Thermally Induced Fracture Performance of Asphalt Mixtures
Licentiate Thesis
Prabir Kumar Das
Division of Highway and Railway Engineering Department of Transport Science
School of Architecture and the Built Environment KTH, Royal Institute of Technology
SE-100 44 Stockholm SWEDEN
September 2012
© Prabir Kumar Das TRITA-TSC-LIC 12-006 ISBN 978-91-85539-91-8
i
Abstract:
A major distress mode in asphalt pavements is low temperature cracking, which results from the contraction and expansion of the asphalt pavement under extreme temperature changes.
The potential for low temperature cracking is an interplay between the environment, the road structure and importantly the properties of the asphalt mixture. The thermal cracking performance of asphalt concrete mixtures can be evaluated by conducting thermal stress restrained specimen tests (TSRST) which is known to be correlated well with the fracture temperatures observed in the field. Although TSRST provides a good estimation of the field performance, it may be unrealistic to implement the obtained results in a design framework.
On the other hand, recent studies showed Superpave indirect tension tests can be used to evaluate fracture performance (fatigue, moisture damage, low temperature cracking, etc.) of the asphalt concrete mixtures. In addition, the obtained elastic and viscoelastic parameters from the Superpave IDT tests can be used as an input parameter to establish a design framework. The study presented in this thesis has a main objective to develop a framework using Superpave IDT test results as input parameters in order to evaluate the low temperature cracking performance of asphalt concrete mixtures. Moreover, the study aims to investigate micro-mechanically the low temperature cracking behavior of bitumen using atomic force microscopy (AFM) as a tool.
The numerical model has been developed by integrating fracture energy threshold into an asphalt concrete thermal fracture model, considering non-linear thermal contraction coefficients. Based on the asphalt concrete mixture viscoelastic properties, this integrated model can predict thermally induced stresses and fracture temperatures. The elastic, viscoelastic and fracture energy input parameters of the model were measured by conducting indirect tension tests and the thermal contraction coefficients were measured experimentally. The proposed model has been validated by comparing the predicted fracture temperatures with the results obtained from TSRST tests. It was found that, while there is a quantitative discrepancy, the predicted ranking was correct. In the measurement of the thermal contraction coefficients it was observed that the thermal contraction coefficient in asphalt concrete is non-linear in the temperature range of interest for low temperature cracking. The implications of having non-linear thermal contraction coefficient were investigated numerically.
In an effort to understand the effect of bitumen properties on low temperature fatigue cracking, AFM was used to characterize the morphology of bitumen. The AFM topographic and phase contrast image confirmed the existence of bee-shaped microstructure and different phases. The bitumen samples were subjected to both environmental and mechanical loading and after loading, micro-cracks appeared in the interfaces of the bitumen surface, confirming bitumen itself may also crack. It was also found that the presence of wax and wax crystallization plays a vital role in low temperature cracking performance of bitumen.
Keywords: Low temperature cracking, asphalt concrete fracture mechanics, viscoelasticity, non-linear thermal contraction coefficient, atomic force microscopy, wax, wax crystallization.
ii
iii
Preface
The work presented in this licentiate thesis has been carried out at the KTH Royal Institute of Technology, at the division of Highway and Railway Engineering.
The Swedish Transport Administration (Trafikverket) is greatly appreciated for financing this study.
I would like to express my sincere gratitude to my supervisors Professor Björn Birgission and Dr. Denis Jelagin for their guidance during this process. Special thanks to Prof. Nicole Kringos for valuable discussions, advice and encouragements.
I am very grateful to Dr. Yuksel Tasdemir for his technical support and advice in the Laboratory. Constructive discussions and comments from Dr. Per Redilius, Mr. Måns Collin, Dr. Jonas Ekbland and Dr. Xiaohu Lu are acknowledged for bringing their insight, experience and ideas to the table.
Thanks go to Prof. Bahia for providing opportunity to perform tests at MARC in University of Wisconsin and PEAB AB for assisting in the slab preparation. Nynas AB is gratefully acknowledged for providing all the bitumen used in this study.
I would like to thanks my colleagues at the department for providing a creative and friendly atmosphere.
Last but not least, my gratitude goes to my wife Soma Ghosh for giving me a life beyond my profession and my family in Bangladesh that always believed in me.
Prabir Kumar Das Stockholm, Sep’12
iv
v
Dedication
To my parents that I dedicate this work.
I cannot ask more from my parents who are simply perfect.
vi
vii
Publications
This Licentiate thesis is based on the following publications:
Journal:
I. Das, P. K., Tasdemir, Y. and Birgisson, B. (2012), Evaluation of Fracture and Moisture Damage Performance of Wax Modified Asphalt Mixtures. International Journal of Road Materials and Pavement Design, Vol. 13 (1), pp. 142-155.
II. Das, P. K., Tasdemir, Y. and Birgisson, B. (2012), Low Temperature Cracking Performance of WMA with the Use of the Superpave Indirect Tensile Test.
International Journal of Construction and Building Materials, Vol. 30, pp. 643-649.
III. Das, P. K., Jelagin, D. and Birgisson, B. (2012), Low temperature cracking model based on HMA fracture mechanics. International Journal of Solids and Structures, under review.
Submitted on: 22-09-2011 Conference:
IV. Das, P. K., Jelagin, D., Birgisson, B. and Kringos, N. (2012), Micro-Mechanical Investigation of Low Temperature Fatigue Cracking Behaviour of Bitumen. 7th RILEM International Conference on Cracking in Pavements, June 20-22, Delft, Netherland.
Other relevant publications:
i. Das, P. K., Jelagin, D. and Birgisson, B. (2009), Warm mix asphalt mixtures for lowering energy consumption during construction of asphalt pavements.
Transportforum, VTI, Linköping, Sweden.
ii. Tasdemir, Y., Das, P. K., and Birgisson, B. (2010), Determination of Mixture Fracture Performance with the Help of Fracture Mechanics. 9th International Congress on Advances in Civil Engineering. Trabzon, Turkey.
iii. Das, P. K., Tasdemir, Y. and Birgisson, B. (2011), Low Temperature cracking Performance of Wax Modified Bitumen and Mixture. XXIVth World Road Congress.
27-30 Sep, Mexico.
iv. Das, P. K., Kringos, N. and Birgisson, B. (2011), Atomic Force Microscopy (AFM) to Characterize the Aging of Asphaltic Materials. KTH-Transport Day, November, Stockholm, Sweden.
viii
ix
Contents
1. Introduction ……… 1
2. Thermally Induced Fracture Model……… 2
3. Summary of Appended Papers……… 6
3.1. Implementation and Evaluation of HMA Fracture Mechanics ……… 6
3.2. Low Temperature Fracture Performance Based on TSRST and Superpave IDT … 8 3.3. Thermally induced fracture model based on Superpave IDT test results……… 10
3.4. Micro-mechanical investigation of low temperature fatigue behavior of bitumen … 14 4. Discussion and conclusions ...……… 16
References….……… 17
x
1
1. Introduction
Low temperature cracking is one of the major distresses in asphalt pavements, since it is irreversible and expensive to repair. It can be a serious problem in cold regions as well as in areas with large daily temperature fluctuation. Whenever the ambient air temperature drops, the asphalt mixtures tend to contract. This continuous contraction phenomenon results in a tensile stress buildup in the mixture which is known as thermally induced stress. If this thermal stress exceeds the tensile strength of the asphalt pavement, a transverse crack will occur at the surface and the corresponding temperature is known as fracture temperature.
Consequently, these cracks will propagate downward through the pavement and increase in frequency with the aging of pavement and additional low temperature cycles. Water infiltration through these cracks can cause problems such as loss of fines or reduced sub-base strength and result in reduce pavement service life and high maintenance cost. Thus, accurate description of the initiation of thermal induced fracture is critical for designing durable pavements in cold regions.
Research regarding thermally induced fracture of asphalt mixtures has been carried out over the years. One of the great outcomes from the low temperature cracking studies is the thermal stress restrained specimen test (TSRST) which was developed at Oregon State University (Kanerva et al., 1994). The thermal cracking performance of asphalt concrete mixtures can be evaluated by conducting TSRST which is known to be correlated well with the fracture temperatures observed in the field. Although TSRST provides a good estimation of the field performance, it may be unrealistic to implement the obtained results in a design framework. On the other hand, Superpave indirect tension tests developed at University of Florida (Buttlar and Roque, 1994) which can be used to evaluate fracture performance of the asphalt concrete mixtures. In addition, the obtained elastic and viscoelastic parameters from the Superpave IDT tests can be used as an input parameter to establish a design framework.
The primary objective of this thesis is to develop and validate a tool using Superpave IDT test results as input parameters in order to predict the thermally induced stress and fracture temperature in order to evaluate the low temperature cracking performance of asphalt concrete mixtures.
The first part of this thesis consists of a comprehensive laboratory study to implement and evaluate fracture performance of asphalt mixtures by using Superpave IDT tests and hot mix asphalt (HMA) fracture mechanics (Paper I). In Paper II, low temperature cracking performances of asphalt mixtures were evaluated by Superpave IDT and TSRST tests.
Consequently, a comparison between Superpave IDT and TSRST test results on low temperature cracking susceptibility was also documented. The main part of the thesis deals with the development and validation of a tool that can predict the thermally induced stress and fracture temperature (Paper III). To do so, the predicted thermally induced stresses and fracture temperatures were compared with the obtained TSRST results, as this test is known to correlate well with the fracture temperature observed in the field. In the proposed model, special attention has been paid to thermal contraction coefficient of asphalt mixtures. This thesis furthermore describes the micro-mechanical investigation of low temperature fatigue behavior of bitumen using atomic force microscopy (Paper IV).
2
2. Thermally Induced Fracture Model
Asphalt concrete consists of an aggregate skeleton that is bound together by a bituminous matrix which is a mixture of bitumen, sand and other types of fillers or modifiers. In current specifications, however, thermal cracking performance is linked only to bitumen properties, neglecting thus the effect of the other components and their interaction on the asphalt mixture viscoelastic and mechanical properties (e.g. creep compliance, relaxation modulus and fracture properties) or thermal properties (e.g. coefficient of thermal contraction). During the last three decades there have been several attempts to develop thermally induced fracture models for asphalt mixtures. In the early seventies and eighties, empirical models had been introduced (Fromm and Phang, 1972; Hass et al., 1987) to estimate crack spacing, where bitumen properties were used as input parameters while mixture properties were not considered. Other existing models such as, for example, COLD (Finn et al., 1986) and CRACK3 (Roque and Ruth, 1990) were able to predict performance of the mixtures at low temperature. However, these two analytical models were not able to determine cracking as a function of time. A model developed in Texas A&M University, known as THERM (Lytton et al., 1983) was able to consider time as a variable, while it was dependent on estimated mixture properties rather than actual mixture properties at low temperatures. By considering all of these problems, a mechanics-based thermal cracking performance model (TCMODEL) was developed during the Strategic Highway Research Program (SHRP A-005) (Hiltunen and Roque, 1994). In TCMODEL, the theory of linear viscoelasticity was used to predict thermally induced stresses which were used to analyze crack growth rate by using Paris’ law (Paris et al., 1961). Researchers in University of Florida (Zhang et al., 2001; Roque et al., 2002) introduced a ‘Hot Mix Asphalt’ (HMA) fracture mechanics framework to characterize load induced cracking performance of asphalt mixtures, also refered to as top-down cracking. A good correlation with field performance was observed using this model, even though the framework completely ignored thermally induced stresses. Recently, Kim et al. (2008) integrated thermally induced stresses in the HMA fracture mechanics framework by incorporating the theory of linear viscoelasticity from the TCMODEL. This integrated framework is known as the HMA thermal fracture model, where dissipated creep strain energy (DCSE) was calculated by considering the effect of thermal loading conditions. The evaluation of that combined system showed a better correlation between the predicted and observed top-down cracking than the results reported by Roque et al. (2004). This work, however, was only done for predicting load and temperature induced top-down cracking, which is associated with high and intermediate temperatures. By considering all those drawback described above, a low temperature cracking model has been implemented by integrating a fracture energy (FE) threshold into the HMA thermal fracture model. The elastic, viscoelastic and fracture input parameters for the proposed model were obtained from performing Superpave indirect tension tests (Buttlar and Roque, 1994).
The different steps of proposed low temperature cracking model can be described by a flow diagram as showing in Fig. 1.
3 No
No
Yes Yes
No Yes
Take the lower time to develop 5mm crack
Total Crack ≥ 100mm
Crack length vs. Time (or Temp.) Calculate: DCSE for Each
Zones and Transfer DCSE to Reference Temperature
Calculate: FE for Each Zones and Transfer FE to
Reference Temperature
DCSE ≥ DCSE limit FE ≥ FE limit
tDCSE (time to crack 5mm) tFE (time to crack 5mm) Input: D0,D1, m, DCSEf,
FE and Tensile Strength
Input: Thermal Contraction Coefficient, Cooling Rate
Define Processing Zone Calculate: Average Stress in each
Calculate: Shift factor, Prony Series Parameters for Relaxation Calculate: Thermal Stress, Creep
strain and Thermal strain
Fig. 1. Flow diagram of thermally induced fracture model for asphalt mixtures
Under transient temperature conditions where temperature varies with time, thermal stress is generally involved and developed due to thermal contraction. The coefficient of thermal contraction could therefore play a vital role to calculate thermally induced stress. The thermal stress can be calculated by using one-dimensional viscoelastic constitutive law, known as the Boltzmann’s Superposition Principle for linear viscoelastic materials.
4
'
'' '0
( )
E d d
d
ξ (1)where ( ) is the stress at reduced time , ( ) is the relaxation modulus at reduced time and is the real time integrating variable. Morland and Lee (1960) introduced the following reduced time, which is able to take into account both effects of temperature gradient and time variations coincidently.
0 (1
' ) 't
T
t dt
a T t
(2)where is the time-temperature shift factor. This shift factor can be determined from Arrhenious equation or Williams-Landel-Ferry (WLF) equation. The other parameter in Eq.
(1), thermal strain rate ̇( ) [i.e., ( ) / ] which is directly related to the coefficient of thermal contraction (α).
T, '
T T( )' (3)
where the rate of change in temperature, ̇( ) ( ) . It can be seen from Eqs. (1) and (3), coefficient of thermal contraction is an input parameter to get thermally induced stress and more importantly, this is a function of temperature.
The area in front of crack tip is divided into a series of process zones. The stress distribution over the process zone is subjected to uniform tension and in the vicinity of a crack it can be expressed by the following linear elastic fracture mechanics equation:
2 2
I R
K a
r r
(4)
where a is crack length, r is distance measured from crack tip and is the remote stress. The stress distribution in the body of course depends on the loading condition and the failure limit of the asphalt mixture. Classical linear elastic fracture mechanics assumes that stresses approach infinity at the crack tip. HMA fracture mechanics, however, recognizes that stresses must be limited to account for the fact that the stress cannot exceed some maximum limit at any point in the mixture. Zhang et al. (2001) suggested in their study that the mixture tensile strength is suitable for use as a stress limit to define the stress distribution near the crack tip, which is implemented in this study.
With respect to the process zone size, an experimental study (Birgisson et al., 2007) for dense graded asphalt mixture indicated that the initial crack length is about 10mm. Accompanied by 5mm of processing zones are assumed, corresponds to one-half the nominal maximum aggregate size. In each process zone, an average value of the thermally induced tensile stress is calculated at small time increments. These average stresses (σavg) are subse uently used along with the rate of creep strain ( cr) to calculate the DCSE over the process zone by using the following equation:
5
0 t
avg cr
DCSE
dt (5)Fracture Energy can be used as another threshold which develops due to the temperature change in the pavement. As, the primary concern of the present study is low temperature cracking performance, thus the FE threshold is included along with the DCSE threshold. The average stresses (σavg) over the process zones and the thermally induced strains ( th) can be used to calculate fracture energy. The total accumulated FE can be determined using the following equation:
0 t
avg th
FE
dt (6)The obtained DCSE and FE are then assigned to each process zone. The process zone near the crack tip is considered to have failed once the total accumulated DCSE reached the DCSE limit or the total accumulated FE reached the FE limit, whichever comes first. At the same time, the increasing crack length is changing the stress distribution along the process zone.
Therefore, in the next step the stresses are redistributed along the processing zones and so on.
6
3. Summary of Appended Papers
3.1. Implementation and Evaluation of HMA Fracture Mechanics (Paper I)
In recent years, Birgisson and Roque (Birgisson et al., 2006; Birgisson et al., 2007; Roque et al., 1999; Roque et al., 2002; Roque et al., 2004; Zhang et al., 2001) have performed comprehensive studies to characterize the crack initiation and crack growth of asphalt mixtures. Central to their work was the development of a viscoelastic fracture mechanics- based framework for predicting the cracking performance of asphalt mixtures, i.e. Hot Mix Asphalt (HMA) fracture mechanics. The asphalt mixture properties used in the HMA fracture mechanics framework include the resilient modulus test, creep compliance and indirect tensile strength, which can all be obtained with the Superpave IDT test (Buttlar and Roque, 1994; Roque and Buttlar, 1992). Central to the HMA fracture mechanics is that no single property is sufficient to predict the cracking performance of a given asphalt mixture.
Rather a model is required that uses fundamental mixture properties. The framework was extended to identify a surrogate property that combines the features of the HMA fracture mechanics framework with field experience, i.e. the Energy Ratio (ER) (Roque et al., 2004).
The load induced fracture and moisture damage of asphalt mixtures were evaluated using this HMA fracture mechanics framework which is based on Superpave IDT test results. Total six different mixtures had been investigated, where the mixtures were produced from two different types of wax modified bitumen (wax FT and wax AB) and aggregates from two different sources (AG1 and AG2). The control mixtures were denoted as AG1-O and AG2-O, where aggregate type AG1 and AG2 mixed with unmodified bitumen (O), respectively and so on.
Fig. 2. Graphical illustration of Energy Ratio of the mixtures 0.0
0.5 1.0 1.5 2.0 2.5 3.0 3.5
AG1-O AG1-AB AG1-FT AG2-O AG2-AB AG2-FT
Energy Ratio (ER)
Mixture ID
7
The following properties were determined from the Superpave IDT test: resilient modulus (MR), creep compliance, tensile strength (St), fracture energy limit (FE) and dissipated creep strain energy limit (DCSEf). In order to evaluate the combined effect on fracture resistance, the Energy Ratio (ER) was calculated and presented in Fig. 2. The results show that the fracture resistance as characterized by the ER which is increased by at least a factor of two for both wax modified mixtures as compared to the unmodified mixtures.
Fig. 3. Graphical illustration of moisture damage ratio (MDR) of energy ratio of the mixtures
Birgisson et al. (2003, 2004) used the concept of ER into moisture damage ratio (MDR) for evaluating the moisture damage in asphalt mixtures. The multiple-parameter MDR is an analytically-based function to quantify damage by combining more than one material property, also considering both dry and wet conditions. This parameter was used to evaluate the loss of resistance to fracture in asphalt mixtures due to moisture damage. Fig. 3 shows the relative change in ER for the conditioned versus unconditioned mixtures as characterized by the Moisture Damage Ratio (MDRER). It can be seen that the wax modified mixtures shows higher MDRER compared to unmodified mixtures, indicating the moisture damage resistance increased significantly. After moisture conditioning, the decrease of ER represents a reduction in fracture resistance due to moisture damage. The mixtures with wax modified bitumen show higher moisture damage ratio of ER compared to unmodified mixtures, indicating the higher fracture resistance than the control mixtures.
It can be concluded that the HMA fracture mechanics framework can be used to evaluate load induced fracture performance of the asphalt mixtures. Thus it would be interesting to study how this framework can be used to determine low temperature cracking performance of asphalt mixtures.
0.00 0.10 0.20 0.30 0.40 0.50 0.60
AG1-O AG1-AB AG1-FT AG2-O AG2-AB AG2-FT
MDRER = ERw / ERd
Mixture ID
8
3.2. Low Temperature Fracture Performance Based on TSRST and Superpave IDT (Paper II)
In Paper II, thermal stress restrained specimen test (TSRST) and Superpave IDT test were used to evaluate the low temperature fracture performance of asphalt mixtures. The same wax modified mixtures were used as described in Paper I. The tests were performed according AASTHO TP-10-93. At the beginning of the test, thermally induced stress increases relatively slow with the decreasing temperature, which is due to the relaxation of the specimen. After crossing a certain temperature (transition temperature), the thermally induced stress is not relaxed and is almost linearly increased until fracture of the specimen.
The temperature at which specimen failed is recognized as fracture temperature and the strength is called fracture strength.
Table 1: Low temperature properties of asphalt mixture obtained from TSRST Mixure ID
Fracture temperature
[°C]
Fracture strength [MPa]
Slope at fracture [MPa/°C]
Transition temperature
[°C]
Transition stress [MPa]
AG1-O -25.7 2.57 0.17 -11.0 1.0
AG1-AB -24.2 2.61 0.18 -9.0 1.2
AG1-FT -24.3 2.80 0.19 -9.2 1.3
AG2-O -25.5 2.88 0.20 -11.9 1.0
AG2-AB -25.3 2.64 0.18 -10.9 1.1
AG2-FT -23.9 2.91 0.21 -9.6 1.4
Note: In the table the control mixtures are denoted as AG1-O and AG2-O, where aggregate types AG1 and AG2 mixed with unmodified bitumen (O) respectively and so on.
The effects on low temperature properties determined from TSRST are summarized in Table 1. An averaging approach was used to present the results, i.e. the numbers presented are obtained as the average value from two tests. Lower fracture temperature represents better performance in low temperature cracking. It can be seen that wax modified mixtures show minor increase in fracture temperature. The maximum difference observed in fracture temperature after wax modification was about 1.6°C. After wax modification, mixture fracture strength, transition stress and slope at fracture increased while transition temperature decreased, indicating a possible negative effect on low temperature cracking performance. This minor negative effect could be because of thermal effect on wax crystallization. However, no significant difference observed using two different types of aggregate, implies aggregate type has no major effect on low temperature cracking.
Superpave IDT tests were conducted at 0°C, -10°C and -20°C and all the fracture parameters obtained from these tests are summarized in Table 2. The resilient modulus (MR) represents the elastic response of the material. At low temperatures, a higher elastic stiffness implies that the asphalt mixture may be more susceptible to thermal cracking. The results at all temperatures show that after wax modification the elastic stiffness of the mixture has
9
increased slightly. Also, after wax modification m-value decreases in most of the cases which may results poor low temperature cracking performance. The lower m-values imply that the mixtures may be less effective in redistributing the stresses which build up in asphalt pavements as the temperature drops.
Table 2: Fracture parameters obtained from Superpave IDT tests Mixture
ID
Test temp.
[°C]
MR
[GPa] m-value FE DCSEf DCSEmin Creep rate (x 10-3) @ 1000s
AG1-O 12.9 0.63 2.51 2.27 2.39 4.463
AG1-AB 0 13.2 0.56 1.98 1.73 0.84 1.222
AG1-FT 15.0 0.60 2.33 2.02 0.82 1.231
AG1-O 18.7 0.65 0.97 0.76 0.19 0.357
AG1-AB -10 18.7 0.48 1.15 0.89 0.19 0.195
AG1-FT 19.7 0.51 1.04 0.80 0.19 0.214
AG1-O 23.3 0.35 0.46 0.25 0.04 0.033
AG1-AB -20 22.7 0.32 0.42 0.23 0.03 0.025
AG1-FT 24.7 0.38 0.51 0.24 0.04 0.027
AG2-O 13.6 0.62 2.72 2.47 3.30 5.707
AG2-AB 0 13.4 0.61 2.70 2.27 1.44 2.481
AG2-FT 13.9 0.61 1.88 1.65 0.94 1.624
AG2-O 17.1 0.63 1.25 1.00 0.31 0.550
AG2-AB -10 18.5 0.43 0.99 0.76 0.21 0.194
AG2-FT 19.9 0.41 0.72 0.52 0.15 0.137
AG2-O 24.7 0.32 0.42 0.23 0.03 0.023
AG2-AB -20 22.9 0.26 0.43 0.20 0.03 0.020
AG2-FT 24.9 0.31 0.45 0.26 0.03 0.025
According to HMA fracture mechanics, the tensile creep rate is related to the rate of damage in tension. As can be seen, the measured creep rate at 1000s is lower for wax modified mixtures, indicating low rate of damage in tension at low temperature. Interestingly, the dissipated creep stain energy at failure (DCSEf) decreased which indicates a lower fracture energy threshold. However, the difference after wax modification is minor at -10°C and - 20°C, which is considered as low temperature. The same phenomenon can be observed for fracture energy (FE) and minimum dissipated creep strain energy (DCSEmin), which are almost same at -20°C.
By comparing all the parameters in Table 2, it can be concluded that each and every parameter shows only minor differences (if at all) at -20°C, which implies wax modification did not improve low temperature fracture properties. It can be seen that the findings are consistence with the TSRST test results. As, Superpave IDT test results can give an overall idea about mixture behavior at low temperature, thus it would be of interest to learn how the obtained elastic and viscoelastic parameters from the Superpave IDT tests can be used as an input parameter to establish a design framework.
10
3.3. Thermally induced fracture model based on Superpave IDT test results (Paper III)
The primary scope of this thesis is develop and validate a tool to predict the thermally induced stress and fracture temperature in order to evaluate the low temperature cracking performance of asphalt mixtures. To do so, a low temperature cracking model has been implemented by integrating a fracture energy (FE) threshold into the HMA thermal fracture model developed earlier by Kim et al. (2008). The elastic, viscoelastic and fracture input parameters for the proposed model were obtained from performing Superpave IDT tests (Buttlar and Roque, 1994). In order to validate the proposed model, the obtained thermally induced stresses and fracture temperature were compared with Thermal Stress Restrained Specimen Test (TSRST) results for three different asphalt mixtures. Furthermore, the coefficient of thermal contraction for all the three asphalt mixtures was measured by conducting glass transition tests. From this, it was found that thermal contraction is non- linear in the temperature range of interest for low temperature cracking. The implications of having non-linear thermal contraction coefficient were investigated numerically.
Three types of asphalt mixture were produced using three different types of bitumen with penetration grade of 35/50, 70/100 and 160/220. These bitumen were mixed with aggregate of the same gradation. Using these aggregates and three types of bitumen total three different types of mixtures were prepared which were denoted as mix35/50, mix70/100 and mix160/220, respectively. The reason of using this three different types of bitumen was to cover wide stiffness range (i.e., from stiff to soft) of bitumen.
Fig. 4. Thermal-volumetric response of the three investigated asphalt mixtures.
-3.0E-3 -2.5E-3 -2.0E-3 -1.5E-3 -1.0E-3 -5.0E-4 0.0E+0
-80 -60 -40 -20 0 20 40
Specific Volume Change, v (ml/g)
Temperature, T ( °C) mix160/220
mix70/100 mix35/50
11
The specific volume change with respect to temperature for the three asphalt mixtures used in this study is presented in Fig. 4, which is known as thermal strain curve. The corresponding slope at a specific temperature of the thermal strain curve is the coefficient of thermal contraction at that temperature. It can be seen from the figure that the slope of the thermal strain curve is changing with the cooling, which indicates the existence of nonlinear trend in the coefficient of thermal contraction. The thermal strain curves also indicate the glass transition of asphalt mixtures occurs over a large temperature range. One may also observe that the thermally induced contraction is higher for mix 160/220. This means this mixture can relatively deform more in similar temperatures compared with the other tested mixtures. On the other hand, mix 35/50 exhibits stiffer behavior than the other two mixtures.
Considering the selection of the bitumen, this behavior is exactly as expected.
The TSRST tests were performed to validate the proposed model. The measured fracture temperature and fracture strength were analyzed statistically and shown in Table 3. The standard deviation (S) and coefficient of variation (CV) was also calculated to monitor the variability of the measurements. As can be seen, fracture temperatures obtained from the tests were quite repeatable while the measured thermal stresses showed significantly higher variance. The fracture temperature of the mixture is an important parameter, as it represents the critical temperature for the pavement to be cracked in winter.
Table 3: Fracture temperature and strength obtained from TSRST
Mixture ID Fracture Temperature (°C) Fracture Strength (MPa)
mean Sa CVb mean Sa CVb
mix35/50 -22.0 0.31 1.31 3.0 0.55 18.03
mix70/100 -26.5 1.25 4.66 2.7 0.24 9.13
mix160/220 -32.0 1.72 5.17 3.0 0.42 13.85
aStandard deviation
bCoefficient of variation (%)
The proposed model was validated by comparing the predicted thermally induced stresses and fracture temperatures with the measured values from the TSRST test. For comparison, the TSRST was selected because this test appears to be the closest in simulating single event thermal cracking in pavement, thus making the model most relevant for being used in a pavement representation. The thermal contraction coefficient is an important required parameter for predicting thermally induced stresses and strains. Thus, both a linear and a non-linear approach were used, to observe the effect of the non-linear relationships on the predicted results.
12
Fig. 5. Comparison of analytical solution with TSRST test results using linear thermal contraction coefficient.
Fig. 6. Comparison of TSRST test results with analytical solutions using nonlinear thermal contraction coefficient.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
-35 -30 -25 -20 -15 -10 -5 0 5
Thermal induced stress (MPa)
Temperature ( °C)
TSRST- mix35/50 TSRST- mix70/100 TSRST- mix160/220 Predicted Stress
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
-35 -30 -25 -20 -15 -10 -5 0 5
Thermal induced stress (MPa)
Temperature ( °C)
TSRST-mix35/50 TSRST-mix70/100 TSRST-mix160/220 Predicted stress
13
It has always been an important issue to predict thermally induced stress correctly. With the help of the proposed model thermally induced stresses were predicted and compared with TSRST test results, which is depicted in Fig. 5 and Fig. 6 . Comparing these two figures, it can be seen that the model can predict thermally induced stress more accurately using non-linear thermal contraction coefficient that the linear. It can be easily seen that the predicted stresses are quite close to the observed stresses in TSRST tests for all of the three different mixes, which means the basic physics of viscoelasticity was followed correctly.
Table 4: Comparison between predicted fracture temperature and TSRST test results.
Mixture ID Fracture temperature, °C
TSRST test Predicted a Predicted b
mix35/50 -22 -13.4 -22.8
mix70/100 -26.5 -19.7 -23.6
mix160/220 -32 -23.8 -27.7
a: using linear thermal contraction coefficient b: using non-linear thermal contraction coefficient
Finally in Table 4, the predicted temperatures were compared with the measured fracture temperatures from the TSRST tests. As can be seen, the same ranking was observed between the fracture temperature obtained from the model and TSRST tests. However, the measured and predicted fracture temperatures were differed more in linear thermal contraction coefficient approach. In non-linear approach, most importantly it can be seen that the predicted fracture temperature shows again the same quantitative ranking with the TSRST test results, but this time much closer to the TSRST fracture temperatures. This certainly indicates the advantage of using a non-linear relationship for predicting the critical fracture temperatures.
Table 5: Observed and predicted fracture temperatures of mixtures used in Paper I and II.
Mixture ID Fracture temperature, °C
Mixture ID Fracture temperature, °C
TSRST test Predicted TSRST test Predicted
AG1-O -26 -19 AG2-O -26 -19
AG1-MW -24 -18 AG2-MW -25 -19
AG1-S -24 -16 AG2-S -24 -15
In addition, those mixtures used in Paper I and II were also used in the model and the predicted and TSRST test results are compared which is shown in Table 5. Interestingly, the same quantitative ranking was again observed. However, the measured and predicted fracture temperatures were differed by several degrees. One of the key reasons may be using linear thermal contraction coefficient because of lacking of glass transition test data.
Thus it can be concluded, correct ranking was observed between the TSRST fracture temperature and predicted critical temperature for all the mixtures.
14
3.4. Micro-mechanical investigation of low temperature fatigue behavior of bitumen (Paper IV)
Atomic force microscope (AFM) is capable of measuring topographic features at atomic and molecular resolutions as compared to the resolution limit of optical microscopy of about 200nm. Moreover, AFM has the advantage of imaging almost any type of surface which opens the window for investigating micro-cracks due to thermal fatigue. From AFM scans, it was found that bitumen is not a homogeneous bulk material as microstructures are observed in almost all the bitumen as shown in Fig. 7. A detailed knowledge of these microstructures is needed to understand the physico-chemistry of bitumen, which can serve as the direct link between the molecular structure and the rheological behaviour. Thermorheological behaviour of bitumen is an important factor for understanding the performance-based optimization of asphalt mixture. Thus, the prediction of the performance of asphalt pavements should also directly be related to this (Lesueur et al., 1996; Loeber et al., 1996).
Structuring may occur at various ranges from molecular to macroscopic but most of the asphalt researchers have concerned themselves with the microstructures of the bitumen itself. From extensive AFM investigations shown in earlier papers (Pauli et al., 2001; Jäger et al., 2004; Schmets et al., 2010), it has been shown that bitumen has the tendency to phase separate under certain kinetic conditions, leading to a predominant clustering of two types of phases, illustrated in Fig 7.
Fig. 7. Topographic 2D (left) and 3D right AFM image (19µm × 19 µm) of bitumen indicating evidence of microstructures.
The existence of microstructures in bitumen matrix has been found in several studies which prove the heterogeneity of bitumen (Masson et al., 2007; Pauli et al., 2011). This inhomogeneity is creating internal interfaces within the bitumen matrix. From mechanical considerations, it is known that the interfaces between two materials with different stiffness properties serve as natural stress inducers. This means that when the material is exposed to mechanical and or environmental loading, these interfaces will attract high stresses and are prone to cracking. Thus, at low temperatures when the bitumen becomes stiff, this induced stress may cause cracking which could propagate through bitumen itself.
15
In this study, approximately 30mg of hot bitumen was carefully spread out on a rectangular silicon bar (60mm × 20mm × 7mm). Then the bitumen film was cooled to room temperature and covered to prevent dust pick-up and annealed for a minimum of 24 h. The bitumen film over the silicon bar was then subjected to thermal fatigue with heating and cooling cycles. A freezer was used to regulate the low temperature at -20°C and a room with controlled temperature at 25°C was used for thawing. In between each thermal cycle the sample was subjected to additional tensile stress. This was done by bending at the mid-point of the silicon bar at a controlled angle of 3 degrees. The sample underwent 15 cycles of thermal loading; after the last freezing, the sample was placed under the AFM to investigate change in micro-structure due to the thermal fatigue.
Fig. 8. AFM images: a) topographic b) amplitude and c) phase contrast of bitumen (15µm × 15µm) before thermal fatigue loading.
Fig. 9. Evidence of micro-crack through AFM scanning: a) topographic b) amplitude and c) phase contrast after thermal fatigue loading
Typical topographic and phase contrast image of the bitumen surface obtained at 25°C are presented in Fig. 8, where one can easily observe the existing of phase separation in the bitumen matrix as described earlier. The rippled microstructures are also observed which are often referred to as bee-structures. The pale and dark lines indicate rise and drop of the topographic profile against the background, which are also known as peaks and valleys, respectively. As, the sample was exposed to thermal cycles and controlled tensile stresses thus the degraded material properties results into micro-cracks (crazing pattern), as depicted in Fig. 9. If this process would continue, these micro-cracks would continue developing and finally form macro-crack.
a) b) c)
a) b) c)
16
4. Discussion and conclusions
The presented study is focused on improving and generalizing previously developed hot mix asphalt fracture models for asphalt mixtures. As the proposed model is mainly based on HMA fracture mechanics framework thus a laboratory study was carried out to validate fracture performance of asphalt mixtures using this framework. To do so, in this case wax modified mixtures were investigated and it was found that HMA fracture mechanics framework can be successfully used to determine load induced fracture performance of asphalt mixtures. For further investigation, load induced fracture at low temperatures was also studied using the same wax modified mixtures and compared with TSRST test results.
One of the key findings was that Superpave IDT test results can give an overall idea about mixture behavior at low temperatures but not a rigid conclusion, thus it would be of interest to learn how the obtained elastic and viscoelastic parameters from the Superpave IDT tests can be used as an input parameter to establish a design framework. Based on this outcome a thermally induced fracture model was implemented and validated. Non-linearity in coefficient of thermal contraction was observed in all the three studied asphalt mixtures and implemented in the proposed model along with linear approach.
The model can predict the thermally induced stress and fracture temperature. It was validated by comparing the predicted stress and fracture temperature with the TSRST test results which can simulate low temperature performance of pavements. Thermally induced stresses were captured well compared with observed TSRST stresses, which mean the basic viscoelastic behavior was followed correctly. It was found that, while there is a quantitative discrepancy, the predicted ranking was correct and the non-linear thermal contraction coefficient resulted in better critical temperature prediction than a constant or linear approach. The parametric study showed that predicted thermal cracking performance of asphalt mixtures significantly depends on the assumption of the coefficient of thermal contraction. Thus, it is very important to know the true thermal-volumetric properties of the mixture.
To understand the low temperature fatigue behavior of bitumen in micro-level, atomic force microscopy was used as a tool. Wax induced micro-structures were observed from AFM scans which lead to a predominant clustering phases. The interface between different phases with different stiffness could generate high stresses due to environmental or mechanical loading. A controlled thermal fatigue cycle and mechanical loading was applied on bitumen specimen and then scanned under the AFM to investigate the effect of it. As expected, the micro-cracks were observed at the interface of the different phases due to stress concentration. This phenomenon proofs bitumen itself may also crack under certain kinetic conditions and the presence of wax in bitumen may affect low temperature properties of bitumen itself.
17
References
Birgisson, B., Roque, R., Page, G., 2003. Evaluation of water damage using hot mix asphalt fracture mechanics. J. of the association of asphalt paving technologists. 73, 424–462.
Birgisson, B., Roque, R., Page, G., 2004. The use of a performance based fracture criterion for the evaluation of moisture susceptibility in hot mix asphalt. Transportation research record: journal of the transportation research board. 1891, 51–61.
Birgisson, B., Wang, J., Roque, R., 2006. Implementation of the Florida Cracking Model into the Mechanistic-Empirical Pavement Design. University of Florida, Gainesville.
Birgisson, B., Sangpetgnam, B., Roque, R., Wang, J., 2007. Numerical Implementation of a strain energy-based fracture model for a HMA materials. Int. J. Road Materials and Pavement Design. 8 (1), 7-45.
Buttlar, W.G., Roque, R., 1994. Development and evaluation of the strategic highway research program measurement and analysis system for indirect tensile testing at low temperatures. Transport Research Record. 1454, 163-17.
Finn, F., Sharif, C.L., Kulkarni, R., Nair, K., Smith, W., Abdullah, A., 1986. Development of Pavement Structural Subsystems. NCHRP Report 291, Transportation Research Board, Washington, D.C.
Fromm, H.J., Phang, W.A., 1972. A Study of Transverse Cracking of Bituminous Pavements.
Proc. Assn. Asphalt Paving Technol. 41, 383-418.
Haas, R., Meyer, F., Assaf G., Lee, H., 1987. A Comprehensive Study of Cold Climate Airport Pavement Cracking. Proc. Assn. Asphalt Paving Technol. 56, 198-245.
Hiltunen, D.R., Roque, R., 1994. A Mechanics-Based Prediction Model for Thermal Cracking of Asphaltic Concrete Pavements. J. Assn. Asphalt Paving Technol. 63, 81-117.
Jäger, A., Lackner, R., Eisenmenger-Sittner, Ch., Blab, R., 2004. Identification of four material phases in bitumen by atomic force microscopy. Road Materials and Pavement Design. 5 (Sup. 1), 9–24.
Kanerva, H., Vinson, T.S., Zeng, H., 1994. Low temperature cracking: field validation of the thermal stress restrained specimen test. SHRP- A- 401.
Kim, J., Roque, R., Birgisson, B., 2008. Integration of Thermal Fracture in the HMA Fracture model. J. Assn. Asphalt Paving Technol. 77, 631-661.
Lesueur, D., Gerard, J.-F., Claudy, P., Létoffé, J.-M., Planche, J.-P., Martin, D., 1996. A structure-related model to describe asphalt linear viscoelasticity. Journal of Rheology. 40 (5), 813–836.
18
Loeber, L., Sutton, O., Morel, J., Valleton, J.-M., Muller, G., 1996. New direct observations of asphalts and asphalt binders by scanning electron microscopy and atomic force
microscopy. J. of Microscopy. 182 (1), 32–39.
Lytton, R.L., Shanmugham U., Garrett, B.D., 1983. Design of Asphalt Pavements for Thermal Fatigue Cracking. Research Report No. FHWA/TX-83/06+284-4, Texas Transportation Institute, Texas A&M University, College Station, Texas.
Masson, J-F., Leblond, V., Margeson, J., Bundalo-Perc, S., 2007. Low-temperature bitumen stiffness and viscous paraffinic nano- and micro-domains by cryogenic AFM and PDM. J.
of Microscopy. 227 (3), 191-202.
Morland, L.W., Lee, E.H., 1960. Stress Analysis for Linear Viscoelastic Materials with Temperature Variation. Transaction of the Society of Rheology. 4, 233-263.
Paris, P., Gomez, M., Anderson, W., 1961. A rational analytic theory of fatigue. The trend in engineering. 13, 9-14.
Pauli, A.T., Branthaver, J.F., Robertson, R.E., Grimes, W., Eggleston, C.M., 2001. Atomic force microscopy investigation of SHRP asphalts. ACS division of fuel chemistry preprints. 46 (2), 104-110.
Pauli, A.T., Grimes, R.W., Beemer, A.G., Turner, T.F., Branthaver, J.F., 2011. Morphology of asphalts, asphalt fractions and model wax-doped asphalts studied by atomic force microscopy. Int. J. of Pavement Engineering. 12 (4), 291-309.
Roque, R., Ruth, B.E., 1990. Mechanism and Modeling of Surface Cracking in Asphalt Pavements. J. Assn. Asphalt Paving Technol. 59, 396-42.
Roque, R., Buttlar, WG., 1992. The development of a measurement and analysis system to accurately determine asphalt concrete properties using the indirect tensile mode. J. Assoc Asphalt Paving Technol. 61, 304–332.
Roque, R., Zhang, Z., Sankar, B., 1999. Determination of crack growth rate parameters of asphalt mixtures using the Superpave IDT, J. Assoc Asphalt Paving Technol. 68, 404–433.
Roque, R., Birgisson, B., Sangpetngam, B., Zhang, Z., 2002. Hot mix asphalt fracture
mechanics: a fundamental crack growth law for asphalt mixtures. J. Assn. Asphalt Paving Technol. 71, 816-27.
Roque, R., Birgisson, B., Drakos, C., Dietrich, B., 2004. Development and field evaluation of energy-based criteria for top-down cracking performance of hot mix asphalt. J. Assn.
Asphalt Paving Technol. 73, 229-260.
Schmets, A., Kringos, N., pauli, T., Redelius, P., Scarpas, T., 2010. On the existence of wax- induced phase separation in bitumen. Int. J. of Pavement Engineering. 11 (6), 555-563.
Zhang, Z., Roque, R., Birgisson, B., Sangpetngam, B., 2001. Identification and verification of a suitable crack growth law. J. Assn. Asphalt Paving Technol. 70, 206-241.
