U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
|This report is an archived publication and may contain dated technical, contact, and link information|
|Publication Number: FHWA-HRT-11-058 Date: December 2011|
Publication Number: FHWA-HRT-11-058
Date: December 2011
Recycling asphalt mixtures became widely practiced in the United States in the 1970s, spurred by high petroleum prices and limited availability caused by the
Arab oil embargo of 1973. The increased availability of cold milling machines also promoted recycling in some areas of the country. By the late 1970s, technology was developed to allow recycle ratios as high as 100 percent, although HMA typically contained at most 25–40 percent RAP. Due to fluctuating petroleum prices, recycling is becoming even more attractive. There is interest in using higher RAP contents and in using RAP in more mixtures.
The presence of RAP binder in a mixture is generally acknowledged to increase the mix stiffness at all temperatures and frequencies of loading, although the amount of stiffening may be negligible at low RAP contents. For mixtures that contain high amounts of RAP, there may be substantial increases in stiffness. At high temperatures, an increase in stiffness is considered advantageous because it helps resist permanent deformation. At low and intermediate temperatures, an increase in stiffness may reduce resistance to cracking.
The current American Association of State Highway and Transportation Officials (AASHTO) guidelines require a change in the binder grade when more than 15 percent RAP is added to a mix.(4) They were developed through National Cooperative Highway Research Program (NCHRP) project 9-12, Incorporation of Reclaimed Asphalt Pavement in the Superpave System, which was concluded in 2000 by the North Central Superpave Center (NCSC) and the Asphalt Institute.(5) The data from that study showed that when 15–20 percent RAP was added to a mix, the stiffening effect of the oxidized RAP binder started to become significant. Dropping the virgin binder grade by one increment on both the high- and low-temperature grades counteracted this stiffening effect, resulting in a mixture that behaved similarly to one without RAP and with the design binder grade. If more than 25–30 percent RAP was added, the effect of the RAP binder became even more significant. At that level, testing the RAP binder was recommended to evaluate what virgin binder should be used or how much RAP could be added with a given virgin binder grade. As more RAP is added to a mix, the amount of RAP binder also increases, potentially stiffening the mix even more.
The extent of blending of the virgin and RAP binders is expected to have an effect on stiffness. If the new and old asphalt binders are homogenized, the mixture may have increased stiffness and may crack at low temperatures. If little blending occurs, the mixture may behave as if it contains only the virgin binder. As a result, low-temperature cracking might not occur, but the lack of stiffening by the RAP binder may contribute to high-temperature rutting. NCHRP project 9-12 found that a significant amount of blending occurred between the hardened RAP binder and the added virgin binder in the materials studied. Total blending of the RAP and virgin binders is highly unlikely, but the NCHRP project 9-12 research showed that enough blending occurred that the mixture testing results were not statistically different from total blending.(5) If little or no blending occurs for any reason, the mix properties will be strongly influenced by the virgin binder grade.
Current specifications in Indiana and many other States conform to the requirements of AASHTO M 323, Standard Specification for Superpave Volumetric Mix Design, and AASHTO R 35, Standard Practice for Superpave Volumetric Mix Design for Hot Mix Asphalt.(4,6) Both standards were revised to incorporate RAP mixes based on the results of NCHRP project 9-12 and input from various State highway agencies as well as the Federal Highway Administration (FHWA) Asphalt Mixture Expert Task Group (ETG). Those specifications include the following three tiers of RAP content based on the mass of RAP in the total mix:
The final NCHRP 9-12 report includes recommendations for different tiers depending on the low-temperature grade of the RAP binder. If the binder was very stiff (PGXX-10 or higher), lower amounts of RAP could be used before changing the binder grade (10 and 15 percent). Conversely, if the RAP binder was softer (PGXX-22 or lower), higher RAP contents could be used (20 and 30 percent).(5) The FHWA Asphalt Mixture ETG determined that there were not enough data points to include this in the AASHTO specifications. The asphalt binder requirements that were subsequently adopted represent a middle ground based on the results of the laboratory testing. They also agree well with the interim recommendations that had been made previously by the Asphalt Mixture ETG based on extensive experience with Marshall mixes.(7)
A regional pooled fund study in the Midwest studied three more RAP sources with contents up to 50 percent. That study showed that the NCHRP results generally held true for the materials tested.(2) The study also included a comparison of plant-produced mixes to a linear blending chart. In two of the three cases evaluated, linear blending worked well. However, in the third case, the mixture was consistently stiffer than expected based on linear blending, which possibly showed the effects of plant production variables.
Anecdotal evidence to date suggests that the AASHTO M 323 tiers generally work well in many cases, but there is also some evidence that these break points may not be appropriate in all cases. The actual amount of blending that occurs in a mixture depends on many factors, including the stiffness of the RAP binder, the compatibility of the virgin and RAP binders, and specifics of the hot mix production, including plant type (batch or drum), type and amount of mixing (pugmill or drum), mixing temperature, mix handling (live bottom trucks or dump trucks, shuttle buggies, windrow and pickup, as well as dumping straight into the paver hopper), etc.(2) Laboratory-produced mixtures may not reflect the effects of all of these factors, so testing plant-produced mixtures, which is not typically done in routine practice, would be more realistic.
As contractors use more RAP, two issues have become increasingly important. The effects of RAP on the low-temperature grade are a concern, particularly to agencies in more severe Northern climates that may have to deal with increased cracking later in the service life of the pavement. The increased stiffness of the mix generally provided by the addition of RAP is beneficial at high temperatures but may be detrimental at low temperatures. Conversely, there is some evidence that the addition of oxidized binders may not have as great an effect on low-temperature properties as it does on high-temperature properties.(2) Experiences in Missouri and Minnesota suggest that recycled shingles do not have as great an effect on low-temperature properties as on high-temperature properties. The effects of RAP may be similar since both may contain highly oxidized binders.(8,9) Another issue for contractors is the RAP content at which the binder grade must be changed. For contractors in Indiana, where this work was conducted, the use of greater than 15 percent RAP required the use of a PG58-28 virgin binder. This binder grade is typically more expensive than PG64-22, which is routinely used with 15 percent RAP or less in that market.
The issue of binder grade changes is also a concern to agencies. If RAP does not stiffen the mix to the extent expected, the resulting mixture may be too soft for the intended purpose. Similarly, if there are cases where the plant-produced mix with RAP is stiffer than expected, as seen in the regional pooled fund study, the mix may be more prone to cracking.(2)
Bonaquist et al. suggests that there are some cases where RAP and shingles do not blend with virgin materials to the extent expected.(10,11) Bonaquist et al. used the complex dynamic modulus (|E*|) of plant-produced mix and |E*| of the binder recovered from the mixture to estimate the blending between the virgin and RAP binder. The binder modulus and mixture volumetric properties were input into the Hirsch predictive model to estimate what the mix modulus would be if total blending occurred during production.(12) It is assumed that the recovered binder from the mix represents complete blending between the virgin and RAP binder. If the measured and estimated mix master curves overlap, the blending of recycled and virgin binders is nearly complete. If the curves do not overlap, there is incomplete blending. Bonaquist et al. has examples of incomplete blending, particularly with shingles, but also with RAP.(10) Since Bonaquist et al.’s technique uses test results from plant-produced mixes, the potential variables introduced by the plant also play a role in the amount of blending that occurs.
Because of the lingering questions about the amount of blending that occurs between the RAP and virgin binders, the effects of RAP on mixture properties (especially at low temperatures), and the point(s) at which the virgin binder grade should be adjusted, the research project summarized in this report was undertaken.