Recycled Materials and Byproducts in Highway Applications—Scrap Tire Byproducts, Volume 7 (2013)

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.

15 TDA has been used in a number of highway applications. The information in this section is organized into three major categories: Bound Applications—Portland Cement and Con- crete, Bound Applications—Asphalt Cement and Concrete, and Unbound Applications—Bases, Embankments, and Fills. Each category has a number of specific applications. Bound ApplicAtions—portlAnd cement And concrete Research was found for the uses of TDA and crumb rubber in a number of portland cement and/or concrete applications including: • Precast portland cement concrete (PCC) applications • Scrap tire fibers in PCC • Crumb rubber in mortars • TDA and crumb rubber in PCC • PCC pavements. precast portland cement and concrete Applications Allen (2004) noted that crumb rubber concrete in precast pan- els was useful because it was lightweight, helped control noise, and improved insulation properties. The major disadvantage to using crumb rubber concrete was the loss of strength. “Concrete Pavement’s New Road Map” (Kuennen 2006) listed those benefits achieved with using crumb rubber in precast panels for PCC pavements: • Improved thermal cycling resistance that provides a concrete that is less likely to crack and shatter under repeated freeze-thaw cycles. • Easier transportation of lightweight precast panels. • Lower overall cost. • Promotion of recycling. In Thailand, Sukontasukkul (2009) evaluated the thermal and acoustical properties of crumb rubber concrete mixes for potential use in sound barrier panels. Materials used in the study included a Type I portland cement, 9.5 mm coarse aggregate, river sand, super plasticizer Type F, and various concentrations and sizes of crumb rubber. Crumb rubber and aggregate properties are shown in Table 22. The gradations used in the study were mostly one-sized. Mixes with the indi- vidual crumb rubber sizes and a combination of the two sizes were used at varying percentages of 10%, 20%, and 30%. Testing used ASTM C642 for density and voids, ASTM C177 for steady state heat flux and thermal transmission mea- surements, and ISO 10534-1 for sound absorption coefficient and impedance measurements using an impedance tube. Den- sity and porosity decreased with increasing amounts of crumb rubber (Table 23). Thermal conductivity is a measure of the heat transmitted through a unit thickness in a direction normal to the surface under steady state conditions. The use of crumb rubber in PCC significantly lowered the thermal conductiv- ity (Table 24). The No. 6 crumb rubber provided the most reduction at 30%, whereas the No. 25 crumb rubber PCC at 10% was almost as low. The rate of heat transfer per unit time (W/h) and the heat resistivity (m2/KW) were calculated using the thermal conductivity and an assumed temperature differ- ence of about 55°F. Sound absorption was measured at low to mid frequency (125, 250, and 500 Hz) and high frequency (1000, 2000, and 4000 Hz) ranges (Table 25). Mixes with crumb rubber absorbed sound better than the control PCC at the higher frequencies. Overall, the crumb rubber mixes improved sound resis- tivity by about 36%. Conclusions were that crumb rubber PCC reduced the unit weight of the mix from 14% to 28%, reduced thermal conductivity, and provided some improved sound absorption. tire derived Aggregate Fibers in portland cement concrete Hernandez-Olivares et al. (2002, 2006) reported on research conducted in Spain that evaluated the use of scrap tire fibers in PCC. The materials used in the study included cement, coarse aggregate (12 to 19 mm), sand (3 to 6 mm), fine aggre- gate (<3 mm), superplasticizer, retarder, polypropylene fibers (0.1% by volume), and crumb rubber (3.5% and 5.0% by vol- ume). The polypropylene fiber was added to minimize the drying shrinkage cracking, but was sufficiently low in con- centration to expect a noticeable impact on the hardened PCC mechanical properties. Scrap tire byproduct was obtained using strip processing to produce rubber fibers with lengths ranging between 0.33 and 0.9 in. This type of cutting process produced about 4% rubber powder. chapter two ApplicAtions

16 ing rubber content were found. The rubber fibers resulted in more energy being absorbed, which tended to disrupt crack propagation. Li et al. (2004) used scrap tire fibers (i.e., small strips of tire rubber) in place of crumb rubber to evaluate the possibil- ity of improving the mechanical properties of scrap tire rub- ber PCC. The hypothesis was based on the proven improve- ment of concrete mechanical properties with the use of fibers. The objective of the research was to determine if tire fibers would provide a similar improvement. Materials used in the study were Type I portland cement, gravel, natural sand, water, and air entrainment admixture. Two geometries of scrap tire byproducts were used. The first was a tire chip with a size of 1 in. by 1 in. by 0.2 in. thick. The second geometry was a long thin rectangle with different lengths (tire fibers). All were 1 in. width by 0.2 in. thick and the lengths varied (2, 3, and 5 in.) (Figure 1). Two types of tires (passenger tires, combination of truck and passenger car tires) with and without steel belts were used. The steel belts were left in the TDS to improve the stiffness of the tire fibers and reduce the cutting cost of the scrap tire byproduct. Fresh concrete properties showed that the workability of the mixes was acceptable with or without the tire byproducts. Testing included a scanning electron microscopy study that was used to evaluate the cement–rubber interface after hydra- tion. A high concentration of calcium oxide crystals was found on the rubber surface. Silicon and aluminum oxides were also observed. Dynamic testing of samples was used to calculate the com- plex modulus of the mixes as well as the ability of the mixes to dissipate elastic energy at low frequency dynamic loads. Typi- cal results of decreasing mechanical properties with increas- Properties Crumb Rubber Aggregates No. 6 (3.36 mm) No. 25 (0.707 mm) No. 6 + No. 25 Coarse Fine Bulk Specific Gravity 0.96 0.62 0.77 2.68 2.43 Bulk Specific Gravity, SSD 0.97 0.62 0.78 2.69 2.47 Apparent Specific Gravity 0.97 0.62 0.78 2.7 2.55 Water Absorption, % 0.92 1.05 0.95 0.25 2.04 Fineness Modulus 4.93 2.83 3.77 — 2.9 After Sukontasukkul (2009). — = data not provided. TABLE 22 PHySICAL PROPERTIES OF FINELy GROUND CRUMB RUBBER After Sukontasukkul (2009). Type Crumb Rubber, % Density, lb/ft3 Porosity, % Control 0 157.9 9.35 No. 6 (3.36 mm) 10 135.5 8.48 20 131.7 6.50 30 126.7 6.38 No. 25 (0.707 mm) 10 130.5 6.79 20 123.0 5.90 30 113.6 5.81 No. 6 and No. 25 (combination) 10 131.1 7.91 20 120.5 7.02 30 118.6 6.09 TABLE 23 POROSITy OF PCC MIxES WITH CRUMB RUBBER Type Crumb Rubber, % Thermal Conductivity, W/ m. K Heat Transfer, W/ h Heat Resistivity, m 2 /KW Control 0 0.531 1514 0.19 No. 6 (3.36 mm ) 10 0.443 1263 0.23 20 0.295 841 0.34 30 0.241 687 0.41 No. 25 (0.707 mm ) 10 0.290 827 0.34 20 0.275 784 0.36 30 0.267 761 0.37 No. 6 + No. 25 10 0.313 893 0.32 20 0.304 867 0.33 30 0.296 844 0.34 After Sukontasukkul (2009). W/m.K − 6.9335 BTU/hour-foot-˚F. TABLE 24 THERMAL PROPERTIES OF PCC MIxES WITH CRUMB RUBBER

17 Air voids were slightly higher when the mixes included the tire byproducts. Testing of the hardened concrete included measurements of compressive strengths, tensile strengths, and modulus of elasticity. Compressive strength was reduced between 35% and 47%, depending on the size and source of the tire fibers (Table 26). The strength decreased with the increasing length of tire fiber. This is contrary to the typical results seen for conventional fiber-reinforced concrete. It was noted that the longer tire fibers tended to agglomerate and be nonuniformly dispersed in the mix during vibration and compaction of the fresh PCC. There was a corresponding decrease in modulus. The range of the decrease was between 10% and 22%. The load and displacement of the samples during the split- ting tension testing showed that the concrete with the tire fibers had a somewhat lower peak load, but a significant increase in the area under the load-displacement curve. This was an indi- cation of an improvement in the toughness of the mix. That is, the tire fibers had the capability of absorbing dynamic loads and resisting crack propagation. Finite element modeling was used to show that the tensile stress concentrations were significantly smaller in the tire fiber mixes than either the control or the tire chip. Retaining the steel in the tire fibers increased the stiffness of the tire byprod- uct used and was considered a desirable property. A restriction on the length (i.e., aspect ratio) to the 2 in. length to maximize benefits and minimize the agglomeration of the tire fibers was recommended. crumb rubber and tire derived Aggregate in mortars yilmaz and Degirmenci (2009) conducted research in Turkey to evaluate the combined use of portland cement (10%), Material Crumb Rubber, % Frequency, Hz 125 250 500 1000 2000 4000 Control 0 23.0 11.5 6.8 24.5 9.1 20.1 No. 6 (3.36 mm) 10 23.0 12.0 12.1 31.5 17.0 25.0 20 23.5 11.3 9.5 37.0 15.1 24.0 No. 25 (0.707 mm) 10 24.5 11.0 10.0 26.5 16.0 27.5 20 24.5 11.3 9.3 29.0 23.5 27.1 No. 6 + No. 25 10 25.1 11.5 9.5 29.0 15.1 24.8 20 25.5 12.2 13.5 30.1 20.3 30.0 After Sukontasukkul (2009). Measurement was carried out in the following condition: temperature: 23.0 ± 2.0 C; pressure: 1013 ± 15 hPa; relative humidity: 50.0 ± 15.0%. Uncertainty of measurement: the uncertainty stated in the table is the expanded uncertainty obtained by multiplying the standard uncertainty by the coverage factor, k = 2. It has been determined in accordance with EA publication EA-4/02 ‘‘Expression of the Uncertainty of Measurement in Calibration” and ‘‘Guide to the Expression of Uncertainty in Measurement”. The obtained values lie within the assigned range of value with a probability of 95%. TABLE 25 ACOUSTICAL PROPERTIES OF PCC MIxES WITH CRUMB RUBBER With steel belt wires 25 mm 76 mm (a) Waste tire chips (b) Waste tire fibers FIGURE 1 TDA chips and fibers used in Li et al. study (2004).

18 Class C fly ash (60% and 70%), and crumb rubber (20% and 30%) for use as an acceptable mortar. The unit weight decreased and water absorption increased with increasing crumb rubber percent. The crumb rubber increased the water to cementitious material ratio. Hardened property results showed that the compressive strength of the mixes decreased with increasing crumb rub- ber content. Compressive strengths were only slightly influ- enced by the TDA size (Table 27). The combined use of the fly ash and the crumb rubber resulted in less of a decrease in strength. The optimum content of 20% crumb rubber, 10% fly ash, and 70% cement provided the best properties for mortars. The flexural strength of the mix was improved with the use of crumb rubber at 20%, but was decreased at 30% crumb rubber. Figure 2 shows the distribution of the particles in the PCC. tire derived Aggregate in portland cement concrete A California Integrated Waste Management Board (Cheng 2006) conducted a study on the use of shredded tires in a dem- onstration slurry cutoff wall (i.e., moisture barrier) in Gridley, California. The function of the cutoff wall was to form a seep- age barrier to stop the migration of water through an impervi- ous barrier. The mix design for the slurry evaluated the use of TDA from 2 in. by 2 in., to 8 in. by 8 in. in sizes. Laboratory mixes were prepared in a drum mixer. A medium scale test was conducted by constructing two slurry walls: one without TDA and one with TDA. Initial testing showed that the larger size particles would not pro- duce an acceptable mix because the slump was difficult to measure and samples for compression testing could not be Waste Tire Type TDA Fiber Di me nsions, in. PCC Properties Slum p, in. Air content, % Com pressive strength, psi Modulus of elasticity, psi Tensile strength, psi Control PCC NA 5.0 4.5 5,656 5,076 464 Mixed Truck and Car Tire Chips With Steel Belt Wi res 1 x 1 x 0.2 4.9 4.1 3,336 4,206 348 2 x 0.2 x 0.2 6.0 5.0 3,626 4,496 377 Car Tire Fibers With Steel Belt Wires 2 x 0. 2 x 0.2 5.8 5.0 3,336 4,351 363 Car Tire Fibers Without Steel Belt Wires 1 x 0.2 x 0.2 5.8 5.0 3,336 3,916 334 3 x 0.2 x 0.2 5.5 5.0 3,046 4,148 305 5 x 0.2 x 0.2 6.0 5.0 2,901 4,134 334 After Li et al. (2004). NA = not available. TABLE 26 FRESH AND HARDENED TDA FIBER PCC RESULTS Mi x Crumb Rubber Range of Sizes, mm Crumb Rubber, % Cement, % Fly Ash, % W/(PC+FA) Dry Unit Weight, kg/m 3 Water Absorption, % Compressive Strength 28 days, MPa Flexural Strength 28 days, MPa Group I 0 to 0.25 20 10 70 0.55 1,078 30 4.63 0.94 0.25 to 0.50 0.50 1,192 26 4.81 1.17 0.50 to 1.0 0.46 1,200 24 4.84 1.40 Group II 0 to 0.25 30 10 60 0.63 1,011 29 3.94 0.93 0.25 to 0.50 0.57 1,096 27 4.37 1.05 0.50 to 1.0 0.52 1,144 25 4.83 1.11 After Yilmaz and Degirmenci (2009). 1 MPa = 145.0377 psi. TABLE 27 MIx VARIABLES AND HARDENED CRUMB RUBBER PCC PROPERTIES

19 fabricated. The minus 2 in. square particle size was selected for use in the project. Both mixes were tested for permeabil- ity and compressive strength. The results showed that the TDA slurry had acceptable properties. The final mix design provided to the contractor was 67% soil, 3% bentonite clay, 5% cement, and 25% tire chips. Field testing of the slurry used a Kelly ball to measure workability because the tire chips tended to stack in the slump cone, which limited the usefulness of this test. The perme- ability testing was conducted on oversized cylindrical speci- mens (12 in. diameter) using ASTM D5084 to determine that the requirements of the permeability did not to exceed 5 × 10-7 cm/s. The full-scale placement was done with standard equip- ment. The contractor did not notice a significant difference in placing the TDA slurry compared with conventional slurries. Lessons learned included: • On-site truck dumps of to-be-mixed TDA resulted in TDA particles being spread around the construction site, which then had to be cleaned up. – Suggested that the TDA be pre-measured and bagged for use on site. • Limit the backfill of the trench to 1 ft below the surface. – Tire particles tended to protrude from the mix and may be a problem with the cover materials. • Clay can be used to finish filling the last 1 ft of the trench. Zheng et al. (2008) evaluated the damping characteristics of rubberized PCC, which is important for dynamic calcula- tions made during the design phase of a project. The damping characteristics were determined for a simply supported beam subjected to a flexural free vibration. Materials used for the PCC beams tested in this study were Type I portland cement, river sand, and a ground rubber with 80% of the particles with a size of 2.36 mm. The coarse ground rubber (tire chips) evaluated had sizes ranging from 15 to 40 mm with exposed steel belt strands and rubber contents of 15%, 30%, and 45% by volume. Fresh PCC properties showed that the slump decreased with increasing ground rubber content. The PCC mixtures with the finer ground rubber were more workable than those with the large size ground rubber. The unit weights, as expected, decreased with increasing TDA content. Hardened properties were determined for modulus of elas- ticity (ASTM C469) using the chord modulus approach. The modulus decreased with increasing rubber content (Table 28). The mix with 15% of the fine ground rubber had the least < 0.5 mm 0.25 to 0.50 mm 0.5 to 1.0 mm Control FIGURE 2 Distribution of crumb rubber in PCC mixes evaluated by Yilmas and Degirmenci (2009). Item Control PCC (1) PCC with Fine Ground Rubber (2) PCC with Coarse Ground Rubber (3) 15% 30% 45% 15% 30% 45% Dynamic Modulus of Elasticity, GPa 43.7 41.2 35.2 31.2 35.4 36.5 32.8 Ratio of Dynamic Modulus Changea, % — 5.72 19.45 28.60 18.99 16.48 24.94 Static Modulus of Elasticity, GPa 31.8 27.1 24.10 22.3 23.10 24.30 22.1 Ratio of Static Modulus Changeb, % — 14.78 24.21 29.87 27.36 23.58 30.50 After Zheng et al. (2008). aRatio of dynamic modulus change = [(2) (1)]/(1) x 100% for fine TDA or [(3) (1)]/(1) x 100% for coarse TDA. bRatio of static modulus change = [(2) (1)]/(1) x 100% for fine TDA or [(3) (1)]/(1) x 100% for coarse TDA. TABLE 28 DIFFERENCE OF MODULUS BETWEEN SCRAP TIRE FIBER CONCRETE AND CONTROL CONCRETE

20 reduction in modulus of any of the ground rubber PCC mixes. The elastic wave modulus was also determined for the beams. Both testing configurations produced similar values for the modulus (Table 29). Ground rubber mixes had lower moduli values than the control PCC. The beams were also used to determine the percent damping (Table 30). Higher percentages of TDA increasingly improved the damping potential. Damping increased with increasing amplitude (acceleration) and also increased with the concen- tration of the ground rubber in the mixes. The magnitude of the damping was dependent on the size of the crumb rubber. Results showed a difference of about 37.4% between dynamic and static modulus for the control PCC mixes (dynamic > static). When ground rubber was used, the differ- ence increased to about 50%. Use of the ground rubber sig- nificantly improved the damping ratios with the coarse ground rubber at 40% providing the most damping. The relationship between damping characteristics and ground rubber percent is nonlinear. The optimal percent would be less than 30% to optimize both the static and dynamic properties. Damping properties of the ground rubber mixes were more sensitive to vibrations response amplitude than the control mix. Ganjian et al. (2009) conducted research using Iranian standard materials. Materials used in the study were crushed siliceous aggregates, portland cement, super plasticizers, and coarse TDA and crumb rubber powder. The scrap tire materi- als were used as a cement replacement. At 5% replacement, the compressive and tensile strengths were only slightly reduced compared with the control. For this use of scrap tires, both the permeability and water absorption of the mixes increased with time and increasing percent of rubber. Oikonomou and Mavridou (2009) conducted research in Greece that evaluated the ability of a fine, primary two-sized crumb rubber (median diameter of 0.3 mm) to enhance PCC resistance to chloride ion penetration. Materials in the study were portland cement, siliceous sand, and crumb rubber with a gradation close to the sand, and a range of admixtures, superplasticizer, SBR latex, and 60% anionic bitumen emul- sion. Water to cement ratios were a constant except for two of the mixes. PCC Mix Elastic Wave Method, GPa Beam Element Method, GPa Difference, % Control 43.7 41.7 4.54 Fine TDA, 15% 41.2 38.2 7.21 Fine TDA, 30% 35.2 32.6 7.30 Fine TDA, 45% 31.2 28.1 9.89 Coarse TDA, 15% 35.4 32.6 7.89 Coarse TDA, 30% 36.5 33.5 8.12 Coarse TDA, 45% 32.8 29.7 9.32 After Zheng et al. (2008). TABLE 29 COMPARISON OF DyNAMIC MODULUS By DIFFERENT METHODS USING GROUND RUBBER CONCRETE Test no. Control Fine Ground Rubber, 15% Fine Ground Rubber, 30% Coarse Ground Rubber, 45% Amplitude acceleration, mg Damping, % Amplitude acceleration, mg Damping, % Amplitude acceleration, mg Damping, % Amplitude acceleration, mg Damping, % Fine TDA 1 8.9 0.45 8.3 0.62 12.1 0.75 10 0.7 2 14.3 0.54 15.7 0.73 18 0.88 21.2 0.98 3 19.4 0.61 25.1 0.85 27.3 1.05 23.7 1.01 4 31.2 0.74 30.5 0.95 34.8 1.17 33.2 1.14 5 40.2 0.74 40 1.12 42.3 1.42 42.1 1.42 Coarse TDA 1 8.9 0.45 12 0.68 9.1 0.89 9.8 0.85 2 14.3 0.54 17.6 0.78 17.3 1.41 20.8 1.39 3 19.4 0.61 32.6 0.93 33.2 1.45 22.3 1.49 4 31.2 0.74 40.2 1.16 38.1 1.61 31.3 1.55 5 40.2 0.74 45 1.38 40.5 1.74 41.5 1.67 After Zheng et al. (2008). 1 g = 32.17405 ft/s2. TABLE 30 DAMPING RATIO FOR PCC MIxES WITH TDA

21 All laboratory tests were conducted for mortar samples. All samples were moist cured for 28 days at 68°F at greater than 95% humidity. Workability (flow) of the mortar decreased with the increasing percent of rubber. Compressive strength, flexural strength, and modulus decreased with increasing crumb rubber content (Tables 31 and 32). Water absorption decreased with increasing crumb rubber percents and at the same resistance as chloride ion penetration increased. The mix with 12.5% crumb rubber had the least reduction in com- pressive strength, although the reduction was still significant. The next best combination of admixture and crumb rubber was with the bituminous emulsion. Conclusions were that crumb rubber concrete would be a good product in applica- tions needing a high resistance to chloride penetration but with a minimal strength requirement. Kaloush et al. (2005) conducted a study where the objec- tives of the Arizona research were to evaluate the proper- ties of crumb rubber concrete and work towards the eventual development of a specification for nonstructural and low loading conditions. Materials included testing compression and three-point bending flexural testing. The crack mouth opening deformation was measured using the three-point beam with an initial notch of 0.5 in. Coefficient of thermal expansion was determined using AASHTO TP60-00. Indi- rect tensile stress was determined using disc specimens with the thickness of about 1 in. with a 4 in. diameter. The ability of the mixes to deform (strain) improved by 39% to 116% when rubber was included in the mix. Findings were that the crumb rubber in the concrete improved the ductility and toughness compared with the control mixes. Although the tensile strength decreased with increasing rubber content, so did the ability of the mix to deform before failure. Rubber concrete was more resistant to length changes owing to tem- perature changes. portland cement concrete pavements Hernandez-Olivares et al. (2007) conducted research in Spain that evaluated the use of the Westergaard equations to calcu- late the minimum thickness for tire rubber-reinforced concrete pavements. Theoretical results indicated that the dynamic loading of a concrete beam can be used to estimate the flex- ural stiffness and fatigue life. The analyses indicated that tire byproducts in PCC slabs on an elastic base result in a slight increase in the design thickness (about 3.5%). Properties Mixture Properties Crumb Rubber, % 0.0 2.5 5.0 7.5 10.0 12.5 15.0 Specific Density, gm/cm3 2.23 2.11 2.03 1.94 1.84 1.76 1.68 Compressive Strength, MPa 40.75 30.92 21.33 16.15 11.12 9.70 8.60 Flexural Strength, MPa 9.00 7.50 5.70 5.25 4.30 3.50 2.90 Absorption of Water by Immersion Under Vacuum, % 8.81 8.25 7.37 7.25 7.03 6.87 6.79 Charge Passed, Coulombs 6,103 5,265 5,080 4,551 4,257 3,956 3,915 After Oikonomou and Mavridou (2009). 1 MPa = 145.0377 psi. TABLE 31 PROPERTIES FOR PCC WITH ONLy CRUMB RUBBER AT VARIOUS CONCENTRATIONS Properties Mixture Crumb Rubber, % 0.0 12.5 0.0 12.5 0.0 12.5 Super Plasticizer, % 1 1 0 0 0 0 Latex, % 0 0 5 5 0 0 Bitu mi nous Em ulsion, % 0 0 0 0 5 5 Specific Density, gm /cm3 2.26 1.79 2.18 1.70 2.21 1.75 Dynamic Modulus of Elasticity, GPa 42.48 15.37 35.20 11.38 39.53 13.47 Compressive Strength, MPa 43.70 13.68 35.83 8.95 42.89 12.79 Flexural Strength, MPa 10.26 4.50 8.36 3.20 10.11 4.10 Absorption of Water by Immersion Under Vacuum , % 7.91 6.25 6.79 4.92 6.96 5.01 Charge Passed, Coulombs 5,910 3,640 5,334 2,824 5,208 2,692 After Oikonomou and Mavridou (2009). 1 MPa = 145.0377 psi. TABLE 32 PROPERTIES FOR PCC WITH A RANGE OF ADDITIVES

22 Bound ApplicAtions—AsphAlt cement And concrete Four main categories for using TDA or crumb rubber in asphalt binders and concrete were found in the literature: • CRM asphalt binders • Crumb rubber and TDA fiber-modified HMA for noise reduction • Crumb rubber and TDA-modified HMA • CRM asphalts in surface treatment applications. It is important that the reader note that there was no stan- dard use of terms for scrap tire use in asphalt applications. For example, VSS (2010a) uses the term “asphalt rubber” to mean styrene-butadiene-styrene latex polymer BASF Nx1118. For this section, CRM will be used to mean scrap tire rubber with steel or fabric removed and in small aggre- gate shapes typically less than 0.5 in. TDA will be used for the larger size shape scrap tire byproducts. CRM asphalt binders can be produced using one of two wet processes (Santucci 2010). The original definition of CRM binder, called asphalt rubber, is defined in ASTM 6114 as “A blend of asphalt cement, reclaimed tire rubber, and certain additives in which the rubber component is at least 15 percent by weight of the total blend and has reacted with the asphalt sufficiently to cause swelling of the rubber particles.” Typical ranges of CRM are 18% to 22%. Extender oils may or may not be used to reduce the viscosity and promote workability. The original wet process requires that the blending and reaction chamber equipment be added to the on-site HMA plant equipment. This type of configuration limits the pro- duction rate of CRM HMA to the production rate of the blending equipment producing the CRM binder. It can be noted that the crumb rubber can substantially dissolve in the asphalt with sufficient time. The extent of the asphalt–rubber interactions will be dependent on the asphalt cement chemis- try, time allowed for interaction (blending), and temperature. In the middle 1980s, terminal blending of the CRM and asphalt started to be produced by introducing the crumb rub- ber into the asphalt as it is loaded at the refinery or blending terminal and shipped to the HMA plant as a finished product (Paramount Asphalt 2008). Initially, the percentage of CRM was kept to about 10%; however, several projects have been placed recently using between 15% and 18% CRM. Termi- nal blending also allows CRM binder to be provided to mul- tiple job sites because the production is not limited to equip- ment located at one HMA plant. Terminal blend CRM binders use a fine mesh crumb rubber to promote the asphalt–rubber interaction and shorten the required interaction time. The dry process of using crumb rubber and larger TDA particles in HMA applications use the byproduct as an aggre- gate and consider there is only a limited interaction between the crumb rubber and/or TDA (FHWA 2010). The perfor- mance of highway applications using the dry process has been mixed. crumb rubber and tire derived Aggregate-modified Asphalt Binders Lee et al. (2008) studied the influence of CRM reaction times on binder properties and molecular size changes using gel permeation chromatography. Seven reactions times (5, 30, 60, 90, 120, 240, and 480 minutes) at 177°C with 10% crumb rubber by weight of binder using a performance grade (PG) 64-22 virgin asphalt were blended with a high shear radial flow mixer. The 30-min reaction time was the reaction time used in field applications in South Carolina. The unmodified asphalt binder was also subjected to the same mixing times and temperatures so that the influence of heat hardening of the asphalt could be factored out of viscosity changes result- ing from the crumb rubber. Gel permeation chromatography analysis was used to quan- tify large molecular size changes with reaction times because these sizes have been shown to have a good relationship to the aging characteristics of the asphalt cement. Molecular sizes increased with increasing aging and reaction times for both the unmodified, but aged, and the CRM-modified asphalt. That is, the inclusion of the CRM did not influence the aging char- acteristics of the virgin binder. The rheology testing showed little change in the PG maximum acceptable summer pavement temperature grading. Viscosities increased substantially up to 60 min of reaction time after which the viscosity was not significantly different. MacLeod et al. (2007) used three grades of virgin binders (150-200 pen, 200-300 pen, and 300-400 pen) modified with one CRM with an average particle size of approximately 0.85 mm to evaluate the impact of the virgin asphalt grades on binder properties. A range of CRM concentrations in each of the virgin binders was evaluated for their influence on the Superpave binder properties and grading. Modified binders were mixed using a low shear mixer for 3 or 6 h at 180°C. The ability to store the CRM binders at elevated temperatures without separation was also evaluated. Results showed that the true Superpave grades had increas- ingly wider ranges of in-service temperatures with increasing CRM concentrations (Table 33, 3 h blending; Table 34, 6 h blending). The maximum summer temperature increased by about 1.2°C to 1.5°C for each 1% of CRM. The low win- ter temperature was decreased by about 0.2°C for each 1% of CRM. The maximum percentage of 10% CRM in the binder was recommended because of the maximum allow- able viscosity for workability (pumpability). This level of CRM would generally result in an increase in the maximum allowable summer temperature about 15°C above that of the virgin binder.

23 VanTimmeren (2009) reported on the initial development of pelletized asphalt-lime-rubber to provide for the easy use of crumb rubber asphalt without having to blend on-site. The pellets can be added at the reclaimed asphalt pavement (RAP) feed in HMA plants. The lime used was at the same percentage typically used as an anti-strip in HMA. Paramount Asphalt (2008) noted that terminal blend tech- nology has been around since the mid-1980s. In this process, the crumb rubber was blended into the asphalt binder at the asphalt terminal or refinery and shipped to the HMA plant as a finished product. Percentages as high as 25% have been used, but a minimum of 10% was typically used. A labora- tory study conducted in 1998 at the University of Nevada, Reno, compared terminal blend HMA with the conventional on-site CRM asphalt process. No statistical difference was seen between the two methods of producing the rubber- modified binders for mix fatigue and rutting. A field evalu- ation of the performance of the terminal blend mix showed very good resistance to thermal cracking, moisture damage, and rutting. xiao et al. (2009) evaluated binder property changes when using warm mix asphalt (WMA) products to reduce the mix- ing and compaction temperatures. WMA products allowed for lower production and placement temperature, which is important in reducing the emissions. A limited number of CRM HMA samples were prepared and used to verify the binder properties used to indicate fatigue life to that of the WMA CRM HMA mixes. Materials used in the project were one virgin asphalt (PG64-22), one minus 40 mesh crumb rubber, and two sources of aggregates (granite, schist) meeting a 12.5 mm gradation. WMA products used were Asphamin (zeolite) and Sasobit (paraffin). The zeolite contained about 21% crystalline water by weight, which produces fine foamed asphalt when mixed with the hot asphalt. The paraffin was a long chain Testing Sample No. 1441 2166 2165 2200 2150 2147 RM content in 200/300 pen asphalt, % 0 6 9 10.5 12 15 180°C mixing temp., mixing time, hours 0 3 3 3 3 3 Standard Tests Penetration at 25°C, 100 g/5 s, dmm 260 162 140 125 114 94 Softening point, °C 37.2 44 47 49.3 52.2 56.9 Flash point, °C 261 290 285 271 — — Superpave Tests Original binder properties Viscosity at 135°C, MPa·s 200 600 996 1,423 3,943 7,890 Dynamic shear (G*/sin), kPa min. 1.0 kPa 1.02 1.05 1.02 1.09 1.01 1.09 Temperature, °C 53 63 67 69 74 82 Rolling Thin Film Oven Test (RTFOT) RTFOT mass loss, % –0.844 –0.880 –0.760 –0.660 –0.750 –0.470 Dy namic shear ( G */sin), kPa min. 2.20 kPa 2.2 2.46 2.25 2.35 2.21 2.34 Temperature, °C 54 60 65 65 69 73 Pressure Aging Vessel (PAV) Residue PAV aging temperature, °C 90 100 100 100 100 100 Dy namic shear ( G */sin), kPa max. 5000 kPa 3,490 4,192 4,331 3,708 4,372 4,543 Temperature, °C 13 10 7 10 7 4 Creep stiffness, S , at 60 s max. 300 MPa 278 254 212 214 208 194 M value at 60 s min. 0.300 0.32 0.307 0.306 0.308 0.305 0.309 Temperature, °C –26.0 –27.0 –27.0 –28.0 –28.0 –29.0 T critical °C –38.0 –39.1 –39.2 –39.9 –39.2 –40.7 Superpave grading PG52-34 PG58-34 PG64-34 PG64-34 PG64-34 PG70-34 True Superpave grading PG53-36 PG60-37 PG65-37 PG65-38 PG69-38 PG73-39 High–low temperature spread using bending beam rheometer low-tem perature parameter, °C 89 97 102 103 107 112 High–low temperature spread using critical cracking temperature, °C 91 99 104 105 108 113 After MacLead et al (2007). High–low temperature spread is the range of temperature represen ted by the PG grade. PG53-36 has a high critical temperature of 56°C and a low critical temperature of –36°C; therefore, the range is 91°C. TABLE 33 SUPERPAVE RESULTS OF 200/300 ASPHALTS MODIFIED WITH 0% TO 15% CRUMB RUBBER MATERIALS WITH 3.0 HOUR MIxING TIME AT 180°C

24 aliphatic hydrocarbon obtained from coal gasification using the Fischer-Tropsch (FT) process. Superpave binder testing showed the viscosity at 135°C increased with the addition of the CRM, however, both WMA products reduced the viscosity when compared with the CRM without the WMA. The anticipated high summer pavement temperature (64°C) showed CRM significantly increased the stiffness of the binder. The binders with the WMA prod- ucts showed an increase in the maximum high temperature of 27% (Asphamin) and 41% (Sasobit). Stiffness at the intermediate temperature (25°C) showed a lower stiff- ness for the CRM binders compared with the unmodified PG64-22. This indicated a more flexible binder at these temperatures than the unmodified binder, which can result in a reduction in fatigue cracking of the pavements. The stiffness at the cold temperature (-12°C) was significantly lower for the CRM mixes with either of the WMA prod- ucts. The WMA CRM binders had stiffness values that were lower than the CRM-only binder by 36% (Asphamin) and 76% (Sasobit). Beam fatigue testing was used to verify the impact of the WMA CRM and CRM-only binder results on the mix fatigue properties (20°C, 10 Hz sinusoidal loading frequency). In general, mixes using CRM showed improved fatigue resis- tance. The large variability in the results made it difficult to evaluate the influence of the WMA products. crumb rubber and tire derived Aggregate- modified hot mix Asphalt for noise reduction Sacramento County (1999) reported on the six-year retention of noise reduction for Sacramento County projects on CRM HMA pavements placed on an expressway. The noise reduc- tion was maintained over six years, primarily in the 500 to 4,000 Hz frequency bands (Table 35). Testing Sam ple No. 1441 2166 2165 2200 2150 2147 CRM content in 200/300 asphalt, % 0 6 9 12 15 0 180°C mi xing te mp ., mi xing ti me (h) 0 6 6 6 6 0 Standard Tests Penetration at 25°C, 100 g/5 s, dmm 260 178 145 118 94 260 Softening point, °C 37.2 42.7 46.2 49.9 58.8 37.2 Flash point, °C 261 283 287 283 283 261 Superpave Tests Original binder properties Viscosity at 135°C, MPa·s 200 593 814 3257 8533 200 Dynam ic shear ( G */sin), kPa mi n. 1.0 kPa 1.02 1.05 1.03 1.01 1.04 1.02 Tem perature, °C 53 61 67 73 82 53 Rolling Thin Film Oven Test (RTFOT) RTFOT mass loss, % –0.844 –0.630 –0.710 –0.810 –0.460 –0.844 Dynam ic shear ( G */sin), kPa mi n. 2.20 kPa 2.2 2.29 2.4 2.36 2.38 2.2 Tem perature, °C 54 60 63 67 74 54 Pressure Aging Vessel (PAV) Residue PAV aging temperature, °C 90 100 100 100 100 90 Dynam ic shear ( G *sin), kPa ma x. 5000 kPa 3490 3727 4702 3486 2989 3490 Tem perature, °C 13 10 7 7 7 13 Creep stiffness, S , at 60s ma x. 300 MPa 278 265 240 193 171 278 M value at 60 s mi n. 0.300 0.32 0.302 0.312 0.319 0.307 0.32 Temperature, °C –26.0 –28.0 –28.0 –27.0 –28.0 –26.0 T critical, °C –38.0 –39.2 –40.2 –39.5 –40.0 –38.0 Superpave grading PG52-34 PG58-34 PG58-34 PG64-34 PG70-34 PG52-34 True Superpave grading PG53-36 PG60-38 PG63-38 PG67-37 PG74-38 PG53-36 High–low temperature spread using bending beam rheo meter low- te mp erature parameter, °C 89 98 101 104 112 89 High–low temperature spread using critical cracking tem perature, °C 91 99 103 106 114 91 After MacLeod et al. (2007). High–low temperature spread is the range of temperature represented by the PG grade. PG53-36 has a high critical temperature of 56°C and a low critical temperature of –36°C; therefore, the range is 91°C. TABLE 34 SUPERPAVE RESULTS OF 200/300 ASPHALTS MODIFIED WITH 0% TO 15% CRUMB RUBBER MATERIALS WITH 6.0 HOUR MIxING TIME AT 180°C

25 WSDOT (2005) evaluated various pavement surfaces for their ability to mitigate tire–pavement noise. WSDOT noted that the sources of traffic noise help explain why noise can be reduced in some areas by surface type selection. Cars emit noise mostly from zero to 2 ft above the pavement surface from the vehicle–pavement interaction. Typical levels of noise are between 72 and 74 dB(A) at 55 mph at a distance of 50 ft. Medium sized trucks emit the most noise between 2 and 5 ft above the surface as a result of vehicle–pavement interactions and engine exhaust noise. Typical levels of noise are between 80 and 82 dB(A) at 55 mph and a distance of 50 ft. Heavy trucks emit the most noise between 6 and 8 ft above the surface from a combination of vehicle–pavement interactions, engine noise, and exhaust stack noise (about 12 to 15 ft above the surface). Typical noise levels are between 84 and 86 dB(A) at 55 mph and a distance of 50 ft. The influence of surface type selection will depend on the number and type of vehicles as well as the location of the noise receptor. WSDOT started the documentation of noise reduction for various pavement surfaces for potential consideration of sur- face design as a source of noise mitigation. To have surface design considered, three objects needed to be met. The first was that the DOT needs to certify that initial noise reduction can be achieved; second, that the noise reductions had a certain life span; and third, that the DOT had to commit to replacing the pavement with one of similar noise quality in the future. All of these requirements require documented proof. Phase one of the study was started in 2005. A summary of WSDOT pavements indicated that a range of benefits and disadvantages could be obtained with different surfaces (Table 36). Dense graded HMA tended to be quieter than average when first constructed and the sounds were at lower and less objectionable frequencies. These pavements gener- ally maintained these characteristics throughout the life of the pavement. Open graded and rubber asphalts showed bet- ter initial reductions in noise but had difficulty in maintain- ing the reduction because of climate-related factors such as surface sanding and studded tire wear in winter conditions. Freitas and Inacio (2009) evaluated the characteristics of various gap graded pavements with different asphalt using cores tested in a laboratory sound tube (Kundt’s tube). In addition to testing the cores, two types of rubberized asphalt mixtures were prepared and tested in the laboratory. The tire- surface noise is one component of the noises produced by vehicles. At speeds between 40 km/h and 110 km/h the tire- surface noise was the most prominent. The characteristics of the pavement that influence noise were aggregate gradation, texture, porosity, age, stiffness, and distresses. Porous surfaces were considered desirable for their ability to absorb noise. Porous surfaces can reduce sound by up to TABLE 35 SUMMARy OF SACRAMENTO COUNTy NOISE STUDy WITH TIME Roadway Pavement Type Time After Paving Completed Change in Noise Levels, dB Leq Alta Arden Expressway Rubberized Asphalt 1 month –6 dB 16 months –5 dB 6 years –5 dB Antelope Road Rubberized Asphalt 6 months –4 dB 5 years –3 dB Bond Road Conventional Asphalt 1 month –2 dB 4 years 0 dB After Sacramento County (1999). Pavem ent Option Initial Noise—New Pavem ent Long-Term Noise Pavem ent Lifespan (life to rem oval) Long- Term Pavem ent Cost Sound quantity (decibel change from average) Dom inant frequency or pitch Rating Sound quantity (decibel change from average) Dom inant frequency or pitch Rating Open Graded Asphalt 2 dB Lower Good 0 to 2 dB Middle/ Higher Fair/Poor Short (4 to 10 years) High Dense Graded Asphalt 1 dB Lower Good 0 to 1 dB Lower Fair Medium (14 to 18 years) Moderate Concrete 0 to +2 dB Depends on surface finish Fair Depends on studded tire damage Higher Poor Long (40 to 50 years) Moderate Rubber Asphalt (specific dB co mp arison unavailable) Lower Good Unknown Middle Fair Nighttime te mp eratures restrict material placem ent High Source : WSDOT (2005). TABLE 36 RESULTS REPORTED FOR WASHDOT NOISE STUDy (2005)

26 6 dB(A) compared with typical dense graded HMA. In porous pavements, sound waves entered the upper layer of the surface and were partially reflected and partially absorbed. The absorp- tion of the sound energy was the result of the viscous losses as the pressure wave pumps air in and out of the voids and the thermal elastic damping. Characteristics that influence absorp- tion, other than porosity, included thickness of the porous layer, flow resistivity that is a function of the aggregate gradation, and the angle of incidence of the sound waves on the surface. Mixes in the study were gap graded mixes with medium air voids (18%) and low air voids (5%). The impedance tube method was used to measure noise absorption. In this test method standing waves were created within the tube using a loudspeaker fed with sound waves (pure tones, sine wave sweeps, etc.) that contains the HMA sample. The pure tone sounds measured the maximum and minimum sound pres- sure in the tube using a microphone that can be moved up and down the inside of the vertically positioned tube. A newer method used a two microphone arrangement that could mea- sure the sound absorption characteristics obtained from the frequency response between the microphones. Typical absorption curves (frequency vs. sound) were char- acterized by the dB value where the absorption curve reaches the first of two maximums and was related to the porosity and flow resistivity of the absorbing material. The frequency at which the measured absorption curve reached it first maxi- mum and is a function of the layer thickness and tortuosity of the material also needs to be included in the characteriza- tion. The normal incidence sound absorption was evaluated for various combinations of aggregate gradation, air voids (porosity), and asphalt with and without CRM. Results for six samples for each set of mix variables were used to obtain average results. All cores were taken from the same slab. The noise measurements were very sensitive to small changes in the mix variables. The average of multiple results showed clear differences between low and medium air void levels. The higher voids clearly absorbed more sound than the lower voids. Changing the asphalt binder had little influence on sound absorption compared with the air voids. Paje et al. (2010) used four test sections, each about 200 m in length, constructed near Barcelona, Spain, to study the acoustical characteristics of gap graded mixes with different binders. Binders used in the study included polymer-modified bitumen, crumb rubber (wet process), crumb rubber binder with 1% TDA (dry process), and crumb rubber binder with 2% TDA. The crumb rubber binder used 9% by weight of binder. The TDA was used to replace fine aggregates in the gap graded HMA. The macrotexture measurements were taken with a laser texture meter for all four sections. Noise measurements were made using the close proximity testing method. Results showed that replicate sound measurements over the length of the test sections had good agreement, but the variability of the acoustical properties of the pavement within a section varied substantially. The variability was attributed to the localized variations in the mix properties that occur with normal construction and HMA production. The standard deviation of replicate measurements was less than 1 dB(A). Significant differences were observed between the different mixtures between 600 Hz and 1 kHz range. The dB(A) in this range showed a tendency to decrease with increasing rubber content. All test sections showed that the frequency spec- trum exhibited a peak of around 800 Hz. The portion of the frequency spectrum between 1,250 and 3,500 Hz represents sound associated with air pumping and other mechanisms related to air flow in and around the tire tread patterns. This is the range of frequencies that are most improved with the use of TDA in the concrete. Data indicated a strong influence by the lack of homoge- neity of the pavement surface on the sound measurements; however, there was evidence of a linear relationship between texture and noise generation. Increasing texture within the test section corresponded to increasing noise. The authors recommend that the variability in the surface texture be con- sidered when measuring pavement noise. crumb rubber and tire derived Aggregate in hot mix Asphalt The Utah DOT (2003) bulletin summarized the pros and cons of CRM HMA applications in Utah. The bulletin noted that terminal blend CRM binder was a more recent approach and could provide better incorporation of the crumb rubber into the asphalt binder. Beneficial properties attributed to the CRM HMA were improved durability of the surface course, reduc- tion in noise of between 6 and 10 dB(A) compared with trans- versely tined concrete pavements, and the reduction of land- filling used tires. Benefits could only be achieved by states with a tire recycling plant; however, Utah did not have one at the time the bulletin was developed. Limitations to the use of CRM binders in Utah were listed as: • Air temperatures at the time of construction. Utah set crit- ical temperatures as a minimum (70°F) and a maximum (>100°F), which would limit the construction season. • Dense graded CRM HMA mixes were not stable when placed in layers of less than 2.5 in. • A solid paving surface underneath was needed for good compaction. • Mixed results were reported with long-term durability in cold freeze-thaw environments. • Potential for debonding and raveling of the CRM HMA surface layer. • Similar in performance to a polymer-modified asphalt but costs from 30% to 50% more. • Needed to compete in the modified asphalt market with- out a special requirement for use.

27 xiao et al. (2007) studied the influence of combinations of CRM, RAP, and RAP-CRM on rutting resistance. Materials included in the study were one aggregate source, and two RAP sources at various concentrations (0%, 15%, 25%, and 30%) that were blended with virgin asphalt to produce a combined final PG64-22 asphalt. The crumb rubber was obtained from two methods of grinding (ambient, cryogenic) used at one of four concentrations (0%, 5%, 10%, and 15%), and at one of three particle sizes (minus 1.4 mm, 0.60 mm, and 0.425 mm). A mechanical mixer was used to prepare the binders at a tem- perature of 350°F for 30 min reaction time using 700 rpms. Results showed an increase in optimum binder content, theoretical maximum gravity, and bulk specific gravity of the compacted sample, all of which increased with crumb rubber content. Increasing the percentage of RAP when using the CRM binder increased indirect tensile strength and rut resis- tance. Increasing the crumb rubber content decreased the indirect tensile strength and creep stiffness but still resulted in increased rut resistance. Little to no difference was seen between the methods of grinding the crumb rubber or the sizes of crumb rubber particles used in this study. Swiss researchers Partl et al. (2010) evaluated the moisture sensitivity of open graded mixes with CRM binders. Materi- als used to prepare two different open graded mixtures with one of two different coarse aggregates (basalt and basalt- expanded clay combination) (Table 37), and one CRM binder (20% by weight) of bitumen. Expanded clay coarse aggre- gate was used to enhance the acoustical properties of the mix. Testing included typical volumetric mix properties and basic CRM binder properties. The coaxial shear test (CAST; Figure 3), developed at the Swiss Federal Laboratories for Materials Testing and Research in 1987, evaluated the lateral deformations of a cored sample with a donut shape under cyclic mechanical loading. Compacted samples were cored out in the center then cut into 50 mm slices so that the outer and inner diameter of the samples are 150 and 58 mm, respec- tively. The samples were glued to an outer steel ring and an inner steel core after being sealed with an epoxy resin to pre- vent the glue from entering the mix voids. The rings were used to constrain the sample from deforming in the lateral direction. A load was applied to the upper surface of the sam- ple and the vertical deformation was recorded by measur- ing the downward movement of the internal steel core. The temperature and moisture condition of the sample could be simultaneously adjusted during load applications. Data col- lected from the dynamic testing was used to calculate the water (moisture) sensitivity index and temperature sensitiv- ity index (Table 38). An elastic-based damage model was used to assess the fatigue resistance of the mixes. Conclusions were that open graded mixes showed better fatigue resistance and reduced moisture damage compared with conventional HMA mixes. Improvements were attributed to both the CRM binder and increased asphalt content. Expanded clay slightly reduced moisture sensitivity and fatigue life, but improved the resistance to temperature changes. Liseane et al. (2010) evaluated the influence of CRM binders on rut resistance with repeated simple shear testing at constant height (RSST-CH) and a loaded wheel tester in Bra- zil. Materials included ambient and cryogenic ground rubber and two virgin asphalts (50/70 and 35/50 pen grades) to pre- pare both terminal blends and field blend asphalts with vari- ous CRM contents (Table 39). Both dense and gap graded mixes were prepared and tested. Results indicated that CRM Testing Properties for CRM Mixes Mineral aggregates Expanded clay Design curve Sieves (mm) Cumulative Percent Passing, % 12 80.12 19.88 100.0 8 73.09 19.21 92.3 4 22.75 1.29 24.0 2 12.86 0.06 12.9 0.42 6.54 0.04 6.6 0.177 5.06 0.04 5.1 0.074 4.60 0.04 4.6 Mix Properties Expanded Clay, % by weight 10 CRM Binder Content, % by weight 10.1 Rubber Content 20% on bitumen Air Void Content, % 15 VMA, % 31.3 VFB, % 52 Binder Properties Penetration, 77 F, 0.1 mm 48 Softening Point, oC 59 Fraass Breaking Point, oC 14 Dynam ic Viscosity, 175 C, MP.s 1,800 Elastic Recovery, 77 F, % 70 TABLE 37 MATERIAL PROPERTIES FOR PARTL et al. (2010) STUDy

28 binders used to prepare gap graded mixes provided the best rut resistance, regardless of test method. The study was too limited to draw conclusions about the influence of the type of grinding or method of blending on the results. crumb rubber-modified Asphalts in surface treatment Applications A number of pavement surface treatments can be prepared with CRM binders: • Chip or cape seals • Rubber emulsion asphalt slurry (REAS) • Crack sealing. Chip or Cape Seals Chip seals are a commonly used preservation and/or mainte- nance highway application that consist of the spray applica- tion of the binder on a clean, dry pavement surface followed closely by an application of small, relatively one or two sized a) b) c) d) FIGURE 3 Coaxial shear test (CAST) for fatigue test (Partl et al. 2010): (a) sample cored out in the center then cut into 50 mm slices; (b) sealed with an epoxy resin; (c) samples are glued to an outer steel ring and an inner steel core; and (d) vertical deformation is recorded. Mix Air Voids, % Water Sensitivity Index* Temperature Sensitivity Index* Calculated Numbers of Cycles to Fatigue Failure Wet Dry Open Graded Basalt Aggregate with CRM Binder 15 0.05 0.24 1,052,004 1,043,738 Expanded Clay and Basalt Aggregate with CRM Binder 15 0.41 0.16 1,047,738 1,031,405 Control 13 0.79 — 1,016,567 944,244 Control 20 1.67 0.35 1,021,024 833,804 After Partl et al. (2010). *Lower index numbers indicate better performance. TABLE 38 RESULTS FROM CAST DyNAMIC TESTING

29 aggregates (chips). The chips are rolled to seat the chips in the binder. The surface may be finished with a fog seal to improve chip retention and to improve the color contrast for striping and marking. When a CRM binder chip seal is placed as an interlayer between an old pavement and a new over- lay to reduce reflective cracking, it has been called a stress absorbing membrane interface. When the same product is used as a chip seal, it has been referred to as a stress absorb- ing membrane. There has been a proliferation of names for the CRM chip seals. International Surfacing Systems (ISS 2007a,b), a major long-time supplier of CRM binders, prepared a guide speci- fication for polymer-modified asphalt rubber (PMAR) stress absorbing membrane and PMAR stress absorbing membrane interlayer. This product used a polymer as well as the crumb rubber in the binder modification. The ISS suggested specification indicates that the selec- tion of the virgin asphalt grade would be modified with the crumb rubber based on climate (Table 40). The recom- mended CRM gradation is shown in Table 41. The com- bined polymer and crumb rubber modifications should be able to produce a modified binder that meets the require- ments in Table 42. Suggested limits for PMAR binder can be found in Table 43. Paramount Asphalt (2008) documented a project in Kern County in Bakersfield, California, which used a PG76-22PM (polymer-modified) and a PG70-22TR (10% tire rubber). The asphalt distributor truck was required to heat up PG70- 22TR to 315°F from the delivery temperature of 285°F because the binder had cooled over the long haul distance. The aggregate used for the chip seal was a 4.75 to 9.5 mm chip, which was precoated with 0.25% PG64-10 for the tire rubber section and 0.5 to 0.7% for the PM section. The appli- cation rates were 0.32 to 0.35 gal/yd2 for the binder and 17 to 19 lb/yd2. The ambient temperature the morning the chip seals were placed started at 68°F. The construction of both sections showed no difference in handling between the two binders. The chip seal was fogged with 0.07 gal/yds2 Topein C Mix Variables Binder Designations ARTB1 ARTB2 ARCB1 ARCB2 Control Asphalt Ce me nt Grade 50/70 pen 50/70 pen 35/50 pen 35/50 pen 50/70 pen Type of Grinding Process Am bient Am bient Cryogenic Am bient NA Concentration of Crum b Rubber, % 20 15 21 21 NA Method of Modifying Binder Term inal blend Term inal blend Continuous Continuous NA Binder Results after RTFOT Change of Mass, % 0.3 0.3 0.9 0.2 0.3 Softening Point Elevation, o C 1.0 2.9 17.2 11.2 4.3 Penetration 25 C,100 g, 5 s, dmm 28.8 25.3 15.5 19.5 22.3 Apparent Viscosity, cP, 175 o C 5,350 1,962 3,025 8,813 96 Retained Penetration, % 72 60.2 92.2 99 43.3 After Liseane et al. (2010). NA = not applicable; RTFOT = rolling thin film oven test. TABLE 39 SUMMARy OF BINDER PROPERTIES FOR TERMINAL AND CONTINUOUS BLEND CRM BINDERS After ISS (2007a,b). Sam = stress absorbing membrane; SAMI = stress absorbing membrane interface. Climate PG Gra ding Cold PG52-28 Moderate PG58-22 Hot PG64-16 or PG70-10 TABLE 40 ISS GUIDE FOR BINDER GRADES FOR SAMS AND SAMIS After ISS (2007a,b) Sam = stress absorbing membrane; SAMI = stress absorbing membrane interface; SBS = styrene-butadiene- st yr ene; PMAR = polymer-m odified asphalt rubber . Sieve Size Reclaim ed Tire CRM Percent Passing SBS Polym er Percent Passing #8 (2.36 mm ) 100 Per PMAR binder manufacturer #10 (2 mm ) 95–100 Per PMAR binder manufacturer #16 (1.18 mm ) 45–75 Per PMAR binder manufacturer #30 (0.600 mm ) 2–20 Per PMAR binder manufacturer #50 (0.300 mm ) 0–10 Per PMAR binder manufacturer #200 (0.075 mm ) — — TABLE 41 SAM AND SAMI CRM AND SBS POLyMER REqUIREMENTS

30 product, which was a standard practice in Kern County. Both products were performing well with no noticeable differ- ence between the two binder products. Updyke (2009) placed terminal blend test sections in Los Angeles County to evaluate the impact of the existing pavement condition on the performance of the chip seal. Before placing the terminal blend binder chip seal, one existing pavement sec- tion used a fully repaired pavement section, the second section only received minor rut filling and edge restoration, and a third section used a pavement fabric. The terminal blend binder was a PG76-22TR that contained 15% crumb rubber. Certificates of compliance were furnished for each truck load of CRM binder. Precoated 9 mm chips were used for all sections. The terminal blend binder over the fabric was the only section that had some construction concerns. The hot CRM binder temperatures resulted in some wrinkling of the fabric and the fabric tended to adhere to the pneumatic roller tires. The quality assurance testing evaluated the aggregate and binder content. Rubber Emulsion Asphalt Slurry (REAS) Phetcharat and Kongsuwan (2003) evaluated asphalt emulsion (CSS-1h emulsion with 60% residual asphalt) with varying con- centrations of crumb rubber (100% passing the 0.6 mm sieve) blended for 10 min in the laboratory at 400 rpm. The general conclusions were that REAS binder helped the flexibility of the surface treatment, which can be expected to be seen in field applications as a higher resistance to reflective cracking. For quality control purposes, VSS (2010b) noted that they test the crumb rubber at least once for every 250 tons; a minimum of once per project. REAS consisted of a combi- nation of passenger and heavy vehicle (high natural rubber) crumb rubber. The steel and fibers, which were considered contaminates, were removed before grinding. Contaminates were limited to no more than 0.01% wire and no more than 0.05% fabric by weight of CRM. The crumb rubber particle length was not more than 3/16 in. The quality control testing recommended by VSS indicates that: • Gradation tests for every truck of CRM delivered to a project are appropriate. • Asphalt modifier, a resinous high flash point aromatic hydrocarbon compound, be tested at least once every 25 tons. Test results should be provided with certificate of compli- ance. Up to 3% talc may be used to prevent CRM from stick- ing together and bagged for use. The CRM needed to be dry so as to not generate foaming when introduced into asphalt and conform to ASTM D297. Updyke (2009) investigated the use of terminal blend binders to produce an asphalt emulsion slurry seal. This pro- cess is a relatively new use for crumb rubber and has limited performance information. Five-year warranties were used to place five projects in California during 2002 and 2004. Four of the projects used asphalt rubber and the fifth used a termi- nal blend. Currently, all projects are performing well and are expected to meet their 10-year design life. After ISS (2007a,b). Test Minim um Maxim um Acetone Extrac t 6.00% 16.00% Ash Content — 8.00% Carbon Black Conten t 28.00% 38.00% Rubber Hydrocarbon 42.00% 65.00% Natural Rubber Content 22.00% 39.00% TABLE 42 RECOMMENDED CHEMICAL COMPOSITION OF CRM Test Range Hot Climate Moderate Climate Cold Climate Apparent viscosity, 347°F (175°C) Spindle 3 @ 12 RPM: cps (ASTM D2669) Min. Max. 1,500 3,500 1,500 3,500 1,500 3,500 Cone Penetration, 77°F (25°C), 150 g, 5 s; 1/10 dm (ASTM D217) Min. Max. 15 40 20 70 25 100 Softening Point, °F (°C) (ASTM D36) Min. 150°F (66°C) 140°F (60°C) 130°F (54°C) Resilience, 77°F (25°C), % (ASTM D3407) Min. 40 30 20 After ISS (2007a,b). Hot climate average July max. @ 110°F (43°C). Average January low @ 30°F ( 1°C) or above. Moderate clim ate average July ma x. @ 100°F (38°C). Average January low @ 15 F–30°F ( 9 C to 1°C). Cold climate average July ma x. @ 80°F (27°C). Average January low @ 15°F ( 9°C) or lower. Note: Certain climates may overlap the above, defined areas. When in doubt of the type of asphalt cement to utilize, always look to the lower penetration materials in the hot temperature range and the higher penetration materials in the low temperature range. TABLE 43 SUGGESTED LIMITS FOR POLyMER-MODIFIED ASPHALT RUBBER BINDER

31 The Los Angeles County (2010) article noted that a REAS surface treatment was higher in cost than conventional slur- ries but was expected to have a 50% longer life, with lasting color contrast for striping and marking, and high skid resis- tance. The estimated usage of scrap tires were more than 78 tires per lane-mile. This surface treatment was included in the 1998 Green Book REAS specifications. Pacific Emulsions, Inc. (2010) described the use of a tire rubber-modified slurry seal (TRMSS) where the emulsion contains 10% crumb rubber. The CRM binder was a termi- nal blend used to formulate a cationic quick set emulsion that works well with negatively charged aggregates. This was produced with a 60% asphalt emulsion and a slurry seal Type II aggregate application rate of 13.5 lb/yd2. The rubber asphalt emulsion was used to produce a slurry seal using a Type I aggregate gradation applied at 9 lb/yd2. Advantages of TRMSS emulsion were that the cationic formulation works in cool weather as well as hot weather. Matthews (2008) reported on the use of REAS over a scrub seal with RAP chips, which was placed in southern California in May 2007. The PASS™ scrub seal was covered with the RAP chips, which was covered with the REAS to produce a cape seal (i.e., chip seal topped with slurry). Some localized traffic damage resulted during the hot summer cli- mate. The contractor’s initial conclusion was that the virgin binder used in the REAS binder may need to be harder to provide better durability at high temperatures. Crack Sealing Symons (1999) summarized the findings from the long-term pavement performance (LTPP) study on crack sealants. The findings of this large-scale research program showed that the most cost-effective sealants were those with rubberized asphalt placed in standard or shallow-recessed “band-aid” configurations, with the recessed method showing the over- all best longevity. The rubberized crack sealant life can be expected to perform for 5 to 8 years. Chehovits and Galehouse (2010) evaluated the environ- mental impact of various pavement preservation choices with wide ranges in typical quantities used and with a range of anticipated life extensions. Crack sealants and fills com- monly use CRM products. Although crack sealants and fill had a lower anticipated life extension, they had the least impact on energy usage and greenhouse gas production (Table 44). unBound ApplicAtions—BAses, emBAnkments, And Fills Barker Lemar (2005) in an Iowa study noted that an ASTM D6270 Type A TDA was used to construct an embankment with maximum 3-ft-thick layers. Additional gradation require- ments were that less than 50% was passing the 37.5 mm sieve and less than 5% was passing the 4.75 mm sieve. TDA for Class II fills were used for infill or as road base subgrade where deeper depths are required. Compaction and placement needed to be done with layer thicknesses limited to between 3 and 10 ft. The authors recommended the TDA have a maxi- mum of 25% passing the 37.5 mm sieve and a maximum of 1% passing the 4.75 mm sieve for bases and infills. Shalaby and Khan (2005) reported on Canadian research that evaluated the use of only large TDA (whole sidewalls) as a base for an unpaved roadway surface for access to a new gravel pit. The project site was northeast of Winnipeg, Manitoba, Canada, and had a subgrade of mostly a boggy area with incoming water flow from a nearby golf course. The subgrade in the area was black organic topsoil (brown clay with fine sand, silt lenses) with a plastic clay subgrade. This material had a high plasticity and ranged from soft at the near surface to firm at a depth of about 12 ft. A total of five layers of tire sidewalls (6 in. per layer) were used and the top layer was covered with 18 in. of granular surface. The sides of the base were also covered with the granular material. Sensors were placed in the base to measure frost penetration and showed that the depth of penetration was about twice that typically seen for the soil. Process Anticipated Life Extension, years Area Energy and GHG Annualized Energy and GHG BTU/yd2 lb CO2/yd 2 BTU/yd2/yr lb CO2/ yd 2/yr Hot Mix AC 5 to 10 46,300 9 4,660 to 9,320 0.9 to 1.8 HIR 5 to 10 38,700 7 3,870 to 7,740 0.7 to 1.4 Chip Seal 3 to 6 7,030 0.9 1,170 to 2,340 0.15 to 0.30 Slurry Seal 3 to 5 3,870 0.4 968 to 1,935 0.10 to 0.20 Crack Seal 1 to 3 870 0.14 290 to 870 0.05 to 0.14 Crack Fill 1 to 2 1,860 0.25 930 to 1,860 0.13 to 0.25 Fog Seal 1 500 0.07 500 0.07 After Chehovits and Galehouse (2010). GHG = greenhouse gas; HIR = hot in-place recycling. TABLE 44 PAVEMENT PRESERVATION PROCESS ENERGy AND GREENHOUSE GAS EMISSIONS PER SqUARE yARD

32 Results showed that the TDA sections had lower insulation values, which was not expected. The lack of insulation value of the tire layer was attributed to the lack of moisture content, which would be frozen throughout the winter in the soil. Laboratory testing was conducted so that a shredded tire road design model could be developed. A one-dimensional constrained compression testing was used to evaluate the resilience of the tire layer at high stress levels during cyclic loading and unloading. Tables 45 and 46 show the information required for the tire layer for a mechanistic–empirical pave- ment design approach. A Poisson’s ratio of 0.3 was assumed. The one-dimensional compressibility of the tire layers were predicted using three equations: The first equation was used to determine compressibility of the TDA: d l s= H z Where: d = compressibility of TDA, l = coefficient of TDA compressibility, H = thickness of TDA layer, and s = stress at a depth of z in the TDA layer. The relationship between stress and the coefficient used in the study was: l s s= -( ) -( )1 1 1 2 2 1h h h Where: h1 = the TDA layer thickness at s1 = 7.3 psi, and h2 = the TDA layer thickness at s2 = 29 psi. The stress at the middle of the TDA layer was obtained using the solution for axis-symmetric loading over an elastic half space: s z q z a z = - +( )     1 3 2 2 1 5. Where: q = uniform applied pressure, and a = radius of the loaded area = 6 in. Engineering Property Range of Values Remarks Size, mm 50 to 300 — Specific Gravity 1.06 to 1.15 — Water Absorption Capacity, % 3.9 Glass belted 9.4 Steel belted Uncompacted Unit Weight, lb/ft3 18.7 to 31.2 For 25–75 mm shred size Compacted Unit Weight, lb/ft3 37.5 to 43.7 For 25–75 mm shred size Compressibility, % 40 At 400 kPa surcharge Poisson’s Ratio 0.3 — Elastic Modulus, psi 159.6 75 mm size tire shreds Internal Friction Angle, o 19 to 25 — Cohesion Intercept, psi 1.2 to 1.6 — Shear Modulus, psi 391.6 75 mm size tire shreds Thermal Conductivity (W/m C) 0.20 to 0.30 At a density range of 0.58–0.79 mg/m3 Permeability (cm/sec) 0.1 Compacted After Shalaby and Khan (2005). 1 kPa = 0.145038 psi. 1 kg/m3 = 0.062428 lb/ft3. TABLE 45 SUMMARy OF ENGINEERING PROPERTIES Size of Tire Shreds, in. Resilient Modulus at Normal Stress of: (psi) 7.3 14.5 29.0 2 50.6 63.2 84.4 6 24.4 34.5 52.1 12 20.0 29.6 46.7 After Shalaby and Khan (2005). TABLE 46 RESILIENT RESPONSE OF SHREDDED TIRES FROM CONSTRAINED COMPRESSION TESTS

33 Previous testing by the researchers indicated that the pre- dicted values were a good representation of the actual defor- mations. Results showed that compression parameters of TDA can be reasonably determined from the one-dimensional con- strained compression laboratory testing. These results can be used to estimate pavement layer responses (Table 47). Initial research by Tatlisoz et al. (1997) used silty soils with TDA to provide improved soil properties (Table 48). The unit weight of the soil–TDA mixtures decreased with increasing TDA content. At the same time, the cohesion increased with increasing TDA content. Soil–TDA mixtures did not show significant long-term deformation, even when soaked. Using TDA with the soil improved the angle of inter- nal friction, regardless of TDA content. Tatlisoz et al. (1998) investigated the use of soil–TDA mixtures in geosynthetic-reinforced earthworks. The three soils used were one coarse grained and two fine grain (two sands and one clay). The clay was eliminated as early test- ing that showed clay–TDA combinations did not provide improved soil properties. One geotextile and two geogrids were used for the reinforcement. The TDA ranged in size from 30 to 110 mm. Testing used direct shear rather than triaxial cell methods because of the large size of the TDA compared with the con- ventional triaxial cell. Pullout testing was conducted using a large steel pullout box to determine the interaction coefficient, which compares the effective strength of the soil–geosynthetic interface to the shear strength of the soil. Results of the laboratory testing showed that soil–TDA mixtures have significantly higher shear strength than soil mixtures alone (Table 49). These mixes did not exhibit peak shear strength and the shear strength continues to increase with increasing displacement. The pull-out forces increase with increasing displacement. The interaction coefficients were generally around 0.5; however, the pullout capacity of the soil–TDA backfill was often equal to or greater than the pullout capacity of the soil-only backfill. Theoretically, the soil–TDA system would require fewer layers to reinforce walls constructed with soil–TDA, which was a function of the higher strength and lower unit weight. The soil–TDA system provided greater resistance against lateral sliding of embankments on geosynthetic layers because of the low active earth pressure and higher interface resistance. There was a greater resistance to bearing failure because of the lighter weight. It is important that these conclusions be validated through large-scale testing. Abichou et al. (2004) evaluated a mechanically stabilized earth (MSE) wall reinforced with geotextiles and geogrids, then backfilled with a sand–TDA mixture (25% TDA by vol- ume). Once constructed, a range of surcharges were applied to the backfill (42, 95, 148, and 200 kPa). Horizontal and verti- cal pressures and displacements were measured with sensors embedded behind the wall. Materials used in the study were a fine, uniformly graded sand and TDA with sizes from 50 to 550 mm. Laboratory results from the earlier study were used to design the MSE. Construction was completed per recommendations by the block wall supplier. The wall was reinforced with two lay- ers of geotextile and two layers of geogrids. Compaction was achieved using a vibratory plate compactor. The wall was allowed to stabilize for three months after construction before the surcharge was placed. Size of Tire Shreds (in.) Coefficient of Tir e Compressibility a (m 2 /MN) Thickness of Tire Shred Em bankm ent ( h ) Stress at Mid-layer Under 550 kPa Tire Pressur e b Stress at z (kPa) Deflection c (nm ) 2 1.1 1.5 12 19.8 6 1.9 1.5 12 34.2 12 2.2 1.5 12 39.6 After Shalaby and Khan (2005). a Stress vs. deflection response. b Boussinesq’s solution. c From one-dimensional consolidation settlement testing . TABLE 47 PREDICTED DEFLECTIONS OF TIRE SHRED EMBANKMENT USING COMPRESSIBILITy PARAMETERS After Tatlisoz et al. (1997). Sample Unit Weight, lb/ft3 c, psi Sandy Silt 116.5 230 30 Silt + 10% TDA 112.0 167 55 Silt + 20% TDA 108.2 793 54 Silt + 30% TDA 103.8 815 53 TABLE 48 SHEAR STRENGTH PARAMETERS OF SANDy SILT AND SANDy SILT–TDA MIxTURES

34 Results showed that conventional geotechnical methods had horizontal and vertical stresses behind the wall that increased with the addition of the surcharge, as expected. Displacements of the wall decreased with increasing depth and were considered small compared with the size of the surcharge. Strains were highest near the wall face. A finite element analysis model reasonably predicted the measured stresses. The frictional characteristics of the soil–TDA mix- ture along the wall were not known. The final conclusion was that conventional design methods could be used for MSE walls with soil–TDA backfill. Cetin et al. (2006) are Turkish researchers who evaluated the use of coarse and fine TDA mixed with a cohesive soil [CL by USCS (unified soil classification system)] at 10%, 20%, 30%, 40%, and 50%. Laboratory testing indicated that the permeability of the mix was dependent on the charac- teristics of the soil used in the mix. The cohesion increased with increasing TDA content up to 40% of either coarse or fine TDA. Conclusions were that the clay–TDA mixture was an acceptable fill material when using a maximum of 20% coarse graded TDA or 30% of the fine graded TDA when used above the ground water table. Neither mix should be used where drainage is needed to prevent the development of pore water pressures during loading under saturated conditions. Huat et al. (2007) evaluated the use of whole tires linked together to form a soil-filled earth wall in Malaysia. Test- ing evaluated the tensile strength (ASTM D4595) of scrap tires. Whole tires with the sidewall removed were loaded in tension at a rate of 2 in./min until failure. The study showed there was an 88% probability that the tensile strength of the available tires would be greater than 4,500 lb-force. Poly- propylene rope was evaluated as a quick, economical, and locally available means of connecting the tires. Tensile test- ing was also conducted for various methods of wrapping and tying the rope (Table 50). A full-scale field test section was constructed to evaluate the use of the proposed earth wall (Figure 4). Reinforcement mats were constructed and backfilled with native cohesive soil. Each layer was compacted with a 1-ton steel wheel roller. A total of 25 layers were used to construct the earth wall. Two vibrating wire earth pressure cells were used to Backfill Geosynthetic Normal Stress, psi Shear Strength, psi Pull-out Force Interaction Coefficient Tire Chips Geotextile 1.16 0.67 3,147 1.51 4.21 2.42 10,116 1.67 7.25 4.19 14,837 1.27 Miragrid 5T 1.16 0.67 3,822 1.95 4.21 2.86 6,969 0.99 7.25 4.18 8,992 0.72 Miragrid 12XT 4.21 2.42 7,868 1.05 Sand Geotextile 1.45 0.97 2,248 0.65 4.35 2.93 10,566 0.93 7.40 4.99 11,690 0.78 Miragrid 5T 1.45 0.97 1,798 0.73 4.35 4.38 6,295 0.63 7.40 4.99 6,969 0.47 Miragrid 12XT 4.35 2.93 5,620 0.61 Sand–Tire Chips Geotextile 1.45 1.86 4,047 0.73 4.35 5.57 9,442 0.54 7.40 5.12 14,837 0.52 Miragrid 5T 1.45 1.86 3,597 0.65 4.35 5.57 8,093 0.47 7.40 9.47 8,992 0.3 Miragrid 12T 4.35 5.57 9,892 0.57 Sandy Silt Geotextile 1.45 2.44 4,946 0.79 4.35 4.10 9,892 0.86 7.40 5.86 19,334 1.15 Miragrid 5T 1.45 2.44 4,047 0.57 7.40 5.86 9,892 0.57 Miragrid 12XT 4.35 4.10 8,992 0.71 Sandy Silt– Tire Chips Geotextile 1.45 7.59 4,496 0.2 4.35 11.43 10,791 0.31 7.40 15.48 17,535 0.38 Miragrid 5T 1.45 7.59 5,395 0.24 7.40 15.48 10,116 0.22 Miragrid 12XT 4.35 11.43 11,016 0.31 After Tatlisoz et al. (1998). TABLE 49 SUMMARy OF PULL-OUT TEST RESULTS

35 compare the theoretical with the field-measured lateral earth pressures. Instrumentation of the wall showed an initial set- tlement of about 15 mm after 10 days. The long-term settle- ment stabilized after this time at 14 mm. The conclusion was that the tire wall was in the Rankine’s active state. Tandon et al. (2007) instrumented and monitored three types of embankment backfill for the Texas DOT. The first section, a soil–TDA (50:50) mixture, was used to construct an embankment with a width of fill of approximately 20 ft, with a layer thickness of 6.6 ft using a sandy lean clay (CL by USCS). Density testing and optimum moisture contents were evaluated during placement. The second section was TDA with a geotextile section of embankment about 41 ft wide with a layer thickness of approximately 5.4 ft, which was con- structed in six 1-foot lifts. The third section used a conven- tional soil to construct an embankment with a width of 88 ft and a height of 12.6 ft, and was constructed in 1-foot-deep lifts. Testing included monitoring of vertical settlement, tem- perature change, and air and water changes. Sensor instru- ments placed during construction included inclinometers, thermal couples with data logger, moisture detection devices, air sampling ducts, and lysimeters. Little settlement was seen because of the attention to compaction during construction. Water did not drain downward through the pavement sur- face, mostly because of the low rainfall in the El Paso area. After 2 years of monitoring, the temperature in the TDA fill was only slightly higher than in the soil–TDA fill (58°F to 84°F and 67°F to 88°F, respectively). Temperatures in the TDA section had less fluctuation than the ambient tempera- tures, which is a function of the insulation properties of the TDA. Air samples were used to identify potentially flamma- ble concentrations per National Fire Protection Association (NFPA) specifications (Table 51). All of the results from the 100% TDA section were well below the specification limits. The conclusion was that there was little potential for envi- ronmental impacts for TDA fill used in the El Paso, Texas, climate region. Wartman et al. (2007) conducted a laboratory study of soil–TDA combinations using both one-dimensional confined compression (shredded tires) and isotropic compression with a triaxial cell (tire chips). TDA dimensions were an average of 1.14 cm thick, 7.53 cm in width, and 17.84 cm in length. Using the USCS, the TDA can be described as poorly graded gravel (GP) with a maximum particle size of 3 cm. The TDA had a specific gravity of 1.07 owing to some metal content still in the TDA. The soil was medium sub angular sand clas- sified as SP. The sand required a moisture content of about 2% to keep it from segregating when mixed with the TDA. The density of the TDA materials was conducted during construction of a field project so that a companion compaction method for laboratory-prepared samples could be developed. Compaction energy of 600 kN/m3 per ASTM D698 was found to provide relatively dense samples for testing (Table 52). Testing determined that the secant modulus at 50 kPa was used to represent the immediate compression modulus. Volume changes in TDA tire chips under saturated, drained, isotropic compression showed that most of the volume strain was the result of the void reduction between rubber particles. The vol- ume change from the compression of the rubber particles them- selves was small and does not contribute to the immediate vol- ume changes (Figure 5). The volume change had a decreasing Sa mp le No. Diam eter of Rope, in. No. of Wraps No. of Knots Maximu m Load, lb 1 0.5 1 1 2,898 2 0.5 1 2 5,171 3 0.5 2 2 11,690 After Huat et al. (2007). TABLE 50 TENSILE STRENGTH OF POLyPROPyLENE ROPE ATTACHMENT FIGURE 4 Construction of MSE wall (after Huat et al. 2007).

36 After Tandon et al. (2007). — = not applicable. Organic Com pound Regulatory Limits Results ppm ·10 3 Immediately dangerous to life or health concentrations, ppm Lower explosive Limit, ppm Upper explosive Limit, ppm 1,4 Dichlorobenzene 1.1 Not noted 62,000 160,000 Ichlorodifluorom ethane 3.1 15,000 — — Ethylbenzene 1.9 800 10,000 67,000 Styrene 1 700 11,000 — Tetrachloroethylene 10.3 150 — — Toluene 43.3 500 12,000 71,000 Trichloroethylene 4 1,000 120,000 400,000 Trichlorofluoro me thane 1.1 N/A — — 1,2,4-trim ethylbenzene 3.4 N/A — — 1,3,5-trim ethylbenzene 2.4 N/A — — m, p-Xylene 6.7 900 11,000 70,000 o-Xylene 2 900 10,000 76,000 TABLE 51 VOLATILE ORGANIC SCAN OF AIR SAMPLE COLLECTED FROM SITE 2 Material kN/m3 Porosityb, n Void Ratiob, e Specific Gravity Dmax, cm Tire Chips 6.46a 0.38 0.62 1.07 3 Tire Shreds 4.74a 0.63 1.71 1.31 Varied Sand 16.04 0.37 0.58 2.59 0.2 aBased on a selected compaction energy of 600 kNm/m3 bAs-compacted porosity and void ratio values. TABLE 52 PHySICAL PROPERTIES OF THE MATERIALS USED IN WARTMAN et al. STUDy (2007) linear relationship with the log of the time and the rate of volume change with time (i.e., slope of regression line) and is dependent on the percentage of TDA in the mixture (Figure 6). The rate of change became steady after 10 to 20 h of testing. Results showed that subbase placement could be used to remove the immediate volume change in the TDA soil mix- tures. The initial compaction requirement that resulted in an increase in the TDA mixture until the compaction was about 10% over that found in the laboratory-achieved densities. Volume changes are to be accounted for in any design. The layer modulus increased with the decreasing volume of the mixtures. Based on consolidation data, a period of 8 weeks FIGURE 5 Proportion of strain owing to change in void space compared with compression of tire particles (after Wartman 2007). 0 5 10 15 20 25 30 35 40 0 200 400 600 vo l,% v, kPa Total Volume Strain Void Volume Strain Strain from void reduction Strain from particle compression

37 was considered to be sufficient to allow for compression of the TDA soil mixture after the subbase surcharge has been added. Long-term field data are needed. Choi et al. (2007) reported using TDA as an insulation layer to minimize frost penetration in Virginia, and allowed for the free draining of water to prevent frost heave in frozen winter conditions. Mills and McGinn (2008) evaluated four options for a lightweight fill (Table 53). Ultimately, the TDA was selected that resulted in the use of 1.6 million scrap tires (about 2 years of discarded tires in New Brunswick). The TDA lightweight fill was placed in two separated 9.6-ft-thick layers within the soil fill. Key factors that made the use of TDA possible were the location of the recycling center and costs subsidized by discarded tire fees. The tire recycling plant was within 160 km of the embankment site. The transportation costs were offset by the use of fees collected for discarding tires, which paid about 97% of these costs. This made the transportation costs of the TDA significantly lower than any of the other options. other ApplicAtions The California Integrated Waste Management Board (2010) identified sound barriers as a use for crumb rubber PCC. The load bearing structural composite tongue and groove build- ing planks (nonrecycled) filled with shredded post-consumer tires can be used to produce a 12-ft-high sound wall with about 250,000 lb of TDA used per mile of installation. The wall com- ponents can be constructed to meet ASTM E884 with a class A rating. A fire retardant can also be incorporated into the cell. Repair Option Benefits Disadvantages 1. Rem oval of Soft Soils and Replacem ent with Im ported Granular Fill Conventional construction techniques Very low risk of future em bankm ent failure Minimizes secondary settlement Eliminates the importance of soil profile and soil param eter assu mp tions Very deep excavation would be a challenge in the soft clays and a large excavation footprint would require gaining access to private land to the south, which was unfavorable. Very large quantity of soft clay soils to dispose of off-site would be a challenge. No instrumentation required Risk of undermining adjacent highway em bankm ent Estimated cost is three times the estimated cost of the TDA option. 2. Stone Colu mn s (installed by a specialty contractor) Will allow em bank me nt to be constructed relatively quickly after treatment using conventional soils. Will allow for the original designed geom etry of the em bankm ent Large quantity of soft clay to be disposed of off-site. Specialty contractor had some concerns that the clay soils ma y be too soft in areas and would not provide the necessary confining pressure for stone columns. Estimated cost is 1.4 times the estimated cost of the TDA option. 3. Geofoam + TDA Can leave soft foundation soils in place Very light weight and easy to construct No wick drains required Geofoam on its own was considerably mo re expensive than Geofoam+TDA. Estimated cost is 1.7 times the estimated cost of the TDA option. 4. TDA Can leave soft foundation soils in place Limits extent of excavation required Most economical option Good track record and ma ny successful cases reported in the literature ASTM standards available for TDA use in civil engineering applications. TDA locally available in NB Wick drains required in co mb ination with TDA Staged construction necessary Instrumentation required to monitor foundation soil’s response to loading. NB = New Brunswick. TABLE 53 COMPARISON OF OPTIONS CONSIDERED By MILLS AND MCGINN (2008) FIGURE 6 Influence of percentage TDA in soil–TDA combina- tion on strain with time (after Wartman 2007). 0 0.25 0.5 0.75 1 1.25 1.5 0.1 1 10 100 Ti m e De pe nd en t vo l,% Time After Load Application, Days 100% TDA 85% TDA 50% TDA Sand