Short-Term Laboratory Conditioning of Asphalt Mixtures (2015)

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.

13 This chapter provides an overview of the experimental design for Phases I and II, field site descriptions, specimen fabrication protocols, and laboratory tests. Experimental Design Phase I Experiment The objectives of the Phase I experiment were to (1) develop a laboratory STOA protocol for asphalt loose mix prior to compaction to simulate short-term asphalt aging and absorption of asphalt mixtures during plant production and (2) identify mixture components and production parameters (henceforth named factors) with significant effects on the performance-related properties of short-term aged asphalt mixtures. Fig ure 2-1 presents a schematic of the short-term aging methodology used for this experiment. Nine field sites in Texas, New Mexico, Connecticut, Wyoming, South Dakota, Iowa, Indiana, and Florida were included in the Phase I experiment, which will be detailed in the following sections. Cores at construction and plant- mixed, plant-compacted (PMPC) specimens were obtained for each mixture at each field site, in conjunction with raw materials including asphalt binder, aggregates, and RAP and RAS used for LMLC specimen fabrication. Based on previous research for NCHRP Project 9-49 (Epps Martin et al. 2014), the laboratory STOA protocols of 2 hours at 275°F (135°C) for HMA and 240°F (116°C) for WMA were used for con- ditioning loose mix prior to compaction. All these asphalt mixtures subjected to short-term aging during production and compaction in the field or laboratory were tested with the MR, E*, and HWTT tests. Binders were extracted from short- term aged asphalt mixtures from three field sites (i.e., Indi- ana, Florida, and Texas II), recovered, and then tested with the DSR and Fourier transform infrared spectroscopy (FT-IR) to characterize their rheological and chemical properties, including continuous performance grade (PG) and complex shear modulus (G*) at 77°F (25°C), and chemical character- istics in terms of the FT-IR carbonyl area (CA). The determi- nation of the continuous PG was helpful in characterizing the change in asphalt binder properties with short-term aging, and the parameter G* at 77°F (25°C) was included in order to provide additional information on binder stiffness measured at the same temperature as the mixture MR stiffness. The mixture and binder results for LMLC specimens fab- ricated using the selected STOA protocols were compared against those for cores at construction and PMPC specimens to validate the laboratory STOA protocol of 2 hours at 275°F (135°C) for HMA and 240°F (116°C) for WMA. Additionally, a second set of comparisons was performed to evaluate the effects of mixture and production factors on the short-term aging of asphalt mixtures for each type of sample, i.e., cores at construction, PMPC specimens, and LMLC specimens. Phase II Experiment The objectives of the Phase II experiment were to (1) evalu- ate the evolution of performance-related properties of asphalt mixtures with intermediate-term aging in the field and labo- ratory, (2) develop a correlation between intermediate-term field aging and laboratory LTOA protocols, and (3) identify mixture components and production parameters with sig- nificant effects on the intermediate- or long-term aging char- acteristics of the asphalt mixtures. Additionally, the aging of WMA relative to HMA was determined in terms of when the WMA stiffness was equal to HMA or when the stiff- ness of WMA was equivalent to the initial stiffness of HMA. Figure 2-2 presents the research methodology used for the Phase II experiment. Seven of the sites from Phase I were included in the Phase II experiment; the Connecticut and Texas II field sites were not used in this experiment because no post-construction cores were obtained due to the traffic concerns of the agency or time constraints of the project. Cores after certain in-service C H A P T E R 2 Research Approach

14 times (ranging from 8 to 22 months) were incorporated in Phase II to quantify aging in the field. To simulate long-term aging in the laboratory, LMLC specimens fabricated with the STOA protocol of 2 hours at 275°F (135°C) for HMA and WMA were further aged after compaction in the environ- mental room or oven (i.e., laboratory LTOA protocol) prior to testing. Two laboratory LTOA protocols were selected based on previous research: 5 days at 185°F (85°C) in accordance with AASHTO R 30 and 2 weeks at 140°F (60°C). Consistent with the Phase I experiment, laboratory mix- ture tests included MR, HWTT, and E*. Binders were extracted from Indiana and Florida long-term aged mixtures (i.e., cores and LMLC specimens with STOA plus LTOA protocols), recovered, and tested with DSR and FT-IR to characterize the change in the rheological and chemical properties. Previous literature indicated that the field aging of asphalt mixtures has been commonly quantified by the in-service time of the pavement at the time of coring. However, a poten- tial issue with using in-service time arose from the differences in construction dates and climates for the various field sites. To address that issue, the concept of cumulative degree-day (CDD; 32°F [0°C] base) was proposed in this project, as expressed in Equation (2-1). TCDD 32 Eq. (2-1)dmax∑( )= − Where: Tdmax = daily maximum temperature, °F. The CDD values for two field sites that were part of NCHRP Project 9-49 (Epps Martin et al. 2014) indicated that an 8-month field aging period in Texas (including the summer) was equivalent to approximately a 12-month aging period in Iowa in terms of increasing mixture stiffness. Therefore, com- pared to the field in-service time, the CDD value provided a more discriminating measure of field aging when comparing Identification of Significant Factors PMPC HMA 2h@275F HMA Stabilize@275F vs. & WMA 2h@240F WMA Stabilize@240F Laboratory Testing Simulation of Plant Aging Construction CoreLMLC HMA vs. WMA High vs. Control TProduction BMP vs. DMP RAP/RAS vs. No RAP/RAS High vs. Low Agg Abs Binder I vs. Binder II Test Sites Texas I New Mexico Connecticut Wyoming South Dakota Iowa Indiana Florida Texas II Figure 2-1. Research methodology for the Phase I experiment. Identification of Significant Factors STOA + LTOA of 2w@140F vs. STOA + LTOA of 5d@185F Lab LTOA vs. Field Aging Construction CoreLMLC HMA vs. WMA High vs. Control TProduction BMP vs. DMP RAP/RAS vs. No RAP/RAS High vs. Low Agg Abs Test Sites Texas I New Mexico Wyoming South Dakota Iowa Indiana Florida Post-Construction Core after x CDD Laboratory Testing Figure 2-2. Research methodology for the Phase II experiment.

15 field sites built in different climates and at various times of the year. Details of the CDD values for field sites included in this project are presented and discussed in the next chapter. To better quantify the evolution of mixture stiffness and rutting resistance, and binder stiffness and oxidation with field and laboratory aging, four binder or mixture property ratios—MR ratio, HWTT rutting resistance parameter (RRP) ratio, DSR G* ratio, and FT-IR CA ratio—were proposed. As expressed in Equation (2-2), the MR ratio is defined as the fraction of the MR stiffness of either field or laboratory long- term aged (LTA) specimens over that of short-term aged (STA) specimens from either the laboratory or field. M Ratio M M Eq. (2-2)R R-LTA R-STA= Where: MR-LTA = MR stiffness of long-term aged specimens includ- ing cores after certain in-service times or LMLC specimens with STOA plus LTOA protocols; and MR-STA = MR stiffness of short-term aged specimens includ- ing construction cores or LMLC specimens with STOA protocols. Because field and laboratory aging produced asphalt mix- tures with increased MR stiffness, the MR ratio was expected to be greater than 1.0. However, the HWTT RRP ratio exhib- ited the opposite trend with aging due to the mixture stiffen- ing effect (i.e., less rutting with aging). Therefore, the HWTT RRP ratio was defined as the ratio of the RRP of STA speci- mens over the RRP of LTA specimens, as expressed in Equa- tion (2-3), in order to also get a ratio greater than 1.0 with aging. HWTT RRP Ratio RRP RRP Eq. (2-3)STA LTA= Where: RRPSTA = HWTT RRP of short-term aged specimens includ- ing construction cores or LMLC specimens with STOA protocols; and RRPLTA = HWTT RRP of long-term aged specimens includ- ing cores after certain in-service times or LMLC specimens with STOA plus LTOA protocols. Similar to the MR ratio, the G* ratio and the FT-IR CA ratio are defined as the fraction of the DSR G* at 77°F (25°C) and FT-IR CA value of extracted and recovered binders from either laboratory or field LTA specimens over that of STA specimens, as expressed in Equations (2-4) and (2-5), respec- tively. In addition, both the G* ratio and the FT-IR CA ratio were expected to be greater than 1.0 since field and laboratory aging were likely to produce asphalt binders with higher stiff- ness and oxidation due to the formation of oxygen-containing C=O bonds. DSR G* Ratio G G Eq. (2-4)LTA* STA*= − − Where: G*–LTA = G* of extracted and recovered binders from long- term aged specimens including cores or LMLC specimens; and G*–STA = G* of extracted and recovered binders from short- term aged specimens including construction cores or LMLC specimens. FT-IR CA Ratio CA CA Eq. (2-5)LTA STA= − − Where: CA–LTA = CA of extracted and recovered binders from long- term aged specimens including cores after certain in-service times or LMLC specimens; and CA–STA = CA of extracted and recovered binders from short- term aged specimens including construction cores or LMLC specimens. MR ratio, HWTT RRP ratio, DSR G* ratio, and FT-IR CA ratio values greater than 1.0 indicate an increase in mixture stiffness and rutting resistance, as well as binder stiffness and oxidation, after long-term aging compared to the short-term aged counterparts. To discriminate asphalt binders or mix- tures with different aging characteristics, mixtures with higher property ratios are considered more sensitive to aging and more likely to exhibit an increase in mixture stiffness and rut- ting resistance as well as binder stiffness and oxidation after a certain level of aging. To further characterize the evolution of binder or mixture properties with field aging, the exponential function shown in Equation (2-6) was used to correlate the measured ratio values of post-construction cores with their corresponding CDD values. As will be explained in Chapter 3, Equation (2-6) provides a good fit of data for field-aged material prop- erties plotted against CDD. Binder or Mixture Property Ratio 1 exp CDD Eq. (2-6) = + α − β    γ p Where: CDD = cumulative degree-days for cores after certain in-service times; and a, b, and g = fitting coefficients. Findings from the Phase I experiment indicated that the laboratory STOA protocol of 2 hours at 275°F (135°C) for HMA and 240°F (116°C) for WMA was representative of cores at construction in terms of mixture stiffness and rutting resistance. Therefore, based on the definitions for binder or

16 mixture property ratios, the correlation between field aging and laboratory LTOA protocols could be made. To sum up, the mixture and binder results obtained in the Phase II experiment for cores versus LMLC specimens with STOA plus LTOA protocols were used to evaluate the effect of field and laboratory aging on mixture properties and to develop a correlation between field aging and laboratory LTOA protocols. Comparisons were also performed to evalu- ate the effect of mixture and production factors on the long- term aging characteristics of asphalt mixtures. Laboratory Tests Based on previous experience with laboratory testing, MR, E*, and HWTT tests were selected in this project to compare the stiffness and rutting resistance of asphalt mixtures with various factors of mixture components, production param- eters, and aging stages. DSR, bending beam rheometer (BBR), and FT-IR tests were included in order to evaluate the change in the rheological and chemical properties of asphalt binders with aging. Details of each laboratory test are given in the following subsections. Resilient Modulus The MR test was conducted through repetitive applica- tions of compressive loads in a haversine waveform along a vertical diametral plane of cylindrical asphalt concrete speci- mens. The resulting horizontal deformations of the specimen were measured by two linear variable differential transducers (LVDTs) aligned along the horizontal diametral plane. An environmentally controlled room at 77°F (25°C) was used for temperature conditioning and testing. The test equipment used to perform the measurements is shown in Figure 2-3. MR stiffness was measured in accordance with the current ASTM D-7369 with a modification consisting of replacing the on-specimen LVDTs with external LVDTs aligned along the horizontal diametral plane (i.e., gauge length as a frac- tion of diameter of the specimen = 1.00). As expressed in Equation (2-7), the MR stiffness was calculated based on ver- tical load, horizontal deformation, and the asphalt mixture’s Poisson ratio. M 0.2732 Eq. (2-7)R P t )( = υ + ∆ Where: MR = resilient modulus of asphalt mixture; P = vertical load; u = Poisson’s ratio; t = specimen thickness; and D = horizontal deformation measured by LVDTs. Dynamic Modulus The E* test was conducted under unconfined conditions using the Asphalt Mixture Performance Tester, shown in Fig- ure 2-4, following the test procedure specified in AASHTO TP 79-13. Superpave gyratory compactor specimens were Figure 2-4. Asphalt Mixture Performance Tester. (a) sample setup in loading frame (b) data acquisition system Loading Pulse Mixture Response 0 100 80 60 40 20 0 –20 –40 –60 200 400 600 800 1000 Figure 2-3. MR test equipment.

17 compacted to a height of 6.7 in. (170 mm) and then cored and trimmed to obtain test specimens with a diameter of 4.0 in. (100 mm) and a height of 6.0 in. (150 mm). Testing was conducted at 39.2°F (4°C), 68°F (20°C), and 104°F (40°C) and three frequencies of 0.1, 1, and 10 Hz for each temperature. Load levels were determined by a trial and error process to ensure that the amplitude of measured verti- cal strains was in the range of 50 to 75 microstrains and to prevent damage to the test specimen. The E* master curve was constructed by fitting the E* val- ues at each temperature/frequency condition to the sigmoidal function described in Equation (2-8), followed by horizon- tally shifting according to the time–temperature shift factor function expressed in Equation (2-9). To further discrimi- nate E* stiffness of asphalt mixtures due to different binder or mixture aging levels, the E* stiffness at 68°F (20°C) and 10 Hz was used as another indicator for asphalt mixture stiffness in addition to the E* master curve. log * 1 1 Eq. (2-8) log E a b ec g fR = + + ( )+ p Where: E* = dynamic modulus of asphalt mixture; fR = reduced frequency; and a, b, c, and g = fitting coefficients of the sigmoidal function. a T TTlog Eq. (2-9)1 2 2 2 3= α + α + α Where: aT = time–temperature shift factor; and a1, a2, and a3 = fitting coefficients of time–temperature shift factor function. Hamburg Wheel-Tracking Test The HWTT (AASHTO T 324) is a laboratory test com- monly used for evaluating rutting resistance and moisture susceptibility of asphalt mixtures. The test consists of sub- merging specimens in water at 122°F (50°C) for one hour prior to subjecting them to 52 passes per minute of a loaded steel wheel. Two replicate specimens were loaded for a maxi- mum of 20,000 load cycles or until the deformation at the center of the specimen reached 12.5 mm, per Texas Depart- ment of Transportation standard specification Tex-242-F, at which point the test was stopped. The HWTT equipment used to perform the measurements is shown in Figure 2-5. Typical results of the rut depth versus number of load cycles consist of three phases: (1) post-compaction, (2) creep, and (3) stripping (Solaimanian et al. 2003). The traditional test parameter for evaluating the rutting performance of asphalt mixtures is the rut depth at a specific number of load cycles (i.e., 5,000, 10,000, 15,000, and 20,000). In this study, the rut depth at 5,000 load cycles was selected as a rutting resistance parameter in the HWTT since a significant amount of the short-term aged mixtures (especially those using soft bind- ers, i.e., low high-temperature PG binders) failed to achieve higher levels of loading. In addition to the traditional HWTT rutting analysis, a novel method developed by Yin et al. (2014) was included in the project to discriminate between asphalt mixtures with different rutting resistance. The Yin et al. RRP value (i.e., visco- plastic strain at the stripping number) was employed as an alternative to the HWTT rutting resistance parameter and accuracy was significantly improved by isolating the visco- plastic strain during the creep phase and excluding any con- tributions from the post-compaction phase due to different specimen AV content or to stripping. The determination of alternative RRP is schematically illustrated in Figure 2-6, and the explanation of the analysis method follows. The methodology to analyze the HWTT results included curve fitting the entire output of rut depth versus load cycles and defining a test parameter to evaluate mixture rutting resistance, as expressed in Equation (2-10): RD ln LC LC Eq. (2-10) ult 1( )= ρ  −βp Where: LC = the number of load cycles at a certain rut depth; LCult = the maximum number of load cycles; RD = rut depth at a certain number of load cycles (mm); and r and b = curve fitting coefficients. Figure 2-5. HWTT equipment.

18 As illustrated in Figure 2-6, the fitted HWTT rut depth curve was composed of one part with negative curvature used to evaluate mixture resistance to rutting due to defor- mation, followed by a second part with positive curvature used to evaluate mixture resistance to stripping. The critical point where the curvature changed from negative to positive was defined as the stripping number (SN), and the load cycle where SN occurred was labeled as LCSN, which can be deter- mined following Equation (2-11): LC LC Eq. (2-11)SN ult 1 e= − β+ β    p The rut depth accumulated before the SN is related to the viscoplastic deformation of the mixture, and this viscoplastic strain could be calculated as the ratio of the rut depth to the specimen thickness at any given number of load cycles up to LCSN. A typical viscoplastic strain versus load cycle curve is also shown in Figure 2-6. Then, the Tseng–Lytton model (Tseng and Lytton, 1989) was employed to fit the viscoplastic strain data using Equation (2-12): vp vp exp LC Eq. (2-12)( )ε = ε − α ∞ λ Where: ε∞vp = saturated viscoplastic strain (treated as a fitting coefficient); and a and l = model coefficients. To better quantify mixture resistance to rutting in the HWTT and compare different mixtures, the RRP (slope of the viscoplastic strain versus load cycle curve at SN) was proposed as an alternative to the traditional parameter (i.e., rut depth at a specific number of load cycles [5,000, 10,000, 15,000, or 20,000]), as expressed in Equation (2-13): vp vpRRP i.e., exp LC LC Eq. (2-13) LC SN SN 1 SN( ) ( )∆ε = α λε − α    ( )λ ∞ λ − λ+ Compared to the traditional parameter, the determination of the RRP isolates the viscoplastic strain during the creep phase and does not include contributions due to specimen AV content or stripping. The RRP should be able to better char- acterize mixture rutting resistance in the HWTT. Specifically, asphalt mixtures with higher RRP values are expected to be more susceptible to rutting than those with lower RRP values. According to laboratory experience with the analysis method, early stripping had been frequently observed for short-term aged asphalt mixtures using softer asphalt binders and/or recy- cled materials, with LCSN observed at less than 3,000 load cycles. These mixtures had a limited duration of the creep phase before stripping occurred and, as a consequence, the determination of the viscoplastic strain was not feasible. Therefore, in this proj- ect, the evaluation of rutting resistance of asphalt mixture by the RRP was only performed for asphalt mixtures having LCSN greater than 3,000 load cycles. Dynamic Shear Rheometer The DSR is commonly used to characterize the viscous and elastic behavior of asphalt binders at medium to high 0.0E+00 1.0E-01 2.0E-01 3.0E-01 4.0E-01 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 0 5000 10000 15000 20000 25000 30000 St ra in R ut D ep th (m m) Load Cycle Measured Rut Depth Predicted Rut Depth Viscoplastic Strain Permanent Strain Stripping Strain Stripping Number LCSN LCST Tseng-Lytton Model εvp = ε∞ vp exp − α LC   λ    RD (LC ) = ρ∗ ln LCu lt LC        1 β Str ipp ing M od el ε st = ε 0 st ex p θ LC − LC SN ( ) [ ]−1 { } Figure 2-6. Determination of alternative HWTT rutting resistance parameter.

19 temperatures. As shown in Figure 2-7, the equipment used in this project was a Malvern Bohlin DSR2. During the test, a 0.04 in. (1 mm) or 0.08 in. (2 mm) thick sample of asphalt binders was placed between two parallel circular plates (0.32 in. [8 mm] or 1.0 in. [25 mm] in diameter). The bot- tom plate was fixed, while the top plate oscillated back and forth across the sample at a given frequency to create shear in the sample. The angular rotation and the applied torque were measured during the test and were then used to calcu- late G* and phase angle (d). G* is the asphalt binder’s total resistance to deformation in repeated shear, and d is the time delay between the applied shear stress and the resulting shear strain. According to the Superpave PG asphalt binder specifica- tions, DSR tests are conducted on unaged, short-term aged, and long-term aged binders. In this project, in order to char- acterize the unaged and short-term aged asphalt binders, the test was performed at high pavement in-service tempera- tures, ranging from 137°F (58°C) to 169°F (76°C), in accor- dance with AASHTO T 315, and the parameter of G*/sin(d) was used to evaluate mixture resistance to rutting. The high-temperature PG of the asphalt binder was determined based on the G*/sin(d) values at different temperatures per ASTM D7643. Besides the traditional DSR testing, the DSR frequency sweep test was incorporated in the experimental design as an alternative to the BBR test in order to determine the asphalt binder’s low-temperature properties. Replacing the BBR test with the DSR frequency sweep test was advanta- geous due to the significantly reduced amount of material required and shorter preparation time for the extracted binder. In this project, the DSR frequency sweep test was conducted at 43°F (6°C), which was the lowest achievable test temperature by the available DSR equipment. To predict the BBR test results (i.e., stiffness and m-value), Equation (2-14)—provided in Strategic Highway Research Program (SHRP) Report A-369 (Anderson et al. 1994)—was used to determine the frequency for the DSR testing at the interme- diate temperature. ( ) = + −     − − p p p T T R t f1 273 2.303 log 250,000 273 Eq. (2-14)d s s 1 Where: Td = test temperature for DSR testing at frequency f, °C; Ts = specified temperature for BBR test, °C; R = ideal gas constant, 8.31 J/°K-mol; ts = specified creep loading time, 60 s; and f = DSR testing frequency, rad/s. Considering the DSR frequency sweep test temperature of 43°F (6°C) and the specified creep loading time of 60 seconds in the BBR test, the relationship between BBR test tempera- ture and DSR testing frequency was established and is sum- marized in Table 2-1. Due to the frequency limitations (0.001 to 100 Hz) of the available DSR equipment at the Texas A&M Transportation Institute, the BBR prediction using the DSR frequency sweep test could not be applied to binders with a low-temperature PG below -28°C. Figure 2-8 presents a typical DSR frequency sweep test result in terms of G* and d versus testing frequency. Given the desired temperature for the BBR tests, the correspond- ing frequency used in the DSR frequency sweep test (fc) could be found using Table 2-1. Then, the G* and d at fc were obtained based on Figure 2-8, referred to as G*(fc) and d(fc), Figure 2-7. DSR test equipment. Table 2-1. Selected frequency for the DSR frequency sweep test. PG BBR TestTemperature (°C) Selected DSR Frequency (Hz) XX-16 −6 0.34 XX-22 −12 4.49 XX-28 −18 67.66 XX-34 −24 1160.72

20 respectively. Then, the creep stiffness for the BBR test was calculated using Equation (2-15). 3 1 0.2sin 2 Eq. (2-15) * c c S G t f f[ ]( )= + δ ( ) ( ) Additionally, the BBR m-value was calculated by taking the first derivative of the G* versus frequency function (shown in Figure 2-8) at fc, as expressed in Equation (2-16): [ ] [ ]( ) ( ) = ∂ ∂ = m G f f f log * log Eq. (2-16) c Bending Beam Rheometer The BBR test is commonly used to characterize the low- temperature relaxation properties of the original asphalt binders after long-term aging. In this project, the BBR test was performed in accordance with AASHTO T 313, and the test equipment is shown in Figure 2-9. During the test, an asphalt binder beam of 4.9 in. (125 mm) long, 0.5 in. (12.5 mm) wide, and 0.25 in. (6.25 mm) thick was supported in a cool- ing bath. After the temperature of the beam achieved the des- ignated testing temperature, a vertical load was applied to the center of the beam for 240 seconds. During loading, the deflection at the center of the beam was measured against time. The low-temperature relaxation properties of the asphalt binders in terms of relaxation stiffness and m-value were cal- culated based on the dimensions, applied loads, and measured deflections of the beam. According to the Superpave PG asphalt binder specifications, BBR tests should be conducted on long-term aged asphalt bind- ers. In this project, the BBR test was conducted on the origi- nal binder after rolling thin-film oven (RTFO) plus pressure aging vessel (PAV) aging in order to characterize their low- temperature rheological properties. Additionally, the stiffness and m-value results measured at different test temperatures were used to determine the continuous low-temperature PG of the asphalt binder in accordance with AASHTO PP 42. Fourier Transform Infrared Spectroscopy The FT-IR analysis has been proven to be an effective tool to determine the compositional changes occurring in asphalt binders with aging. During the process of oxidation, changes occur in the chemical bonds and molecular structure of the asphalt binder; polar oxygen-containing functional com- pounds, which contain infrared active carbonyl C=O bonds, are formed (Michalica et al. 2008; Jia et al. 2014). Therefore, the amount of asphalt aging can be quantified by measuring the change in the amount of carbonyl C=O bonds. In addition to DSR and BBR tests, the infrared spectra anal- ysis was also performed on short-term aged and long-term aged asphalt binders in order to characterize their chemi- cal properties. Figure 2-10 presents the Thermo Scientific y = -0.0498x2 + 0.4133x + 7.482 R² = 1 y = 0.8849x2 – 8.3059x + 37.43 R² = 0.9996 20 25 30 35 40 45 50 6.9 7.2 7.5 7.8 8.1 –1.50 –0.50 0.50 1.50 2.50  ,  Lo g G * (P a) Log(f), Hz G* δ Figure 2-8. DSR frequency sweep test result example. Figure 2-9. BBR test equipment. Figure 2-10. FT-IR test equipment.

21 Nicolet 6700 FT-IR spectrometer used in the project. Dur- ing the test, a Teflon-coated spatula was used to apply the molten asphalt sample (approximately 0.5 g) to the reflec- tion surface of the prism, which was preheated in an oven at 302°F (150°C). The CA was measured using the attenuated total reflectance method (effective path length was about 4 µm). The CA was defined as the integrated peak area from 1820 to 1650 cm–1, measured in arbitrary units, as a surro- gate of asphalt oxidation level (Jemison et al. 1992). Asphalt binders with higher CA values were expected to have experi- enced a greater level of aging compared to those with lower CA values. Field Sites Table 2-2 provides a summary of the field sites used in this project. A detailed construction report for each field site is pro- vided in Appendix A. Based on a literature search on asphalt Location & Climate Date Plant Type TProduction Asphalt Aggregate & Additives Mixtures Factors Texas I FM 973 Wet–No Freeze 1/12 CFD 325°F (163°C) HMA 325°F (163°C) HMA+RAP/RAS 275°F (135°C) Foaming 275°F (135°C) Evotherm 270°F (132°C) Evotherm+RAP/RAS 5.2% AC PG 70-22 (Styrene– Butadiene– Styrene [SBS]) PG 64-22 w/ 15%RAP / 3%RAS Limestone HMA HMA + 15%RAP + 3%RAS WMA Foaming WMA Evotherm DAT WMA Evotherm + 15%RAP + 3%RAS HMA vs. WMA RAP/RAS vs. No RAP New Mexico IH 25 Dry–No Freeze 10/12 CFD 345°F (174°C) HMA 315°F (157°C) HMA + RAP 285°F (141°C) Foaming 275°F (135°C) Evotherm 5.4% AC PG 76-28 (SBS) PG 64-28 w/ 35%RAP Siliceous Gravel 1% Versabind HMA HMA + 35%RAP WMA Foaming + 35%RAP WMA Evotherm 3G + 35%RAP HMA vs. WMA RAP vs. No RAP Connecticut IH 84 Wet–Freeze 8/12 CFD 322°F (161°C) HMA 312°F (156°C) Foaming 5.0% AC PG 76-22 (SBS) w/ 20%RAP Basalt HMA + 20%RAP WMA Foaming + 20%RAP HMA vs. WMA Wyoming SR 196 Dry–Freeze 8/12 CFD 315°F (157°C) HMA 255°F & 275°F (124°C & 135°C) Evotherm 275°F & 295°F (135°C & 146°C) Foaming 5.0% PG 64-28 (polymer) Limestone 1% Lime HMA WMA Foaming WMA Evotherm 3G HMA vs. WMA Production Temperature (WMA) South Dakota SH 262 Dry–Freeze 10/12 CFD 310°F (154°C) HMA 275°F (135°C) Foaming 270°F (132°C) Evotherm 280°F (138°C) Advera 5.3% PG 58-34 (SBS) w/ 20%RAP Quartzite 1% Lime HMA + 20%RAP WMA Foaming + 20%RAP WMA Evotherm 3G + 20%RAP WMA Advera + 20%RAP HMA vs. WMA Iowa Fairgrounds Wet–Freeze 6/13 CFD 295°F & 325°F (146°C & 163°C) Low Abs HMA 295°F & 310°F (146°C & 154°C) High Abs HMA 265°F & 295°F (129°C & 146°C) Low Abs Foaming 260°F & 290°F (127°C & 143°C) High Abs Foaming 5.0% (0.9% AC Limestone) and 7.0% (3.2% AC Limestone) PG 58-28 w/ 20%RAP 0.9% AC Limestone + Field Sand 3.2% AC Limestone + Field Sand Low Abs HMA + 20%RAP High Abs HMA + 20%RAP Low Abs WMA Foaming + 20%RAP High Abs WMA Foaming + 20%RAP HMA vs. WMA Production Temperature Aggregate Absorption Florida Parking Wet–No Freeze 8/13 CFD 306°F (152°C) Low Abs HMA 308°F (153°C) High Abs HMA 272°F (133°C) Low Abs Foaming 267°F (131°C) High Abs Foaming 5.1% (0.6% AC Granite) and 6.8% (3.7% AC Limestone) PG 58-28 w/ 25%RAP 0.6% AC Granite 3.7% AC Limestone 0.5% Liquid Antistrip Low Abs HMA + 25%RAP High Abs HMA + 25%RAP Low Abs WMA Foaming + 25%RAP High Abs WMA Foaming + 25%RAP HMA vs. WMA Aggregate Absorption Table 2-2. Summary of field sites. (continued on next page)

22 Location & Climate Date Plant Type TProduction Asphalt Aggregate & Additives Mixtures Factors Texas II Local Dry–No Freeze 4/14 CFD BMP 310°F (154°C) HMA (Binder V, CFD) 315°F (157°C) HMA (Binder V, BMP) 310°F (154°C) HMA (Binder A, CFD) 315°F (157°C) HMA (Binder A, BMP) 6.2% PG 64-22 Limestone HMA (Binder V, CFD) @ 5.9% AC HMA (Binder V, BMP) @ 6.4% AC HMA (Binder A, CFD) @ 5.9% AC HMA (Binder A, BMP) @ 6.4% AC Plant Type Asphalt Source Indiana Residential Wet–Freeze 8/13 CFD BMP 300°F (149°C) HMA + RAP (CFD) 305°F (152°C) HMA + RAP (BMP) 271°F (133°C) Foaming + RAP (CFD) 273°F (134°C) Advera + RAP (BMP) 5.8% PG 64-22 Limestone HMA + 25%RAP (CFD) HMA + 25%RAP (BMP) WMA Foaming + 25%RAP (CFD) WMA Advera + 25%RAP (BMP) HMA vs. WMA Plant Type Table 2-2. (Continued). plants and a preliminary laboratory experiment (Appendixes B and C, respectively), the following factors—besides climate (wet–freeze, dry–freeze, wet–no freeze, and dry–no freeze)— were considered when selecting field sites that included a wide spectrum of materials and production parameters: • Aggregate type (high water absorption capacity [AC] and low AC) • Asphalt source (fast aging and slow aging) • Recycled materials (RAP and RAS) • WMA technology (WMA and HMA) • Plant type (batch plant and drum plant) • Production temperature (high and low) Certain factors such as asphalt grade and nominal maxi- mum aggregate size (NMAS) had to be treated as co-variables. All field sites included at least one of the factors listed above. Figures 2-11, 2-12, and 2-13 summarize the continuous high-temperature performance grades, low-temperature per- formance grades, and intermediate-temperature performance grades for 11 different binders used in nine field sites, respec- tively. For each binder type presented in Figures 2-11 and 2-12, two bars describe the results; the bar on the left of each pair represents the specified PG provided by the material supplier, while the bar on the right represents the PG measured in the laboratory. As illustrated in Figure 2-11, for 9 out of 11 binders included in the study, the high-temperature PG results obtained from the DSR testing on unaged and/or RTFO aged binders matched the specified grades. The only two exceptions were Texas I and New Mexico 76 (PG 76-28) binders, which showed a higher measured PG. A similar trend in the low-temperature PG results from labo- ratory testing versus specified values is shown in Figure 2-12. TXI NM 76 NM 64 CT WY SD IA FL IN TXII-A TXII-V Specified 70.0 76.0 64.0 76.0 64.0 58.0 58.0 58.0 64.0 64.0 64.0 Graded 76.8 82.0 66.0 76.9 67.0 59.8 62.4 62.1 66.5 64.4 64.2 52 58 64 70 76 82 88 H ig h- Te m pe ra tu re P G Figure 2-11. Continuous high-temperature PG results.

23 TXI NM 76 NM 64 CT WY SD IA FL IN TXII-A TXII-V Specified –22.0 –28.0 –28.0 –22.0 –28.0 –34.0 –28.0 –22.0 –22.0 –22.0 –22.0 Graded –23.8 –30.8 –27.4 –24.6 –25.4 –34.9 –29.2 –28.3 –21.6 –24.4 –24.0 –40 –34 –28 –22 –16 Lo w -T em pe ra tu re P G Figure 2-12. Continuous low-temperature PG results. TXI NM 76 NM 64 CT WY SD IA FL IN TXII-A TXII-V Graded 18.2 12.0 13.8 23.1 19.2 11.1 18.0 21.6 19.5 22.5 19.5 9 12 15 18 21 24 In te rm ed ia te -T em pe ra tu re P G Figure 2-13. Continuous intermediate-temperature PG results. For most of the binders, the low-temperature PG results obtained from the BBR tests on RTFO plus PAV aged bind- ers matched the grades specified by the supplier. However, a slightly higher PG was exhibited by the laboratory test results for the New Mexico 64 (PG 64-28) binder (-27.4 instead of -28.0) and Wyoming binder (-25.4 instead of -28.0), while the opposite trend was shown for the Florida binder (-28.3 instead of -22.0). The continuous PG results for the 11 binders used in the nine field sites are summarized in Table 2-3. Specimen Fabrication To fabricate LMLC specimens, aggregates and binders were heated to the specified plant mixing temperature (Tm) and then mixed with a portable bucket mixer. Afterwards, the loose mix was conditioned in the oven following the labora- tory STOA protocol of 2 hours at 275°F (135°C) for HMA and 240°F (116°C) for WMA prior to compaction with the Superpave gyratory compactor. Trial specimens were fabri- cated to ensure specimens were obtained with AV contents of 7 ± 0.5 percent. To simulate the intermediate-term mix- ture aging in the field for the Phase II experiment, LMLC specimens were conditioned following a STOA protocol of 2 hours at 275°F (135°C) for both HMA and WMA and then further aged in accordance with the selected laboratory LTOA protocols in an environmental room or oven prior to being tested. In total, 522 LMLC specimens with 7 ± 0.5 percent AV content were fabricated for the nine field sites that included 40 mixtures. For PMPC specimens, loose mix was taken from the trucks before leaving the plant and maintained in the oven for 1 to 2 hours at the field compaction temperature prior to compac- tion (Epps Martin et al. 2014). As mentioned previously, cores were obtained at construction for all nine field sites. Addition- ally, one or more sets of post-construction cores were acquired

24 Table 2-3. Continuous PG summary. Field Site Specified PG Continuous PG* Texas I 70-22 76.8–23.8 (18.2) New Mexico 76-22 82.0–30.8 (12.0) New Mexico 64-28 66.0–27.4 (13.8) Connecticut 76-22 76.9–24.6 (23.1) Wyoming 64-28 67.0–25.4 (19.2) South Dakota 58-34 59.8–34.9 (11.1) Iowa 58-28 62.4–29.2 (18.0) Indiana 64-22 66.5–21.6 (21.6) Florida 58-22 62.1–28.3 (19.5) Texas II (Binder A) 64-22 64.4–24.4 (22.5) Texas II (Binder V) 64-22 64.2–24.0 (19.5) *Values in parentheses are intermediate grades. from the seven field sites for the Phase II experiment. In total, 352 cores and 160 PMPC specimens were tested in this project. For LMLC specimens, the total time between fabrication and completion of testing or the beginning of LTOA was approximately 2 weeks. After LTOA of LMLC specimens, test- ing was completed within an approximately 2-week period. This time frame was also applicable to the cores and PMPC specimens when cores and specimens from one field site (which included three to four mixtures) at a time arrived at the labo- ratory. When cores and specimens from more than one field site arrived at the same time or if other delays occurred due to equipment availability or testing schedules, these specimens were stored in a controlled-temperature room at 68°F (20°C) and tested within 2 to 5 months. Since more than one laboratory prepared and tested speci- mens, round robin testing was conducted to ensure consis- tency, repeatability, and reproducibility of the laboratory test results. The round robin testing was limited to MR, E*, and flow number tests, none of which are currently assessed under AASHTO Materials Reference Library (AMRL) accreditation. The HWTT was not included, as all partici- pating laboratories are AMRL accredited for this test. The round robin testing included (1) compaction and testing of specimens using loose mix prepared by the University of California–Davis (UC Davis) and (2) testing of specimens prepared and compacted at UC Davis and sent to the other two participating laboratories. Testing of specimens prepared from supplied loose mix assessed consistency, repeatabil- ity, and reproducibility between participating laboratories of mixing, compacting (gyratory), sawing/trimming, AV content testing, and actual performance-related testing. Test- ing of specimens prepared at UC Davis and sent to parti- cipating laboratories assessed performance-related testing only. Details of the round robin study can be found in Appendix D.