Bitumen PG 58-10: Specification, Mild Climate Fit and Export Supply

PG 58-10 means the binder has been shown by test to satisfy the AASHTO M320 high-temperature criteria at an average seven-day maximum pavement design temperature of 58 °C, and its low-temperature criteria at a minimum pavement design temperature of −10 °C. Both figures describe the pavement, not the air above it. Neither is assigned by inspection, inferred from a penetration grade or carried over from a refinery data sheet. Both are measured on a rheometer, and the low-temperature pair is measured on binder that has been artificially aged twice before it is tested at all.

The grade name states one temperature; the report carries three. The other two are not laboratory choices and cannot be negotiated: 28 °C for the fatigue measurement on pressure ageing vessel residue, and 0 °C for the pair of bending beam readings on that same residue. Both fall out of the numbers 58 and −10 by rule. The dynamic shear work at 58 °C is done twice over, on the binder as produced and again after the rolling thin film oven. All three temperatures are derived in full further down this page, and the reason to derive them is practical rather than academic: a result printed without the temperature it was measured at — or at a temperature belonging to a neighbouring grade — is not evidence about PG 58-10 at all.

This is the least demanding grade in ordinary trade

The useful temperature interval — the high grade minus the low grade — is 58 − (−10) = 68 °C. Only PG 52-10 at 62 °C and the rarely traded PG 46 grades sit below it. Both ends are close to the edge of what the standard can ask: −10 is the warmest low-temperature step that exists in M320, and 58 °C is one step above the lowest high-temperature grade in practical use. Nothing here implies polymer modification, selective crude sourcing or a difficult production run.

That is a statement about the grade, not a criticism of it. There are real climates for which 58-10 is the honest and correct answer, and buying a stiffer binder for those sites spends money on protection the road cannot use. But it does mean that a PG 58-10 certificate, on its own, tells a buyer remarkably little — and it means the two most valuable questions on this page are about what to ask instead.

A requirement and an assigned grade are not the same document

This distinction matters more on PG 58-10 than on any other grade, and it is routinely missed because the two look identical on paper.

• PG 58-10 as a requirement. A project specification asks that the M320 criteria be satisfied at 58 °C and at −10 °C. A very large share of the ordinary straight-run paving binder in world trade satisfies that, including binders whose own assigned grade is far wider. A binder graded PG 64-22 meets a PG 58-10 requirement comfortably: it is stiffer than necessary at 58 °C, and at 0 °C it is far more relaxable than the bending beam criteria demand.
• PG 58-10 as an assigned grade. The two-number label a laboratory writes on a grading report is the pair of six-degree steps the binder actually reaches and no further. Assigning −10 therefore carries a specific and easily missed meaning: the binder was evaluated and did not satisfy the −16 step. Its critical low temperature lies somewhere between −10 and −16 °C. Assigning 58 on the high side carries the mirror meaning: the binder did not reach 64 °C.

Most straight-run paving binders grade a good deal colder than −16 °C at the low end. A binder whose assigned low grade is −10 is therefore not typical material. It is reporting unusually poor relaxation capacity at low temperature — the signature of a waxy crude, a heavily processed or partly blown product, or binder that has already been through more heat history than it should have. In a climate that never approaches freezing, none of that may matter to the finished road. It matters a great deal to what else you should be asking, because the same causes that flatten the low end tend to reappear as poor ageing behaviour, early surface embrittlement and ravelling.

So read the header of the report carefully. A grading report showing PG 64-22 supplied against a PG 58-10 requirement is compliance with margin at both ends. A grading report whose assigned grade is PG 58-10 is compliance with no margin at either end, and deserves a question. And a third case exists that looks like both: a laboratory that tested only at the temperatures the specification named, and therefore reports conformity rather than an assigned grade at all. Confirm which of the three you are holding before you read anything else on it.

What the grade does not tell you

A PG grade is a climatic envelope for the binder and nothing more. It says nothing about crude source, wax content or asphaltene fraction. It is not a mix design, it carries no test for adhesion or moisture damage, and it describes ageing only as far as the PAV protocol reaches. On a grade where four of the five performance criteria are undemanding, everything M320 does not cover carries proportionally more of the risk — which is the argument, made in full further down this page, for putting conventional consistency and ageing lines into the contract alongside the PG requirement rather than in place of it.

The high-temperature tests: both at 58 °C

The 58 needs no calculation, because it is itself the bath temperature for the dynamic shear rheometer (AASHTO T315 / ASTM D7175). Two separate acceptance lines are read there. The first uses the binder as it leaves the tank and asks for G*/sin δ of at least 1.00 kPa. The second uses residue from the rolling thin film oven — 85 minutes at 163 °C in a moving film under air, to AASHTO T240 / ASTM D2872 — and lifts the same parameter to at least 2.20 kPa. That step up is not severity for its own sake. Hot-mix production stiffens the binder before a wheel ever touches it, and M320 wants its rutting criterion applied to the material that will actually be in the road rather than to the material that went into the mixer.

So the aged line is the one that decides, since clearing 2.20 kPa after the oven implies clearing 1.00 kPa before it. What matters on PG 58-10 is how little deciding it does. Fifty-eight degrees is one rung above the softest high-temperature grade anyone trades in export quantities, and a straight-run vacuum residue of ordinary paving consistency reaches it from most crude slates without assistance. Expect offers to clear this line rather than compete on it. If you want a summer-end number that actually separates one offer from another, the continuous high grade does that job and the pass mark does not.

The intermediate temperature: 28 °C for this grade

The fatigue criterion is measured at neither grade temperature. M320 puts it at the midpoint of the two and then adds four degrees. Worked through for the pair 58 and −10:

• (H + L) ÷ 2 + 4
• (58 + (−10)) ÷ 2 + 4
• 48 ÷ 2 + 4
• 24 + 4 = 28 °C

The requirement at 28 °C is that pressure ageing vessel residue keeps G*·sin δ at or below 5000 kPa. Two features of that parameter routinely confuse people. It is a product, not a quotient, unlike the rutting parameter. And a low value is what you want: an aged binder that is stiff and elastic at the same time stores strain energy under each wheel pass rather than dissipating it, and eventually returns that energy as a crack.

Because the intermediate temperature depends on the mean of the two grade numbers, several different grades share one value. 28 °C is also the intermediate temperature for PG 64-16 and for PG 70-22, since all three pairs sum to 48. It moves in three-degree steps as either end changes: PG 58-16 is tested for fatigue at 25 °C and PG 58-22 at 22 °C. A report showing the fatigue parameter measured at 25 °C is a report on a different grade, whatever the header says — that is not a rounding difference, it is a different test on a different sample condition.

The engineering point is more important than the paperwork one. At 28 °C a paving binder is soft, G*·sin δ is a small number, and the 5000 kPa ceiling is cleared by an enormous margin. The fatigue criterion is not screening candidates at this grade. Fatigue cracking is still a genuine failure mode on thin pavements over weak subgrades, and on this grade the protection against it has to come from the structural design, the layer thicknesses and the mix — not from the binder specification.

The low-temperature test: 0 °C for a −10 grade

Add ten to the low grade and you have the bending beam temperature. Here that lands on 0 °C — the freezing point itself, not ten degrees below it. Buyers who notice the gap usually assume the laboratory has been generous with them. It has not, and the section below explains why the warmer bath measures the colder condition exactly.

Mechanically the apparatus is undramatic. A moulded beam of PAV-aged binder, 125 mm long, 12.5 mm wide and 6.25 mm deep, rests on two supports in a temperature-controlled fluid bath while a constant midspan load of roughly 980 mN is applied to it. Deflection is logged out to 240 seconds, and both reported figures are taken from the reading at 60 seconds. Creep stiffness S is capped at 300 MPa, because a binder stiffer than that converts the layer’s thermal contraction into more tensile stress than the surrounding mix can carry. The m-value — the local slope of log stiffness against log time at that same instant — has a floor of 0.300, which is the standard asking whether the binder can still let stress go once it has built. Stiffness and relaxation are two separate questions, and M320 insists on both answers.

Why ten degrees warmer measures the colder condition

The offset comes from time–temperature superposition, and the cleanest way to understand it is in terms of the stiffness curve rather than the physics of the binder. Plot binder stiffness against loading time on logarithmic axes and you get a curve. Change the temperature and the curve does not change shape — it slides sideways. Bitumen is thermorheologically simple, which is exactly the statement that cooling and waiting are interchangeable: any stiffness reachable by cooling can be reached instead by holding the load longer, and the exchange rate between the two is a measurable shift factor.

For paving binders in this temperature region, the accepted exchange is that 60 seconds of loading at L + 10 °C gives the same stiffness as roughly two hours of loading at L. Sixty seconds to two hours is a factor of 120 — a little over two decades of time — so a ten-degree change in temperature is buying about two decades of loading time. Two hours is the physically meaningful dwell, because a pavement does not touch its annual minimum for an instant; it sits near it through the coldest hours of a night while thermal stress accumulates and the binder tries to relax it.

Running the beam at 0 °C for 60 seconds therefore returns the same information as running it at −10 °C for two hours, and it does so with two practical advantages: the deflection stays inside the range the instrument can resolve accurately, and an acceptance test that would have taken two hours per specimen takes one minute. Note also that the bending beam temperature depends only on the low grade, so 0 °C is the correct test temperature for the entire −10 family — PG 52-10, PG 58-10, PG 64-10, PG 70-10 and PG 76-10 are all graded from one bath temperature.

Two things follow for a buyer. On a PG 58-10 certificate, BBR at 0 °C is the expected and correct entry, and seeing it should reassure you rather than worry you. If BBR is reported at −10 °C, the laboratory has tested at L rather than at L + 10, which is the condition for a low grade the standard does not contain. If there is no BBR result at all, the binder has not been performance graded, whatever the offer, the tank label or the drum says.

The honest limitation at this grade

The superposition argument is sound, and at −22 or −34 the resulting test is genuinely demanding. Here it is not, and the reason is where the ladder stops rather than where the offset lands. Testing at 0 °C asks a paving binder to stay relaxable at a temperature at which it is, by its own standards, still warm. Ordinary binders pass by a wide margin — and so do binders that would fail badly at −6 or −12 °C. A clean bending beam line on a PG 58-10 report is evidence of compliance, not evidence of quality, and the two are far apart here.

If a mild-climate project genuinely wants assurance about low-temperature brittleness, M320 cannot supply it at this grade because there is no warmer step to test at. The assurance has to come from outside the two-number label: from the continuous low grade, which reports the actual temperature at which the criteria were met rather than a rounded step, or from a European-system test such as the Fraass breaking point to EN 12593, which measures brittleness directly and appears on many refinery certificates already.

A 68 °C useful temperature interval is a description of a climate before it is a description of a binder. It says the pavement is never very hot and never cold, which is an unusual combination: the same mechanisms that keep winter mild usually let summer run hot, and the ones that cap summer usually let winter bite. Four mechanisms produce the exception, and they are worth naming because a buyer can check whether the project sits under one of them.

Cold currents and coastal upwelling

Where a cold ocean current runs along a subtropical coast, or wind-driven upwelling brings deep water to the surface, the sea acts as a fixed low-temperature reservoir a few kilometres from the road. Summer air arriving over that water is capped near the sea surface temperature, while the latitude keeps winter frost-free. The Canary Current along Atlantic Morocco and Iberia, the Humboldt Current along the Peruvian and northern Chilean coast, and the Benguela along the Namibian and Angolan coast all produce this profile. Coastal Peru is the extreme case: a desert with almost no rain, almost no seasonal swing, and pavement temperatures that live in a band a temperate engineer would find implausible.

Persistent marine stratus

The same upwelling coasts, and the marine-layer coasts of central California and the Atlantic islands, sit under low cloud for a large part of the warm season. That matters more than the air temperature alone suggests. The standard pavement temperature model works from air temperature with a latitude correction, and the latitude correction stands in for the solar gain typical of that latitude. A road under stratus for most of the summer receives considerably less than that, and its surface runs closer to air temperature than the model implies. This is one of the few situations in which the derived high-temperature grade may be conservative rather than optimistic.

Altitude at low latitude

The highland tropics produce the narrowest annual ranges on earth. Altitude caps the daytime maximum through the lapse rate; latitude caps the nocturnal minimum because there is no winter to fall into. Quito, Bogotá, Nairobi, Addis Ababa, Antananarivo and Kunming all sit in this band, along with a great many secondary cities in the Andes, the East African highlands and the Ethiopian plateau. One caution belongs with it: thin, dry highland air radiates heat away efficiently at night and admits intense solar radiation by day, so the diurnal range can be far wider than the annual range, and both ends deserve checking against real pavement data rather than monthly means.

Small oceanic islands

An island small enough that no point is far from the sea takes the ocean’s annual cycle almost unmodified. The Azores, Madeira, the Canaries, Cape Verde, Mauritius, Réunion and most of the Pacific island states behave this way. There is an operational consequence as well as a climatic one: on a small island network the entire road stock sits in one climate, so a single binder grade can serve the whole programme, and the logistics of importing one grade in container quantities usually outweigh any argument for splitting the order.

Temperate southern maritime

Coastal New Zealand, Tasmania and coastal Victoria, the Western Cape, and the coastal strip of Uruguay and Buenos Aires province combine genuinely moderate summers with winters that are cool rather than cold. These are the classic mild maritime climates of the southern hemisphere, and they sit at the temperate end of the same envelope the tropical highlands occupy from the other direction.

A note on which of these markets actually buys in PG

Several of these regions specify binders in the European system rather than the American one. Atlantic Iberia, the Atlantic Moroccan coast and the Macaronesian islands work to EN 12591 penetration grades; New Zealand and much of southern Africa have their own traditions. Performance grading reaches these markets mainly through internationally specified or donor-funded projects, through agencies that have adopted Superpave deliberately, and through Latin American road authorities that use PG grades directly. It is worth confirming at the enquiry stage which document actually governs the cargo, because a project can be described in PG in the tender and in EN penetration grades in the contract, and the two require different certificates.

When 58 is the honest answer and when it is an optimistic one

Everything above describes sites where PG 58-10 is right. The uncomfortable half of the picture is that a project can also arrive at this grade by accident, and the arithmetic that produces it looks identical either way. A pavement design temperature is a computed number, and it is computed from choices: which weather station, which period of record, which reliability level, what depth in the layer, and whether a traffic adjustment was applied afterwards. Change any of those and the answer can move a full grade step.

The reason to care is that the risk here is strongly asymmetric, and it is asymmetric in the opposite direction to a cold-climate grade. Under-specifying the high end costs rutting, shoving at junctions and bleeding in the first two or three summers — a structural failure that ends in milling and replacement. Over-specifying the high end costs a modest difference in price and, at a −10 low grade, effectively nothing in low-temperature performance, because a stiffer binder still clears the bending beam criteria at 0 °C with room to spare. On a PG 58-22 project that trade does not exist: there, a stiffer binder actively causes thermal cracking, so over-specifying is a real penalty. Here it is not. When the site data is uncertain, checking upward is cheap and being wrong downward is expensive.

The table below sets out the specific levers, what each one does, and what it means when the answer comes out as PG 58-10.

68 °C carries no implication of polymer modification whatsoever

The working rule in the trade puts the ceiling for an unmodified straight-run binder from a suitable crude at a useful temperature interval of roughly 92 °C; past that, holding both ends at once normally takes a polymer. At 68 °C, PG 58-10 sits four grade steps inside that boundary. It should be produced neat, quoted neat and priced neat. If an offer for this grade arrives described or priced as a polymer modified binder, ask what the modification is being bought for, because nothing in the grade name calls for it.

The headroom above it is the grade’s most useful commercial property

Because the span starts so low, PG 58-10 can absorb traffic adjustments without ever leaving neat-binder territory. Under Superpave practice the high-temperature half of the grade answers to loading as well as to climate: add one six-degree step where traffic is slow but still moving, two where it stands. Applied to this grade:

• One step to PG 64-10 — span 74 °C, routinely produced neat.
• Two steps to PG 70-10 — span 80 °C, still frequently produced neat from a suitable crude.
• Three steps to PG 76-10 — span 86 °C, achievable neat from some sources and modified from others.

Compare that with the same exercise starting from a cold-climate grade, where a single traffic bump can push the span past 92 °C and turn a neat order into a modified one. Either way the instruction is the same: fix the traffic bump before the grade is written into the contract, not after it. It decides which binder is being bought, which feedstock can make it and what the price is based on. Standing loads deserve the closest attention, because a parked or crawling axle is not simply a slow-moving one — the dwell is longer by orders of magnitude, and a binder graded for traffic that keeps moving has nothing to offer against it.

If the specification is written to M332

Some agencies grade to AASHTO M332 instead, in which case a traffic letter is appended — PG 58S-10, PG 58H-10, PG 58V-10 or PG 58E-10 — and the rutting check changes character entirely. Out goes the G*/sin δ line on RTFOT residue; in comes multiple stress creep recovery to AASHTO T350, run at the high grade temperature of 58 °C and reported as non-recoverable creep compliance (Jnr) with percent recovery beside it. Jnr at the 3.2 kPa stress level must not exceed 4.5 kPa⁻¹ for S, 2.0 kPa⁻¹ for H, 1.0 kPa⁻¹ for V or 0.5 kPa⁻¹ for E — smaller means more rutting resistance — and a second limit holds the increase in Jnr between the 0.1 kPa and 3.2 kPa stress levels to 75 %, catching binders that behave respectably until the load gets serious. The M332 counterpart of an M320 PG 58-10 is PG 58S-10.

One observation specific to this grade: an S designation at 58 °C is a very undemanding requirement, and an H, V or E designation at 58 °C is unusual enough to be worth a second look. Traffic heavy enough to justify V or E normally comes with slow or standing loads, and those are exactly the conditions under which the climatic grade itself is usually bumped. A specification asking for severe traffic performance at the mildest high-temperature grade in the ladder is not wrong, but it is uncommon, and confirming it costs one message.

Because the label is not selective here, put the selective lines in the contract

This is the practical conclusion of the whole page. On PG 58-22 the low end does the screening; on PG 64-10 the RTFOT rutting line does it. On PG 58-10 nothing in M320 does much screening at all, so a compliant certificate is compatible with a wide range of very different materials. The protection is not exotic, and none of it is expensive:

• The continuous grade at each end. The M320 envelope is six degrees wide at both ends, so the label conceals a lot. A continuous grading report gives the actual temperatures at which the criteria were met — 61.7 and −13.4, say — rather than the rounded steps. The laboratory already calculated both figures in order to assign the label, so asking for them costs nothing. On this grade the low continuous figure is the more revealing of the two, for the reason set out at the top of this page.
• Penetration at 25 °C and softening point. ASTM D5 and ASTM D36 are on every refinery certificate in this trade already. Together they tell you what kind of binder is behind a PG pass that four of five criteria could not have prevented, and they are the language most of these markets actually contract in.
• RTFOT mass loss, plus retained penetration and softening point rise. Oxidation is where a heavily reprocessed binder eventually gives itself away, and the climates that call for PG 58-10 give it no respite: a site that is mild all year has no cold season to slow the reaction down, and the surface is expected to stay flexible through every month of it.
• Solubility in trichloroethylene to ASTM D2042. The anti-adulteration line. It has nothing to do with performance grading and everything to do with whether the product is genuine bitumen. Its absence from a certificate is a finding rather than an oversight.
• Specific gravity to ASTM D70. The figure that converts the cargo between tonnes and cubic metres, and therefore the one that matters at the discharge port.

A grading report must be tied to your cargo

Grading is performed on a sample, and a sample belongs to a batch. A report issued against last quarter’s production is a document about that production and nothing else, however impressive it looks in a PDF. What belongs in your file is a DSR and BBR sheet carrying the same tank or batch identifier as the Certificate of Analysis for your parcel, with both the sampling date and the testing date printed on its face. The temptation to recycle an old report is strongest precisely here, on a grade so easily met that a fresh test would almost certainly have agreed with the stale one — which is an argument for asking, not an argument for letting it pass.

Working temperatures come from the binder’s own viscosity curve

The 3 Pa·s ceiling at 135 °C is a handling requirement rather than a performance one: it confirms the binder can be pumped and will coat aggregate. For an unmodified binder the plant temperatures are read off the binder’s own viscosity–temperature plot, at the conventional equiviscous targets of roughly 0.17 Pa·s for mixing and roughly 0.28 Pa·s for compaction. Take those from the viscosity data supplied with the cargo rather than from a handbook table: two binders that both satisfy the 3 Pa·s ceiling at 135 °C can still want noticeably different settings at the plant.

Two points are specific to the climates this grade serves. Mild maritime sites are windy and often damp, and wind speed cools a mat far faster than air temperature alone predicts — a 15 °C breezy coastal afternoon can give a shorter compaction window than a still 8 °C inland morning. The correct response is to plan rolling patterns and paver speed around it, not to compensate by raising the binder temperature, which ages the binder and costs exactly the flexibility a mild-climate surface is expected to keep for twenty years. And the equiviscous approach is not valid for polymer modified binders, so if a traffic bump moves the order to a modified PG 76-10, the supplier’s recommended range replaces the calculation.

A penetration grade cannot be converted into a PG grade

No conversion factor, table or rule of thumb turns 60/70, 80/100 or EN 70/100 into PG 58-10, and none is waiting to be found. A penetration figure is one needle, at one temperature, into unaged binder. It says nothing about how the material behaves at 58 °C, nothing about how it behaves at 0 °C, and nothing whatever about its condition after twenty hours in a pressure ageing vessel — which is the state three of the five M320 criteria are judged in.

The temptation is sharper on this grade than on any other, and it is worth naming. Because a PG 58-10 requirement is so easily met, it is tempting for everyone involved to assume a standard 60/70 or 80/100 cargo will satisfy it and to proceed without testing. Most of the time that assumption would be correct. It is still not evidence, and it produces nothing to hand the engineer when the certificate is asked for. Two cargoes in the same penetration band from different crudes can continuous-grade several degrees apart at both ends, and the one that comes in at 57 °C on the high side is a PG 52 binder no matter how ordinary it looks. When the contract says PG, one document settles it and nothing else does: a grading report carrying the measured DSR and BBR values at the temperatures derived on this page, against the batch on the vessel. A data sheet, an assurance or a grade written on a drum is a claim nobody has measured.