2.2.5. GPC Test

An Agilent PL-GPC220 gel permeation chromatography was used to dissolve the asphalt specimens in a tetrahydrofuran (THF) solvent to form a solution at a specific concentration, with a specimen volume of 100 <sup>μ</sup>L and a flow rate of 1 mL·mL<sup>−</sup>1. The column used to separate the asphalt specimens was maintained at 35 ◦C. The molecular weight and molecular weight size distribution of the asphalt were determined experimentally. Figure 4 shows the principle of GPC. In the GPC test, the separation mode was not based on the molecular weight, but molecules' apparent size and molecular aggregation in a specific solution. After the asphalt sample was adequately dissolved, it was introduced into a group of columns through the injection mechanism as a molecular filtration system. The chromatographic column was filled with a cross-linked gel containing surface pores, which were different in size and played the role of molecular filters. Therefore, larger-size molecules (LMSs) were not able to enter smaller pores, and smaller molecules (MMSs) fit most pores and remained for a longer time.

**Figure 4.** Gel permeation chromatography (GPC) schematic diagram.

GPC research results can show the differences of asphalt in molecular size distribution. In other words, GPC provided the "fingerprint" of each asphalt and provided a reasonable explanation for the changes of its macro-properties in combination with the knowledge

of the relative molecular weight of asphalt components. Equations (2) and (3) show the formulae for the calculation of *M*<sup>n</sup> (number average molecular weight) and *M*<sup>w</sup> (heavy average molecular weight), respectively [18]:

$$M\_{\rm n} = \frac{\sum H\_{\rm i}}{\sum (H\_{\rm i}/M\_{\rm i})} \,\tag{2}$$

$$M\_{\mathbf{w}} = \frac{\sum M\_{\mathbf{i}} H\_{\mathbf{i}}}{\sum H\_{\mathbf{i}}},\tag{3}$$

where *H*<sup>i</sup> is the peak height, *M*<sup>i</sup> is the molecular weight, *M*<sup>n</sup> is the number average molecular weight, and *M*w is the heavy average molecular weight.

#### 2.2.6. FTIR Spectroscopy Test

FTIR can be used to detect chemical functional groups in solids, gases, and liquids. Many researchers have now used FTIR spectroscopy to characterize the aging and regeneration behavior of asphalt and the content of polymers [19]. In this study, a Breaker Alpha FTIR spectrometer with a spectral acquisition range of 500–4000 cm−<sup>1</sup> was used to determine the functional groups of each asphalt specimen. OMNIC software was used for the baseline correction and smoothing of the IR curves to evaluate the regenerative effect of the regenerator on the aging asphalt with the use of the characteristic functional group (C=O and S=O) indices and the equation method [20] to calculate the sulfoxide group index (IS=O) and the carbonyl group index (IC=O) with the equations as in Equations (4) and (5):

$$I\_{S=O} = \frac{A\_{1032}}{A\_{2923} + A\_{2852}},\tag{4}$$

$$I\_{\mathcal{C}=O} = \frac{A\_{1700}}{A\_{2923} + A\_{2852}}.\tag{5}$$

#### 2.2.7. AFM Test

A Bruker Dimension Icon AFM was used to observe the surface morphologies of the asphalt specimens to obtain the asphalt morphology map. The test was performed in the tap mode with a scanning area of 20 μm × 20 μm. Figure 5 shows the schematic diagram of the AFM working principle. Figure 6 shows the asphalt AFM specimens, and the asphalt specimens were hot-cast. Firstly, the asphalt was heated to the flowing state, and then, a small amount of asphalt was dropped on the center of a slide and heated to make it flow freely on the slide. Finally, it was tested after natural cooling.

**Figure 5.** Schematic diagram of the atomic force microscopy (AFM) working principle.

**Figure 6.** AFM specimens.

*Rq* and *Ra* could characterize the roughness in the asphalt microscopic appearance and were described as Equations (6) and (7), respectively [21]:

$$R\_{\emptyset} = \sqrt{\frac{\sum\_{i}^{N} Z\_{i}^{2}}{N}},\tag{6}$$

$$R\_d = \frac{1}{N} \sum\_{j=1}^{N} |Z\_j|\_{\prime} \tag{7}$$

where *Rq* is the average root mean square of the planar adhesion deviation, *Ra* is the arithmetic mean of the absolute values of the adhesion deviation measured in the average plane, and *Zi* denotes the corresponding adhesion deviation. In order to further quantitatively investigate the microscopic property change of the regenerated asphalt, in this paper, the AFM morphological images were pre-processed and quantitatively calculated. Figure 7 shows the image processing and calculation process.

**Figure 7.** *Cont*.

**Figure 7.** Quantitative analysis process of the asphalt micromorphological map processing: (**a**) original morphology; (**b**) pre-processing; (**c**) identification; (**d**) calculation.

#### **3. Analysis of Test Results**

#### *3.1. Analysis of Physical Properties Test Results*

As shown in Figure 8, the performance of the regenerated asphalt improved with the increase of the regenerator dose during the aging asphalt regeneration process. The needle penetration and the ductility increased significantly. In contrast, the softening point showed a decreasing trend. When the regenerator dose was 6%, the indicators of RA6 and MA tended to be close to each other, which indicated that the three primary indicators of the regenerated asphalt met the technical requirements of 70# base asphalt.

**Figure 8.** Physical properties of the regenerated asphalt at different doses.

#### *3.2. Viscosity and Activation Energy Analysis*

The asphalt viscosity was closely related to the construction and ease. As seen in Figure 9, the asphalt viscosity tended to decrease, as the dosage of the regenerator increased. The viscosities difference between RA6 and MA was slight, indicating that the viscosity of the regenerated asphalt with a dose of 6% was close to the base asphalt. It indicated that the regenerator had the effect of reducing the viscosity of aging asphalt. The viscosity activation energy reflected the energy required for the regenerated asphalt to reach the flow state, in order to characterize the construction and ease of the regenerated asphalt. The smaller the viscosity activation energy, the better the construction and ease. Figure 10 shows that the viscosity activation energy of asphalt increased after aging. In addition, the activation energy of RA2, RA4, and RA6 decreased by 1.1%, 3.2%, and 4.6%, respectively, compared with that of RA0, indicating that the regeneration agent can reduce the viscosity of the aged asphalt and improve the construction. In addition, the activation energy of the viscosity of aged asphalt was similar to that of the base asphalt, when the dose of the rejuvenator was 6%.

**Figure 9.** Viscosities of different asphalts.

**Figure 10.** Activation energies of different asphalt viscosities.

#### *3.3. Analysis of GPC Test Results*

3.3.1. Molecular Weight and Polydispersity

GPC provided valuable data about molecular weight distribution. Table 4 gives the GPC parameters of each asphalt specimen, including *M*<sup>n</sup> (number average molecular weight), *M*w (heavy average molecular weight), and polydispersity coefficient (PD = *M*w/*M*n). Generally, the larger the polydispersity index is, the wider the molecular weight distribution is. In addition, polydispersity represents the degree of component migration [22]. As shown in Table 4, the *M*n value decreased with the increasing dose of the regenerator, which means

that the asphalt formed smaller molecules through physical or chemical reactions during the regeneration process. By observing changes in the *M*<sup>w</sup> and PD values of various asphalts, we can see that the *M*<sup>w</sup> and PD of RA6 were lower than those of other asphalts, indicating that the regeneration agent led to lower *M*w and PD values and had a degrading and diluting effect on the aged asphalt macromolecules.


**Table 4.** GPC parameters of different asphalt specimens.
