The same federal-sponsorship report states two different recycling rates for the same material, four pages apart, and the difference between them is the whole subject. The abstract of the 14th annual industry survey states that more than 99 percent of reclaimed asphalt pavement is put back to use. The executive summary states that more than 89 percent of asphalt mixture reclaimed from old asphalt pavements goes back into new asphalt pavements, with the remaining 11 percent used in other engineering applications such as unbound aggregate bases.
Neither figure is wrong. They count different things. The higher number measures whether the milled material avoided a landfill. The lower number measures whether it returned to the layer it came from, carrying its aged binder back into a new mixture where that binder does structural work. Only the second one describes recycling in the sense engineers mean, and the eleven-point gap between them is where the technical constraints live.
From milling head to stockpile to weigh hopper
Reclaimed asphalt pavement begins as the output of a milling machine working an existing surface. EPA’s AP-42 emissions documentation describes the reclamation sequence plainly: old pavement is removed from the road base, transported to the plant, and crushed and screened to the appropriate size for further processing, after which it is heated and mixed with new aggregate and the proper amount of new asphalt cement to produce a mixture meeting the required quality specifications.
FHWA’s own recycling program page fills in the plant floor that summary skips. It shows a dedicated RAP bin at a mobile plant and a screen positioned specifically to prevent large RAP particles from entering the mixture. AP-42 adds that at newer facilities the reclaimed material also may be transferred to a separate heated storage bin. By the time of that document’s 2004 revision, virtually all new plants already carried RAP processing capability as standard.
Fractionation, in which the reclaimed material is screened into separate coarse and fine streams instead of used as a single blend, accounts for about 21 percent of RAP use nationwide according to the 2023 survey. That figure is a reasonable proxy for how much of the industry has moved past treating RAP as a bulk commodity and started treating it as two aggregates with different binder contents.
What the national numbers show, and what the survey can carry
For the 2023 construction season the survey estimates 96.1 million tons of reclaimed asphalt pavement used in asphalt mixtures, a 71.6 percent increase over 2009, against total asphalt mixture tonnage that grew only 22.7 percent across the same period, so RAP is growing three times faster than the market it feeds. About 171 million tons remained stockpiled nationwide at the end of that season. The estimated displacement was 4.8 million tons of asphalt binder, roughly 26.4 million barrels, and more than 91 million tons of aggregate, valued in total at about $4.5 billion.
The figure that deserves the most attention is the smallest one. The average percentage of RAP in asphalt mixtures rose from 15.6 percent in 2009 to 21.9 percent in 2023. Fourteen years of sustained federal encouragement, industry promotion and rising binder prices moved the national average by about six percentage points.
The survey’s own accounting of itself is worth carrying alongside its numbers. It is voluntary. For 2023 it drew responses from 201 companies representing 1,079 production plants across 49 states and the District of Columbia, and the response base moved: 17 percent fewer companies and 21 percent fewer plants than the prior season, with 18 percent of responding companies not having participated the year before. Roughly 5 percent of respondents were not association members. National totals are scaled from that base and cross-checked against state asphalt pavement association tonnage reports, which is a defensible method and not a census. Year-over- year changes of a point or two should be read as such. For the 2024 season the association has announced more than 101 million tons of RAP used and warm-mix technologies at about 40 percent of production, with the full report following the same annual pattern.
Specifications cap RAP because its binder has already been aged twice
The reason a mixture cannot simply be made of old pavement is chemical, and the standard laboratory sequence for grading asphalt binder explains it precisely. The rolling thin-film oven test, adopted for asphalt work out of the Strategic Highway Research Program in 1992, conditions binder for 85 minutes and simulates the condition of the asphalt binder immediately after the pavement is constructed. The pressurized aging vessel then puts the same residue through a 21-hour procedure that includes 20 hours of conditioning at elevated temperature and pressure, which the federal course manual describes as simulating 5 to 10 years of in-service aging.
Binder recovered from a road milled at year 25 has been through both of those processes and then another decade or two of the real thing. It is stiffer than anything a supplier sells. Add enough of it and the composite binder in the new mixture moves toward brittleness, which shows up as low-temperature cracking and fatigue cracking rather than as rutting. How much can be tolerated is the question every specification answers.
The tiered answer in AASHTO M 323, as summarized in FHWA’s state of the practice review, is straightforward at the low end and stops being straightforward quickly. Below 15 percent RAP, no change in binder selection is required. Between 15 and 25 percent, the virgin binder is selected one grade softer, so a project that would ordinarily specify PG 64-22 takes PG 58-28 instead. Above 25 percent, the designer is directed to blending charts, which requires recovering and testing the RAP binder itself rather than applying a rule. That is the real cliff in the specification: not a prohibition, but the point at which the mix designer takes on laboratory work and technical responsibility that a table used to carry.
Due to the stiffening effect of the aged binder in RAP, the specified binder grade may need to be adjusted.
High-RAP mixtures are demonstrably buildable
The evidence that higher percentages work is not theoretical. FHWA’s recycling program documents field projects placed during the 2007 and 2008 construction seasons across multiple states using 25 to 50 percent RAP content. On the production side, EPA’s plant documentation puts the normal processing ceiling for a counterflow drum mix plant at half the mixture by weight, with no emissions penalty it could observe at that level.
That same document is a useful corrective on a claim often made in the opposite direction. It had been suggested that processing RAP in parallel-flow drum plants raises organic compound emissions, because the reclaimed material meets the hot gas stream in the mixing zone. EPA’s assessment is that the supporting data are very weak, and that although a relationship appears between RAP content and condensible organic particulate emissions, 89 percent of the variation in the data was the result of other unknown process variables. Plant type genuinely matters to how RAP is introduced, which is covered in how asphalt is made. It has not been shown to matter much to what comes out of the stack.
The tools that raise the ceiling, and an awkward finding about them
A producer aiming to carry more reclaimed material typically starts by fractionating the stockpile, then reaches for one of two techniques the survey actually quantifies: dropping to a softer virgin grade, or adding a recycling agent that partially restores the aged binder’s rheology. Mixture tons produced using softer binders are estimated at 16 percent of RAP mixture production, and tons produced using recycling agents at 5 percent.
Then it reports something that ought to trouble anyone who assumes the practice is technically driven. Softer binders and recycling agents were reported nationwide, but there was little correlation between the level of RAP and reclaimed shingle content used and the use of softer binders or recycling agents. The countermeasures are not tracking the exposure. Either the binder grade adjustment is being applied where it is not needed, or it is not being applied where it is, or the reporting categories are too coarse to see the relationship. All three possibilities are worth an agency’s attention.
Warm-mix technologies interact with this directly, since lower production temperatures reduce the additional aging a mixture takes on during manufacture and improve compaction at higher recycled contents. The trade-offs there, including the ones the environmental case tends to skip, are treated in warm-mix asphalt.
Reclaimed shingles show what a stalled recycled stream looks like
Reclaimed asphalt shingles are the useful counterexample, because the same industry, the same plants and the same survey handle them and the trajectory is different. Use rose 18 percent in 2023 to an estimated 797,000 tons, and that increase still left utilization about 59 percent below the 2014 peak. About 1.18 million tons sat stockpiled, down 17 percent from the year before.
The survey records that decline without attributing a cause, and the honest position is that the published national data document the retreat rather than explain it. The reason in any given state sits in that state’s mixture specification history and its asphalt pavement association reporting, not in the national totals. What the trajectory establishes is the general point: a recycled stream can go backwards for reasons unconnected to landfill economics, and the tiered treatment of recycled binder exists because the consequences of getting the blend wrong arrive years after the certificate of completion.
What is actually binding
Specifications permit more reclaimed material than the industry uses. The national average sits under 22 percent while the specification framework provides a documented route past 25 percent, and federally documented field projects have been built and monitored as high as 50 percent. Use of RAP was reported by 99.5 percent of survey respondents, and only two producers reported landfilling any, a combined 62,000 tons amounting to 0.2 percent of the total. The constraint is therefore neither disposal pressure nor a shortage of material.
What binds is variability and the cost of controlling it. A stockpile with unknown gradation and unknown binder content forces conservative design; fractionation, moisture control, recovered binder testing and tighter quality control all cost money at the plant and produce a benefit that lands on the agency’s balance sheet twenty years later. That split between who pays and who gains is a procurement problem rather than a materials problem, and it recurs throughout pavement materials coverage and in the wider question of what happens to construction waste on a highway project. The percentage will move when specifications reward the producer who does the laboratory work instead of merely permitting it.