Recycled Asphalt Materials: Binder Availability and Its Impact on Mix Performance (2026)
Chapter: 4 Summary and Suggested Research
CHAPTER 4 SUMMARY AND SUGGESTED RESEARCH
This chapter presents a summary of the project’s key findings, conclusions, and suggested areas for future research and implementation.
4.1 Summary
4.1.1 Development of RBA Test Method
This project proposed a laboratory test method for determining the RBA of asphalt mixtures containing RAM. This method, referred to as the Glass Beads method, was adapted from existing literature and refined based on proof-of-concept evaluations of mixtures prepared with LP-RAP samples under different aging conditions. The method introduced borosilicate glass beads as tracers for virgin aggregate during RAM mixture preparation. These glass beads were pre-blended with virgin aggregate, dry-mixed with RAM, and combined with virgin binder. After mixing, the glass beads were manually collected and subjected to solvent extraction and recovery to obtain the binder coating their surface, representing a blend of virgin binder and activated RAM binder.
The binder extracted from the glass beads was tested using a DSR to determine PGH. Virgin binder and fully blended virgin-RAM binder, obtained by solvent extraction and recovery of the same mixture without glass beads, were also tested to represent 0% and 100% RBA conditions, respectively. If no RAM binder was activated, the glass bead-extracted binder would resemble the virgin binder, exhibiting comparable PGH values and indicating 0% RBA. If RAM binder was fully activated, the glass bead-extracted binder would mimic the fully blended virgin-RAM binder, corresponding to 100% RBA. Partial activation of RAM binder would result in the glass bead-extracted binder with PGH values between those of the virgin binder and the fully blended virgin-RAM binder. RBA was calculated based on the relative PGH differences among these three binders. This method can typically be completed within three days, including one day for mixture preparation and collection of glass beads, one day for binder extraction and recovery, and one day for PGH testing.
4.1.2 RBA Sensitivity Evaluation of Field RAM Asphalt Mixtures
The proposed Glass Beads method was applied to asphalt mixtures prepared with a wide range of RAM sources, contents, and mix design and production variables. The RBA values obtained ranged from 48% to 100%. For mixtures containing a single RAM type (RAP or RAS), RBA decreased with increasing RAM content and with more aged RAM sources, suggesting that higher amounts and stiffness of RAM binder can limit its activation and mobilization within the mixture. Mixtures with RAS exhibited RBA values within the range observed for RAP asphalt mixtures, despite the higher stiffness associated with RAS binders. These findings suggest that intrinsic properties of RAM beyond binder stiffness may also affect RBA. One potential factor is aggregate particle agglomeration. Sieve analysis revealed that RAS is significantly less prone to agglomeration than RAP, which may improve activation of the RAS binder and contribute to higher RBA values. Additional factors may include improved dispersion of RAS due to its finer particle size and the higher proportion of virgin binder in the mixture, which can facilitate mobilization of the RAS binder during mixing.
The sensitivity of RBA to common mix design and production variables was evaluated using two high-RAM mix designs: one containing 40% RAP and another with 20% RAP and 3% RAS. The changes in the RBA values for each variable are summarized in Table 24.
Table 24. Summary of RBA Sensitivity Evaluation Results
| Sensitivity Evaluation Variables | Change in RBA Value | |
|---|---|---|
| 40% RAP Mix Design | 20% RAP, 3% RAS Mix Design | |
| Change virgin binder from PG 64-22 (Source A) to PG 64-22 (Source B) | Increase by 30% | Increase by 22% |
| Change virgin binder from PG 64-22 (Source A) to PG 58-28 | Increase by 17% | Increase by 9% |
| Add a bio-based RA | Increase by 27% | Increase by 17% |
| Change virgin aggregate from low-absorptive GRN to high-absorptive LMS | Increase by 25% | Increase by 20% |
| Increase mixing temperature by 16.7°C | Increase by 52% | Increase by 25% |
| Decrease mixing temperature by 16.7°C with WMA | Decrease by 1% | Decrease by 18% |
| Simulate silo storage | Increase by 4% | Decrease by 22% |
For both mix designs, changing the virgin binder from PG 64-22 to another PG 64-22 from a different source or PG 58-28 increased RBA values. This indicates that virgin binder grade and source can influence RBA, and this effect is not solely related to binder stiffness, as the PG 58-28 binder produced RBA values between those of the two PG 64-22 binders. Adding a bio-based RA also resulted in increased RBA, suggesting that the softening and rejuvenation effects of RA enhance activation and mobilization of RAM binder. Virgin aggregate type was another influential factor for RBA. Mixtures prepared with high-absorptive LMS aggregates exhibited higher RBA values than those using low-absorptive GRN aggregates, when the same RAM and virgin binder were used. This is likely due to the higher virgin binder content required for LMS mixtures to account for aggregate absorption during volumetric mix design, and that the additional virgin binder facilitates RAM binder mobilization during mixing, resulting in higher RBA values.
Increasing the mixing temperature by 16.7°C raised RBA values for both mix designs. Conversely, reducing the mixing temperature by 16.7°C with WMA produced similar RBA for the 40% RAP mix design but a lower RBA for the 20% RAP, 3% RAS mix design. Laboratory simulated silo storage produced mixed results: the 40% RAP mix design showed a slight increase in RBA, while the 20% RAP, 3% RAS mix design exhibited a significant reduction. Exposure to elevated temperatures during laboratory simulated silo storage may enhance RAM binder activation, but it can also cause additional aging, restricting binder mobilization. The net impact may depend on the balance between these two competing factors, as evidenced in the differences between the two mix designs.
Overall, these findings confirm that RBA is an intrinsic property of the asphalt mixture, influenced not only by RAM type, source, and content, but also by mix design and production variables that affect RAM binder activation and mobilization. These variables should be considered when selecting RBA for asphalt mixtures containing RAM.
4.1.3 RBA Impact Analysis for High-RAM Asphalt Mixtures
Considering RBA requires adding extra virgin binder to compensate for inactive RAM binder in the mixture, with the amount determined by the RBA value, RAM content, and RAM binder content. This project evaluated the performance and cost implications of this approach using four
high-RAM asphalt mixtures prepared with two mix designs: one incorporating RAP only from Georgia and the other combining RAP and RAS from Wisconsin. Each design utilized two virgin binders: an unmodified binder and a PMA binder. Volumetric analysis showed that the additional virgin binder reduced RBR, air voids, VMA, and the D/P ratio, while increasing VFA. As a result of these changes, mixtures at A-OBC did not comply with the existing volumetric requirements specified by the corresponding state DOT. Index-based performance testing indicated that the added virgin binder significantly improved cracking resistance but reduced rutting resistance. The impact on rutting resistance was generally minor except for the Wisconsin high-RAP/RAS mixture containing an unmodified binder. According to the suggested test criteria in AASHTO PP 127, all mixtures exhibited good rutting resistance but poor cracking resistance at V-OBC with 100% RBA. When adjusted to A-OBC at reduced RBA conditions, three mixtures achieved balanced rutting and cracking resistance, while the Wisconsin high-RAP/RAS mixture with an unmodified binder showed satisfactory cracking resistance but marginal rutting resistance.
AMPT-based performance testing revealed that the additional virgin binder decreased mixture stiffness and improved fatigue cracking resistance although in some cases, the improvement was not statistically significant while accounting for test variability. FlexPAVE simulations further demonstrated that for three out of the four mixtures evaluated, increasing virgin binder to raise V-OBC to A-OBC improved predicted pavement fatigue cracking performance and could enable the use of thinner asphalt layers in new construction projects from a structural design perspective.
The material cost analysis showed that the selected Georgia high-RAP asphalt mixtures were approximately 15% more expensive at A-OBC compared to V-OBC, while Wisconsin high-RAP/RAS asphalt mixtures showed a cost increase ranging from 9% to 12%, depending on the virgin binder used. Despite these increases, all asphalt mixtures at A-OBC remained more cost-effective than their corresponding virgin asphalt mixtures and had material costs comparable to mixtures containing 25% RAP at 100% RBA. LCCA of a hypothetical overlay project indicated that Georgia high-RAP asphalt mixtures would require an estimated life extension of two to three years to offset the additional material costs associated with the extra virgin binder at A-OBC. For Wisconsin high-RAP/RAS asphalt mixtures, the required life extension ranged from one to four years, depending on the virgin binder type and the service life of the mixture at V-OBC.
Overall, these findings indicate that considering RBA is a viable approach for improving cracking resistance to achieve balanced performance in high-RAM asphalt mixtures with unmodified or PMA binders. They also underscore the importance of verifying rutting resistance and, when necessary, making additional mix design adjustments to enhance rutting resistance for RBA-adjusted asphalt mixtures, especially those using unmodified binders. Furthermore, this approach may necessitate relaxing current volumetric requirements for RBA-adjusted asphalt mixtures to account for changes in air voids, VMA, and other properties when transitioning from V-OBC to A-OBC. While this approach will increase material costs due to the additional virgin binder required, potential life-cycle savings could offset these expenses through improved pavement performance and reduced maintenance and rehabilitation needs. Although these findings are based on four high-RAM mixtures from Georgia and Wisconsin, they are expected to be applicable to similar mixtures from other states; however, verification in future research is warranted.
4.1.4 Development of Research Implementation Materials
Based on the promising findings of this project, several supporting materials were developed to facilitate the implementation of the RBA approach for asphalt mixtures containing RAM. These materials, provided in the appendixes, include: a draft test method for determining the RBA using the Glass Beads method (Appendix C); a draft test method for evaluating the impact of RBA on performance properties of asphalt mixtures containing unmodified, PMA, and rubberized asphalt binders (Appendix D); proposed revisions to AASHTO PP 127, incorporating updated guidance on the RBA approach for producing and evaluating balanced and durable high-RAM asphalt mixtures (Appendix E); proposed revisions to AASHTO M 323 (Appendix F) and AASHTO R 35 (Appendix G) to incorporate RBA into the Superpave volumetric mix design system; and a practical implementation guide document for state DOTs (Appendix H).
4.2 Suggested Research
To enhance the reliability and applicability of the refined Glass Beads method to determine RBA, future research should address the following limitations and areas for improvement in the test procedure:
- Variability in Solvent Extraction and Recovery: Solvent extraction and recovery can introduce variability in PGH results for both the extracted binder from the glass beads and the fully blended virgin-RAM binder, which in turn affects the accuracy of RBA calculations. Automated extraction and recovery procedures may reduce this variability; however, they require specialized equipment that may not be readily available to state DOTs and asphalt contractors for routine use.
- Selection of Extraction Solvents: TCE has proven effective for extracting asphalt binders from mixtures containing several unmodified binders (PG 58-28, PG 58S-28, and PG 64-22) as well as a PG 58V-28 PMA binder evaluated in the project. However, it was ineffective for mixtures containing a PG 76-22 PMA binder. This finding highlights the need for future research to develop solvent selection guidelines tailored to various types of PMA binders, including rubberized asphalt and other modified binders not evaluated in this project.
- Differentiation in Binder PGH Values: Accurate RBA calculations require sufficient differentiation in PGH values between the virgin binder and the fully blended virgin-RAM binder. This condition may not be met in mixtures with low RAM content or when the virgin and RAM binders have similar PGH values. Alternative test methods should be explored for such scenarios.
- Operating Procedures and Precision and Bias Statements: Comprehensive ruggedness evaluations and inter-laboratory studies are recommended to support broader implementation of the refined Glass Beads method. These efforts will help standardize operating procedures and establish precision and bias statements to guide assessments of within-laboratory repeatability and between-laboratory reproducibility.
Further research and implementation efforts are needed to assist state DOTs in establishing appropriate RBA value(s) tailored to local materials and conditions, verifying their impact on mixture performance for various types of asphalt binders, and integrating the RBA approach into existing mix design and production practices for asphalt mixtures containing RAM. For agencies that use volumetric properties such as air voids and VMA as pay factor items for acceptance, future
research is needed to guide the revision of production targets and tolerances for RBA-adjusted asphalt mixtures, especially those with high RAM contents.
