Anda di halaman 1dari 8

Construction and Building Materials 31 (2012) 212–219

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials


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

Differentiating between damage and thixotropy in asphalt binder’s fatigue tests


F.E. Pérez-Jiménez, R. Botella ⇑, R. Miró
Technical University of Catalonia, Department of Transport and Regional Planning, C/Jordi, Girona 1-3, B1 215, 08034 Barcelona, Spain

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

Article history: When testing asphalt binders under cyclic conditions, a drop in modulus is observed. This decrease has
Received 11 October 2011 been explained by an increase in the temperature of materials due to energy dissipation, thixotropy, and
Received in revised form 22 December 2011 damage. This study presents the application of a cyclic uniaxial tension–compression test to analyze the
Accepted 23 December 2011
causes and effects of that behavior on the fatigue properties of asphalt binders.
Available online 25 January 2012
The authors conducted both time and strain sweep tests and concluded that the steady-state complex
modulus depended only on the strain applied and was not influenced by the prior loading history, i.e.
Keywords:
both data sets converged to the same complex modulus values after a short number of cycles when
Asphalt binder
Fatigue
applying the same strain. In addition, by increasing and decreasing strain alternatively it was found that
Strain sweep the initial complex modulus drop is mostly caused by thixotropic effects rather than irreversible damage
Thixotropy and can be recovered. The recovery of material properties was faster and complete when the binder was
Uniaxial tension–compression test mechanically forced to its initial state instead of allowing it to rest (healing).
Dissipated energy density Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction The problem with tests using the DSR comes when testing low
penetration graded binders or employing low test temperatures. In
The viscoelastic properties of bituminous mixtures make them these cases, it is very difficult to avoid detachment between the as-
the most common paving materials used worldwide. The constitu- phalt binder and the equipment steel plates [9,13]. Furthermore,
ent material that provides these properties is the asphalt binder, when applying low strain levels long time periods are required
i.e. a byproduct of the refining process of crude oil in petroleum to obtain the number of cycles to failure.
refineries [1,2]. One of the major forms of distress affecting bitumi- As a result, some authors are currently developing new test pro-
nous pavement performance is fatigue cracking, where the asphalt cedures to obtain the fatigue law of asphalt binders in a quicker
binder plays a crucial role [3–8]. and simpler way. Among them are strain sweep tests, in which
The fatigue behavior analysis of asphalt binders is typically loading cycles are applied with increasing strain amplitude. These
based on obtaining their fatigue law by cyclic strain-controlled tests are performed to obtain the complex modulus evolution with
tests. The most common device used for that purpose is the strain. Moreover, using viscoelastic continuum damage models
Dynamic Shear Rheometer (DSR) [9–11]. This equipment applies (VECDs), the asphalt binder’s fatigue law can be estimated [14].
shear cycles at constant strain amplitude and records the shear Traditionally, it is stated that the evolution of complex modulus
stress necessary to maintain this strain level. When the shear stress during strain-controlled fatigue tests undergoes three different
drops to half the initial value, it is assumed that the specimen has stages or phases [15].
failed, and a number of cycles to failure is thus obtained for a given In phase I, a sudden drop in complex modulus is observed,
strain level. The performance of several tests at different strain which is often explained by an increase in the temperature of the
amplitudes allows obtaining as many pairs of strain-cycles values material due to the energy released during the test, an initial adap-
as tests performed. This type of test method is known as time sweep tation and a time-dependent change in viscosity, also known as
test. By plotting these data points together, a relationship is found thixotropy [16–18].
between the strain and number of cycles, Eq. (1) [12]: Thixotropy is defined as the continuous decrease of viscosity
with time when flow is applied to a sample that has been previ-
c ¼ a  N b ; ð1Þ ously at rest and the subsequent recovery of viscosity in time when
the flow is discontinued [19]. Bituminous materials are known to
where N is the number of cycles to failure, c is the applied strain, experience thixotropic effects. For this reason, some authors have
and a and b are the fitting parameters (i.e. different for each binder). studied the influence of thixotropy on the behavior of these mate-
rials [18,20–23].
⇑ Corresponding author. In phase II, the modulus remains constant or decreases linearly
E-mail address: ramon.botella@upc.edu (R. Botella). with the number of cycles. Di Benedetto et al., state that fatigue

0950-0618/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.12.098
F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219 213

damage in the form of microcracks takes place during this stage


[15].
In phase III, the complex modulus drops suddenly, leading to
complete failure of the specimen. What is believed to happen is
that at some point the microcracks generated in phase II intercon-
nect, producing macrocracks that make the specimen fail [15].
When testing asphalt mixtures incorporating ductile or modi-
fied binders, the 50% stiffness loss criterion sometimes leads to
the wrong number of cycles to failure in strain-controlled fatigue
tests [24]. These types of mixtures lose stiffness as the fatigue test
progresses but their specimens show almost no damage when
their stiffness modulus drops to half the initial value.
Knowledge of asphalt binder behavior change due to tempera-
ture, thixotropy or damage is very helpful in evaluating fatigue
behavior and adjusting VECD models. For this reason, a series of Fig. 2. Example of the displacement input signal in a two cycle per step increasing
strain sweep test.
fatigue tests using both time and strain sweeps were carried out.
In some strain sweep tests, strain was equally increased and de-
creased to observe the complex modulus recovery and differentiate
between damage and change in behavior.

2. Test methods

2.1. Test setup

The test procedure consisted of a cyclic uniaxial tension–compression test.


Specimens were 20 mm diameter, 39.5 mm high cylinders of asphalt binder
fabricated using a cylindrical mold. The asphalt binder was heated in the oven
before pouring into the mold. Specimens were allowed to cool to ambient
conditions and then cooled in the environmental chamber at 10 °C for at least
24 h. After removing from the mold, they were glued to a servohydraulic press
using a commercial cyanoacrylate glue. The glue applied hardened in less than
15 min, but the waiting period between gluing and testing was established in 1 h
to assure good binding. The test equipment applied a sinusoidal displacement
signal and recorded the resultant load. Strain applied was calculated using the fol-
lowing equation:

D Fig. 3. Example of the strain semi-amplitude evolution in both the increasing and
e¼ ; ð2Þ the up & down strain sweep tests.
h

where D is the displacement and h is the height of the specimen, which was fixed at
39.5 mm.
All tests were carried out at 10 °C and 10 Hz. The final test setup is shown in
Fig. 1.

2.2. Time sweep test

Time sweep tests are the most common fatigue test method. They consist of
applying a constant cyclic amplitude displacement or strain signal to the test
specimen until failure takes place. These tests allow to obtain the fatigue law of
the material using Eq. (1), as explained in Section 1. Their main drawback is the
large amount of test time and number of tests required to calculate a reliable
fatigue law.

Fig. 4. Computation of the complex modulus using the ellipse shaped test data.

In these tests, the complex modulus and the dissipated energy density decrease
with the number of cycles applied, particularly in the first cycles. After that, they
remain nearly constant or slightly decreasing linearly until failure, causing a rela-
tively sudden drop in these parameters. Failure is often assumed to occur when
the stiffness modulus drops below half the initial value.

2.3. Increasing strain sweep test

This test procedure consists of applying a sinusoidal displacement signal (ten-


sion–compression) to the specimen at a frequency of 10 Hz. A 60 lm of displace-
ment amplitude is applied for a specified number of cycles; then the amplitude
becomes 60 lm higher (120 lm total amplitude) for the same number of cycles,
and the process repeats itself. The test ends when complete failure of the specimens
takes place. For this study, the strain step duration was fixed in 5000 cycles. Figs. 2
Fig. 1. Uniaxial tension–compression cyclic test setup. and 3 show an example of the displacement input signal and the evolution of the
214 F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219

Fig. 5. Complex modulus versus number of cycles for all binders.

strain semi-amplitude1 per cycle throughout the test, respectively. This test proce- 1
S ¼ jðx1  y2 þ x2  y3 þ . . . þ xn1  yn þ xn  y1 Þ  ðy1  x2 þ y2  x3 þ    þ yn1  xn þ yn  x1 Þj;
dure was called EBADE test, after the Spanish words for strain sweep test (Ensayo 2
de BArrido de DEformaciones). ð4Þ

2.4. Up & down strain sweep test where S is the area enclosed by the data points in [x]  [y] units and xi, yi are the coor-
dinates of the n data points arranged clockwise or counterclockwise.
The up & down strain sweep test was performed to evaluate the recovery of
the complex modulus of the binder. The test starts the same as the increasing
strain sweep test. When the second strain step is over, after 10,000 cycles, the 3. Results
strain amplitude decreases to the initial value, i.e. that applied in the first step,
and the specimen is allowed to rest for 30 min. Then, the test starts again from 3.1. Test materials
the first step to the second and third and again strain is decreased step by step
to the initial value. This process is repeated, each time achieving higher strain
steps. To clarify the concept, Fig. 3 shows the evolution of the strain amplitude
Four different binders were employed, i.e. B 60/70, B 40/50,
throughout the test. B 13/222 and B 60/70 after aging by Rolling Thin-Film Oven Test
(RTFOT)3 [25,26], hereafter referred to as B 60/70 RTFOT.
2.5. Test parameters
3.2. Time sweep tests and fatigue laws
The most important parameters that can be computed during the test are: the
stress semi-amplitude, rmax, complex modulus, jE⁄j, and dissipated energy density,
DED, during each cycle. Due to the delay between stress and strain, an ellipse is The results from in the time sweep tests are summarized in
formed in the stress vs. strain plot. Using the maximum stress, rmax, and strain Fig. 5. The evolution of the complex modulus with the number of
semi-amplitudes, emax, it is possible to obtain the complex modulus, jE⁄j, applying cycles is plotted for each strain amplitude applied. All tests were
Eq. (3):
conducted at 10 °C and 10 Hz.
rmax Specimens tested at 0.08% strain semi-amplitude strain neither
jE j ¼ : ð3Þ showed an initial drop in complex modulus nor failed before
emax
2,000,000 cycles. By contrast, all tests carried out at higher strains
Fig. 4 illustrates the complex modulus computed using Eq. (3).
The dissipated energy density is proportional to the area of the ellipse described
showed the three fatigue stages. In phase I the complex modulus
in the tension–compression graph. To compute this area from the test data, Gauss
area formula was employed (Eq. (4)). 2
The two numbers represent the penetration grade range at 25°C in dmm for each
binder. Therefore, lower grades imply harder binders.
1 3
The strain semi-amplitude is half the strain amplitude (peak to valley) and it is This aging procedure is described in the ASTM standard D 2872-04 and is used to
the value plotted in all the graphical representations. simulate the aging process undergone by asphalt binders during mixture fabrication.
F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219 215

strain semi-amplitude at which the dissipated energy density


dropped to zero; at this point the complex modulus was almost
zero as well.
Comparison of the complex modulus and number of cycles
curves for all the binders tested, Fig. 7, revealed that B 13/22 had
the highest initial modulus and failed at the lowest strain ampli-
tude, 350 MPa and 0.61%, respectively (7th strain step). The other
three binders exhibited a similar behavior among them, although
B 60/70 RTFOT had a slightly higher initial modulus and failure
strain amplitude, 170 MPa and 0.76% respectively (10th strain
step). Therefore, this asphalt binder has a slightly better fatigue
behavior than B 40/50 and B 60/70. These results, from strain
sweep tests, point to the same direction as those obtained using
the conventional method based on time sweep tests to character-
ize fatigue behavior.
Regarding the dissipated energy density, Fig. 7, this magnitude
increased with the strain amplitude applied and within each step
showed the same behavior as the complex modulus, that is an ini-
Fig. 6. Fatigue laws obtained from the time sweep test data.
tial drop followed by a stabilization at lower strain amplitudes.
At the same strain level, harder binders achieved higher dissi-
pated energy densities. Hence, at the first steps B 13/22 obtained
Table 1 the highest values. However, this binder was also the first to fail.
Numerical values for the parameters calculated in the strain sweep tests. On the other hand, B 60/70 RTFOT reached the highest strain
 amplitude before failing, and thus, it had the highest dissipated en-
Binder Failure strain (%) E (MPa)
1st
DEDmax (J/m3)
ergy density. B 60/70 and B 40/50 exhibited a similar behavior at
B 13/22 0.53 345 5061
the first steps but failed at lower strain amplitudes than B 60/70
B 40/50 0.68 174 4476
B 60/70 0.68 146 4470 RTFOT.
B 60/70 RTFOT 0.76 180 4951 The dissipated energy density plots illustrate the strain ampli-
tude at which the binder failed since at this point this parameter
dropped to zero immediately.
decreased rapidly with the number of cycles, in phase II it re-
mained constant or decreased slightly with a much higher number
of cycles, and finally in phase III it dropped to almost zero in a short
period of time. The initial drop in complex modulus and the final
failure described in phase III did not occur for tests performed at
very low strain levels (0.08%) and asphalt binders B 13/22 and
B 60/70, Fig. 5.
B 40/50 and B 60/70 failed at around 1,000,000 cycles at 0.15%
strain. The number of cycles to failure show by B 13/22 at this
strain level was around 150,000 cycles. To reach the 1,000,000 cy-
cles with this binder, a test was carried out at 0.11% strain. B 60/70
and B 40/50 had lower initial complex moduli, between 120 and
150 MPa, while B 13/22 and B 60/70 RTFOT had initial complex
moduli between 300 and 370 MPa and 156 and 170 MPa,
respectively.
The fatigue laws were obtained by plotting strain versus num-
ber of cycles to failure in a double logarithmic scale, see Fig. 6.
The difference between B 13/22 and the other three binders can
be clearly seen. Its fatigue law was at the lowest position in the
graph and had the lowest slope. As for the softest binders, B 60/
70 RTFOT exhibited a slightly better behavior than the other two.
Its complex modulus was higher and its fatigue law was at the
highest position in the graph. In spite of having different penetra-
tion grades B 60/70 and B 40/50 had very similar fatigue laws.

3.3. Increasing strain sweep tests

In the increasing strain sweep tests, the complex modulus de-


creased with the number of cycles and the strain amplitude. Dur-
ing each strain step (constant strain amplitude) the complex
modulus loss tended to stabilize as it did in the time sweep tests.
Table 1 contains the average values for the three most important
parameters, failure strain,
average of the complex modulus during
the first strain step, E1st , and maximum dissipated energy density Fig. 7. Complex modulus and dissipated energy density versus number of cycles for
reached during the test, DEDmax. Failure strain was defined as the all binders.
216 F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219

The increasing strain sweep test does not directly provide the density. It seemed that an increase in strain amplitude did not
fatigue law of the material but it is a simple and quick procedure cause, the previous cycles to affect the binder’s phase II complex
to evaluate its fatigue behavior. Two important parameters can modulus or dissipated energy density at the given strain ampli-
be obtained: the initial complex modulus at low strain levels and tude. That would be the same as saying that for a given strain
the strain semi-amplitude at which the material fails. When test- amplitude (fixed temperature and frequency) there was a stabiliza-
ing asphalt binders with the same initial complex modulus, higher tion value for complex modulus and dissipated energy density
maximum strain semi-amplitudes imply better fatigue behaviors. independent of the previous loading history.
Furthermore, the maximum value of dissipated energy density The questions that arise are: what would happen if the strain
during a test can be used to compare binders with different initial amplitude was decreased instead of increased? Would the binder
complex moduli. recover its complex modulus? To answer these questions the
In order to design a bituminous mixture, it is necessary to con- authors designed the up & down strain sweep test described in
sider the modulus and ductility that must be provided by the bin- Section 2.4.
der. These two properties are closely related to the parameters
obtained in the increasing strain sweep test. 3.5. Up & down strain sweep test

3.4. Time sweep versus strain sweep When this test was first performed on B 40/50, its complex
modulus decreased as the strain amplitude increased, in this case
In order to compare the results obtained in the time and strain until the 5th step. However, when the strain amplitude started to
sweep tests, data from both tests were plotted on the same graph decrease, the complex modulus returned to the same values as at
for each binder. Fig. 8 summarizes these results for B 13/22 and the previous strain amplitude. Finally, when the first strain step
B 60/70 RTFOT. was again reached, the complex modulus had the same value as
The most interesting observation about these graphs is that at the beginning of the test. After allowing the specimen to rest
when the analyzed parameter stabilized during a strain step, the for 30 min, the test was repeated and the same result was found,
value at which it stabilized was extremely close to the stabilization Fig. 9. Therefore, the complex modulus loss observed when
value (phase II) obtained in the time sweep tests performed at the increasing strain was not caused by damage but by thixotropy
same strain amplitude. In other words, the complex modulus at resulting from binder’s viscoelastic behavior (Section 1).
which the binder stabilized during each strain step in the strain Apparently there were two phenomena affecting the complex
sweep tests was not affected by the previous strain amplitudes ap- modulus variation. One had and instant effect when the strain
plied to it. A similar situation happened to the dissipated energy amplitude was increased or decreased while the other was related

Fig. 8. Complex modulus and the dissipated energy density for B 13/22 and B 60/70 RTFOT from time and strain sweep test data.
F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219 217

Fig. 9. First up & down strain sweep test performed on B 40/50. Complex modulus and dissipated energy density data.

Fig. 10. Up & down strain sweep test performed on B 13/22. Complex modulus Fig. 11. Up & down strain sweep test performed on B 60/70. Complex modulus
data. data.

with the time dependent viscous drag force that increased or de- graph for the same test clarifies this concept, Fig. 9. Note the expo-
creased with the number of cycles as the binder reached a steady nential decay and the logarithmic rise when the strain was in-
state for the given strain amplitude. The dissipated energy density creased and decreased, respectively.
218 F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219

 Specimen damage and failure occur when the material under-


goes strain levels close to the strain semi-amplitudes that cause
complete failure in increasing strain sweep tests.
 During the time sweep test, strain distribution may not be
totally homogeneous in the whole specimen; hence, at some
points strain can reach the critical value. Crack propagation
could start at those points and then progress throughout the
fracture area. In addition, higher strain levels lead to higher
probabilities of crack propagation and therefore to lower num-
bers of cycles to failure.
 The fatigue-related properties afforded by a given asphalt bin-
der added to a bituminous mixture can be evaluated by analyz-
ing the initial complex modulus and maximum strain semi-
amplitude reached in increasing strain sweep tests. Comparison
Fig. 12. Fracture surface belonging to one of B 13/22 specimens. of similar initial complex modulus binders shows that the one
with the highest failure strain would be the most suitable for
a fatigue resistant asphalt mixture.
To validate these results, tests were conducted for all binders,  Healing is often defined as the partial recovery of complex mod-
starting with two-step high triangles and as high as necessary to ulus upon resting after subjecting the binder to degradation
make the specimen fail, Figs. 10 and 11. The test performed on (damage). However, recovery was complete and much quicker
B 60/70 revealed that even when the complex modulus dropped when the binder was mechanically forced to its initial state.
below half the initial value, it returned to its initial value when
the strain amplitude was decreased. Higher strain amplitudes led
to rapid failure of the specimens. Failure strains in the up & down References
tests were very close to those obtained in the increasing strain
sweep tests. [1] Krishnan JM, Rajagopal K. On the mechanical behavior of asphalt. Mech Mater
Since the material is not totally homogeneous, the strain distri- 2005;37(11):1085–100.
[2] Masad E, Tashman L, Little D, Zbib H. Viscoplastic modeling of asphalt mixes
bution in the specimen is not uniform. Thus, at some point strain with the effects of anisotropy, damage and aggregate characteristics. Mech
can be higher than the average strain applied to the whole speci- Mater 2005;37(12):1242–56.
men, and even higher than the failure strain. Then, from that point [3] Boussad N, DesCroix P, Dony A. Prediction of mix modulus and fatigue law
from binder rheology properties. J Assoc Asphalt Paving Technol 1996;65:40.
the fracture would propagate through the specimen.
[4] Reese R. Properties of aged asphalt binder related to asphalt concrete fatigue
The fracture surfaces of some time sweep test specimens were life. J Assoc Asphalt Paving Technol 1997;66:604.
analyzed. In some cases, the fracture initiated at a singular point [5] Soenen H, De La Roche C, Redelius P. Fatigue behaviour of bituminous
and from there it propagated through the material. Fig. 12 clearly materials: from binders to mixes. J Road Mater Pavement Des 2003;4:7–27.
[6] Castro M, Sánchez JA. Estimation of asphalt concrete fatigue curves – a damage
shows the fracture origin surrounded by several concentric semi- theory approach. Construct Build Mater 2007;22:1232–8.
circular marks typical of fatigue processes called beach marks [7] Reyes Ortiz OJ, Hernández Noguera JA, Pérez Jiménez FE. Relation between the
[12]. Finally, the surface of the rapid fracture area was extremely fatigue behaviour of bituminous and bituminous asphalt mixture. In: Asphalt –
road for life, Euroasphalt and Eurobitumen congress; 2008.
flat. In other words, a small defect in the specimen caused the fail- [8] Chailleux E, Bodin D, de La Roche C, Leguern M, Vignard N. Fatigue behaviour of
ure even when the average strain semi-amplitude applied to the bitumen in tension–compression loading mode: rheological analysis and
specimen was much lower than the failure strain obtained in the comparison with mix fatigue. In: Advanced testing and characterization of
bituminous materials; 2009.
increasing strain sweep test for that binder. [9] Bahia HU, Zhai H, Bonnetti K, Kose S. Non-linear viscoelastic and fatigue
The healing effect was reduced in all cases. Particularly, B 60/70 properties of asphalt binders. J Assoc Asphalt Paving Technol 1999;68:1.
which failed at the 6th triangle, did not show a significant increase [10] AASTHO. Guide for mechanistic-empirical design of new and rehabilitated
pavement structures; 2002.
in complex modulus from before and to after the 30-min rest [11] Reyes Ortiz OJ. Análisis del fallo a fatiga de ligantes hidrocarbonados.
period. Parámetros que definen su comportamiento. Ph.D. thesis. Universidad
Politécnica de Cataluña; 2009.
[12] Hertzberg RW. Deformation and fracture mechanics of engineering materials.
4. Conclusions 4th ed. Wiley; 1996.
[13] Anderson DA, Le Hir YM, Marasteanu MO, Planche J, Martin D. Evaluation of
fatigue criteria for asphalt binders. Transport Res Record: J Transport Res
The results obtained in this study led to the following Board 2001;1766:48–56.
conclusions: [14] Johnson CM. Estimating asphalt binder fatigue resistance using an accelerated
test method. Ph.D. thesis. University of Wisconsin – Madison; 2010.
[15] Di Benedetto H, Soltani AA, Chaverot P. Fatigue damage for bituminous
 The initial drop in complex modulus and dissipated energy mixtures. In: Mechanical tests for bituminous materials; 1997.
density is not related to damage or increase in specimen [16] De La Roche C, Marsac P. Caractérisation expérimentale de la disipation
thermique dans un enrobé bitumineux sollicité en fatigue. In: First
temperature but to thixotropic effects that govern binder international Eurobitume and Euroasphalt congress; 1996.
transition from one flow rate to another. It is important to [17] Rompu JV, Benedetto HD, Gauthier G, Gallet T. New fatigue test on bituminous
note that, since the test frequency was kept constant binders and mastics using an annular shear rheometer prototype and waves
propagation. In: Advanced testing and characterization of bituminous
throughout the test, the velocity of the equipment moving materials; 2009.
parts, and therefore the flow rate within the specimen, [18] Shan L, Tan Y, Underwood S, Kim YR. Application of thixotropy to analyze
increased or decreased with increasing or decreasing the fatigue and healing characteristics of asphalt binder. Transport Res Record: J
Transport Res Board 2010;2179:85–92.
strain semi-amplitude.
[19] Mewis J, Wagner NJ. Thixotropy: colloids, polymers and surfactants. Special
 In the increasing strain sweep test, an increase in strain semi- issue in honour of Brian Vincent. Adv Colloid Interf Sci 2009;147–148:214–27.
amplitude resulted in a sudden increase in dissipated energy [20] Coombs CE, Traxler RN. Rheological properties of asphalts – IV. Observations
density, followed by a rapid drop until it reaching the same sta- concerning the anomalous flow characteristics of air-blown asphalts. J Appl
Phys 1937;8(4):291–6.
bilization value observed in the time sweep test performed at [21] Traxler RN. A review of the rheology of bituminous materials. J Colloid Sci
the same strain. 1947;2(1):49–68.
F.E. Pérez-Jiménez et al. / Construction and Building Materials 31 (2012) 212–219 219

[22] Romberg JW, Traxler RN. Rheology of asphalt. J Colloid Sci 1947;2(1):33–47. [25] AASTHO. ASTM D 2872-04: Standard method of test for effect of heat and air
[23] Majidzadeh K, Schweyer HE. Non-newtonian behavior of asphalt cements. on a moving film of asphalt (rolling thin-film oven test); 2009.
Proc Assoc Asphalt Paving Technol 1965;34:20–44. [26] UNE-EN 12607-1:2007. Bitumen and bituminous binders. Determination of
[24] Pérez-Jiménez FE, Miró R, Martínez A, Botella R, Reyes O, Valdés G. False failure the resistance to hardening under influence of heat and air. Part 1: RTFOT
in flexural fatigue test. In: 2nd Workshop on four point bending, 24–25th method; 2007.
September 2009.

Anda mungkin juga menyukai