3.1. AFM Study
Atomic force microscopy has been used for more than 25 years to study the microstructure of bitumen. Particular attention of the researchers was focused on the so-called “bee-like” structure [
24]. Creep measurements demonstrated that the microstructure of the “bee-like” phase has 40–50% higher stiffness than the surrounding matrix phase [
25]. Studies of the binders’ resistance to low-temperature cracking after applied load showed the appearance of cracks at the “bee”-matrix phase interface [
3], which allowed the visualization of theoretical ideas justifying the possibility of binder cracking under climatic factors and transport loading due to its inherent heterogeneity [
26].
Figure 1a shows the surface topography of bitumen Bit-A, which illustrates the structural heterogeneity of bitumen and the presence of a pronounced microheterogeneity in it. The surface of the bitumen sample examined 2 days after preparation showed the presence of a “periphase” surrounded by “bee-like” structures with a maximum size of 5–7 μm, with a tendency to form “star bee-like” structures. Storage of this bitumen at ambient temperature for 1.5 months led to an increase in the maximum size of the “bee-like” structures by 1.5–2 times—up to 15–20 μm.
The ©ntroduction of APDDR into bitumen primarily had an effect on the formation of a “bee-like” structure.
For MB with 10 wt.% APDDR content, a significant decrease in the length of “bees” to the maximum 2–3 μm was observed (
Figure 1b). When the APDDR content in the MB was increased to 15 wt.% and higher, no “bees” were observed even after 1.5 months of sample storage (
Figure 1c).
The structure of the fragmented PEM particles observed using AFM was compared with the SEM data for the modified binder, where the APDDR content was 15 wt.%. The measurements were carried out 2 days after MB preparation (3 minutes’ mixing at 160 °C).
In
Figure 2a,b, one can see an elongated APDDR particle of the size of about 10 μm. When the image is zoomed in (
Figure 2c,d), it is clearly seen that the particle consists of two different-sized fragments: the upper one is about 4÷5 µm long, and the lower one is about 2 µm. These fragments are connected by strands up to 2µm-long and a few tenths of a micron thick. We can also assume that the upper fragment of this APDDR particle consists of at least three parts. It can be hypothesized that this particle is the decay product of a larger particle, whose fragments were interconnected by strands.
Smaller fragments of the original APDDR particles of 100÷1000 nm are clearly visible in
Figure 3. Similar particles of the same size and similar (self-similar) agglomerative structure were observed in SEM images after washing the MB with solvent [
17].
The images in
Figure 4 illustrate the stage preceding the separation of the 50÷200 nm APDDR fragments from a larger particle (
Figure 4a,b) and the formation of the spatial structure (
Figure 4c,d). The data presented in
Figure 3 and
Figure 4 indicate the formation of a physical spatial network of rubber fragments. Since the distances between the fragments are comparable to their fragment sizes, the percolation threshold’s conditions are satisfied.
Figure 5 shows AFM images of the MB surface (the MB is based on Bit-A and hybrid powder; the Bit-A/hybrid powder is 85/15 wt.%). The measurements were carried out two days after sample preparation.
In a series of zoomed images, the different stages and mechanisms of hybrid particle disintegration can be observed. In
Figure 5a,b at the top left, a round-shaped hybrid particle of 2–3 μm in size can be observed connected by strands with smaller fragments up to 0.2 μm in size. One such fragment is shown in
Figure 5c,d. The appearance of these fragments correlates very well with SEM images of the APDDR fragments [
17]. Elements of an agglomerative structure are also traced.
The AFM images of thin films observed on the surface of the MB with a 90 wt.% Bit-A+10 wt.% APDDR composition are shown in
Figure 6. In our opinion, the separation of such films occurs as a result of multidirectional swelling forces from the surface of microblocks with a denser structure than others. In our opinion, these images have similarities with the fragments of APDDR particles in the form of films observed earlier using electron scanning microscopy [
17].
Figure 6b shows a three-dimensional image of the MB surface, from which we can see that the film is partially located on the bee structure. Thus, it is verified that the films were formed by the rapid decay of PEM particles, because the formation of the “bee” takes longer.
The results of AFM studies confirm the main conclusions drawn based on SEM studies [
17]: already at the early stage of interaction, PEM particles break down into micro- and nanofragments with an agglomerative structure, similar (self-similar) to that of the original modifier particles given in [
17], as well as present in the form of thin films. Additional information obtained on the basis of AFM images concerns the formation of gel structures on the basis of broken PEM particles. Additional information obtained on the basis of AFM images concerns the formation of gel structures on the basis of the broken PEM particles. Such a spatial network of nano- and microfragments should hinder the processes of diffusion and crystallization of the waxes present in bitumen. The result is the disappearance or a significant decrease in the size of “bee-like” formations, which is observed at PEM concentrations above 10 wt.%. A more homogeneous structure of the modified binder is created.
3.2. Rheological Tests
The structure of binders resulting from the interaction of rubber particles (APDDR) and hybrid particles (APDDR-SBS) obtained by high-temperature shear grinding with hot bitumen determines the rheological and, therefore, the performance properties of such binders. Rheological tests according to Superpave standards were performed on aged binder specimens. The aging was carried out for 85 min at 163 °C in a rolling thin-film oven (RTFO) [
27] and 20 h at 100 °C in a pressure-aging vessel (PAV) [
28]. This aging simulates the processes of asphalt concrete mixture preparation and as well as the 7-year operation service of the pavement. Earlier, it was shown that aged samples of PEM-modified bitumen at a PEM concentration of 12–15 wt.% in the binder showed an improvement in bitumen resistance to all types of pavement defects in the whole range of operating temperatures. However, it was noted that reducing the PEM concentration to 10 wt.% and less led to some improvement in low- and high-temperature parameters but worsened some parameters of fatigue-cracking resistance of the modified samples compared to bitumen [
29]. These data justified the PEM concentration in bitumen for rheological tests as 12.5 wt.%. In this case, rheological tests were conducted for unaged bitumen to see the change in binder properties immediately after the introduction of the modifier.
Among the parameters of resistance to rutting the most promising for the characteristic of bituminous binders is considered parameter J
nr (unrecoverable creep compliance), determined in the test for resilience to multiple cycles of creep–recovery [
20]. The parameter J
nr is calculated as the ratio of the average unrecovered creep strain over 10 test cycles to the applied stress level. The second important parameter is the elastic recovery in percent at a given load (R
3.2) calculated as the average value of the elastic recovery for 10 cycles of creep–recovery at a 3.2 kPa stress level.
Table 4 shows the data of unrecoverable creep compliance (J
nr) and elastic recovery at a 3.2 kPa stress level (R
3.2) of the MSCR test for two bitumen and modified binders, the composition of which is given in
Table 3. Additionally, in
Table 4 are the data for the upper operating temperature of the no-aged bitumen binders, which is defined from a dynamic (oscillatory) shear test [
21] as the temperature at which rutting parameter G*/sinδ equals 1 kPa.
A test for multiple stress creep recovery was conducted to characterize the samples in terms of resistance to rutting during the operation of the pavement in the summer under the influence of moving traffic. The introduction into both bitumen samples of a 12.5% powder elastic modifier (PEM) led to a sharp decrease in unrecoverable creep compliance (J
nr) at almost all load levels. Increasing the SBS content of PEM reduced J
nr and increased elastic recovery (R) (
Table 4 shows the elastic recovery (R
3.2) for the 3.2 kPa stress level) both compared to bitumen and to SBS-free APDDR. MSCR tests carried out early (see [
29], for example) for RTFO-aged bitumen samples at 64 °C showed that the elastic response was no more than 5–6%, whereas for modified binders, the elastic response exceeded 70%, which showed that the final formation of the spatial mesh occurs during the time the modified asphalt mixture is brought to the paving site. J
nr values after RTFO aging ensure that the modified binder can be used for pavements with maximum traffic. The good resistance to multiple cycles of creep and recovery at a sufficiently high test temperature (64 °C) of Bit-B, which has a much higher penetration compared to Bit-A, seems to be due to differences in production technology and chemical composition.
The resistance to low-temperature cracking was determined by the ABCD method. The results are presented in
Table 5.
From
Table 5, it can be seen that the specimen-cracking temperature (T
ABCD) decreased with the introduction of PEM. The greatest decrease (9.8 °C) was observed for the hybrid-modified binder with 20 wt.% SBS. Only a slight decrease in fracture temperature (up to 0.5 °C) was observed for the hybrid powder with 5 wt.% SBS compared to the SBS-free APDDR-modified binder, which decreased the T
ABCD of both bitumens by about 4 °C. It has previously been shown that the decrease in the fracture temperature in the ABCD test using PEM compared to bitumen is also characteristic of aged samples [
29]. At the same time, the ABCD data show that at the moment of fracture, the modified samples have a much higher fracture stress (σ) than the original bitumen. Thus, it is confirmed that a sufficiently short mixing time (in this case, 3 min at 160 °C) already leads to the formation of a new binder structure.
Data of
Table 4 and
Table 5 show a significant expansion of the operating temperature range, as well as an increase in resistance to rutting and low-temperature cracking.
The resistance of specimens to fatigue cracking was evaluated using a linear amplitude sweep (LAS) test.
Figure 7 shows the dependences of complex modulus (G*) versus shear strain (γ) recorded during testing at 16 °C. It can be seen that for the modified binders, there was no sharp drop in the complex modulus (G*), which was observed for both bitumen samples at γ of 12–18%. This may indicate the formation of a spatial network in the modified binders, the existence of which is most pronounced during fatigue tests.
Figure 8 shows the dependence tangent of phase angle (tg¦Ä) versus shear strain (γ) recorded during testing at 16 and 7 °C. As can be seen, the modified binders in all cases show less sensitivity to cyclic deformation.
Table 6 shows the number of cycles to failure (N
f) at strains of 2.5 and 5%, calculated according to [
22] based on LAS test data. As can be seen, the modification significantly improved the resistance to cyclic strain. Recall that
Table 6 shows the data obtained for no-aged specimens of the modified binders. After RTFO and after PAV aging, the best results were observed for the modified binder based on hybrid PEM particles.
The comparative rheological tests of bitumen samples and PEM-modified binder produced in conditions close to the temperature–time conditions of road mixture production (3 min mixing at T = 160 °C) and not subjected to additional temperature influence showed that even such a short time of interaction of the hot bitumen and PEM provides an expansion of the performance temperature range of modified bitumen and its increased resistance to cyclic loads.
The improvement in rheological indicators confirms that the rapid degradation of micronized powders of elastic modifiers into micro- and nanofragments leads simultaneously to the formation of a new structure of the modified binder that is more resistant to external influences. It can also be assumed that the improvement in low-temperature and fatigue-cracking resistance may be due to energy absorption and crack growth stopping in the presence of micron and submicron PEM elastic fragments in accordance with the mechanism of increasing impact toughness in plastics [
30,
31,
32].
The results obtained confirm the effective recycling of worn-out tire rubber by the HTSG method in obtaining PEM for further use to modify bitumen directly in the production of road asphalt mixtures or to reduce the time and energy costs for the preparation of modified bitumen binder for road construction.