Bitumen Adhesion and Anti-Stripping: Testing, Treatment and Specification
What stripping is, and why no binder test predicts it
Stripping is the loss of adhesion between the bitumen film and the aggregate surface in the presence of water. The binder does not fail and the aggregate does not fail. The bond between them fails.
Once that bond goes, the mix stops behaving as a structural layer. It is the binder-to-stone adhesion that turns a pile of graded rock into a load-spreading slab; remove it and what remains is graded rock with some grease in it. The road then shows ravelling, potholing, wheelpath deformation or a base course that crumbles in the hand when a core is broken open — damage that is routinely written up as a binder quality problem when it is nothing of the kind.
This is the single most common cause of premature pavement failure that appears nowhere in any binder specification. Penetration, softening point, ductility, flash point, solubility, specific gravity, absolute and kinematic viscosity, the TFOT and RTFOT ageing residues, and even the full Superpave DSR and BBR suite — every one of those is measured on the binder alone, in the absence of aggregate and in the absence of water. Not one of them can tell you whether the binder will stay attached to the stone on your job. A cargo can pass every line of ASTM D946 or IS 73 and still strip off a granite in six months.
What it looks like on the road
- Ravelling — the surface loses fines first, then coarse aggregate, giving a swept, open look. It usually starts in the wheelpath or wherever water stands.
- Potholes that come back — a pothole that reappears in the same place after patching is normally a stripping problem in the layer underneath, not a patching problem.
- Rutting without shear — rutting with no shoulder heave and no binder flushing at the edges of the rut. The layer is losing volume because it is disintegrating, not deforming.
- Damage that starts at the bottom — the classic moisture-damaged core shows a sound top and a friable base where water has been sitting. Cracking that appears to be bottom-up fatigue is often stripping.
- Dull brown aggregate on a broken core — the fastest field diagnosis there is. Break a core face by hand: a healthy mix fractures through the stone and shows black-coated faces, while a stripped mix separates at the interface and shows clean, dull, uncoated aggregate.
Why water wins at the interface
Bitumen is largely non-polar. A minority of polar functional groups — carboxylic acids, sulfoxides, phenolics and nitrogen bases — do almost all of the sticking, adsorbing onto charged sites on the mineral surface. Water is small, strongly polar, and present in vastly greater molar quantity than those functional groups. Given access and time, it competes for the same adsorption sites and it wins. Adsorption studies have consistently shown the awkward part of this: the carboxylic acid functions that adsorb most strongly onto aggregate are also the ones most easily displaced by water, while the weaker-adsorbing nitrogen bases give a more water-resistant bond. That asymmetry is the chemical reason amine-based additives work at all.
Why the aggregate decides
Because adhesion happens on the mineral surface, and mineral surfaces are not alike. Silica-rich rock carries surface silanol groups that ionise negatively at normal pore-water pH. The binder functions doing most of the bonding are acidic and also carry negative character. Like charges do not bond, and there is no ion available to bridge them, so the joint is weak from the moment it is made and water only has to find it.
Carbonate rock is the opposite case. Calcite and dolomite present calcium and magnesium at the surface, and those ions react with the binder’s carboxylic acids to form calcium and magnesium soaps that are effectively insoluble in water. The bond is not merely a physical wetting; it is a chemical one that water cannot easily reverse. That single difference in surface chemistry explains most of the field record, and it is why the same binder can perform faultlessly against limestone and fail against quartzite two valleys away.
Six routes by which water breaks the bond
Moisture damage is not one process. Six mechanisms are recognised in the pavement literature, and they normally act together on the same pavement — which is part of why no single laboratory test captures the whole story.
Detachment
A microscopically thin layer of water works between binder and aggregate with no visible break in the coating. The film peels off the stone intact, which is why stripped aggregate on a core often looks clean rather than damaged.
Displacement
Water reaches the mineral surface through a break in the film — a pinhole, a sharp aggregate edge, an incompletely coated face — then spreads along the interface outward from that point.
Pore pressure
Water trapped in a partially saturated mix is pressurised by every passing wheel. Repeated pressure cycles scour the binder film and drive water deeper into the interface.
Hydraulic scour
At the surface, a tyre presses water into the voids ahead of the contact patch and sucks it back out behind. The alternating action erodes the film mechanically, independently of chemistry.
Spontaneous emulsification
Water and binder form an emulsion in place, a process accelerated by mineral fines and by some surfactants. The film softens and loses cohesion with no single point of attack.
Film rupture and desorption
Thin films over sharp asperities rupture under strain, and polar binder molecules already adsorbed on the mineral surface are progressively replaced by water molecules competing for the same sites.
Aggregate mineralogy and stripping propensity
The practical consequence of the surface chemistry described above is that rock type, more than any other factor, sets the level of risk on a job. This table is the anticipation tool: it tells you what test programme and what treatment budget to plan for before a single specimen is made.
| Aggregate / rock type | Dominant chemistry | Behaviour of the surface in water | Typical propensity | Practical note |
|---|---|---|---|---|
| Limestone, dolomite | Calcareous — calcium and magnesium carbonate | Basic; supplies cations that form water-resistant soaps with binder acids | Low | The reference good aggregate. Porous or clay-bearing limestones can still strip |
| Basalt, dolerite, gabbro | Basic to intermediate igneous, low free silica | Basic; generally good affinity for binder | Low to moderate | Weathered basalts carry secondary clay minerals — check the quality of the fines, not just the rock name |
| Granite, granodiorite | Siliceous — quartz and feldspar rich | Acidic; silanol sites ionise negative and repel the binder’s acid functions | High | The classic problem aggregate. Assume treatment and full testing will be required |
| Quartzite, chert, flint | Close to pure silica, often glassy | Strongly acidic and hydrophilic, with smooth fracture faces | High to very high | Poor mechanical key as well as poor chemistry — two failure paths at once |
| Rhyolite, felsite | Siliceous volcanic | Acidic | High | Behaves broadly like granite; treat it the same way |
| Sandstone, siliceous cement | Silica grains in a silica cement | Acidic and frequently porous | High | Absorption also changes effective binder content, which compounds the problem |
| Sandstone, calcareous cement | Silica grains in a carbonate cement | Mixed, depending on which phase is exposed | Moderate | Behaviour depends on how much carbonate is present on fresh fracture faces |
| River gravel, uncrushed | Variable, often siliceous with a weathered rind | Variable and frequently poor | Moderate to high | Crushing to expose fresh faces improves both the chemistry and the interlock |
| Air-cooled slag | Calcium silicate, high lime content | Basic | Low | Good adhesion, but steel slags in particular need a separate volumetric stability and expansion check |
The standard tests for moisture damage and adhesion
Adhesion tests fall into two families: those run on loose coated aggregate, which are fast, cheap and visual; and those run on compacted specimens, which are slower, quantitative and are what specifications actually accept.
| Test | Standard | Sample form | Water conditioning | Result reported |
|---|---|---|---|---|
| Moisture-induced damage (modified Lottman) | AASHTO T283 | Compacted specimens at 7.0 percent air voids, two subsets | Vacuum saturation to 70–80 percent, optional freeze at −18 °C for at least 16 h, then 60 °C water for 24 h | Tensile strength ratio (TSR): conditioned indirect tensile strength divided by dry |
| Effect of moisture on asphalt concrete | ASTM D4867 / D4867M | Compacted specimens at about 7 percent air voids | Vacuum saturation to 55–80 percent, optional freeze, then 60 °C water for 24 h | Tensile strength ratio (TSR) |
| Water sensitivity of bituminous specimens | EN 12697-12 | Compacted specimens, two sets of equal void content | Wet set held in water at 40 °C for 68–72 h | ITSR, the indirect tensile strength ratio, in percent |
| Affinity between aggregate and bitumen, rolling bottle | EN 12697-11 Method A | Loose single-size coated chippings in bottles of water | Rolled continuously at ambient temperature, coverage assessed at the intervals the standard sets: 6, 24, 48 and 72 h | Degree of binder coverage in percent, by visual estimate at each interval |
| Boiling water stripping | ASTM D3625 / D3625M | Loose coated mix | Boiled in water for 10 minutes, then decanted and cooled | Retained coating in percent, by visual estimate |
| Coating and stripping, static immersion | AASHTO T182 | Loose coated aggregate | Immersed in water at 25 °C for 16–18 h | Retained coating reported as above or below 95 percent |
| Immersion-compression | ASTM D1075 / AASHTO T165 | Compacted cylindrical specimens, two sets | 49 °C water for 4 days, or 60 °C water for 24 h under the alternative conditioning | Index of retained compressive strength, in percent |
| Hamburg wheel-track | AASHTO T324 | Compacted slab or gyratory specimens | Loaded by a steel wheel while submerged in water at the 50 °C default set in T324, unless the agency specifies otherwise | Rut depth against pass count, and the stripping inflection point |
| Surface free energy / contact angle | No single consensus trade standard | Binder and aggregate characterised separately | Not a soak test; the work of adhesion in the presence of water is computed | An energy ratio or compatibility index, used in research and forensic work |
The four methods you will actually be asked for
Two of these are acceptance tests on compacted mix and two are screening tests on loose mix. Using one in the other’s place is the most common mistake made with adhesion testing.
AASHTO T283 — moisture-induced damage and the tensile strength ratio
T283 is the workhorse. Specimens are compacted from the design mix to 7.0 percent air voids — the void level matters enormously, because a denser specimen simply will not admit water — and divided into two subsets of at least three. The dry subset is left at room temperature. The conditioned subset is vacuum saturated to a degree of saturation of 70 to 80 percent, optionally frozen at −18 °C for a minimum of 16 hours where the agency requires a freeze cycle, then held in a 60 °C water bath for 24 hours, and finally brought to 25 °C for two hours. Both subsets are then loaded in indirect tension at 50 mm per minute at 25 °C. The tensile strength ratio is the average conditioned strength divided by the average dry strength.
Is 80 percent a standard requirement or a specification choice? Both, depending on which document governs your job, and the distinction matters. T283 is a test method: it defines specimen preparation, conditioning and loading, and it sets no acceptance limit whatsoever. The familiar 0.80 comes from AASHTO M 323, the Superpave volumetric mix design specification, which requires a mixture to achieve a minimum tensile strength ratio of 0.80 when tested to T283. Where M 323 or an agency specification adopting it applies, 80 percent is a requirement. Everywhere else it is a very common specification choice — and not a universal one. Individual highway agencies write their own limits, and those limits are not all 80: some sit at 70, some at 85, some vary whether the freeze cycle is applied, and several impose a minimum absolute dry indirect tensile strength alongside the ratio.
That last point deserves emphasis. A high ratio calculated from two low strengths is not a good result — a weak mix that is uniformly weak divides into a flattering number. Always read the dry tensile strength alongside the TSR, and be suspicious of a report that gives you the ratio without the two strengths that produced it.
ASTM D3625 — boiling water stripping, as a screen
A loose coated mix is boiled in water for 10 minutes, the water is decanted, the material is cooled, and an operator estimates the percentage of aggregate surface still covered by binder. It needs a hotplate, a beaker and a few minutes, which is precisely its value: it can be run in a plant laboratory or on site, and it will tell you within the hour whether a combination is obviously in trouble.
Two cautions come with it. First, the assessment is a visual estimate, so it is operator-dependent and reproduces poorly between laboratories. It separates clearly acceptable from clearly unacceptable and is unreliable in the middle, where most real decisions live. Second, it runs on loose mix, so it says nothing about compacted behaviour, air voids or pore pressure under traffic. Use it to compare additive candidates quickly, to check an incoming delivery against a previously approved combination, or to confirm that nothing has changed since the mix design. Do not offer it as the acceptance result on a specification that calls for a tensile strength ratio. Note also that ASTM D3625 is a practice, not a test method, and it says so on the cover.
EN 12697-11 — affinity between aggregate and bitumen, including the rolling bottle
EN 12697-11 covers the affinity between aggregate and bitumen and contains more than one procedure. The one usually meant in practice is Method A, the rolling bottle test. A single-size coarse fraction is coated with the binder under test, cured, then placed in bottles with water and rolled continuously on a bottle roller at ambient temperature. At set intervals — typically 6, 24, 48 and 72 hours — a bottle is removed and the operator estimates the percentage of the aggregate surface still covered.
The output is a coverage-against-time curve rather than a single number, and that is genuinely more informative than a one-shot screen. A pair holding 90 percent at 6 hours and falling to 30 percent at 72 hours is telling you something different from a pair that drops to 60 percent immediately and then holds. The standard also provides a static method and a boiling water method for the same purpose. As with the boiling test the assessment is visual, so report the estimates at each interval and the operator, rather than reducing the whole thing to a headline figure.
ASTM D1075 — immersion-compression
Two sets of compacted cylindrical specimens are made from the same mix. One set is tested dry. The other is immersed in water — at 49 °C for four days, or at 60 °C for 24 hours under the alternative conditioning — then returned to 25 °C. Both sets are loaded in unconfined compression at 25 °C, and the result is the index of retained strength, the conditioned compressive strength as a percentage of the dry. AASHTO T165 covers the same conditioning, with AASHTO T167 covering the compressive strength measurement itself.
A minimum index of retained strength of 70 percent appears in a great many agency and project specifications, but it is a specification choice reported here as common practice, not a requirement of D1075 — the method sets no acceptance limit at all. Read the limit off the document that governs your job. Immersion-compression predates T283 and has largely been displaced by it in new specification writing, but it survives in older national standards and in some airfield and industrial paving documents. If your project specification names it, run it; do not substitute a TSR result and expect the engineer to accept the swap without a written approval.
Where AASHTO T182 fits
The static immersion test in AASHTO T182 holds loose coated aggregate in water at 25 °C for 16 to 18 hours and asks a single binary question: is the retained coating above or below 95 percent? It is even coarser than the boiling test, and it is included here mainly because it still appears in legacy specifications. The old ASTM counterpart, D1664, has been withdrawn. If a document you are working to cites D1664, raise it — that specification has not been maintained.
Anti-stripping treatments and typical dosing
There are three chemical treatments in routine use and three non-chemical ones that are often more effective and always cheaper. The dose figures below are typical industry practice, not standard requirements — no standard specifies an anti-stripping dose, for the reason set out under the table.
| Treatment | Typical dose | Dosed into | How it works | Practical constraints |
|---|---|---|---|---|
| Liquid amine anti-stripping agent | Commonly 0.3–0.5 percent by mass of binder; the working range across suppliers and jobs runs roughly 0.2–1.0 percent | The binder, by in-line injection or into the tank with recirculation | Amine head groups adsorb onto the mineral surface and present a bitumen-compatible tail, producing a bond that water displaces less readily | Degrades with prolonged hot storage; corrosive and skin-sensitising to handle; effectiveness is specific to each binder-aggregate pair and must be verified by test |
| Hydrated lime, Ca(OH)2 | Commonly 1.0–1.5 percent by mass of dry aggregate; some agencies specify up to 2 percent | The aggregate: dry onto damp aggregate, as a slurry, or into the pugmill | Calcium ions react with the binder’s carboxylic acids to form insoluble calcium soaps; lime also stiffens the mastic and slows oxidative ageing | Needs a silo, accurate feeding and dust control at the plant; caustic dust hazard; some specifications require a marination period before mixing |
| Portland cement as mineral filler | Typically 1–2 percent by mass of aggregate, within the filler fraction | The aggregate or the filler stream | Supplies calcium, raises pH at the interface and stiffens the mastic | Stiffens the mix and can reduce workability and low-temperature flexibility; alkaline burn and hexavalent chromium hazards in handling |
| Compaction to target density | Not an additive: compact to the density the project specification sets, which normally carries in-place voids below the vulnerable band discussed under the table | The construction process | Removes the water pathway. A mix water cannot enter cannot be stripped from within | Costs nothing in materials, but demands roller pattern control, mix temperature control and honest density testing |
| Aggregate substitution or blending | Job-specific | The mix design | Replacing or blending part of a siliceous aggregate with a calcareous one changes the chemistry of the pair at source | Haul distance and cost normally decide this rather than chemistry; often the right answer where a quarry is close |
| Polymer modification | Job-specific; the PMB grade is set by the project specification | The binder | Higher cohesion and a stiffer, tougher film resist scour and pore-pressure damage, and some systems improve wetting of the aggregate | Improves a mix’s tolerance of moisture but does not repair a chemically incompatible pair, and does not remove the need for a TSR test |
Using the treatments properly, and handling them safely
The treatments work. Most of the failures attributed to them are failures of dosing point, storage or verification rather than failures of chemistry.
Liquid amine additives
These are fatty amidoamines, polyamines and imidazolines formulated as surfactants for hot bitumen. They are dosed as a percentage of binder mass, not mix mass, which is the first thing to get right when converting a supplier recommendation into a plant setting: 0.4 percent of binder in a mix carrying 5 percent binder is 0.02 percent of the mix.
Dose them into the binder as close to the point of use as the plant allows, through an in-line injection system with a calibrated pump, or into a tank that can be recirculated. Two failure modes recur. The first is inadequate mixing: an additive poured into the top of a tank and left to find its own way down does not distribute, and the first tanker load out is not the same product as the last. The second is thermal degradation. Amines lose activity when held at binder storage temperature for extended periods, so pre-dosed binder that has spent weeks hot in a terminal tank may arrive materially less effective than the dosing record implies. Where dosed binder will be stored, agree a maximum storage time and temperature with the additive supplier, and re-run the acceptance test on the stored material rather than assuming the dose survived the wait.
Hydrated lime
Lime does two jobs, which is why many agencies prefer it. It supplies calcium at the aggregate surface, converting the binder’s acid functions to insoluble soaps; and as a very fine filler it stiffens the mastic and slows oxidative ageing of the binder. There are three normal ways to introduce it: as dry lime onto damp aggregate, as a lime slurry onto the aggregate, or directly into the pugmill. Dry lime onto damp aggregate is the common plant method because the surface moisture helps the lime adhere to the stone rather than blowing through to the baghouse. Some specifications require a marination period — the treated aggregate is stockpiled for a defined time before mixing to let the reaction proceed. Where that is specified, it is a real requirement and not a formality; skipping it changes the result.
Lime addition needs plant capability: a silo, a weigh feeder capable of accurate delivery at around 1 percent of aggregate mass, and dust control at every transfer point. A contractor without that equipment will use a liquid additive whether or not the specification prefers lime, so settle the point at tender stage rather than after mobilisation.
Portland cement
Cement is used where lime is unavailable or where the mix design already calls for added filler. It supplies calcium and raises the pH at the interface, and it stiffens the mastic. The stiffening is the catch: cement is a more aggressive stiffener than lime, and overdosing costs workability and low-temperature flexibility. It is also common in cold recycling and emulsion-bound mixes, where it does additional work on setting behaviour that has nothing to do with adhesion. Treat any cement addition as a mix design decision with volumetric consequences, not as a bolt-on cure.
Density is the treatment nobody invoices for
Moisture-damage research — the SHRP-era work on water sensitivity that introduced the idea of a "pessimum" void content, and the agency studies that followed it — reports that compacted mixes are most vulnerable somewhere in the region of 8 to 12 percent air voids, with the mix behaving as effectively impermeable below roughly 7 percent and as free-draining above roughly 15 percent. Treat those as approximate research findings and not as specification limits: no standard defines a vulnerable void band, the boundaries move with gradation and nominal maximum size, and the only void figure binding on your job is the in-place density requirement in the project specification. The mechanism behind the finding is sound and worth understanding — in the middle band the mix is permeable enough for water to get in but too tight for it to drain out, so water sits at the interface and is pressurised by every axle. Since most real pavements are compacted into or near that middle band, chasing the specified density is the highest-return anti-stripping measure available on any job and costs nothing in materials. Void terminology and the volumetric targets themselves are covered on asphalt mix design basics.
The same logic applies to the rest of the water management. Aggregate that arrives at the mixer still carrying moisture is partly wet when the binder reaches it. A pavement edge that ponds, a blocked subsoil drain or a failed shoulder will strip a mix that performs perfectly two hundred metres further on. Freeze-thaw cycling prises the film off mechanically as ice expands at the interface, which is exactly why the freeze cycle exists in AASHTO T283 and why cold-climate agencies keep it in their version of the test.
Safety: this is where people actually get hurt
- Amine additives are corrosive and sensitising. They are typically classified for skin corrosion, serious eye damage and skin sensitisation. Dosing points need chemical splash goggles with a face shield, nitrile or neoprene gauntlets, a chemically resistant apron, and an eyewash and safety shower within reach — not at the other end of the plant.
- Dose through a closed connection. Typical short-term storage for a paving grade sits at 150 to 165 °C in normal practice, and adding additive through an open hatch on a tank at that temperature exposes the operator to the amine and to hot binder fumes at the same time. Use in-line injection or a sealed tank connection. Grade-by-grade storage and pumping windows are set out in the bitumen heating temperature guide.
- Never introduce anything containing water into hot bitumen. Water flashes to steam instantly and drives a violent foam-over out of the tank hatch. This is a mechanism that kills people, not an inconvenience. Confirm the additive is anhydrous, confirm the receiving tank has ullage, and recirculate rather than sparging with air.
- Hydrated lime is caustic. Calcium hydroxide gives a saturated solution at around pH 12.4. Dry dust burns moist skin and eyes chemically and irritates the respiratory tract. Handle it in a sealed silo with pneumatic transfer, extract dust at transfer points, and use goggles and respiratory protection when bags are opened.
- Cement carries an additional hazard. Alongside the same alkaline burn risk, soluble hexavalent chromium in cement is a recognised cause of allergic contact dermatitis; a number of jurisdictions require chromate-reduced cement for this reason.
- Never dose additive into an open drum over a flame heater. Localised overheating against a drum wall carbonises binder, and an open drum being heated is the worst possible place to introduce a reactive liquid.
How to run an adhesion evaluation for a specific job
Five steps. Done in this order they cost very little and settle the question before the plant starts, which is the only time the answer is cheap.
Define the actual pair
Record the binder source and grade, the aggregate source and the specific fractions, the filler, and the additive if one is proposed. Everything that follows is valid only for that combination. A change to any single element creates a new pair and invalidates the previous result.
Screen fast and cheaply first
Run ASTM D3625 boiling water or the EN 12697-11 rolling bottle on the untreated pair. This costs almost nothing, takes hours rather than days, and tells you immediately whether you are dealing with a routine job or a problem that will need a treatment budget and time in the programme.
Run the acceptance test on the real mix
AASHTO T283 or EN 12697-12 on specimens made at the design gradation and binder content, compacted to the air void level the method specifies. This is the number the engineer accepts or rejects, and screening results will not substitute for it.
If it fails, treat and re-test at more than one dose
Test two or three dose levels rather than one. A single passing result tells you the mix passed; a dose-response curve tells you whether you have margin or are sitting on the limit, and margin is what survives normal production variability.
Lock it into the approved mix design and re-test on change
Record binder source, aggregate source and face, additive product name and dose. Re-test when any of them changes — a new quarry bench, a new refinery cargo, a different additive brand at the same nominal dose. These changes happen mid-project far more often than anyone plans for.
Adhesion is a property of the pair, not of the binder
This is the part that has procurement consequences, and it is the part most often got wrong in a purchase contract.
Two cargoes of 60/70 with materially identical Certificates of Analysis, tested against the same granite with the same gradation, can return different tensile strength ratios. Nothing is wrong with either certificate. The polar chemistry that governs adhesion — the acid number, the distribution of nitrogen bases, the sulfoxide content, all of which vary with crude slate and processing — is not measured on a COA, is not controlled by the grade, and is not part of any penetration, viscosity or performance grade specification. The grade controls consistency, not compatibility.
What a Certificate of Analysis does and does not tell you
A good Certificate of Analysis certifies four things: consistency (penetration, softening point, viscosity), purity (solubility, water content, spot test), safety (flash point) and ageing behaviour (TFOT or RTFOT residue). Those are real and they matter, and a COA missing any of them is a finding worth raising before shipment. But the certificate contains no adhesion information at all, and it cannot, because adhesion is not a property that can be measured on binder in isolation. There is no test that could be added to fix this. Adhesion only exists once there is an aggregate.
What this means for a supplier’s claims
It follows that no bitumen supplier can warrant a tensile strength ratio without your aggregate in front of them. An offer that guarantees TSR above 80 percent, with no reference to a specific aggregate source, is guaranteeing something the supplier has not tested and cannot control. Treat that claim exactly as you would treat any other unverifiable assertion in an offer: ask which aggregate it was measured against, by which laboratory, to which method, on what date. If the answer does not exist, the guarantee does not either.
What you can reasonably ask a supplier for
- Source consistency and change notification. A commitment that the refinery source will not change without notice is worth more to your adhesion result than any adhesion claim, because a source change is what silently invalidates an approved mix design.
- Retained samples of the shipped cargo. Sealed samples drawn at loading to ASTM D140 and held by both parties let you verify the mix design against the material actually delivered rather than against a sample sent ahead of the order.
- Additive disclosure where binder is pre-dosed. If the binder is supplied already treated, you need the additive product name, the dose as a percentage of binder mass, the dosing date, and the storage temperature history. Without those you cannot judge whether thermal degradation has reduced the effective dose in transit.
- Cooperation on additive selection. Suppliers see many aggregate types and often know which additive families work on rock like yours. That is useful input. It is not a substitute for testing your pair.
Where the obligation belongs in the contract
Put the adhesion requirement where it can actually be met. The tensile strength ratio requirement belongs in the mix design approval and the construction specification, tested by the party who controls both materials — normally the contractor, on the actual binder and the actual aggregate, before production starts. The binder purchase contract should specify the grade, the test methods, the Certificate of Analysis content, a source-change notification clause and a retained-sample clause. Writing a TSR obligation into a binder supply contract creates a clause nobody can perform and that will not survive a dispute.
The cost asymmetry
A T283 series on one mix costs a small fraction of a single day of paving. Rebuilding a stripped base course costs the layer, the traffic management, the programme and the argument about who pays. Very few decisions in a paving project have that ratio, and the test happens to be one that has to be done before production anyway. Testing the actual pair on the actual job is not caution; it is the only thing that gives you an answer at all.
Frequently asked questions about adhesion and stripping
What is stripping in asphalt and what causes it?
Stripping is the loss of adhesion between the bitumen film and the aggregate surface caused by water. Water displaces the binder from the mineral surface by detachment, displacement through a break in the film, pore pressure under traffic, hydraulic scour, spontaneous emulsification and desorption of adsorbed binder molecules. The result is a mix that loses cohesion and disintegrates: ravelling, recurring potholes and a friable layer at the bottom of a core.
Is the 80 percent TSR minimum a standard requirement or a specification choice?
Both, depending on the governing document, and the distinction is worth stating precisely. AASHTO T283 is a test method and sets no acceptance limit at all. The 0.80 minimum comes from AASHTO M 323, the Superpave volumetric mix design specification, so where M 323 or an agency specification adopting it applies, 80 percent is a requirement. Elsewhere it is a common specification choice: individual agencies set limits at 70, 80 or 85 percent, vary whether the freeze cycle applies, and some also impose a minimum absolute dry indirect tensile strength alongside the ratio.
Which aggregates are most prone to stripping?
Siliceous and acidic aggregates: quartzite, chert, flint, granite, granodiorite, rhyolite, siliceous sandstone and uncrushed river gravel with a weathered rind. Their surfaces ionise negatively and repel the acidic functions in bitumen that do most of the bonding. Calcareous aggregates such as limestone and dolomite perform far better because calcium and magnesium ions react with the binder’s carboxylic acids to form soaps that water cannot easily reverse. Basalt and air-cooled slag are generally good. Mineralogy predicts risk but does not decide the outcome, so a clay-bearing limestone can still strip.
Can a bitumen Certificate of Analysis predict stripping?
No, and no additional test on the binder could make it do so. A COA certifies consistency, purity, safety and ageing behaviour, all measured on the binder alone in the absence of aggregate and water. Adhesion does not exist as a property until there is an aggregate present, so it cannot appear on a certificate for the binder. It has to be measured on the actual binder-aggregate pair, on the actual job.
How much anti-stripping agent is typically added?
Liquid amine agents are commonly dosed at 0.3 to 0.5 percent by mass of binder, with the range across suppliers and jobs running roughly 0.2 to 1.0 percent. Hydrated lime is commonly 1.0 to 1.5 percent by mass of dry aggregate, and some agencies specify up to 2 percent. Portland cement used as filler is typically 1 to 2 percent by aggregate mass. These are typical industry practice, not standard requirements. No standard specifies a dose, because the effective dose depends on the binder, the aggregate and the additive together. Establish it by testing two or three dose levels during mix design.
Hydrated lime or a liquid amine additive, which should I use?
Lime does two jobs, supplying calcium at the aggregate surface and stiffening the mastic while slowing oxidative ageing, and it does not degrade in hot storage. Liquid amines need no plant equipment beyond a dosing pump and are far easier for a contractor without a lime silo, weigh feeder and dust control. The practical constraint is usually the plant rather than the chemistry. Whichever you use, the choice has to be validated on your own pair by the acceptance test in your specification.
Can the ASTM D3625 boiling water test replace AASHTO T283?
No. D3625 is a practice run on loose coated mix and assessed by visual estimate, so it is operator-dependent and reproduces poorly between laboratories, and it says nothing about compacted mix behaviour, air voids or pore pressure under traffic. It is an excellent screen for comparing additives quickly or confirming that an incoming delivery matches a previously approved combination. It is not an acceptance test, and offering a boiling test result against a specification calling for a tensile strength ratio will not be accepted.
Does polymer modified bitumen prevent stripping?
It helps and it does not cure. A polymer modified binder has higher cohesion and a tougher film, which resists hydraulic scour and pore-pressure damage better than an unmodified binder, and some systems wet the aggregate more effectively. But polymer modification does not change the surface chemistry of a siliceous aggregate, so a chemically incompatible pair remains incompatible. PMB mixes are still tested to the same moisture susceptibility requirement, and they still fail it against difficult aggregates without an anti-stripping treatment.
Testing a binder against your aggregate?
Send the binder grade you need, the aggregate source and rock type, and the adhesion test and limit written into your project specification, along with quantity, packing, destination port and Incoterm. The offer will be built against the specification you are actually working to.