Bitumen PG 64-10: Specification, Tropical Climate Fit and Export Supply

PG 64-10 means the binder has been shown by test to satisfy the high-temperature criteria of AASHTO M320 at an average seven-day maximum pavement design temperature of 64 °C, and the low-temperature criteria at a minimum pavement design temperature of −10 °C. The word to hold onto is shown. A performance grade is the output of a laboratory programme, not a judgement someone forms about a product, and it cannot be arrived at by looking at a penetration figure or by lifting a number off an old refinery data sheet. Two of the four measurements behind the label are made on binder that has been artificially aged twice before it is ever loaded.

Three test temperatures appear on the resulting report, and the grade name discloses one of them. Rheometry at 64 °C covers the rutting criteria, applied first to original binder and then repeated on rolling thin film oven residue. The fatigue criterion is applied to pressure ageing vessel residue at 31 °C. The bending beam work is done on that same PAV residue at 0 °C. Neither of the two unstated figures is a laboratory’s choice — both fall out of a formula — so the conditions printed at the top of any certificate can be checked by anyone with a calculator and no laboratory access whatever. Both derivations are worked through in full further down this page.

−10 is the floor of the ladder, not a point on it

The low-temperature side of M320 is a ladder of six-degree steps, and it has an end: −10, −16, −22, −28, −34, −40, −46. −10 is the warmest rung that exists. There is no PG XX-4 and there is no PG XX-0, so a project in Singapore, Lagos or Doha cannot specify anything less demanding on the winter side even if the climate would justify it. When you buy PG 64-10 you are buying the least the standard is capable of asking about cold behaviour.

That has a consequence worth stating without decoration. The bending beam test for this grade is run at 0 °C, and at 0 °C an ordinary straight-run paving binder is nowhere near 300 MPa of creep stiffness and comfortably above an m-value of 0.300. In practice the low-temperature half of a PG 64-10 specification screens out almost nothing. It is not a mistake in the standard — it is simply the standard running out of range at the warm end. But it means the low grade in the name carries no discriminating information in a tropical tender: every compliant binder looks identical there, including binders that differ enormously in every respect that will actually affect the road.

Where the discrimination has gone instead

Since the winter side cannot separate one offer from another, the separation has to come from somewhere else. For this grade it comes from three places, and they are the backbone of the rest of this page:

• The high-temperature grade itself — whether 64 °C is genuinely the right design temperature for the site, which on a Gulf or equatorial coast is a question that deserves checking rather than assuming.
• Traffic loading — the Superpave grade bump, and under AASHTO M332 the traffic letter and the Jnr limit that goes with it. In markets where ports, border crossings and congested urban arterials generate slow and standing loads, this is the decision.
• Properties M320 does not grade at all — ageing behaviour beyond the PAV, adhesion and moisture resistance in a wet tropical climate, and simple consistency checks that tell you what kind of binder you have been offered.

The boundary of what a PG label covers

Everything a performance grade certifies concerns the binder’s response to temperature. Nothing outside that is in scope. Crude source, wax content and asphaltene fraction are not reported. Adhesion to a particular aggregate is not tested, and neither is moisture damage or stripping — which in a monsoon or equatorial climate is one of the leading reasons a pavement comes back early. A grade is not a mix design and cannot be read as one. Two cargoes both correctly labelled PG 64-10 may share nothing but a rheology envelope, and can differ substantially in how they handle at the plant, how fast they age and how well they stick to wet stone. Against that, the grade does deliver one thing no penetration or viscosity grade offers: a high-temperature capability that has actually been measured on aged binder, at the temperature the pavement will reach.

The two tests at 64 °C

The high-temperature grade is the one figure in the name that is not derived from anything: 64 °C is the dial setting on the rheometer. Two criteria are read there. Original binder must give G*/sin δ of at least 1.00 kPa. Rolling thin film oven residue — binder rolled for 85 minutes at 163 °C in a continuous air stream to AASHTO T240 — must give at least 2.20 kPa at the same 64 °C. The two limits are not one question asked twice at different strictness. The unaged limit asks whether the binder leaves the tank fit for purpose; the residue limit asks whether it is still fit for purpose after the mixing plant has taken its first bite out of it, which is the condition in which the road actually receives it. Since anything reaching 2.20 kPa aged has already passed 1.00 kPa unaged, only the residue line ever decides an outcome. On PG 64-10 that single line carries almost the whole specification, and most of the rest of this page is an argument about what a buyer should do about that.

Deriving the intermediate temperature: 31 °C

The fatigue criterion is placed by formula — average the two grade temperatures, add four. For PG 64-10 the working is short enough to do in the margin of the certificate:

• (H + L) ÷ 2 + 4
• (64 + (−10)) ÷ 2 + 4
• 54 ÷ 2 + 4
• 27 + 4 = 31 °C

At 31 °C the PAV residue must show G*·sin δ of no more than 5000 kPa. Two features of that parameter catch people out. It is a product rather than a quotient, unlike the rutting parameter; and the direction of the limit is reversed, so here a low number is the good one. Behind that inversion is a simple physical picture: a wheel passing over the layer puts work into the binder, and an aged binder that is both rigid and springy banks that work as strain energy rather than losing it as heat. Energy that has been banked has to come out somewhere, and it comes out as a crack.

Where that temperature lands is the whole story at this grade. Because L is as warm as the standard permits, 31 °C sits at the warm end of the intermediate temperatures met in ordinary trade: PG 64-16 is tested at 28 °C and PG 64-22 at 25 °C, on an identical summer requirement. Six degrees is a large difference for a material as thermally sensitive as bitumen, and G*·sin δ at 31 °C is a fraction of what the same binder would return at 25 °C. The 5000 kPa ceiling is therefore cleared with a wide margin by essentially any straight-run binder that has got this far down the certificate. As a screen it is inert.

Two consequences follow, one clerical and one structural. Clerically: a report offered as PG 64-10 that shows the fatigue parameter at 25 °C or 28 °C was graded against a −22 or a −16 low grade, not against yours, and that is a different test rather than a rounding difference. Structurally: fatigue cracking is a genuine failure mode on tropical pavements — thin structures, heavy and frequently overloaded axles, weak and often saturated subgrades — and M320 is not protecting against it here, because the criterion has been placed at a temperature at which the binder is too soft to fail it. Fatigue protection on a PG 64-10 project comes from layer thickness, subgrade preparation and mix design. It does not come from the binder specification, and no amount of reading the certificate will change that.

Deriving the bending beam temperature: 0 °C

The second derived figure comes from a flat rule — the beam is tested at L + 10, which for a −10 grade puts the bath at 0 °C. Ten degrees warmer than the number in the name looks like a concession and is not one; the reasoning is set out immediately below. What is true is that the criterion ends up undemanding at this grade, for a quite separate reason taken up at the end of this section.

The instrument is straightforward. A rectangular beam of PAV-aged binder is supported at both ends in a cold bath, loaded at its centre, and its deflection is tracked. Two numbers are taken at the 60-second mark. Creep stiffness S is capped at 300 MPa: a binder stiffer than that converts the thermal contraction of a cooling pavement into more tensile stress than the surrounding mix can carry. The m-value — the slope of the log-stiffness against log-time curve at that same 60 seconds — has a floor of 0.300: whatever stress does build, the binder has to be able to bleed it away rather than hold on to it.

Why a reading at 0 °C verifies a −10 °C requirement

The offset is an application of time–temperature superposition, and the easiest way to see it is to stop thinking about temperature and start thinking about clock time. Bitumen does not have a stiffness; it has a stiffness that depends on how long you lean on it. Cooling the binder and loading it for longer are two routes to the same number, and the exchange rate between them can be measured.

Now ask what the test is imitating. Thermal cracking is not an impact event. A pavement sheds heat across a whole night, tensile stress accumulates in the bound layer for hours, and the binder spends those hours trying to relax it away. A laboratory test faithful to that would hold the beam at −10 °C and keep loading it for something on the order of two hours per specimen — unusable as a routine acceptance test, and at that duration the deflection also drifts towards the edge of what the instrument resolves reliably. The SHRP researchers spent ten degrees instead of two hours: 60 seconds of loading at L + 10 °C returns the same stiffness as roughly two hours of loading at L. So the bath is set to 0 °C, the reading is taken at one minute, and the requirement being verified is still the one at −10 °C.

This is the most misread line on a PG certificate, and it is worth stating flatly. On a PG 64-10 report, BBR at 0 °C is correct, and seeing it should reassure you rather than worry you. A BBR reported at −10 °C means the laboratory tested at L instead of at L + 10, which would be the right condition for a −20 low grade that does not exist in the M320 ladder, and its numbers cannot be measured against the M320 limits at all. And if the document carries no bending beam result of any kind, then whatever is printed at the top of it, the binder in front of you has never been performance graded.

Correct arithmetic, easy question

None of that rescues the criterion at this grade, and the reason has nothing to do with the offset. The offset is applied identically at every grade; what differs is where it starts from. A −34 grade puts the beam at −24 °C, and there the test genuinely hurts. A −10 grade puts the beam at 0 °C, which is a temperature at which a paving binder is, by its own standards, still working comfortably. Straight-run binders clear both limits by a mile — and so do binders that would come apart at −6 or −12 °C, which is exactly the population a discriminating test would have to separate.

So read a clean bending beam line on a tropical certificate for what it is: proof of compliance, not proof of quality. The two are far apart at this grade, and no amount of correct arithmetic closes the gap. The consequence is not that the buyer is helpless, but that the evidence has to be found elsewhere on the report and elsewhere in the standard — at the high-temperature end, where this grade is genuinely tested, and under AASHTO M332 where the traffic letter and the Jnr limit sharpen exactly that end. Both are dealt with in the sourcing section below.

The case for a −10 low grade rests on a single mechanism working in two different ways. Water vapour is a thermal blanket. Where the air is humid, radiative cooling at night is heavily suppressed: heat leaving the pavement is absorbed and re-radiated back by the atmosphere instead of escaping to space. The desert interior, with dry air and clear skies, loses that heat freely and can drop twenty degrees or more overnight. A humid coast cannot. The pavement is warm when the sun sets and it is still warm at dawn.

On the Gulf and Red Sea coast the mechanism is humidity plus the thermal mass of the sea. In the equatorial belt it is humidity plus the absence of a seasonal cycle: the sun is high all year, day length barely changes, and the difference between the hottest and coolest month can be smaller than the difference between morning and afternoon. Either way the outcome is the same. The pavement has a demanding summer requirement and effectively no winter requirement at all.

Where PG 64-10 fits, region by region

• The Gulf and Red Sea coastal strip. Dubai, Abu Dhabi, Doha, Kuwait City, Dammam, Manama, Muscat, Jeddah. Extreme summer surface temperatures, high humidity through much of the year, and winter nights that stay comfortably above freezing. The inland plateau a few hundred kilometres away is a different climate and usually a −16 grade.
• Coastal and equatorial West Africa. Lagos, Accra, Abidjan, Douala, Libreville, Cotonou. Year-round heat, very high humidity, heavy rainfall, and no cold season in any meaningful sense.
• Coastal East Africa and the Indian Ocean. Mombasa, Dar es Salaam, Maputo, Djibouti, the Comoros and Mauritius. Maritime moderation at the top end, nothing at all at the bottom.
• Southeast Asia and the archipelagos. Vietnam, Cambodia, southern Thailand, Malaysia, Singapore, Indonesia and the Philippines. Monsoon rainfall, permanent heat, and pavement temperatures that stay high across the whole year rather than for one season.
• Coastal South Asia and the islands. The Kerala, Tamil Nadu, Andhra and Odisha coasts, coastal Sri Lanka, Bangladesh and the Maldives, where the winter minimum is a non-event even inland from the shore.
• The Caribbean, Central America and northern South America. Island and littoral networks with the same profile: heat, humidity, salt air and intense rainfall, with no low-temperature requirement to design against.

Check the high number before you accept it

Because the low side is uncontroversial, the discussion in a tropical tender collapses onto the 64. It should not be assumed. A dark surface under vertical equatorial sun, or on a Gulf coast in July, can exceed a 64 °C seven-day maximum pavement design temperature, and the standard ladder has two steps above it — 70 and 76 — that exist precisely for this situation. Coastal Gulf specifications commonly sit at PG 70-10 or PG 76-10 for surface courses rather than at 64, and the correct grade for a specific project comes from the road authority’s own temperature data at its own chosen reliability level, not from a regional habit. If the pavement design temperature was derived at a lower reliability than the agency normally uses, or from an airport weather station some distance inland, the 64 deserves a question.

Peak temperature and time at temperature are not the same thing

Here is the point that separates a tropical PG 64 from a continental one, and it is not visible anywhere in the grade name. The high-temperature grade is built on the average seven-day maximum pavement temperature — a measure of the peak. It says nothing about how long the pavement spends near that peak over a year.

A PG 64-22 pavement in a continental interior reaches its design temperature for a few weeks in mid-summer and spends the rest of the year cool or cold. A PG 64-10 pavement on the equator is hot in February and hot in August. The cumulative hours spent above any rutting-critical temperature can be several times greater for the same nominal high grade. The practical implications run in two directions:

• Rutting is a longer exposure, not a shorter one. Permanent deformation accumulates with load repetitions applied while the binder is soft. The same 64 °C grade delivers noticeably less protection when the binder spends the whole year in that condition. This is one of several reasons a traffic grade bump is more often justified in the tropics than the traffic count alone would suggest.
• Oxidative ageing runs continuously. Oxidation is temperature-driven and does not pause for a season that never comes. The PAV protocol conditions binder for 20 hours at 2.1 MPa and is generally described as representing several years of in-service ageing, but that calibration was developed against temperate service. Treat it as an index rather than as a prediction of service life in a hot, permanently warm climate.

Two failure modes the grade does not address at all

In wet tropical service, the two problems that most often bring a pavement back for early rehabilitation are moisture damage — stripping of binder from aggregate under saturated conditions, accelerated by pore pressure under traffic — and surface ravelling on aged, over-oxidised binder films. M320 tests neither. Moisture sensitivity belongs to the mix, and is addressed through tests on the compacted mixture and where necessary an anti-stripping additive or hydrated lime. Ravelling resistance is largely an ageing question. Neither is a reason to change the binder grade, and both are reasons to treat the grade as one input to a design rather than as the design.

74 °C carries no implication of polymer modification

As a working rule across the industry, a useful temperature interval of up to about 92 °C can be reached with an unmodified straight-run binder, provided the crude is suitable. Beyond roughly 92 °C, polymer modification is normally required to hold both ends at once. At 74 °C, PG 64-10 sits three grade steps inside that boundary and should not be priced as a modified product. If an offer for this grade arrives described as polymer modified, ask why. Either the specification called for modification for a reason not visible in the grade name, or you are being quoted something the project did not ask for.

The real advantage of this grade is the headroom above it

Because the span starts low, PG 64-10 has unusual room to move for traffic. Under Superpave practice the high-temperature grade answers to loading as well as to climate. Slow transient loads add one step of six degrees; standing loads add two; very high traffic volumes put a further step on the table. The low grade is never touched, so every bump widens the span. Applied here, that gives:

• One step to PG 70-10 — span 80 °C, still routinely produced neat.
• Two steps to PG 76-10 — span 86 °C, still inside the span rule, though whether it comes neat depends on the feedstock rather than on the span, and a good deal of PG 76 in trade is supplied polymer modified.
• Three steps to PG 82-10 — span 92 °C, at the practical edge, and normally a polymer modified binder.

A caution belongs with that list, because the span rule answers only half the question. The span describes how hard it is to hold both ends at once; it says nothing about whether a given refinery stream can reach the top end at all. For a high grade of 64 °C or below, the span is effectively the whole answer, which is why PG 64-10 is a straightforward neat product. Above it the binding constraint migrates to the high end itself, and a supplier who cannot reach 2.20 kPa on residue at 76 °C with what comes out of the vacuum tower has to do something to the material — either start from a much stiffer base or add a polymer. Read the span as a lower bound on difficulty, not as a guarantee that a neat product exists.

Compare that with the same exercise starting from PG 64-22, where a single traffic bump to PG 70-22 already reaches 92 °C and moves the order into modified supply. In markets built around ports, container corridors, border crossings and congested urban intersections — which describes a great deal of the coastal tropical trade — this headroom is the most valuable commercial property of the −10 low grade. It means the traffic problem can usually be solved by buying a stiffer neat binder rather than by buying a modified one.

Resolve the traffic adjustment before the grade goes into the contract

A grade bump changes the product being bought, the crude that can produce it and the price basis, so it belongs at the enquiry stage and not after. Standing loads deserve particular attention on tropical projects because they are so common: a container yard, a weighbridge queue, a bus stop, a signalised approach on a hot afternoon. Rutting under a stationary axle is not a slower version of rutting under a moving one — the load is applied for orders of magnitude longer, and a binder graded for moving traffic has no answer to it.

M332 and the Jnr route

Some project specifications are written to AASHTO M332 instead of M320. The grade then gains a traffic letter — PG 64S-10, PG 64H-10, PG 64V-10, PG 64E-10 — and the high-temperature acceptance changes character entirely. In place of the G*/sin δ limit on RTFOT residue, the binder goes through a multiple stress creep recovery test to AASHTO T350, which loads and unloads the sample repeatedly at 64 °C and reports non-recoverable creep compliance Jnr alongside percent recovery. Jnr is capped according to the letter: 4.5 kPa⁻¹ for S, 2.0 for H, 1.0 for V and 0.5 for E, a lower number meaning greater rutting resistance. A second limit caps the change in Jnr between the 0.1 kPa and 3.2 kPa stress levels at 75 %, which discards binders whose response falls apart once the stress is raised.

The important structural difference is that M332 handles traffic by tightening Jnr at 64 °C rather than by raising the test temperature. On this grade that is a genuine improvement, because the criterion it replaces is the one line in the M320 set that was doing any work. Jnr is not present on an M320 certificate and there is no way to back it out of one, so the governing standard has to be settled at enquiry rather than discovered at inspection. PG 64S-10 is the direct counterpart of M320 PG 64-10; H, V and E are the same climate with progressively harder traffic written into it.

The specific risk at this grade: a pass that proves very little

The four slack criteria described earlier create a commercial exposure that does not exist at colder grades. A binder that has been heavily processed or over-oxidised can reach 64 °C on the high side without difficulty, and it will then clear the fatigue criterion at 31 °C and both bending beam criteria at 0 °C, because those tests are being asked at temperatures where nothing fails. The screens that would have caught it — the m-value at −6 or −12 °C, the fatigue parameter at 25 °C — are simply not being run. A certificate reading PG 64-10 across every line can therefore be entirely truthful and still be describing a binder you would not want in a road.

The protection is not exotic. Ask for three things alongside the grading report:

• The continuous grade at both ends. The M320 envelope is six degrees wide at each end, so a pass tells you less than it appears to. A continuous grading report gives the actual temperature at which each criterion was met — 66.4 and −19.8 for a binder offered against a PG 64-10 requirement, say — rather than the rounded step. On the high side the continuous grade tells you how much margin you have against a hot spell at the plant or a RAP-stiffened blend. On the low side it tells you something else: most straight-run paving binders continuous-grade far colder than −10, so a binder whose true low grade sits only just below −10 is unusual and worth a question about how it was made.
• Penetration, softening point and specific gravity. These are on every refinery certificate in this trade already. They cost nothing, and together they tell you whether the material behind the PG label is a normal paving binder or something that has been pushed hard to reach a high grade.
• RTFOT mass loss and the retained consistency figures. Ageing behaviour is where a heavily processed binder shows itself, and in a permanently warm climate it is the property that determines whether the surface is still flexible in year eight.

Tie the report to the tank

A PG grading report describes the batch it was cut from and nothing else. Production varies — the high-side continuous grade of a single refinery stream can move a degree or two between campaigns as the feed and the cut point change — and at 64 °C a degree or two is the entire margin. Last quarter’s report is therefore a document about last quarter. The test to apply is one of identity: does the grading report carry the same tank or batch number as the Certificate of Analysis issued for your cargo, and does it show a sampling date and a test date on its face? If those three things do not line up, the report is describing somebody else’s bitumen, however impressive the letterhead.

Setting plant temperatures: the 3 Pa·s line is not the answer

The rotational viscosity limit at 135 °C is a handling check and nothing more. It establishes that the binder will pump out of a tank and wet an aggregate. It is not a mixing temperature and should never be used as one. The working temperatures come from a different place: the viscosity–temperature chart of the specific binder in the tank, read across at the conventional equiviscous targets of roughly 0.17 Pa·s for mixing and 0.28 Pa·s for compaction. The word specific is doing work there — a generic table for the grade is not a substitute for the viscosity data of the cargo actually loaded.

Two points belong specifically to the climates this grade serves. The first is a rare piece of good news: a mat laid at 35 °C ambient in humid air sheds heat far more slowly than the same mat in a temperate autumn, so the compaction window is generous and there is correspondingly little excuse for over-heating binder to buy working time. The second is a limitation. The equiviscous method is not valid for polymer modified binders, whose viscosity is non-Newtonian, so the moment a traffic bump carries the order to a modified PG 76-10 or to PG 82-10, the calculation is replaced by the supplier’s written recommended range.

A penetration grade cannot be converted into a PG grade

There is no arithmetic that turns 60/70 into PG 64-10. This matters more here than almost anywhere, because the coastal tropical trade is overwhelmingly a penetration-grade trade: refineries supplying these markets produce and certify 60/70 and 80/100, and a PG 64-10 order is usually one of those production streams performance graded on a specific batch. A penetration figure is a needle pushed into unaged binder at 25 °C. It is a useful consistency check and it is not a rheological measurement: it predicts nothing about G*/sin δ at 64 °C, nothing about creep and relaxation at 0 °C, and nothing whatever about the material once the PAV has finished with it.

There is a temptation peculiar to this grade to treat the conversion as safe in one direction — to argue that since the low side asks so little, any 60/70 must be a PG 64-10. The low half of that argument is usually right and the high half is not. A 60/70 from one source may continuous-grade at 66 °C on the high side and from another at 62 °C, in which case it is a PG 58-10 and not a PG 64-10 at all, and no amount of reasoning about the tropics changes that. Where a specification is written in PG, the only thing that discharges it is a grading report showing the actual DSR and BBR results at the temperatures set out on this page. Absent that report, a PG grade on an offer is a claim about a measurement that nobody made.