Mechanical and Fracture-Mechanical Properties of Asphalt-Concrete Interfaces

(ProQuest: ... denotes formula omitted.)

INTRODUCTION

The bond of cement-bound or bituminous or other materials is produced very often in civil engineering. Examples for this are roads, airfields, bridge parts, industrial floors, and canals, whether it concerns a new building or restoration. Weak points of such buildings may often be the material bond. Cracks mostly run in the interface of the bond.

The bond between old and new concrete1-6 and between old and new asphalt7-10 have been investigated several times. Adding a layer of concrete to an old asphalt surface is called thin or ultra-thin whitetopping. An old asphalt pavement is milled and a concrete layer is installed. In the 1990s, the first experiences with this procedure were made,11-15 and today these asphalt-concrete interfaces can often be found in road construction.16-20

The bond between asphalt and concrete should have high strength; it should have a high stiffness and little deflection under loading, which should subsequently lead to a longer service life. If the resistance of the bond against cracks in the interface is reduced, the service life will be reduced, which leads to higher costs. In addition, the concrete mixture design has to be optimized for low drying shrinkage and low curling potential to minimize the stresses at the concreteasphalt interface.

The factors that influence a lasting bond are complicated. Steigenberger15 gave a comprehensive overview for the design of asphalt-concrete interfaces with an emphasis on their durability. A significant factor is the design of the surface of the bond, which is realized by mechanical procedures (for example, with water-jet, milling, or cleaning) or changed by means of chemical treatments (adhesive agents) to improve the bond. Such substances and procedures are offered by many companies and not only do not lead (as scientific investigations on the bond between old and new concrete have pointed out) to the desired effect, but may even lead to a bigger crack sensitivity and a weakening of the bond2 in some cases.

The central question is if the bond of the layers can be characterized by the bond properties. A possibility is the pullout test. The obtained maximum force divided by the cross section of the specimen is the so-called adhesive tensile strength. With this procedure, the reason for the obtained result cannot be judged, that is, whether the bond separation occurs by brittle or ductile fracturing, and whether a small amount of energy (brittle fracture) or a large amount of energy (ductile fracture) is necessary for the separation of the bond. Hence, the adhesive tensile strength test is an insufficient method to characterize the bond of materials and determine the crack resistance, as formulated clearly by Schulz.21 Nevertheless, this procedure was used in different national and international standards, and no better alternative was offered in the technical literature in the past.

In 1986, an easy test method for the fracture mechanical characterization of materials and the material bond was developed,22,23 which was called the wedge splitting test. This method is described comprehensively in the following and does not have the aforementioned disadvantages of the adhesive tensile strength test. The wedge splitting test is able to distinguish between brittle and ductile material separation. From the achieved load-displacement curve, the slope of the curve that describes the elastic properties, the notched-bar tensile strength following the adhesive tensile strength, and the fracture energy and crack resistance, respectively, are obtained. Hence, with one measurement, several mechanical and fracture mechanical properties are obtained, which is very profitable.23,24 For finite element simulations as well as further ratings, these values are reasonable and necessary. The crack opening mode (Mode I) in addition to the shear mode (Mode II) of the bond is tested in road construction by material scientists, as it is done in Germany and Austria according to the method of Leutner.25 Especially in shaped surfaces of interfaces, the crack is not a shear mode crack (Mode II) but runs at an angle to the shaped surfaces in the crack opening mode (Mode I) when external shear is applied. In the shear test according to Leutner25 there is no pure Mode II crack, but a mixed mode (I + II) crack that is predominantly Mode I.26,27 Hence, the test according to Leutner25 does not seem adequate, and therefore has not been carried out in the present investigation.

In this work, the bond properties and the crack resistance of different interfaces between asphalt and concrete were examined, and different chemical pretreatments were tested and assessed. From this study, a better insight into the mechanisms of cracking of asphalt-concrete interfaces was obtained and important factors on the mechanical properties of the bond were detected.

From this basic knowledge, strategies for the advancement and optimization for practical applications of thin whitetopping can be derived.

RESEARCH SIGNIFICANCE

Whitetopping and ultra-thin whitetopping is used for repairing rutted asphalt pavements. Scientific investigations about the bond asphalt-concrete are scarce. Only the pullout test is performed, but this method does not allow fracture mechanical assessment of the bond, which is necessary for computational mechanics. In this study, different pretreatments of milled asphalt are investigated at different temperatures by means of the wedge splitting test, which allows determination of mechanical and fracture mechanical properties. Choosing the appropriate pretreatment increases the crack resistance of the bond and leads to a reduction of costs because of a longer life cycle.

TEST METHODS AND THEIR EVALUATION

Pullout test

The adhesive tensile strength of the bond is determined by pulling out a glued stamp plunger, as used in this study, or sometimes by means of mechanical grips. The diameter of the stamp plunger is standardized in Austria to 50 mm (1.97 in.), hence for the correct load transmission, appropriate specimens can only be slightly larger. A two-component epoxide adhesive was used.20 Loading is force-driven, and testing was performed either on the site by spot drilling the pavement, as shown in Fig. 1(a), or in the laboratory by taking drill cores, as shown in Fig. 1(b). Only the adhesive tensile strength is determined-no fracture mechanical properties can be obtained.

Wedge splitting test according to Tschegg22

Figure 2 schematically shows the arrangement of the specimen and the loading device of this testing method.22,23 This method greatly reduces the probability for unstable crack propagation in cube-shaped and ashlar-shaped test bodies as well as in drilling core specimens using simple mechanical or hydraulic test machines with displacement control. It is especially suitable for carrying out fracture tests in material bond breaking in a quasi-brittle or brittle manner.28

The direct way of the forces as well as their short ways in the test equipment and in the test specimens make the test equipment extremely stiff-only very little energy is stored in the test equipment and, hence, when testing brittle materials, a stable crack propagation can be achieved.

The loading device, consisting of two load transmission pieces with movable rolling bodies and a slender wedge, is inserted into the groove of the ashlar-shaped test body. The force is transmitted from the wedge to the specimen by rolling bodies with such low frictional forces that they can be neglected.23

The force F^sub M^ from the test machine is transmitted by a slender wedge (favorable wedge angles are 5 to 15 degrees) to a high horizontal force F^sub H^ and a small vertical force component FV during the fracture test. While the force F^sub H^ leads to Mode I tensile fracturing with some bending portion, F^sub V^ stabilizes the crack propagation in the plane, which is defined by the starter notch and the linear-shaped support of the specimen.23

At the horizontal plane of the splitting force F^sub H^, electronic displacement gauges are mounted at both ends of the rectangular groove by means of an aluminium frame (refer to Fig. 2(b)), which measures the load displacement d (crack mouth opening displacement [CMOD]). The two displacement gauges allow measuring an average value of the load displacement on both sides of the specimen and serve as crack propagation detectors.23 If the crack does not propagate vertically, the test is invalid and therefore not used for any analysis.

The force F^sub M^ and load displacement d are recorded during the test. From F^sub M^, the horizontally acting splitting force F^sub H^ can be calculated with the wedge angle a according to the equation

... (1)

The load-displacement curve (F^sub H^ as a function of CMOD or δ) describes the dependence of the splitting force on the load displacement. This curve characterizes the fracture behavior of a layer bond completely if the crack propagates along the interface. The following typical mechanical and fracture-mechanical properties can be obtained from this curve in an easy way:

1. From the maximal force F^sub max^, a notched-bar tensile strength σ^sub KZ^ can be derived with the help of the linear-elastic theory (also refer to Reference 23). This value agrees in some respect with the adhesive tensile strength of the pullout test; and

2. The area under the load-displacement curve corresponds to the fracture work G, which is determined by means of numerical integration. If G is divided by the normal projection of the fractured surface A, the specific fracture work or energy G^sub f^ is obtained. This is a fracture-mechanical material property that is independent of size and shape of the test specimen and is a measure of the resistance against crack propagation, provided that the specimen dimensions are not chosen too small (size effect).24,29

EXPERIMENTAL PROGRAM

East of Austria, an asphalt access road to a mixing plant was milled (maximum aggregate particle size of the asphalt: 22 mm (0.87 in.); more information on the asphalt composition was not available), the residual thickness was 100 mm (3.94 in.). The milling pattern was 8 mm (0.32 in.) deep and 12 mm (0.47 in.) wide. Milling was performed in December 2004; the 100 mm (3.94 in) thick concrete layer was applied 3 days after milling.

The following pretreatments of the milled asphalt surface were studied:

1. Cleaning with high-pressure water, no adhesion agent was applied. Immediately before concreting, water was broomed in place. The bond is called "no" in the following;

2. Cleaning with high-pressure water as an adhesion agent, cement grout (water-cement ratio [w/c] = 0.5) was broomed in place. The layer thickness was smaller than 0.2 mm (0.008 in.). Immediately afterward, the concrete was installed. This pretreatment is called "cement grout" in the following;

3. Cleaning with high-pressure water as an adhesion agent, a mixture of cement grout and plastic dispersion was broomed in place (labeled "cement grout + dispersion" in the following). The thickness of the layer was less than 0.3 mm (0.012 in.). Immediately afterward, the concrete whitetopping was performed; and

4. Cleaning with high-pressure water as an adhesion agent, plastic dispersion was broomed in place (labeled "dispersion" in the following). The thickness of the layer was less than 0.3 mm (0.012 in.). Immediately afterward, the concrete whitetopping was performed.

The concrete layer (two aggregate fractions: 0 to 4 mm [0 to 0.16 in.] and 4 to 8 mm [0.16 to 0.32 in.]; cement content: 450 kg/m^sup 3^ [3.76 lb/gal.]; w/c: 0.39; degree of compactibility: 1.58; air content: 5.2%; compressive strength: 50 MPa [7252 psi]; and splitting tensile strength: 4.4 MPa [638 psi]) had a thickness of 100 mm (3.94 in.). The cross section of the whitetopping pavement is shown in Fig. 3 including the dimensions. After approximately 4 days, the pavement was opened for traffic.

Test specimen

After approximately 4 months of exposure to traffic load, test specimens were taken from the pavement. Drilling cores with a diameter of 300 mm (11.8 in.) and a height of approximately 200 mm (7.9 in.) were taken first. Prismatic specimens for the adhesive tensile strength test were cut out of these drilling cores and the metal cylinders were glued to stiffen the sides of the specimens (refer to Fig. 4(a)). The prismatic specimens for the wedge-spitting test were sawed out from the drilling cylinders, too. These test specimen dimensions are in Fig. 4(b).

The prismatic specimens for the fracture-mechanical characterization of the old asphalt and the concrete were prepared in the same way.

Test conditions

The adhesive tensile strength test was performed forcedriven at a speed of 100 Newtons per second (N/s). The ambient temperature was approximately 20 °C (68 °F). Three measurements were performed for each pretreatment, and then the average value was calculated.

The wedge-splitting test according to Tschegg22 is displacement controlled. The horizontal splitting speed was 0.5 mm (0.02 in.) per minute. The wedge angle was 10 degrees. The testing temperatures were -10, 0, 10, and 22 °C (14, 32, 50, and 71.6 °F). The specimens were placed into a freezer at the desired temperature. For the test itself, no climate chamber was necessary because the testing time was only a few minutes and had no significant influence on the measurement. Two identical tests were performed in each case, and then the average values were calculated.

RESULTS

Pullout test

Figure 5 shows the adhesive tensile strength for the different pretreatments. The adhesive tensile strengths for no, cement grout, and cement grout + dispersion are approximately 1.2 MPa (174.0 psi), and it seems that the two pretreatments do not increase the adhesive tensile strength compared with no. The value for dispersion, which is 0.8 MPa (116.0 psi), is even lower than the others. The measured values scatter strongly-this is another disadvantage of this test.21

In the pullout test of no and cement grout, the crack often did not form in the interface between asphalt and concrete, but in the bond of the affixed steel stamp and the asphalt. The adhesive tensile strength of an asphalt-concrete bond seems to be quite higher than the bond of steel, glue, and asphalt.

The assessment of the fractured surfaces is as follows: In the case of no and cement grout, the fractured surfaces consist of approximately 50% asphalt-like structure (the surface is more black than white = asphalt structure), approximately 25% concrete-like structure (the surface is more white than black = concrete structure) and 25% broken aggregate. In the case of cement grout + dispersion and dispersion, the fractured surfaces contain approximately 50 to 80% white dispersion film mixed with cement grout, and the rest is asphalt-like structure and broken aggregate.

Wedge splitting test

Figure 6 shows the notched-bar tensile strength σ^sub KZ^ at temperatures for -10, 0, 10, and 22 °C (14, 32, 50, and 71.6 °F), except concrete where the temperatures were -10 and 10 °C (14 and 50 °F). With increasing temperatures, the notchedbar tensile strength decreased. The highest notched-bar tensile strength of the bond was achieved at -10 °C (14 °F) for no with 3.2 MPa (464.1 psi), and the lowest value with 2.6 MPa (377.1 psi) for dispersion. Compared with asphalt with 4.7 MPa (682.7 psi) and concrete with 7.0 MPa (1015.3 psi), the values of the bond amount to approximately 60% or 40%, respectively. At temperatures of 10 °C (50 °F), the values of the bond lie between 1.3 to 1.6 MPa (188.5 to 232.1 psi) and indicate a higher strength compared with asphalt with 2.0 MPa (290.1 psi). The values of the bond are 75% in comparison to asphalt and 20% to concrete with 7.5 MPa (1087.8 psi).

Figure 7 shows the specific fracture energy Gf at temperatures for -10, 0, 10, and 22 °C (14, 32, 50, and 71.6 °F). The specific fracture energy of concrete is approximately 200 J/m2 (13.7 ft.lbf/ft^sup 2^) and hardly changes with temperature. Asphalt has a Gf value of 420 J/m2 (28.8 ft.lbf/ft^sup 2^) at -10 °C (14 °F) and is thus much more ductile than concrete. At 10 °C (50 °F), the maximum is reached at approximately 1000 J/m2 (68.5 ft.lbf/ft^sup 2^) and then drops to a value of 300 J/m2 (20.6 ft.lbf/ft^sup 2^) at 22 °C (71.6 °F). The specific fracture energy of the interface is lower than that of asphalt for every test temperature. The highest values of the interface fracture energy are reached for no, and the lowest values are reached for dispersion.

The fractured surfaces of the specimen tested with the splitting wedge method are substantially larger in comparison with the pullout test and, hence, give more information for the assessment of the progress of the fracture. The fractured surfaces of the specimen of no and cement grout tested at -10 °C (14 °F) show an asphalt structure with approximately 70% of the fractured surfaces, with the rest being 20% of broken aggregate and 10% with concrete structure. For dispersion and cement grout + dispersion, the fractured surfaces are characterized by a white film, which comes from the dispersion, which is approximately 80% of the fractured surface. The rest consists of broken aggregate and a small part of asphalt structure.

At 0 °C (32 °F), the fractured surfaces of no and cement grout show a 40% asphalt structure with 20% of broken aggregates and 40% of concrete structure. For cement grout + dispersion, the fractured surfaces is the same as for -10 °C (14 °F). For the dispersion treatment, the fractured surface consists of approximately 50% dispersion film, 40% of asphalt structure, and the rest is broken aggregates.

The fractured surfaces of no, cement grout, and cement grout + dispersion do not change in the test temperature range of 10 to 22 °C (50 to 71.6 °F) any more. The fractured surface is a little rougher and more jagged compared with the temperature range of -10 to 0 °C (14 to 32°F). In the case of dispersion, the crack runs for 70% in the asphalt structure, and the surface is also more jagged.

DISCUSSION

Adhesive properties and fracture energy of bond

Figures 6 and 7 depict the measuring results obtained with the wedge splitting test at certain temperatures. At lower temperatures than investigated, the fracture mechanical properties of concrete and asphalt remain almost constant. At higher temperatures than investigated, the properties of concrete also remain almost constant; asphalt becomes plastic and very ductile so that fracture mechanical investigations do not seem to make sense.26 Two specimens of each bond (an example of each pretreatment) were tested at different temperatures. The low number of the tested specimens can lead to random errors and, hence, to larger deviations because of inhomogeneous material and different layer thickness of the interface. Hence, it would be useful to test a larger number of specimens (at least three) in the future. Nevertheless, material properties can be identified as a function of the test temperature, and values in between may be estimated.

Figure 8 shows the notched-bar tensile strength σ^sub KZ^ as a function of the temperature (from -10 to 22 °C [14 to 71.6 °F]). The notched-bar tensile strength values are highest in the range of -10 to 0 °C (14 to 32 °F). The bond is strong (approximately 3 MPa [435 psi]), however, how large the resistance against a crack in the interface is cannot be determined. Furthermore, how distinct the softening behavior of the bond is cannot be determined.

From 0 to 10 °C (14 to 50 °F), the notched-bar tensile strength drops to half of the value-only 0.7 MPa (101.5 psi) at 22 °C (71.6 °F). This is a typical decrease of the notchedbar tensile strength or the adhesive tensile strength that was also already found in interface bond of asphalt + asphalt.26 The behavior of the specific fracture energy G^sub f^ is quite different (refer to Fig. 9). The low specific fracture energy values are typical for low temperatures and point to a low resistance against crack propagation-the interface is brittle. In the medium temperature range (5 to 15 °C [41 to 59°F]), the specific fracture energies are high, the interface is tough, and they lead to large crack propagation resistance. At ambient temperature, the specific fracture energy drops again, and though the interface is ductile, the interface cannot resist the crack opening force, and crack propagation takes place in contrast to 5 °C (59 °F), where the crack did not grow.

The specific fracture energy of concrete changes only slightly in the entire temperature range, as was expected (refer to Fig. 9). The mechanical properties of the asphalt change extremely with temperature. The course of the curve in Fig. 9 shows similar behavior as that of the bond, which, however, has substantially lower fracture energies.

It must be pointed out that the overall trends of the notch tensile strength and the specific fracture energy in dependence of the temperature are completely different (Fig. 8 and 9). Notched-bar tensile strength or adhesive tensile strength values cannot assess the real fracture mechanical quality of the bond in practice, but the crack propagation resistance can be derived from the force displacement diagrams.

If damage and crack growth in the interface are simulated with finite element methods, it is necessary not only to determine the notched-bar tensile strength or adhesive tensile strength, but also the fracture energy and, if possible, the elastic and plastic behavior. This is another reason why the tests on the bond should be carried out with the wedge splitting test instead of the pullout test. This experimental procedure requires only a little more expenditure, but the output of information is much more detailed, that is, several mechanical and fracture mechanical material properties can be determined.

When comparing the fracture mechanical properties of the interfaces with those of asphalt or concrete specimens (Fig. 8 and 9), the values should be real material properties. A rough estimation indicates that the plastic zones of the tested specimen were small. Hence, the values can be used as material properties.

Characterization of fractured surfaces

Using dispersion or cement grout + dispersion as an adhesive agent, the fractured surfaces show a white layer. The layer thickness is thicker than with cement grout. On one hand, the dispersion is not absorbed by the concrete and the asphalt, and, on the other hand, this layer is brittle and makes for a glassy structure.30 Therefore, the crack finds its quickest way with lowest energy in the interface and therefore propagates along this hard layer. In the fracture process zone, less microcracks form in the asphalt and concrete, and the crack resistance is small, as the measuring results have shown.

If the bond is tested fracture-mechanically after the asphalt layers have been compacted by rollers, a few aggregate particle fractures are detected at low temperatures and none at ambient temperature.26 The number of aggregate particle fractures also decreases in the present asphalt concrete bond with increasing temperature. At ambient temperature, however, there are still broken aggregate particles, and, hence, the fracture appearance is different from that of bond after compacting.

Especially in winter, the bitumen has similar mechanical properties as the aggregates themselves, and if a pavement is milled, the surface of the pavement is damaged, which leads to partially fractured aggregate particles. At higher temperatures, such as 22 °C (71.6 °F), broken aggregate particles may still be observed, although the asphalt is quite plastic. It is supposed that fracturing of the aggregate particles results from milling at low temperatures. If milling is performed at ambient temperature, damage is less severe and the number of the broken aggregate particles is smaller. Further research on the damage of milled pavements could be considered.

At higher temperatures, the fractured surfaces are rough and cleft in comparison to lower temperatures, where they are smoother and contain less broken aggregate particles. Because of the similar mechanical properties of the aggregates and the bitumen at low temperature, the aforementioned fracture appearance is observed. At higher temperatures, the bitumen is more plastic, and the aggregate particles are pulled out of the crack surface and, hence, do not break.

Increasing fracture resistance of bond

The characterization of the fractured surfaces showed that the crack partially runs in the asphalt structure, in the broken aggregate particles, and/or in the dispersion layer at all temperatures. Therefore, the mechanical and fracture mechanical properties of the asphalt concrete bond will be compared with those of asphalt in the following.

The specific fracture energy G^sub f^ and the notched-bar tensile strength σ^sub KZ^ are normalized for the different bond types to the values of the asphalt (Gf Asp and σ^sub KZ Asp^)(G^sub f^/G^sub f Asp^ and σ^sub KZ^/σ^sub KZ Asp^). At all temperatures, the normalized σ^sub KZ^ value (adhesive tensile strength values) of every bond is approximxately 50 to 70% of the maximum value. The σ^sub KZ^ values are in the range of asphalt. Hence, an increase of the notched-bar tensile strength or the adhesive tensile strength seems to be very costly.

The σ^sub KZ^ values, however, are not relevant for the cracking in the bond, rather, it is the specific fracture energy G^sub f^ that characterizes the crack propagation. The normalized G^sub f^ values range from approximately 25 to 40% of the homogeneous asphalt. These values are much more relevant for road construction purposes and characterize the crack propagation properties much better. Hence, it is reasonable that the fracture energy G^sub f^ should be raised.

For an old-new concrete bond, cracks in the interfaces are a well-known problem and were comprehensively examined by means of mechanical and fracture-mechanical tests.30-32 Tschegg and Stanzl30 investigated different chemical pretreatments, for example, epoxy mortar, hard and soft dispersions, and cement paste, and it was shown that the relation G^sub f^/G^sub f Concr^ was 15 to 40%. Despite further technical development, the crack propagation resistance of the bond could not be increased by chemical treatments. Progress may be achieved by special shaping of the surface of the old-new concrete interface.32 If it is adjusted to the maximum aggregate particle size of the new concrete, the specific fracture energy is increased.33 The substrate (old concrete) is formed by milling such that the width of the milled grooves equals the maximum aggregate particle size. Then the crack has to take the longest way in the interface and needs maximum fracture energy. With the optimum surface shape, a fracture energy being 90% of the concrete should be reached.

First experiences were obtained15 with surface shaping of the asphalt concrete bond (thin whitetopping): the asphalt did not have any surface shape, was shaped by means of water jetting, or was milled very finely. The relation of G^sub f^/G^sub f Asp^ lies at 10 to 40% and is similar to the bond of old and new concretes. Hence, it would be useful to influence depth and width of the grooves when milling pavements to obtain the maximum fracture energy.33

With an optimum surface roughness of the asphaltconcrete bond, high crack resistance will be reached in a temperature range from 5 to 20 °C (22.5 to 68 °F). At -10 °C (14 °F) and lower, however, it is to be expected that the asphalt is brittle and glassy and, hence, the crack no longer runs along the shaped interface, and brittle cracking consuming much less energy occurs.

SUMMARY AND CONCLUSIONS

Interfaces between milled asphalt and concrete as used for whitetopping with different pretreatments (without an adhesive agent no; with cement grout; with cement grout + dispersion; and with dispersion as adhesive agent) have been tested at different temperatures, -10, 0, 10, and 22 °C (14, 32, 50, and 71.6 °F), by means of the pullout test and the wedge splitting test according to Tschegg.22 The following results have been achieved:

1. The adhesive tensile strength obtained from the pullout test is 1.2 MPa (145.0 psi) for no, cement grout, and cement grout + dispersion, and 0.8 MPa (116.0 psi) for dispersion. The measurement values scatter strongly;

2. The notched-bar tensile strength σ^sub KZ^ values of the wedge splitting test are summarized in Table 1;

3. The specific fracture energies obtained from the wedge splitting test are summarized in Table 2;

4. A visual rating of the fractured surfaces is given in Table 3. The surfaces are more rough and cleft for all pretreatments compared with no pretreatment;

5. The pullout test allows the determination of the adhesive tensile strength only, which is insufficient for a fracture mechanical characterization, whereas the wedge splitting test-with only slightly more experimental effort-gives mechanical and fracture mechanical properties as well;

6. The adhesive tensile strength and notched-bar tensile strength, respectively, and the specific fracture energy show a completely different behavior as a function of temperature. The notched-bar tensile strength decreases with increasing temperature, and the specific fracture energy has a pronounced maximum in a medium temperature range;

7. The crack resistance is higher for no pretreatment than for any chemical pretreatment;

8. As a matter of the properties of the dispersion layer, the crack finds the fastest way through the interface with lowest energy consumption and runs along the hard layer. The specific fracture energy is small;

9. It is supposed that milling leads to damaging the aggregate particles; and

10. The crack resistance could be increased by an optimal surface roughness of the milled pavement, that is, the depth and width of the grooves should be adjusted to the maximum aggregate particle size of the concrete.33

ACKNOWLEDGMENTS

The authors wish to express their gratitude and sincere appreciation to the Austrian Cement Association for financing parts of this research work.