS ratios of LS are always higher than
effects. For example, although M58-28+ EVA on SS has those of SS irrespective of the conditioning times.
much higher Gⴱ values of 212 and 230 kPa, it shows a simi- This analysis shows that the effects of conditioning time, in
lar or weaker bonding strength than other samples with a this research, are insignificant, though this result may be attrib-
W/D YSS ratio of 0.23. uted to the reaction rate of the tap water used. That is, most
3. W/D YSS ratio is insensitive between 6- and 12-h condition- reactions between the tap water, asphalt, and minerals may be
ings because the W/D YSS ratio has no pattern such as in- complete in 6 h, and after 6 h few or no reactions may exist.
Therefore, caution regarding conditioning time should be used in
virtue of the reaction rate and chemicals.
Table 4. Selection of Percentage
Test combinations Total number
Percentage selected as having of the lowest Study of Film Thickness „Tasks 1 and 3…
of LVE Gⴱ the lowest COV COVs
Asphalt film thicknesses are related to durability in hardening and
85% SS+ M58-28− T1 , T4 19 aging 共Read and Whiteoak 2003; Roberts et al. 1996兲 and can
SS+ M58-28+ EVA− T2 , T3 , T4 , T5 range from 5 to 15 ␮m in an asphalt mixture 共Read and Whit-
SS+ SBS PG64-28− T1 , T4 eoak 2003兲. The range from 5 to 15 ␮m is produced upon real
LS+ M58-28− T1 , T4 , T5 contact between aggregates so that in that range not only the
LS+ M58-28+ EVA− T1 , T4 , T5 asphalt-aggregate interface but also aggregate-aggregate friction,
LS+ SBS PG64-28− T1 , T2 , T3 , T4 , T5 lubrication, and wear are a real concern. The science dealing with
90% SS+ M58-28− T5 4 this friction, lubrication, and wear is called tribology 共Czichos
SS+ M58-28+ EVA− none 1978兲. The tribological effects are beyond the scope of this re-
SS+ SBS PG64-28− T3 , T5 search, because the simplified nature of the experiment cannot
LS+ M58-28− T2
simulate real contact between aggregates. Therefore, a just mea-
surable film thickness, 0.1 mm, was chosen without tribological
LS+ M58-28+ EVA− none
effect. Table 8 and Fig. 6 compare the results between 1 and
LS+ SBS PG64-28− none
0.1-mm film thicknesseses for Tasks 1 and 3; the analysis of this
95% SS+ M58-28− T2 , T3 7
data follows:
SS+ M58-28+ EVA− T1
1. Most results for the thin film thickness 共0.1 mm兲 show
SS+ SBS PG64-28− T2 weaker durability against moisture effects except EVA modi-
LS+ M58-28− T3 fied asphalt. This is especially true for LS. These results are
LS+ M58-28+ EVA− T2 , T3 reasonable because the asphalt-aggregate interface will expe-
LS+ SBS PG64-28− none rience more impact of torsion as the thickness is thinner and
Note: Boldface 19 indicates the largest one among the total numbers 共19, thus the interfacial structure is weaker. The interfacial struc-
4, 7兲 of the lowest COVs in the 85, 90, and 95% LVE Gⴱ, respectively. ture on SS was substantially damaged on both 1- and 0.1-mm

272 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 5. W/D LVE Gⴱ Ratio of Each Task
W/D LVE Gⴱ ratio
Dry LVE Gⴱ Average W/D LVE ratio 共wet/dry兲
Rock disk Asphalt binder Average Standard COV Standard COV
type type Task 共kPa兲 deviation 共%兲 Average deviation 共%兲
SS M58-28 T1 143 3.45 2.4 1.05 0.029 3
T2 146 7.90 5.4 1.20 0.093 8
T3 92 6.47 7.0 1.07 0.169 16
T4 150 4.92 3.3 1.08 0.067 6
T5 160 5.01 3.1 1.19 0.023 2
M58-28+ EVA T1 231 21.13 9.1 1.09 0.034 3
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

T2 229 16.37 7.1 1.18 0.065 6

T3 99 22.42 22.6 1.15 0.102 9
T4 249 0.57 0.2 1.09 0.027 2
T5 313 23.94 7.6 1.06 0.039 4
SBS PG64-28 T1 137 7.55 5.5 1.08 0.068 6
T2 110 1.91 1.7 1.19 0.049 4
T3 58 7.66 13.2 1.19 0.053 4
T4 132 2.43 1.8 1.13 0.016 1
T5 142 11.70 8.2 1.14 0.006 0
LS M58-28 T1 134 19.28 14.4 1.04 0.022 2
T2 149 1.40 0.9 1.13 0.042 4
T3 82 5.75 7.0 1.07 0.077 7
T4 153 0.95 0.6 1.02 0.026 3
T5 160 12.07 7.5 1.06 0.045 4
M58-28+ EVA T1 226 10.09 4.5 1.01 0.050 5
T2 234 4.21 1.8 1.15 0.029 3
T3 141 9.00 6.4 1.20 0.153 13
T4 306 19.73 6.4 1.01 0.029 3
T5 317 9.94 3.1 1.02 0.066 6
SBS PG64-28 T1 106 12.32 11.6 1.28 0.061 5
T2 117 2.24 1.9 1.14 0.066 6
T3 73 18.48 25.3 1.19 0.120 10
T4 153 2.97 1.9 1.06 0.045 4
T5 166 8.63 5.2 1.08 0.046 4

film thicknesses so that the difference between them is insig- Study of Chemical Effects „Tasks 1, 4, and 5…
nificant. However, the structure on LS was not damaged as
much, so that the impact of torsion may play a more impor- Sodium chloride 共NaCl兲 and calcium chloride 共CaCl2兲 are com-
tant role for determining moisture effects. monly used as icing melter to melt ice on highways during the
2. EVA modified asphalt binders show a different trend from winter season. Therefore, the study of the effects of these chemi-
other asphalt binders. In this case, 0.1 mm of M58-28 cals on asphalt materials is important for practical purposes. For
+ EVA on SS shows an increased result of 0.44 substantially this study, 0.5 mol concentration was arbitrarily selected and tap
higher than the 1-mm result of 0.23. In addition, the differ- water was used as the mixing solvent. Table 9 and Fig. 7 compare
ence between 0.1 and 1 mm on LS is also insignificant if results between the tap water, NaCl solution, and CaCl2 solution
compared with other results. This result may be related to the in Tasks 1, 4, and 5; the analysis of these data follows.
crystallization of polyethylene in EVA; asphalt and SBS are 1. Most results of the NaCl solution indicate the lowest W/D
not crystallizable. The crystallization in polyethylene may YSS ratio among the solutions. Thus NaCl provides adverse
give the structural strength relatively independent of the film effects in the asphalt-aggregate interface. This could be at-
thickness, and tribological effects may work between the tributed to the quick ionization of NaCl.
crystallized structure and aggregates. 2. The results of the CaCl2 solution on the SS show the highest
From this analysis, the effect of film thickness is significant. W/D YSS ratio among the solutions. In addition, the results
However, the generalization of this result is limited because tri- on the LS appear tolerable because the ratio is near 1.0. Thus
bological effects were not studied in this research. It is possible CaCl2 has a tolerable to beneficial effects on the asphalt-
that tribological effects could also be significant in determining aggregate interface. This could be attributed to the ion satu-
the strength of the asphalt mixture. ration and nonsolubility of calcium.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010 / 273

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 6. W/D YSS Ratio of Each Task
W/D YSS ratio
Dry YSS W/D YSS ratio
Rock disk Asphalt binder Average Standard COV Standard COV
type type Task 共kPa兲 deviation 共%兲 Average deviation 共%兲
SS M58-28 T1 18.2 0.45 2.5 0.23 0.059 25
T2 18.4 1.19 6.5 0.41 0.102 25
T3 35.5 1.52 4.3 0.21 0.063 30
T4 18.2 0.88 4.8 0.20 0.102 50
T5 17.4 0.85 4.9 0.53 0.104 19
M58-28+ EVA T1 30.9 2.80 9.1 0.23 0.071 30
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

T2 26.0 1.30 5.0 0.23 0.048 21

T3 26.1 3.88 14.9 0.44 0.079 18
T4 29.2 3.00 10.3 0.20 0.005 2
T5 32.7 2.37 7.2 0.69 0.200 29
SBS PG64-28 T1 27.9 0.69 2.5 0.50 0.106 21
T2 21.2 0.00 0.0 0.45 0.137 31
T3 33.1 0.82 2.5 0.38 0.097 25
T4 25.2 1.08 4.3 0.41 0.035 8
T5 27.5 1.18 4.3 1.02 0.130 13
LS M58-28 T1 17.7 1.14 6.4 1.22 0.030 2
T2 19.0 1.25 6.6 1.09 0.185 17
T3 33.1 2.19 6.6 0.34 0.092 27
T4 19.5 1.41 7.2 0.74 0.327 44
T5 18.5 1.77 9.6 1.16 0.126 11
M58-28+ EVA T1 29.9 1.28 4.3 1.15 0.029 2
T2 29.1 1.42 4.9 1.14 0.049 4
T3 31.7 1.54 4.9 1.13 0.193 17
T4 33.6 3.05 9.1 0.79 0.270 34
T5 32.7 2.79 8.5 0.99 0.064 6
SBS PG64-28 T1 26.4 1.99 7.5 1.24 0.053 4
T2 24.2 1.03 4.3 1.37 0.139 10
T3 42.2 3.19 7.6 0.77 0.162 21
T4 28.3 1.83 6.5 1.11 0.113 10
T5 29.5 0.72 2.4 1.06 0.093 9

Table 7. Results Comparing Conditioning Time

W/D YSS ratio Average

Asphalt Conditioning Film Average Gⴱ at phase angle
Rock disk binder time thickness Dry YSS Average Standard deviation wet YSS at wet YSS
type type 共h兲 共mm兲 Solution 共kPa兲 W/D YSS ratio of W/D YSS ratio 共kPa兲 共deg兲
SS M58-28 6 1 Tap 18.2 0.23 0.059 127 74
12 1 Tap 18.4 0.41 0.102 147 73
M58-28+ EVA 6 1 Tap 30.9 0.23 0.071 212 66
12 1 Tap 26.0 0.23 0.048 230 67
SBS PG64-28 6 1 Tap 27.9 0.50 0.106 126 69
12 1 Tap 21.2 0.45 0.137 111 68
LS M58-28 6 1 Tap 17.7 1.22 0.030 118 74
12 1 Tap 19.0 1.09 0.185 143 73
M58-28+ EVA 6 1 Tap 30.0 1.15 0.029 195 68
12 1 Tap 29.0 1.14 0.049 228 68
SBS PG64-28 6 1 Tap 26.4 1.24 0.053 115 69
12 1 Tap 24.2 1.37 0.139 114 70

274 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010

J. Mater. Civ. Eng. 2010.22:267-276.

 1.8 1.4

W/D Ratio of YSS (with Max and Min error bar)

W/D Ratio of YSS (with Max and Min error bar) 1.6
1.2
1.4
1
1.2

1 0.8

0.8 0.6

0.6
0.4
0.4
0.2
0.2

0 0
6hrs 12hrs 6hrs 12hrs 6hrs 12hrs 6hrs 12hrs 6hrs 12hrs 6hrs 12hrs 1mm 0.1mm 1mm 0.1mm 1mm 0.1mm 1mm 0.1mm 1mm 0.1mm 1mm 0.1mm

M58-28 M58-28+EVA SBS PG64-28 M58-28 M58-28+EVA SBS PG64-28 M58-28 M58-28+EVA SBS PG64-28 M58-28 M58-28+EVA SBS PG64-28
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

SS LS SS LS

Fig. 5. Comparison between 6- and 12-h wet conditionings Fig. 6. Comparison between 1- and 0.1-mm film thickness

From this analysis, the effect of deicing materials is significant moisture damage. Another potential limitation is that for ag-
even in tap water. However, a further study may need to be con- gregates with an extremely low porosity, the test setup might
ducted to determine what quantities of these materials are best in not be applicable because the small pores of aggregates will
practice because the arbitrarily selected 0.5 M was not scientific. not permit water to access the asphalt-aggregate interface in
Also, instead of solutions using tap water, competent chemical a reasonable time.
solutions using distilled water may be needed to have a better 2. The developed parameter, defined as W/D YSS ratio 共=wet-
reference point for reproducibility. conditioned YSS/ dry-conditioned YSS兲, is sensitive en-
ough to evaluate moisture effects of the asphalt-aggregate
combinations used in this research. Because of this limit re-
lating to using materials, more tests should be run with vari-
Conclusions ous asphalts and aggregate types for the further research.
3. The use of LVE Gⴱ as a performance-related parameter is
This research program was performed to develop a simple test shown to be insensitive to moisture effects. This raises some
and functional parameter for evaluating the effect of aggregates’ concerns regarding the common practice of using LVE rhe-
surface and water conditioning on the rheology of asphalts films. ology to evaluate moisture effects. Thus caution should be
Discussions and conclusions of this process and the results fol- exercised regarding the use of LVE rheology to evaluate
low: moisture effects on the asphalt-aggregate bond.
1. The DSR moisture damage test can evaluate moisture effects 4. Identifying a standard wetting solution for conditioning is
in the asphalt-aggregate interaction and can differentiate the necessary for further study. The standard wetting solution
properties between different materials. Since this test was should be reactive and representative enough of the actual
developed as a simplification of the asphalt-aggregate inter- pavement application.
action, its application to real asphalt mixtures may be lim- 5. For relating asphalt mixture performances to the results of
ited. The key benefit of this test is its ability to isolate the DSR moisture damage test, tribological effects may be con-
physical and chemical effects of the interactions from mix- sidered to complement the study of the unknown and inex-
ture variables which commonly confound the evaluation of plicable factors between aggregates.

Table 8. Results Comparing Film Thickness

W/D YSS ratio
Asphalt Conditioning Film
Rock disk binder time thickness Dry YSS Average Standard deviation Average Gⴱ Average phase angle
type type 共h兲 共mm兲 Solution 共kPa兲 W/D YSS ratio of W/D YSS ratio at wet YSS 共kPa兲 at wet YSS 共deg兲
SS 6 M58-28 1 Tap 18.2 0.23 0.059 127 74
6 0.1 Tap 35.5 0.21 0.063 84 72
M58-28+ EVA 6 1 Tap 30.9 0.23 0.071 212 66
6 0.1 Tap 26.1 0.44 0.079 96 66
SBS PG64-28 6 1 Tap 27.9 0.50 0.106 126 69
6 0.1 Tap 33.1 0.38 0.097 58 69
LS M58-28 6 1 Tap 17.7 1.22 0.030 118 74
6 0.1 Tap 33.1 0.34 0.092 74 72
M58-28+ EVA 6 1 Tap 29.9 1.15 0.029 195 68
6 0.1 Tap 31.7 1.13 0.193 145 66
SBS PG64-28 6 1 Tap 26.4 1.24 0.053 115 69
6 0.1 Tap 42.2 0.77 0.162 73 69
Note: FT⫽film thickness; CT⫽conditioning time; Sol.⫽solution; and Tap⫽tap water.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010 / 275

J. Mater. Civ. Eng. 2010.22:267-276.

 Table 9. Results Comparing Chemical Effects
W/D YSS ratio
Conditioning Film Average
Rock disk Asphalt time thickness Dry YSS Average Standard deviation Average Gⴱ phase angle
type binder type 共h兲 共mm兲 Solution 共kPa兲 W/D YSS ratio of W/D YSS ratio at wet YSS 共kPa兲 at wet YSS 共deg兲
SS M58-28 6 1 0.5M NaCl 18.2 0.20 0.102 137 72
6 1 Tap water 18.2 0.23 0.059 127 74
6 1 0.5M CaCl2 17.4 0.53 0.104 161 71
M58-28+ EVA 6 1 0.5M NaCl 29.2 0.20 0.005 231 65
6 1 Tap water 30.9 0.23 0.071 212 66
6 1 0.5M CaCl2 32.7 0.69 0.200 283 61
SBS PG64-28 6 1 0.5M NaCl 25.2 0.41 0.035 128 68
6 1 Tap water 27.9 0.50 0.106 126 69
Downloaded from ascelibrary.org by Universiti Sains Malaysia on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.

6 1 0.5M CaCl2 27.5 1.02 0.130 138 64

LS M58-28 6 1 0.5M NaCl 19.5 0.74 0.327 132 72
6 1 Tap water 17.7 1.22 0.030 118 74
6 1 0.5M CaCl2 18.5 1.16 0.126 145 73
M58-28+ EVA 6 1 0.5M NaCl 33.6 0.79 0.270 262 65
6 1 Tap water 30.0 1.15 0.029 195 68
6 1 0.5M CaCl2 32.7 0.99 0.064 275 64
SBS PG64-28 6 1 0.5M NaCl 28.3 1.11 0.113 137 69
6 1 Tap water 26.4 1.24 0.053 115 69
6 1 0.5M CaCl2 29.5 1.06 0.093 152 68

tion using the dynamic shear rheometer.” Ph.D. dissertation, Univ. of

Wisconsin, Madison, Wis.
W/D Ratio of YSS (with Max and Min error bar)

1.4 Cho, D. W., and Bahia, H. U. 共2007兲. “Effect of aggregates’ surface and
1.2 water on rheology of asphalt films.” Transportation Research Record.
1998, Transportation Researcch Board, Washington, D.C., 10–17.
1
Czichos, H. 共1978兲. Tribology—A systems approach to the science and
0.8 technology of friction, lubrication and wear, tribology series, 1,
0.6 Elsevier Science, New York.
Delgadillo, R., Cho, D. W., and Bahia, U. H. 共2006兲. “Nonlinearity of
0.4
repeated creep and recovery binder test and the relationship with mix-
0.2 ture permanent deformation.” Transportation Research Record. 1962,
0
Transportation Research Board, Washington, D.C., 3–11.
Huang, S., Robertson, R. E., Branthaver, J. F., and Petersen, J. C. 共2005兲.
NaCl+Tap

Tap

Tap
CaCl2+Tap

NaCl+Tap

Tap

CaCl2+Tap

NaCl+Tap

CaCl2+Tap

NaCl+Tap

Tap

CaCl2+Tap

NaCl+Tap

Tap

Tap
CaCl2+Tap

NaCl+Tap

CaCl2+Tap

“Impact of lime modification of asphalt and freeze-thaw cycling on

the asphalt-aggregate interaction and moisture resistance to moisture
M58-28 M58-28+EVA SBS PG64-28 M58-28 M58-28+EVA SBS PG64-28 damage.” J. Mater. Civ. Eng., 17共6兲, 711–718.
SS LS Ishai, I., and Yuval, R. 共2002兲. “Reformulation of asphalt cements for
paving.” J. Trans. Eng., 128共2兲, 111–122.
Fig. 7. Comparison between NaCl, tap water, and CaCl2 solution Kim, Y. R., Lutif, J. S., Bhasin, A., and Little, D. N. 共2008兲. “Evaluation
of moisture damage mechanisms and effects of hydrated lime in as-
phalt mixtures through measurements of mixture component proper-
ties and performance testing.” J. Mater. Civ. Eng., 20共10兲, 659–667.
Majidzadeh, K., and Brovold, F. N. 共1968兲. “State of the art: Effect of
References water on bitumen-aggregate mixture.” Special Rep. No. 98, National
Research Council, Washington, D.C.
Anderson, D. A., et al. 共1994兲. “Binder characterization and evaluation Petersen, J. C., et al. 共1994兲. “Binder characterization and evaluation.”
volume 3: Physical characterization.” Strategic highway research pro- Strategic highway research program (SHRP), Vol. 1, National Re-
gram (SHRP), National Research Council, Washington, D.C. search Council, Washington, D.C.
Ang, A. H., and Tang, W. H. 共2007兲. Probability concepts in Read, J., and Whiteoak, D. 共2003兲. The Shell Bitumen handbook, 5th Ed.,
engineering—Emphasis on applications to civil and environmental Thomas Telford, London.
engineering, 2nd Ed., Wiley, New York. Roberts, F. L., Kandhal, P. S., Brown, E. R., Lee, D., and Kennedy, T. W.
Asphalt Institute. 共1987兲. “Cause and prevention of stripping.” Asphalt 共1996兲. Hot mix asphalt materials, mixture design, and construction,
pavement educational series No. 10, 2nd Ed., College Park, Md. 2nd Ed., NAPA Education Foundation, Lanham, Md.
Bausano, J., and Williams, R. C. 共2009兲. “Transitioning from AASHTO Solaimanian, M., Harvey, J., Tahmoressi, M., and Tandon, V. 共2002兲.
T283 to the simple performance test using moisture conditioning.” J. “Test methods to predict moisture sensitivity of hot-mix asphalt pave-
Mater. Civ. Eng., 21共2兲, 73–82. ments.” Proc., Moisture Sensitivity of Asphalt Pavement: A National
Cho, D. W. 共2008兲. “Study of asphalt-aggregate bond in simple simula- Seminar, Transportation Research Board, Washington, D.C., 77–109.

276 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / MARCH 2010

J. Mater. Civ. Eng. 2010.22:267-276.