Evaluation of the Moisture Susceptibility of WMA Technologies (2014)

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.

16 Field Projects The following factors were considered in selecting field proj- ects (including a wide spectrum of materials and field condi- tions in this study): climate (wet and dry, freeze and no-freeze), aggregate type, binder type, inclusion of recycled materials (RAP or recycled asphalt shingles [RAS]), and WMA tech- nology. Materials and cores from four field sections in Iowa, Texas, Montana, and New Mexico were selected based on these considerations. During construction, raw materials and loose plant mix were acquired on site, conditioned and compacted, and evaluated based on the selected performance parameters. PMFC cores were obtained at all four field projects at con- struction and after 6 months and 12 months in service from the Iowa field project, after 8 months in service from the Texas field project, and after 6 months in service from the Montana field project. The three field projects where performance was monitored with time by taking PMFC cores represent the three extreme climates for moisture susceptibility as follows: • Iowa = wet and freeze-thaw (F/T) = high rainfall with some F/T cycles = northern and northeastern states that may be susceptible after 1,000 days or 2–3 years. • Texas = hot and wet = high temperatures, high rainfall, and high relative humidity = southeastern states that may be susceptible after 1,000 days or 2–3 years. • Montana = cold and multi-F/T = low temperatures, some rainfall, and multiple F/T cycles = intermountain western states that may be susceptible after 100 days or 3–4 months. In addition to these three extreme climates for moisture sus- ceptibility, a risk to mixture durability and performance is posed by late-season construction in almost all United States climates (where mixtures may be susceptible after 100 days), the use of aggregates prone to moisture damage, and entrap- ment of moisture beneath an impenetrable surface layer or treatment. The four field projects are summarized in Table 2-1 and introduced in the following subsections with additional details provided in Appendix D. Climate data, including cumulative plots of degree days (base 32°F [0°C]), freez- ing days, and wet days and corresponding coring dates throughout March 2013, are summarized in Figures 2-1, 2-2, and 2-3. As shown in Figure 2-1, for cumulative degree days (base 32°F [0°C]), aging over 8 months that included a summer in the hot/wet Texas climate was similar to aging over 12 months that included a summer in the wet/F/T Iowa climate. Aging after 6 months over the winters in the wet/F/T Iowa climate and the cold/multi-F/T Montana cli- mate were also similar in terms of this climatic parameter. In terms of cumulative freezing days, as shown in Figure 2-2, the cold/multi-F/T Montana climate was significantly more severe than the wet/F/T Iowa climate, and the hot/wet Texas climate experienced almost no freezing days. Oppo- site trends are shown in Figure 2-3 in terms of cumulative wet days, with the hot/wet Texas climate showing the most precipitation followed by the wet/F/T Iowa climate and the essentially dry cold/multi-F/T Montana climate. Even though performance was not monitored with time by taking PMFC cores for the New Mexico field project, the climate is dry like Montana, cold during the winter like Iowa, and rela- tively hot during the summer with cumulative degree days between Texas and Iowa, as shown in Figures 2-3, 2-2, and 2-1, respectively. In addition to the climate data, traffic data were also esti- mated in terms of cumulative equivalent single-axle loads (ESALs) throughout the project, as shown in Figure 2-4 with corresponding coring dates. These estimated cumulative ESALs were determined based on inputs of 2011 annual average daily traffic (AADT), truck percentage (assumed constant), and annual growth rate (used to calculate assumed constant compound monthly growth rate) for each field project and assumed 50-percent directional factor and route type (Major Mixed Truck Route [Type I]) (Titus-Glover et al. 2010). Annual C H A P T E R 2 Research Approach

17 growth rates were not available for Texas and New Mexico, so assumed values of 2.5 percent and 0 percent, respec- tively, were used. The New Mexico assumed annual growth rate was based on decreasing AADT counts from the New Mexico DOT. As shown in Figure 2-4, there is a significant difference in the traffic between those field projects on interstate highways (Montana and New Mexico) and those on other types of facilities (US highway in Iowa or busy FM road in Texas). Iowa Field Project The Iowa field project was in Union and Adams Counties on US Route 34. A quartzite aggregate, two limestone aggregates, Figure 2-1. Summary of cumulative degree days (base 32F) for field projects. 0 5000 10000 15000 20000 25000 0 2 4 6 8 10 12 14 16 18 20 C um u la tiv e D eg re e D ay s (°F -d ay s) Field In-Service Times (months) Iowa Texas Montana New Mexico PMFC cores at construction PMFC cores after summer at 8 months PMFC cores after summer at 12 months PMFC cores after winter at 6 months Location and Environment Condition Location Construction Completion Date Mixtures Aggregates Asphalt Binder Additives Field Compaction Temperature (°F) Coring Dates RAP RAS Anti-Strip Agent Iowa (Wet, Freeze) US 34, near Corning Sep. 2011 HMA+RAP Quartzite, Limestone, Field Sand PG 58-28 17% None None 295-300 Sep. 2011 Mar. 2012 Sep. 2012 Evotherm 3G+RAP 240-248 Sasobit+RAP 235-240 Montana (Dry, Freeze) IH 15, near Dillon Oct. 2011 HMA Siliceous Modified PG 70-28 None None 1.4% Lime 310-315 Oct. 2011 Apr. 2012 Evotherm 3G 270-280 Sasobit 275-280 Foaming 270-275 Texas (Wet, No- Freeze) FM 973, near Austin Jan. 2012 HMA Limestone, Field Sand Modified PG 70-22 None None None 275-285 Jan. 2012 Sep. 2012 Evotherm DAT 230-235 Foaming 240-250 New Mexico (Dry, No- Freeze) IH 25, near Truth or Consequences Oct. 2012 HMA+RAP Siliceous Gravel Modified PG 64-28 35% None 1% Versabind 285-290 Oct. 2012 Evotherm 3G+RAP 255-260 Foaming+RAP 265-270 Table 2-1. Summary of field projects.

18 0.0E+00 5.0E+04 1.0E+05 1.5E+05 2.0E+05 2.5E+05 3.0E+05 3.5E+05 4.0E+05 0 2 4 6 8 10 12 14 16 18 20 C um ul at iv e E SA Ls Field In-Service Times (months) Iowa Texas Montana New Mexico Figure 2-4. Summary of traffic information for field projects. 0 50 100 150 200 250 300 350 0 2 4 6 8 10 12 14 16 18 20 C um ul at iv e F re ez in g D ay s ( da ys ) Field In-Service Times (months) Iowa Texas Montana New Mexico Figure 2-2. Summary of cumulative freezing days for field projects. 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 14 16 18 20 C um u la tiv e Pr ec ip ita tio n (in ch es ) Iowa Montana New Mexico Field In-Service Times (months) Texas Figure 2-3. Summary of cumulative wet days for field projects.

19 and the optimum binder content was determined as 4.6 percent (by weight of the total mixture). Evotherm® 3G, Sasobit®, and the Madsen Eco-Foam II foaming process were used as WMA technologies in this field project. The compaction temperatures of WMA used in the Montana field project were significantly higher than those in the Iowa and Texas field projects. Thus, off- site PMLC specimens were fabricated following the condition- ing protocol proposed based on resilient modulus (MR) data from the Iowa and Texas field projects and were tested using MR to validate the laboratory conditioning protocol. The construction of the pavements was completed in Octo- ber 2011, and PMFC cores at construction and after 6 months in service were obtained from this field project. Climate data, including cumulative plots of degree days (base 32°F [0°C]), freezing days, and wet days and corresponding coring dates, are summarized in Figures 2-1, 2-2, and 2-3, respectively. This field project represents the cold and multi-F/T extreme cli- mate for moisture susceptibility. Texas Field Project The Texas field project was on FM 973, near the Austin Bergstrom International Airport. Four different fractions of a limestone aggregate and a field sand were used and com- bined. The gradation of combined aggregate is presented in Figure 2-5. A washed sieve analysis was also conducted to verify the gradation of the combined aggregates, and two tri- als were again used to adjust the gradation of the combined aggregates. A PG 70-22 binder with a specific gravity of 1.033 was used in this project, and the optimum binder content was determined as 5.2 percent (by weight of the total mixture). Evotherm DAT™ and a foaming process were used as WMA technologies in this field project. Evotherm DAT™ has been designed to enhance coating, adhesion, and workability at lower production temperatures. In order to treat the binder with this chemical additive, the binder was heated to the and field sand and RAP were used and combined. The grada- tion of the combined aggregate is presented in Figure 2-5. A washed sieve analysis was also conducted to verify the gradation of the combined aggregates, and two trials were again used to adjust the gradation of the combined aggregates. The asphalt binder used in this project was a PG 58-28 binder with a specific gravity of 1.0284. The optimum binder content was determined as 5.4 percent (by weight of the total mixture). Evotherm® 3G and Sasobit® were selected as the WMA tech- nologies for this project. Evotherm® 3G is a combination of surfactants, waxes, processing aids, polymers, acids, and other materials intended to reduce frictional forces between the binder and aggregate. Sasobit® is a crystalline, long chain ali- phatic polymethylene hydrocarbon, identical to paraffin waxes that are found in crude oil, except that it has a higher molecular weight. Given its ability to lower the viscosity of the binder at high temperatures, Sasobit® may improve the binder flow dur- ing the mixing process and laydown operations. Both WMA additives were blended at 0.4 percent by weight of binder at the plant. The construction of the pavements was completed in Sep- tember 2011, and PMFC cores at construction, after 6 months in service, and after 12 months in service were obtained from this field project. Climate data, including cumulative plots of degree days (base 32°F [0°C]), freezing days, and wet days and corresponding coring dates are summarized in Figures 2-1, 2-2, and 2-3, respectively. This field project represents the wet and F/T extreme climate for moisture susceptibility. Montana Field Project The Montana field project was on IH 15, near the Idaho bor- der. Three different fractions of a siliceous gravel aggregate and 1.4 percent lime were used and combined. The gradation of the combined aggregate is presented in Figure 2-5. A PG 70-28 binder with a specific gravity of 1.034 was used in this project, 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% #200 #100 #50 #30 #16 #8 #4 3/8" 1/2" 3/4" 1" Iowa+RAP Texas Montana Sieve Size New Mexico+RAP C um ul at iv e R et ai ne d (% ) Figure 2-5. Aggregate gradations from field projects.

20 RAP, a PG 64-28 binder with a specific gravity of 1.02 was used, while a PG 76-28 binder with a specific gravity of 1.00 was used for the HMA without RAP, and the total binder content was determined as 5.4% (by weight of the total mixture). The construction of the pavements was completed in Octo- ber 2012, and only PMFC cores at construction were obtained from this field project. Climate data including cumulative plots of degree days (base 32°F [0°C]), freezing days, and wet days are summarized in Figures 2-1, 2-2, and 2-3, respectively. Summary of Compaction Temperatures Used in the Field Projects Compaction temperatures used in the Iowa, Texas, Montana, and New Mexico field projects are summarized in Table 2-2. Laboratory Tests and Specimen Fabrication Laboratory Tests Based on previous experience in evaluating asphalt mix- ture stiffness and moisture susceptibility in the laboratory, one nondestructive test and two destructive tests were selected to quantify the mixture stiffness in dry and wet conditions and the loss of strength and stiffness after moisture conditioning. The destructive tests were (1) determination of indirect tensile (IDT) strength in dry conditions and after moisture condition- ing to determine TSR and (2) the Hamburg Wheel-Tracking Test (HWTT) that indicates mixture resistance to both mois- ture susceptibility and rutting. The nondestructive test was mixing temperature (Tm) and the additive was blended at 5 percent by weight of binder. Foamed binder was produced on site by injecting 5-percent water and air into the hot binder inside a special expansion chamber. In the laboratory, a foam- ing device that simulates the air-atomized mixing at the plant was used to produce foamed binder/mixtures with 5% water, as shown in Figure 2-6. The construction of the pavements was completed in Janu- ary 2012, and PMFC cores at construction and after 8 months in service were obtained from this field project. Climate data, including cumulative plots of degree days (base 32°F [0°C]), freezing days, and wet days and corresponding coring dates, are summarized in Figures 2-1, 2-2, and 2-3, respectively. This field project represents the hot and wet extreme climate for moisture susceptibility. New Mexico Field Project The New Mexico field project was on IH 25, in Sierra County. Three fractions of a siliceous gravel aggregate and 1% Versabind (a low-grade Portland cement) were used and combined. The gradation of the combined aggregate is presented in Figure 2-5. A washed sieve analysis was also conducted to verify the gra- dation of the combined aggregates, and three trials were used to adjust the gradation of the combined aggregates. Evotherm® 3G and a foaming process were used as WMA technologies in this field project. Thirty-five percent of RAP was included in the mixture for the control HMA and WMA. In addition, another control HMA without RAP was constructed using the same aggregates to discriminate the effect of recycled materials on mixture performance. For those mixtures with Figure 2-6. Laboratory foaming process.

21 tures. IDT strength at 77°F (25°C) was determined for both dry specimens and for wet specimens moisture conditioned according to AASHTO T 283 with partial vacuum saturation, one freeze-thaw cycle, and soaking in warm water, as shown in Figure 2-7. All laboratory-compacted specimens were fab- ricated to a diameter of 6 inches (150 mm) and a height of 3.75 inches (95 mm) in the Superpave gyratory compactor to target air void contents of 7±0.5%. In this project, the TSR was determined as the ratio of the average of three IDT strength determination of MR with testing conducted in dry conditions and after moisture conditioning to determine MR-ratio. IDT Strength and TSR IDT strength in both wet and dry conditions and the result- ing TSR of wet-to-dry IDT strengths was selected as one of the destructive tests given that it is the most common national standard test to evaluate moisture susceptibility of asphalt mix- Location and Environmental Condition Mixture Type Specimen Type PMFC (°F) Onsite PMLC 0-1 h (°F) Onsite PMLC 1-2 h (°F) LMLC (°F) Offsite PMLC (°F) Iowa (Wet, Freeze) HMA+RAP 295-300 N/A 295-300 295 295 Evotherm 3G+RAP 240-248 N/A 240-248 240 240 Sasobit+RAP 235-240 N/A 235-240 240 240 Montana (Dry, Freeze) HMA 310-315 N/A 315 N/A 275 Evotherm 3G 270-280 N/A 275 N/A 240 Sasobit 275-280 N/A 279 N/A 240 Foaming 270-275 N/A 271 N/A 275 Texas (Wet, No- Freeze) HMA 270-285 275 275 275 275 Evotherm DAT 230-235 225 225 240 240 Foaming 240-250 225 250 235 275 New Mexico (Dry, No- Freeze) HMA+RAP 285-290 N/A 295 275 275 Evotherm 3G+RAP 255-260 N/A 275 240 240 Foaming+RAP 265-270 N/A 275 240 275 HMA 330-335 N/A 320 275 275 Table 2-2. Summary of compaction temperatures (Tc) from field projects. (Santucci, 2010) Figure 2-7. Modified lottman test by AASHTO T 283.

22 In this project, MR stiffness at 77°F (25°C) was measured by ASTM D7369 with the modification of replacing on-specimen linear variable differential transducers (LVDTs) with LVDTs aligned along the horizontal diametral plane (gauge length as a fraction of diameter of the specimen = 1.00) to reduce costs, as shown in Figure 2-8. For each specimen, MR stiffness was measured twice, rotating the specimen 90 degrees after the first measurement. MR stiffness was first determined for dry specimens, and then these same specimens were moisture- conditioned according to AASHTO T 283 with partial vacuum saturation, one freeze-thaw cycle, and soaking in warm water, as shown in Figure 2-8, and tested again to determine MR stiffness for wet specimens. All laboratory-compacted specimens were fabricated to a diameter of 6 inches (150 mm) and a height of 2.4 inches (61 mm) in the Superpave gyratory compactor to target air void contents of 7±0.5%. In this project, the MR-ratio was determined as the ratio of the average of three MR stiffness results obtained from three specimens tested in wet condition to the average of three MR stiffness results obtained from three specimens tested in dry condition. The MR-ratio values and the wet MR stiffnesses were considered in this project to com- pare WMA and HMA in terms of moisture susceptibility. A precision and bias statement was not available for the MR-ratio, because although the wet and dry stiffness measurements were conducted on the same specimen, the d2s value was not likely any larger than that for TSR (9.3%) (Azari 2010). Therefore, a 10% difference was considered for this project for identifying significant differences between mixture types and specimen types for the MR-ratio where only one replicate value was pro- duced from each set of six specimens. HWTT The HWTT by AASHTO T 324 was selected as the other destructive test because of its recent adoption by several states results obtained from three specimens tested in wet condition to the average of three IDT strength results obtained from three specimens tested in dry condition. The TSR values and the wet IDT strengths were considered in this project to compare WMA and HMA in terms of moisture susceptibility. As only one rep- licate TSR value was produced from each set of six specimens, the TSR results for different mixture types or different speci- men types were compared to each other based on the preci- sion and bias statement that indicates a d2s acceptable range of two results with more than a 95% confidence level of 9.3% (Azari 2010). In AASHTO M 323, the threshold for TSR by AASHTO T 283 is a minimum of 0.80, or 80%. Some agencies also specify a minimum value of dry, wet, or both IDT strength val- ues in addition to or instead of a limit on the TSR. Some of these minimums include • Nevada: – Unmodified binder: 60 psi for dry IDT strength (48 psi wet IDT strength assuming TSR ≥ 80%). – Modified binders: 90 psi for dry IDT strength (72 psi wet IDT strength assuming TSR ≥ 80%). • Tennessee: – Unmodified binder: 80 psi for wet IDT strength. – Modified binders: 100 psi for wet IDT strength. • Texas: – 85 psi for dry IDT strength (68 psi wet IDT strength assuming TSR ≥ 80%). Resilient Modulus (MR) and MR-ratio MR stiffness in both wet and dry conditions and the resulting MR-ratio of wet-to-dry MR stiffnesses was selected as the non- destructive test given its cost effectiveness in providing an accu- rate indicator of moisture susceptibility in terms of stiffness. (a) Data Acquisition System (b) LVDT Setup (c) Loading Frame Setup Figure 2-8. MR test equipment.

23 Current specifications when available for the HWTT for the states where the field projects were located were as follows: • Iowa: – Water temperature during test: 122°F (50°C). – Minimum SIP of 10,000 or 14,000 load cycles depend- ing on ESAL level (i.e., <3M or ≥3M, respectively). • Texas: – Water temperature during test: 122°F (50°C). – Variable cycles to failure depending on the binder per- formance grade (PG 64 or lower, 10,000 load cycles; PG 70, 15,000 load cycles; PG 76, 20,000 load cycles). – 0.5 inch (12.5 mm) max allowable rut depth. • Montana: – Water temperature depending on the binder performance grade (14°C lower than the high-temperature perfor- mance grade). – 20,000 cycles to failure; 0.51 inch (13 mm) max allowable rut depth. For all of these states except Iowa, the HWTT specifications focus on limiting rutting, not moisture susceptibility. Specimen Fabrication To fabricate LMLC specimens, aggregates and binder were heated to the specified mixing temperatures independently and then mixed with a portable mixer. Afterwards, HMA and WMA loose mixes were conditioned (1) in the oven with vari- ous protocols for the WMA laboratory-conditioning experi- ment and (2) for 2 h at 275°F (135°C) and 240°F (116°C) for HMA and WMA, respectively, for the WMA moisture- susceptibility (including the effects of anti-stripping agents) and WMA performance-evolution experiments. Specimens were then compacted with the Superpave gyratory compactor (SGC) at the compaction temperatures shown in Table 2-3. Trial speci- mens were fabricated to ensure specimens were obtained with air void contents of 7±0.5%. To simulate field aging in early life in the WMA performance-evolution experiment, compacted specimens were further conditioned following various aging protocols in an environmental room or oven prior to being tested. In total, almost 500 LMLC specimens with 7±0.5% AV were fabricated for the Iowa, Texas, Montana, and New Mexico field projects that included 13 mixtures (4 HMA and 9 WMA). PMFC cores were obtained at construction for the Iowa, Texas, Montana, and New Mexico field projects. Addition- ally, PMFC cores after 6 months and 12 months in service from the Iowa field project, after 8 months in service from the Texas field project, and after 6 months in service from the Montana field project were also acquired. To fabricate onsite PMLC specimens, loose mixes were taken from the trucks before leaving the plant and maintained in the oven for 1–2 h to simultaneously evaluate rutting and moisture susceptibil- ity of asphalt mixtures. The HWTT was conducted at 122°F (50°C), and the stripping inflection point (SIP) and strip- ping slope, as shown in Figure 2-9, were calculated to com- pare WMA and HMA in terms of moisture susceptibility. All laboratory-compacted specimens were fabricated to a diam- eter of 6 inches (150 mm) and a height of 2.4 inches (61 mm) in the Superpave gyratory compactor to target air void con- tents of 7±0.5%, and the cylindrical specimens were tested as shown in Figure 2-9 for a maximum of 20,000 passes or until 0.5 inch (12.5 mm) of deformation occurred. Because a precision and bias statement was not available for the selected HWTT test results, the average differences in SIP and the stripping slope for all Texas mixtures that exhib- ited stripping were calculated as approximately 2,000 load cycles and 0.2 µm/cycle, respectively, for use as correspond- ing d2s values in this analysis. Thus, these thresholds were used for identifying significant differences between mixture types and specimen types for these performance parameters. Figure 2-9. Hamburg wheel-tracking test. (a) Equipment with Loaded Specimens (b) Typical Deformation Behavior with Load Cycles

24 schedules are encountered, storage of these specimens at cold temperatures (≤ 68°F [20°C]) is proposed for delays that can stretch from 1 to 4 months. For the Iowa field project, speci- mens were stored at 77°F (25°C), and average increases of 30% in dry MR stiffness were noted. For the Montana field project, specimens were stored at 68°F (20°C), and dry MR stiffnesses did not change. Experiment Designs WMA Laboratory Conditioning The goal of the WMA laboratory-conditioning experiment was to propose conditioning protocols consisting of a com- bination of time and temperature that produce WMA LMLC and offsite PMLC specimens calibrated to PMFC field cores. Figure 10 presents the research method used for this experi- ment. In this experiment, LMLC and offsite PMLC specimens with different laboratory-conditioning protocols, PMFC at the temperature shown in Table 2-2 prior to compaction. In total, more than 250 PMFC cores and more than 150 onsite PMLC specimens from the Iowa, Texas, Montana, and New Mexico field projects were tested in this project. To fabricate offsite PMLC specimens, loose mixes were transported to the laboratory in buckets that were then reheated in an oven to the specified conditioning temperature prior to compaction. In total, almost 250 offsite PMLC specimens were fabricated from the four field projects. For LMLC and offsite PMLC specimens, the total time between fabrication and completion of testing or the begin- ning of LTOA was approximately 2 weeks. After LTOA of LMLC specimens, testing was also completed within an approximately 2-week period. Such timeframes are also pos- sible for onsite PMLC specimens and PMFC cores when one field project (that includes 3 to 4 mixtures) at a time is arriv- ing at the laboratory. However, when these types of speci- mens are arriving from more than one field project or other unavoidable delays due to equipment availability or testing PMFC Cores @ Construction 2 h @ 275 F 4 h @ Tc 2 h @ Tc 4 h @ 275 F LMLC Specimens 2 h @ Tc + 16 h @140 F + 2 h @ Tc Onsite PMLC Specimens Reheat + 4 h @ 275 F Reheat + 2 h @ Tc Reheat to Tc Offsite 16 h @ 140 F + Reheat + 2 h @ Tc Dry MR Dry MR Binder G* & Mixture Anisotropy HMA Iowa Evotherm Sasobit HMA Texas Evotherm Foaming Note: Tc: compaction temperature. Figure 2-10. Flowchart for WMA laboratory-conditioning experiment. Location and Environmental Condition Mixture Type Laboratory-Conditioning Protocols 2 h @ Tc 2 h @ 275°F 4 h @ Tc 2 h @ Tc + 16 h @ 140°F + 2 h @ Tc 4 h @ 275°F Iowa (Wet, Freeze) HMA+RAP X X X X X Evotherm 3G+RAP X X X X X Sasobit+RAP X X X X X Texas (Wet, No-Freeze) HMA X X X X X Evotherm DAT X X X X X Foaming X X X X X Table 2-3. WMA laboratory-conditioning test plan for LMLC specimens.

25 were also used to prepare offsite PMLC specimens. The lab- oratory conditioning protocol for offsite PMLC specimens was proposed based on MR data from the Iowa and Texas field projects. Because the Montana compaction temperatures for both HMA and WMA were significantly higher than those for Iowa and Texas, offsite PMLC specimens from the Mon- tana field project were fabricated following the proposed protocol, as well as one consisting of the same condition- ing time at Tc, and tested with MR to validate the proposed protocol. Field cores at construction and onsite PMLC specimens were expected to have similar stiffnesses as they experienced approximately the same level of binder aging. However, their performance in MR tests was significantly different, as described subsequently, and thus binder was extracted and recovered from these specimens to measure the difference in binder stiffness with the dynamic shear rheometer (DSR). In addition, images were acquired from the same specimens through a novel method (Zhang et al., 2011) to evaluate the effect of aggregate orientation by different compaction methods on mixture stiffness. Finally, the effect of total AV on the stiffness of the specimens was also evaluated. WMA Moisture Susceptibility The goal of the moisture-susceptibility experiment was to evaluate moisture susceptibility of WMA in comparison with HMA; Figure 2-11 and Table 2-5 present the research method and test plan, respectively, for this experiment. In this experi- ment, all laboratory-compacted specimens (LMLC, onsite PMLC, and offsite PMLC) were tested to determine wet and dry MR stiffness and MR-ratio, HWTT SIP and stripping slope, and wet and dry IDT strengths and TSR. PMFC cores were also evaluated in terms of all of these same moisture-susceptibility cores, and onsite PMLC specimens were tested to determine dry stiffness in terms of MR, as well as compactability in terms of number of SGC gyrations (N) required in specimen fabri- cation to achieve 7±0.5% AV. Analysis of variance (ANOVA) and Tukey-Kramer Hon- estly Significant Differences (HSD) tests were conducted at a 5% significance level to compare the conditioned LMLC and PMLC specimens (both on site and off site) with the PMFC cores for each mixture type and each selected per- formance parameter, while accounting for the variability in the MR stiffness results. Initially, in addition to the main factor of interest conditioning protocol, the effect of ori- entation (i.e., rotating the specimen 90 degrees after the first measurement) as well as the interaction effect between orientation and conditioning protocol was also tested by using a more sophisticated ANOVA analysis (a split plot design analysis). As shown in Tables 2-3 and 2-4, five different conditioning protocols were selected for LMLC specimens prior to com- paction, and four different ones were applied to offsite PMLC specimens after reheating to the specified conditioning temper- ature. For LMLC specimens, the conditioning protocol of 2 h at Tc was used because it was proposed by the recently completed NCHRP Project 9-43, and 4 h at 275°F (135°C) was proposed because it is the current standard in the state of Texas. The comprehensive conditioning protocol of 2 h at Tc followed by 16 h at 140°F (60°C) and 2 h at Tc was proposed during a WMA workshop (Harrigan, 2012b) held in May 2011, in Irvine, California. The other two protocols used were derived by combining common conditioning temperatures and times. For offsite PMLC specimens, the conditioning protocol of reheating to Tc was proposed as the least amount of con- ditioning time/temperature possible prior to compaction. Additionally, three protocols proposed for LMLC specimens Location and Environmental Condition Mixture Type Laboratory-Conditioning Protocols R @ Tc R + 2 h @ Tc R + 16 h @ 140°F + 2 h @ Tc R + 4 h @ 275°F Iowa (Wet, Freeze) HMA+RAP X X X X Evotherm 3G+RAP X X X X Sasobit+RAP X X X X Texas (Wet, No-Freeze) HMA X X X X Evotherm DAT X X X X Foaming X X X X Montana* (Dry, Freeze) HMA X Evotherm 3G X Sasobit X Foaming X Note: R: reheat. *Also included proposed protocol of reheating to 240°F for WMA (except Foaming) and 275°F for HMA and WMA Foaming. Table 2-4. WMA laboratory-conditioning test plan for offsite PMLC specimens.

26 moisture-susceptibility performance for the same specimen type while accounting for the variability in those tests with multiple replicates. For those tests without multiple replicates, d2s values for the acceptable range of two results or similar values defined based on data from this project as described previously were used in the comparisons. parameters, except wet MR stiffness and MR-ratio. Offsite PMLC and LMLC specimens for moisture testing were fabricated to mimic the early-life behavior of the mixture, based on the results of the laboratory-conditioning experiment. ANOVA and Tukey’s HSD tests were conducted at a 5% significance level to compare WMA with HMA in terms of Wet MR TSR MR-ratio Wet IDT Strength PMLCLMLCPMFC @Construction After X months in service HWTT-SIP & Stripping Slope TSR Wet IDT Strength HWTT-SIP & Stripping Slope Onsite Offsite HMA Iowa Evotherm Sasobit HMA Montana Evotherm Foaming Sasobit HMA Texas Evotherm Foaming HMA New Mexico Evotherm Foaming Figure 2-11. Flowchart for WMA moisture-susceptibility experiment. WMA Field Project Mixture Type LMLC As Designed Onsite & Offsite PMLC Cores at Construction Cores after Winter (6 months IA, MT) Cores after Summer (12 months IA, 8 months TX) Dry MR Wet MR TSR HWTT Dry MR Wet MR TS R HWTT Dry MR TSR HWTT Dry MR TSR HWTT Dry MR TSR HWTT Iowa US 34 (Wet, Freeze) Wet/F/T Evotherm 3G +RAP X X X X X X X X X X X X X X X X X Sasobit +RAP X X X X X X X X X X X X X X X X X HMA +RAP X X X X X X X X X X X X X X X X X Montana IH 15 (Dry, Freeze) Cold/multi- F/T Sasobit - - - - X X X X X X X X X X - - - Evotherm 3G - - - - X X X X X X X X X X - - - Foaming - - - - X X X X X X X X X X - - - HMA - - - - X X X X X X X X X X - - - Texas FM 973 (Wet, No- Freeze) Hot/Wet Evotherm DAT X X X X X X X X X X X - - - X X X Foaming X X X X X X X X X X X - - - X X X HMA X X X X X X X X X X X - - - X X X New Mexico IH 25 (Dry, No- Freeze) Hot/Dry Evotherm 3G +RAP X X X X X X X X X X X - - - - - - Foaming +RAP X X X X X X X X X X X - - - - - - HMA +RAP X X X X X X X X X X X - - - - - - Table 2-5. WMA moisture-susceptibility test plan.

27 ANOVA and Tukey’s HSD tests were conducted at a 5% significance level to compare WMA with HMA in terms of moisture-susceptibility performance for the same specimen type while accounting for the variability in those tests with multiple replicates. For those tests without multiple replicates, d2s values for the acceptable range of two results or similar values defined based on data from this project as described previously were used in the comparisons. Effect of Specimen Type Using the same data as the WMA moisture-susceptibility experiment where WMA was compared with HMA for each specimen type, different specimen types were also compared for each mixture type to examine important differences in (1) LMLC specimens used in mix design, (2) PMLC speci- mens used in QA, (3) PMLC specimens reheated for offsite compaction, and (4) field performance as determined by laboratory testing of PMFC cores. Data used in this analy- sis were the same as those collected for the WMA moisture- susceptibility experiment but regrouped and reanalyzed for a different comparison. Effect of Anti-Stripping Agents To evaluate the use of anti-stripping agents as an aid to improve WMA moisture susceptibility, hydrated lime and a common liquid anti-stripping agent were added to Texas and Iowa HMA and WMA in LMLC specimens in a separate experiment. Figure 2-12 presents the method used. Hydrated lime was added through the dry process in a proportion of 1% by the weight of dry aggregates, removing that same 1% of material passing the No. 200 sieve to preserve the grada- tion of the mix design. A liquid anti-stripping (LAS) agent was added at 0.5% by weight of binder by blending with the binder prior to mixing with the aggregates. In total, an additional 108 LMLC specimens were fabricated to compare with the LMLC specimens designed and used in the WMA moisture-susceptibility experiment. The laboratory test plan for this anti-stripping experiment is shown in Table 2-6. In this experiment, to assess the effectiveness of the hydrated lime and the LAS agent, dry and wet IDT strength and TSR and dry and wet MR and MR-ratio were measured. Moisture con- ditioning for all wet specimens was done following AASHTO T 283 with one freeze-thaw cycle. LMLC Wet MR TSR MR-ratio Wet IDT Strength As Designed + Liquid Anti- Stripping Agent (0.5% by Binder) + Lime (1% by Mixed) HMA Iowa Evotherm Sasobit HMA Texas Evotherm Foaming Figure 2-12. Flowchart for anti-stripping agent experiment. WMA Field Project Mixture Type As Designed + Hydrated Lime + Liquid Anti- Stripping Agent MR-ratio TSR MR-ratio TSR MR-ratio TSR Iowa US 34 (Wet, Freeze) Wet/F/T Evotherm 3G+RAP X X X X X X Sasobit+RAP X X X X X X HMA+RAP X X X X X X Texas FM 973 (Wet, No-Freeze) Hot/Wet Evotherm DAT X X X X X X Foaming X X X X X X HMA X X X X X X Table 2-6. Anti-stripping agent test plan for LMLC.

28 field project were tested to determine dry MR to evaluate the change in mixture stiffness with aging in the field. Addition- ally, onsite PMLC specimens from these field projects were tested to indicate the initial stiffness of HMA and WMA pavements in their early life. The same set of LMLC speci- mens from Iowa and Texas were aged at 140°F (60°C) over a series of aging periods (1 week, 2 weeks, 4 weeks, 8 weeks, and 16 weeks) prior to being tested to determine MR, with the same specimen tested repeatedly in this nondestructive test to reduce variability and specimen fabrication efforts. These aging periods were selected based on the long-term conditioning for loose mix included in AASHTO T 283 and on previous research in Texas that indicated 4 weeks (1 month) at 140°F (60°C) aged HMA mixtures to stiffnesses similar to those in HMA pavements after approximately 1 year in Texas climate conditions (Glover et al. 2005). Thus, the selected aging times might reflect 1 to 4 years under Texas conditions and likely 2 to 8 years in milder climates in the United States. Results from the first phase were used to define an aging period at which WMA reached a dry MR stiffness equivalent to that of an HMA control section. This aging time was defined as tA for use in the second phase. In the second phase, as shown in Table 2-8, HMA and WMA mixture properties were evaluated after LTOA at dif- ferent periods in terms of IDT strength in dry and wet condi- tion and TSR, dry and wet MR and MR-ratio, and HWTT SIP and stripping slope. Selected LTOA protocols included tA of 2 weeks at 140°F (60°C), as defined in the first phase of the experiment; a longer aging time, tB, of 16 weeks, also at 140°F (60°C) to represent several years in service in the field; and the standard LTOA at 185°F (85°C) for 5 days, as is included in AASHTO R 30. Materials from the Iowa, Texas, and New Mex- ico field projects were used in this phase. LMLC specimens were fabricated following the proposed laboratory-conditioning ANOVA and Tukey’s HSD tests were conducted at a 5% significance level to compare the different specimen types in terms of moisture-susceptibility performance for the same mixture type while accounting for the variabil- ity in those tests with multiple replicates. For those tests without multiple replicates, d2s values for the acceptable range of two results or similar values defined based on data from this project as described previously were used in the comparisons. WMA Performance Evolution Results from the laboratory-conditioning experiment indi- cated that the initial stiffness of the WMA is less than the stiff- ness of conventional HMA and that this gap can be reduced with increased elapsed time in the field. The goal of the WMA performance-evolution experiment was to determine the time when (or if) the properties of HMA and WMA converge and evaluate the performance of WMA as compared with HMA in the early life of the pavement. The hypothesized results in terms of the changes in HMA and WMA stiffness in the field and in the laboratory as long-term aging time increases are shown in Figures 2-13 and 2-14. Figure 2-15 presents the two-phase research method used for this experiment. In the first phase, as shown in Table 2-7, the vulnerabil- ity of WMA in terms of moisture susceptibility was deter- mined in terms of a critical age or time period to reach a dry MR stiffness equivalent to that of an HMA control sec- tion. Changes in HMA and WMA dry MR stiffness in the field and laboratory were evaluated separately, followed by the correlation of mixture aging in these two conditions. PMFC cores at construction and after 6 or 8 months in service, respectively, from the Iowa and Texas field proj- ects and those after 12 months in service from the Iowa Figure 2-13. Hypothesized evolution of field mixture stiffness with time. Figure 2-14. Hypothesized evolution of laboratory mixture stiffness with LTOA time.

29 Location and Environmental Condition Mixture Type LTOA Protocols @140°F 1 week 2 weeks 4 weeks 8 weeks 16 weeks Iowa US 34 (Wet, Freeze) HMA+RAP*† X X X X X Evotherm 3G+RAP*† X X X X X Sasobit+RAP*† X X X X X Texas FM 973 (Wet, No-Freeze) HMA* X X X X X Evotherm DAT* X X X X X Foaming* X X X X X * Onsite PMLC specimens and PMFC cores at construction and after 6 or 8 months in service were tested for dry MR. † PMFC cores after 1 year in service were tested for dry MR. Table 2-7. Phase one of the WMA performance evolution test plan for dry MR tests of LMLC specimens. Location and Environmental Condition Mixture Type LMLC Specimens tA = 2 weeks @ 140°F tB = 16 weeks @ 140°F* 5 days @ 185°F Iowa US 34 (Wet, Freeze) HMA+RAP†‡ - X - Evotherm 3G+RAP†‡ - X - Sasobit+RAP†‡ - X - Texas FM 973 (Wet, No-Freeze) HMA† X X X Evotherm DAT† X X X Foaming Process† X X X New Mexico IH 25 (Dry, No-Freeze) HMA+RAP† X - X Evotherm 3G+RAP† X - X Foaming+RAP† X - X * HWTT test was not performed. † PMFC cores at construction were tested for dry/wet MR, HWTT, and dry/wet IDT tests. ‡ PMFC cores after winter at 6 months in service were tested for dry MR, HWTT, and dry/wet IDT tests. PMFC cores after summer at 8 months in service were tested for dry MR, HWTT, and dry/wet IDT tests. Table 2-8. Phase two of the WMA performance evolution test plan for MR, HWTT, and IDT strength tests of LMLC specimens. Onsite PMLC LMLC PMFC 1st Step IA & TX LMLC 1 weeks @ 60°C 2 weeks @ 60°C 4 weeks @ 60°C 8 weeks @ 60°C Dry MR @ Construction After winter at 6 months (IA) After summer at 8 months (TX) and 12 months (IA) 2 weeks @ 60°C 16 weeks @ 60°C 5 days @ 85°C 2nd Step IA, TX, & NM Dry/Wet IDT TSR 16 weeks @ 60°C Dry/Wet MR MR-ratio HWTT SIP & Stripping Slope Dry/Wet IDT & TSR Dry/Wet MR & MR-ratio No LTOA No LTOA HMA Iowa Evotherm Sasobit HMA Texas Evotherm Foaming HMA New Mexico Evotherm Foaming Figure 2-15. Flowchart for WMA performance evolution experiment.

30 In both phases, ANOVA and Tukey’s HSD tests were con- ducted at a 5% significance level to compare WMA with HMA in terms of moisture-susceptibility performance for the same specimen type while accounting for the variability in those tests with multiple replicates. For those tests without multiple repli- cates, d2s values for the acceptable range of two results or simi- lar values defined based on data from this project as described previously were used in the comparisons. protocol prior to compaction defined in the laboratory- conditioning experiment and then long-term aged by the selected protocols after compaction and prior to testing to determine wet and dry MR stiffness and MR-ratio, HWTT rut depth at a specific number of load cycles, HWTT SIP, and wet and dry IDT strength and TSR. These moisture- susceptibility parameters were then used to compare WMA and HMA in their early lives.