Influence of Recycling Agents Addition Methods on Asphalt Mixtures Properties Containing Reclaimed Asphalt Pavement (RAP)

Abstract

Recycling agents (RAs) are used to restore the physicochemical properties of the aged asphalt binder existing in the reclaimed asphalt pavement (RAP) material. The best location for adding the RAs in the manufacturing process of the asphalt mixtures containing RAP has always been of concern to asphalt mixture researchers. In this study, vegetable, paraffinic and aromatic extract oils were used as RAs. The effect of RA location on the performance properties of the asphalt mixtures containing different percentages of the RAP material was investigated by adding the RAs in three different locations, including in the virgin binder, in the hot RAP material, and in the final mixture. For this aim, the rutting, cracking, and moisture sensitivity of the mixtures containing different RAs at different locations was investigated. The results showed that the best location for adding the RAs is different for various RAs. The best place for adding the paraffinic, aromatic extract, and vegetable oils in terms are in the virgin binder, in the RAP, and the final mixture, respectively. Therefore, using RAs in their appropriate location can improve the mechanical properties of asphalt mixtures containing RAP.

1. Introduction

Despite environmental and economic advantages, the use of high contents of reclaimed asphalt pavement (RAP) in newly produced asphalt mixtures has faced some problems due to the difficulties in determining the degree of blending between the recycling agents (RAs) and aged and virgin asphalt binder [1] Moreover, using high contents of the RAP materials leads to a reduction in the resistance of the mixtures to low-temperature cracking due to the brittleness of the aged asphalt binder in the RAP. The negative effect of the aged asphalt binder can be made up using RA or a softer virgin asphalt binder [2]. On the other hand, using a softer asphalt binder is not recommended in asphalt mixtures containing RAP material higher than 25%. Therefore, using RA is mandatory in asphalt mixtures containing high RAP contents. Finding the optimum RA dosage, homogeneous distribution of the RA in the mixtures, proper diffusion of the RA in the aged asphalt binder, and enough blending degree between the RAP binder and the RA are reported as the challenges of using RA in the RAP-blended mixtures [3]. In addition to the type and chemical properties of the RAs, it is reported that the blending method and the place of adding RA in the asphalt mixture affect the blending degree, the RA diffusion, and the uniformity of the RA in the mixture, which greatly affect the final performance of the mixtures [4].

There have been some research studies on the effect of the location of adding an RA into asphalt mixtures on the performance of the asphalt mixtures. Some studies show that the location of adding the RA does not have a significant effect on the performance of the RA and the asphalt mixtures containing RAP [5,6,7]. A study showed that the location of adding the soybean as RA has no significant effect on the volumetric properties of asphalt mixtures [5]. In 2021, Juan et al. [6] investigated the effect of the RA location on the cracking performance of asphalt mixtures containing 30% and 50% of the RAP. For this aim, they chose a bio-oil as RA and mixed it once with the virgin asphalt binder and once with the RAP material at two RA locations. No significant effect of the RA location on the crack resistance of the mixture was observed [6]. Another study in 2020 showed that the location of adding tall oil as RA does not have any significant effect on the indirect tensile strength (ITS) and stiffness modulus of the asphalt mixtures with 60% RAP [8].

On the other hand, some research studies indicate the significant effect of the location of RA addition on the performance of the asphalt mixtures containing RAP. Zaumanis et al. [2,3], in limited research on tall oil as RA, showed that spraying the RA on the cold RAP material in the RAP conveyor belt is the most effective method of using RA in 100% RAP asphalt mixtures [2,3]. However, it was shown in another study that adding a modified vegetable oil directly to the RAP may lead to incomplete blending between the RAP asphalt binder and the virgin asphalt binder [9]. Yu et al. investigated the effect of the location of adding a specific RA on the performance of asphalt mixtures containing 30% RAP. Spraying to the cold RAP and blending with virgin asphalt binder were selected as the methods of adding RA to the mixtures. They showed that spraying the RA directly to the RAP material results in higher rutting and moisture resistance [1]. In 2019, Haghshenas et al. [10] added aromatic extract as RA once to asphalt binder and once to a mixed with RAP in mixtures containing 65% RAP. The results showed that adding the RA directly to the RAP material extremely softened the mixtures, which led to a high rutting and bleeding potential in the mixtures. Mixing RA with the virgin asphalt binder before adding it to the mixture resulted in acceptable rutting and cracking performance [10]. In a study, Xuan Lu et al. investigated the effect of RAs on the properties of warm mix asphalt containing RAP. In this work, a RA named Syvaroad was applied. The RA was added once to the virgin binder and once directly to the RAP. Unlike the specimen containing 50 percent of RAP, the specimens containing 25 and 75% RAP were improved in terms of rutting resistance upon adding the RA to the virgin binder. The researchers also stressed the need for more extensive research to investigate this finding [11]. Kaseer et al. investigated the effect of the way of adding RAs on the properties of RAP binder. In this research two methods were used to add RAs to the asphalt mixtures. In the first method, the vegetable oil-based RA was directly added to the virgin binder before mixing the asphalt with RAP and virgin aggregates. In the second method, the RA was added to the RAP five minutes before the RAP was mixed with aggregates and virgin binder. The results showed that at the mixture temperature of 140 °C, adding the RA to the virgin binder had a better effect on the blending of aged and virgin binder, compared to adding the RA to the RAP. In the discussed research, no performance test on asphalt mixtures was investigated, and only the blending of virgin binder and RAP binder was evaluated [9].

The review of the previous research shows the existing conflicting information on the effect of the RA location on the performance of the asphalt mixtures containing RAP, and thus, more investigations are needed in this regard. Due to the variety of the RAs with different physicochemical properties, investigating the effect of the RA location on the volumetric and functional properties of the mixtures containing low and high contents of the RAP can help fill research gaps in this area.

Study Objectives

The main purpose of this study is to investigate the location of adding different RAs on the performance properties of the asphalt mixtures containing different percentages of the RAP. For this purpose, three different RAs, including paraffinic oil, aromatic extract, and vegetable oil, were added in different locations of the asphalt mixtures containing 25% and 50% of RAP. Mixing with virgin asphalt binder, adding to the RAP material, and adding to the final mixture were selected as three different methods of adding RA to the mixtures. The effect of the RA location was studied by investigating the volumetric and performance properties of the mixtures using rutting (Hamburg wheel tracking (HWT)), low and medium-temperature cracking (Semi-Circular Bending (SCB)), and moisture susceptibility (ITS) tests.

2. Materials and Methods

2.1. Materials

2.1.1. Asphalt Binder

An asphalt binder with a penetration grade of 60/70 (performance grade (PG) 64-22) was selected as the virgin asphalt binder. The physical properties of the RAP asphalt binder were also studied by conducting the specific gravity, penetration, softening point, loss of heating, ductility, flash point, and rotational viscosity tests on the recovered RAP asphalt binder, whose results are shown in

Table 1. The properties of virgin and RAP asphalt binder.

Test	Standard	Results
		Virgin Binder	RAP Binder
Specific gravity	ASTM D70 [12]	1.03	1.04
Penetration (0.1 mm)	ASTM D5 [13]	66	29
Softening point (°C)	ASTM D36 [14]	50	61
Loss of heating (%)	ASTM D1754 [15]	0.75	0.75
Ductility (cm)	ASTM D113 [16]	>100	34
Flashpoint (°C)	ASTM D92 [17]	305	280
Rotational viscosity @135 °C (Pa·s)	AASHTO T316 [18]	0.360	2.783

.

2.1.2. Aggregate

Limestone aggregates with physical properties shown in

Table 2. The characteristics of virgin and RAP aggregates.

Physical Properties	Standard	Values (%)		Specification Limit (%)
		Virgin Aggregates	RAP Aggregates
Coarse aggregate specific gravity	ASTM C127 [19]	2.659	2.593	-
Fine aggregate specific gravity	ASTM C128 [20]	2.639	2.464	-
Los Angeles abrasion (%)	ASTM C131 [21]	21.5	23.5	<30
Water absorption (%)	ASTM C128 [20]	0.7	0.8	<2.5
Fractured particles in one face	ASTM D5821 [22]	98	97	>50
Fractured particles in two faces and more	ASTM D5821 [22]	95	94	>80
Sand equivalent	AASHTO T176 [23]	72	69	>50

were used as virgin aggregates. The RAP material was obtained from the top layer of a highway pavement in Tehran, which was constructed 10 years ago. The RAP aggregates were also extracted, and their physical properties were measured, and the results are shown in Table 2. As can be seen, the specific gravity of the RAP aggregates was slightly lower than that of the virgin aggregates. Other properties of the RAP and virgin aggregates were almost similar to each other. The gradations of the RAP aggregates and the final mixture are also depicted in Figure 1.

2.1.3. Recycling Agents (RAs)

Aromatic extract, paraffinic oil, and vegetable oil were used as RAs to restore the physicochemical characteristics of the aged asphalt binder existing in the RAP. Some physical properties of RAs, such as kinematic viscosity, flash point, color, and specific gravity, were measured, and the results are shown in

Table 3. The physical properties of the three different RAs used in this study.

Property	Standard	RA
		Vegetable Oil (V)	Paraffinic Oil (P)	Aromatic Extract Oil (A)
Kinematic viscosity at 100 °C (cSt)	ASTM D445 [24]	15.5	13	60
Flashpoint (°C)	ASTM D92 [17]	290	250	300
Color	-	Clear yellow	Brown	Dark green
Specific gravity at 15 °C	ASTM D1298 [25]	0.925	0.900	0.995

.

For determining the optimum content of the RA, the RAP asphalt binder should be extracted from the RAP materials, which is carried out according to AASHTO T164 [26]. The RAP asphalt binder should then be separated from the solvent (Trichloroethylene) using the rotary evaporation and Abson methods based on ASTM D5404 [27] and ASTM D1856 [28], respectively. The optimum dosage of RAs was selected by measuring the penetration (ASTM D5) and softening point (ASTM D36) of the aged binder containing virgin asphalt binder and different percentages of the RAs. For this aim, the virgin and recovered RAP asphalt binders were mixed at 160 °C for 10 min. Then, different percentages of the RAs were added to the blend. The RA dosage that changes the penetration and softening point of the blend of aged and virgin binder to be similar to the virgin asphalt binder was selected as the optimum dosage.

Table 4. The optimum dosage of the RAs used in this study.

RA ID	RA Type	75% RAP Binder + 25% Virgin Binder	50% RAP Binder + 50% Virgin Binder
V	Vegetable oil	4.5	5.2
P	Paraffinic oil	5.5	7.2
A	Aromatic extract oil	9.7	12.9

shows the optimum dosage of each RA for different blends.

2.2. Methodology

2.2.1. Experimental Plan

In this study, the asphalt mixture performance tests were conducted on 21 different asphalt mixtures, including one control mixture without RAP and twenty mixtures containing two different RAP contents (25% and 50%) and three different RAs (aromatic extract, paraffinic, and vegetable oils) added in three different locations (RAP, Virgin binder and final mixture). Afterward, the optimum RA location was determined using the HWT, SCB, and ITS tests. The details of the experimental plan of this study are shown in Figure 2.

2.2.2. Mix Design

The Superpave mix design (SMD) was carried out based on AASHTO M323 [29] and AASHTO R35 [30] to determine the optimum asphalt binder content (OBC) of the mixtures. The asphalt content in RAP materials was determined 5.6% based on the total weight of RAP according to AASHTO T308 [31]. The SMD procedure was conducted on all 21 mixtures. The details of the mixtures and the results of the mix-design procedure are shown in

Table 5. Results of the Superpave mix design (SMD).

Mixture ID	Mixture Composition	OBC	Gmm	Gmb	VMA	VFA	D/B
C	100% Virgin	4.91	2.484	2.385	14.37	72.16	0.89
R25	75% Virgin + 25% RAP	5.42	2.438	2.340	15.35	73.94	0.80
R25V-R	75% Virgin + 25% RAP + V (R) 1	5.29	2.437	2.340	15.27	73.80	0.81
R25V-B	75% Virgin + 25% RAP + V (B)	5.21	2.441	2.343	15.06	73.43	0.82
R25V-M	75% Virgin + 25% RAP + V (M)	5.26	2.440	2.342	15.14	73.57	0.82
R25P-R	75% Virgin + 25% RAP + P (R)	5.33	2.437	2.340	15.30	73.86	0.80
R25P-B	75% Virgin + 25% RAP + P (B)	5.25	2.440	2.342	15.13	73.56	0.82
R25P-M	75% Virgin + 25% RAP + P (M)	5.28	2.440	2.342	15.15	73.60	0.82
R25A-R	75% Virgin + 25% RAP + A (R)	5.35	2.438	2.340	15.29	73.83	0.81
R25A-B	75% Virgin + 25% RAP + A (B)	5.28	2.441	2.343	15.12	73.54	0.82
R25A-M	75% Virgin + 25% RAP + A (M)	5.32	2.441	2.343	15.16	73.61	0.82
R50	50% Virgin + 50% RAP	5.86	2.413	2.316	15.57	74.32	0.78
R50V-R	50% Virgin + 50% RAP + V (R)	5.42	2.416	2.319	15.07	73.46	0.81
R50V-B	50% Virgin + 50% RAP + V (B)	5.31	2.423	2.326	14.73	72.84	0.84
R50V-M	50% Virgin + 50% RAP + V (M)	5.35	2.419	2.322	14.91	73.16	0.83
R50P-R	50% Virgin + 50% RAP + P (R)	5.51	2.413	2.316	15.26	73.79	0.80
R50P-B	50% Virgin + 50% RAP + P (B)	5.38	2.418	2.321	14.97	73.28	0.82
R50P-M	50% Virgin + 50% RAP + P (M)	5.43	2.416	2.319	15.08	73.48	0.81
R50A-R	50% Virgin + 50% RAP + A (R)	5.60	2.412	2.316	15.38	73.98	0.79
R50A-B	50% Virgin + 50% RAP + A (B)	5.51	2.419	2.322	15.05	73.42	0.82
R50A-M	50% Virgin + 50% RAP + A (M)	5.54	2.416	2.319	15.18	73.65	0.81

¹ The character in brackets is about how the RAs are used (R: RAP, B: Binder, and M: Mixture). OBC: Optimum asphalt. Binder Content. Gmm: Theoretical maximum specific gravity. Gmb: Bulk specific gravity. VMA: Voids in the Mineral Aggregate. VFA: Void Filled with Asphalt Binder. D/B: Dust/Binder.

.

It can be seen that the OBC is increased by 10% and 19% when 25% and 50% of the RAP materials are used without RA, respectively. The increase in the OBC is mainly due to the lack of participation of the RAP asphalt binder in the new asphalt mixture. On the other hand, it is observed that the OBC declines significantly when RAs are used to produce the asphalt mixtures containing RAP, which shows the improvement in the participation of the RAP asphalt binder in the newly produced mixture when RAs are used. The results also indicated that the RA location did not affect the volumetric properties and the OBC of the mixtures.

2.2.3. Sample Preparation

All conditions of sample preparations were kept constant except the method of adding the RAs to the mixtures. For preparing the mixtures, the virgin aggregates were kept in a 160 ℃ oven for 12 h. The RAP and virgin asphalt binder were also placed in a 160 ℃ oven for 2 h before mixing with RAP. According to Zaumanis et al. [8], the total mixing time should be 4 min for all mixtures. The RAs were added to the mixtures in three different locations as follows:

Adding RAs directly to the RAP: In this method, the RAs were directly added to the heated RAP material and mixed for 1 min. The virgin aggregates were then added to the blend of RAP and RA and mixed for 1 min. Finally, the virgin asphalt binder was added and mixed for 2 more minutes.

Adding RAs to the virgin asphalt binder: In this method, the virgin asphalt binder was mixed with the RAs for 10 min using a low-shear mixer. Simultaneously, the heated virgin aggregates and RAP were mixed for 1 min. The blend of virgin asphalt binder and RA were then added to the RAP and aggregates and mixed for 3 min.

Adding RAs to the final mixture: In this procedure, the heated virgin aggregates and RAP materials were mixed for 1 min. Then, the virgin asphalt binder was added to the mixture and mixed for 2 min. The RAs were added to the final mixtures and mixed for another 1 min.

The specimens of the HWT and SCB tests were placed in a 135 ℃ oven for 4 h before compaction to simulate the short-term aging based on AASHTO R30 [32]. The short-term aged SCB specimens were then placed in a 95 ℃ oven for 72 h to apply long-term aging to the mixtures [33]. The Superpave gyratory compactor was then employed to produce cylindrical specimens for conducting the HWT, ITS, and SCB tests. It should be noted that the short-term aging conditions were used for HWT tests, and the long-term aging conditions were utilized for the SCB tests. The curing of the ITS specimens was performed based on AASHTO T283 [34].

2.2.4. Testing Program

Hamburg Wheel Tracking (HWT) Test

The rutting performance of the mixtures was determined by conducting the HWT test. To do so, the specimens were short-term aged and compacted to reach 7 ± 0.5% air void. The tests with two replicates of each mixture were conducted at 50 ℃, and in total, 42 samples were prepared and tested according to AASHTO T324 [35]. In this study, the HWT test was conducted under dry conditions. The test temperature (50 ℃) is kept constant during the test by an air-heating chamber. The Rut Depth (RD) after the 20,000 wheel path was reported as the rutting resistance. The test was repeated for two replicates of each mixture, and the mean values of the tests were reported.

Semi-Circular Bending (SCB) Test

The low and medium-temperature cracking of the mixtures were studied using the SCB test at −12 ℃ and +25 ℃ according to AASHTO TP 124 [36] and FHWA-ICT-15-017 [37], respectively. The cylindrical specimens were compacted to reach 7 ± 0.5% air void and cut into four SCB specimens. The geometry of the SCB specimens and the testing setup are shown in

Table 6. The details of the SCB specimens and SCB tests.

RA ID	RA Type
Radius (R)	75 mm
Thickness (T)	50 mm
Noth length (a)	15 mm
Noth width (t)	1.5 mm
Geometry factors (Yi)	3.059
Distances of span support	120 mm
Constant load line displacement at +25 ℃	50 mm/min
Constant load line displacement at −12 ℃	0.7 mm/min

. A total of 168 SCB specimens were prepared. The test was conducted by applying a monotonic load with a constant loading rate, as described in Table 6. The Flexibility Index (FI) and fracture toughness (KI) were used to describe the cracking performance of asphalt mixtures. The FI index was developed at the Illinois Center for Transportation (ICT) to describe the cracking performance of asphalt mixtures at medium temperatures and is calculated using equation 1. The fracture toughness was used to describe the cracking resistance of asphalt mixtures at minus temperatures and is calculated using Equation (2).

$$FI=\frac{G_f}{\left|m\right|}\times 0.01$$

(1)

$$K_{IC}=\frac{P_{max}{Y}_i\sqrt{\pi a}}{2RT}\times 0.01$$

(2)

where G_f is the fracture energy (J/m²), m is the slope at the inflection point of the post-peak load versus displacement curve, P_max is the peak load (KN), Y_i is the shape factor, which is obtained from a finite element analysis, and a, R and T are the notch length, radius, and thickness of the SCB specimen, respectively.

Indirect Tensile Strength (ITS) Test

The moisture sensitivity was determined by performing the ITS test on the wet and dry specimens. Six replicates were prepared for each mixture (three replicates for each dry and wet ITS test), and in total, 126 cylindrical specimens were compacted to reach 7 ± 0.5% air void, according to AASHTO T 283 [34]. The Tensile Strength Ratio (TSR), which was calculated by dividing the ITS of wet specimens by that of dry specimens, was used as a criterion for describing the moisture susceptibility of the mixtures. The ITS and TSR parameters are calculated using Equations (3) and (4), respectively.

$$\mathit{ITS}=\frac{2000P}{\mathsf{\pi }\mathrm{td}}$$

(3)

$$\mathit{TSR}=\frac{{\mathit{ITS}}_{\mathit{wet}}}{{\mathit{ITS}}_{\mathit{dry}}}\times 100$$

(4)

where the ITS is the indirect tensile strength (kPa), TSR is the tensile strength ratio, P is the maximum load (N), t is the mean thickness of the test specimen (mm), and d is the specimen diameter (mm).

3. Results and Discussion

3.1. Hamburg Wheel Tracking (HWT) Test Results

The results of the HWT tests are presented in Figure 3. It can be seen that increasing the RAP content results in a reduction in the rut depth of asphalt mixtures, and the rut depth of all RAP-blended mixtures is lower than the control mixture, regardless of the RA type and location. The rut depth of the mixtures decreases by 44% and 60% when the RAP content is increased to 25% and 50%, respectively. This performance is mainly due to the higher stiffness of the aged RAP asphalt binder [38]. The rut depth of all RAP-blended mixtures is lower when the mixture is not modified with RA. In other words, using RA in all mixtures has impaired the rutting performance of the RAP-blended mixtures, which is due to the reduction in the stiffness of the RAP binder as a result of adding RA.

The effect of the RA location on the rutting performance of the mixtures varies depending on the RA type. The aromatic extract oils perform better and soften the aged RAP asphalt binder when added directly to the RAP material. The diffusion of the RAs in the aged asphalt binder is affected by its viscosity. The higher the viscosity of the RA, the lower the diffusion rate [39]. As the aromatic extract oil has a high viscosity, its direct contact with the RAP materials will lead to a better diffusion in the aged asphalt binder, which leads to a softer mixture [10]. On the other hand, when the aromatic extract oil is added to the virgin asphalt binder, some of this oil may not come in contact with the aged bitumen. Therefore, adding the aromatic extract oil to the RAP result in a softer mixture and higher rut depth. For example, the rut depth of the R50A-R mixture increased by 39% compared to the R50A-B. The lowest rut depth is attributed to the mixtures in which the aromatic extract is added to the virgin asphalt binder.

The paraffinic oil has the highest effect on the rutting performance of the mixtures when it is mixed with the virgin binder. The RAs perform differently in terms of aging [40,41]. It was indicated in previous studies that paraffinic oils are more susceptible to oxidation than other RAs [40,41]. It seems that the direct contact of this RA with the hot RAP material leads to higher oxidation, and mixing this RA with the virgin asphalt binder before using it in the mixture, can reduce the effect of oxidation. However, more investigations are needed in this regard. Mixing the paraffinic RAs with the virgin asphalt binder leads to a higher rut depth than the other two locations. For example, the rut depth in the R50P-B mixture is 47% higher than the R50P-R.

When the vegetable oil is added to the virgin asphalt binder, it cannot diffuse completely to the aged binder and cannot soften the RAP binder, which results in better rutting resistance and lower rut depth. Adding this oil directly to the RAP material results in better blending and diffusion, which lead to a decrease in the rutting resistance of the mixtures. The highest rut depth in the vegetable oil occurs when it is added to the final mixture. Using this RA in the final mixture leads to a higher softness of the mixtures compared to the other two locations. When using vegetable oils in the final mixture, a layer of the RA covers the components of the mixture, which probably lubricates the mixture and leads to higher softness. Examining and proving this issue requires more investigations in the future. For instance, the rut depth of the R 50V-M mixture is 38% and 12% higher than the R50V-B and R50V-R mixtures, respectively.

3.2. Semi-Circular Bending (SCB) Test Results

3.2.1. Intermediate-Temperature Cracking

The FI is used in this study to describe the intermediate cracking performance of the asphalt mixtures. The FI results are presented in the bar charts in Figure 4. As expected, a significant reduction in the FI value is observed by increasing the RAP content. The FI index is decreased by 27% and 59% when the RAP content is increased by 25% and 50%, respectively. This inferior performance of the RAP-blended mixtures is due to the brittleness of the aged asphalt binder in the RAP [42], which can be compensated by adding RAs. RAs help the flexibility of the aged RAP asphalt binder and prevent premature cracking [43]. It can be observed that using RAs can increase the FI index to just lower than the control mixture. The FI index is less than 5 for all mixtures containing RAP material. The minimum value of the FI index for the long-term aged asphalt mixtures is 5 [33,44].

As can be seen in the results, the FI is influenced significantly by the RA location. The effect of RA location also is highly dependent on the RA type. The aromatic extract is more effective when added directly to the RAP material. For example, it is seen that the FI index of the R50A-R mixture is 15% higher than the R50A-B mixture. The high viscosity of aromatic extract leads to less diffusion and blending when this RA is added to virgin asphalt binder or the final mixture than adding the RA directly to the RAP. As a result, when the aromatic extract is added directly to the RAP, it reduces the stiffness of the RAP asphalt binder and consequently increases the FI index. A similar result was obtained by Haqshanas et al. [10] in another study. The best place to add the paraffinic oil is in the virgin asphalt binder. Paraffinic oils are very sensitive to aging [40,41]. It seems that the direct addition of this RA to the RAP and its direct contact with the hot RAP and then with virgin aggregates leads to more oxidation in this RA. After long-term aging, mixtures in which this oil was added to the virgin binder had a larger FI index value. It can be observed that the FI index is increased when the paraffinic oil is added to the virgin asphalt binder compared to the direct addition to the RAP. For example, the FI value of the R50P-B mixture increased by 19% compared to the R50P-R. Adding vegetable oil to the hot RAP results in better distribution and complete diffusion of RA to the RAP binder, which will result in a softer binder and, subsequently, a more resistant mixture to cracking. On the other hand, adding vegetable oil to the virgin asphalt binder results in the lowest FI value because of the incomplete diffusion of the RA to the aged RAP binder. Adding vegetable oil to the final mixture leads to an increase in the FI index, which is probably due to the lubrication effect of this RA when added to the final mixture. For example, the FI value of the R50V-M mixture increased by 23% compared to the R50V-B.

3.2.2. Low-Temperature Cracking

Investigating the low-temperature cracking performance of RAP-blended mixtures is crucial, as cracking is one of the weaknesses of RAP-blended mixtures. Fracture toughness is one of the parameters used for describing the low-temperature performance of asphalt mixtures [45]. The fracture toughness results of the mixtures containing different RAs applied at different locations are shown in Figure 5. As expected, using the RAP materials results in an increase in the fracture toughness parameter [46,47]. The reason is the increase in the stiffness of the mixtures owing to the higher stiffness of the RAP asphalt binder, which leads to an increase in the peak load and, subsequently higher fracture toughness. The fracture toughness value of the R25 and R50 mixtures are increased by 34% and 58% compared to the control mixture, respectively. Using RAs results in a reduction in the fracture toughness parameter. However, even after modifying with RA, all the RAP-blended mixtures have higher fracture toughness than the control mixture.

The effect of the RA’s location on the fracture toughness is similar to their effect on the rutting resistance. In the mixtures containing aromatic extract, the highest fracture toughness is related to the mixture in which the RA is added to the virgin asphalt binder. On the other hand, the lowest fracture toughness occurs when the RA is added directly to the RAP material. The reason is that the aromatic extract diffuses in the aged asphalt binder more effectively when they are added directly into the hot RAP material, which results in a softer mixture and lower fracture toughness. For example, the fracture toughness of the R50A-B is 28% higher than the R50A-R mixture. In the mixtures containing paraffinic oil, the highest and lowest fracture toughness is related to the mixtures that the RA added to the RAP and virgin asphalt binder, respectively. The fracture toughness of the R50P-R mixture is 27% higher than that of the R50P-B mixture. In the mixtures containing vegetable oil, the highest fracture toughness belongs to the mixtures in which the vegetable oil had been added to the virgin asphalt binder. Using vegetable oil in the final mixture softened the mixture and decreased the fracture toughness. For example, the fracture toughness of the R50V-B mixture is 19% higher than that of the R50V-M mixture.

3.3. Indirect Tensile Strength (ITS) Test Results

The ITS test results of the dry and wet specimens are presented in Figure 6. The results show that both dry and wet specimens have higher ITS values when the RAP content is increased, which can be due to the higher stiffness of the aged asphalt binder in the RAP materials. On the other hand, it can be seen that adding RAs has led to a lower stiffness in the mixture and declines the ITS parameter.

The location of the RAs affects both dry and wet ITS values. It can be seen that, similar to the other test results, adding vegetable oil to the final mixture can increase the softness of the mixtures and subsequently decrease the ITS values. Among the mixtures containing paraffinic oil, the highest ITS is related to the mixtures in which the RA is added directly to the RAP material. On the other hand, the mixtures containing aromatic extract have higher ITS value when the aromatic extract is mixed in the virgin asphalt binder. It is seen that the wet and dry ITS results follow the same trend. However, the rate of the changes in the wet ITS values is less than that of the dry ITS.

The TSR values are calculated and presented in Figure 7. The results show that all TSR values are higher than the acceptable threshold (80%) of AASHTO T283 [34]. The reason is that both RAP and virgin aggregates are limestones which are classified as hydrophobic aggregates and less sensitive to moisture damage. However, it can be seen that using RAP material in the newly produced mixtures has led to a slight reduction in the TSR parameter. The reason can be due to the negative effect of the aged RAP asphalt binder on the cohesion of the total asphalt binder in the mixture and the adhesion between the RAP materials and virgin aggregates [48,49].

The effect of RAs on the moisture resistance of the RAP-blended mixtures is highly dependent on the RA type. The weakest moisture performance among RAs is related to vegetable oil, which can be owing to the existence of the chemical functional groups in this oil [50]. The paraffinic oil and aromatic extract can improve the moisture resistance of the mixtures if used in the correct location, which approves the previous research results conducted by Haghshenas et al. [51]. As all the mixtures have high resistance against moisture damage, the effect of RA location on the TSR value cannot be followed as the changes between the mixtures are too low. However, even the small changes approve the results of the rutting and cracking tests, and it is shown that the best locations for adding the vegetable, paraffinic and aromatic extract oils in terms of moisture resistance are in the final mixture, virgin binder, and in the RAP, respectively.

3.4. Statistical Analysis

The analysis of variance (ANOVA) was performed in order to investigate the significance of the results of the dependent variables. The rut depth, FI, fracture toughness, and dry and wet ITS were considered the dependent variables, and the RAP content and the type and locations of the RAs were considered the independent variables. The results of the statistical analysis are presented in

Table 7. The ANOVA results: (a) RD, (b) FI, (c) KI, (d) Dry ITS, and (e) Wet ITS.

Source	Type III Sum of Squares	Degree of Freedom (df)	Mean Square	F-Values	p-Values
(a) RD
Corrected Model	130.071	20	6.504	105.983	0.000
Intercept	2553.265	1	2553.265	41,608.329	0.000
RAP	79.327	2	39.663	646.358	0.000
RAs	1.199	2	0.599	9.769	0.001
location	3.613	2	1.807	29.439	0.000
Error	1.289	21	0.061
Total	2942.826	42
Corrected Total	131.360	41
R Squared = 0.990 (Adjusted R Squared = 0.981)
(b) FI
Corrected Model	24.380	20	1.219	17.545	0.000
Intercept	1013.920	1	1013.920	14,593.106	0.000
RAP	16.020	2	8.010	115.286	0.000
RAs	0.646	2	0.323	4.647	0.015
location	0.481	2	0.241	3.464	0.041
Error	2.918	42	0.069
Total	1203.980	63
Corrected Total	27.299	62
R Squared = 0.860 (Adjusted R Squared = 0.793)
(c) KI
Corrected Model	0.427	20	0.021	32.150	0.000
Intercept	21.257	1	21.257	32,037.527	0.000
RAP	0.219	2	0.110	165.060	0.000
RAs	0.018	2	0.009	13.207	0.000
location	0.013	2	0.006	9.691	0.000
Error	0.028	42	0.001
Total	27.033	63
Corrected Total	0.454	62
R Squared = 0.939 (Adjusted R Squared = 0.909)
(d) Dry ITS
Corrected Model	305,844.090	20	15292.205	33.845	0.000
Intercept	25,981,777.277	1	25,981,777.277	57,503.065	0.000
RAP	179,651.845	2	89,825.923	198.803	0.000
RAs	3235.914	2	1617.957	3.581	0.037
location	4091.422	2	2045.711	4.528	0.017
Error	18,976.982	42	451.833
Total	32,307,920.077	63
Corrected Total	324,821.072	62
R Squared = 0.942 (Adjusted R Squared = 0.914)
(e) Wet ITS
Corrected Model	171,427.857	20	8571.393	19.447	0.000
Intercept	18,665,108.256	1	18,665,108.256	42,346.988	0.000
RAP	95,880.956	2	47,940.478	108.766	0.000
RAs	9057.862	2	4528.931	10.275	0.000
location	3245.562	2	1622.781	3.682	0.034
Error	18512.168	42	440.766
Total	23,031,073.211	63
Corrected Total	189,940.025	62
R Squared = 0.903 (Adjusted R Squared = 0.856)

. The significance level of the results is shown by the p-value so that the p-value less than 0.05 shows that the differences between the results are significant with a 95% of confidence level. It can be seen that the p-values for all parameters are less than 0.05, which means that the results of this study are statistically significant.

4. More Discussion on the Results

There exist some important discussions about the results in addition to what was stated in the previous sections that are presented separately in this section.

The results of this study show that regardless of the testing method and the location of adding the RA, various RAs cannot improve the performance of RAP-blended mixtures to be as desirable as the control mixture. For example, the FI indices of all RAP-blended mixtures were lower than that of the control mixture, even after modification with RA. Two issues can be raised in this case. Firstly, the mix-design of all RAP-blended mixtures is carried out by the SMD method. However, it seems that this volumetric procedure cannot guarantee the performance of the RAP-blended mixtures. Therefore, using performance-based mix design procedures such as the balanced mix design (BMD) is strongly recommended for RAP-blended mixtures. Secondly, in this study, conventional asphalt binder tests such as penetration and softening point are used to find the optimum RA content. The penetration, softening point, performance grade, and BMD are typically used for finding the optimum RA content. Although the RA content found by the conventional tests can suggest an estimation of the optimum RA content, the exact RA content that guarantees the performance of the asphalt mixture can be found by performance-based methods such as the performance grade and the BMD procedures.

Another important result of this study is that the performance of RAs is dependent on the location of adding the RA. For example, the HWT test results showed that if all the RA are added directly to the RAP material, the mixtures containing the aromatic extract have the highest rut depth. On the other hand, if the RAs that are added to virgin asphalt binder are considered, the paraffinic oil will result in the highest rut depth. Therefore, it can be concluded that comparing the performance of mixtures containing different RAs cannot be correct when all RAs are added in the same location. It is recommended that the appropriate location for adding each RA should be determined by its manufacturer so that each RA can be used in its best location.

5. Conclusions

In this study, the effect of the RA’s location on the performance properties of RAP-blended mixtures was investigated. For this purpose, vegetable, paraffinic and aromatic extract oils were used as RAs, which were added in three different locations, including in the virgin asphalt binder, in the hot RAP material, and the final mixture. The rutting, cracking, and moisture sensitivity parameters were investigated by conducting the HWT, SCB, and wet and dry ITS tests. The conclusions are as follows:

-

Adding RAP material stiffens the asphalt mixtures, and using RAs compensates for the stiffening effect of the RAP materials. Therefore, the rut depth is decreased by adding RAP and using RAs to increase the rut depth.

-

Increasing the RAP content results in a substantial reduction in the crack resistance of asphalt mixtures at intermediate temperatures. Using RAs improves the cracking resistance of the mixtures; however, the effectiveness of the RAs depends on their location.

-

The wet and dry ITS test results showed that the moisture resistance of the mixtures does not change significantly by changing the location of the RA. The results show that all TSR values are higher than the acceptable threshold (80%).

-

Based on the study results, the best location for adding the RAs highly depends on the RA type. The results of all tests of this study show that the best place for adding the vegetable, paraffinic and aromatic extract oils in terms are in the final mixture, virgin binder, and in the RAP, respectively.