The asphalt used in the test is a shell styrene–butadiene–styrene (SBS)-modified asphalt processed by Suzhou Sanchuang Pavement Engineering Co., Ltd., and its performance indexes are shown in Table 1 (where it can be seen that the selected indexes fall within the specification values). The selected warm mixing agent is Sasobit, produced by the Sasol group in South Africa. It consists of fine white solid particles at room temperature, with a structural mixture of long-chain saturated hydrocarbons. It belongs to the class of organic warm mixing agents that reduce viscosity. Sasobit basic characteristics are shown in Table 2.

SBS-modified asphalt was subjected to laboratory short-term aging (RTFO) and long-term aging (PAV) to obtain the aged asphalt required for the test. The aged asphalt was mixed with the new SBS-modified asphalt in five proportions of 0, 20, 50, 80, and 100%. The mixer was used for 30min, and then 1.5% Sasobit warm mix agent was added. Finally, a high-speed shear was used to fully mix the system at 160 °C, with a rotating speed of 4500 rad for 40 min (Fan, 2020). Further details are given in Table 3.

The penetration, softening point, and 5°C ductility of asphalt samples were tested according to the “Test Code for Asphalt and Asphalt Mixture in Highway Engineering” (JTG E20-2011).

VTNR20-010V-I NMR crosslinking density imaging analyzer (Suzhou NewMai Analytical Instrument Co., Ltd.) was used for ¹H spectrum analysis. The operating temperature was 25°C, and the system was kept warm for 20 min before the test. Carr–Purcell–Meiboom–Gill (CPMG) pulse sequence was selected for the test. The echo time was set to 0.05 ms, the number of echoes was 5,000, and the scanning superposition times were 32.

An atomic force microscope produced by Bruker was used. PeakForce QNM mode and 0.4 N/m probe were selected. The scanning dimensional range was 10 μm × 10 μm, and the scanning rate was 1 Hz. The hot casting method was used to prepare the sample. First, the asphalt sample was heated to the flowing state, and then a few samples were taken out and dropped in the center of the glass slide, which was obliquely placed in an oven at 135°C for 10min-heating to make the surface flow naturally into a smooth and flat film. Then, surface morphology and detection force curve were recorded by AFM at room temperature (25°C).

NICOLET IS5 instrument was used for FTIR tests. KBr smear method was used to prepare the sample. First, the asphalt was mixed with CS₂ in a ratio of 1:19. Then, it was gently dropped in the middle of the KBr window and then put into the UV drying oven to dry CS₂. Finally, the FTIR test was carried out. The scanning range of the infrared spectrum was 400-4000 cm⁻¹, and the scanning times were set to 32.

Figure 1 shows the results in terms of indexes such as penetration, softening point, and ductility. It can be seen from Figure 1A that when the content of new asphalt gradually changes from 0 to 100%, the penetration and ductility of new and old mixed asphalt are improved, while the softening point is decreased. Among them, the change in ductility is the most marked, with an increase of 144%. In contrast, the softening point was slightly affected and decreased by only 13%. This is mainly because there is a high content of asphaltene in the aged asphalt, which makes the asphalt hard and the viscosity higher. After adding new asphalt, the content of asphaltene and gum can be effectively adjusted to restore the performance of the old asphalt. In terms of macro performance, the material becomes soft, and the ductility, correspondingly, increases. From a comparison between Figures 1A and B, it can be seen that Sasobit warm mix agent has an adverse impact on the low-temperature performance of new and old mixed asphalt, in which the maximum ductility loss can reach 24.8%. The main reason is that Sasobit warm mix agent is an organic viscosity-reducing material. When the temperature is higher than its melting point, it will adsorb the saturated components in the asphalt. When the temperature is lower, it will form a network lattice structure in the asphalt. Therefore, it will harden the asphalt macroscopically and change the elasticity behaviour, which is reflected in the decrease of ductility and penetration and in the increase of softening point (Zhang and Zhang, 2016). When there is a higher fresh asphalt content (more than 50%) in the mixed asphalt, the effect of Sasobit is more significant, mainly because there is a higher saturated component content, which improves the effect of Sasobit warm mix. Altogether, the penetration ranged from 35.3 to 52.6 mm in the absence of Sasobit, and from 33.4 to 42.5 mm in its presence. The softening point was lowered in the range 89.9–78.6 °C and 91.6–82.9 °C in the absence and presence of Sasobit, respectively, while the ductility increased from 12.3 to 30.1 cm (no Sasobit) or from 12.7 to 27.6 cm (presence of Sasobit).

Asphalt is mainly composed of C- and H-compounds and their derivatives, and the LF-NMR technique can characterize the difference between ¹H and other molecular bond energies according to the relaxation time T ₂ . When T ₂ is short, this indicates that ¹H can be closely connected with other molecules, the bond energy is greater, and the degree of freedom is smaller. The relative content of hydrogen protons can be expressed by the signal amplitude covered by T ₂ peak at different relaxation times (i.e., the integral area of T ₂ interval).

When the number of LF-NMR inversion fitting iterations is set to 1,000, there is only one peak on each T ₂ inversion spectrum curve (peaks with a relative area less than 2% are not considered). It can be seen from Figure 2A that in the absence of warm mix agent, the T ₂ diffraction peak of each group of mixed asphalt has basically the same starting time, but the ending time is delayed with the increase of the proportion of new asphalt. The reason is that during the aging process of asphalt, some bond energy is released due to the breaking of light alkyl components, low molecular weight compounds volatilization, and C=C cracking/reorganization. At the same time, it is also found that the larger the proportion of new asphalt, the lower the peak value, i.e., the smaller the corresponding diffraction peak area. This is because when the content of C and H remains unchanged, it will react with oxygen during asphalt aging, resulting in an increase in the relative content of oxygen and a decrease in the relative content of hydrogen.

By observing Figures 2B–F, it can be found that after adding Sasobit, the ending time of T ₂ diffraction peak moves backward at varying degrees. The most significant backward movement is observed in the mixed asphalt with the highest proportion of new asphalt. Among them, 1-s backward shift is within 0.02 ms, and 5-s backward shift reaches 0.08 ms. This shows that when Sasobit warm mix adsorbs the saturated components in asphalt, the bond energy between ¹H and other molecules is reduced so as to achieve the effect of viscosity reduction. The higher the content of saturated components, the better the viscosity reduction effect. The reason why Sasobit warm mix can increase both peak value and area of T ₂ is mainly because it is composed of C- and H-based polymers, which can increase the content of hydrogen in new and old mixed asphalt.

In the asphalt morphology obtained by AFM, the “bee-like structure” is generally considered the most prominent feature. This concept was first proposed by Loeber et al. (2003). After that, Li (Li et al., 2015) conducted AFM tests on a variety of asphalts and found that the “bee-like structure” was formed by precipitation during cooling because asphaltene could not be well integrated with other asphalt components. Figure 3 represents the height variation diagram of different phase structure areas for non-aged asphalt. It can be found that its micro surface structure has different shapes, and the peak height relationship of each phase is as follows: (white) bee structure > continuous phase structure > (black) bee structure.

In the infrared spectrum of asphalt, methylene (-CH₂-), seen at 1460 cm⁻¹, and methyl (-CH₃), seen at 1375cm⁻¹, are not easy to undergo chemical or physical changes in the aging process of asphalt due to their stable properties. The aging degree of asphalt mainly depends on the structural index of carbonyl (C=O), seen at 1700cm⁻¹, and sulfoxide (S=O), seen at 1030cm⁻¹. For the calculation formulas of carbonyl (I _C=O), sulfoxide (I _S=O), and SBS (I _SBS) structure indexes, see Eqs 1–3 (Jiang, 2018): I C = O = A b s o r p t i o n   p e a k   a r e a   n e a r   1700 c m − 1 A b s o r p t i o n   p e a k   a r e a   n e a r   1460 c m − 1 + A b s o r p t i o n   p e a k   a r e a   n e a r   1375 c m − 1 . (1) I S = O = A b s o r p t i o n   p e a k   a r e a   n e a r   1030 c m − 1 A b s o r p t i o n   p e a k   a r e a   n e a r   1460 c m − 1 + A b s o r p t i o n   p e a k   a r e a   n e a r   1375 c m − 1 . (2) I S B S = A b s o r p t i o n   p e a k   a r e a   n e a r   966 c m − 1 + A b s o r p t i o n   p e a k   a r e a   n e a r   699 c m − 1 A b s o r p t i o n   p e a k   a r e a   n e a r   1460 c m − 1 + A b s o r p t i o n   p e a k   a r e a   n e a r   1375 c m − 1 . (3)

These indexes were calculated through the aid of OMNIC software. The specific location and range of characteristic peaks for each functional group are shown in Table 5.

Figure 5 shows the infrared spectra for the samples under investigation. In Figure 5A, the dotted lines show the absorption peaks of carbonyl (C=O, 1700 cm⁻¹) and sulfoxide (S=O, 1030 cm⁻¹). For sample 5, the characteristic peaks of carbonyl and sulfoxide groups are not clear, but the characteristic peaks of these two functional groups become more prominent with the increase of the aged asphalt proportion (i.e., going back from sample 5 to sample 1, cf. Table 3). By observing Figures 5B–F, it can be found that, after the warm mix agent is added, there is no new characteristic peak or shift in the position of the characteristic peaks, indicating that there is no chemical reaction between Sasobit warm mix agent and new/old mixed asphalt. Sasobit warm mix agent can weaken the characteristic peaks of carbonyl and sulfoxide groups of new and old mixed asphalt, indicating that the warm mix agent can improve the aging state of asphalt.

Table 6 lists the values for the structure indexes defined in Eqs 1–3. It can be found that, at an increasing proportion of new asphalt (from sample one to sample 5), I _C=O decreased from 0.0331 to 0.0035, with a decrease of 89.4%, and I _S=O decreased from 0.0341 to 0.0106, with a decrease of 68.9%, indicating that the proportion of carbonyl and sulfoxide groups also gradually decreases in the process of asphalt regeneration. After adding Sasobit warm mix agent, I _C=O and I _S=O are reduced, indicating that warm mix agent can improve the aging state of the materials under investigation. However, I _SBS were all between 0.04–0.06, and there is no clear trend in I _SBS when going from samples 1 to 5, neither in the absence nor in the presence of Sasobit. Actually, tests carried out in this work simulate asphalt aging under laboratory hot oxygen conditions and lack the influence of ultraviolet on SBS under natural light conditions (Marsac et al., 2014).