The Effects of Reclaimed Asphalt Pavement Modification on the Delta T_c Parameter for PG58-XX and PG64-XX Asphalt Binders

Abstract

The use of reclaimed asphalt pavement (RAP) in asphalt mixtures has increased due to its economic and environmental benefits. However, RAP integration can negatively impact the durability and performance of asphalt binders, particularly at low temperatures. This study evaluates the effects of RAP modification on the $\Delta T_C$ parameter, a key indicator of binder brittleness and resistance to non-load-related cracking, focusing on PG XX-34 and PG XX-28 grades commonly used in Kansas. Laboratory testing was conducted on virgin and RAP binders subjected to Rolling Thin-Film Oven (RTFO) and Pressure Aging Vessel (PAV) aging. Blended binders were prepared with RAP replacement levels of 15%, 25%, and 40%. The critical temperatures $T_{C,m}$, $T_{C,S}$, and $\Delta T_C$ values were calculated using data from Bending Beam Rheometer (BBR) testing. The results showed that increasing RAP content generally led to more negative $\Delta T_C$ values, indicating reduced relaxation capacity and higher susceptibility to thermal cracking. RAP source variability also affected performance, with some sources causing more severe deterioration than others. These findings highlight the limitations of conventional linear blending assumptions and underscore the need for improved RAP characterization in binder selection. The study recommends limiting RAP replacement to 25% unless the RAP source demonstrates favorable properties, incorporating $\Delta T_C$ thresholds (−2.5 °C and −5.0 °C) into binder specifications, and further investigating RAP–virgin binder interactions to enhance long-term pavement performance. The findings support the potential adoption of $\Delta T_C$ as a specification criterion for binder evaluation, helping agencies like the Kansas Department of Transportation (KDOT) balance binder durability and RAP use.

1. Introduction

The Performance Grade (PG) specification, outlined in AASHTO M320 [1], has been the standard method for evaluating asphalt binders used in pavement construction across the United States for over two decades. PG provides a structured approach to binder selection, ensuring suitability for road construction based on climate and loading conditions. However, with the growing incorporation of reclaimed asphalt pavement (RAP), recycled asphalt shingles (RASs), and polymer-modified binders, concerns have arisen regarding the adequacy of PG specifications in predicting long-term performance, particularly under aging conditions [2,3].

The use of RAP has gained popularity due to its cost-effectiveness and environmental benefits. By reusing aged asphalt binder and aggregates, RAP reduces the demand for virgin materials, conserving natural resources and minimizing the environmental impact of asphalt production [4,5,6]. Additionally, RAP incorporation lowers material costs while maintaining satisfactory performance in many cases. Studies have shown that well-designed asphalt mixtures containing RAP can exhibit performance comparable to virgin mixtures, particularly in terms of rutting resistance [7,8]. However, higher RAP content can significantly alter binder properties, leading to an increase in the effective low-temperature PG of the mixtures [9,10]. These challenges underscore the need for alternative parameters to evaluate RAP-modified binders more effectively and ensure long-term pavement durability.

Researchers have proposed various approaches to address durability challenges in asphalt binders. Anderson et al. [3] proposed $\Delta T_C$ as a parameter to assess non-load-related cracking in asphalt binders. $\Delta T_C$ is the difference between the temperature at which the binder becomes too stiff and the temperature at which it can no longer relax stress, helping predict how likely the binder is to crack at low temperatures. Yan et al. [11] further investigated the effects of oxidative aging on $\Delta T_C$, while other studies examined the impact of RAP modification on $\Delta T_C$ of virgin binders [2,5]. Sharma et al. [12] conducted an experimental study evaluating $\Delta T_C$, designed to assess the short- and long-term performance of binders with varying levels of RAP, analyzing changes in their chemical composition and characteristics.

The $\Delta T_C$ parameter has emerged as a key metric for evaluating the durability and cracking potential of asphalt binders, particularly connecting aging and RAP content. Introduced by Anderson et al. [3], $\Delta T_C$ is defined as the difference between the critical temperatures at which the binder meets the stiffness ($T_{C,S}$) and relaxation ($T_{C,m}$) limits in the Bending Beam Rheometer (BBR) test. The BBR test measures the creep stiffness ($S$) and the creep rate ($m$-value), these values are used to determine the critical temperatures $T_{C,S}$ and $T_{C,m}$, which form the basis for calculating $\Delta T_C$.

This parameter provides insights into a binder’s ability to relax stress and resist non-load-related cracking, such as thermal cracking. More negative $\Delta T_C$ values indicate reduced relaxation properties and higher susceptibility to cracking, making it a useful tool for assessing long-term binder performance [11,13,14]. Researchers have recommended $\Delta T_C$ thresholds of −2.5 °C as a warning level and −5.0 °C as a critical failure limit to ensure binders maintain adequate stress relaxation and resistance to premature cracking [3].

While different agencies use different criteria on the level of use for RAP, the Kansas Department of Transportation (KDOT) permits the use of high RAP content in asphalt mixtures, with levels reaching up to 40%, making it essential to evaluate the impact of RAP on binder performance. While RAP offers economic and environmental benefits, its high-aged binder content can significantly alter the relaxation properties of asphalt binders, potentially leading to increased stiffness and higher susceptibility to low-temperature cracking. Traditional binder specifications, such as AASHTO M320 [1], do not adequately account for these effects, creating a need to explore alternative performance parameters.

This study examines the impact of incorporating different percentages of RAP (15%, 25%, and 40%) on the $\Delta T_C$ parameter for asphalt binders with high temperature grading of 64 and 58, commonly used in Kansas. Given that $\Delta T_C$ is a promising indicator of binder brittleness and stress relaxation capacity at low temperatures, this research aims to assess its feasibility as a specification parameter for RAP-containing binders. The evaluation includes $\Delta T_C$ analysis of virgin and RAP binders under short- and long-term aging conditions, followed by the determination of $\Delta T_C$ using the BBR test. This approach provides insight into how RAP content affects low-temperature performance and cracking susceptibility. The findings will support the development of guidelines for RAP usage and binder selection in Kansas, addressing both pavement performance and sustainability considerations.

2. Materials and Methods

2.1. Materials

This study utilized both virgin binders and RAP binders to evaluate their performance under various blending and aging conditions. The KDOT provided four virgin binders and four RAP binders recovered from different highway projects across the state, as summarized in

Table 1. Materials used and sources.

Original Binders (Virgin)
Binder ID	PG Grading
V1	PG 64-34 (FHR)
V2	PG 64-28 (P66)
V3	PG 58-28 (Valero)
V4	PG 58-34 (FHR)
Reclaimed Asphalt Pavement (RAP)
RAP ID	Source
R1	21-2159 31 KA 5850-01
R2	21-2158 50-028 KA 5886
R3	21-2157 54-106 KA 5843-01
R4	21-2156 70-028 KA 5507-01

.

The four virgin binders used in this study represent a range of PG classifications commonly used in Kansas. They were supplied by Flint Hills Resources (FHR), Phillips 66 (P66), and Valero, providing a diverse selection of materials for evaluation.

The four RAP binders were recovered from asphalt pavement milling operations on different highway projects across Kansas. The properties of each RAP material vary based on its depth of removal and the historical pavement treatments before reclamation, which influenced their aging levels and rheological characteristics. The key features of each RAP source are summarized as follows:

• RAP1: Variable-depth cold milling, composed of a combination of chip seal, crack sealant, surface recycle, and hot mix asphalt (HMA).
• RAP2: 2-inch cold mill removal, containing 2012 crack sealant, ~2005 SM-9.5T PG 70-28 HMA (1.5-inch layer), and ~2005 SM-19A PG 70-28 HMA (0.5-inch layer).
• RAP3: 1-inch cold mill removal from two locations:
• Haskell Co.: Some 2014 crack sealant, 2013 SR-12.5A PG 58-28 (1-inch layer).
• Gray Co.: 2017 UBAS-Novachip (~5/8-inch layer), 2012 surface recycle (~3/8-inch layer).
• RAP4: 2.5-inch cold mill removal, containing 2013 chip seal, 2008 UBAS-Novachip (~5/8-inch layer), 2006 UBAS-Novachip (~5/8-inch layer), and ~2002 SM-9.5T PG 70-28 HMA (1.25-inch layer).

These RAP sources differ significantly in aging, and previous surface treatments, affecting their stiffness, viscosity, and relaxation properties, ultimately influencing the performance of RAP-modified asphalt mixtures.

2.2. Experimental Procedures

To simulate the aging conditions experienced during asphalt production and service life, this study employed Rolling Thin-Film Oven (RTFO) aging for short-term aging and Pressure Aging Vessel (PAV) aging for long-term aging. RTFO aging was applied to all virgin binders to replicate the aging that occurs during mixing and placement, while PAV aging was used to simulate extended oxidative aging over the binder’s service life. Given that RAP binders had already undergone significant in-field aging, they were subjected to PAV aging only. These standardized protocols ensured a consistent evaluation of binder aging effects across different RAP replacement levels.

The RTFO aging process simulates the short-term aging of asphalt binders that occur during production and placement. Following AASHTO T240-22 [15], binder samples (35 ± 0.5 g) were poured into glass containers and conditioned for 60 min before being aged in the RTFO chamber as shown in Figure 1. The aging process was conducted at 163 °C for 85 min under an air-pressure environment, ensuring uniform oxidation and volatilization effects. The RTFO apparatus accommodates up to eight bottles simultaneously during the aging process.

The PAV aging procedure replicates long-term oxidative aging, simulating in-service binder aging over pavement life, as per AASHTO R28-22 [16]. RTFO-aged virgin binders and RAP binders (without RTFO aging) were heated and poured into stainless steel PAV pans (50 ± 0.5 g each) before being subjected to 2.1 ± 0.1 MPa of air pressure for 20 h. The PAV system allows for simultaneous processing of up to ten pans, ensuring controlled long-term aging conditions as shown in Figure 2.

A Ross 100 L high-shear laboratory mixer, depicted in Figure 3, was utilized to achieve homogeneous blending of virgin and RAP binders at 15%, 25%, and 40% replacement levels. Virgin and RAP binders were measured, heated to the designated temperature, and blended at 5000 rpm for 2 min to ensure consistency without spillage. Vacuum Degassing Oven (VDO) treatment was applied only to pure PAV-aged RAP samples to remove entrapped air.

The BBR test was used to evaluate the low-temperature performance of the asphalt binders by measuring flexural creep stiffness ($S$) and the creep rate ($m$-value), which indicate the binder’s ability to resist thermal cracking and relax stress, respectively. According to AASHTO T313 [17], the maximum allowable stiffness is 300 MPa, and the minimum acceptable m-value is 0.300. These values were used to calculate the corresponding critical temperatures $T_{C,S}$ and $T_{C,m}$, as discussed in the next section.

All BBR tests were conducted under controlled conditions, with specimen preparation, standardization, and temperature conditioning following manufacturer and AASHTO T313 [17] specifications, as illustrated in Figure 4. Due to material limitations, testing was primarily performed on single replicates, with additional replicates used as needed for verification.

2.3. Parameter Calculation

The $\Delta T_C$ parameter is calculated as the difference between $T_{C,S}$ and $T_{C,m}$, serving as an indicator of an asphalt binder’s aging resistance and low-temperature performance, particularly in RAP-modified binders, Equation (1). These critical temperatures are determined through interpolation, where $T_{C,S}$ represents the temperature at which the binder meets the stiffness criterion (300 MPa), Equation (2), and $T_{C,m}$ corresponds to the temperature at which the $m$-value reaches the specification limit of 0.3, Equation (3).

$$\Delta T_c=T_{c,S}-T_{c,m}$$

(1)

$$T_{c,S}=T_1+\left[{\displaystyle \frac{(T_1-T_2)\times (Log(300)-Log(S_1))}{Log(S_1)-Log(S_2)}}\right]-10$$

(2)

$$T_{c,m}=T_1+\left[{\displaystyle \frac{(T_1-T_2)\times (0.300-m_1)}{m_1-m_2}}\right]-10$$

(3)

In the equations, $S_1$ and $S_2$ represent the creep stiffness values at temperatures $T_1$ and $T_2$, respectively, measured in MPa. Similarly, $m_1$ and $m_2$ denote the creep rates at the corresponding temperatures $T_1$ and $T_2$. The temperatures $T_1$ and $T_2$ are expressed in degrees Celsius (°C) and represent the points at which the stiffness $S$ and the slope $m$-value pass and fail the specified criteria, respectively. These parameters collectively contribute to the calculation of the $\Delta T_C$ value, which serves as an indicator of the binder’s low-temperature performance and aging resistance.

Thresholds have been established to evaluate binder performance based on $\Delta T_C$ values. A $\Delta T_C$ greater than −2.5 °C is considered acceptable (safe category), indicating sufficient relaxation properties. Binders with $\Delta T_C$ values between −2.5 °C and −5.0 °C fall within a warning range (watch category), exhibiting increased brittleness and susceptibility to cracking. A $\Delta T_C$ below −5.0 °C is classified as a critical failure threshold (critical category), suggesting severe susceptibility to low-temperature cracking. These criteria allow for a more comprehensive assessment of binder durability, particularly in high-RAP mixtures, where oxidation and increased stiffness may impact long-term pavement performance.

3. Results

3.1. Pure Virgin Binders

Figure 5 illustrates the critical temperatures $T_{C,S}$ and $T_{C,m}$ and the corresponding $\Delta T_C$ values for the four virgin binders evaluated under RTFO and PAV aging conditions. The results indicate that all virgin binders are controlled by the $m$-value, meaning that their performance is primarily dictated by their ability to relax thermal stress rather than stiffness. Physically, this implies that these binders have reduced relaxation properties, making them more susceptible to thermal cracking as they age.

Among the binders, only virgin binder (V2), with performance grading of PG 64-28, falls into the critical category, with a $\Delta T_C$ value of −6.27 °C, which is below the failure threshold of −5.0 °C, indicating significant stress relaxation loss after aging and increased cracking susceptibility. This aligns with its high $T_{C,m}$ value of −22.26 °C, suggesting that V2 has poor stress relaxation at low temperatures, making it more prone to brittle cracking as it ages. Additionally, V2 has a $T_{C,S}$ value of −28.53 °C, indicating that while its stiffness is relatively high, its inability to relax thermal stresses is the dominant factor leading to its highly negative $\Delta T_C$. Binder V3 (PG 58-28), which has a $\Delta T_C$ of −1.17 °C, also shows reduced stress relaxation with a $T_{C,m}$ of −26.09 °C, though it remains within the safe range.

In contrast, V1 (PG 64-34) and V4 (PG 58-34) demonstrate $\Delta T_C$ values within the safe range at −0.54 °C and −0.42 °C, respectively, indicating better relaxation properties and lower susceptibility to low-temperature cracking. This is supported by their more negative $T_{C,m}$ values, which suggest stronger stress relaxation capabilities at low temperatures. Additionally, their stiffness-related $T_{C,S}$ values are more negative, indicating a lower risk of embrittlement.

The differences in $\Delta T_C$ values reflect the aging characteristics of the binders, particularly the impact of PG on stress relaxation. PG XX-34 binders (V1 and V4), designed for colder climates, tend to retain their relaxation capacities better under aging conditions, resulting in less negative $\Delta T_C$ values. Conversely, the higher stiffness and poorer relaxation properties of PG XX-28 binders (V2 and V3) under aging highlight their greater susceptibility to low-temperature cracking.

3.2. Pure RAP Binders

Figure 6 presents the critical temperatures $T_{C,S}$ and $T_{C,m}$ and the corresponding $\Delta T_C$ values for the four RAP binders after PAV aging, confirming that all are $m$-controlled, meaning their low-temperature performance is dictated by stress relaxation rather than stiffness. Among the RAP binders, RAP1 exhibited the most negative $\Delta T_C$ value at −16.91 °C, indicating severe relaxation loss and the highest cracking susceptibility. RAP4 demonstrated the least negative $\Delta T_C$ value at −10.67 °C, suggesting slightly better low-temperature performance. RAP2 and RAP3, with $\Delta T_C$ values around −15 °C, fall between RAP1 and RAP4 in cracking resistance. Since all RAP binders have $\Delta T_C$ values far below the −5.0 °C threshold, they are classified as critical for low-temperature performance.

The age and composition of RAP binders significantly affect $\Delta T_C$ values, with older and more heterogeneous materials exhibiting poorer relaxation properties. RAP1, sourced from variable-depth milling, contains highly aged treatments, leading to severe oxidation and embrittlement. RAP2 consists of 2005–2012 pavement layers, contributing to stiffness and reduced stress relaxation. RAP3, with a shallow 1-inch milling depth, likely incorporated more aged materials, reducing its relaxation capacity, despite the presence of some newer layers. RAP4, removed from a deeper 2.5-inch milling, likely incorporated less aged materials, resulting in higher $\Delta T_C$ and slightly better performance than the others.

These findings underscore the importance of evaluating RAP binder properties before incorporation into asphalt mixtures, particularly in cold climates where thermal cracking is a major concern. Given their severely compromised relaxation capacity, modification strategies such as blending with softer virgin binders, rejuvenators, or polymer additives should be considered to improve long-term performance.

3.3. Blended Mixtures

The influence of increasing RAP binder content on the low-temperature performance, $\Delta T_C$, was evaluated by blending 15%, 25%, and 40% RAP binders with RTFO-PAV aged virgin binders. A total of 48 samples were prepared, including 0% RAP (pure virgin) and 100% RAP (pure RAP) for comparison.

Figure 7 shows that increasing RAP content generally resulted in more negative $\Delta T_C$ values for V1, indicating reduced relaxation capacity and greater susceptibility to low-temperature cracking. Exceptions were observed at 15% RAP replacement for RAP1, RAP3, and RAP4, where $\Delta T_C$ remained relatively unchanged. Among the 12 samples, 5 fell into the watch category, including all RAP2 blends and the 40% RAP blends with RAP1 and RAP3. The remaining samples remained in the safe category. RAP2 produced the lowest (most negative) $\Delta T_C$ values, while RAP4 had minimal impact on $\Delta T_C$, maintaining values similar to the pure virgin binder. The standard deviation (SD) for the blended binders was 1.3, with a coefficient of variation (COV) of 79%, indicating considerable variability in $\Delta T_C$ across the blended binders.

Figure 8 and Figure 9 show that increasing RAP content also led to higher critical temperatures $T_{C,S}$ and $T_{C,m}$ for most samples, reflecting increased stiffness and reduced relaxation properties. However, one RAP2 sample (40%) deviated from this trend, showing a decrease in $T_{C,S}$ value. Overall, RAP2 blends exhibited the most significant risk, while RAP4 blends had the least impact on low-temperature performance.

Figure 10 shows the effects of blending RAP binders with V2 on $\Delta T_C$ values. Unlike the trends observed for V1, the $\Delta T_C$ behavior for V2 was inconsistent, with six samples showing higher $\Delta T_C$ values and the remaining six showing lower $\Delta T_C$ values as RAP content increased. Of these, five samples fell into the watch category, while seven samples were below the critical limit of −5 °C, indicating a significant risk of low-temperature cracking. The SD for the blended binders was 2.6, with a COV of 43%, reflecting moderate variability in $\Delta T_C$ values across these samples.

Figure 11 and Figure 12 illustrate the changes in critical temperatures $T_{C,S}$ and $T_{C,m}$ for these blends. Unlike V1, most V2 samples exhibited a cooling trend in critical temperatures, particularly for $T_{C,m}$. This cooling trend was most pronounced in blends with RAP3 and RAP4, where lower $T_{C,m}$ values were observed compared to the pure virgin binder. Since all of these samples are $m$-controlled, the change in $T_{C,m}$ has a greater influence on $\Delta T_C$ values, potentially explaining the higher $\Delta T_C$ values observed in some RAP3 and RAP4 samples.

These results suggest that the variability in $\Delta T_C$ and critical temperatures for V2 blends is strongly influenced by the RAP source, with RAP3 and RAP4 contributing to a more significant reduction in $T_{C,m}$ and better $\Delta T_C$ performance relative to RAP1 and RAP2. However, the number of samples below the critical $\Delta T_C$ threshold emphasizes the need for careful evaluation of RAP content and source when designing mixtures with PG 64-28 binders.

Figure 13 shows the change in $\Delta T_C$ values of V3 binders blended with RAP at various contents. Similarly to V1, most samples exhibited a decline in $\Delta T_C$ (more negative values) as RAP content increased, indicating reduced relaxation capacity. An exception was observed for 15% RAP1, where $\Delta T_C$ remained relatively stable. Of the twelve samples, six fall within the watch category, while two samples (40% RAP1 and 40% RAP2) exceeded the critical threshold. The remaining samples stayed within the safe range. The SD for the blended binders was 1.8, with a COV of 49%, indicating moderate variability in $\Delta T_C$ for these samples.

Figure 14 and Figure 15 illustrate changes in critical temperatures $T_{C,S}$ and $T_{C,m}$ with increasing RAP content. A general warming trend was observed for $T_{C,m}$, with increased RAP content, except for blends with 15% RAP1 and 15% RAP3, where $T_{C,m}$ values remained relatively unchanged. For $T_{C,S}$, a cooling trend was observed for most samples, but this had no significant effect on the overall cooling trend of $\Delta T_C$.

These findings suggest that increasing RAP content generally leads to a decline in $\Delta T_C$ values for most samples, highlighting reduced stress relaxation capacity. However, the cooling trend observed in $T_{C,S}$ for most samples did not appear to influence the overall $\Delta T_C$ trend, as $T_{C,m}$ primarily drives $\Delta T_C$ changes. This indicates that while RAP incorporation increases stiffness, the observed decline in $\Delta T_C$ is predominantly influenced by the ability of the binder to relax stresses (controlled by $T_{C,m}$).

Figure 16 presents the changes in $\Delta T_C$ values for V4 binders as RAP content increases. As expected, higher RAP content led to a more negative $\Delta T_C$, reflecting reduced stress relaxation capacity. Out of twelve samples, six fell into the watch category, while three samples, 40% and 25% RAP2, and 40% RAP3, exceeded the critical threshold. The remaining samples remained in the safe category. Among the RAP sources, RAP2 blends consistently exhibited the most negative $\Delta T_C$ values, indicating a higher susceptibility to low-temperature cracking. The SD for the blended binders was 2.5, with a COV of 62%, highlighting substantial variability in $\Delta T_C$ for these samples.

Figure 17 and Figure 18 illustrate the corresponding changes in critical temperatures $T_{C,S}$ and $T_{C,m}$ with increasing RAP content. A consistent warming trend was observed across all 12 samples, with increases in both stiffness and slope critical temperatures compared to the pure virgin binder. This trend indicates that higher RAP content significantly increases binder stiffness and reduces its ability to relieve thermal stresses. These findings align with the overall reduction in $\Delta T_C$ values, emphasizing that RAP incorporation has a pronounced effect on both critical temperatures and binder performance.

The results for V4 further highlight the impact of RAP source variability. RAP2 blends showed the most significant adverse effects, followed by RAP3 at higher replacement levels, while RAP4 demonstrated the least impact on $\Delta T_C$ and critical temperatures. This underscores the importance of carefully selecting RAP sources and replacement levels to balance stiffness and relaxation properties in asphalt mixtures.

Table 2. Sample categories.

		V1	V2	V3	V4
RAP1	40%	W	C	C	W
	25%	S	C	S	W
	15%	S	C	S	W
RAP2	40%	W	C	C	C
	25%	W	C	W	C
	15%	W	C	W	W
RAP3	40%	W	C	W	C
	25%	S	W	W	W
	15%	S	W	S	S
RAP4	40%	S	W	W	W
	25%	S	W	W	S
	15%	S	W	S	S

S: safe; W: watch; C: critical.

presents the classification of the 48 samples into three distinct categories based on their $\Delta T_C$ values: safe, watch, and critical categories. Among the samples, the following groupings were made:

• Fourteen samples were categorized as safe, reflecting adequate relaxation properties.
• Twenty-two samples fell within the watch category, indicating increased susceptibility to low-temperature cracking.
• Twelve samples were classified in the critical category, showing significant risks for cracking due to poor relaxation capacity.

All 12 samples blended with V2 were classified in either the watch or critical category, likely due to the low $\Delta T_C$ value of pure V2 (−6.27 °C), which made the binder more prone to poor performance when combined with RAP. Additionally, all 12 samples blended with RAP2 and the 40% RAP blends with RAP1 and RAP3 fell into the watch or critical category, highlighting their adverse effects on base binder relaxation properties.

Notably, all samples with 40% RAP were classified either in the watch or critical categories, except for V1 + RAP4, indicating that 40% RAP replacement is generally unsuitable for maintaining acceptable $\Delta T_C$ values, especially with more aged or stiff RAP sources. These results emphasize the importance of optimizing both RAP content and source selection to minimize adverse impacts on low-temperature performance.

Table 3. BBR creep stiffness ($S$) values in MPa for the blended binders at −18 °C.

		V1	V2	V3	V4
RAP1	40%	355	396	410	261
	25%	207	318	367	170
	15%	264	265	309	210
RAP2	40%	197	228	248	153
	25%	179	249	272	146
	15%	161	232	269	144
RAP3	40%	265	250	322	212
	25%	238	274	249	145
	15%	235	218	256	108
RAP4	40%	245	294	353	230
	25%	202	271	314	150
	15%	147	317	299	133

shows how BBR creep stiffness ($S$) varies with increasing RAP content for the blended binders. In general, increasing RAP content from 15% to 40% results in higher stiffness values for most combinations. This trend indicates that binders become stiffer with more RAP, which is consistent with the expectation that aged RAP binders contribute to increased stiffness in the blended binder.

4. Discussion

4.1. Impact of RAP Content on Binder Performance

The incorporation of RAP binders into virgin binders significantly influenced $\Delta T_C$ values, with trends varying based on RAP content and source. In general, higher RAP content resulted in more negative $\Delta T_C$ values, reflecting reduced stress relaxation capacity and increased susceptibility to low-temperature cracking.

For V1, V3, and V4, $\Delta T_C$ consistently declined as RAP content increased, confirming that stiffness and oxidation effects from RAP binders negatively impact relaxation properties. However, at 15% RAP replacement, $\Delta T_C$ remained relatively stable for RAP1, RAP3, and RAP4, suggesting that at lower RAP levels, the virgin binder retains its dominant properties, mitigating the negative effects of RAP aging. In contrast, V2 exhibited inconsistent $\Delta T_C$ behavior, with some samples showing improvement while others deteriorated. This variability was primarily driven by changes in $T_{C,m}$, which dictates $m$-controlled behavior.

A critical observation was that all 40% RAP blends, except V1 + RAP4, fell into either the watch or critical categories, reinforcing that high RAP content significantly reduces stress relaxation capacity. RAP2 blends consistently exhibited the most negative $\Delta T_C$ values, leading to higher cracking risk, while RAP4 had minimal impact on $\Delta T_C$, maintaining values similar to the pure virgin binder. The variability in $\Delta T_C$ trends among different RAP sources underscores the importance of RAP characterization before incorporation into asphalt mixtures, considering factors such as the following:

• Material Age: Older RAP materials, such as RAP2 (2005–2012 pavement layers), showed the most severe deterioration, leading to significantly lower $\Delta T_C$ values and greater cracking susceptibility.
• Milling Depth: RAP4, removed from a deeper 2.5-inch milling, incorporated less aged materials, resulting in higher $\Delta T_C$ values and better relaxation capacity than RAP1, RAP2, and RAP3. Conversely, RAP3, with a shallower 1-inch milling depth, likely included more aged materials, worsening its low-temperature performance at higher RAP contents.
• Binder Source and Composition: The presence of highly oxidized surface layers, crack sealants, and recycled treatments in RAP1 and RAP2 contributed to their poor performance, whereas RAP4’s UBAS-Novachip layers helped retain better stress relaxation properties.
• RAP Percentage in Blends: Higher RAP content (40%) had a pronounced effect on $\Delta T_C$, especially in RAP1, RAP2, and RAP3 blends, whereas lower RAP percentages (15–25%) mitigated some of these negative effects.

4.2. Evaluation of $\Delta T_C$ as a Specification Parameter

The results confirm the suitability of $\Delta T_C$ as a robust parameter for assessing binder performance, particularly in evaluating aging effects and RAP incorporation on low-temperature behavior. $\Delta T_C$ combines stiffness $T_{C,S}$ and relaxation $T_{C,m}$ to provide a comprehensive measure of thermal cracking susceptibility. The use of $\Delta T_C$ thresholds effectively distinguished binder performance in this study. Samples with $\Delta T_C$ > −2.5 °C were classified as safe, while those between −2.5 °C and −5.0 °C were categorized as watch, and values below −5.0 °C were considered critical. These thresholds highlighted the increased risk of cracking with higher RAP content and allowed for a clear assessment of the impact of RAP sources.

While $\Delta T_C$ is a promising specification parameter, it should be complemented with additional evaluations to account for RAP variability. For example, despite high RAP content, RAP4 blends maintained better relaxation properties, suggesting that factors like milling depth, aging history, and material composition influence $\Delta T_C$ behavior. These findings indicate that $\Delta T_C$ alone may not fully capture performance variations across different RAP sources, necessitating a more comprehensive approach to binder evaluation.

4.3. Practical Implications

The results support the feasibility of adopting $\Delta T_C$ as a specification parameter for evaluating binder performance, particularly in RAP-incorporated mixtures. $\Delta T_C$ provides a reliable measure of low-temperature cracking risk, offering state agencies such as KDOT a practical tool for binder assessment.

To maintain acceptable $\Delta T_C$ values, RAP replacement should generally be limited to 25%, as higher levels increase cracking susceptibility. However, RAP4 exhibited minimal $\Delta T_C$ impact even at 40%, suggesting that higher RAP percentages may be feasible with well-preserved materials. In contrast, RAP2 significantly degraded $\Delta T_C$, highlighting the need for caution with highly aged RAP sources.

It is suggested that agencies such as KDOT should consider adopting $\Delta T_C$ thresholds of 2.5 °C and −5.0 °C as warning and critical limits, respectively, while prioritizing evaluations of RAP source variability. Field validation is necessary to confirm these recommendations and ensure alignment with Kansas-specific conditions. In addition to field validation, a more detailed investigation into the effects of RAP source properties including age, composition, and milling depth is needed to better understand their impact on binder performance and $\Delta T_C$ values. Further investigating RAP-virgin binder interactions will also be essential to enhance long-term pavement performance.

5. Conclusions and Recommendations

This study evaluated the impact of RAP content and source on binder performance using the $\Delta T_C$ parameter and explored its feasibility as a specification criterion for the KDOT. Based on the findings, the following conclusions and recommendations are drawn:

• Among the virgin binders, V2 exhibited a critical $\Delta T_C$ value of −6.27 °C, indicating a high risk of cracking. The remaining virgin binders demonstrated acceptable performance, with $\Delta T_C$ values above −2.5 °C. In contrast, all RAP binders fell into the critical category, with $\Delta T_C$ values ranging from −10.67 °C to −16.91 °C, reflecting their poor relaxation capacity and higher stiffness.
• RAP source significantly influenced $\Delta T_C$ of blended binders. RAP2 (containing highly oxidized 2005–2012 pavement layers) exhibiting the most negative $\Delta T_C$ values, while RAP4 (milled from deeper, less aged layers) had the least impact. These differences highlight that $\Delta T_C$ is influenced by both RAP content and source properties, such as age, composition, and milling depth.
• Blended samples with V2 consistently fell into the watch or critical categories, attributed to its already low $\Delta T_C$ value, which made it more susceptible to performance deterioration when combined with RAP.
• RAP replacement at 40% consistently resulted in critical $\Delta T_C$ values across all binders and sources (except V1 + RAP4), indicating that high RAP levels are generally unsuitable unless the RAP source exhibits favorable properties.

The following recommendations are made for better binder blending practices:

• RAP replacement levels should be limited to 25% unless the RAP source demonstrates favorable performance, such as RAP4. Binder specifications should also include $\Delta T_C$ thresholds (−2.5 °C and −5.0 °C) to classify binder performance and reduce low-temperature cracking risks.
• RAP sources should be evaluated based on their aging history, milling depth, and material composition before use to optimize mixture performance and mitigate cracking susceptibility.