Crumb Rubber Modifier in Asphalt Pavement

Chapter 3

Materials and Mix Design

Each agency has employed slightly different CRM technologies in HMA. Though Arizona and California use similar suppliers of CRM and asphalt rubber, there are differences in the CRM composition size and gradation. Florida experimented with one of the Arizona technologies, then made modifications that include smaller rubber particle size, lower amounts of rubber, and lower blending temperatures. This chapter briefly describes the materials and mix design procedures. Additionally, it also summarizes accepted procedures used in the States by local agencies.

Materials

Crumb Rubber Modifier (CRM)

Tire rubber, the principal component of CRM, is primarily a composite of natural rubber, synthetic rubber, and carbon black. Historically, passenger tires contained approximately 20 percent natural and 26 percent synthetic rubber, whereas truck tires contained approximately 33 percent natural and 21 percent synthetic rubber. SUP6SUP Industry sources today indicate that passenger car tires typically contain approximately 16 percent natural and 31 percent synthetic rubber, whereas truck tires contain approximately 31 percent natural and 16 percent synthetic rubber. (7 Other sources of raw material for CRM include peel from over-the-road vehicles and buffings (a by-product of the retreading process).

Raw material may be delivered to the processing plant as whole, cut, or shredded tires or buffing waste; the form depends on the capabilities of the processing plant. Whole tires require the least amount of preprocessing but are bulky and limit shipping capacity. Tires that have been minimally processed-typically cut, split, or sectioned-improve handling and shipping. Shredded tire rubber approximately 150 mm (6 in) square is the preferred form of raw material for producing CRM. Buffing waste, because of its small size and generally high quality, is typically diverted to other rubber manufacturing processes. The quality of the raw material is a critical factor in producing a "quality" CRM and is inevitably the responsibility of the CRM processor.

Although there are several methods for processing scrap tires, the primary goal of each is to reduce the size and separate the steel belting and fiber reinforcing from the rubber. Processing scrap tires into CRM may generally be divided into two general categories: ambient grinding/granulating and cryogenic grinding.

As the name implies, ambient grinding/granulating involves tearing and shearing at room temperature. The ambient process consists of a series of crackermills or granulators, screeners, conveyors, and various types of magnets to remove steel as necessary. A schematic of a typical crackermill grinding system is shown in figure 2. The crackermill process is currently the most common and productive method of producing CRM. The end product is usually an irregularly shaped particle with a large surface area, varying in size from 4.75 mm (.187 in) to 0.425 mm (0.017 in) (i.e., the No. 4 to No. 10 sieve). These particles are typically referred to as ground CRM. The granulator produces a cubical, more uniformly shaped particle with lower surface area over a range of sizes, usually from 9.5 mm to 2 mm (i.e., 3/8 in to No. 10 sieve), called granulated CRM. Micro-milling, also an ambient and sometimes slurry process, yields finely ground particles ranging in size from 425 to 75 microns (i.e., No. 40 to No. 200 sieve). ^(8,9)

Cryogenic grinding (or separation) is accomplished at extremely low temperatures (-87°C to -198°C [-125°F to -325°F]) by submersing the scrap tire rubber in liquid nitrogen. Below the glass transition temperature (- -620C [- -800F]) the rubber is very brittle and easily fractured in a hammer mill to the desired size. Reportedly, the surface is glasslike, and thus has a much lower surface area than ambiently ground CRM of similar gradation. A schematic of the basic cryogenic grinding process appears in figure 3 ^(8,9)

Specifying CRM may be done in terms of physical and/or chemical properties. The most commonly specified properties include the following: size/gradation, specific gravity, steel and fiber contents, acetone extract, ash, carbon black, rubber hydrocarbon, and natural rubber content. There may also be an upper limit set for moisture content to prevent problems when mixing the CRM at elevated temperatures routinely encountered at a hot-mix plant. A list of CRM producers/blenders used in the three States is shown in appendixes B and C.

Asphalt and CRM Interaction ⁽¹⁰⁾

When CRM is added to asphalt cement at elevated temperatures, the rubber particles tend to swell. The extent and rate of swelling is dependent on a number of factors: chemical and physical properties of the asphalt cement and CRM; mixing conditions such as time, temperature, and degree of agitation; and additives. As the rubber particles swell, the interparticle distance between them is reduced, which results in an increase in viscosity of the CRM-asphalt blend. Preferential absorption of asphalt components by the rubber also contributes to the swelling. Asphalt cement is composed of a variety of petroleum fractions typically classified as asphaltenes, resins, cyclics, and saturates. The quantity of the fractions varies widely with crude source and refining process. Laboratory test data suggest that rubber is more likely to absorb the cyclic fraction than the asphaltene fraction. Therefore, in material selection, compatibility of the asphalt cement and CRM is a key consideration.

Many CRM characteristics affect the interaction with asphalt cement. They include the following: CRM size, gradation, quantity, surface area, and chemical composition; and contaminants such as water, fiber, mineral, and metal. Generally, as rubber content increases, the viscosity of the asphalt-rubber blend increases. Finer size CRM materials tend to "react" more quickly and produce higher viscosities than do CRM with larger particle sizes because of increased surface area. The major CRM compositional effect on asphalt-rubber physical properties is attributed to the total rubber hydrocarbon content and natural rubber content. Asphalt-rubber blends with high total rubber hydrocarbon tend to make the material more ductile because of the high natural rubber content. Preliminary research at Western Research Institute (WRI) suggests that asphalt crude source is the most important variable affecting physical properties of the asphalt-rubber blend. ⁽¹¹⁾

Contaminants may affect processing and/or physical properties of the blend. Excessive moisture causes foaming and may lead to overflow of production or storage vessels. Fiber contaminants tend to increase viscosity, softening point, and resilience, and decrease penetration and ductility. Metal contaminants tend to accelerate wear on pumps and other construction application equipment.

Extender oils may be added to asphalt-rubber blends to soften the asphalt and decrease the low temperature stiffness of the blend. The primary purpose of the addition of an extender oil is to minimize the absorption of the lighter asphalt fractions by the CRM. Generally, aromatic or naphthanic oils are specified.

Design Methodologies-DOT's

As mentioned earlier, an interview form that addressed critical issues affecting mix design was sent to State Highway Agency (SHA) personnel prior to the visits. Of particular interest were the following: type of CRM technology used; CRM source, composition, size, gradation and specification(s); and mix design procedures and criteria.

The general state-of-the-use of CRM for hot mix asphalt concrete in these three States may be summarized as follows:



• Currently, all three States use the "wet" process exclusively, and do so with open-, dense-, and/or gap-graded mixes. California terminated use of the "dry" process technology in 1992 because of erratic performance of the product and poor technical support from those marketing the technology.
  
  
• Maximum size of the CRM particles used in these "wet" process applications ranges from 2.36 mm to 0.18 mm (i.e., passing the No. 8 sieve to passing the No. 80 sieve).
  
  
• Ambiently ground CRM is much more widely used than is cryogenically ground CRM. One CRM supplier in California cryogenically separates the rubber from the steel, then grinds the material at ambient temperature.*
  
  
• Specifying the CRM is based primarily on gradation, although the California and Florida specifications do include some chemical properties as well as requirements for a minimum percentage of natural rubber. Rarely, however, do any of the States or CRM suppliers conduct tests to verify the CRM properties, physical or chemical.
  
  
• Only California specifies the use of extender oils.
  
  
• Laboratory preparation of asphalt rubber binders is reasonably consistent from State to State, as all reportedly blend at higher temperatures (~150°C to 175°C [~300°F to 350°F]) for 10 to 45 minutes to ensure adequate reaction of the QRM and asphalt cement. Furthermore, all the States use a minimum value or range of viscosity as a means of quality control. There is not, however, widespread agreement as to the need for agitation (i.e., initial, continuous, or intermittent) during laboratory blending or the length of time beyond which the material is unsuitable for use.
  
  
• In terms of mix design, two general approaches have been taken to select a total binder content for CRM mixes. Both Arizona and Florida, which use the Marshall method, adjust the unmodified mix binder content by formula or some preselected percentage, or select the CRM binder content that corresponds to a higher air void content (e.g., 5 percent air voids). Arizona typically uses 20 percent CRM (by weight of asphalt cement) for both open- and gap-graded mixes. Florida uses CRM mixes for both dense- and open-graded friction courses. The rubber contents are 5 and 12 percent (by weight of asphalt cement) for the dense- and open-graded mixes, respectively.
  
  Depending upon binder type, California uses 14 to 23 percent CRM in dense-, gap and open-graded mixes. Using the Hveem method of mix design, California currently selects total binder content of CRM mixes at 3 to 4 percent air voids, depending on climate and traffic.
  
  
• Application of conventional mix design criteria varies from State to State. For its gap-graded mixes, Arizona has no minimum Marshall stability requirement. If the mix has low stability, engineering judgment is used to make appropriate changes in the job mix formula. Florida has a minimum Marshall stability requirement of 6.67 kN (1,500 lb) for its dense-graded mixes. There are additional criteria for Marshall flow, air void content, and voids in the mineral and effective asphalt content. Although California uses the Hveem method mix design for its conventional mixes, it does not specify a minimum stabilometer value:
  
  Note: Detailed specifications are given in appendix D.
  
  Arizona
  
  

  The Arizona SHA makes exclusive use of the "wet" process for both open- and gap-graded mixes where the maximum size of the CRM is 2.0 mm (passing the No. 10 sieve). For both the open- and gap-graded mixes, 20 percent CRM (by weight of asphalt cement) is blended with an AC-10. CRM is specified by gradation only and no extender oil is specified. The asphalt-rubber blend is evaluated in terms of viscosity, cone penetration, softening point, and resilience (ASTM D-297 test methods). Typical binder contents (by total weight of mix) for the open- and gap-graded mixes are 9 to 10 and 6.5 to 8.5 percent, respectively. Binder content for the gap-graded mixes is selected at 5 percent air voids. For the open-graded mix, binder content is selected based on the following formula:
  
  If Marshall stability of the gap-graded mix is less than 4.45 kN (1,000 lb), the job-mix formula is adjusted by reducing the binder content or adding an admixture such as hydrated lime or Type II portland cement. More detailed information on Arizona SHA's use of CRM technology is shown in table 3 on page 3-6.

  
  
  California
  
  

  Like Arizona, CALTRANS uses the "wet" process exclusively for open-, dense-, and gap-graded mixes. Most of the CRM-HMA applications, however, use the gap-graded mix. Because of poor performance and inadequate technical support, the State no longer uses the "dry" process. The 2.36 mm (passing the No. 8 sieve) CRM that CALTRANS requires is typically supplied by Atlos, Baker, and BAS Recycling, Inc. (see appendix B). Unlike Arizona, CALTRANS specifies the CRM in terms of gradation, production process, specific gravity, and rubber content. CRM content depends on binder type:
  
  Type 1 binder uses 14 to 20 percent CRM by weight of binder; while type 2 binder uses 17 to 23 percent CRM by weight of binder, with 2 to 6 percent extender oil by weight of binder permitted. The asphalt-rubber blend is specified in terms of viscosity. Base asphalt cements used, regardless of mix types, are AR-i 000, AR-2000, and AR-4000. Typical binder contents (by weight of dry aggregate) for the gap-, dense-, and open-graded mixes are 8, 6.5, and 7.2 percent. Mix design for the gap- and dense-graded mixes is generally conducted in accordance with the Hveem procedure with modifications. For gap-graded mixes, the binder content is selected at 3 to 4 percent air voids in hot climates and 3 percent in mountain climates; and there is no requirement for a minimum Hveem stabilometer value. For CRM open-graded mixes, the binder content is 1.25 X binder content of the unmodified mix. More detailed information on CALTRANS use of CRM technology is shown in table 4.

  
  
  Florida
  
  

  State of Florida uses "wet" process CRM technology for both open- and dense-graded friction courses. For dense-graded mixes, the nominal maximum size of the CRM is 0.18 mm (passing the No. 80 sieve); for open-graded friction courses, the nominal maximum size is either 0.425 mm or 0.18 mm (passing the No. 40 or No. 80 sieve). Typical CRM content is 5 percent and 12 percent (by weight of asphalt cement) for the dense- and open-graded mixes, respectively. An AC-30 is used for both open- and dense-graded mixes. Florida requires that final processing of the CRM be ambient grinding. Additionally, the CRM is specified in terms of specific gravity, moisture content, metal contaminants, acetone extract, hydrocarbon content, ash content, carbon black, and natural rubber content. Specifications for the asphalt-rubber blend include blending time, temperature, and minimum viscosity. Typical binder contents (by total weight of mix) are 7.1 and 6.5 percent for the open- and dense-graded mixes, respectively. For dense-graded mixes, a minimum Marshall stability of 6.67 kN (1,500 lb) is required, as is flow in the range of 0.031 cm to 0.055 cm (8 to 14 [0.01 in]). Additional criteria include the following: minimum voids in the mineral aggregate (VMA) of 15.5 percent; air void content of 4 to 6 percent; and minimum effective asphalt content of 5.5 percent. More detailed information on Florida's use of CRM technology is shown in table 5.
  
  A side-by-side summary of the individual SHA approaches to CRM mix design is shown in table 6.
  
  

  Design Methodologies-Local Agencies
  
  

  In addition to discussions with SHA personnel in Arizona, California, and Florida, major contractors and material suppliers in these States were also visited to obtain the industry perspective on CRM technology. Personnel from the larger cities and counties in these three States also provided input that is noteworthy.
  
  Several contractors in Arizona, including International Surfacing, Inc. (ISI), FNF Construction, and Western Technologies, provided tours of their facilities and information with regard to CRM specifications, mix design, and plant operations. The city of Phoenix provided information on its specification for asphalt concrete overlays that require either polymer or rubber modification. Information provided by these groups is summarized in appendix E.
  
  Similarly, visits with blending contractors (e.g., Manhole Adjusting and Granite Construction) also resulted in information on local agency experience in southern California. The specifications developed by American Public Works Association (APWA) for use in southern California are shown in appendix D and summarized in appendix E. Type A uses whole scrap tire rubber, while types B through D allow the use of natural rubber (often ground-up tennis balls). Type B is the process used by Manhole Adjusting.
  
  

  Summary
  
  

  Based on the interviews with SHA personnel, producers/suppliers, and paving contractors in Arizona, California, and Florida, CRM mix design issues may be divided into two general categories: CRM technology and specifications, and mix design procedures and criteria. Within the former there are specifications for not only the CRM, but also the asphalt-rubber blend, which vary in terms of degree and enforcement.
  
  As noted earlier in this chapter, all three States make exclusive use of the "wet" process technology. This is logical from both historical and field performance perspectives. The wet process technology was developed by Charles H. McDonald, the Sahuaro Petroleum and Asphalt Co., and the Arizona Refining Co. In California, exclusive use of the wet process is largely the result of erratic performance and poor technical support of the dry process technology. The CRM used in all three States is primarily ambiently ground, and the maximum size varies from 2.36 mm to 0.18 mm (No. 8 sieve and No. 80 sieve). All States specify, at a minimum, the CRM gradation. California and Florida specify some chemical and physical properties for the CRM, as well. Interestingly, however, neither State currently verifies the CRM properties with any testing in-house. CRM content varies from a minimum of 5 percent in Florida to a maximum of 23 percent in California, depending on mix type.** Also, CRM has been used success-fully with dense-, gap-, and open-graded mixes. Laboratory preparation and specification of the asphalt-rubber blend are a bit more uniform as all States mix at somewhat higher temperatures (El 50°C to 175°C kz300°F to 350°F]) and use viscosity as a measure of quality control. Only California requires the use of extender oils.
  
  In terms of mix design, there are similarities as well. For both CRM dense- and gap-graded mixes, whether Marshall or Hveem design, binder content is selected at a particular void content. For CRM open-graded mixes, binder content, including CRM, is determined by formula-usually by increasing the unmodified binder content by some predetermined factor. Application of mix design criteria is somewhat less consistent from State to State. Only Florida has minimum Marshall stability requirements. California, which uses the Hveem method, does not specify a minimum stabilometer value.
  
  Ironically, some cities, counties, and producers/suppliers have far more comprehensive specifications for both the raw material, for example ORM, and the asphalt-rubber blend than do some of the State highway agencies. Enforcement of these specifications, however, varies dramatically.
  
  * It should be noted that early potential applications did not permit the use of cryogenically ground CRM. Back to report
  
  ** Note that CRM content is expressed as percent by weight of asphalt cement.
  
  

[home library membership newsletter FAQ links contact events]

---

copyright © 1998 Rubber Pavements Association