Currently, the investment in the maintenance of the Spanish roads is low because of the economic crisis. As a result, roads are in a poor state of conservation (1), specifically the surface layer, which is the most exposed to traffic, weather, and other effects of external agents. For these reasons, it is indispensable to improve the design and construction criteria of this layer in order to produce more durable and, thus, more environmentally friendly pavements. In this regard, it is important to consider the deterioration processes taking place when those pavements are exposed to extreme conditions, such as the use of salts during winter maintenance practices, to better understand the pavement damaging mechanisms.

More than 25,000 km of Spanish roads are exposed to winter maintenance, and NaCl is the most common deicer applied (2). However, this product has a negative impact on motor vehicles, transportation infrastructure, as well as on the environment (3). Particularly evident is the effect of this compound on the durability of asphalt pavements. Freeze–thaw cycles cause the expansion of water, which results in the decrease of the pavement’s indirect tensile strength but, combined with salts, especially chlorides, it is also responsible of the hardening of the binder and, therefore, the loss of pavement elasticity. In addition, the combination of freeze-thawing and salting causes either the interfacial damage between binder and aggregate or the fracture of the binder, which produces a greater increase of weight loss (4–5). The latter is considered the main cause of asphalt pavement deterioration, and it is directly related to the low adhesion between the aggregate and the binder in asphalt mixtures (6–9). However, the mechanisms which produce the low adhesion between the aggregates and the binder are still unclear.

The so-called stripping phenomenon, or asphalt-aggregate bond failure, can be produced by a large number of factors, including composition, porosity and roughness of the aggregates, wettability between bitumen and aggregates, chemical composition at the interface (10), exposure aggregate history (e.g., freshly crushed versus days of exposure to environmental weathering after crushing), asphalt viscosity or traffic cyclic loading (11). However, the most accepted hypothesis is that the aggregate surface chemistry is the main responsible of the stripping process (7, 11). Fromm (12) reported that stripping is high in mixes with acidic granites with hydroxylated SiO₄. Water vapor, which is emitted during the production of asphalt mixtures, can form strong hydrogen bonds with siliceous surfaces, forming silanols. This hydrated silica molecules can cause the replacement of the bitumen polar parts (13). Vuorinen and Hartikainen (14) showed a similar behavior in Al₂O₃. In addition, mixtures of aggregates containing alkali metallic elements, like Na and K, exhibit relatively high stripping sensitivity. In contrast, metallic elements, such as Ca, Mg, and Fe, have a good aggregate-binder adhesion (15).

Despite the importance of the aggregate properties in the asphalt mixture durability, they are usually underestimated during road pavement design and production (2). Aggregates are considered the most stable component of the mixture (high resistance to deformation, durability and chemical inertia, 16). In contrast, asphalt binder is chemically more complex and expensive (6€/tone of aggregate versus 500€/tone of asphalt binder, 2), and consequently it gets more attention by producers, technicians and researchers. However, it is important to better understand the influence of the aggregate properties in the asphalt mixture behavior. In this regard, petrographic techniques, such as optical and fluorescence microscopy, can provide adequate information about the process of asphalt mixture damaging, and the interaction between the aggregates and the binder, especially from the aggregate point of view (2).

Optical polarized light microscopy has been widely applied in the study of the physical-mechanical properties of the aggregates (17–21). In addition, fluorescence microscopy has been extensively performed to estimate the damage of building materials, because it can detect the cracks produced by external agents, such as water or extreme temperature values (22, 23). This technique has been applied on concrete (24–29), but it has barely been used in the study of asphalt mixtures. Eriksen (30) used optical and fluorescence microscopy on asphalt mixtures to estimate their homogeneity of compaction. Broekmans (31) evaluated the effect of studded tires on asphalt pavements with fluorescence microscopy, obtaining different failure patterns depending on the type of aggregates used in the mixture. Poulikakos and Partl (32, 33) studied the microstructure of asphalt mixtures by fluorescence microscopy and its relationship with the mechanical effects of compaction and fatigue. The microscopic observations, supported by field data, concluded that the performance quality of mixes depends on the shape of voids, and interlocking and binder adhesion of the aggregates.

The aim of this paper is to evaluate with petrographic techniques the effect of the stripping phenomenon (low binder-aggregate adhesion) on the durability of asphalt mixtures, paying especial attention to the interaction between the aggregates and the binder of the mixtures, and to discuss the use of optical and fluorescence microscopy in their characterization. It is important to point out that Spanish roads are subjected to extreme climatic conditions, very cold winters with frequent maintenance operations, and high temperatures during the summer. Lately, these temperatures can often reach up to 40° C owing of the recurring heat waves registered in the Iberian Peninsula (34, 35). In addition, the most common asphalt mixtures used on surface layers in Spain are open-graded. The presence of voids (up to 12%, 36) and the low rate of rainfalls, produce the retention of the NaCl spread during winter inside the asphalt mixture. NaCl then begins to precipitate from the solution to form crystals within the pores (37, 38), and this crystal growth is capable of producing tensile hoop stresses above 2500 psi (17.2 MPa) (39, 40). For these reasons, the effect of low and high temperatures (winter and summer seasons, respectively), both combined with NaCl, has been taken into account in this paper as they have been poorly studied (41, 42).

With the purpose of estimating the combined effect of the NaCl, extreme climatic conditions and traffic load, two types of asphalt mixtures used at NW Spain were prepared in the laboratory. The mixtures where exposed to freeze-thaw and thermal-stress cycles with a saturated solution of NaCl to simulate the effect of the seasonal thermal variations and the winter maintenance practices on road surfaces. After the durability tests, the samples were tested with the wheel tracking method to simulate the effect of the traffic load.

Finally, the samples treated and tested in the laboratory were analyzed by optical and fluorescence microscopy so as to evaluate the mechanisms of the damage caused. In addition, the surface properties (morphology and composition), as well as the mechanical resistance of the aggregates used in the asphalt mixtures, were also determined to round off the results collected from the petrographic analysis.

The data collected from the petrographic analysis and the mechanical tests performed in the laboratory showed that the surfaces of the asphalt mixtures were damaged by the induced effect of NaCl, extreme climatic conditions, and traffic load.

Both tests, freeze-thawing and thermal-stress with NaCl, produced the deterioration by cracking of the asphalt mixture coarse aggregate with a similar mechanism. The binder acted as an impermeable wall. Thus, NaCl solution could only penetrate in the mixture through the aggregate, producing an intense and quick damage (Figure 5). The penetration of the NaCl solution inside the aggregate was possible because of the presence of previous or initial porosity in the mixture. This porosity was mainly caused by the low adhesion between the aggregate and the binder (Figure 5a), the presence of air bubbles in the binder originated during the mixture production (Figure 5b) and the internal structure and mineral composition of the coarse aggregate (Figure 5d). In addition, the effect of the traffic aggravated the deterioration of the asphalt mixtures exposed to both, NaCl and extreme climate conditions. The deformation of the internal structure of the mixture (Figure 3) caused the cracking of its components, especially the coarse aggregates, which were weakened by the effect of the NaCl and extreme climatic conditions.

In spite of the similar damaging mechanism exhibited by both durability tests, the effect of thermal-stress (high temperature) was more harmful on the mixture surface (Figure 5c). This test affected not only the aggregate, but also the binder. It was observed that the wheel tracking tests after thermal-stress decreased the deformability of the mixtures analyzed (Figure 3). This phenomenon could result from the possible aging of the binder after its exposure to high temperatures (54–56). In addition, petrographic analysis showed the surface loss of the binder (Figure 5c), causing a higher exposure of the aggregate surface to the effect of the NaCl solution and increasing the intensity of the cracking, in contrast to the lower amount of cracks detected after the freeze-thaw cycles.

Even though the effect of NaCl used during winter on asphalt pavements has been widely studied by researchers and Transportation Agencies (3–5, 57–59), no current standards consider their effects in asphalt mixtures. In addition, the effect of that salt during the summer weather conditions has been poorly studied (41, 42). For these reasons, it is recommended to consider both, winter and summer weather conditions, in future studies of road surface durability, especially in countries as Spain, with extreme temperature changes between winter and summer.

The low adhesion between the aggregates and the binder was one of the main degradation mechanisms of the asphalt mixtures. The amphibolite aggregate, with an “a priori” higher quality than the schist aggregate because of its lower water absorption coefficient and higher mechanical resistance (Table 3), showed an intense degradation by the effects of the NaCl, extreme climatic conditions and traffic because of its lower adhesion with the binder. In spite of the importance of this process in the durability of the mixture, the causes of this low adhesion are unclear. The most accepted explanation is that rocks with high silica content have low adhesion with asphalt binders (12–13, 15, 60, 61). Other compositional factors that hinder an optimal adhesion include minerals with K and Na (albite, biotite or muscovite) because of their higher solubility (Table 2; 10, 13). Conversely, minerals rich in non-soluble Ca and Mg, such as actinolite or plagioclase (Table 2) favor the adhesion with the binder (62). However, the amphibolite aggregate showed lower adhesion than the schist aggregate, giving rise to a challenging conundrum. It follows that, in this case, the low adhesion with the binder cannot be explained only by the presence of minerals with soluble ions which can rapidly react with water molecules displacing the asphalt binder (10, 13). Therefore, additional properties from the aggregate surface should be taken into in account in this process.

It is argued that these additional properties comprise the surface roughness (Table 2) and the water absorption coefficient (used as a proxy of the connected porosity), as both parameters are much higher in the schist than in the amphibolite aggregate (see Tables 2 and 3, respectively). This surface roughness and “porosity” would improve the penetration of the binder inside the aggregate contributing to its mechanical interlocking when it hardens (6, 63). In addition, the angularity of the aggregates can also influence the adhesion with the binder. Angular shapes on aggregates increase the potential for asphalt film rupture producing trapped porosity in the mixture, as well as low adhesion between the binder and the aggregate (6, 64). The amphibolite aggregate shape, more angular, could be pivotal in the low adhesion of this material with the binder. For this reason, it is recommended to study parameters such as surface roughness and shape, combined with the mineralogical and chemical composition of the aggregates to better understand the causes of low adhesion with the binder.

Other important factors responsible of the asphalt mixture deterioration were the mineral habit and the fabric of the coarse aggregates. While the cracking of amphibolite coarse aggregates was generally related to this low adhesion (Figure 3a), the schist aggregate damaging was related to the delamination and cracking of its mica domains (Figure 3c). The foliated structure of the schist aggregate was also responsible of the deeper deterioration caused by the traffic effect in the schist asphalt mixture compared to the amphibolite mixture (Figure 3e). Therefore, both factors must be taken into account as selection criteria to improve the mechanical resistance and durability of asphalt mixtures.

Finally, the petrographic analysis (65) is not included in Spanish or European current standards to evaluate the quality of the aggregates used on roads surface (16, 66). However, the optical polarized light microscopy and the fluorescence microscopy, as it has been showed in this paper, are good techniques to evaluate the durability and the quality of the asphalt mixtures, especially the coarse aggregates.

The main conclusions of this study can be summarized as follows: