1. Introduction
Epoxy asphalt concrete (EAC) is composed of asphalt, epoxy resin, a curing agent, and aggregates, and is characterized by good waterproofing performance, rutting resistance, high strength, and fatigue resistance. EAC is widely used in orthotropic steel bridge deck pavement engineering [
1,
2]. The steel plates of a steel bridge exhibit larger deformation under traffic loading than a concrete bridge deck or ordinary pavement. Hence, a steel bridge deck pavement should have good deformation compatibility, which requires an EAC with high deformation resistance. The traffic load causes flexural deformation in the steel bridge deck pavement, generating large shear and tensile stresses in the pavement that make the mixture prone to fatigue cracking [
3,
4]. After cracks appear in the pavement, water can enter the interior of the pavement along the cracks and accelerate pavement damage. Therefore, it is very important to study the factors that influence EAC flexural cracking behavior.
Mineral materials, which include aggregates and fillers comprising over 90% of the total mass of an asphalt mixture, significantly affect the mixture performance. Aggregates interlock to form a supporting skeletal structure in an asphalt mixture. Fillers are mixed with asphalt into a mastic that fills in the skeletal voids, which is the key to ensuring the cohesion of the mixture [
5,
6]. Kollmann et al. [
7] used a cohesive zone model to simulate the initiation and propagation of microcracks in an asphalt mixture in an indirect tensile test. The fracture failure of the asphalt mixture was found to result from the continuous development and expansion of microcracks and microvoids in the concrete mixture under loading. The existence of air voids results in the development of more distinct stress concentrations, which in turn, lead to reduced load-bearing capacities. The damage distribution showed that the aggregate–mortar interface exhibited more significant damage than the bulk fracture in the mortar itself. As the temperature increases, the ductility of the asphalt mortar increases. The mixture showed pronounced deformation before damage, and the damage distribution was more uniform. Low temperatures such as −10 °C were associated with significant localized damage in the mixture due to brittle fracture behavior. Inside the asphalt mixture, the interfacial bond strength between the aggregates and the asphalt is an important factor in the flexural cracking behavior of the asphalt mixture. Studies have shown that using angular aggregates with a rich surface texture and high roughness can enhance the interfacial bond strength between the aggregate and the asphalt, which in turn significantly improves the fatigue cracking resistance of the asphalt mixture [
8,
9,
10]. Haskett et al. [
11] analyzed the shear friction and interlocking behavior of aggregates. The mineral material skeleton formed by the interlocking of coarse and fine aggregates was found to be conducive to improving the internal friction and load-bearing capacity of the material, thereby increasing the resistance to bending and shear stresses from traffic loading. Golalipour et al. [
12] showed that improving the gradation design and reducing the air void percentage and number of voids in a mineral aggregate enhanced the rutting resistance of an asphalt mixture.
The aggregate strength characteristics also affect the flexural cracking behavior of asphalt mixtures. Moreno and Rubio [
13] carried out an asphalt fatigue cracking test (UGR-FACT) to evaluate the fatigue cracking behavior of a diabase asphalt mixture and a limestone mixture. The test results showed that the diabase aggregates did not fracture because of a high crushing strength, and the corresponding asphalt mixture had a higher resistance to fatigue cracking. In comparison, the limestone asphalt mixture exhibited a few aggregate fractures and a short fatigue life, indicating that high-strength aggregates provide higher resistance against crack propagation. Mahmoud [
14] integrated the discrete element method with image processing technology to model and analyze the modulus, compressive strength, and indirect tensile strength of five aggregates including granite, hard limestone, soft limestone, gravel, and sandstone. The effect of the aggregate strength on the fracture property of the asphalt mixture was evaluated, and the results showed that the crushing of coarse aggregates reduced the internal force between particles. Thus, the mechanical properties and the load-bearing capacity of the asphalt mixture were reduced. Aliha et al. [
15] investigated the influence of limestone and siliceous aggregates on the fracture performance of an asphalt mixture using a semicircular bending (SCB) test. The limestone aggregates had a higher strength and stiffness and a correspondingly higher fracture resistance than the siliceous aggregates. Li and Marasteanu [
16] conducted an SCB fracture test to compare the influence of using granite and limestone on the fracturing of an asphalt mixture. Granite is more resistant to crushing than limestone. The test results showed that while a significant portion of the fracture passed through the aggregate particles in the mixture made with limestone, most of the fracture in the mixture made with granite passed through the interface with the mastic. Under the same low-temperature conditions, the granite asphalt mixture had a higher peak load and fracture energy and exhibited higher fracture resistance than the limestone asphalt mixture.
In addition to the shape and surface roughness of the aggregates affecting the performance of the asphalt mixture, the mechanical properties of the aggregates also influenced the mixture cracking resistance, which has attracted the attention of researchers. Vishalakshia et al. [
17] performed tests to comparatively evaluate the strength and cracking behavior of ordinary and high-strength concrete containing five different types of aggregates. The aggregate type was found to strongly influence the strength of high-strength concrete, but had no significant influence on ordinary-strength concrete. Generally, asphalt mixtures, especially epoxy asphalt mixtures, exhibit clear elastic characteristics and higher flexural strengths at low temperatures, similar to cement concrete. The influence of the relative strength of aggregate strength and mastic strength of cement concrete on cracking characteristics has reference significance. Ipek [
18] evaluated the effect of artificial aggregates on the mechanical properties, fracture parameters, and bond strength of cement concrete. The mechanical and fracture properties of concretes with similar strengths were significantly affected by the incorporation of artificial lightweight aggregates. Thus, as the concrete strength increases, the influence of aggregates becomes increasingly significant.
A crosslinking reaction between epoxy resin and asphalt can be used to create a three-dimensional network. The high cohesion and strength of epoxy asphalt significantly improve the properties of EAC including the strength and viscoelasticity over those of ordinary asphalt mixtures [
19]. In the literature, EACs have been designed and evaluated mainly considering the properties of an ordinary asphalt mixture. Little attention has been paid to the effects of aggregate strength characteristics on the EAC fracture resistance or the corresponding correlation.
Considering the influence of aggregates on the properties of ordinary asphalt mixtures and cement concrete, aggregates are expected to significantly affect the EAC properties. The effects of aggregate properties on the EAC mechanical properties were evaluated in this study by comparative tests on three types of EACs containing aggregates of different strengths. First, a Marshall stability test was carried out, followed by a three-point bending test to evaluate the flexural tension resistance and impact toughness of the three types of EACs. Finally, a single-edge notched beam bending test was used to evaluate the fracture toughness of three types of EACs. After the beam bending tests, data from the specimen fracture cross sections were collected, the aggregate fracture area was statistically quantified, and the fracture morphology of the aggregate particles was analyzed to comprehensively evaluate the effect of aggregate strength on EAC flexural cracking behavior. To evaluate the effects of cyclic loading, four-point bending fatigue tests were conducted on three types of EACs.
3. Methods
3.1. Marshall Stability Test
The Marshall stability test is used to characterize the compression resistance and deformation properties of an asphalt mixture at high temperatures. This test empirically reflects the compressive load-bearing capability of EACs with different stone aggregates and was used to comprehensively evaluate the internal cohesion of the mixture and the shear friction of the aggregates. In the experiment, four specimens were tested for each type of EAC.
3.2. Stone Beam Bending Test
An epoxy asphalt pavement is subjected to relatively high bending stress; consequently, high flexural tensile strengths are required for the aggregates inside the mixture. Aggregate strength indexes include the compressive strength and the crushing value. As aggregate fracture failure is mainly determined by the flexural tensile properties, the flexural tensile strength of aggregate source stone has considerable engineering significance [
21,
22]. A three-point bending test was carried out on the aggregate source stone to determine the flexural tensile strength characteristics of the different aggregate types. Three types of raw stone materials, diabase, granite, and limestone, were cut into prismatic beam specimens with dimensions of 250 mm (length) × 30 mm (width) × 35 mm (height), as shown in
Figure 2. The three-point stone beam bending test was conducted according to “Standard Test Methods of Bituminous Mixtures for Highway Engineering” (JTG E20-T0715-2011) using a MTS 810 material test system (MTS Systems Co., Eden Prairie, MN, USA) at a test temperature of 15 °C and a loading rate of 50 mm/min. In the experiment, six specimens were tested for each type of stone.
3.3. Three-Point Bending Test
A three-point bending test was conducted to evaluate the effect of the three aggregate types on the EAC flexural cracking behavior. Under continuous loading of the beam specimen, cracks formed and developed inside the mixture, eventually resulting in specimen fracture. The impact toughness (
AK) of a material reflects its ability to resist deformation and fracturing under an impact load. Energy is dissipated from the fracturing of a beam specimen under the impact load. The load–displacement curve of the test data was used to calculate the EAC impact toughness by the area integral method using Origin software (Origin 2018, OriginLab, Northampton, MA, USA). The area under the loading portion of the load deflection curves, up to the maximum load, was measured from the curves presented in
Figure 3. Zhang [
23] conducted three-point and four-point bending tests and determined a good positive correlation between
AK and fatigue performance in an EAC. In the present study, EACs with three types of aggregates were formed by the slab roller compacting method. After curing, the EACs were cut into beam specimens with dimensions of 250 mm in length, 30 mm in width, and 35 mm in height. The fracture property differences among the EACs, the base asphalt mixture, and the SBS asphalt mixture with the same diabase aggregate were investigated. The tests were carried out according to the “Standard Test Methods of Bituminous Mixtures for Highway Engineering” (JTG E20-T0715—2011) at test temperatures of 15 °C and −10 °C and a loading rate of 50 mm/min. The test device is shown in
Figure 4. In the experiments, six specimens were tested for each mixture, as shown in
Figure 5.
3.4. Single-Edged Bending of Mixtures
The fracture toughness of a mixture characterizes its ability to withstand crack propagation and is an indicator of the fracture performance of the mixture. Studies on asphalt mixtures have shown that the crack resistance increases with the fracture toughness [
24,
25,
26]. The fracture toughness of asphalt mixtures is significantly influenced by the material composition characteristics and is commonly characterized by the fracture toughness index (
KIC), which can be determined by a three-point bending test with a single-edge notched beam [
27]. The maximum load
Pmax is obtained experimentally and used to calculate
KIC using the following formula recommended by the American Society for Testing Materials (ASTM):
where
S is the specimen span (
S = 200 mm);
is the length of the prefabricated notch;
h is the specimen height; and
b is the specimen width.
The beam specimens of the different EACs were 250 mm in length, 30 mm in width, and 35 mm in height. The prefabricated notch had a depth of 5 mm in the height direction and a width of 1 mm, so that the ratio of the notch depth to the specimen height (a0/h) was approximately 0.143. The test was carried out at −10 °C and 15 °C at a loading rate of 1 mm/min. In the experiment, six specimens were tested for each EAC.
3.5. Digital Image Processing
The cross section of the asphalt mixture beam after flexural fracture contained voids, asphalt mastic, and aggregate. The aggregate fracture was related to the EAC fatigue cracking performance using digital image processing (DIP) to statistically quantify the aggregate fracture area on the specimen cross section. ImageJ is a Java-based image processing software developed by the National Institutes of Health (Version: 1.52q, NIH, Bethesda, MD, USA). The scale and units are set, and the image is processed using binarization, enhancement, and contour feature extraction. Then, the number of aggregates on the cross section is automatically extracted, and the fracture area of aggregates is statistically quantified, as shown in
Figure 6. A high-definition digital camera was used for image acquisition, and the contour of the binarized image was enhanced using Photoshop software (Photoshop CC 2017, Adobe Systems, Inc., San Jose, CA, USA). In the experiment, six specimens were tested for each EAC.
3.6. Four-Point Bending Fatigue Test
Fatigue tests were conducted to understand the influence of cyclic loading on the flexural cracking behavior of epoxy asphalt mixtures on a steel bridge deck pavement. EAC samples were subjected to a four-point bending fatigue test assessed according to the American Association of State Highway Transportation Officials (AASHTO) T 321-03 at a temperature of 15 °C and a loading frequency of 10 Hz in the strain control mode using a NU-14 fatigue testing machine (Cooper Research Technology Ltd., Ripley, Derbyshire, UK). The failure point is defined as the point in the load cycle at which the specimen exhibits a 50% reduction in stiffness relative to the initial stiffness. The beam specimens used for the fatigue test were characterized by dimensions of 380 mm (length) × 63.5 mm (width) × 50 mm (height) and were obtained by cutting 400 × 300 × 70 mm slabs prepared using a laboratory roller compactor. The test device and beam specimens are shown in
Figure 7. In the experiment, four specimens were tested for each EAC.
The EAC mixture sample exhibited an endurance limit at a strain level lower than 500 με, and its modulus basically became steady upon reaching 60–70% of the initial stiffness, after which there was no significant change in its fatigue trend. On this basis, the fatigue property differences among the EAC samples containing diabase, granite, and limestone were evaluated at strain levels of 600 με, 900 με, and 1200 με. The influences of aggregate strength on the fatigue and rheological properties of the EACs were analyzed.
The initial stiffness and phase angle of EACs were tested at the 50th load cycle in the four-point bending test according to AASHTO T 321-03, at a loading frequency of 10 Hz in the strain of 600 με, and at temperatures of 0 °C, 15 °C, 30 °C, 45 °C, and 60 °C, respectively. The specimens were cured four hours before each temperature test.