Recycled Asphalt Pavement (RAP): Binder, Blending and Design

What the material actually is

Every tonne of RAP delivers two things at once. It delivers aggregate — stone that was crushed, graded and accepted once already, and that still occupies volume in the new mix. And it delivers binder — commonly in the region of four to six percent of the mass of the RAP, coating that stone, and already oxidised. That range is a description of industry practice and not a figure any standard sets, and the binder content of the actual stockpile has to be measured rather than assumed from it. A mix design that treats RAP as aggregate alone will be over-asphalted. A mix design that treats it as binder alone is meaningless. The whole discipline of RAP design is the accounting between those two contributions.

Three terms cause confusion and are worth separating at the outset:

• RAP (reclaimed or recycled asphalt pavement) is the North American term and the one used by AASHTO. RA (reclaimed asphalt) is the European term used in EN 13108-8, the material specification for reclaimed asphalt within the EN 13108 series. They describe the same material.
• RAP percentage is the mass of RAP as a fraction of the mass of the whole mixture. Binder replacement, sometimes written as ABR or RAP binder ratio, is the mass of RAP binder as a fraction of the total binder in the mixture. These are two different numbers and confusing them is the single most common error in RAP work. The arithmetic converting one into the other is set out further down this page.
• RAS (recycled asphalt shingles) is a different material entirely. Roofing shingle binder is air-blown and far harder than road binder, it is present at a much higher content, and it blends far less readily. Nothing on this page about RAP should be applied to RAS.

Route one: cold milling

The dominant production route. A cold planer runs a drum fitted with replaceable carbide-tipped picks across the pavement at a set depth, and the material is thrown onto a conveyor and into a truck in one pass. It is fast, it produces material that is already broken down to something close to a usable size, and it leaves a textured surface ready for overlay.

What it does to the material is the part that matters technically. The drum does not simply detach the mix; it fractures it. Milled RAP is consistently finer than the mixture it came from, with more material passing the 4.75 mm and 0.075 mm sieves than the original design gradation contained, because the picks crush aggregate as well as separating it. The degree of breakdown depends on the tool spacing on the drum, the drum rotation speed and the forward speed of the machine: a machine run fast in a low gear to produce a fine texture produces finer RAP. That extra fine material carries a disproportionate share of the binder, which is why gradation and binder content of milled RAP have to be measured rather than inherited from the original job records.

The second thing milling does is mix layers. A single full-depth pass through a surface course, a binder course and part of a base course produces one stream containing three different design gradations, three different binder contents and three different ages of binder. Selective milling — taking each layer off in a separate pass and stockpiling it separately — requires an extra machine pass and produces RAP that is very much easier to design with. Where a job is going to run high RAP contents, that decision is made at the milling stage, not at the plant.

Route two: breaking out and crushing slab

The alternative is to break the pavement out in slabs with a hydraulic breaker or an excavator bucket, haul it to a yard, and reduce it in a crushing and screening plant. A jaw or impact crusher takes the slab down, and screens split the product into fractions.

The advantages are real. The crusher setting gives direct control over top size, screening gives direct control over gradation, and the operator can crush to the size the mix actually needs rather than accepting whatever a milling drum produced. Aggregate degradation can be lower than in milling if the crushing is done in stages rather than in a single aggressive reduction.

The disadvantages are equally real. Breaking out and hauling slab is a separate operation with its own plant and yard requirement. Slabs taken full-depth carry every bound layer together, so unless the layers are separated during removal the same layer-mixing problem appears here as in full-depth milling. And slab removal brings the material closest to the ground, which is where the contamination risk lives: unbound base, subgrade soil, geotextile and geogrid fragments all travel with a slab if the excavator is not disciplined about where it stops.

Route three: plant returns and mix waste

Out-of-specification production, load rejects, silo and drum cleanouts, and laboratory samples all come back to the yard as asphalt that has never been laid. This material has one strong advantage — it has been through short-term ageing during mixing but not through years of field oxidation, so its binder is materially softer than milled RAP binder — and one strong disadvantage: the quantity is small, intermittent and completely unrepresentative of the milled stockpile it is usually tipped onto. Keep it separate or accept that it is degrading the consistency of whatever it is added to.

Why the production route matters to the binder decision

None of this is aggregate housekeeping. Every one of the differences above lands on the binder question:

• The finer the RAP, the more binder it carries per tonne. Binder content tracks surface area, so a finely milled RAP contributes more aged binder at the same RAP percentage than a coarsely crushed one, and therefore pushes the blend harder.
• The older the layer, the harder its binder. A surface course that has been exposed to sun and air for twenty years yields far harder binder than a base course of the same age that has been sealed underneath it. Mixing the two produces an average nobody measured.
• Layer mixing destroys the ability to predict. A stockpile containing three courses has a binder hardness that varies with which part of the pile the loader is in. That variability, not the average, is what makes a high-RAP design impossible to hold.

The mixture-level consequences of all this — volumetrics, gradation and binder demand — are covered on asphalt mix design basics. This page follows the binder.

One stockpile, one material

The governing principle is that a stockpile has to stop changing before it can be characterised. A pile that receives new material every day from a different job has no properties — it has a history. The workable practice is to build a pile to a defined quantity, close it, sample and test it, and then use it as a designed input. Anything else means the design is describing a material that no longer exists by the time the plant draws on it.

Physical stockpile discipline follows from that:

• Separate by source and by layer. Surface course RAP, binder and base course RAP, and plant returns are three different materials. So is RAP from a different road with a different aggregate.
• Build low and wide, not conical. A conical pile segregates: coarse particles roll to the toe, fines stay near the apex. Since binder content tracks fineness, a segregated pile has a binder content that varies with where the loader digs.
• Stockpile on a paved, drained pad. RAP built on ground picks up soil at the base of the pile, and every bucket taken from the bottom carries some of it.
• Keep piles low enough that they do not consolidate. RAP re-agglomerates under its own weight, and warm weather accelerates it. A tall pile that has stood through a hot season can develop a solid core that has to be re-crushed before it will feed.
• Shed water. Crown or cover the pile. Water is the single largest constraint on how much RAP a given plant can run, for the reason set out below.

Fractionation is what makes high RAP contents possible

Fractionating means screening the processed RAP into two or more size fractions — commonly a coarse and a fine split at around 4.75 mm or 9.5 mm, sometimes three fractions — and stockpiling and metering each one separately.

The reason is binder content. Binder coats surface area, and fine material has far more surface area per unit mass than coarse material. A fine RAP fraction therefore carries substantially more binder per tonne than a coarse fraction taken from the same pavement. In an unfractionated pile those two populations are mixed, and any segregation in the pile turns directly into a swing in the binder content of the material actually being fed. Fractionating separates them, so each stream has a much tighter binder content and gradation, and the design can then use the two in controlled proportions to hit both the gradation target and the binder replacement target independently.

Fractionation is normal practice at plants running meaningful RAP contents and is required by some agency specifications above a threshold RAP percentage. It is not a requirement of any product standard, and no standard prescribes the screen size at which the split should be made. Treat the specific numbers as plant practice; treat the principle as settled.

Moisture, and why it caps the RAP percentage

This is the constraint that most often decides how much RAP a plant can actually run, and it is a heat balance question rather than a materials question.

RAP cannot be put through the burner flame. Direct flame contact burns the aged binder, produces heavy blue smoke and hydrocarbon emissions, and destroys the very binder the design is counting on. So RAP is introduced downstream of the flame: through a RAP collar part-way along a parallel-flow drum mixer, into the mixing zone of a counterflow or double-barrel drum, or into the pugmill or weigh hopper of a batch plant. It is therefore heated indirectly, by heat transferred from virgin aggregate that has been deliberately superheated above the target mixing temperature so that the combined material lands on target after the cold, wet RAP is added.

The arithmetic is unforgiving. Using the physical constants normally applied to this calculation — a specific heat of about 0.9 kJ per kilogram per kelvin for mineral aggregate, which is a typical value that varies with mineralogy rather than a standard constant, 4.18 kJ per kilogram per kelvin for water, and a latent heat of vaporisation of about 2260 kJ per kilogram at 100 °C — heating one kilogram of dry aggregate from 20 °C to 160 °C requires roughly 0.9 × 140 = 126 kJ. Removing one percentage point of moisture from that same kilogram requires heating 0.01 kg of water from 20 °C to 100 °C (0.01 × 4.18 × 80 = 3.3 kJ) and then vaporising it (0.01 × 2260 = 22.6 kJ), a total of about 26 kJ. In other words, each single percentage point of stockpile moisture adds roughly a fifth again to the burner duty for that material, and a RAP arriving at five percent moisture roughly doubles it.

Three consequences follow directly. First, wet RAP does not merely consume fuel; it caps the RAP percentage, because there is a limit to how hot the virgin aggregate can be superheated before the plant, the baghouse or the aggregate itself is damaged. Second, covering or crowning a RAP stockpile is one of the most effective single measures available in the yard. Third, the practical RAP ceiling is a property of the specific plant configuration — a batch plant without a RAP preheater is far more constrained than a counterflow drum designed for it — so any statement that a plant "can run 40 percent RAP" is a statement about that plant, not about the material.

Variability is what disqualifies a RAP, not the mean

A stockpile with a five percent binder content is designable. A stockpile whose binder content averages five percent while individual samples run from 3.8 to 6.2 percent is not, because the virgin binder addition is calculated as the difference between the total binder demand and the RAP contribution, and an uncertainty of plus or minus a percentage point in the RAP contribution is enormous when total binder is around five percent of the mix.

So characterisation must report scatter, not just an average. Take a set of samples spread across the pile rather than from the working face, sample to AASHTO T 2 and reduce to test size to AASHTO R 47, and run binder content and gradation on each sample separately. Report the mean and the standard deviation. Agencies that permit high RAP contents generally impose both a sampling frequency and a variability criterion, and those criteria come from the agency specification — no product standard sets them.

The commercial reading of this is straightforward: variability in the stockpile is paid for in virgin binder. A plant that cannot predict its RAP contribution has to design conservatively, which means either accepting a dry mix in the bad batches or adding virgin binder to cover the worst case in every batch.

Contamination, and one item that is a legal question

Foreign matter in RAP does not blend, does not bond and is not aggregate. The routine offenders are unbound base and subgrade soil from the bottom of the excavation, concrete from adjacent structures, joint and crack sealant, thermoplastic road marking, chip seal, geotextile and geogrid fragments, and general site rubbish including wood and plastics. Joint sealant is the persistent problem because it does not disperse: it agglomerates, smears through the plant and shows up as discrete soft inclusions in the mat.

EN 12697-42 gives the method for determining the amount of foreign matter in reclaimed asphalt, and EN 13108-8 sets the declared categories against which a reclaimed asphalt is described. The structure of that classification is worth knowing even outside Europe: heavy foreign matter such as concrete, brick, stone and metal is treated separately from light material such as wood and plastics, and the permitted proportion of light material is very much tighter, because a small mass of floating plastic represents a large volume of defect in a finished mat.

One contamination question is not technical at all. Old pavements can be bound with coal tar rather than bitumen. Coal tar contains polycyclic aromatic hydrocarbons at concentrations that make the material a regulated hazardous waste in many jurisdictions, and it must not be recycled hot. Where the age and history of a road make tar plausible, screening for polycyclic aromatic hydrocarbons is a health and legal step that comes before any question of mix design. Field spray screening kits are widely used as a first pass; the confirmatory method and the action limits are set by the local regulator, not by any asphalt standard.

What ageing does to a binder

Bitumen ages in two regimes. Short-term ageing happens in a few hours during mixing, storage and laydown: light components volatilise and oxidation runs fast because the binder is a thin film at mixing temperature, and the mixing windows for the paving grades are set out on the heating temperature guide. This is the regime simulated by the rolling thin-film oven test, AASHTO T 240 / ASTM D2872, which holds the binder at 163 °C for 85 minutes in a moving film with an air jet. Long-term ageing happens over years in the pavement, as atmospheric oxygen diffuses into the binder film. This is the regime simulated by the pressure ageing vessel, AASHTO R 28, which holds the rolling thin-film residue for 20 hours at 2.10 MPa at 90, 100 or 110 °C depending on the grade.

Chemically, oxygen attacks the reactive sites in the molecule — benzylic carbons and sulfur linkages — producing ketones, other carbonyl compounds and sulfoxides. These oxidation products are strongly polar. They associate with each other, and the effect at the colloidal scale is that material migrates from the maltene fractions (the aromatics and resins that keep the asphaltenes dispersed and the binder mobile) into the asphaltene fraction. The binder becomes stiffer, more elastic, and progressively less able to relax stress.

Two things about this are worth stating plainly because they govern what rejuvenators can and cannot do later on this page. First, it is a chemical change and not a physical one: the carbonyl and sulfoxide groups formed in the road are real functional groups in real molecules, and they do not go away. Second, the field ageing that a twenty-year-old surface course has undergone is substantially more severe than a laboratory pressure ageing vessel cycle. Recovered RAP binder is routinely harder than a laboratory-aged binder of the same original grade.

What that looks like on the tests you already read

• Penetration falls. Binder recovered from an old surface course is commonly reported well down into the low tens of decimillimetres by EN 1426 / ASTM D5 — that is, harder than the hardest paving grade in EN 12591, and in the region of, or beyond, an oxidised industrial grade. That observation comes from published recovered-binder data and from industry practice; no standard sets a penetration value for recovered binder, so measure it.
• Softening point rises. For scale: EN 12591 specifies a ring and ball softening point by EN 1427 of 46 to 54 °C for a 50/70 and 43 to 51 °C for a 70/100. Recovered RAP binder frequently sits above the top of both bands.
• The high-temperature continuous grade rises. Under AASHTO M 320, the high-temperature criterion is a dynamic shear rheometer result to AASHTO T 315: G* divided by sin delta of at least 1.00 kPa on the original binder and at least 2.20 kPa on the rolling thin-film residue, measured at 10 radians per second. Aged RAP binder satisfies that criterion at temperatures far above any grade that would be ordered for the same road.
• The low-temperature grade rises too, which is the damaging direction. The bending beam rheometer, AASHTO T 313, requires a creep stiffness of not more than 300 MPa and an m-value of at least 0.300 at 60 seconds, tested at the low grade temperature plus 10 °C. Ageing raises stiffness and lowers the m-value, so the temperature at which a recovered binder still passes is much warmer than it was when new. The mixture’s tolerance of cold has gone.
• The intermediate-temperature criterion tightens. M 320 limits G* multiplied by sin delta on pressure-aged residue to 5000 kPa, and this fatigue-related criterion is frequently the one that constrains a high-RAP blend.

Delta Tc: the number that describes lost relaxation

Ageing does not raise stiffness and destroy the m-value at the same rate. The m-value — the slope of the creep stiffness curve, which is a measure of how quickly the binder relaxes stress — deteriorates faster. That asymmetry is captured by Delta Tc, defined as the difference between the low critical temperature determined by the 300 MPa stiffness criterion and the low critical temperature determined by the 0.300 m-value criterion, both from the bending beam rheometer. As a binder ages, m-value control takes over and Delta Tc becomes increasingly negative.

It matters because it correlates with non-load-associated cracking — block cracking and surface ravelling on pavements that are not fatigued. Research work published by the Asphalt Institute and the United States Federal Highway Administration proposed a cracking warning at Delta Tc of −2.5 °C and a cracking limit at −5.0 °C, and a number of agencies have since adopted one or both as a specification requirement for recovered or laboratory-aged binder. State the origin honestly when you use them: these are research-derived criteria adopted by some agencies, not requirements of AASHTO M 320, which contains no Delta Tc limit. The reason they appear in a RAP discussion is that a blend can meet every M 320 criterion and still have a poor Delta Tc, and high RAP contents are one of the routes to that outcome.

The consequence in one sentence

Recover the binder from a produced RAP mixture and grade it, and you will get a harder grade than the one on the delivery note. That is not a supply failure and it is not an argument about the Certificate of Analysis — it is the RAP binder doing exactly what the design assumed it would do. It is also the entire reason a specification asks for a softer virgin grade at higher RAP contents: the grade step is there to cancel a shift the designer already knows is coming.

The black rock question

The design cannot proceed until one question is answered: how much of the RAP binder actually joins the binder system of the new mixture, and how much of it stays where it is, acting as a stiff coating on what is effectively just another aggregate? Three positions are possible.

Full blending. All of the RAP binder mobilises at mixing temperature and mixes homogeneously with the virgin binder. The composite binder is a true blend in proportion to the two contributions, and the virgin binder addition is reduced by the full RAP binder contribution. This is the assumption on which the AASHTO M 323 tiers and blending charts are built, and it is the assumption under which almost all production mix designs are actually calculated.

No blending — the black rock case. None of the RAP binder mobilises. Each RAP particle behaves as a piece of aggregate that happens to be coated in stiff black material, contributing volume and surface but nothing to the binder system. Under this assumption the mixture needs its full binder content in virgin binder, and the RAP contributes no binder at all.

Partial blending, which is what actually happens. At the particle scale there is a gradient. The outermost binder on a RAP particle is thinnest, reaches temperature fastest and mixes freely. The binder deepest in the film and in the crevices between agglomerated particles never fully softens within the mixing time available. Between those two there is a diffusion zone in which virgin binder and aged binder interpenetrate partially. How far that diffusion goes depends on mixing temperature, mixing time, how finely the RAP is divided, how soft the virgin binder is, and whether a recycling agent is present to accelerate it.

What the evidence actually supports

The question was tested directly. Work carried out under NCHRP Project 9-12, published as NCHRP Report 452 and the associated guidance on incorporating reclaimed asphalt pavement in the Superpave system, compared mixtures made three ways: with RAP as it is; with a "black rock" control in which the RAP binder had been extracted away and the bare RAP aggregate re-mixed with virgin binder; and with a total-blending control in which the recovered RAP binder was blended with virgin binder before mixing. The finding that shaped current practice is that RAP mixtures behaved much more like the total-blending case than like the black rock case, and that at low RAP contents the three cases were difficult to distinguish at all. That result is what justifies the tiered structure in AASHTO M 323 — no action at low RAP, a fixed grade step at intermediate RAP, measurement at high RAP.

Which assumption is dangerous, and in which direction

This is the part that is usually left out, and it is the part with consequences. Designing on full blending is not the conservative choice. If you assume all of the RAP binder is available, you subtract all of it from the virgin binder demand and therefore add less virgin binder. If blending is in reality only partial, the mixture has less effective binder than the design believed. The failure mode is a dry mixture: high air voids, permeability, poor compaction, ravelling, top-down cracking and short fatigue life. Every one of those is a durability failure that appears years after handover, which is why it is under-diagnosed.

The opposite error is far easier to see and far easier to correct. Design as though nothing blends, add full virgin binder, and the mixture is over-asphalted: it ruts, it flushes and it is tender under the roller. That shows up in the first hot season.

The practical positions taken in response are worth knowing because they explain why agency limits differ so much from one another. Some agencies simply cap RAP percentage. Many now cap binder replacement instead, which is the more honest control because it addresses the binder directly. Some apply an availability factor of less than one to the RAP binder, effectively designing between the two extremes. Some require a rejuvenator above a threshold. Some require mixture performance testing on the produced mixture rather than trusting any blending assumption. All of these are agency policy, and none of them is a product standard requirement.

Two plant-side levers push a real mixture towards full blending, and both belong in the discussion when a high-RAP design is being agreed: finer, fractionated RAP, because thin films on small particles mobilise faster; and a softer virgin binder, because it wets and diffuses into the aged film more readily than a hard one. The second of those is the same lever the specification is already pulling for a different reason, which is a useful coincidence.

First, the two numbers that are not the same

The specification tiers are written in RAP as a percentage of the mixture. The blending calculation works in RAP binder as a percentage of the total binder, which is the binder replacement ratio. Convert between them before doing anything else:

Binder replacement = (RAP percentage in the mixture × RAP binder content) ÷ total binder content

• A mixture with 25 percent RAP, where the RAP has a binder content of 5.2 percent and the total binder content is 5.0 percent: the RAP contributes 0.25 × 5.2 = 1.30 percent binder by mass of mixture, so binder replacement is 1.30 ÷ 5.0 = 26 percent.
• A mixture with 40 percent RAP, RAP binder content 4.8 percent, total binder content 5.3 percent: 0.40 × 4.8 = 1.92 percent, so binder replacement is 1.92 ÷ 5.3 = 36 percent.
• A mixture with 15 percent RAP, RAP binder content 5.5 percent, total binder content 5.0 percent: 0.15 × 5.5 = 0.825 percent, so binder replacement is 16.5 percent.

Note that the second and third examples move in opposite directions: RAP percentage and binder replacement are not proportional to each other, because the ratio between the RAP binder content and the mixture binder content changes the conversion. A rich, fine RAP replaces more binder per tonne than a lean, coarse one. This is exactly why agencies have been migrating from RAP caps to binder replacement caps.

The blending relation

For performance grading, the relation is linear in critical temperature. Writing R for the RAP binder fraction of total binder:

T(blend) = R × T(RAP) + (1 − R) × T(virgin)

Rearranged the two ways it is used in practice:

• Case 1 — the virgin grade is fixed and you want the allowable RAP: R = (T(blend) − T(virgin)) ÷ (T(RAP) − T(virgin)).
• Case 2 — the RAP content is fixed and you want the required virgin grade: T(virgin) = (T(blend) − R × T(RAP)) ÷ (1 − R).

Three practical points about this. Use continuous critical temperatures, not grade labels. A binder graded PG 64-22 has a continuous high critical temperature somewhere between 64 and 70, and using 64 when the true figure is 68.3 throws the answer away. Run the calculation separately at the high, intermediate and low critical temperature — three calculations, three windows, and the design window is the intersection of all three. The chart is the plot of that equation: critical temperature on the vertical axis, RAP binder fraction on the horizontal, with the specification limit drawn as a horizontal line. Where the blend line crosses the limit is the boundary of the allowable window, and it can be read off as easily as it can be solved.

Worked example, performance grade

Take a target blend of PG 64-22. The recovered RAP binder grades out at, say, a high critical temperature of 88 °C and a low critical temperature of −12 °C (rounded here so the arithmetic can be followed; a real calculation uses the unrounded continuous values). The virgin binder available is a PG 58-28, one grade softer than the target at both ends.

• High temperature. R = (64 − 58) ÷ (88 − 58) = 6 ÷ 30 = 0.20. So 20 percent binder replacement is the point at which the blend reaches a high critical temperature of 64. Below that, this virgin binder plus this RAP does not make the high grade; above it, the blend exceeds it.
• Low temperature. R = (−22 − (−28)) ÷ (−12 − (−28)) = 6 ÷ 16 = 0.375. So 37.5 percent binder replacement is the point at which the blend reaches a low critical temperature of −22. Above that, the blend fails the low temperature requirement.

The allowable window is therefore 20 to 37.5 percent binder replacement, and both ends are real constraints pointing in opposite directions. Check the arithmetic at each end: at R = 0.20, high = 0.20 × 88 + 0.80 × 58 = 17.6 + 46.4 = 64.0, and low = 0.20 × (−12) + 0.80 × (−28) = −2.4 − 22.4 = −24.8, comfortably colder than the −22 required. At R = 0.375, high = 33.0 + 36.25 = 69.25 and low = −4.5 − 17.5 = −22.0, exactly on the limit.

Convert the window back into RAP percentage using a RAP binder content of 5.2 percent and a total binder content of 5.0 percent: RAP percentage = R × 5.0 ÷ 5.2, giving 19.2 percent at the lower bound and 36.1 percent at the upper bound. That is the number the plant actually sets on the RAP feeder.

One honest caveat on the lower bound. Relying on RAP to reach the high-temperature grade means the mixture fails the rutting requirement whenever the RAP feed runs light. Unless the RAP proportion is tightly controlled, treat the high-temperature crossing as information rather than as a design target, and pick a virgin grade that does not need RAP to make the grade it is being bought for. The intermediate critical temperature, governed by the AASHTO M 320 limit of 5000 kPa on G* multiplied by sin delta, has to be run as a third calculation and frequently produces the tightest window of the three at high binder replacement.

Worked example, penetration grade

Most export binder is bought as a penetration grade, and the same logic applies with a different relation. Penetration does not blend linearly — the logarithm of penetration does. The relation used throughout European practice, and given for reclaimed asphalt in EN 13108-1 Annex A, is that the logarithm of the blend penetration is the proportion-weighted mean of the logarithms of the component penetrations. Softening point, by contrast, blends approximately linearly.

Take a RAP whose recovered binder has a penetration of 20 dmm by EN 1426, and a target blend in the 50/70 band. Order a 50/70 virgin binder (take it as 60 dmm) at 20 percent binder replacement and the blend is:

log(pen) = 0.20 × log(20) + 0.80 × log(60) = 0.2602 + 1.4225 = 1.6827, giving a blend penetration of 48 dmm — out of the 50/70 band and into 35/50.

Order a 70/100 instead (take it as 85 dmm) at the same 20 percent replacement:

log(pen) = 0.20 × log(20) + 0.80 × log(85) = 0.2602 + 1.5435 = 1.8037, giving 64 dmm — squarely inside 50/70.

That is the whole commercial argument in two lines of arithmetic. To land a 50/70 blend at 20 percent binder replacement, you buy a 70/100. The buyer who insists on ordering the grade written on the road specification gets a mixture a full grade too hard, and it will crack.

How a RAP mix design is actually sequenced

• 1. Characterise the stockpile. Binder content, extracted aggregate gradation, moisture, variability and foreign matter, on a closed pile, with scatter reported alongside the means.
• 2. Fix the target RAP percentage against the agency cap, the plant’s heat balance and the layer the mixture is for.
• 3. Convert to binder replacement using the arithmetic above, and check it against the agency’s binder replacement cap if it has one, because that cap and the RAP cap will not agree.
• 4. Decide the tier. Below the first threshold, no grade change. Between thresholds, one grade softer at both ends. Above the upper threshold, recover the RAP binder and grade it.
• 5. Run the blending calculation at all three critical temperatures where the tier requires it, and take the intersection of the three windows.
• 6. Complete the volumetric design on the combined aggregate, including the RAP aggregate gradation, with the RAP binder counted in the total binder content. The virgin binder addition is the difference.
• 7. Verify on produced mixture. Extract and recover binder from plant-produced material and grade it, and run the mixture performance tests the specification requires. This is the only step that tests the blending assumption rather than restating it.
• 8. Re-verify on change. A new stockpile, a new virgin cargo, a change in the RAP feed proportion or a change in mixing temperature all invalidate the design. RAP designs go out of date faster than virgin designs do.

Extraction and recovery solvents: the most serious hazard in this workflow

Dissolving binder out of RAP is the highest-risk operation described on this page, and it is the one most often waved through, because it happens in a laboratory rather than on a site. The solvents historically used for extraction are chlorinated: trichloroethylene, tetrachloroethylene (perchloroethylene) and dichloromethane (methylene chloride). All three are classified as carcinogenic or suspected carcinogenic in most jurisdictions, and all three act as central nervous system depressants at airborne concentrations well below the level at which the smell becomes alarming, so odour is not a warning. The common substitutes are not benign either: toluene is a reproductive toxicant, and n-propyl bromide carries its own neurotoxicity and reproductive concerns.

What that means for a laboratory running RAP extractions:

• Extraction and recovery are fume cupboard operations, not bench operations. A centrifuge extractor or a rotary evaporator running open on an open bench releases solvent vapour continuously throughout the run rather than in a single puff. Local exhaust ventilation belongs on the apparatus itself, and its face velocity should be tested and recorded rather than assumed.
• The vapour is heavier than air and collects low down — in sinks, floor channels, sample stores, pits and basement laboratories. A room ventilated by a ceiling extract does not clear it, and a solvent store below grade is the classic place where a dangerous concentration accumulates unnoticed.
• Skin contact matters as much as inhalation for the chlorinated solvents. Thin nitrile examination gloves are not adequate protection against dichloromethane or trichloroethylene, which permeate them in minutes. Gloves have to be selected from the manufacturer’s permeation data for the specific solvent in use and changed on a time schedule, not when they look dirty.
• Occupational exposure limits are set by your regulator and differ substantially between jurisdictions. Establish the limit that applies where the laboratory sits and monitor against it. Do not carry a number across from another country’s list, and do not treat any figure quoted on a web page, including this one, as the applicable limit.
• Spent solvent, wash solvent and extraction residue are hazardous waste. They are not drain disposal and they are not general waste. Establish the disposal route before the first sample is run, not after the drums have accumulated.
• Where only a binder content is needed, the solvent can be eliminated entirely. The ignition method of ASTM D6307 / AASHTO T 308 is a genuine substitution control — the strongest kind — and it should be the first option considered. The trade-off is that it destroys the binder, so as the recovery table above sets out, it forecloses grading. That is a decision to take with the design tier in view on day one, and it is a safety decision as much as a technical one. The ignition furnace brings its own hazards: it runs at furnace temperature, its hot surfaces and hot baskets burn, and it vents combustion products that need extraction.

Coal tar and polycyclic aromatic hydrocarbons

The contamination section above treats this as a legal question. It is equally a health question, and the controls come before any testing. Polycyclic aromatic hydrocarbons in coal tar are skin carcinogens and photosensitisers, meaning that exposed skin that is then exposed to sunlight burns severely. Where the age or history of a pavement makes tar plausible:

• Treat the pavement as tar-bound until screening says otherwise. The screening result governs the method of removal, not merely the fate of the material afterwards.
• Do not generate dust or fume from suspect material before the result is in. Dry cutting, dry milling, crushing and any hot processing all mobilise polycyclic aromatic hydrocarbons. Water suppression and wet cutting are the interim control.
• Cover skin and prevent transfer. Contaminated gloves, sleeves and overalls carry the material to the face, to the vehicle cab and home. Separate storage of work clothing and washing facilities on site are the controls that stop that.
• Establish the disposal route before the pavement is broken out. Discovering that material is regulated hazardous waste after it is loose in a yard is a much worse position than discovering it beforehand.

Respirable dust at the milling, crushing and screening face

RAP aggregate is rock, and most road aggregate contains crystalline silica. Milling, crushing, screening, loading and pile handling all generate respirable dust, and RAP fines are the fraction that carries most of the binder as well as most of the dust.

• Water suppression at the source — on the milling drum and its conveyor, and at every crusher and screen transfer point — is the primary control, and it is the one that fails quietly when a nozzle blocks.
• Enclosed cabs with filtered, positive-pressure ventilation for milling machines, loaders and crusher operators, with the filters treated as a maintenance item.
• Respiratory protection selected against the measured concentration and face-fit tested, rather than issued by habit. A dust mask handed out at the gate is not a control.
• Never dry sweep a RAP pad. It re-suspends exactly the fine fraction you are trying not to breathe. Wet methods or industrial vacuum.

Water meeting superheated aggregate, and hot binder generally

The heat balance section above explains why virgin aggregate is deliberately superheated above the target mixing temperature before cold, wet RAP is added. The consequence is that the plant is holding material substantially hotter than the finished mixture, and free water arriving into it is the violent failure mode: one volume of water becomes roughly 1,700 volumes of steam, and the expansion happens beneath a hot mass rather than at its surface.

• Free water must never enter the RAP feed. Ponded water in a pile, a thawing frozen crust and an uncovered pile after heavy rain are the three routes. Drained, crowned and covered pads are a safety control as well as a fuel measure.
• Do not charge visibly wet or ponded material, and treat surging, blowback or pulsing at the RAP collar as a stop condition rather than something to work through.
• Keep people clear of the RAP collar, the drum discharge and the pugmill during charging. Steam eruption ejects hot material, and the ejection path is exactly where an operator stands to look at a feed problem.

Hot binder burns are the routine injury on any asphalt plant, and RAP work does not change their character. Bitumen at mixing temperature adheres to skin and continues to burn after contact. It must not be pulled off. Cool immediately with copious clean water and leave the bitumen in place for removal under medical supervision. Face protection, gauntlets and closed sleeves belong on anyone opening a hatch, taking a sample or working near a discharge.

One further plant hazard is created by a mistake the material section warns against on technical grounds. RAP must never be introduced through the burner flame. Beyond destroying the aged binder the design depends on, direct flame contact on aged binder produces heavy blue smoke and hydrocarbon fume — an emissions breach and an inhalation exposure in the same event.

What this section is and is not

This is a description of the hazards that RAP work specifically creates, written so that they can be recognised. It does not replace the safety data sheet for the actual solvent, recycling agent or binder in use, the risk assessment for your site or laboratory, or the exposure limits, monitoring requirements and waste rules your regulator sets. Those three documents govern; this page does not.

What they are

The term covers several things that are sold under different names: recycling agent, which is the term used by ASTM D4552 / D4552M, the practice for classifying hot-mix recycling agents; rejuvenator; softening agent; and various proprietary designations. ASTM D4552 sorts hot-mix recycling agents into six grades on the basis of viscosity at 60 °C, and also imposes requirements on flash point, on saturates content and on the change in viscosity after the rolling thin-film oven test. ASTM D4887 covers the preparation of viscosity blends for hot recycled bituminous materials, which is the companion procedure for working out how much agent a blend needs.

Chemically, the products in use fall into a few families: high-aromatic petroleum extracts and process oils; naphthenic oils; tall oil and tall oil pitch derivatives from wood pulping; and a growing set of bio-derived esters made from vegetable oils, including used cooking oil derivatives. What they have in common is that they are rich in the maltene components — aromatics in particular — that oxidation has depleted.

What they are for

Oxidation moves material out of the maltene fractions and into the asphaltene fraction, and it is the loss of the dispersing medium as much as the growth of the asphaltenes that stiffens the binder and destroys its ability to relax stress. A recycling agent restores the dispersing medium. It re-solvates the associated asphaltene structures, restores the colloidal balance towards where it was, and in doing so brings the rheology back:

• Viscosity falls and penetration rises. The binder becomes workable again at normal mixing temperatures.
• The high-temperature critical temperature falls, which is usually the intended effect — a high-RAP blend generally has far more high-temperature grade than it needs.
• The low-temperature properties improve. Creep stiffness falls and, more importantly, the m-value recovers, so the low critical temperature moves back down.
• Delta Tc frequently improves, which is the property most directly associated with the block cracking that high-RAP mixtures are criticised for.
• Mixture workability and compactability improve, which has knock-on benefits for in-place density and therefore for durability.

What they cannot restore

This is the part that gets oversold, and it is worth being blunt about.

• They do not reverse oxidation. The carbonyl and sulfoxide groups formed in the road are covalently bonded functional groups. A recycling agent dilutes and re-solvates them; it does not un-oxidise them. A binder that has been restored to a target penetration is not the same material as a virgin binder at that penetration, and it should not be assumed to behave identically over twenty years.
• They cannot fix anything that is not a binder problem. Contamination, a poor extracted aggregate gradation, degraded aggregate, excess dust or a variable stockpile are unaffected by any dose of any agent.
• They are limited by diffusion. The agent has to migrate into the aged binder film on the RAP particle, and that takes time and temperature. Immediately after mixing the agent is concentrated where it was introduced, not distributed through the film. A recovered blend tested straight after mixing can show a softer result than the particle-scale reality, and some agencies require a conditioning or silo storage period for exactly this reason.
• They can be overdosed, and an overdose cannot be taken back. Too much agent takes the high-temperature grade below what the traffic requires, and the mixture ruts, flushes and shoves. Because the effect of the agent is not fully realised at the moment of mixing, the temptation to add more is real and the correction is not available afterwards.
• The long-term re-ageing behaviour is still being established. There is credible evidence that some rejuvenated systems re-age towards their pre-treatment condition faster than a virgin binder ages, and the field record on modern bio-derived agents is still short. This is an open question and should be described as one.

Dosage: there is no standard dose

No standard specifies a rejuvenator dose, for the same reason no standard specifies an anti-stripping dose: the effective dose depends on the specific aged binder, the specific virgin binder and the specific agent acting together. Doses are normally expressed as a percentage by mass of the RAP binder — not of the total binder, and not of the mixture — and that basis has to be stated, because the three numbers differ by large factors and mixing them up is a routine plant error.

The way to establish a dose is by rheological trial: recover binder from the RAP, blend it with the intended virgin binder at the design proportion, add the candidate agent at two or three dose levels, and grade each blend. That gives a dose-response curve rather than a single passing result, and a curve tells you whether the design has margin or is sitting on a limit. Record the agent product name, the dose and the basis in the approved mix design so that none of the three can be changed quietly at the plant.

Rejuvenator or softer virgin grade?

They do the same job by different routes, and the choice is usually practical rather than technical.

• A softer virgin grade needs no extra plant equipment, no dosing pump, no separate storage and no additional approval. It is bought as a graded product with a Certificate of Analysis against a recognised standard, which makes it far easier to verify and to contract for. Its limit is the softest grade the market supplies: once binder replacement is high enough that even the softest available grade cannot bring the blend back, the softer-grade route has run out.
• A recycling agent reaches further, and it targets the aged binder directly rather than diluting the system with more fresh binder. What it requires in return is plant equipment, dosing accuracy, an additional approval and an additional variable in the design.

The error to avoid is using both without redoing the calculation. A design that specifies a virgin grade two steps softer and a rejuvenator has applied two corrections for one problem, and the result is an over-soft binder system that will rut. If an agent is introduced, the blending calculation is rerun with the agent in it and the virgin grade is reconsidered from scratch.

One final distinction, because the same word is used for both: rejuvenators are also sold as surface-applied maintenance treatments, sprayed onto an existing pavement to soften the binder in the top few millimetres and delay ravelling. That is an entirely different application with different products, different rates and different acceptance criteria, and nothing on this page applies to it.

The grade ordered is not the grade the road specification names

This is the headline, and it deserves to be stated flatly because it causes so much avoidable trouble. A plant running meaningful RAP contents will order a softer virgin grade than the grade written into the pavement specification, and that is deliberate. A 50/70 road specification with 20 percent binder replacement is served by a 70/100 virgin binder. At higher replacement, or against a harder recovered RAP binder, it may be served by a 100/150. In performance grading terms, a PG 64-22 pavement specification is served by a PG 58-28 virgin binder in the middle M 323 tier, and by whatever the blending chart returns above it.

What is being specified in the pavement document is the blend. What is being bought in the purchase order is one component of that blend. They are different materials, and the second is deliberately softer than the first.

The document trail has to be internally consistent

The practical failure this produces is a false non-conformity. A quality engineer holding a pavement specification that says 50/70 receives a Certificate of Analysis showing a penetration of 82 dmm, checks it against the wrong table, and raises a claim against a cargo that is exactly what was ordered. Cargoes have been rejected on this basis, and the argument is entirely avoidable.

• The purchase order names the virgin grade and the standard it must satisfy — EN 12591, ASTM D946, IS 73, AASHTO M 320 or whatever governs — and nothing else. It does not name the pavement grade.
• The Certificate of Analysis is read against the ordered grade. Every check — penetration, softening point, flash point, solubility, ageing residue — belongs to the grade on the purchase order.
• The approved mix design carries the link between the two, showing the RAP percentage, the binder replacement ratio, the recovered RAP binder properties and the blending calculation that produced the virgin grade selection. That document is what makes a softer purchase defensible to an auditor.
• Put the RAP obligations where they can be performed. The blend requirement belongs in the mix design approval, tested by the party that controls both the RAP and the virgin binder. It does not belong in the binder supply contract. No supplier can warrant a blended grade against a RAP stockpile the supplier has never seen and does not control.

Source consistency matters more on a RAP job, not less

The virgin grade in a RAP design was selected against two things: a recovered RAP binder with measured properties, and a virgin binder with assumed properties. Move either one and the blend moves. A change of refinery source shifts the base binder’s ageing behaviour and its position within its own grade band, and both feed straight into the blend. On a virgin-only job that shift is absorbed by the grade band. On a high-RAP job it is one of two variables that were carefully balanced against each other.

So the clauses that are worth arguing for on a RAP job are the ordinary ones, applied more strictly: notification before any change of source, sealed retained samples drawn at loading so the design can be checked against what actually arrived, and a Certificate of Analysis that reports the measured figures rather than restating the specification limits. Sampling procedure is covered on the sampling page.

Softer grades have handling consequences

Ordering a softer grade is not a free move on the logistics side, and the differences are worth planning for rather than discovering.

• Lower softening point. EN 12591 puts the ring and ball softening point at 46 to 54 °C for a 50/70 and 43 to 51 °C for a 70/100. In a hot climate that difference is meaningful for drummed material: stacking, sun exposure and yard storage all need more care with the softer grade, and the softening point on the certificate is the figure to check. Packaging options are covered on the packaging page.
• Different heating and pumping windows. A softer grade reaches pumping and mixing viscosity at lower temperatures, so the storage and handling temperatures set for a harder grade are wrong for it, and holding it needlessly hot ages it before it ever reaches the mixer — which on a RAP job is precisely the wrong direction. Grade-by-grade windows are on the heating temperature guide.
• Do not over-soften as a safety margin. The instinct to order one grade softer again "to be safe" produces a blend below the required high-temperature grade, and the mixture ruts in its first hot season. The blending calculation has an upper bound as well as a lower one.

The quantity moves too

RAP reduces the virgin binder required per tonne of mixture, and the arithmetic is simple enough to do on the purchase order. A mixture at 5.0 percent total binder with 30 percent binder replacement needs 3.5 percent virgin binder by mass of mixture. Over 1000 tonnes of mixture that is 35 tonnes of virgin binder instead of 50. Plan the tonnage, the packing and the shipment against the virgin figure, not the total binder figure — a routine and damaging planning error. Conversions between mass, volume and drums are covered on the tonnage and volume conversions page.

What this site supplies, and what it does not

Stated plainly, because it is more useful than pretending otherwise. This site supplies virgin bitumen. Penetration grades, viscosity grades, performance grades, oxidised grades, cutbacks, emulsions and polymer modified binder, shipped in drums, jumbo bags, flexitanks and bulk.

We do not supply RAP, reclaimed asphalt or any recycled aggregate. We do not supply rejuvenators, recycling agents, softening agents, warm mix additives or natural asphalt. We do not supply milling machines, crushers, screens, asphalt plant, dosing equipment or any other plant or equipment. We do not carry out extraction, recovery, binder grading or mix design, and we do not provide inspection, laboratory testing or certification services. Those are the province of your milling contractor, your plant, your laboratory, your design engineer and your appointed inspection agency, and anyone offering them alongside a binder cargo is offering something they are unlikely to be able to perform.

What is genuinely useful from a binder supplier on a RAP job is narrower and more valuable than a bundled service: the virgin grade your blending calculation actually calls for, supplied against a named standard; a Certificate of Analysis carrying the measured penetration, softening point, flash point, solubility and ageing residue rather than a restatement of the specification limits; retained sealed samples from the loaded cargo; and notice before the source changes. Those four things let your laboratory hold the design together. Nothing else a supplier can offer will.