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.
174 Evaluating the Activation of Asphalt Binder from Recycled Asphalt Shingles in Asphalt Concrete Abstract The use of recycled asphalt shingles (RAS) to replace a por- tion of the virgin asphalt binder in asphalt mixtures has been increasing over the past several years because of environmen- tal and economic motivations. Yet, since shingle asphalts are much stiffer than paving asphalts, there are concerns that the shingle asphalt may not be completely activated as a binding material. This could result in under-asphalted mixtures; also, the composite binder created by the blended shingle asphalt and virgin asphalt may not have suitable characteristics to resist cracking under load and environmental conditions in pavements. In this study, the activation of shingle asphalt was inves- tigated and estimated using three experiments. The first experiment involved mixing heated aggregate and ambient- temperature RAS without any additional asphalt to determine if any RAS binder was transferred to the virgin aggregate. The second experiment evaluated the compactability of laboratory-prepared mixtures containing 5% RAS at mixing and compaction temperatures ranging from 250Â°F to 350Â°F. A third experiment involved laboratory performance tests to evaluate the effect of mixing and compaction temperatures, the effect of separate RAS components on mixture properties, and laboratory versus plant mixtures with and without RAS. The results of this study indicate that increasing the mix- ing temperature increases the stiffness and reduces the crack- ing resistance of mixtures containing RAS. The hypothesis that higher mixing temperatures increase the activation of RAS binder was partially supported by statistical differences in mixture performance test results. However, higher mixing temperatures alone have also been shown to affect laboratory performance tests. Increasing the mixing time and/or storage time may additionally increase activation of shingle asphalt. Overall, the addition of RAS to a mixture, regardless of how well the RAS binder might be activated, appears to negatively impact cracking resistance. Introduction Since the 1980s, a small number of asphalt concrete pro- ducers have used recycled asphalt shingles (RAS) as a com- ponent in asphalt paving mixtures to replace a portion of the virgin asphalt binder and fine aggregate material. In 2003, the U.S. Environmental Protection Agency (EPA) began to tar- get construction and demolition (C&D) debris as a part of its Resource Conservation Challenge. Tipping fees at landfills were increased for C&D debris as an incentive to reduce, reuse, and recycle (U.S. Environmental Protection Agency 2003). The annual tonnage of asphalt shingles disposed into U.S. landfills each year has been estimated to be 10 million tons of PC shingles and one million tons of manufacturerâs waste shingles (Sandler 2003). According to an estimate by the EPA, asphalt shingles account for about 8% of all building-related debris and 1% to 10% of all C&D debris generated annually (Sandler 2003). With shingle waste tipping fees escalating to over $60 per ton, roofing contractors and shingle manufactur- ers have been searching for more economical ways to dispose of waste shingles (Brunswick County, North Carolina 2015; California Waste Services 2016; Dane County Department of Public Works 2015; Halifax C & D Recycling, Ltd.). In response to the Resource Conservation Challenge and the rising cost of asphalt binder through much of the past 10 years, recycling of shingles in asphalt paving mixtures has been increasing in popularity, and more state highway agen- cies are allowing its use. In 2007, 15 state highway agencies allowed the use of RAS in hot mix asphalt (HMA), 11 of these states had adopted specifications for routine use of RAS, and eight of these 11 states allowed only the use of manufacturerâs waste RAS (MWâRAS) (Krivit 2007). By 2013, 23 state high- way agencies had adopted specifications for routine use, 10 of which allowed either post-consumer RAS (PCâRAS) or MWâRAS (Willis 2013). A 2014 National Asphalt Pavement Association survey estimated that from 2009 to 2014, the amount of RAS used in HMA increased from 702,000 tons A P P E N D I X C
175 to nearly 2.0 million tons in the United States (Hansen and Copeland 2015). A number of experts in the asphalt pavement community have voiced concerns about how to account for the quantity and quality of asphalt from RAS. Asphalt used in the man- ufacture of shingles is commonly air-blown to significantly increase the asphaltâs viscosity. PCâRAS shingles, which are removed from roofs, are further oxidized from years of direct solar radiation. If the shingle asphalt is not completely acti- vated and blended when RAS is used in an asphalt mixture, the mixture will have a low total asphalt content. This could potentially lead to difficulties in placement and compaction and, ultimately, an increase in susceptibility to cracking and raveling. Also, underestimating the amount of activation and blending of the shingle asphaltâor not compensating for the higher stiffness of the shingle asphaltâcan increase the paving mixtureâs stiffness beyond its ability to relax strains. If not used properly, potential savings gained by using recycled shingles can be quickly negated by a decrease in pavement life. Objective The objective of this research was to evaluate the activation of shingle asphalt and assess the contribution of the individ- ual components within RAS on the properties and perfor- mance of asphalt concrete mixtures. Scope To meet the above objective, several laboratory experi- ments were conducted using asphalt mixtures containing 5% RAS by weight of total aggregate blend. The first experiment involved mixing heated aggregate and ambient temperature RAS without any additional asphalt. The second experiment evaluated the activation of shingle asphalt by mixing and compacting a 5% RAS mixture at temperatures ranging from 250Â°F to 350Â°F. Mixture performance testing was conducted to assess the effect of the shingle components on the perfor- mance properties of the mixtures prepared at the different temperatures. Dynamic Modulus, Texas Overlay, Indirect Tensile (IDT) Creep Compliance and Strength, and Energy Ratio testing were used to gain a better understanding of the effect of mixing temperature on the activation of shingle asphalt and the effect each shingle component had on labo- ratory performance (the experiment included separating the RAS binder from the granules and fiber). Materials and Mix Designs In 2012, NCAT began a pavement preservation experiment on a local Alabama county road (Lee County Road 159). Lee County Road 159 dead-ends into a quarry and asphalt plant, and road traffic is primarily closely monitored truck traffic. This allows for analysis of traffic data on the performance of each preservation technique throughout the duration of the experiment. The road was segmented into 25 100-ft sections, and each section was treated with a different pavement pres- ervation technique. Crack sealing, chip seals, microsurfacing, fog seals, and thin-lift asphalt overlays were all included in the preservation experiment. A thin-lift overlay mixture with 100% virgin materials constructed in Section 19 and a 5% RAS thin-lift overlay mixture constructed in Section 24 were chosen as the designs to target and modify for the purposes of this research (Willis 2013). A 50% RAP thinlay constructed in Section 23 was also sampled for comparison testing of plant mix samples. All of these sections contained the same PG 67-22 virgin binder. Four mix designs were developed for this study. The first two replicated as closely as possible the 5% RAS 4.75 mm nominal maximum aggregate size (NMAS) thin-lift over- lay and the virgin 4.75 mm NMAS thin-lift overlay used on the pavement preservation experiment on Lee County Road 159. These two mixtures were reproduced at NCATâs laboratory using the same aggregate materials and PCâRAS stockpiles as the test sections. The 5% RAS mixture was designed as 5% RAS by weight of the aggregate. The qual- ity control data of the thin-lift overlay mixtures were used as targets for the aggregate gradations of the laboratory- produced mixtures. These two mixtures were designed at a mixing temperature of 325Â°F and are identified herein as Mixture 5S and Virgin. Table C-1 shows the average gradations for the aggregate stockpiles, as well as the gradations of the RAS and RAS aggregate. The Gsb of the RAS was determined according to AASHTO PP 53-09, which recommended determining the Gse of RAS using AASHTO T 209. AASHTO PP 53-09 notes that shingle granules are not very absorptive, and therefore, Gsb and Gse should be relatively equal. The fiber content of the RAS was determined during the sieve analysis of the post- extraction aggregates by removing clumps of fiber from the sieves and weighing them separately. The extracted shingle asphalt was recovered from the trichloroethylene (TCE) sol- vent using rotary evaporation. The shingle asphalt was tested and graded as a rolling thin-film ovenâaged binder, accord- ing to AASHTO M 320-10: Standard Specification for Perfor- mance-Graded Asphalt Binder. The high temperature grade was determined to be 148Â°C and the low temperature grade was estimated to be +2Â°C. Table C-2 shows the aggregate gradations and mix for- mulas for the four mixtures designed in the laboratory and the three plant-produced mixtures. All mixtures were designed according to AASHTO M 323-13 and AASHTO R 35 for 3 to 10 million equivalent single-axle loads (ESALs) and an Ndesign of 75 gyrations. The virgin asphalt binder
176 used for these mixtures was a PG 67-22, and the optimum binder content was determined at a mixing temperature of 325Â°F and a compaction temperature of 300Â°F. Activation of 100% was assumed when designing the mixtures with shingle asphalt. Table C-3 shows the volumetric properties obtained for the 4.75 mm asphalt mixtures. The volumetric properties of the four mixtures were reasonably consistent. Experimental Plan The activation of shingle asphalt and its effects were inves- tigated using several experiments. The two main questions concerning activation are: Does the shingle asphalt soften or melt during mixing and blend with the virgin asphalt? And how much of the shingle asphalt is activated and contributes to the total asphalt content of the mixture? The following Sieve Size Percent Passing English Metric Calera LS 820s Shorter Sand EAP Baghouse Fines Hydrated Lime PCâRAS (aggregate) PCâRAS (as is) 3/8 in. 9.5 mm 100.0 100.0 100.0 100.0 100.0 99.3 No. 4 4.75 mm 96.7 99.5 100.0 100.0 99.3 81.7 No. 8 2.36 mm 68.6 91.8 100.0 100.0 97.7 70.7 No. 16 1.18 mm 42.6 72.3 100.0 100.0 82.9 52.1 No. 30 0.6 mm 28.4 42.0 100.0 100.0 59.5 31.3 No. 50 0.3 mm 19.3 14.3 99.7 99.8 49.1 19.6 No. 100 0.15 mm 13.7 4.0 98.5 98.3 39.4 9.8 No. 200 0.075 mm 10.1 1.0 93.6 97.1 27.4 2.6 Asphalt content 18% Fiber content 2%a Note: EAP = emulsified asphalt prime. aApproximately 2% by weight; approximately 5% to 10% by volume. Table C-1. Aggregate particle size distributions. Sieve Size (mm) Percent Passing Virgin Mixture 5S Mixture 5A Mixture 16B Virgin Plant QC 5% RAS Plant QC 50% RAP Plant QC 9.5 100.0 100.0 100.0 100.0 100 100 100 4.75 97.4 97.7 97.7 97.7 98 98 96 2.36 74.7 77.6 77.6 76.6 72 76 74 1.18 50.6 54.7 54.7 53.2 50 55 56 0.6 32.5 35.5 35.5 34.2 31 34 39 0.3 18.8 20.9 20.9 19.4 16 19 24 0.15 12.1 13.8 13.8 12.4 11 13 16 0.075 8.7 10.0 10.0 9.0 8.4 10.3 11.4 Table C-2a. Aggregate gradations of mixtures. Material Virgin(%) Mixture 5S (%) Mixture 5A (%) Mixture 16B (%) Virgin Plant QC (%) 5% RAS Plant QC (%) 50% RAP Plant QC (%) Calera limestone 74 63 63 66 69 64 34 Shorter sand 25 30 30 32 30 30 15 Hydrated lime 1 1 1 1.1 1 1 1 Baghouse fines 0 1 1 1.1 0 0 0 Oxford PCâRAS 0 5 5a 16b 0 5 0 aExtracted shingle aggregates and fibers used in place of whole asphalt shingles. bRecovered shingle asphalt used to replace 16% of total virgin binder content. Table C-2b. Aggregate blend proportions.
177 sections describe three experiments conducted to address these questions. Experiment 1: Dry Mixing with RAS The first experiment examined activation by mixing aggre- gate and RAS without any additional asphalt. The hypothesis was that the RAS particles would break because of mixing, and the RAS asphalt would soften as heat transferred to the particles. As the RAS binder heated up, it would activate and begin to coat the aggregate and provide visual cues that activa- tion was occurring. Visual cues thought to indicate the activa- tion of the shingle asphalt were transfer of asphalt from the shingles to aggregate particles (evident by discolored aggre- gates) and agglomeration of nonshingle particles and shingle particles. The experiment consisted of mixing a 5% RAS mixture at 300Â°F without any virgin binder. After mixing for two minutes and then cooling the material, a sieve analysis was performed to separate the particle sizes of the mixture for closer inspec- tion. As shown in Figure C-1a, there was little to no asphalt coating on any of the aggregate, and intact shingle pieces were evident and coated with very fine aggregate. Figure C-1b shows the larger particles of the mixture after being washed to remove the very fine mineral matter, according to AASHTO T 11: Standard Method of Test for Materials Finer Than 75-Âµm (No. 200) Sieve in Mineral Aggregates by Washing. This indi- cates that the larger shingle pieces were not activated enough to cause any of the aggregate to stick to the shingle pieces. Additionally, a replicate sample was mixed in the same manner and then placed in an oven for 2 h at 275Â°F to simulate aging, according to AASHTO R 30. After this short-term con- ditioning period, the overall color of the aggregate changed and some of the large shingle pieces showed dark spots where the shingle asphalt was activated enough to absorb the fine aggregates, as shown in Figure C-1c. The change in color was more notable in the fine aggregates, as shown in Figure C-2a and Figure C-2b. This change in color during the condition- ing process indicates the occurrence of activation after mix- ing and during storage time of mixtures. Experiment 2: Effect of Mixing Temperature on Activation of Shingle Asphalt The second activation experiment involved mixing and compacting Mixture 5S over a range of temperatures from Property Superpave Criteria Mixture 5S Mixture 5A Virgin Mixture 16B Pb na 6.3 6.2 6.4 6.4 Va 4.0â6.0 4.1 4.0 4.7 a 4.3 VMA >16.0 16.3 15.6 16.3 16.0 VFA 66â77 74.8 74.0 71.0 72.9 D/B 1.5â2.0 1.88 2.01 1.72 1.78 Gmm na 2.461 2.475 2.472 2.473 Gmb na 2.360 2.375 2.355 2.366 Note: na = not applicable. aMatched to QCair voids of thin-lift overlay control mixture Table C-3. Design volumetric properties of 4.75 mm mixtures. (a) (b) (c) Figure C-1. Dry mixing of aggregates and RAS: (a) ASTM No. 4 retained aggregates after dry mixing, (b) washed ASTM No. 4 retained aggregates after dry mixing, and (c) ASTM No. 4 retained aggregates after dry mixing and aging.
178 250Â°F to 350Â°F. The laboratory-prepared mixtures were short-term aged for mechanical property testing for 4 h, according to AASHTO R 30-02: Mixture Conditioning of Hot Mix Asphalt (HMA). The %Gmm results from mixing and compacting Mixture 5S at six temperatures are shown in Figure C-3. The plot shows a trend of increasing %Gmm with higher mixing tempera- tures, which could have been caused by either increasing activation and blending of the shingle asphalt or from better compactability because of a lower viscosity of the asphalt at higher temperatures. An analysis of variance showed that temperature had a high influence on the compaction of these specimens (p = 0.0001). The data was further analyzed using Tukeyâs Test of group- ing with a 95% confidence interval. The results shown in Fig- ure C-3 demonstrate the groupings according to Tukeyâs Test. The mid-range temperatures were placed into Group B, and the highest and lowest temperatures were placed in Group A and Group C, respectively. According to the hypothesis, these groupings could indi- cate different levels of activation of the RAS asphalt with little to no activation at the lowest temperature, minor activation at (a) (b) Figure C-2. Dry mixing of aggregates and RAS: (a) passing ASTM No. 16 aggregates after dry mixing, and (b) passing ASTM No. 16 aggregates after dry mixing and aging. Group C Group B Group A 97.5 97.0 96.5 96.0 95.5 95.0 94.5 94.0 G m m o f Sa m pl es ( % ) Figure C-3. Compaction results from increasing mixing temperatures for Mixture 5S.
179 the mid-range temperatures, and more complete activation at the highest temperature. However, NCHRP Project 9-39 also studied the effect of temperature on compaction results and concluded that temperature had a highly significant effect on the compaction (%Gmm) of mixtures, even without recycled materials (West et al. 2010). Mixture 5S had a similar impact from increasing temperature, and therefore, this experiment did not provide clear evidence that activation of the shingle asphalt was affected by the higher temperatures. Experiment 3: Effect of RAS on Mixture Properties Mixture performance tests were also conducted on a series of mixtures to: (a) determine the effects of the RAS compo- nents, (b) further evaluate the effects of mixing and compaction temperatures, and (c) compare the properties of laboratory- prepared mixtures to plant-produced mixtures. The mixture performance tests included dynamic modulus, Texas overlay, IDT creep compliance and strength, and energy ratio. Nine mix- tures were evaluated in this experiment; the results were orga- nized into three groups for comparisons, as shown in Table C-4. PMLC is the abbreviation for plant-mixedâlaboratory- compacted. These mixtures were produced for the pavement preservation experiment on Lee County Road 159. These mixes were sampled at the plant and were stored in buckets for laboratory compaction and testing. The buckets of each mixture were reheated at 320Â°F for 3 to 4 h, and four to six specimens were separated into smaller pans. The smaller pans were then placed back into the oven at 320Â°F until they reached the compaction temperature of 300Â°F. These mix- tures were included as a comparison between plant-produced and laboratory-produced mixtures. The laboratory-prepared mixtures were short-term aged for mechanical property testing for 4 h, according AASHTO R 30-02: Mixture Conditioning of Hot Mix Asphalt (HMA). Long-term aging was excluded from the scope of this research because the pavement preservation project was only 2 years old at the time of testing, and a preliminary comparison of laboratory results and field evaluation was desired. The three groups were selected to test the various hypotheses. Group 1 was selected to investigate the effects of each shingle component separately. The hypothesis for Group 1 was that the addition of the shingle asphalt would cause a change in the stiffness and cracking resistance of the mixture, and the shingle fibers would increase the cracking resistance of the mix- ture. Group 2 was selected to test the effect of mixing tempera- ture on the performance properties of Mixture 5S and compare those results to the plant-produced Mixture 5S PMLC. The hypothesis for Group 2 was that increasing the mixing temper- ature would change the stiffness and cracking resistance of the mixture. The change in stiffness and cracking resistance could be caused by increased activation and blending of the shingle asphalt at the higher temperatures, or it could also be caused by increased aging from the higher temperatures. It was also hypothesized that the plant mixture would have different per- formance test results, possibly because of increased activation of the shingle asphalt assisted by the more thorough mixing conditions, but also because of differences in short-term aging of the mixture from the more vigorous mixing in the plant. Group 3 was selected to compare the plant-produced mixtures to the matched laboratory-produced mixtures. It was hypoth- esized that the plant-produced mixtures would be different than the laboratory-produced mixtures because of differences in mixing and aging between the laboratory and the plant. The QC data and the volumetric properties of the laboratory mix designs are shown in Table C-5. The percentage of the mix- ture that is virgin binderânot including any activated shingle asphaltâis shown as the virgin binder content. The total binder content (TBC) is the total percentage of the mixture that is asphalt, assuming 100% activation of recycled asphalt. The plant-produced virgin mixture and the laboratory- produced virgin mixture both had 4.7% air voids at 75 gyra- tions. AASHTO M 323-12 allows for 4.0 to 6.0% design air Mixture Group 1: Shingle Components Group 2: Mixing Temperature Group 3: Comparison to PMLC Mixture 5S mixed at 250Â°F X Mixture 5S mixed at 300Â°F X X Mixture 5S mixed at 350Â°F X X X Virgin X X Mixture 5A X Mixture 16B X Mixture 5S PMLC X X Virgin PMLC X Mixture 50R PMLC a X Note: PMLC = plant-mixedâlaboratory-compacted. a50% fine-fractionated RAP with 6% asphalt equaling 55% of total asphalt from RAP binder. Table C-4. Grouping of mixtures for analysis of mixture properties.
180 voids for 4.75-mm NMAS designs because of their low per- meability. The ignition method (AASHTO T 308) was used to determine the binder content of the plant-produced mix- tures stored for this study. Except for TBC, the plant-produced 5% RAS mixture and the laboratory Mixture 5S 325Â°F had very similar volumetric properties. The 5% RAS QC data listed TBC at 6.0%, but AASHTO T 308 testing showed the mixture to have a 6.3% TBC. The difference is possibly caused by organic fiber or deleterious materials in the RAS that were burned off and indi- cated a higher binder content. TBC of 6.3% matches the design binder content of the laboratory-produced Mixture 5S 325Â°F. Assuming 100% activation of the shingle asphalt, the total binder content for Mixture 5A and Mixture 5S 325Â°F were within 0.1%. The Gmm of Mixture 5A was higher than Mix- ture 5S 325Â°F. This could be caused by the extracted shingle aggregates and fibers because they are no longer conglomer- ated together and are free to fill smaller voids in the aggregate structure. Also, this could indicate that the Gsb of the aggre- gates and fines needs to be further evaluated when separated from the shingles. Mixture 16B and the Virgin Mixture had the same total binder content and nearly the same Gmm. The Gmm for these mixtures more closely matches the Gmm of Mixture 5A, which may also indicate that the conglomerated shingle aggregates, fibers, and asphalt were not adequately separated and dis- persed in Mixture 5S 325Â°F. Analysis of Mixture Performance Testing Dynamic Modulus (E*) Figure C-4 shows the E* master curves for the mixtures in Group 1, and Table C-6 shows the Tukeyâs Test of Group 1 E* data. The analysis of these results showed that the E* values for Variable Plant-Produced QC Data Laboratory-Produced Mixture 5% RAS PMLC Virgin PMLC 50% RAP PMLC Mixture 5S 325Â°F Mixture 5A Virgin Mixture 16B TBC 6.0a 6.2 5.4 6.3 6.2 6.4 6.4 VBC 5.0b 6.2 2.7 5.2 6.2 6.4 5.4 Pbe 5.6 5.8 4.6 5.3 5.0 5.1 5.0 Gmm 2.455 2.449 2.486 2.461 2.475 2.472 2.473 Avg. Gmb 2.365 2.335 2.395 2.360 2.375 2.355 2.366 Avg. Va 3.7 4.7 3.7 4.1 4.0 4.7 4.3 Avg. Gsb 2.663 2.665 2.648 2.638 2.638 2.635 2.635 Avg. VMA 16.5 17.8 14.4 16.2 15.6 16.3 16.0 Avg. VFA 78 74 75 75 74 71 73 D/B ratio 1.8 1.4 2.5 1.9 2.0 1.7 1.8 Note: VBC = virgin binder content. aBinder content later proven to be 6.3% using ignition oven (AASHTO T308). bEstimated using known RAS properties. Table C-5. Quality-control data compared to the matched laboratory designs. Figure C-4. RAS component effects on dynamic modulus.
181 the Virgin Mixture, Mixture 5A (RAS aggregate and fibers only), and Mixture 5S 300Â°F were not statistically different at any temperature or loading frequency. Therefore, this indicates that the RAS binder in Mixture 5S 300Â°F was not sufficiently activated to affect the mixture stiffness. However, Mixture 5S 350Â°F and Mixture 16B had significantly higher E* values compared to Virgin Mixture and Mixture 5A at the higher temperaturesâlower frequencies. Therefore, the increased stiffness in Mixture 5S 350Â°F and Mixture 16B indi- cate activation and blending of the shingle asphalt with the virgin binder. This is consistent with conclusions from other studies, which found that increasing the blend percentage of shingle asphalt increasingly stiffens the composite binder (Foo et al. 1999, Mallick et al. 2000, Kriz et al. 2014). Table C-7 shows the Tukeyâs Test of Group 2 E* data, and Figure C-5 shows the E* master curves for the mixtures in Temperature Mix ID 0.1 Hz 1 Hz 10 Hz 0.01 Hz Mixture 5S 350Â°F A A A Mixture 16B A B A A 4Â°C Mixture 5A A B C A B A B Virgin Mixture B C A B A B Mixture 5S 300Â°F C B B Mixture 5S 350Â°F A A A Mixture 16B A A B A B 20Â°C Mixture 5A B B A B Virgin Mixture B B A B Mixture 5S 300Â°F B B B Mixture 5S 350Â°F A A A A Mixture 16B B B B B 40Â°C Mixture 5A C C B C C Virgin Mixture C C B C C Mixture 5S 300Â°F C C C C Table C-6. Tukeyâs Test of dynamic modulus values for Group 1. Temperature Mix ID 0.1 Hz 1 Hz 10 Hz 0.01 Hz Mixture 5S PMLC A A A 4Â°C Mixture 5S 350Â°F A A A B Mixture 5S 300Â°F B B C Mixture 5S 250Â°F B B B C Mixture 5S PMLC A A A 20Â°C Mixture 5S 350Â°F A A A Mixture 5S 300Â°F B B B Mixture 5S 250Â°F C B B Mixture 5S PMLC A A A A 40Â°C Mixture 5S 350Â°F A A B A Mixture 5S 300Â°F B B C B Mixture 5S 250Â°F C C D C Table C-7. Tukeyâs Test of the dynamic modulus values for Group 2.
182 Group 2. Analysis of the results for this group shows that increasing the mixing temperature significantly increases the E* of the mixture. This could be because of the increased activation and blending of the shingle asphalt caused by the increased mixing and compaction temperatures, or it could be because of the stiffening/aging of the asphalt caused by the increased mixing and compaction temperatures. As noted in Group 1, Mixture 5S 300Â°F was not significantly differ- ent from the Virgin Mixture. Mixture 5S PMLC (produced at 325Â°F) and Mixture 5S 350Â°F were not significantly different except at the highest temperature and 10 Hz frequency. This may indicate that a higher mixing temperature is needed for laboratory-prepared mixtures to simulate the conditions of the plant-produced mixture. Table C-8 shows the Tukeyâs Test of Group 3 E* data, and Figure C-6 shows the E* master curves for the mixtures in Group 3. The analysis of this group also showed that Mix- ture 5S 350Â°F and Mixture 5S PMLC were statistically simi- lar at all temperatures and frequencies. Similarly, the E* of the Virgin Mixture closely matched that of the Virgin PMLC Figure C-5. Temperature effects on the dynamic modulus of mixtures containing RAS. Temperature Mix ID 0.1 Hz 1 Hz 10 Hz 0.01 Hz Mixture 50R PMLC A A A Mixture 5S PMLC B B B 4Â°C Mixture 5S 350Â°F B C B C B Virgin Mixture C D B C B Virgin PMLC D C B Mixture 50R PMLC A A A Mixture 5S PMLC B B B 20Â°C Mixture 5S 350Â°F B B C B Virgin Mixture C C B Virgin PMLC C C B Mixture 50R PMLC A A A A Mixture 5S PMLC B B B B 40Â°C Mixture 5S 350Â°F B B B B Virgin Mixture C C C C Virgin PMLC D D D D Table C-8. Tukeyâs Test of the dynamic modulus values for Group 3.
183 mixture except at the highest test temperature. These are important comparisons for relating the laboratory perfor- mance testing results to the field performance on Lee County Road 159. Mixture 50R PMLC was significantly stiffer than the other mixtures. This was likely because of the high per- centage of recycled asphalt replacing the virgin binder in the mixture and its lower total binder content. The Dynamic Modulus Test results indicated an increased stiffness in Mixture 5S 350Â°F and Mixture 16B (RAS binder only), which may indicate increased activation and blending of the shingle asphalt in those mixtures. Conversely, Mix- ture 5A, Virgin Mixture, and Mixture 5S 300Â°F were not sig- nificantly different, indicating no significant activation of the RAS at 300Â°F. Mixture 5S 250Â°F had a decreased stiffness at the high temperaturesâlow frequencies. The laboratory-prepared mixtures were not significantly different from the correspond- ing PMLC mixtures. A higher mixing temperature or addi- tional aging may be needed to adequately match E* results of laboratory-produced mixtures to plant-produced mixtures. Texas Overlay Tester The number of cycles to failure using the AMPT overlay tester was determined in two ways: first, by the traditional Tex-248-F 93% load reduction method as developed by Zhou and Scullion (2005); and second, by the normalized load Ã cycle method as reported by Ma et al. (2014). Results for these two methods were analyzed separately. Table C-9 gives the Tukeyâs Test of each group individually and as a combined group. Figure C-7 compares the cycles to failure for each group using bar charts. Figure C-7a compares the mixtures in Group 1. Mixture 5A had the highest number of cycles to failure in this group, but its OT results were not statistically different from the Virgin Mix- ture or Mixture 5S 300Â°F for both evaluation methods. There- fore, the hypothesis that the shingle fine aggregates and fibers would increase the cracking resistance could not be verified. In addition, this statistical grouping indicates that the shingle asphalt in Mixture 5S 300Â°F may not have activated or blended enough to affect the cracking resistance of the mixture. Yet, the RAS was likely contributing to the mixture in some way because a decrease in the effective binder content would likely decrease the cracking resistance of the mixture. OT cycles to failure for Mixture 5S 350Â°F and Mixture 16B were signifi- cantly lower than Mixture 5A. This may support the hypothesis that the recycled shingle asphalt contributes to a lower resistance to high strains, but the OT cycles to failure for Mixture 16B was not significantly lower than for the Virgin Mixture. An analysis of the mixtures in Group 2, as shown in Fig- ure C-7b, yielded similar results as the E* testing. A statis- tically significant difference was seen between Mixture 5S 300Â°F and Mixture 5S 350Â°F. This supports the hypothesis that increased mixing temperatures decrease the high strain tolerance of RAS mixtures, which could be attributed to more complete activation and blending of RAS binder or simply because of the increased aging at 350Â°F. Also consistent with the E* results, the OT results for Mixture 5S 350Â°F were sta- tistically similar to Mixture 5S PMLC. Figure C-7c compares OT results for the mixtures in Group 3. Although the bar chart shows a large difference between the laboratory-prepared Virgin Mixture and Virgin PMLC, Tukeyâs Test indicated that the results are not statistically different because of the high variability of the test for both methods of analysis. Tukeyâs Test also showed that results of Mixture 5S PMLC, Mixture 5S 350Â°F, and Mixture 50R PMLC (50% RAP) were not significantly different. Figure C-6. Dynamic modulus curves for Group 3.
184 The hypothesis that the shingle fine aggregates and fibers would increase the cracking resistance could not be verified because Mixture 5A was not significantly different from the Virgin Mixture. Although the shingle asphalt in Mixture 5S 300Â°F and Mixture 5S 250Â°F may not have activated or blended enough to affect the cracking resistance of the mix- tures, they were not adversely affected by a lack of effective asphalt. OT cycles to failure for Mixture 5S 350Â°F and Mix- ture 16B were significantly lower than Mixture 5A, suggest- ing that when the RAS asphalt is activated, the resistance to high strains decreases. There was not a significant difference between Mixture 5S PMLC, Mixture 5S 350Â°F, and Mix- ture 50R, but they had statistically lower OT cycles to failure Group 1 NLC 93%LR Mixture 5A A A Mixture 5S 300Â°F A B A Virgin A B A Mixture 16B B A B Mixture 5S 350Â°F C B Group 2 NLC 93%LR Mixture 5S 250Â°F A A Mixture 5S 300Â°F A A Mixture 5S 350Â°F B B Mixture 5S PMLC B B Group 3 NLC 93%LR Virgin PMLC A A Virgin B B Mixture 5S 350Â°F C C Mixture 5S PMLC C C Mixture 50R PMLC C C All Mixtures NLC 93%LR Virgin PMLC A A Mixture 5A A A Mixture 5S 300Â°F A B A Mixture 5S 250Â°F A B A Virgin A B A B Mixture 16B B C A B C Mixture 5S 350Â°F C D B C D Mixture 5S PMLC D C D Mixture 50R PMLC D D Note: NLC = normalized load cycle. LR = load reduction. Table C-9. Tukeyâs Test of the overlay tester cycles to failure results. (a) 12,000 10,000 8,000 6,000 4,000 2,000 0 14,000 Cy cl es to F ai lu re (b) 12,000 10,000 8,000 6,000 4,000 2,000 0 14,000 Cy cl es to F ai lu re (c) 12,000 10,000 8,000 6,000 4,000 2,000 0 14,000 Cy cl es to F ai lu re Figure C-7. Overlay testing results: (a) Group 1, shingle components; (b) Group 2, mixing temperature, and (c) Group 3, comparison to PMLC.
185 than the Virgin Mixture, which also supports the hypothesis that the activation of RAS asphalt reduces a mixtureâs crack- ing resistance. Indirect Tension Testing Creep compliance testing was performed at four tempera- tures: â20Â°C, â10Â°C, 0Â°C, and 10Â°C. These data were ana- lyzed in two setsâ[â20Â°C, â10Â°C, and 0Â°C] and [â10Â°C, 0Â°C, and 10Â°C]âfor comparison. The base virgin asphalt was a PG 67-22, and according to AASHTO T 322-07, the creep compliance analysis should be performed at â20Â°C, â10Â°C, and 0Â°C. Analyses were also conducted using tests at â10Â°C, 0Â°C, and 10Â°C because a majority of the mixtures contained aged asphalt, which has been shown to affect mixture results for low-temperature properties (Bonaquist 2011, Abbas et al. 2013, Zhao et al. 2014). After analyzing both data sets, the [â10Â°C, 0Â°C, and 10Â°C] set was found to have a lower standard error than the [â20Â°C, â10Â°C, and 0Â°C] data set. Therefore, the IDT strength test- ing was performed at 0Â°C, and only the [â10Â°C, 0Â°C, and 10Â°C] data set was used for comparisons between mixtures. It should also be noted that Mixture 50R PMLC and the Virgin PMLC mixture were only tested at â20Â°C, â10Â°C, and 0Â°C, and the IDT strength testing was performed at â10Â°C. There- fore, Mixture 50R PMLC and the Virgin PMLC were analyzed using the data from â20Â°C, â10Â°C, and 0Â°C but with a refer- ence temperature of â10Â°C, and they were excluded from the comparison of mixture IDT strengths. The estimated critical pavement temperature for low- temperature cracking is shown in Table C-10. The critical temperatures were determined using the LTSTRESS work- book. But when the critical temperatures were plotted, the temperatures did not line up with the thermal stress curves. After further investigation, it was determined that the calcu- lation of the critical temperature assumes that the thermal stress curve is linear between 100 and 1,000 psi when plotted as semilogarithmic [x-log(y)]. While this was a close approxi- mation, the semilogarithmic curve was concave down, and therefore, the critical temperature was overestimated. To get a more accurate calculation of the critical stress, the thermal stress curve was assumed to be a quadratic function; so a qua- dratic regression was used to determine the best-fit equation for the curves between 100 and 1,000 psi. The sum of squares due to error (SSE) and the total sum of squares were calcu- lated for each point with a stress between 100 and 1,000 psi, and Microsoft Excelâs Solver function was used to maximize the R2 value. This resulted in more reasonable estimates of the critical thermal cracking temperature, based on the thermal stress curves. As evident in Table C-10, there is no relationship between IDT strength and critical temperature. The shift of the ther- mal stress curve has a greater impact on the critical tem- perature than changes in the strength. The Virgin PMLC, the laboratory-prepared Virgin Mixture, and Mixture 5S 250Â°F had critical temperatures below â17.0Â°C. This group of results indicates that at the mixing temperature of 250Â°F, the RAS binder is not activated. Mixture 5S 300Â°F, Mixture 5A, and Mixture 16B had critical temperatures of about â15Â°C. Although the slightly higher critical temperature for Mix- ture 5S 300Â°F could be caused by minor activation of the RAS binder, it would not be the case for Mixture 5A since it did not contain any RAS binder. Mixtures 5S PMLC, Mixture 5S 350Â°F, and Mixture 50R PMLC had the highest critical tem- peratures (â13.3Â°C and above), perhaps indicating that the recycled binders were more fully activated and had a negative effect on thermal cracking resistance. These mixtures were also shown to have higher E* values than the other mixtures and to be significantly less tolerant of high strains, based on OT results. Since the LTSTRESS analysis of the creep compliance mas- ter curve yields only one combined result, it is not possible to conduct a statistical analysis of critical temperatures. There- fore, bar charts of the creep compliance at certain loading times were used for simple comparisons among the mixture groups. Loading times of 31,600 s, 1,000 s, and 31.6 s were chosen as the approximate midpoints of the data at each tem- perature tested: 10Â°C, 0Â°C, and â10Â°C, respectively. Mix ID IDT Strength (psi) LTSTRESS Critical Temperature (Â°C) Quadratic Regression Critical Temperature (Â°C) Virgin PMLC 383.6 -19.1 -18.4 Mixture 5S 250Â°F 270.5 -17.8 -17.3 Virgin Mixture 272.1 -18.0 -17.3 Mixture 5S 300Â°F 271.6 -16.6 -15.5 Mixture 5A 284.2 -15.3 -15.0 Mixture 16B 275.6 -15.6 -15.0 Mixture 5S PMLC 308.3 -14.4 -13.3 Mixture 5S 350Â°F 282.6 -13.7 -12.7 Mixture 50R PMLC 422.1 -13.6 -12.3 Table C-10. IDT Critical temperature results.
186 Figure C-8 shows little differences in the compliance values among Group 1 at the two lower temperatures. At the high- est test temperature (+10Â°C), the creep compliance values for Mixture 5S 350Â°F and Mixture 16B are somewhat lower than the other mixes in the group, possibly indicating activation and blending of the shingle asphalt and, consequently, a less compliant (stiffer) composite binder. Figure C-9 shows the data for Group 2. There is a clear decrease in creep compliance with increased mixing temper- ature, and Mixture 5S PMLC has an even lower creep compli- ance than the laboratory-produced mixtures. The decrease in creep compliance could be because of increased activation of the stiffer shingle asphalt. This is consistent with the results seen in the E* and overlay testing. Yet, data from NCHRP Report 648 also showed a reduction in the creep compliance results for higher mixing and compaction temperatures for virgin mixtures (West et al. 2010). Therefore, the reduction in creep compliance may be impacted by aging of the virgin binder at the higher mixing temperatures. Further testing may be necessary, comparing the impact of mixing tempera- ture on virgin binder and composite binders. Figure C-10 shows the data for Group 3. Mixture 5S PMLC had a lower creep compliance than Mixture 5S 350Â°F. The Virgin PMLC mixture had a lower creep compliance than the laboratory-produced Virgin Mixture. The lower creep com- pliance of the PMLC mixtures could indicate more aging during plant production. Mixture 50R PMLC was shown to have the lowest creep compliance. This was consistent with the higher stiffness seen in the E* and OT results. Three apparent groupings of the mixtures were observed for the critical temperatures. The virgin mixtures and RAS mixtures at low mixing temperature had the lowest criti- cal temperatures. The stiffer mixtures, as shown by E* and OT results, had higher critical temperatures. The shift in the Figure C-8. Group 1 creep compliance midpoint values. Figure C-9. Group 2 creep compliance midpoint values.
187 thermal stress curves may have been caused by increased activation and blending of the shingle asphalt. Yet, there may be other factors affecting the shift because Mixture 5A had a higher critical temperature than the virgin mixtures. The RAS fibers may have contributed to a lower compliance of the mixture. The creep compliance values for Mixture 5S 350Â°F and Mix- ture 16B are somewhat lower than the other mixes in Group 1, possibly indicating activation and blending of the shingle asphalt. Increasing the mixing temperature decreased the creep compliance, which could be because of increased acti- vation and blending of the shingle asphalt or increased aging of the virgin binder at higher mixing and compaction tem- peratures. The PMLC mixtures both had lower creep compli- ance results compared to their laboratory-produced matches. This indicates that the plant-produced mixes were aged more than the 4-h loose mix aging of the laboratory-prepared mixtures. Energy Ratio Results Energy ratio and related properties were determined using the software program Indirect Tension Test at Low Tempera- tures (ITLT) developed by Roque et al. (2004) at the Univer- sity of Florida and the Florida Department of Transportation. ITLT uses three replicate specimens for analysis and returns a single result for each property of the mixture. The resulting properties are shown in the following bar graphs. The differ- ent colored bars show the three groups. Some mixtures are in more than one group. The DCSEf result for each mixture is shown in Figure C-11. For Group 1, Mixture 5A had a higher DCSEf, which may indicate that the shingle fibers contribute to this cracking resistance parameter. The Virgin Mixture, Mixture 5S 300Â°F, and Mixture 16B had similar results. Mixture 5S 350Â°F had a lower DCSEf, which could indicate increased brittleness caused by activation and blending of the shingle asphalt. The Figure C-10. Group 3 creep compliance midpoint values. Group 1 D C SE f ( kJ /m 3) Group 2 Group 3 Figure C-11. Dissipated creep strain energy at failure (kJ/m3).
188 data from Group 2 showed a decreasing trend of DCSEf with increasing mixing temperature, and Mixture 5S PMLC had the lowest DCSEf. This indicates increased brittleness with higher mixing temperature, which may be caused by increased activation and blending of the shingle asphalt or because of additional aging caused by the higher temperatures. For Group 3, the DCSEf of the plant mix with RAS was most simi- lar to the laboratory RAS mix prepared at 350Â°F. However, the DCSEf of the Virgin Mixture and the Virgin PMLC mixture in Group 3 were not similar. The Virgin PMLCâs DCSEf was much higher than the laboratory-produced Virgin Mixture. All of the mixtures met the University of Florida recom- mended criteria of DCSEf >0.75 kJ/m3 except Mixture 50R PMLC, which had a negative value for DCSEf. DCSEf is cal- culated from resilient modulus and fracture energy. IDT fracture energy of Mixture 50R PMLC was 0.0 kJ/m3 for two of the samples and 0.1 kJ/m3 for one of the samples tested. Fracture energy for the other mixtures ranged from 2.2 to 5.3 kJ/m3. Mixture 50R PMLC has been shown to be much stiffer than the other mixtures from the E* results (Fig- ure C-6) and IDT creep compliance results (Figure C-10). The creep compliance rates for each mixture are shown in Figure C-12. The results show a clear divide between the mixtures that have been shown to be stifferâMixture 16B, Mixture 5S 350Â°F, Mixture 5S PMLC, and Mixture 50R PMLCâand the mixtures that were not significantly differ- ent from the virgin mixture (Mixture 5A, Mixture 5S 250Â°F, Mixture 5S 300Â°F, Virgin Mixture, and Virgin PMLC Mix- ture). The mixtures with lower creep compliance results are likely stiffer because of increased activation and blending of the shingle asphalt. The energy ratio result for each mixture is shown in Fig- ure C-13. In Group 1, Mixture 16B and Mixture 5S 350Â°F had the highest energy ratio. Mixture 5A had a slightly higher energy ratio than the Virgin Mix and Mixture 5S 300Â°F. In Group 2, Mixture 5S 250Â°F and Mixture 5S 300Â°F had energy ratios that were much lower than Mixture 5S 350Â°F and Mix- ture 5S PMLC. In Group 3, the Virgin PMLC Mixture had a higher energy ratio than the laboratory-produced Virgin Mixture. Mixture 5S 350Â°F had a higher energy ratio than Mixture 5S PMLC, but both were higher than the Virgin mix- tures. Mixture 50R PMLC had a negative energy ratio because of the negative DCSEf result. Although energy ratio has been proposed as an indicator of top-down cracking, there is growing skepticism about its ability to properly characterize very stiff mixtures. Very stiff mixtures tend to have high energy ratio values because creep compliance parameters are in the denominator of the energy ratio equation. As creep compliance decreases (stiffness/brittleness of the mix- ture increases), energy ratio increases. This seems counterintui- tive if higher stiffness is associated with brittleness. Field Performance of Thin Overlays on LR 159 The field performance of the three thinlay mixtures included in this study continues to be monitored as part of NCATâs ongoing pavement preservation research. Table C-11 shows the cracking performance of these sections and the untreated control sections after 5 years of traffic. No other distresses are significant. The inbound and outbound lanes of this roadway are monitored separately because of the significant difference in loading of the two directions due to the quarry and asphalt plant located at the end of the road. These data show that the virgin thinlay section has performed extremely well, with only about Â¼ of the crack- ing in the virgin thinlay compared to the pretreatment condition. The 50% RAP thinlay has provided a substantial preservation benefit compared to the control sections but has not performed as well as the virgin thinlay. Similarly, Group 1 Group 2 Group 3 Figure C-12. Creep compliance rate.
189 the 5% RAS thinlay has performed well with less crack- ing than the pretreatment cracking in the inbound direc- tion but an increased amount of cracking in the outbound direction, although the moderate amount of cracking that now exists in the heavily loaded outbound direction after 5 years is considered to be good performance. Overall, the virgin thinlay section is clearly performing the best with the 5% RAS and 50% RAP sections performing similarly, which is consistent with OT results, creep compliance rate, and dissipated creep strain energy at failure but not con- sistent with energy ratio. West et al. (2015) and Willis et al. (2016) reported similar relationships between these labora- tory tests and field performance of other closely monitored test sections. Conclusions and Recommendations The results from the dry mix experiment seem to indicate that RAS was not activated during a short mixing time at a moderate mixing temperature. However, after condition- ing the dry mix with RAS for 2 h at elevated temperatures, a minor amount of activation occurred to allow some of the RAS binder to adhere to fine aggregate. Increasing mixing and compaction temperatures may increase the activation and blending of shingle asphalt in mix- tures containing RAS. Although activation of the RAS binder increases the density of laboratory compacted samples and may indicate more effective binder in volumetric property analyses, the activation of the RAS binder also increases the stiffness and Group 1 Group 2 Group 3 Figure C-13. Energy ratio. Table C-11. Field cracking performance of thinlay mixtures and control sections. Section Number Section Description Inbound Outbound 60,000 accumulated ESALs 800,000 accumulated ESALs Aug. 2012 Aug. 2017 Change in 5 years Aug. 2012 Aug. 2017 Change in 5 years Cracking (% of lane) Cracking (% of lane) 3 Control, low distress 7.6 36.0 +28.4 0.5 19.5 +19.0 4 Control, medium distress 7.5 36.3 +28.8 24.8 96.7 +71.9 19 Virgin thinlay 5.8 1.5 -4.3 27.2 6.1 -21.1 23 50% RAP 1.3 4.9 +3.6 9.3 17.6 +8.3 24 5% RAS 2.2 1.5 -0.7 4.4 15.8 +11.4
190 brittleness of mixtures containing RAS, making them more susceptible to cracking. The hypothesis that higher mixing tem- peratures increase the activation of RAS binder was partially supported by statistically significant differences in the results of the performance tests for the mixes prepared at different tem- peratures. The confounding effect is that higher mixing and compaction temperatures would also tend to age the virgin binder which could cause differences in performance test results. Based on the available data on laboratory and plant pro- duced mixtures, it is not possible to determine if the differences in mix performance test results (E*, OT, creep compliance rate, and DCSEf) were caused by more activation of the RAS binder or because of more short-term aging at the higher mixing and compaction temperatures. However, the fact that several per- formance test results of the mix with 5% RAS (Mix 5S) mixed at 300Â°F were not statistically different from the mix with RAS aggregates and fibers, but were different from the mix with only RAS binder (Mix 16B) provides some evidence that the RAS binder was not fully active when mixing at 300Â°F. Comparison of the laboratory-produced mixtures to their respective PMLC mixtures demonstrated the PMLC mixtures to be slightly stiffer than the laboratory-produced mixtures conditioned according to the short-term aging protocol in AASHTO R 30. Further laboratory aging or increased mix- ing temperatures may be needed to better match laboratory- prepared mixtures to plant-produced mixtures. Field performance of thin overlays with a mixture with all virgin materials, a mixture containing 5% RAS, and a mixture containing 50% RAP showed that these mixtures containing recycled materials were less cracking resistant than the virgin mix overlay. The Texas Overlay Tester results and the indi- rect tensile creep compliance rate determined as part of the energy ratio method were consistent with the observed 5-year field performance of these three thin overlays. Recycled asphalt shingles can be used in asphalt mixtures to offset a portion of the virgin binder. However, the effect of RAS on the performance of the asphalt mixture must be considered. Although increasing mixing temperatures and/or longer storage times may cause greater activation of the recy- cled shingle asphalt, that should not be considered a desirable outcome in most cases. There is no better argument for the need to use mixture performance tests in mix design and QA than the challenges of using RAS in asphalt mixtures. Bibliography Abbas, A., U. Mannan, and S. Dessouky. Effect of Recycled Asphalt Shin- gles on Physical and Chemical Properties of Virgin Asphalt Binder. Construction and Building Materials, Vol. 45, 2013, pp. 162â172. Bonaquist, R. Effect of Recovered Binders from Recycled Shingles and Increased RAP Percentages on Resultant Binder PG. Wisconsin Highway Research Program, Madison, 2011. Brunswick County, North Carolina. Solid Waste and Recycling Accepted Items and Tipping Fees, 2015. http://www.brunswickcountync.gov/ solidwaste/tipfees/. Accessed Jan. 31, 2016. California Waste Services. Construction and Demolition (C&D) Recycling and Current Rates, 2016. http://www.californiawasteservices.com/ cd.html. Accessed Jan. 31, 2016. Christenson, D. LTSTRESS, English units version software. June 2013. Dane County Department of Public Works. Dane County Landfill, 2015. https://www.countyofdane.com/pwht/recycle/landfill.aspx. Accessed Jan. 30, 2016. Foo, K. Y., D. I. Hanson, and T. A. Lynn. Evaluation of Roofing Shin- gles in Hot Mix Asphalt. Journal of Materials in Civil Engineering, Vol. 11, No. 1, 1999, pp. 15â20. Halifax C & D Recycling, Ltd. Tipping Fees. http://halifaxcdrecycling.ca/ tipping-fees. Accessed Jan. 31, 2016. Hansen, K., and A. Copeland. Asphalt Pavement Industry Survey on Recycled Materials and Warm-Mix Asphalt Usage: 2014. National Asphalt Pavement Association, Lanham, Md., 2015. Krivit, D. Recycling Tear-Off Asphalt Shingles: Best Practices Guide. Con- struction Materials Recycling Association, Eola, Ill., 2007. Kriz, P., D. Grant, B. Veloza, M. Gale, A. Blahey, and J. Brownie. Blending and Diffusion of Reclaimed Asphalt Pavement and Virgin Asphalt Binders. Road Materials and Pavement Design, 2014, Vol. 15, pp. 78â112. Ma, W., N. H. Tran, A. J. Taylor, and Z. Li. Proposed Improvements to Overlay Test for Determining Cracking Resistance of Asphalt Mixtures. Presented at 93rd Annual Meeting of the Transportation Research Board, Washington, D.C., 2014. Mallick, R. B., M. R. Teto, and W. S. Mogawer. Evaluation of Use of Manufactured Waste Asphalt Shingles in Hot Mix Asphalt. Technical Report 26, Chelsea Center for Recycling and Economic Develop- ment, Chelsea, Mass., 2000. Roque, R., B. Birgisson, C. Drakos, and B. Dietrich. Development and Field Evaluation of Energy-Based Criteria for Top-Down Crack- ing Performance of Hot Mix Asphalt. Journal of the Association of Asphalt Paving Technologists, Vol. 73, 2004, pp. 229â260. Sandler, K. Analyzing Whatâs Recyclable in C&D Debris. BioCycle, 2003, pp. 51â54. U.S. Environmental Protection Agency. Estimating 2003 Building- Related Construction and Demolition Materials Amounts. 2003. West, R. C., D. E. Watson, P. A. Turner, and J. R. Casola. NCHRP Report 648: Mixing and Compaction Temperatures of Asphalt Binders in Hot-Mix Asphalt. Transportation Research Board of the National Academies, Washington, D.C., 2010. West, R. C., N. H. Tran, A. J. Taylor, and J. R. Willis. Comparison of Laboratory Cracking Test Results with Field Performance of Moderate and High RAP Content Surface Mixtures on the NCAT Test Track. Presented at 8th RILEM International Symposium on Testing and Characterization of Sustainable and Innovative Bituminous Materials, Ancona, Italy, 2015. Willis, J. R. Reclaimed Asphalt Shingles Characterization: Best Practices. NCAT Report 13-07. National Center for Asphalt Technology, Auburn University, Auburn, Ala., 2013. Willis, J. R., A. J. Taylor, and T. M. Nash. Laboratory and Field Evalua- tion of Florida Mixtures at the 2012 National Center for Asphalt Technology Pavement Test Track. Transportation Research Record: Journal of the Transportation Research Board, No. 2590, 2016, pp. 65â73. http://dx.doi.org/10.3141/2590-08. Zhao, S., B. Bowers, B. Huang, and X. Shu. Characterizing Rheologi- cal Properties of Binder and Blending Efficiency of Asphalt Paving Mixtures Containing RAS Through GPC. Journal of Materials in Civil Engineering, Vol. 26, No. 5, 2014, pp. 941â946. Zhou, F., and T. Scullion. Overlay Tester: A Rapid Performance Related Crack Resistance Test. No. FHWA/TX-05/0-4467-2, Texas Trans- portation Institute, Texas A&M University System, College Station, 2005.