**1. Introduction**

Currently, the use of sodium chloride (NaCl) is widespread in a large number of countries, as a de-icing agen<sup>t</sup> on winter roads. The impact that salt has on the environment has been extensively studied; the melting salt leads to the mobilization of heavy metals such as lead, cadmium, copper and zinc [1–4], along with increased chlorides (Cl−) [5,6]. However, salt is still used, due to its low cost, versatility and physical properties [7,8]. These physical properties make their use propitious in temperatures up to −21 ◦C, but as García [7] indicated, from −5 ◦C, in order to be more effective, they can be used in combination with calcium chloride (CaCl2).

Shi et al. [9] noted that the value of negative impacts on the environment and on vehicles must be taken into account when calculating the costs of de-icing agents, including NaCl. In this trend of reducing costs and negative impacts, Klein-Paste et al. [10] indicated that the amount of NaCl that is used in practice could be reduced by 40%, to prevent the pavement from being slippery, because the salt spread on the roads causes a weakening of the ice. Another aspect studied along the line of reducing the negative impact of NaCl is the creation of tools to decide more e fficiently when salt should be applied to roads [11,12].

There are also studies on the behaviour of bituminous mixtures in contact with the di fferent de-icing agents. Wang et al. [13] noted that, compared to other anti-icing agents, such as sand and quartz dust, salt does not produce a polished surface. Hassan et al. [14] indicated that the value of Indirect Tensile Strength (ITS) for asphalt mixtures exposed to salt after 25 freeze–thaw cycles is similar to other de-icing agents, such as potassium acetate, sodium formate or urea. However, after 50 cycles, the ITS result is even better for the mixtures exposed to salt than those submerged in distilled water.

Feng et al. [15] studied the impact of salt on the performance of bituminous mixtures when it remains submerged under seawater subjected to freeze–thaw cycles. They simulate this fact by adding salt to bitumen and subjecting various types of mixtures (AM-16, OGFC-19 and AC-16) to freeze–thaw cycles; their results show a decrease in the Water Sensitivity Test (ITSR), with results being lower in the case of the AC mixture.

Juli-Gándara et al. [16] investigated how NaCl influences the mechanical properties of three types of asphalt mixtures: a hot mix asphalt with conventional bitumen; and two porous mixtures, one manufactured with a conventional binder and another with a modified binder. The e ffect of the salt is analysed by three di fferent processes: immersing the specimens in salt water; adding salt as aggregate into the mixture; and submerging the aggregate in water with a certain concentration of salt, drying it and then making the mixture with it. The results show that the hot mix asphalt is scarcely a ffected when it is submerged in salt water.

The e ffects of freeze–thaw cycles in the bituminous mixture are widely studied. Goh and You [17] used an image-processing technique to indicate that the removal of fine and coarse aggregates on the surface increases when the asphalt mixture undergoes more freeze–thaw cycles. Tarefder et al. [18] tested an AC mixture to ITS and Fatigue Test after 5, 10, 15 and 20 freeze–thaw cycles, and the results showed a reduction of 30.7% in the fatigue life and a small amount of reduction of ITSR after five cycles. Özgan and Serin [19] indicated through the void ratio (Vh), the void ratio filled with asphalt (Vf) and the void ratio inside mineral aggregate (VMA) parameters; the ultrasonic velocity test; and the Marshall Stability (MS) that the e ffect of the freeze–thaw cycles on the asphalt concrete is highly important, especially for the hot mix asphalt design. Islam and Tarefder [20], who investigated the sti ffness and tensile strength degradation behaviour of asphalt concrete on long-term freeze–thaw samples in the laboratory, indicated that the flexural sti ffness decreases with the number of freeze–thaw cycles, whereas the ITS does not change significantly with the number of freeze–thaw cycles. Teltayev et al. [21] investigated, in laboratory conditions, the e ffect of cyclic freezing and thawing on the characteristics of the neat bitumen and bitumens modified with di fferent polymers, as well as stone mastic asphalt concretes. Their results of the SMA-20 Bit 100/130 show a 78% decrease in the value of ITS and a 270% increase in the Rut Depth (RD) after 50 freeze–thaw cycles.

Until now, no research has covered the results of mechanical behaviour such as mechanical strength, permanent deformations, dynamic modulus and fatigue life, when an asphalt concrete mixture is subjected to temperature changes, freeze–thaw cycles and the impact of salt. The purpose of this research is to fill this gap. The temperature interactions have been carried out for dry specimens, specimens submerged in distilled water and specimens submerged in salt water (NaCl).
