Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios (2020)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

4 More than 90% of highways and roads in the United States are built using HMA or WMA mixtures. In the early 1990s, the Federal Highway Administration (FHWA) estimated that more than 90 million tons of asphalt pavements are milled off roads each year during resurfacing projects, and over 80% of RAP is recycled in new mixtures (FHWA 1993). Subsequent studies and surveys showed that this trend is increasing, with more recent estimates of more than 99% of some 76.2 million tons of RAP reused, making it the most recycled product in the United States. In addition, about 1 million tons of RAS are now used in paving applications (Copeland 2011; NAPA 2018). The use of RAP in HMA dates back to the early 1900s, with renewed focus on research and implementation in the 1970s and 1980s due to dramatic increases in the cost of oil, and thus asphalt binders and fuel, needed to produce asphalt pavements. Newcomb and Epps (1981) reviewed the technologies developed during this early period of recycling, which included drum mix plants, cold milling machines, vibratory compaction rollers, cold and hot in-place recy- cling techniques, and mix design methods, to increase RAP contents to maximize economic and environmental benefits. These symbiotic benefits are substantial and include conservation of natural resources (aggregate, binder, fuel, etc.), reduction in energy consumption, and reduction in emissions (including greenhouse gases). For example, in a relatively high 25% RAP content HMA mixture, Robinett and Epps (2010) indicated 10% energy savings, 10% emissions reduc- tions, and 20%–25% conservation of natural resources that translated into reduced production and construction costs. In 2017, cost savings totaled approximately $2.2 billion with recycled materials replacing virgin materials (NAPA 2018). Interest in recycling waned during the 1990s and recycling was not considered in the Strategic Highway Research Program (SHRP). Thus recycling technologies remained largely unchanged until 2008, when the cost of petroleum products significantly increased again. So highway agen- cies and the paving industry have developed a renewed interest in achieving higher RBRs in asphalt mixtures through the use of larger percentages of RAP and/or the addition of RAS from either manufacturer waste asphalt shingles (MWAS) or tear-off asphalt shingles (TOAS) for the same economic and environmental benefits. To provide an overall indication of the possible binder contribution from these recycled materials, RBR is defined according to Equation 1. [ ]( ) ( )= × + × ×100 Equation 1RBR Pb P Pb P Pb RAP RAP RAS RAS total where PbRAP = binder content of the RAP, PRAP = percentage of RAP by weight of mixture, PbRAS = binder content of the RAS, C H A P T E R 1 Introduction

Introduction 5 PRAS = percentage of RAS by weight of mixture, and Pbtotal = binder content of the combined mixture. In spite of the symbiotic benefits, state DOTs limit the use of RAP and/or RAS in asphalt mix- tures for reasons that include variability of the recycled materials and concerns about long-term mixture performance. In addition, mix design is more complicated and more time consuming, particularly with high RBRs between 0.3 and 0.5. The potential for the following construction and performance issues is also increased as RBRs increase and corresponding mixtures become stiffer and more brittle: • Compactibility/workability in cool weather, • Low-temperature cracking with accumulation of thermally induced stresses, • Fatigue cracking and microdamage accumulation leading to crack initiation and propagation with repeated loading, • Reflection cracking with repeated loading and daily/seasonal thermal stresses, and • Raveling with subsequent aging or moisture damage. Thus, the environmental and economic benefits must be compared to the potential increased risks associated with construction and performance to ensure that engineering benefits can also be realized. Mitigation of these construction and performance issues can be addressed through mix design with the use of higher binder contents, material selection with the use of softer binders that may be polymer modified, or additives such as recycling agents, as long as there are not compatibility concerns and mixture resistance to rutting and moisture damage is main- tained. Mitigation through the use of recycling agents includes the following (Tran et al. 2012; Mogawer et al. 2013; Im and Zhou 2014): • Partial restoration of stiffness and reversal of embrittlement caused by the addition of recycled materials at high RBRs, • Improvement in cracking resistance of mixtures with high RBRs without adversely affecting rutting resistance, and • Improvement in compactibility/workability (in some cases). Using lower production temperatures through the use of WMA technologies will also affect these construction and performance issues and possibly offset benefits of reduced aging with decreased blending of base and recycled binders and/or generate possible compatibility concerns with WMA additives, recycling agents, and base and recycled binders. Recycling agents were used in HMA in the early period of widespread recycling in the 1970s and 1980s for the purpose of realizing all three types of benefits—environmental, economic, and engineering. As RBRs continue to increase in the current period of widespread recycling, the use of recycling agents holds promise once again with proper understanding of their effectiveness in partially restoring rheology, its evolution with aging in HMA and WMA mixtures in both the laboratory and the field, and stiffness and cracking resistance of these binder blends and corre- sponding mixtures. Mix design procedures, including component material characterization to ensure that binder blends are restored as much as possible rheologically, specimen fabrication protocols to simulate field conditions, and production and construction best practices (includ- ing handling of recycled materials—fractionation and drying, for example), are needed to ensure adequate performance when using recycling agents. 1.1 Project Overview and Objectives This final report completes NCHRP Project 09–58, “The Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios.” Figure 1 illustrates how Phase 1, the two parts of Phase 2 (A and B), and Phase 3 contributed to the overall study that included five

Figure 1. NCHRP 09–58 overview.

Introduction 7 field projects in Texas (TX), Nevada (NV), Indiana (IN), Wisconsin (WI), and Delaware (DE). Phase 2B was requested based on concerns with the Phase 2A results that explored a limited set of materials combinations and indicated limited recycling-agent effectiveness with aging. Phase 2B explored some more fundamental tools and a wider range of materials combinations that included the use of an improved base binder, a softer base binder, three engineered recycling agents, and higher RBRs with TOAS. Table 1 presents the materials combinations explored throughout both parts of Phase 2 and Phase 3 that evolved based on a continuous review of results to select combinations that consider the limitations of materials, time, and budget and most efficiently meet the study objectives to • Assess the effectiveness of recycling agents at a selected dose to partially restore binder blend rheology and improve mixture cracking performance without adversely affecting rutting resistance, • Evaluate the evolution of recycling-agent effectiveness with aging, and • Recommend evaluation tools for assessing the effectiveness of recycling agents initially and with aging for mixtures with high RBRs in specific climatic regions. Many of the results for the materials combinations in Table 1 were intermediate results that are not presented in this report. During Phase 2B, coordination of field projects and procurement of materials for Phase 3 was also completed to tie the laboratory results to field performance. Each of the materials combinations listed in Table 1 that included the following recycling agent types and binder and mixture testing results facilitated engineering rejuvenated binder blends and corresponding mixtures and the development of evaluation tools for assessing the effectiveness of recycling agents initially and with aging for binders and mixtures with high RBRs: • Recycling agent types: – A1 and A2—aromatic extracts. – P—paraffinic oil. – T1 and T2—tall oils. – V1—vegetable oil. – V2 and V3—modified vegetable oils. – B1 and B2—reacted bio-based oils. These recycling agents are proprietary products labeled by generic descriptors that define their origin. Petroleum-based recycling agents in this study included aromatic extracts (A) that are traditional recycling agents refined from crude oil as a by-product of lube oil processing with dominant polar aromatic oil components and paraffinic oils (P) that are also refined as a by-product of lube oil processing, but may have similar performance as recycled engine oil bottoms (REOBs). Bio-based recycling agents are derived from plant life rather than petro- leum and in this study included tall oils (T) that are by-products of paper processing from pine trees, vegetable oils (V) and simple derivatives such as esters, and other bio-based oils (B) that appear to be chemically reacted, usually to reduce impacts of oxidative aging on rhe- ology. Based on limited data, V1 is a vegetable oil that consists of a mixture of glycerides and fatty acids, and V2 and V3 are engineered (modified) vegetable oils. B1 and B2 are reacted bio-based oils that consist of fatty amine derivatives and bio solvents. • Binder testing results: – PGL—low-temperature PG. – PGH—high-temperature PG. – DTc—the difference in continuous PG temperature for stiffness and relaxation properties in the bending beam rheometer (BBR); i.e., the critical temperature when S equals 300 MPa minus the critical temperature when m-value equals 0.30. – G-R—Glover-Rowe parameter and Black space analysis. – Td = 45°—crossover temperature.

TX 70-22 P (ΔTc -4.9) — — — Base Binder — — — — — — — — — — — 2.7% T1 (field) — — — — — — — — — — — — — — 0.3 0.2 TX 0.1 TX MWAS Control (no recycling agent) — — — — — — — — — — — — — — 2.7% T1 (field) — — — — — — — — — — — — — — TX 64-22 (ΔTc -4.6) — — — Base Binder — — — — — — — — — 2.7% T1 (field) — — — — — — — — — — 0.28 0.1 TX 0.18 TX MWAS DOT Control (no recycling agent) — — — — — 2.7% T1 (field) — — — — — — — 0.28 0.1 TX 0.18 TX MWAS 4.5% T1 (PGL) — — — — — — — — — 5.5% A1 (PGL) — — — — — — — — 4.0% V1 (PGL) — — — — — — — — — — — 4.0% B1 (PGL) — — — — — — — — — — — 12.5% T1 (ΔTc) — — — — — — — — — — 9.5% A1 (ΔTc) — — — — — — — — — — — — 8.5% V1 (ΔTc) — — — — — — — — — — — — 7.0% B1 (ΔTc) — — — — — — — — — — — — 6.0% T1 (PGH) — — — — — — — — — — — — — — 6.5% A1 (PGH) — — — — — — — — — — — — — 5.5% V1 (PGH) — — — — — — — — — — — — — — 6.5% B1 (PGH) — — — — — — — — — — — — — — NOTE: Gray shading indicates TX field project material combinations; — = not applicable. aAt rolling thin film oven (RTFO), 20 pressure aging vessel (PAV), and 40 PAV aging. bLong-term oven aging including binder master curve, G-R, Tδ = 45°, and FT-IR. cDescribed in Chapter 2. Materials Combinations Binder Testing Mortar Testing (PG) Mixture Testing Base Binder Recycled Materials Recycling Agent Dose and Type (Dose Selection Methodc) Rheology Chemical Characterization FT-IRa Agingb C I M R |E *| I- FI T BB R m (S liv er ) H W TT A PA U TS ST S- V EC D R BR R A P R A S G -R a T δ = 45 °a SAR-AD, MDSC Table 1. Materials combinations explored in NCHRP 09–58.

Materials Combinations Binder Testing Mortar Testing (PG) Mixture Testing Base Binder Recycled Materials Recycling Agent Dose and Type (Dose Selection Methodc) Rheology Chemical Characterization FT-IRa Agingb C I M R |E *| I- F IT B B R m (S liv er ) H W T T A P A U T SS T S- V E C D R B R R A P R A S G -R a T δ = 45 °a SAR-AD, MDSC TX 64-22 (ΔTc -4.6) 0.4 0.4 TX — Control (no recycling agent) — — — — — — — — — — — — — 7.5% T1 (PGL) — — — — — — — — — — — — 9.5% A1 (PGL) — — — — — — — — — — — 0.5 0.25 TX 0.25 TX MWAS Control (no recycling agent) — — — — — — — — — — — 7.5% T1 (PGL) — — — — — — — — — — — 9.0% T1 (PGH) — — — — — — — — — — — — — Control (no recycling agent) — — — — — — — — — — — 11.5% T1 (PGL) — — — — — — — — — — — 13.5% T1 (PGH) — — — — — — — — — — — — — NH 64-28 (ΔTc +1.2) — — — Base Binder — — — — — — — — — 2.7% T1 — — — — — — — — — 6.0% A1 — — — — — — — — — — — — 0.28 0.1 TX 0.18 TX MWAS Control (no recycling agent) — — — — — — — — — — — — — 2.7% T1 (field) — — — — — — — 7.5% T1 (PGH) — — — — — — — — — — — — 0.4 0.4 TX — Control (no recycling agent) — — — — — — — — — — — — 6% A1 (PGL) — — — — — — — — — — — 0.5 0.25 TX 0.25 TX TOAS Control (no recycling agent) — — — — — — — — — — — 12.5% T1 (PGL) — — — — — 15.5% T1 (PGH) — — — — — — — — — — 17.5% V1 (PGH) — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — NOTE: — = not applicable. aAt RTFO, 20 PAV, and 40 PAV aging. bLong-term oven aging including binder master curve, G-R, Tδ = 45°, and FT-IR. cDescribed in Chapter 2. (continued on next page)

NH 64-28 (ΔTc +1.2) 0.5 0.4 NH 0.1 CA TOAS 9.0% T1 (PGH)d — — — — — — — — — — 9.0% V2 (PGH)d — — — — — — — — — 0.4 TX 0.1 TX TOAS 14.0% T1 (PGH)d — — — — — — — — — 14.0% V2 (PGH)d — — — — — — — — — 0.7 0.7 NH — Control (no recycling agent) — — — — — — — — — — — — — — — 8.0% B1 (PGH)d — — — — — — — — — — NV 64-28P (ΔTc -3.6) — — — Base Binder — — — — — — — — — 2.7% T1 — — — — — — — — — — — 0.5 0.25 TX 0.25 TX TOAS 11% T1 — — — — — — — — — — — — — — 0.33 0.3 NV — Control Blend — — — — — — — — — 2.0% T2 (field) — — — — — — — — — 2.0% A2 (field) — — — — — — — — — 0.33 0.3 NV — 3.5% T2 (PGH) — — — — — — — — — — — — 5.5% A2 (PGH) — — — — — — — — — — — — IN 64-22 (ΔTc -1.2) 0.28 0.1 TX 0.18 TX MWAS Control (no recycling agent) — — — — — — — — — — 2.0% T1 (PGL/ΔTc) — — — — — — — — — — 2.0% A1 (PGL/ΔTc) — — — — — — — — — — 1.0% V1 (PGL/ΔTc) — — — — — — — — — — 1.0% B1 (PGL/ΔTc) — — — — — — — — — — 5.0% T1 (PGH) — — — — — — — — — — — — — — — 6.5% A1 (PGH) — — — — — — — — — — — — — — 3.5% V1 (PGH) — — — — — — — — — — — — — — — 4.0% B1 (PGH) — — — — — — — — — — — — — — — NOTE: Gray shading indicates NV field project material combinations; — = not applicable. aAt RTFO, 20 PAV, and 40 PAV aging. bLong-term oven aging including binder master curve, G-R, Tδ = 45°, and FT-IR. cDescribed in Chapter 2. dEstimated recycling agent dose to match continuous PGH (from blending charts). Materials Combinations Binder Testing Mortar Testing (PG) Mixture Testing Base Binder Recycled Materials Recycling Agent Dose and Type (Dose Selection Methodc) Rheology Chemical Characterization FT-IRa Agingb C I M R |E *| I- F IT B B R m (S liv er ) H W T T A P A U T SS T S- V E C D R B R R A P R A S G -R a T δ = 45 °a SAR-AD, MDSC Table 1. (Continued).

NOTE: Gray shading indicates IN field project material combinations; — = not applicable. aAt RTFO, 20 PAV, and 40 PAV aging. bLong-term oven aging including binder master curve, G-R, Tδ = 45°, and FT-IR. cDescribed in Chapter 2. dEstimated recycling agent dose to match continuous PGH (from blending charts). IN 64-22 (ΔTc -1.2) — — — Base Binder — — — — — — — — — IN 58-28 (ΔTc -8) — — — Base Binder — — — — — — — — — — — — — — — 0.32 0.25 IN 0.07 IN MWAS DOT Control (no recycling agent) — — — — — — — — — 0.42 0.14 IN 0.28 IN MWAS 3.0% T2 (field) — — — — — — — — — 0.42 0.14 IN 0.28 IN MWAS 6.5% T2 (ΔTc) — — — — — — — — — — 0.28 IN 0.14 IN MWAS 8.0% T2 (PGH)d — — — — — — — — — — — 0.5 0.36 IN 0.14 IN MWAS 9.5% T2 (PGH)d — — — — — — — — — — — 0.7 0.7 IN — 10.0% T2 (PGH)d — — — — — — — — — — — MN 58-28 (ΔTc 0) — — — Base Binder — — — — — — — — — — — — — — 0.5 0.25 TX 0.25 TX TOAS 16.5% T1 (PGH) — — — — — — — — — — — — 16.5% V1 (PGH) — — — — — — — — — — — Materials Combinations Binder Testing Mortar Testing (PG) Mixture Testing Base Binder Recycled Materials Recycling Agent Dose and Type (Dose Selection Methodc) Rheology Chemical Characterization FT-IRa Agingb C I M R |E *| I- F IT B B R m (S liv er ) H W T T A P A U T SS T S- V E C D R B R R A P R A S G -R a T δ = 45 °a SAR-AD, MDSC (continued on next page)

NOTE: Gray shading indicates WI and DE field project material combinations; — = not applicable. aAt RTFO, 20 PAV, and 40 PAV aging. bLong-term oven aging including binder master curve, G-R, Tδ = 45°, and FT-IR. cDescribed in Chapter 2. WI 58-28 (ΔTc -3.4) — — — — — — — — — — — — — — — — — — 0.22 0.22 WI — DOT Control (no recycling agent) — — — — — — — 0.31 0.31 WI — Recycled Control (no recycling agent) — — — — — — — — WI 52-34 (ΔTc +0.4) 0.31 0.31 WI — Recycled Control (no recycling agent) — — — — — — — WI 58-28 (ΔTc -3.4) 0.31 0.31 WI — 1.2% V2 (field) — — — — — — 0.31 0.31 WI — 5.5% V2 (PGH) — — — — — — 0.5 0.5 WI — 9% V2 (PGH) — — — — — — — DE 64-28 (ΔTc -0.1) 0.34 0.17 DE 0.17 MWAS DOT Control (no recycling agent) — — — — — — — — — 0.41 0.24 DE 0.17 MWAS 0.8% T2 (field) — — — — — — — — — 0.41 Recycled Control (no recycling agent) — — — — — — — — — — 0.41 0.24 DE 0.17 MWAS 8.5% T2 (PGH) — — — — — — — 0.5 0.33 DE 10% T2 (PGH) — — — — — — — Materials Combinations Binder Testing Mortar Testing (PG) Mixture Testing Base Binder Recycled Materials Recycling Agent Dose and Type (Dose Selection Methodc) Rheology Chemical Characterization FT-IRa Agingb C I M R |E *| I- F IT B B R m (S liv er ) H W T T A P A U T SS T S- V E C D R B R R A P R A S G -R a T δ = 45 °a SAR-AD, MDSC Table 1. (Continued).

Introduction 13 – SAR-AD—saturates, aromatics, resins–asphaltene determinator fractions. – CII—colloidal instability index. – TPA—total pericondensed aromatics. – Tg—glass transition temperature by modulated differential scanning calorimeter (MDSC). – Tg End—high-end temperature of the glass transition by MDSC. – FT-IR—Fourier transform-infrared spectra. • Mixture testing results: – CI—coatability index. – MR—resilient modulus. – G-Rm—mixture Glover-Rowe parameter and Black space analysis. – FI and CRI—flexibility index and cracking resistance index by Illinois Flexibility Index Test (I-FIT). – Sm, m-valuem—creep stiffness and relaxation rate by BBR for mixtures (BBRm) or sliver test. – N12.5—number of load cycles to 12.5-mm rut depth by asphalt pavement analyzer (APA) and Hamburg wheel-tracking test (HWTT). – CRIEnv—environmental cracking resistance index by uniaxial thermal stress and strain test (UTSST). – DR, Nf at G R = 100—average reduction in pseudo stiffness up to failure and the number of load cycles for specific rate of damage accumulation by simplified viscoelastic continuum damage (S-VECD) fatigue test. 1.2 Key Results from Phase 1 The use of RAP/RAS in new HMA and WMA mixtures is a sustainable engineering practice that reduces production and construction costs and protects the environment by conserving natural resources and decreasing energy consumption and emissions. As the percentage of RAP/ RAS increases, these benefits also increase. State DOTs and contractors alike have recognized these benefits, but to ensure engineering benefits are also realized, state DOTs and other highway agencies require recycled mixtures to meet the same mix design and performance standards as mixtures with all virgin materials by specifying maximum RAP/RAS contents and allowing the use of a softer (substitute) base binder (Epps Martin et al. 2015). This study aimed to explore the effectiveness of using recycling agents to partially restore rheology and thus allow for an increase in the allowable RAP/RAS content. This section provides a summary of the following key results identified in the literature review and surveys in Phase 1 and remaining issues that were addressed to some extent in this study: • Separation of RAP and RAS Contributions to RBR: The Phase 2 laboratory experiment designs include specification of an overall RBR and the contribution from RAP and RAS separately as RAPBR and RASBR according to the following equations (NCAT 2014): 100 Equation 2RAPBR Pb P Pb RAP RAP total [ ]= × × [ ]= × ×100 Equation 3RASBR Pb P Pb RAS RAS total where PbRAP = binder content of the RAP, PRAP = percentage of RAP by weight of mix, PbRAS = binder content of the RAS, PRAS = percentage of RAS by weight of mix, and Pbtotal = binder content of the combined mixture.

14 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios • Predominant Use of RAP at High RBRs: Despite the widespread acceptance across state DOTs of the use of recycled materials in HMA and WMA mixtures, survey results reported in the first interim report indicated that most DOTs do not commonly use a high percent- age of RAP (60% use 11%–20%, or approximately 0.1–0.2 RAPBR, and 23% use 21%–30%, or approximately 0.2–0.3 RAPBR) and do not commonly use a high percentage of RAS (65% use 0%–3%, or approximately 0–0.13 RASBR, and 29% use 4%–6%, or approximately 0.17–0.26 RASBR). Surveys also indicated that state DOTs and contractors predominantly use RAP in mixtures with high RBRs since this recycled material is more readily available compared to RAS, but concerns remain with respect to material variability. • Increased Use of Recycling Agents at High RBRs: Recycling agents were used in HMA in the early period of widespread recycling in the 1970s and 1980s toward realizing the environmen- tal, economic, and engineering benefits. Despite this long history of use, survey results indi- cated that more than 80% of state DOTs do not use or do not allow the use of recycling agents in mixtures. According to these DOTs, the main barriers to using recycling agents in recycled mixtures are the lack of experience and, most importantly, the absence of tests and criteria to determine dose and/or to assess the performance of mixtures with recycling agents. These shortcomings become more pronounced as RBRs increase and recycling agents are required to partially restore rheology. Based on the survey, state DOTs considering mixtures with high RBRs are predominantly exploring the use of tall oils as recycling agents. • Characterization of Binder Blend Rheology: With aging, the stiffness of a binder increases and the phase angle decreases. With the addition of recycling agents that partially restore or reju- venate aged binder rheology, and not just soften the material, the aging process is expected to be reversed, with the stiffness decreasing and phase angle increasing. Both of these processes (aging and rejuvenation) are illustrated with the G-R parameter in Black space. Other chemi- cal and rheological parameters provide additional tools for assessing the effectiveness of recy- cling agents in restoring binder blend rheology initially and with aging. FT-IR spectroscopy and determination of carbonyl area (CA) can be used to track oxidative aging of the binder blend. Binder blend master curves, determined through dynamic shear rheometer (DSR) iso- thermal frequency sweeps tests, can also be used to determine other rheological indices, such as crossover temperature (Td = 45°). • Representative Characterization of Binder Blends: To characterize the aged, recycled binder in RAP/RAS and to quantify the effect of blending these recycled binders with a base binder, extraction and recovery are required. However, this process alters the recycled binder proper- ties due to incomplete extraction, remaining solvent, possible binder-solvent reaction, and binder aging due to high temperatures during the process. The mortar procedure defined by the latest draft of AASHTO T XXX-12 Estimating Effect of RAP and RAS on Blended Binder Performance Grade without Binder Extraction (www.arc.unr.edu/Outreach.html) provides a more representative method for characterizing the effect of the aged, recycled binders on a base binder with or without recycling agents as compared to the binder blend with complete blend- ing after extraction and recovery. These mortar results agree with mixture results and field performance in terms of low-temperature cracking when recovered PG binder grades do not. • Primary Concern of Mixture Cracking Resistance: As expected, survey results indicated that rutting resistance of mixtures with high RBRs is not a concern unless higher recycling agent doses are used. Of greater concern in recycled mixtures is fatigue, reflective, and thermal cracking since cracking resistance decreases with aging, and mixtures with high RBRs are expected to have lower cracking resistance due to their aged, stiff, and brittle binders. Most existing models/tests to predict crack growth in mixtures are generally empirical or phenom- enological in nature and include indirect tensile (IDT) strength, thermal stress restrained specimen test (TSRST), beam fatigue, and overlay test (OT). More recent mechanistic-based approaches and associated tests, such as the UTSST, the S-VECD approach, and the energy- based mechanistic (EBM) method, provide improved characterization tools for evaluating cracking resistance of mixtures with high RBRs. These approaches can be used along with the

Introduction 15 semicircular bend (SCB) test recommended by NCHRP Project 9–57, “Experimental Design for Field Validation of Laboratory Tests to Access Cracking Resistance of Asphalt Mixtures” (Zhou and Newcomb 2015) after long-term oven aging (LTOA) to evaluate the effectiveness of recycling agents in improving cracking resistance for mixtures with high RBRs. In addi- tion, differences in laboratory specimen fabrication and field production and construction of mixtures with recycling agents must be considered with respect to short-term laboratory aging protocols for use in mix design and quality-assurance testing. • Evolution of Recycling-Agent Effectiveness: Although many studies have shown the effect of recycling agents in improving the cracking resistance of mixtures with high RBRs, the surveys indicated that state DOTs remain concerned with the evolution of recycling-agent effective- ness with aging and the resulting long-term performance of these mixtures. Binder and mor- tar rheology and mixture stiffness and cracking resistance results after laboratory LTOA can be subsequently tied to field project locations in terms of climate and construction date, which play a role in the blending of the binder components through diffusion. This is one approach to evaluating long-term performance of mixtures with high RBRs and recycling agents. A companion approach that predicts long-term performance uses a computational pavement oxidation model that is based on fundamentals of heat and mass transfer together with mea- sured binder oxidation kinetics and rheological hardening properties to provide changes in binder rheology as a function of time and depth below the surface. The model is founded on local climate and weather data as well as parameters for the specific binder used in the pavement. For sufficiently oxidized binder blends, data suggest that the model also provides meaningful durability calculations for polymer-modified binders. This model provides a tool to capture the unknown effect of recycling agents on binder oxidation kinetics and resulting evolution of recycling agent effectiveness with aging. 1.3 Recent Relevant Literature A comprehensive literature review was presented in the first interim report (Epps Martin et al. 2015), and a list of recent relevant literature organized by discussion area was provided in the revised second interim report (Epps Martin et al. 2017). Table 2, Table 3, Table 4, Table 5, and Table 6 provide a summary of the literature used throughout the study that indicates the dependency of recycling-agent effectiveness with short- and long-term aging on its type and dose and on the type of recycled materials (RAP versus RAS). 1.4 Scope of Final Report Chapter 1 of this final report begins with a brief history of the use of recycled materials in HMA and WMA mixtures, the construction and performance challenges associated with using high percentages of these materials through high RBRs, and the use of recycling agents to overcome these challenges. An NCHRP 09–58 overview with objectives and a summary of key results from Phase 1 and recent relevant literature is also provided, followed by the scope of this final report, which includes a list of associated publications to date. Next, the field projects and associated materials and selected laboratory tests and specimen fabrication protocols are introduced. Chapter 2 summarizes the key results completed in Phase 2, including the following: • Development of recycling agent dose selection method, • Chemical compatibility of binder blends, • Rheological balance of binder blends, • Representative binder blending, • Mixture cracking resistance by S-VECD, and • Comparison of specimen types.

16 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios Author(s) and Year Laboratory Test(s) Main Findings Shen and Ohne (2002) PG and penetration Nonlinear reduction in PGH with increased recycling-agent dose, but linear reduction in PGL. Shen et al. (2007) PG Linear reduction in PGH and PGL of RAP binder with increased recycling-agent dose. Tao et al. (2010) PG and penetration Linear reduction in PGH and PGL, and linear increase in penetration, with increased recycling-agent dose. Linear increase in ductility with increased recycling-agent dose, up to 10% dose. Beyond 10%, a slower increase in ductility with recycling-agent dose. A specific type of recycling agent is required to rejuvenate aged polymer-modified binders to ensure long-term durability. Tran et al. (2012) PG Linear reduction in PGH and PGL of RAP/RAS binders with increased recycling-agent dose. Oliveira et al. (2013) DSR frequency sweep and penetration Linear reduction in penetration at 25°C of RAP binder with increased recycling-agent dose. Recycling-agent addition decreased |G*| and increased δ in DSR test. Zaumanis et al. (2014) PG and penetration Linear reduction in PGH and PGL of RAP binder with increased recycling-agent dose, but nonlinear reduction in intermediate-temperature PG (PGI). Organic products (waste vegetable oil and distilled tall oil) require much lower doses compared to petroleum products (aromatic extract and waste engine oil) to deliver the same effect on PG. Yu et al. (2014) DSR frequency sweep and PG Recycling-agent addition decreased |G*| and increased δ in DSR and decreased S and increased m-value in BBR, depending on aged binder source and recycling-agent type. Waste vegetable oil is much more effective than aromatic extract. Ali (2015) PG and rotational viscometer Linear reduction in PGH and viscosity with increased recycling-agent dose. Bio-based oil and petroleum distillate products require lower doses compared to other products to deliver the same effect on PGH and viscosity. Zhou et al. (2015) PG Linear reduction in PGH and PGL of RAP/RAS binders with increased recycling-agent dose, only when the dose is 20% or less. - agent dose. Mohammadafzali et al. (2015) PG Petroleum-based and bio-based recycling agents accelerated aging in the blends, while a paraffinic-based recycling agent slowed down aging. Bio-based oils increased aging the most. Alavi and He et al. (2015) dynamic shear modulus The addition of a petroleum-based recycling agent decreased the stiffness of binder blends with 25% and 40% RAP and 15% RAS for five different base binders. Pradyumna and Jain (2016) Marshall stability, tensile strength ratio, and resilient modulus Rejuvenated mixtures that contain recycling agents with higher colloidal instability index (CII) had better moisture susceptibility and load spreading properties than those containing recycling agents with lower CII. Nayak and Sahoo (2016) DSR frequency sweep The plant-based oil performed better in terms of fatigue resistance then the naphthenic aromatic recycling agent, while this was reversed for rutting performance. Karki and Zhou (2016) DSR frequency sweep and PG Linear reduction in PGH and PGL of RAP/RAS binders with increased recycling agent dose. Higher dose is required to restore PGH than PGL. Recycling-agent addition decreased |G*| and increased δ depending on the dose. Osmari et al. (2017) rotational viscometer and DSR frequency sweep Petroleum-based recycling agents require higher doses than waste cooking oil and castor oil to deliver the same effect on viscosity. Waste cooking oil and castor oil had more impact in reducing |G*| than petroleum-based recycling agent. Tabatabaee and Kurth (2017) SARA Fractionation Blends with vegetable oils have a lower CII than blends with aromatic extracts. Beyond 20%, a nonlinear decrease in PGH with recycling Table 2. Previous research on the effect of recycling agents on rejuvenated binder blends.

Author(s) and Year Laboratory Test(s) Main Findings Mallick et al. (2010) dynamic modulus Recycling-agent addition dropped the stiffness of 100% RAP mixture at high loading frequencies (5 Hz and 10 Hz) but increased the stiffness at lower loading frequencies (1 Hz and 0.1 Hz) at the highest testing temperature (54.4°C). Uzarowski et al. (2010) dynamic modulus Recycling-agent addition significantly dropped the stiffness of rejuvenated mixtures. O’Sullivan (2011) dynamic modulus Recycling-agent addition decreased the stiffness of 80%, 90%, and 100% RAP mixtures to below the stiffness of the virgin mixture. Tran et al. (2012) dynamic modulus After short-term aging, recycling-agent addition dropped the stiffness of the rejuvenated RAP/RAS mixtures closer to that of the virgin mixture. After long-term aging, rejuvenated mixtures with recycling agent appeared to age faster than the RAP/RAS mixtures without recycling agent. Mogawer et al. (2013) dynamic modulus Recycling agent addition dropped the stiffness of rejuvenated mixtures closer to that of the virgin mixture. Rejuvenated mixtures with RAS and RAP/RAS showed less significant reduction in stiffness after incorporating the recycling agent, as compared to RAP-only rejuvenated mixtures. Im et al. (2014) dynamic modulus Recycling-agent addition dropped the stiffness of the rejuvenated mixtures at high testing temperature (40°C) and low frequency ranges but did not affect the stiffness at lower temperatures (4°C and 20°C). Alavi and He et al. (2015) dynamic shear modulus The addition of a petroleum-based recycling agent decreased the stiffness of fine aggregate mixtures with 25% and 40% RAP and 15% RAS for five different base binders. Haghshenas et al. (2016) dynamic modulus The petroleum-based recycling agent had a greater impact in reducing |E*| than soybean oil and tall oil. Table 3. Previous research on the effect of recycling agents on stiffness of recycled asphalt mixtures. Author(s) and Year Laboratory Test(s) Main Findings Lin et al. (2011) IDT Recycling-agent addition improved cracking resistance, depending on recycling-agent type. Tran et al. (2012) energy ratio test Recycling-agent addition improved fracture properties of rejuvenated mixtures. Texas OT Recycling-agent addition increased the average number of cycles to failure. Mogawer et al. (2013) OT Recycling-agent addition improved the cracking performance of rejuvenated RAP/RAS mixtures, depending on recycling-agent type. Yan et al. (2014) four-point bending Recycling-agent addition highly improved the fatigue cracking resistance of rejuvenated 30%, 40%, and 50% RAP mixtures, depending on recycling-agent type. Rejuvenated mixtures containing recycling agents with higher CII had better fatigue cracking resistance. Im et al. (2014) OT Recycling-agent addition increased the average OT number of cycles to failure, from approximately 110% to 300%, depending on recycling-agent type. Cooper et al. (2015) SCB Mixtures rejuvenated with napthenic oil exhibited better fracture resistance at intermediate temperature than those rejuvenated with vegetable oil. Ali (2015) OT Recycling-agent addition increased the average OT number of cycles to failure as compared to the virgin and 100% RAP mixture. Nabizadeh et al. (2017) I-FIT Recycling-agent addition increased the FI. Aromatic extract was more effective than tall oil and soybean oil. Espinoza- Luque et al. (2018) I-FIT Recycling-agent addition increased the FI, depending on recycling-agent dose. Cooper et al. (2015) SCB Recycling agent adversely affected the fracture resistance of the rejuvenated mixtures with RAS. Table 4. Previous research on the effect of recycling agents on intermediate-temperature cracking resistance of recycled asphalt mixtures.

Table 6. Previous research on the effect of recycling agents on rutting resistance and moisture susceptibility of recycled asphalt mixtures. Author(s) and Year Laboratory Test(s) Main Findings Shen et al. (2004) dynamic stability test (wheel-tracking rut test) Recycling-agent addition significantly decreased the dynamic stability by a range of 20% to 50% depending on recycling-agent dose (2% to 7.4%). Shen et al. (2007) APA Recycling-agent addition and the use of a softer binder decreased the rut depth, but the rut depths of rejuvenated mixtures with recycling agent were smaller than those using a softer base binder. tensile strength ratio (TSR) Recycling-agent addition and softer binder usage did not affect moisture susceptibility compared to the virgin mixture. Uzarowski et al. (2010) APA Recycling-agent addition significantly dropped the rutting resistance of rejuvenated mixtures. Lin et al. (2011) Marshall stability With increasing recycling agent dose from 10% to 40%, the reduction in rejuvenated mixture stability ranged from 25% to 55% depending on recycling- agent type. Tran et al. (2012) APA Recycling-agent addition increased mixture susceptibility to rutting, but with rut depths less than 5.5 mm to withstand at least 10 million equivalent single axle loads. TSR Recycling-agent addition did not negatively affect the TSR value. Mogawer et al. (2013) Hamburg wheel- tracking device (HWTD) Recycling-agent addition increased mixture susceptibility to rutting and moisture damage in rejuvenated RAP/RAS mixtures. Yan et al. (2014) Marshall stability, wheel-tracking rut test Rejuvenated mixtures containing recycling agents with higher CII had better rutting resistance. Im et al. (2014) HWTD Rutting and moisture susceptibility of rejuvenated mixtures with RAP/RAS and recycling agents depend on the type and dose of recycling agent. Espinoza- Luque et al. (2018) HWTD Recycling-agent addition increased mixture susceptibility to rutting, depending on recycling-agent dose. Cooper et al. (2015) HWTD Recycling-agent addition did not negatively affect the rutting or moisture susceptibility of the rejuvenated mixtures with RAS. Author(s) and Year Laboratory Test(s) Main Findings Shen et al. (2004) TSRST Recycling-agent addition significantly improved low- temperature fracture properties, depending on recycling-agent type. Tran et al. (2012) IDT Recycling-agent addition reduced the critical failure temperature of rejuvenated RAP/RAS mixtures. Mogawer et al. (2013) TSRST Recycling-agent addition considerably improved the low-temperature cracking resistance of rejuvenated RAP/RAS mixtures. Zaumanis et al. (2013) IDT creep compliance Recycling-agent addition increased the low- temperature creep compliance (and thus reduced low- temperature cracking potential) of rejuvenated 100% RAP mixture. IDT Recycling-agent addition increased indirect tensile strength and fracture energy, depending on recycling- agent type. Hajj and Souliman et al. (2013) TSRST Recycling-agent addition improved the low- temperature cracking resistance with a decrease in the fracture stress and microcracking. Yan et al. (2014) three-point bending Recycling-agent addition improved the low- temperature cracking resistance, depending on recycling-agent type. Cooper et al. (2015) TSRST Recycling agent adversely affected the low- temperature performance of the rejuvenated mixtures with RAS. Table 5. Previous research on the effect of recycling agents on low-temperature cracking resistance of recycled asphalt mixtures.

Introduction 19 Chapter 3 presents a summary of field performance of high RBR mixtures in the field projects and a comparison of field and corresponding laboratory performance toward development of thresholds for cracking resistance. Chapter 4 provides more detailed laboratory performance results of high RBR binder blends and associated mixtures in terms of the following challenges associated with the evaluation of the effectiveness of recycling agents in high RBR binder blends initially and with aging: • Binder blend rheology with aging, • Binder blend aging prediction, • Recycling-agent characterization, • Mixture performance, and • Recycled binder availability. Chapter 5 describes the following practical tools developed in this study and incorpo- rated into a draft AASHTO standard practice to facilitate the evaluation of the effectiveness of recycling agents in high RBR binder blends and corresponding mixtures initially and with aging: • Component materials selection guidelines, • Recycling-agent dose selection method and materials proportioning, • Binder blend rheological evaluation tools, • Mixture performance evaluation tools, and • Recycled binder availability factor. A discussion of laboratory aging and climate effects is also provided. Chapter 6 provides conclusions and suggested research and implementation activities for other considerations outside the scope of this study to conclude this final report. Construction reports for the five field projects are presented as Appendices A through E, and Appendices F through I provide additional data on binder blend aging prediction, recycling-agent characterization, an economic analysis of the use of RAP in asphalt mix- tures, and a draft AASHTO standard practice to increase RBR in asphalt mixtures by using recycling agents. During this study, the following papers were published and provide additional details on specific topics presented in this final report and other collaborative efforts: • Kaseer, F., E. Arámbula-Mercado, L. Garcia Cucalon, and A. Epps Martin (2018a) “Perfor- mance of Asphalt Mixtures with High Recycled Materials Content and Recycling Agents.” International Journal of Pavement Engineering. • Garcia Cucalon, L., F. Kaseer, E. Arámbula-Mercado, A. Epps Martin, N. Morian, S. Pournoman, and E. Y. Hajj (2018) “The Crossover Temperature: Significance and Application towards Engineering Balanced Recycled Binder Blends.” Road Materials and Pavement Design, https:// doi.org/10.1080/14680629.2018.1447504. • Oshone, M., J. Sias Daniel, E. Dave, R. Rastegar, F. Kaseer, and A. Epps Martin (2018) “Explor- ing Master-Curve Based Parameters to Distinguish between Mix Variables,” Proceedings of the International Society for Asphalt Pavements (ISAP) 13th Conference on Asphalt Pavements, Fortaleza, Brazil, June 19–21. • Arámbula-Mercado, E., A. Epps Martin, and F. Kaseer (2018a) “Case Study on Balancing Mixtures with High Recycled Materials Contents,” Proceedings of the International Confer- ence on Advances in Materials and Pavement Performance Prediction (AM3P), Doha, Qatar, April 16–18.

20 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios • Kaseer, F., L. Garcia Cucalon, E. Arámbula-Mercado, A. Epps Martin, and J. Epps (2018b) “Practical Tools for Optimizing Recycled Materials Content and Recycling Agent Dosage for Improved Short- and Long-Term Performance of Rejuvenated Binder Blends and Mixtures,” Journal of the Association of Asphalt Paving Technologists 87. • Morian, N., E. Y. Hajj, S. Pournoman, J. Habbouche, and D. Batioja-Alvarez (2018) “Low Temperature Behavior of Asphalt Binders, Mortars, and Mixtures with High Recycled Materials Content,” Journal of the Association of Asphalt Paving Technologists 87. • Pournoman, S., E. Y. Hajj, N. Morian, and A. Epps Martin (2018) “Impact of Recycled Materials and Recycling Agents on Asphalt Binder Oxidative Aging Predictions,” Transportation Research Record. • Menapace, I., L. Garcia Cucalon, F. Kaseer, E. Masad, and A. Epps Martin (2018a) “Appli- cation of Low Field Nuclear Magnetic Resonance to Evaluate Asphalt Binder Viscosity in Recycled Mixes,” Construction and Building Materials 170, 725–736, https://doi.org/10.1016/ j.conbuildmat.2018.03.114. • Kaseer, F., F. Yin, E. Arámbula-Mercado, A. Epps Martin, J. Daniel, and S. Salari (2018c) “Development of an Index to Evaluate the Cracking Potential of Asphalt Mixtures Using the Semi-Circular Bending Test Construction & Building Materials,” Construction and Building Materials 167, 286–298, https://doi.org/10.1016/j.conbuildmat.2018.02.014. • Arámbula-Mercado, E., F. Kaseer, A. Epps Martin, F. Yin, and L. Garcia Cucalon (2018b) “Evaluation of Recycling Agent Dosage Selection and Incorporation Methods for Asphalt Mixtures with High RAP and RAS Contents,” Construction and Building Materials 158, 432–442, https://doi.org/10.1016/j.conbuildmat.2017.10.024. • Menapace, I., L. Garcia Cucalon, F. Kaseer, E. Arámbula-Mercado, A. Epps Martin, E. Masad, and G. King (2018b) “Effect of Recycling Agents in Recycled Asphalt Binders Observed with Microstructural and Rheological Tests,” Construction and Building Materials 158, 61–74, https://doi.org/10.1016/j.conbuildmat.2017.10.017. • Garcia Cucalon, L., G. King, F. Kaseer, E. Arámbula-Mercado, A. Epps Martin, T. F. Turner, and C. J. Glover (2017) “Compatibility of Recycled Binder Blends with Recycling Agents: Rheological and Physicochemical Evaluation of Rejuvenation and Aging Processes,” Indus- trial Engineering and Chemistry Research 56 (29), 8375–8384, https://doi.org/10.1021/ acs.iecr.7b01657. • Kaseer, F., F. Yin, E. Arámbula-Mercado, and A. Epps Martin (2017a) “Stiffness Character- ization of Asphalt Mixtures with High RAP/RAS Contents and Recycling Agents,” Trans- portation Research Record 2633, 58–68, http://dx.doi.org/10.3141/2633-08. • Yin, F., F. Kaseer, E. Arámbula-Mercado, and A. Epps Martin (2017) “Characterizing the Long-Term Rejuvenating Effectiveness of Recycling Agents on Asphalt Blends and Mix- tures with High RAP and RAS Contents,” Road Materials and Pavement Design 18 (Sup 4), 273–292, http://dx.doi.org/10.1080/14680629.2017.1389074. • Carvajal Munoz, J. S., F. Kaseer, E. Arámbula, and A. Epps Martin (2015) “Use of the Resil- ient Modulus Test to Characterize Asphalt Mixtures with Recycled Materials and Recy- cling Agents,” Transportation Research Record 2506, 45–53, http://dx.doi.org/10.3141/ 2506-05. 1.5 Experiment Design This section provides the experiment design in terms of field projects and associated materials and of selected laboratory tests and specimen fabrication protocols. Different materials and testing combinations were used for each issue explored and each tool developed, as described in Chapter 5 based on the results presented in Chapter 2, Chapter 3, and Chapter 4.

Introduction 21 1.5.1 Field Projects and Materials In selecting field projects for use in the Phase 2 laboratory experiments and the Phase 3 field experiments, consideration was given to obtaining a range in each of the following factors to make the conclusions of this study as comprehensive as possible: • Recycling agents by category, as defined in Table 7, for comparison of types; • RAPBR and RASBR; • Environmental zone, as defined by the SHRP Long-Term Pavement Performance Program (SHRP-LTPP) and shown in Figure 2; and • Traffic volume. Since the laboratory experiments were tied to field projects to facilitate Phase 3, selection of a field project in a specific environment simultaneously resulted in selection of the materials (aggregate; base binder; recycled materials; and any additives, including recycling agents) based on the materials selected by the respective DOT. Other eligibility requirements for field proj- ects besides location on a highway, arterial, or collector facility in North America included the following: • A high RBR between 0.3 and 0.5; • A virgin test section (with no recycled materials) if possible; • A DOT control test section (with recycled materials but without recycling agents) at the maxi- mum allowed by the DOT without recycling agents; • Multiple recycling agents if possible; and • A minimum number of WMA, anti-stripping, and other additives. Construction of a new asphalt mixture layer was required for this study because it provided the only opportunity for fabrication of reheated plant-mixed, laboratory-compacted (RPMLC) specimens that capture field blending of base binders, recycled materials, and recycling agents and a starting point for tracking performance of cores to validate laboratory aging protocols critical to evaluating mixture cracking resistance and its evolution with aging. Category Types Description Paraffinic oils Waste engine oil Waste engine oil bottoms Valero VP 165® Storbit® Refined used lubricating oils. Aromatic extracts Hydrolene® Reclamite® Cyclogen L® ValAro 130A® Refined crude oil products with polar aromatic oil components. Napthenic oils SonneWarmix RJ™ Ergon HyPrene® Engineered hydrocarbons for asphalt modification. Triglycerides and fatty acids Waste vegetable oil Waste vegetable grease Brown grease Oleic acid Derived from vegetable oils. Tall oils Sylvaroad™ RP1000 Hydrogreen® Paper industry by-products. Same chemical family as liquid antistrip agents and emulsifiers. Table 7. Recycling-agent categories and types (Willis and Tran 2015).

22 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios Table 8 through Table 12 provide details for the test sections in the field projects constructed in TX, NV, IN, WI, and DE (Figure 2), respectively, which include those used in Phase 2 from TX, NV, and IN and in Phase 3 from WI and DE. Construction reports for the TX, NV, IN, WI, and DE field projects are presented in Appendices A through E, respectively. Each of these field projects provided the following key elements that facilitated the development of evaluation tools for assessing the effectiveness of recycling agents initially and with aging for binder blends and corresponding mixtures with high RBRs in different environmental zones: • TX—constructed early in study, poor softer (substitute) binder; • IN—high RASBR, high RBR, poor softer (substitute) binder; • NV—polymer-modified binder [no softer (substitute) binder], two recycling agents; • WI—engineered recycling agents; and • DE—WMA additive, high RBR, no softer (substitute) binder. Field activities for the constructed field projects included gathering component materials (virgin aggregate; base binder; recycled materials; and any additives, including recycling agents) and plant mix for fabrication of laboratory-mixed, laboratory-compacted (LMLC) and RPMLC specimens, respectively, and procuring cores at construction and after approximately 1 year to verify specimen fabrication and aging protocols, validate relationships between binder and mixture properties, and evaluate the effectiveness of recycling agents with aging. A third set of pavement cores for the TX, NV, and IN field projects after approximately 2 years were procured in Phase 3. All of these materials were available through cooperation with the state DOTs, con- tractors, and state asphalt paving associations. A general field-performance assessment by visual survey was completed at each coring period in cooperation with the associated DOTs, and a summary is provided in Chapter 3. Figure 2. SHRP-LTPP environmental zones and constructed field projects.

Introduction 23 Mixture Type/Test Section Virgin DOT Control (0.28 RBR) +0.5% WMA Rejuvenated (0.28 RBR) +2.7% T1 Binder PG 70-22P 64-22 64-22 Binder contenta 4.9% 4.9% 4.9% RAP/RAS content — 10% RAP/ 5% MWAS 10% RAP/ 5% MWAS RBR — 0.28 (0.1 RAP + 0.18 RAS) 0.28 (0.1 RAP + 0.18 RAS) Recycling-agent type and doseb — — 2.7% T1 WMA doseb — 0.5 — NOTE: — = not applicable. aTotal binder in the mixture (virgin/base + recycled). bBy percentage of total binder in the mixture. Table 8. Mixture characteristics for the TX field project. Mixture Type/Test Section Virgin DOT Control (0.32 RBR) Rejuvenated (0.42 RBR) +3% T2 Binder PG 64-22 58-28 58-28 Binder contenta 5.9% 5.8% 5.8% RAP/RAS content — 28% RAP/ 2% MWAS 16% RAP/ 8% MWAS RBR — 0.32 (0.25 RAP + 0.07 RAS) 0.42 (0.14 RAP + 0.28 RAS) Recycling Agent Type and Doseb — — 3% T2 NOTE: — = not applicable. aTotal binder in the mixture (virgin/base + recycled). bBy percentage of total binder in the mixture. Table 9. Mixture characteristics for the IN field project. Mixture Type/Test Section Virgin DOT Control (0.15 RBR) Recycled Control (0.33 RBR) Rejuvenated (0.33 RBR) +2% T2 Rejuvenated (0.33 RBR) +2% A2 Binder PG 64-28P 64-28P 64-28P 64-28P 64-28P Binder contenta 5.37% 5.04% 4.6% 4.5% 4.6% RAP content — 15% 33% 33% 33% RBR — 0.15 0.33 0.33 0.33 Recycling-agent type and doseb — — — 2% T2 2% A2 NOTE: — = not applicable. aTotal binder in the mixture (virgin/base + recycled). bBy percentage of total binder in the mixture. Table 10. Mixture characteristics for the NV field project.

24 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios For the laboratory experiments, materials were selected from those used in the field projects and expanded with additional base binders from NH and MN; RAP from NH; TOAS from TX and CA; and A1, V1, V3, B1, B2, and P recycling agents as shown in Table 1 and summarized as follows: • Phase 2A: – High 0.3 to 0.5 RBRs with all RAP and RAP/RAS combinations with equivalent RAPBR and RASBR. – Traditional aromatic (A1, A2) and greener alternative tall oil (T1, T2) recycling agents. – TX field project materials (expanded with additional base binders, recycled materials, and recycling agents): TX PG 70–22P, TX PG 64–22, NH PG 64–28, and NV PG 64–28P base binders; TX RAP, TX MWAS, and TX TOAS; T1 and A1. – NV field project materials: NV PG 64–28P base binder, NV RAP, and T2 and A2 for binder and mortar experiments only. • Phase 2B: – High 0.3 to 0.5 RBRs with all RAP and balanced RAP/RAS combinations based on PGH. – Improved and softer IN PG 64–22, MN PG 58–28, and IN PG 58–28 base binders. – Aromatic extract (A2), tall oil (T2), vegetable oil (V1), modified vegetable oil (V2), and reacted bio-based oil (B1) recycling agents. Mixture Type/Test Section DOT Control (0.22 RBR) (PG 58-28) Recycled Control (0.31 RBR) (PG 58-28) Recycled (0.31 RBR) (PG 52-34) Rejuvenated (0.31 RBR) (PG 58-28) +1.2% V2 Binder PG 58-28 58-28 52-34 58-28 Binder contenta 5.6% 5.4% 5.4% 5.4% RAP content 27% 36% 36% 36% RBR 0.22 0.31 0.31 0.31 Recycling- agent type and doseb — — — 1.2% V2 NOTE: — = not applicable. aTotal binder in the mixture (virgin/base + recycled). bBy percentage of total binder in the mixture. Table 11. Mixture characteristics for the WI field project. Mixture Type/Test Section DOT Control (0.33 RBR) +0.4% WMA Rejuvenated (0.41 RBR) +0.8% T2 Rejuvenated (0.41 RBR) +0.8% T2 +0.25% WMA Binder PG 64-28 64-28 64-28 Binder contenta 5.4% 5.4% 5.4% RAP/RAS content 20% RAP/ 4% MWAS 29% RAP/ 4% MWAS 29% RAP/ 4% MWAS RBR 0.34 (0.17 RAP + 0.17 RAS) 0.41 (0.24 RAP + 0.17 RAS) 0.41 (0.24 RAP + 0.17 RAS) Recycling- agent type and doseb — 0.8% T2 0.8% T2 WMA doseb 0.4% — 0.25% NOTE: — = not applicable. aTotal binder in the mixture (virgin/base + recycled). bBy percentage of total binder in the mixture. Table 12. Mixture characteristics for the DE field project.

Introduction 25 – NV field project materials: NV PG 64–28P base binder, NV RAP, and T2 and A2. – IN field project materials: IN PG 64–22 and IN PG 58–28 base binders, IN RAP and IN MWAS, and T2. • Phase 3: – High 0.3 to 0.5 RBRs with all RAP and balanced RAP/RAS combinations based on PGH. – Virgin WI mixture. – Improved and softer WI PG 52–34 base binders. – Aromatic extract (A1), tall oils (T1 and T2), modified vegetable oils (V2 and V3), reacted bio-based oils (B1 and B2), and paraffinic oil (P) recycling agents. – WI field project materials: WI PG 58–28 and WI PG 52–34 base binders, WI RAP, and V2. – DE field project materials: DE PG 64–28 base binder, DE RAP and DE MWAS, and T2. 1.5.2 Laboratory Tests and Specimen Fabrication Protocols The laboratory parameters and tests shown in Table 13, Figure 3, and Figure 4 were selected based on the results from Phase 1 and a continuous review of the literature. For all tests, a mini- mum of two replicate specimens was utilized with at least three replicates for MR/FI testing. Air voids (AVs) for all mixture specimens were determined by AASHTO T 166. Mixture specimens included LMLC specimens, RPMLC specimens, and field cores extracted immediately after con- struction and approximately 1 year, 2 years, and 3 years after construction. Prior to mixture performance testing, specimen fabrication protocols were established for use in the laboratory to address recycling agent incorporation in binders for mixtures and aging. These protocols are described subsequently, followed by additional details for those tests without standards. 1.5.2.1 Recycling-Agent Incorporation Protocol The most common practice for incorporating recycling agents in mixtures is to follow the producer recommendation for dose and proportion of the recycling agent with respect to the base binder. In most cases, when the recycling-agent dose by weight of total binder is 2.0% or lower, the recycling agent is added to the mixture without modifying the amount of base binder (i.e., by addition), whereas when the recycling-agent dose is more than 2.0%, the base binder content is reduced by the recycling-agent amount (i.e., by replacement). In this study, the replacement practice led to incomplete aggregate coating by the binder in mixtures containing RAS at recycling agent doses as low as 5.5%. Thus, for two mixtures at high recycling-agent doses, three recycling-agent incorporation protocols that ranged from 100% replacement to 100% addition were evaluated in terms of aggregate coatability using a modified water absorption method developed in NCHRP Proj- ect 09–53, “Properties of Foamed Asphalt for Warm Mix Asphalt Applications” (Newcomb et al. 2015b). This method is based on the assumption that a completely coated aggregate sub- merged in water for a short period of 1 h cannot absorb water because water cannot penetrate through the binder film covering the aggregate surface. Conversely, a partially coated aggregate is expected to have detectable water absorption because water can penetrate and be absorbed by the uncoated portions of the particle. The resulting CI is calculated as the relative difference in saturated surface dry (SSD) water absorption for the uncoated and coated coarse aggregate fraction (larger than 9.5 mm). Larger CI values indicate better aggregate coating. Figure 5 presents the CI results for a 0.4 RBR mixture with PG 64–22 base binder, TX RAP, and 9.5% A1, and a 0.5 RBR mixture with PG 64–28 base binder, TX RAP at 0.25 RAPBR, TX TOAS at 0.25 RASBR, and 12.5% T1. For the 0.4 RBR mixture, the CI values remained at 100% even after replacing the base binder by half the recycling-agent amount, and above 95% for replacing the base binder by the full recycling-agent amount. However, for the 0.5 RBR mixture, the CI value decreased significantly, especially when replacing the base binder by the

26 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios Performance Issue Binder Parameter and Test Mortar Parameter and Test Mixture Parameter and Test Recycling-agent dose selection PGH after S Aging, PGL after L Aging per AASHTO T 315, T 313, and M 320 — — Rheological balance and effectiveness evolution with aging PGH after S Aging per AASHTO T 315 & M 320 ∆Tc @ Tlow after L Aging per AASHTO PP 78 G-R & Tδ=45° @ Tint with Aging by DSR Master Curve per AASHTO T 315 CA Growth by FT-IR with Aging Oxidation Kinetics and G-R/CAg HS by FT-IR and DSR Master Curve ∆Tc @ Tlow after L Aging per AASHTO PP 78 MR @ Tint after S, L Aging per ASTM D7369 Externally across Diameter |E*|, φ@ Thigh, Tint and Tlow after S, L Aging per AASHTO T 342 G-Rm @ 20°C, 5 Hz Rutting resistance and balanced mixture PGH after S Aging per AASHTO T 315 and M 320 PGH after S Aging per Draft AASHTO N12.5 by HWTT^ & APA Jr^ @ Thigh after S Aging per AASHTO T 324 Fatigue cracking resistance G-R @ Tint with Aging by DSR Master Curve per AASHTO T 315 PGI after L Aging per Draft AASHTO FI & CRI by I-FIT @ Tint after S and L Aging per AASHTO TP 124 DR & Nf@GR = 100 by S-VECD and |E*| @ Tint after L Aging per AASHTO TP 107 Low-temperature cracking resistance PGL after L Aging per AASHTO T 313 and M 320 PGL after L Aging per Draft AASHTO CRIEnv by |E*| and UTSST @ Tlow after L Aging per AASHTO T 342 and Draft AASHTO Sm & m-valuem by BBRm @ Tlow after L Aging per AASHTO TP 125 Chemical compatibility CII and TPA by SAR-AD Tg and Tg End by MDSC CA Growth by FT-IR with Aging — — NOTE: S = short-term aging; L = short- and long-term aging; HS = hardening susceptibility; — = not applicable. ^ For limited number of mixtures. Table 13. Laboratory parameters and tests. full recycling agent amount and thus reducing the base binder content from 4.9% to 4.3% and resulting in a significant number of coarse aggregate particles left visibly uncoated (Figure 6). Based on these limited coatability observations and practicality concerns, the recommended practice for incorporation of recycling agents in mixtures with RAS and more than 5.0% recy- cling agent is addition of the full recycling agent amount (i.e., by addition) with a mandatory requirement to ensure adequate mixture rutting resistance. For mixtures with only RAP or those with RAS and less than or equal to 5.0% recycling agent, the recommended practice for incorporation of recycling agents is reduction of the base binder by the full recycling-agent amount (i.e., by replacement). Additional validation of the recommended 5.0% recycling-agent dose threshold for mixtures with RAS should be completed for mixtures with various optimum binder contents since the amount of total binder in the mixture and other factors, such as binder availability/contribution of the recycled materials and RBR, will likely have an effect on the CI. Additional details are included in Arámbula-Mercado et al. (2018b).

Introduction 27 (a) (b) (c) (d) (e) Figure 3. Binder tests: (a) DSR for PGH, G-R parameter, Tc = 45ç, and G-R/CAg HS; (b) BBR for DTc ; (c) FT-IR for G-R/CAg HS; (d) SAR-AD for CII and TPA (Boysen and Schabron 2015); and (e) MDSC for Tg and Tg End (TA Instruments 2012). (a) (b) (c) (d) (e) (f) (g) (h) Figure 4. Mixture tests: (a) MR; (b) |E*| for G-Rm, S-VECD, UTSST; (c) HWTT for N12.5; (d) APA for N12.5; (e) I-FIT for FI and CRI (Al-Qadi et al. 2015); (f) S-VECD for DR and Nf@GR = 100; (g) UTSST for CRIEnv; and (h) BBRm for Sm and m-valuem.

28 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios Table 14 provides a summary of the specimen fabrication aging protocols and guidelines for recycling-agent blending by addition or replacement. As an example, for a mixture with 30% RAP (PbRAP = 5.0%) and 3% RAS (PbRAS = 18%) or 0.28 RAPBR and 0.1 RASBR, 5.4% total binder content, and a 10 kg batch weight: total binder = 540 g (10,000 × 5.4%) = 150 g RAP binder (540 × 0.28) + 54 g RAS binder (0.1 × 0.1) + 336 g base binder IF recycling-agent dose = 4% (100% replacement): recycling agent = 22 g (540 × 4%) base binder = 314 g (336 g – 22 g) total binder = 540 g IF recycling-agent dosage = 9% (100% addition + ensure adequate mixture rutting resistance): recycling agent = 49 g (540 × 9%) base binder = 312 g (336 – ½ × 49) total binder = 564 g (540 + ½ × 49) Figure 5. Coatability index for 0.4 RBR and 0.5 RBR mixtures. Figure 6. Coatability results for various recycling- agent incorporation methods: virgin aggregate (left), aggregates after replacing the base binder by the full recycling-agent amount (middle), and aggregates with no replacement of the base binder (right).

Introduction 29 1.5.2.2 Aging Protocols Critical or representative aging protocols were used for specimen fabrication for each lab- oratory test across the pavement temperature spectrum, with short-term aged binders (after RTFO) or mixtures (short-term oven aging [STOA]) evaluated for stiffness, cracking resistance at intermediate temperatures, and rutting resistance at high temperatures, and long-term aged binders (RTFO and PAV) or mixtures (STOA and LTOA) evaluated for stiffness and cracking resistance at intermediate and low temperatures. Standard AASHTO T 240 and AASHTO R 28 binder aging protocols in the RTFO and 20-h PAV at 100°C, respectively, were used in the laboratory experiments, along with an extended 40-h PAV aging at 100°C for tracking recy- cling agent effectiveness with aging. Recommended revisions to AASHTO R 30 from NCHRP Project 09–52, “Short-Term Laboratory Conditioning of Asphalt Mixtures” (Newcomb et al. 2018) that include mixture oven-aging protocols of 2 h at 135°C (275°F) on loose mix for STOA prior to compaction and an additional LTOA of 5 days at 85°C (185°F) per AASHTO R 30 for compacted specimens were utilized. This STOA protocol developed and verified in NCHRP Project 09–49, “Performance of WMA Technologies: Stage 1—Moisture Susceptibility” and NCHRP 09–52 by Epps Martin et al. (2014), Yin et al. (2013), Yin et al. (2014a), and Yin et al. (2015) was further verified for mixtures with recycling agents by comparing MR results at 25°C (77°F) for LMLC specimens after laboratory aging with those for cores at construction and after 1 year of field aging, as described in the first interim report (Epps Martin et al. 2015). 1.5.2.3 Laboratory Tests For binders, the influence of recycling agents on rheological balance and recycling agent effectiveness with aging was assessed using standard PGH and PGL temperatures per AASHTO M 320 and other chemical, physicochemical, and rheological properties. Chemical oxidation was tracked using changes in the FT-IR spectrum, and chemical compatibility was evaluated using the SAR-AD and an MDSC. For each aging state, an attenuated total reflectance FT-IR spectrometer was used to collect absorbance data from 600 cm-1 to 4,000 cm-1, and changes in this spectrum were monitored with an emphasis on the carbonyl and sulfoxide regions with peaks at 1,700 cm-1 and 1,032 cm-1, respectively. The SAR-AD technique divides the binder blend by polarity into an expanded set of eight chemical fractions (two saturates, three asphaltenes, two aromatics, and resins) as compared to the traditional four by ASTM D4124 (Boysen and Schabron 2013, 2015). The asphaltene deter- minator separates the asphaltenes into three fractions by solubility using an evaporative light Table 14. Specimen fabrication protocol for preparing high RBR mixtures. Mixing Dry RAP and RAS for 6 h to 8 h at 60°C (140°F) Dry virgin aggregates overnight at mixing temperature Mix RAP and RAS with virgin aggregates Heat base binder and RAP/RAS/aggregate blend at mixing temperature 2 h before mixing Blend recycling agent with base binder using the 100% addition method when recycling agent dose is greater than 5.0% and RAS is used with a mandatory requirement to ensure adequate mixture rutting resistance; otherwise, use the 100% replacement method Heat base binder/recycling-agent blend at mixing temperature for 10 min before mixing Mix RAP/RAS/aggregate blend with base binder/recycling-agent blend Short-term conditioning Condition the loose mix for 2 h at 135°C (275°F) for HMA and WMA with recycling agent or 2 h at 116°C (240°F) for WMA without recycling agent Compaction Compact the loose mix to the target AV at 135°C (275°F) Long-term aging Age the compacted specimens for 5 days at 85°C (185°F) Cutting/Coring Cut/core specimens to final testing dimensions

30 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios scattering (ELS) detector and a variable wavelength absorbance detector set at 500 nm. The least soluble fractions are presumed to represent larger asphaltene agglomerations. The CII and the TPA were calculated as follows, with lower chemical compatibility associated with a larger CII and TPA indicating the total amount of asphaltenes in the binder: [ ]= + + Equation 4CII Saturates Asphaltenes Aromatics Resins [ ]= 500 Equation 5TPA Asphaltenes ELS Asphaltenes nm MDSC testing was conducted by cooling from 165°C (329°F) to −90°C (−130°F) at 0.5°C per minute and then heating again to 165°C (3,295°F) at 0.5°C per minute. Output parameters from the MDSC results included glass transition temperature (Tg) determined by inflection and the high-end temperature of the glass transition (Tg End). In an extensive binder aging evaluation, binder blends were characterized by CA growth rate by FT-IR after standard 20-h PAV aging per AASHTO R 28, for an extended 40-h PAV, and in a forced draft oven at different temperatures and for multiple durations. CA was calculated as the area beneath the FT-IR spectrum from 1,650 cm-1 to 1,820 cm-1 with a baseline from 1,524 cm-1 to 1,820 cm-1. CA growth (CAg) was then determined as the difference between CA at a specific aging state and CA for a reference or tank (CAtank) condition. For long-term aging evaluations in this study, CAtank was defined as after RTFO aging. The binders were also tested in a DSR to determine master curves of shear complex modulus (|G*|) and binder phase angle (d) by conducting isothermal frequency sweeps at different temperatures. These chemical and rheo- logical results were used together to develop and assess the effects of recycling agents on binder oxidation kinetics and resulting HS, which are key inputs for modeling binder aging during the in-service life of a pavement. Rheological indices such as the G-R parameter and crossover temperature (Td = 45°) were also calculated from DSR master curves at intermediate temperatures to evaluate the effectiveness of recycling agents initially and with aging. Td = 45° was determined from DSR master curves at 10 rad/s as the temperature at which the storage modulus (G′) is equal to the loss modulus (G″) and the phase angle is 45 degrees. The G-R parameter was calculated as follows at 15°C and 0.005 rad/s from both DSR properties (|G*| and d) that can be plotted in Black space to assess both the effects of adding aged recycled materials and partially restoring the stiffness and flex- ibility by the inclusion of recycling agents: [ ]( )− = ∗ δ δ Equation 6 2 G R G cos sin These testing conditions were used to tie with inadequate ductility of 5 cm to 3 cm that cor- relates to G-R parameter values between 180 and 600 kPa, respectively, and relates to cracking onset and significant cracking, respectively, in the field (Kandhal 1977). At low temperatures, the difference between the S-controlled and m-controlled PGL grades (DTc) that Anderson et al. (2011) found correlated with the G-R parameter was also determined as an indicator of base binder quality and thus the starting point in evaluating recycling agent effectiveness with aging by G-R in Black space. Finally, the G-R/CAg HS binder parameter combines chemical oxidative aging and its effect on rheology (both stiffness and embrittlement) and is calculated as follows: [ ] [ ] [ ] ( )− = ∆ − ∆ Equation 7G R HS Log G R CAg

Introduction 31 Since some of the selected recycling agents contain large amounts of carbonyl, CAg was moni- tored during aging to define oxidative changes, rather than the more traditional CA, which represents the total carbonyl peak area. For mortars, laboratory testing followed the latest draft of test method AASHTO T XXX-12 Estimating Effect of RAP and RAS on Blended Binder Performance Grade without Binder Extraction (www.arc.unr.edu/Outreach.html). In this procedure, mortar and binder samples are tested in the DSR and BBR by AASHTO T 315 and T 313, respectively, to quantify the effect of blending recycled binder with base binder in terms of continuous PG grade, allowing for an estimation of binder blend properties at critical pavement temperatures with commonly avail- able equipment and without the need for the time-consuming and hazardous binder extrac- tion and recovery process that may impact binder properties. The following three samples are each tested at low, intermediate, and high critical PG temperatures after appropriate or critical aging in the RTFO or RTFO and PAV: • Base binder; • Voidless Mortar A with the same base binder and a single size RAP (or RAS) from a single source; and • Voidless Mortar B with the same base binder, the same total binder content as Mortar A, and recovered aggregate from the same RAP (or RAS) material (using the ignition oven). With known base binder properties, this procedure determines the change in continuous PG grade of the binder blend with the addition of recycled materials. The effect of recycling agents on selected binder properties was also evaluated by adding this component to the base binder and both mortars. In addition to PG, DTc was also determined for mortars. For mixtures, the influence of recycling agents on rheological balance and recycling-agent effectiveness evolution with aging were assessed using standard MR stiffness tests at 25°C (77°F) per ASTM D7369, with linear variable differential transformers (LVDTs) externally attached across the diameter and standard HWTT per AASHTO T 324 and APA Junior tests per AASHTO T 340 at 50°C. The evolution of recycling agent effectiveness in improving cracking resistance of mixtures with high RBRs was evaluated with respect to intermediate-temperature rheology and cracking resistance, respectively, using |E*| at 20°C and 5 Hz per AASHTO T 342 to explore mixture Black space and G-Rm by Equation 8, the standardized S-VECD approach per AASHTO TP 107 with the asphalt mixture performance tester (AMPT), and the I-FIT at 25°C (77°F) per AASHTO TP 124 (Al-Qadi et al. 2015). [ ]( )− = ∗ φ φ Equation 8 2 G R E cos sin m The conditions for determining G-Rm were different from those for the binder for practicality such that master curves are not required, for consistency with a frequency selected close to the inflection point (peak of mixture phase angle master curve), and for analysis purposes to allow for statistical evaluation (Mensching et al. 2015, Mensching et al. 2016a, Mensching et al. 2016b). To overcome the limitations of the FI per AASHTO TP 124, including difficulty in determin- ing the inflection point, moderate variability, and inability to characterize brittle mixtures, an alternative SCB cracking parameter to rank mixture cracking resistance at intermediate tem- peratures was also developed and used in this study using the same I-FIT test procedure and data. The CRI was calculated as follows: [ ]= Equation 9 max CRI G P f

32 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios ∫ [ ]( )= = Equation 10G W A P du A f f where Wf = work of fracture (Joules), P = load (kN), U = load-line displacement (mm), A = ligament area (mm) = the ligament length × the thickness of the specimen, and Pmax = peak load (kN). This alternate parameter includes peak load to differentiate mixtures that may have similar Wf and Gf values but exhibit different behavior in terms of embrittlement. Brittle mixtures with low Gf and/or high Pmax will have lower CRI. Additional details on the development of the CRI are presented in Kaseer and Yin et al. (2018c). The evolution of recycling agent effectiveness in improving cracking resistance of mixtures with high RBRs was preliminarily evaluated with respect to low-temperature cracking using the BBR mixture (BBRm) or sliver test per AASHTO TP 125 to produce a low-temperature mix- ture Black space and the UTSST. This recently developed UTSST test enhanced the traditional TSRST per AASHTO TP 10 that measures only thermal stress under a constant cooling rate in a restrained mixture specimen until fracture. The development of the UTSST methodolo- gies permits the determination of thermo-volumetric (i.e., coefficient of thermal contraction), thermo-viscoelastic (i.e., stiffness–temperature relationship), crack initiation, and fracture properties of asphalt mixtures using thermal stress and thermal strain measurements (ASTM 2018). More detailed information regarding the test setup, sample fabrication, and mixture results can be found in the literature (Alavi et al. 2013; Hajj et al. 2013b; Mensching et al. 2014; Morian 2014). Figure 7(a) presents the layout of the UTSST apparatus. Briefly, the thermal stress and thermal strain measurements are obtained, respectively, from restrained and unrestrained specimens simultaneously subjected to a cooling rate of 10°C/hr starting from an initial tem- perature of 20°C. A minimum of two replicates for the restrained specimen were tested for each evaluated mixture, and the same unrestrained specimen was tested twice, once for each of the restrained specimen replicate tests. The following five characteristic stages of material behavior are identified from the stiffness-temperature relationship and thermal stress curve (Figure 7[b], Figure 7[c]): • Viscous softening: From this stage, the relaxation modulus of the asphalt mixture increases rapidly with decreasing temperature. • Viscous-glassy transition: At this stage, the glassy properties of the material overcome the viscous properties. • Glassy hardening: At this stage, the behavior of the material is purely glassy. • Crack initiation: In this stage, micro-cracks occur in the specimen due to the induced thermal stresses while the material behavior is glassy. • Fracture: At this stage, the asphalt mixture specimen breaks due to the propagation of micro- cracks by the induced thermal stresses (i.e., macro or global failure). In this study, the primary evaluation was performed by examining the stresses and tempera- tures corresponding to the fracture, crack initiation, and viscous softening stages. Further information regarding the thermo-viscoelastic properties and the stiffness-temperature rela- tionship can be found in the literature (Alavi et al. 2013; Hajj et al. 2013b; Morian et al. 2014; Alavi and Hajj 2014; Mensching et al. 2014; Alavi and Morian et al. 2015).

Introduction 33 Additional characterization resulting from the UTSST measurements was also developed to summarize mixture low-temperature cracking resistance and combine specific aspects of the thermal stress and thermal strain relationships with those of the thermos-viscoelastic property regions and recognize the benefits of certain mixture characteristics. A cracking resistance index (CRIEnv) was determined through calculations based on the measured thermal stress and strain plots, as indicated in Figure 8(a) and Equation 11, including an environmental correction factor, FEnv, as defined in Equation 12, that relates mixture cracking resistance to that required by the environment, as shown in Figure 8(b) and Figure 8(c). ( ) [ ]= + + + + +     σ σ × 1 Equation 11CRI A A A A A A A A FEnv v i v v i p p i vgt f Env [ ]= − − Equation 12F A A Env vgt F vgt crit (c) (a) (b) Figure 7. (a) UTSST setup; (b) measured thermal stress and strain; (c) calculated UTSST modulus and associated characteristic stages.

34 Evaluating the Effects of Recycling Agents on Asphalt Mixtures with High RAS and RAP Binder Ratios (a) (b) (c) Figure 8. (a) Thermal stress and strain plots with CRI parameters; (b) environmental adjustment parameters when Tcritical < TFracture; (c) environmental adjustment parameters when Tcritical > TFracture.

Introduction 35 where CRIEnv = UTSST cracking resistance index including the environmental adjustment factor; Av = area of viscous behavior, i.e., area of stress–strain up to viscous softening; Ai = area of crack initiation, i.e., area of stress–strain from viscous softening up to crack initiation; Ap = area of crack propagation, i.e., area of stress–strain from crack initiation up to ulti- mate fracture; svgt = thermal stress at viscous–glassy transition; sf = thermal stress at fracture; Avgt-F = area under the thermal stress–strain plot between the viscous–glassy transition temperature and the fracture temperature of the restrained UTSST specimen; and Avgt-crit = area under the thermal stress–strain plot between the viscous–glassy transition temperature and the required environmental temperature at a given location. In this configuration, increased levels of low-temperature cracking resistance are indicated by larger values of the CRIEnv. For instance, a mixture may exhibit limited resistance to crack propagation (i.e., low Ap) but may show higher levels of overall cracking resistance (i.e., CRI) if the mixture exhibits a high level of crack initiation resistance (i.e., Ai). By similar logic, the cracking resistance of a mixture would increase with larger measured fracture stress, sf. However, the overall resistance would be reduced by an elevated stress level at the viscous–glassy transition stage, svgt, which would indicate increased stress levels coupled with glassy or brittle behavior. The addition of the environmental factor, FEnv, provides a simplified correction to acknowl- edge if the mixture will fracture above or below the required environmental temperature, defined as the DOT-selected PGL for the respective field projects (−22°C for TX, NV, and IN; −28°C for WI and MN).