-
1Graduate Student, University of Wisconsin at Madison 2
Researcher, University of Wisconsin at Madison 3 Former Graduate
Student, University of Wisconsin at Madison 4 Professor, University
of Wisconsin at Madison
Effect of Mineral Filler on Damage Resistance Characteristics of
Asphalt Binders
Ahmed Faheem1, Haifang Wen2, Lawrence Stephenson 3, and Hussain
Bahia4
ABSTRACT
Numerous studies have indicated that the addition of mineral
filler to an asphalt binder increases the stiffness of the binder.
The stiffening ratio and change in rheological properties have
attracted researchers to report data and model the changes due to
physical and sometimes mineralogical nature of fillers. There is,
however, limited information about the effects of fillers on the
damage resistance of binders to permanent strain accumulation and
fatigue. In this study, the effects of filler content and type on
the damage resistance of mastics (filler-binder system) were
investigated. The mastics and binders were tested to evaluate the
effects of type and content of the fillers on fatigue and rutting
performance of mastics. Two binders and two fillers of different
mineralogy, limestone (basic) and granite (acidic), were included
in the study. Two filler contents, 25% and 50%, were used by the
volume of asphalt binder.
Based on the tests results, it is evident that the presence of
fillers significantly increases the complex shear modulus and
fatigue life of binders as compared to those of the base binders.
It is also found that the fatigue life of mastics was significantly
larger than that of binders. The limestone filler was found to have
more positive effects on the fatigue resistance than the granite
filler. For creep and recovery measurements, tests were conducted
at three different temperatures, 52oC, 58oC and 64 oC. The addition
of the fillers enhanced the resistance to rutting, in terms of
total terminal strain and non-recoverable compliance. The
binder-filler interactions need to be considered in estimating the
performance of mastics and asphalt mixture.
KEYWORDS: Mineral Filler, Mastic, Fatigue, Rutting, Shear
Modulus
-
2
BACKGROUND
According to the American Society of Testing and Materials
(ASTM) standard D-242, 70% or larger of the mineral fillers
particles pass the No. 200 sieve (75 m) (1). Typically, for
dense-graded AC mix, a filler-asphalt ratio (F/A) of 0.6-1.2 by
weight is specified (2). Commonly, fillers are considered as part
of the aggregate system. However, observing any asphalt mix, it is
apparent that fillers are actually embedded in the asphalt binder
in such a way that a mastic system (filler + asphalt) is
effectively binding the relatively coarser aggregates. It has been
commonly reported that mineral filler plays an important role in
the construction and performance of hot mix asphalt (HMA) pavements
(3,4,5,6). In addition, the nature and quantity of mineral fillers
are especially important in specialty mixes like stone matrix
asphalt (SMA) mixes in which the mineral filler contributes
significantly to compactibility, impermeability, and in-service
pavement performance (3,4,5,6) .
Numerous researchers have attempted to give guidance on effect
of fillers. As early as 1947, Rigden et al introduced the concept
of fixed and free asphalt in mastics (7). Rigden measured the bulk
volume of compacted dry samples of fillers and considered the
asphalt required to the fill the voids in the dry compacted bed as
fixed asphalt while asphalt in excess of that fixed is considered
as free asphalt (7). Anderson and Goetz (1973) concluded that
different fillers will have different reinforcing effects,
depending on both the nature of filler and the type of asphalt (8).
In an extensive study on dust collector fines, Anderson et al.
(1982) indicated that the nature and extent of the physio-chemical
interaction need to be further studied (9). The authors
specifically called for further study to define the relationship
between dust behavior and heat of immersion measurements.
Different methods have been tried by others to study the effects
of chemical composition of the fillers on mastics and/or asphaltic
mixtures. Kandhal (1981) and Anderson et al. (1982), measured the
pH values of a diluted water solution of fillers and concluded the
pH values could hardly be related to behavior of fillers (9,10).
Dukatz and Anderson (1980) used a cone and plate viscometer to
-
3
construct master curves for several filled asphalts. Their
findings confirmed that mastics may be characterized as linear
viscoelastic materials. However, their measurements indicated that
different types of fillers produced different stiffening effects
and the effects cannot be justified solely on the basis of
gradation of the filler. The authors noted that effects of the
physio-chemical interactions have to be included to offer
reasonable explanations (11).
Anderson et al. (1992) showed that the fillers can change the
failure stresses and concluded that addition of mineral fillers did
not affect the temperature shift factors of the rheological
response. However, the addition of fillers did change the frequency
dependency by shifting the relaxation times to longer times and
stiffening the asphalt (12). In 1997, Kavussi et al. conducted a
study on four different fillers to characterize the role of mineral
fillers in asphalt mixtures. The authors concluded that the higher
the Rigden voids, the more stiffening is the filler. Furthermore,
the viscosity consistently increased with the increase of the
filler content regardless of the physical properties of each filler
type (13).
Only recently, papers were published on the subject of effect of
fillers on failure and damage resistance properties. Kandhal et al.
conducted several tests on the mineral fillers in an attempt to
characterize the fillers and relate the characteristics to their
effects on HMA performance (14). The tests on fillers included
Rigden Voids (British Standards-BS 812), Penn State Modified Rigden
Voids, Particle Size Analysis, Methylene Blue Test (Ohio DOT
Procedure), Plasticity Index (AASHTO T90), and German Filler Test
(Koch Materials Company Procedure). Kandhal et al. indicated that
the influence of the filler on the performance of the HMA can be
best measured as follows (14):
- For permanent deformation, D60 and Methylene Blue are
recommended.
- For fatigue cracking, no test is recommended - For stripping,
D10 and Methylene Blue are
recommended
The above studies show that the interaction between filler and
binder is still not fully understood. This is due to the complex
nature of this interaction. Other studies tried to understand the
role
-
4
of the filler by characterizing the mastic properties, so that
it is understood as a unit made up of two components. Shashidhar et
al. (1998) investigated the factors affecting the stiffening
potential of mineral fillers. In this study, they introduced two
parameters, the maximum packing fraction, and the generalized
Einstein coefficient. The maximum packing fraction is defined as
the maximum amount of fillers that can be added to asphalt without
prompting the emergence of air voids. The generalized Einstein
coefficient is the stiffening rate of the mastics as a function of
filler addition. The results showed that the stiffness of the
mastic can be predicted with the use of these parameters as they
take into account the agglomeration, degree of dispersion, and
asphalt- filler interface contribution (15).
Modeling of asphalt mastic is another step towards understanding
the effect of mineral filler based on the properties of the mastic
as a unit. Abbas et al. (2005) conducted a study to model the
mastic stiffness using discrete element method (DEM) and
micromechanics models. The study aimed to simulate the dynamic
mechanical behavior of asphalt mastics. The DEM results captured
the stiffening behavior of asphalt mastics as a function of the
volumetric concentration of mineral fillers. The DEM results
exhibited a high rate of stiffening that is typically observed in
experimental measurements of mastics at relatively low volume
concentrations of fillers. Compared to the DEM results, the
micromechanics-based models were not sensitive to the dynamic shear
modulus of the asphalt binder, and the models underestimated the
stiffening effect of the mineral fillers (16).
From the above it is clear that the factors affecting the change
in the binder performance due to the introduction of mineral filler
are in need for more research. In this study, the research team has
used advanced testing methods to evaluate the key filler factors
that affect the performance of the mastic in relation to damage
caused by known traffic distresses.
MATERIALS AND TESTING PROGRAM
Two asphalt binders and two mineral fillers were used in this
study. The binders used in this study were of grade PG 70-22,
and
-
5
PG 58-28. The binders were aged in the rolling thin film oven
(RTFO) before adding fillers. The mineral fillers used were
limestone and granite which were selected to represent basic and
acidic mineral aggregates, respectively. The filler-to-asphalt
ratio in the asphalt-filler mastic was 0.25 and 0.5 by volume,
respectively. The ratio was selected to observe the extent to which
the filler will affect the properties of the original binder. The
specific gravity of the granite filler is 2.49, and that of the
limestone is 2.60. The filler contents were selected to cover the
current range of dust to asphalt ratio used in the industry. The
grain size analysis for the fillers was conducted using a LS
Particle Size Analyzer which applies light scattering techniques to
a dilute solution in distilled water. The results of the particle
size analysis are shown in Table 1.
Table 1: Results from LS Particle Size Analyzer on Granite and
Limestone Fillers
Filler Amount Mean Median S.D. C.V. S.S.A. d10 d50 d90 Type % m
m m cm/g m m m
Granite 100 5.53 1.66 10.5 189% 5422 0.6 1.66 14 Granite 100
5.52 1.67 10.4 189% 5432 0.6 1.67 13.9 Granite 100 5.54 1.66 10.6
190% 5412 0.599 1.66 13.9 Granite 100 5.54 1.66 10.6 190% 5412
0.599 1.66 13.9
Limestone 100 6.96 1.58 12.8 183% 4309 0.575 1.58 19.8 Limestone
100 7.09 1.59 12.9 183% 4230 0.576 1.59 20.3 Limestone 100 7.27 1.6
13.2 182% 4125 0.576 1.6 21
(Average) 100 6.21 1.63 11.6 187% 4906 0.589 1.63 16.7
(C.V.) 0.00% 13.60% 2.40% 11.60% 2.00% 13.10% 2.20% 2.40% 20.60%
Maximum) 100 7.27 1.67 13.2 190% 5432 0.6 1.67 21 Minimum) 100 5.52
1.58 10.4 182% 4125 0.575 1.58 13.9
The mastics were prepared by mixing in small containers (150 ml)
the heated binders with a weighed amount of fillers. Filler was
added in steps and mixed thoroughly by hand with a heated spatula.
The resulting batches of mastics were used for all testing (fatigue
and repeated creep) and were kept stored in covered containers at
ambient temperature while not in use. Tests were done in duplicates
to assure repeatability of results.
-
6
Rigden Voids Ratio Tests
In this test method, the volume of the voids in a dry-compacted
bed of mineral dust (Rigden voids) is determined by compacting the
dust in a small mold. The void volume in dry compacted fines
(Rigden voids) is sensitive to changes in gradation and other
properties of the fines, and, therefore, the dry compaction test
has been proposed as a test for monitoring the uniformity of the
fines collected in HMA facilities. Rigden voids can also be used to
estimate the stiffening effect of the fines when mixed with asphalt
cement (17). Figure 1 shows a schematic of the test apparatus.
Figure 1. Schematic of Rigden Voids Test Apparatus (17)
Fatigue Test
For the fatigue test, the base binders and the mastics were
tested under repeated shear cyclic loading at 28oC using a dynamic
shear rheometer (DSR). This temperature was selected because it is
close to the average grade temperature and because the stiffness of
the mastic at this temperature is well within the range of the
optimum rheomoeter stress capability. The complex shear modulus
(G*) and phase angle () as functions of number of loading cycles
were measured (18). The tests were conducted with a sinusoidal
shear stress of 300kPa and a frequency of 10Hz. All the samples
were
-
7
tested at the same stress level and temperature for the sake of
comparison. The fatigue resistance is characterized by the amount
of loading cycles required to drop the complex shear modulus (G*)
of the sample by 50%. The samples were also characterized in terms
of the initial value of the G* multiplied by sin().
Creep and Recovery Test
To evaluate the contribution of filler to the rutting resistance
of the mastic, repeated creep and recovery testing was conducted in
accordance with the procedure of the National Cooperative Highway
Research Program (NCHRP) project 9-10 (18). The loads were applied
for 1 second, followed by 9 seconds of rest period. The accumulated
permanent deformation at the end of the 100th loading cycle was
used to evaluate the rutting resistance of the asphalt mastic. The
samples were subjected to a sequence of loading and unloading at a
shear stress of 100 Pa. The creep and recovery tests were done at
three different temperatures: 52C, 58C, and 64C. The temperatures
52C and 58C represent two high temperatures regions in Wisconsin
while 64C would specifically account for the effects of heavy
traffic and/or slow traffic speeds.
RESULTS AND ANALYSIS
Rigden Test Results
Rigden fractional voids ratio tests were done on granite and
limestone fillers. The literature review has indicated that mineral
fillers have stiffening power directly related to the output of the
Rigden's voids test (13). The results from the Rigden tests are
shown on the bar chart in Figure 2.
-
8
Regiden Voids
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
1 2
Filler Type
Vo
id P
erce
nta
ge
Trial 1
Trial 2
Granite Limestone
Figure 2 Rigden Voids Test Results
Figure 2 shows that the granite filler is consistently showing
higher voids ratio than the limestone filler. The Rigdens void
ratio for the granite filler is 33.36% while limestone filler has a
Rigdens void ratio of 29.51%. It is important to note that the
differences in fractional voids are rather small and could be
considered negligible. It is also recognized that the mineralogy is
significantly different and thus if there is an interactive effects
with the binders, then some difference should be observed in the
stiffening and possible damage accumulation behavior. The following
sections include comparison of the effects of fillers on modulus,
fatigue and repeated creep results of the binders.
Complex Shear Modulus The complex shear modulus of the binders
and mastics were measured in this study as part of the fatigue
testing (initial modulus). To examine the effects of adding filler
to the asphalt, the stiffening power of the filler is determined.
The stiffening power is simply calculated by dividing the value of
the initial complex shear modulus in a fatigue test by that of the
base asphalt. Figure 3 shows the stiffening power due to the
addition of filler. The results indicate that the shear modulus G*
was increased after the mineral fillers were added. The stiffening
power of fillers used in this study, however, is binder-specific.
For PG 70-22 binder, there is no
-
9
difference in stiffening power between granite and limestone
fillers. The increase of filler content from 25% to 50% slightly
increased the stiffening power. For PG 58-28 binder, the stiffening
powers of both fillers are significantly larger than those for PG
70-22. In addition, for PG 58-28 binder, limestone filler has
significantly larger stiffening power than granite filler. Increase
of filler content also significantly increased the stiffening power
of fillers when mixed with PG 58-28 binder.
Stiffening Power of Fillers
0
20
40
60
Filler Content and Type
Stif
fen
ing
Po
wer
PG 70-22PG 58-22
PG 70-22 3.99 5.62 2.88 5.62
PG 58-22 9.61 27.54 16.85 48.01
25% granite 50% Granite 25% Limestone 50% Limestone
Figure 3. Stiffening Power of Fillers Used at 28C.
The stiffing effect of mineral filler has been evaluated in the
literatures by various means. Some studies proposed theoretical
models to predict the stiffening effect of the filler using binder,
as well as filler properties. The MarionPierce model is typically
used to determine the stiffening effect of filler in terms of
viscosity. Huang et al (2007) modified the model to quantify the
stiffening effects of the filler on the complex shear modulus. The
model is as follows (3):
2
** 1
=m
bindermastic GG
(1)
Where, G*binder, binder complex modulus (MPa); G*mastic , mastic
complex shear modulus (MPa); m maximum packing factor; , volume
fraction of filler added to the asphalt. Lesueur and Little (1999)
concluded that most mineral fillers have m of approximately 63%
(3,19).
-
10
Another model used to estimate the stiffening effect of the
filler is the Nelsen Model (1968). This model incorporates set of
properties to predict the filler effect (20):
P
P
m
c
VB
ABV
G
G
+
=1
1 (2)
Where, A=KE-1, and
AGG
GG
B
m
P
m
P
+
=1
(3)
Pm
m V2
11
+= (4)
Gc is the magnitude of complex shear modulus of composite
(mastic); Gm is magnitude of the complex shear modulus of matrix
(binder); Gp is magnitude of the shear modulus of particle
(filler); Vp is the volume fraction of filler; KE is the
generalized generalized Einstein coefficient; and m is the maximum
volumetric packing fraction. The constant B is approximately equal
to 1.0 for a very large Gp /Gm ratio, as in this case the filler
particles are much stiffer than the matrix in the mastics. The
Belgian Road Research Center (BRRC) developed a model to predict
the stiffening effect of the filler in terms of the softening
point. The stiffening effect of the filler is measured through the
ring and ball test (21).
))1(100(
2.1021&
KV
KBR
F += (5)
Where, K = f/b, f = volume-% of filler, b = volume-% of bitumen
and VF = % voids (Rigden).
-
11
Although the model was developed for changes in R&B
softening point, it was later verified for modulus (3). In this
study, equation 5 is modified to use the complex shear modulus
instead of the ring and ball values, following the work by Huang et
al (3):
))1(100(
2.1021*
KV
KG
F += (6)
These three models were used to estimate the complex shear
modulus of the mastics and to compare the estimates with the
measured moduli. Figure 4 and Figure 5 show a comparison of the
measured verses the predicted G* using the three models shown above
for both binders used in the study
Evaluation of Prediction Models for PG 70-22
Nelseny = 8.3524x - 1E+08
BRRCy = 3.3432x - 4E+07
Marion-Piercey = 10.108x - 1E+08
0.E+00
1.E+07
2.E+07
3.E+07
4.E+07
5.E+07
6.E+07
7.E+07
8.E+07
9.E+07
1.E+08
0.E+00 1.E+07 2.E+07 3.E+07 4.E+07 5.E+07 6.E+07 7.E+07 8.E+07
9.E+07 1.E+08
Predicted G*
Act
ual
G*
Marion-PeirceModelNelsen Model
BRRC Model
Line of Equality
Figure 4. Evaluation of Results From Various Prediction Models
for
PG 70-22 Figure 4 shows that all three models give relatively
similar trends. For PG 70-22 binder, at lower filler concentration,
the estimated shear moduli of mastics are very close to the
measured values. At higher concentration, all three models
significantly over estimate the value of the G*. This is
demonstrated by the slope of the linear equations for all models,
as shown in Figure 4. The predicted and
-
12
measured shear moduli of mastics for PG 58-28 are shown in
Figure 5.
Evaluation of Prediction Models For PG 58-28
Nelseny = 1.2317x - 171888
Marion-Piercey = 0.6072x - 193449
BRRCy = 0.2007x + 141177
0.E+00
1.E+06
2.E+06
3.E+06
4.E+06
5.E+06
6.E+06
0.E+00 1.E+06 2.E+06 3.E+06 4.E+06 5.E+06 6.E+06
Predicted G*
Act
ual
G*
Marion-Peirce ModelNelsen ModelBRRC ModelLine of Equality
Figure 5. Evaluation of results form various Prediction Models
for
PG 58-28
Figure 5 paints a different picture when compared to PG 70-22.
Again all models managed to predict the value of G* within small
difference of the actual value, while at higher concentrations the
ability of the models differ. The Nelson model outperforms all the
other models in predicting the value of G*, as the prediction
values were very close to those of the actual ones. On the other
hand, both Marion-Pierce and Nelson models manage to show a slope
that is close to the line of equality, while the Marion-Pierce
model consistently under estimates the value of the G*. The BRRC
model failed to match the results measured from tests. Figure 5
shows that the type of binder plays an important role in the
effectiveness of the prediction models, which in turns, indicate
the importance of including binder-filler interaction in such
models.
Although the data set is small to make the observation
conclusive, it shows that relying on the filler physical properties
only in estimating the stiffening effect of the filler could be
limiting the effectiveness of the prediction models. The
physico-chemical
-
13
interactions between the filler and the binder need to be
represented in any model used for predicting the stiffening power
of mineral fillers. In fact, recent study by BRRC research group (
21) indicates that the BRRC model (Equation5) has already been
modified with a filler reactivity factor to include interactive
effects. As indicated in the background, many other researchers
have observed filler reactivity and recommended modification of
stiffening effects models based on mineralogy and/or binder
chemistry.
Fatigue
The fatigue tests were conducted using a parallel plate DSR to
evaluate the effects of fillers on the resistance to fatigue.
Figure 6 depicts examples of the results for the base binder (no
Filler) and the 50 % filler mastics for binder PG 70-22. The
fatigue life was calculated as the number of cycles to reach a 50%
drop in G*. The results of G* as a function of cycles clearly
indicate significant improvements in the fatigue performance of
asphalt binder after adding the fillers, as seen in Figure 6. The
dropping rate of G* for asphalt binder without filler was much
faster than those of binder with 50% granite and limestone fillers.
Comparing the performance of the granite and limestone mastic, the
G* for 50% granite mastic dropped faster than that of 50% limestone
mastic.
-
14
Comparison of the Fatigue life of Binder and Mastic
0.E+00
5.E+06
1.E+07
2.E+07
2.E+07
3.E+07
0 10000 20000 30000 40000 50000 60000
Number of Cycle
Com
plex
Mod
ulus
(P
a)
50% Granite
50% Limestone
No Filler
Figure 6. Sample Results for Shear Modulus during Fatigue
Tests
The results for the number of cycles to failure for the
different filler concentrations are shown in figure 7 and figure 8.
Adding fillers to the binder results in increasing the resistance
to fatigue. The increase of filler content from 25% to 50% also
increased the number of cycles to failure.
-
15
Change in Fatigue Life For PG 70-22
No Fill
25% Granite
50% Granite
25% Limestone
50% Limestone
0
10000
20000
30000
40000
50000
60000
70000
Num
ber
of C
ycle
s
No Fill 1400
25% Granite 11885
50% Granite 38600
25% Limestone 45200
50% Limestone 56299
Binder
Figure 7 Fatigue Life at Various Filler Contents and Type for
PG
70-22
The results in Figure 7 for PG 70-22 clearly indicate that the
limestone improved the fatigue life more significantly than the
granite filler. This means that although the stiffening effect of
the limestone is similar to the granite, its physical
characteristics are allowing better interaction with the binder and
thus more pronounced resistance for crack initiation and
propagation. However, another explanation that could be offered is
related to the mineralogical property of limestone. It is known the
affinity of binder to the limestone filler is relatively high which
improves adhesion at the interface. This is clearly an interactive
effect that highlights the importance of the adhesive bond created
at the interface with the limestone based filler. This mineralogy
hypothesis is corroborated by many studies on limestone effects
conducted by Little et al. (20). The significance of this
interaction needs to be further evaluated and perhaps modeled to
take into account of using specific interactive fillers to improve
damage resistance. Similar findings were found on the effects of
fillers on PG 58-28, as shown in Figure 8. However, the improved
fatigue resistance of PG 58-28 is not as significant as that of PG
70-22, especially at 25% filler content.
-
16
Change in Fatigue Life For PG 58-28
No Fill
25% Granite
50% Granite
25% Limestone
50% Limestone
0
200
400
600
800
1000
1200
1400
1600
Num
ber
of C
ycle
s
No Fill 461
25% Granite 536.5
50% Granite 841
25% Limestone 613
50% Limestone 1220
Binder
Figure 8 Fatigue Life at Various Filler Contents and Type for PG
58-
28
According to the Superave PG Specification, G*.sin () is an
indicator of the resistance of binder to fatigue cracking. The
results of the fatigue testing collected in this study showed that
the addition of mineral filler consistently generally increased
G*.sin () of the binder. Figures 9 and 10 show that G*.sin () of
the mastic was greatly enhanced due to the addition of mineral
filler depending on the binder type, and filler type and
concentration.
-
17
Effect of filler on G*.sin() for PG 70-22
No Fill
25% Granite
50% Granite25%
Limestone
50% Limestone
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
22000
G*.
Sin
(
) kP
a
No Fill 3657
25% Granite 14309
50% Granite 18278
25% Limestone 15654
50% Limestone 19022
Binder
Figure 9. Effects of Fillers on Measured Values of G*sin for PG
70-
22
For PG 70-22 in Figure 9, when the filler content increased from
25% to 50%, G*.sin () also increased significantly. However, G*.sin
() appears not to be sensitive to filler type, as the values of
G*.sin () are very close at each filler content regardless of the
type.
-
18
Effect of filler on G*.sin() for PG 58-28
No Fill 25% Granite
50% Granite
25% Limestone
50% Limestone
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
G*.
Sin
(
) kP
a
No Fill 939
25% Granite 823
50% Granite 3109
25% Limestone 1313
50% Limestone 3682
Binder
Figure10. Effect of Fillers on Measured Values of G*sin for PG
58-
28
Figure 10 shows that the value of G*sin is not sensitive to the
filler type, similar to the observation made earlier for PG 70-22.
However, Fgures 9 and 10 show the dependency of the filler effect
on the asphalt type. At 25% filler concentration, the value of
G*sin showed no change compared to the unfilled asphalt.
The test results indicated that the measured fatigue performance
was affected by the fillers. In addition, this effect is not simply
dependent of the presence of filler, but it greatly depends on
filler type and concentration as well as the binder type/chemistry.
Furthermore, the G*sin values fail to show the larger effect of the
mineralogy and the superior effects of the limestone compared to
the granite. Considering the nature of fatigue data collected, it
is believed that the fatigue life from the cyclic tests could more
realistically reflect the effect of fillers on the fatigue
performance of materials.
Rutting
Each binder or mastic was tested for creep at three different
temperatures, namely 52oC, 58oC, and 64oC. The terminal strains
were recorded for each filler content. The terminal strain was
-
19
calculated to evaluate the permanent deformation at the
different testing temperatures, and filler contents. Figure 11
shows the comparison of the terminal strains.
Terminal Strain for Different Levels and Types of Fillers
0
2
4
6
8
10
12Te
rmin
al S
tra
in
PG70-22 52C PG70-22 58C
PG70-22 64C PG58-28 52C
PG58-28 58C Binder B 64C
PG70-22 52C 1.7297 0.4857 0.0495 0.2813 0.0461
PG70-22 58C 4.3546 1.0880 0.2200 1.1735 0.1140
PG70-22 64C 10.6050 2.8600 0.5300 2.5742 0.2500
PG58-28 52C 0.7825 0.3520 0.1055 0.3975 0.1855
PG58-28 58C 1.9550 0.8430 0.2045 1.0150 0.3970
Binder B 64C 4.3650 1.9000 0.4415 2.3200 0.9085
No Filler 25% granite 50% granite 25% lime 50% lime
Figure 11 Terminal Strains for Different Filler Content and
Type
The overall results prove that the existence of the fillers
enhanced the binders capability of resisting the accumulation of
permanent deformation at all three testing temperatures. As
expected, the terminal strain at any given filler content increased
with the increase of the test temperature. With the increase of
filler content, the binders resistance to rutting was greatly
increased. Looking at the relative values, it is also seen that the
binder-filler interaction and the mineralogy do not have
significant influence. The speculation is that the physical
stiffening effect is much higher than the effects at the interface
because binders are very soft and the rigidity of the fillers
controls the stiffening.
The non-recoverable compliance also demonstrated the effects of
filler on the rutting performance of binder, as shown in Figures 12
and 13.
-
20
PG 70-22 - Non Recoverable Compliance Jr (1/Pa)
0.E+00
5.E-05
1.E-04
2.E-04
2.E-04
3.E-04
3.E-04
Test Temp
Jr (
1/P
a)
25% granite
50% granite
25% limestone
50% limestone
25% granite 5.00E-05 1.06E-04 2.80E-04
50% granite 5.00E-06 1.30E-05 4.90E-05
25% limestone 2.30E-05 1.12E-04 2.52E-04
50% limestone 1.00E-06 1.00E-05 2.00E-05
52 58 64
Figure 12. Non-recoverable Compliance Test Results for PG
70-22
In Figure 12 for PG 70-22, with the increase of test
temperature, the non-recoverable compliance also increased. When
the same filler content is used, the mastic with limestone filler
had less non-recoverable compliance than mastic with granite
filler, except for 25% filler content at 58oC. When the filler
content increased from 25% to 50%, the non-recoverable compliance
was significantly reduced. Therefore, the non-recoverable
compliance seems to be sensitive to both the filler type and
content. The results for PG 58-28 are shown below in Figure 13.
PG58-28 - Non Recoverable Compliance Jr (1/Pa)
0.E+00
5.E-05
1.E-04
2.E-04
2.E-04
3.E-04
3.E-04
Test Temp
J r (
1/P
a)
25% granite
50% granite
25% limestone
50% limestone
25% granite 3.50E-05 8.50E-05 1.90E-04
50% granite 1.30E-05 2.00E-05 4.50E-05
25% limestone 4.50E-05 1.00E-04 2.50E-04
50% limestone 1.80E-05 4.00E-05 8.50E-05
52 58 64
Figure 13. Non-recoverable Compliance Test Results for PG
58-28
In Figure 13, PG 58-28 shows a similar behavior to that for PG
70-22, except that the limestone mastics show higher
non-recoverable compliance consistently at all temperatures than
granite mastics.
-
21
Summary and Conclusions
Two mineral fillers, limestone and granite, were added to two
asphalt binder and their effects on damage resistance
characteristics were investigated using fatigue and creep recovery
tests. It was found that:
An increase in filler content resulted in an increase in G*
(stiffness of the mastic), as expected.
The examination of three prediction models highlighted the need
to include the interaction between the binder and filler in such
models to be able to accurately predict the mastic performance.
The fatigue life (under stress-controlled conditions) of the
mastic is improved significantly when using both types of fillers.
However, the limestone causes more improvement than that of the
granite. This can be ascribed to the mineralogical property of
filler.
The use of mineral filler significantly improved the mastics
resistance to the accumulation of permanent deformation. The
effects, however, are not highly dependent on the filler
mineralogy.
In the past, the role of mineral fillers on damage resistance
characteristics were not well studied and there is only limited
information published on the subject. The findings from this study
shed some lights on this important topic.
References
1- Standard Specification for Mineral Filler for Bituminous
Paving Mixtures, American Society of Testing and Materials,
Designation 242, 2000.
2- Roberts, F. L., P.S. Kandhal, E. R. Brown, D-Y Lee, and T. W.
Kennedy, Hot Mix Asphalt Materials, Mixture Design, and
Construction, National Center for Asphalt Technology, Second
Edition, 1996.
-
22
3- B. Huang, X. Shy, and X. Chen, Effects of mineral fillers on
hot-mix asphalt laboratory-measured properties. International
Journal of Pavement Engineering, Vol. 8, pp 1-9, 2007
4- Prithvi S Kandhal, Cynthia Y. Lynn, Characterization tests
for mineral fillers related to performance of asphalt paving
mixtures, Transportation Research Record Nov, 1998 1638 101-110
5- A. Tayebali, G. Aalpass, N. Khosla, Effect of Mineral Filler
Type and Amount on Design and Performance of Asphalt Concrete
Mixtures, Transportation Research Record: Journal of the
Transportation Research Board, No. 1609, Transportation Research
Board of the National Academies, Washington, D.C., 1998, pp.
36.
6- L A Cooley Jr,, E. R. Brown, D. E. Watson,Evaluation of OGFC
Mixtures Containing Cellulose Fibers, National Center for Asphalt
Technology, 2000
7- P. J. Rigden, The Use of Fillers in Bituminous Road
Surfacing. A Study of Filler-Binder Systems In Relation To Filler
Characteristics. Journal of the Society of Chemical Industry,
September, 1947
8- D. Anderson, and W. H. Goetz, "Mechanical behavior and
reinforcement of mineral filler-asphalt mixtures." Proc.,
Association of Asphalt Paving Technologists, Vol. 42, 3766.
1973.
9- D. Anderson, J.P. Tarris, and D. Brock, Dust Collector Fines
and Their Influence on Mixture Design. Journal of the Association
of Asphalt Paving Technologists, Vol. 51, 1982
10- P. S. Kandhal, Evaluation of Baghouse Fines in Bituminous
Paving Mixtures. Journal of the Association of Asphalt Paving
Technologists, Vol. 50, 1981
11- E.L Dukatz, and D.A. Anderson, The Effect of Various Fillers
on the Mechanical Behavior of Asphalt and Asphaltic Concrete.
Journal of the Association of Asphalt Paving Technologists, Vol.
49, 1980
12- D.A. Anderson, H.U. Bahia, and R. Dongre, Rheological
Properties of Mineral Filler - Asphalt Mastics and its
-
23
Importance to Pavement Performance. American Society for Testing
and Materials, STP 1147, 1992
13- A. Kavussi, and R. G. Hicks, Properties of Bituminous
Mixtures Containing Different Fillers, Asphalt Paving Technology,
Volume 66, pp 153, 1997
14- P.S. Kandhal, and F. Parker, Jr. NCHRP Report 405: Aggregate
Tests Related to Asphalt Concrete Performance in Pavements,
Transportation Research Board, National Research Council,
Washington, DC, 1998.
15- N. Shashidhar, P. Romero, Factors Affecting the Stiffening
Potential of Mineral Fillers, Transportation Research Record:
Journal of the Transportation Research Board, No. 1638,
Transportation Research Board of the National Academies,
Washington, D.C., 1998, pp. 94.
16- A. Abbas, E. Masad, T. Papagiannakis, A. Shenoy, Modeling
asphalt mastic stiffness using descrete element analysis and
micromechanics-based models, International Journal of Pavement
Engineering, Vol. 6, June 2005, pp. 2005
17- Method for Determining Volume of Voids in Compacted Filler
or Fines, Kansas Test Method KY-64, Kansas Department of
Transportation, 2005.
18- H. U. Bahia, D. I. Hanson, M. Zeng, H. Zhai, M. A. Khatri,
Characterization of Modified Asphalt Binders in Superpave Mix
Design, National Cooperative Highway Research Program, Report 459,
2001.
19- D. Lesueur, D. Little, effect of hydrated lime on rheology,
fracture, and aging of bitumen, Transportation Research Record
1661, 1999
20- Dallas N. Little, J Claine Petersen, Unique Effects of
Hydrated Lime Filler on the Performance Related Properties of
Asphalt Cements: Physical and Chemical Interactions Revisited,
Journal of Materials in Civil Engineering. Vol. 17, no. 2, pp.
207-218. Mar.-Apr. 2005
21- S.Vansteenkiste, Properties of Fillers, Belgian Road
Research Center, 2005. Presentation.
Table of Contents