Bitumen Test Methods: The Complete Reference

Bitumen is not characterised by its chemistry in commercial trade. It is characterised by how it behaves under a set of standardised physical tests, each of which fixes a temperature, a load, a duration and a geometry. Change any one of those and the number changes. That is why a specification line is only complete when it names three things: the property, the limit and the test method.

This has a direct commercial consequence. If your contract says "viscosity minimum 350" without naming a method, it is unenforceable. Absolute viscosity at 60 °C is measured in poise and runs into the thousands. Kinematic viscosity at 135 °C is measured in centistokes and runs in the hundreds. Rotational viscosity is measured in pascal-seconds and runs below ten. All three are "viscosity". A supplier could satisfy the literal words of that clause with a product that fails your project entirely.

The three questions that make a number mean something

Every value on a bitumen certificate should survive three questions. Most disputes in this trade are a failure of one of them, and the second is the one almost nobody asks.

• By which method? The designation fixes the apparatus, the load, the duration and the geometry. ASTM D5 and EN 1426 both measure penetration and are close in procedure; they are still two documents, and a project specification names one of them.
• On which material? The binder as supplied, the residue from a short-term ageing procedure, or the residue from a pressure ageing vessel are three different materials. The same property measured on each gives three different numbers, and a specification limit belongs to exactly one of them. This is the subject of the ageing section below and it is the single most misread part of any certificate.
• At which temperature and reference condition? Viscosity at 60 °C and at 135 °C are unrelated numbers. Density to ASTM D70 may be reported at 15.6 °C or 25 °C, EN 15326 at 25 °C and IS 1202 at 27 °C. Softening point is measured in a water bath below 80 °C and in glycerin above it. A value quoted without its reference condition is incomplete.

The four standard systems

• ASTM — the American Society for Testing and Materials. Dominant in the Americas, the Middle East and much of Asia, and the system most bitumen export documentation is written in.
• AASHTO — the American Association of State Highway and Transportation Officials. Runs in parallel with ASTM, is often near-identical in procedure, and is the system that carries the Superpave performance grading specifications. Where a US state or a donor-funded highway project writes a specification, it usually writes it in AASHTO numbers.
• EN — the European standard series, published nationally as BS EN, DIN EN and so on. Required for material placed on the European market and referenced across North Africa and parts of the CIS.
• IS — the Indian Standard series from the Bureau of Indian Standards. Governs the Indian market, which is one of the largest bitumen import markets in the world.

The methods are usually equivalent in intent and very close in procedure, but they are not always numerically interchangeable, and in a few cases they are not even measuring the same quantity. Solubility is the clearest example: ASTM D2042 dissolves the sample in trichloroethylene and EN 12592 dissolves it in toluene. Both report a percentage and both are called solubility. Where a project specification names a system, test to that system.

The unit arithmetic: one conversion that is legitimate, three that are not

Viscosity units cause more contractual confusion than any other line on a certificate, and the arithmetic is worth stating once.

• 1 poise = 0.1 Pa·s, so 3 Pa·s is 30 poise. This is a unit change, not a test change.
• 1 centistokes = 1 mm²/s. Again a unit change only.
• Dynamic viscosity = kinematic viscosity × density, at the same temperature. In working units, dynamic viscosity in mPa·s equals kinematic viscosity in mm²/s multiplied by density in g/cm³. The condition matters: the density must be the density at the test temperature, not the density at 15 °C on the certificate.
• What cannot be converted: viscosity at 60 °C into viscosity at 135 °C; penetration at 25 °C into any viscosity; and a penetration grade into a performance grade. Each of these requires a different measurement, not a calculation, because the ratio between them depends on the temperature susceptibility of the individual binder.

What this page does, and what belongs elsewhere

This is a cross-reference: what each method is, what it physically does, what material it runs on and what a poor result predicts. The limits live with the products. Grade tables for penetration, viscosity and performance grades are on the specifications page and the individual grade pages; the full AASHTO M320 sequence with test temperatures by grade is on the performance grading guide; drawing the sample in the first place is on the sampling procedure page. Nothing here restates those. This page is the layer underneath all of them.

Bitumen hardens by oxidation. It hardens quickly and severely in the few minutes it spends as a thin film on hot aggregate in a mixing plant, and then slowly for the rest of its life in the pavement. Because the binder that matters is the aged binder, the standards do not test only the material in the tank. They deliberately manufacture aged material in the laboratory and test that instead.

That produces three distinct materials, and every line on a certificate belongs to exactly one of them.

• The binder as supplied. Also called original binder, unaged binder or tank binder. Penetration, softening point, ductility, flash point, solubility, density and the viscosity lines are normally reported on this material.
• Short-term ageing residue. The material left after a thin film oven test (ASTM D1754) or a rolling thin film oven test (ASTM D2872). It stands for the binder as it comes out of the mixer and goes down behind the paver. Mass change, retained penetration, viscosity ratio, residue ductility and the DSR rutting criterion on aged binder are reported on this material.
• Long-term ageing residue. The material left after the pressure ageing vessel (ASTM D6521), which is run on short-term ageing residue, never on fresh binder. It stands for the binder after several years in service. The DSR fatigue criterion and both BBR criteria are reported on this material and on nothing else.

Why this is the most misread part of a certificate

Take retained penetration, which is the line that predicts premature cracking better than anything else on a paving certificate. It is not a measurement; it is a ratio. It is the penetration of the ageing residue expressed as a percentage of the penetration of the binder as supplied. ASTM D946 requires a minimum of 52 % for grade 60-70. So a binder that penetrated 65 dmm as supplied and 34 dmm after the thin film oven test has retained 52 % — exactly on the floor, and a legitimate pass. A certificate that shows only "penetration 34" with no context has reported a failing paving grade. A certificate that shows only "retained penetration 52 %" with no residue value behind it has reported arithmetic nobody can check.

The same trap sits under every aged-residue line. A softening point that has risen ten degrees is alarming as an as-supplied value and unremarkable as a residue value. A viscosity at 60 °C that has quadrupled is a rejection as an original value and precisely the IS 73 limit as a ratio across the ageing step. A number with no material basis is not a result, and it is not a discrepancy either — it is an unanswered question.

TFOT and RTFOT: the same idea, two different films

Both procedures hold the binder at 163 °C in a ventilated oven and both are trying to reproduce the mixing plant. They differ in how the film is presented to the air, and that difference is the whole reason the industry moved.

The thin film oven test, ASTM D1754, places 50 g of binder in a flat-bottomed pan on a rotating shelf, giving a static film about 3.2 mm deep, and holds it for five hours. The surface of that film oxidises and forms a skin. Below the skin, the binder is protected. The result is a residue that is aged unevenly, and the procedure under-represents what happens to a binder that is genuinely being turned over.

The rolling thin film oven test, ASTM D2872, puts a fixed mass of binder into open-ended glass containers on a rotating carriage, with an air jet blowing into each container as it passes. The bottles turn at 15 rpm and the air flow is 4000 mL/min. The binder is continuously re-coated over the inside of the bottle, so fresh material is presented to the air throughout, and the whole sample is aged instead of only its surface. It takes 85 minutes rather than five hours, and it correlates better with plant ageing for the same reason.

RTFOT has one known limitation, and it is worth knowing because it explains an entire European standard. Very stiff binders and heavily modified binders may not flow enough to roll inside the bottle. The film does not renew, and the test quietly reverts to a slow, uneven version of TFOT. That is why EN 12607-3 exists: a rotating flask procedure for binders the rolling bottle cannot handle. If you are buying a hard grade or a highly modified binder and the certificate reports RTFOT, it is a fair question whether the material actually rolled.

Which specification uses which, and why IS 73 changed

The choice is not the laboratory's. It is written into the specification the cargo is sold against, and mixing them up is a real defect in a certificate.

• ASTM D946, penetration-graded asphalt binder, specifies the thin film oven test (D1754). This is why a 60/70 export certificate normally shows a TFOT line.
• ASTM D3381, viscosity-graded asphalt, does both. The AC grades are graded on the viscosity of the original binder at 60 °C and their residue requirements come from the thin film oven test. The AR grades are graded on the viscosity of the rolling thin film oven residue at 60 °C. The grade designation itself tells you which material carries the number.
• EN 12591, paving grade bitumens, specifies resistance to hardening by EN 12607-1, the rolling thin film oven test, and then sets requirements for change of mass, retained penetration, softening point after hardening and increase in softening point. Those four lines are all residue lines. The numeric limits differ by grade and are on the specifications page.
• IS 73 moved Indian paving bitumen off penetration grading and onto viscosity grading, and the current revision, IS 73:2013, specifies the aged properties on RTFOT residue — the viscosity ratio at 60 °C and the ductility of the residue. There is no Indian rolling thin film oven method; the standard references ASTM D2872 directly. IS 9382, which is often quoted alongside it, is the Indian thin film oven test.
• AASHTO M320 and ASTM D6373, the performance grading specifications, use RTFOT for short-term ageing and then the pressure ageing vessel on that residue for long-term ageing. Every criterion in the specification names its material explicitly.

The practical consequence is short. A certificate reporting a thin film oven test against a viscosity-graded or performance-graded specification has used the wrong ageing procedure, and the residue values it reports cannot be compared with the limits it is quoted against. That is not necessarily fraud; it is more often a laboratory running the procedure it owns. It is still a reason to ask before the cargo loads rather than after.

Mass change, and why a gain is not a mistake

The ageing procedures report mass change, and buyers routinely assume the number must be a loss. It need not be. Two things happen at once in the oven: volatile material leaves the binder, which reduces mass, and oxygen is chemically taken up, which increases it. The reported figure is the net of the two, so a well-refined paving grade with little volatile content can show a mass gain. Specifications recognise this and normally cap the change in either direction. What should worry you is a large loss, because that says volatile material was present that has no business in a paving grade — and the first cross-check for that is the flash point, not the ageing line.

Cutbacks and emulsions: the residue is made, not aged

For cutbacks and emulsions the certificate carries a residue too, but it is produced by driving off solvent or water rather than by ageing. The tests that matter are the tests on that residue, because that is the binder that will still be on the road next year.

For cutbacks, ASTM D402 distils the sample under controlled conditions and the residue is then tested for penetration, ductility and solubility. For emulsions there is a choice, and the choice changes the answer. ASTM D6997 recovers residue by distillation, which uses heat. ASTM D6934 recovers it by evaporation, and ASTM D7497 by a low-temperature evaporative technique. For an unmodified emulsion the three are broadly comparable. For a polymer-modified emulsion they are not: distillation temperatures can damage the polymer, so a residue recovered by distillation can test worse than the same emulsion recovered gently. A residue specification on a modified emulsion should name the recovery method, and a certificate should state which one was used.

The one-line rule to write into a contract

Require that every property derived from an ageing or recovery step is reported with the procedure that produced the material, by designation, and that the underlying values are shown alongside any ratio or percentage. In practice that means three lines rather than one: the as-supplied value, the residue value, and the ratio between them. It costs the laboratory nothing, it makes the arithmetic checkable, and it removes the single most common ambiguity in bitumen documentation.

Penetration — ASTM D5 / AASHTO T 49 / EN 1426 / IS 1203

A sample is conditioned in a water bath at 25 °C. A standard needle carrying a total load of 100 g is released onto the surface for exactly 5 seconds, and the depth it reaches is measured in tenths of a millimetre. That depth is the penetration value. A result of 65 means the needle sank 6.5 mm.

What it tells you: consistency at one temperature, under one load, for one duration. The needle is not measuring hardness in the metallurgical sense; it is measuring the combined elastic and viscous response of the binder to a small load over five seconds. That is also its limitation. It says nothing about how the binder behaves at 60 °C or at −10 °C, which is precisely why viscosity grading and performance grading were developed. Penetration can also be run at other temperatures where a specification calls for it — 4 °C and 46 °C both appear — and a value at a non-standard temperature must say so, because it is not comparable with the 25 °C figure.

What a failure predicts: above the band, a softer binder than the mix was designed for, with rutting and bleeding as the hot-weather consequence. Below the band, a stiffer binder with better rut resistance but worse fatigue and thermal cracking resistance, and a higher mixing temperature requirement at the plant.

Softening point, ring and ball — ASTM D36 / AASHTO T 53 / EN 1427 / IS 1205

Two brass rings are filled with bitumen and a 3.5 g steel ball is placed on each. The assembly sits in a bath heated at a controlled 5 °C per minute. The softening point is the temperature at which the binder sags far enough to touch a plate 25 mm below.

The bath liquid is part of the method and it matters commercially. ASTM D36 and EN 1427 use freshly boiled distilled water for softening points from 30 to 80 °C and glycerin above 80 °C, up to about 157 °C. A laboratory that only has a water bath physically cannot complete this test on an oxidized grade, because the water boils before the specimen softens. If you are buying an oxidized 85/25, 90/15, 95/25 or 115/15, the softening point line on the certificate has to have come out of a glycerin bath, and it is a fair question to ask.

What it tells you: the approximate temperature at which the binder stops behaving as a solid. For a buyer in a hot climate this is the line to scrutinise. Bitumen has no true melting point — it softens progressively — so this is a defined-condition test rather than a physical constant, and a softening point from a different method is not comparable.

Read it with penetration, never alone. The two together describe the temperature susceptibility of the binder, and the relationship between them is the cheapest diagnostic available on any certificate. A high softening point sitting next to an in-band penetration is the signature of blown or blended material in something sold as straight-run. The penetration index calculation that formalises this check, with a worked example, is set out on the fraud prevention page.

Absolute viscosity at 60 °C — ASTM D2171 / AASHTO T 202 / EN 12596 / IS 1206 (Part 2)

Bitumen is drawn through a calibrated capillary tube at 60 °C under a partial vacuum of 300 mm Hg, and the flow time is recorded. The vacuum is not an incidental detail: at 60 °C a paving grade will not flow through a capillary under gravity in any usable time, so the method applies suction to make the measurement possible. The result is reported in poise.

What it tells you: stiffness at a temperature close to a hot pavement surface in service. This is a better predictor of rutting than penetration at 25 °C, which is exactly why Indian paving bitumen moved to viscosity grading under IS 73, and why the ASTM AC grades are defined on it.

Kinematic viscosity at 135 °C — ASTM D2170 / AASHTO T 201 / EN 12595 / IS 1206 (Part 3)

Flow through a capillary under gravity at 135 °C. The result is reported in centistokes, which is millimetres squared per second.

What it tells you: behaviour near mixing temperature — whether the binder will pump and coat aggregate properly at the plant. The same method is run at 60 °C on cutback grades, where the solvent makes the material fluid enough for gravity flow at that temperature. A kinematic viscosity quoted without its temperature is therefore genuinely ambiguous on a cutback certificate.

Rotational viscosity — ASTM D4402 / AASHTO T 316 / EN 13302

A spindle rotates in the sample at a controlled temperature and the torque required is measured, giving viscosity directly in pascal-seconds. Under the Superpave specifications a binder must not exceed 3.0 Pa·s at 135 °C so that it remains pumpable; 3.0 Pa·s is 30 poise. The test is commonly run at 135 °C and 165 °C together, because two points define the viscosity-temperature relationship used to set mixing and compaction temperatures.

Newtonian and non-Newtonian: why the method has to change for modified binders

An unmodified bitumen at 135 °C behaves close to a Newtonian liquid: its viscosity does not depend much on how fast it is being sheared, so a capillary result and a rotational result agree, and a capillary number is a property of the material. A polymer modified binder is not Newtonian. It is shear-thinning, meaning its apparent viscosity falls as the shear rate rises. Run it through a capillary and the result depends on the shear rate the capillary happened to impose, which is not controlled and not reported.

That is the reason the rotational method is specified for modified binders, and the reason a rotational viscosity result is incomplete without the spindle, the rotational speed and the temperature. Both ASTM D4402 and EN 13302 require those to be reported. A PMB certificate showing "viscosity at 135 °C" with a number and nothing else has reported a value that cannot be reproduced. The same physics applies with more force to crumb rubber modified binders, which are suspensions rather than solutions, and where the spindle geometry and the settling behaviour of the rubber both influence the reading.

Grade limits for all of these live with the products: see penetration grades, viscosity grades and grade equivalence.

Solubility — ASTM D2042 / AASHTO T 44 / EN 12592 / IS 1216

A weighed sample is dissolved in solvent and filtered. True bitumen dissolves; mineral matter does not. The insoluble fraction is weighed and the result is reported as the percentage that dissolved. A minimum of 99.0 % is the normal requirement, and it is the requirement ASTM D946 sets for penetration grades.

The solvent is part of the method. ASTM D2042 uses trichloroethylene. EN 12592 uses toluene. ASTM D5546 is a toluene method using a centrifuge, and ASTM D7553 uses n-propyl bromide, which exists because trichloroethylene is restricted in a number of jurisdictions on health grounds. The results are close but the methods are not interchangeable, so a solubility figure should name its solvent as well as its designation.

What it catches, and what it does not. This is the anti-adulteration line, and it catches extension with mineral filler, sand, dust or any other insoluble material. It is the most useful cheap test on the certificate for that purpose. But it is worth being precise about its limit: an organic extender dissolves in trichloroethylene and passes. Used engine oil, heavy refinery residues, slop and other soluble hydrocarbons will not show up here at all. A certificate reporting 99.6 % soluble has ruled out mineral extension and has said nothing whatever about organic extension.

What does catch organic extension is a pattern rather than a single line: an unexpectedly low flash point, a large mass loss in the ageing test, a penetration and softening point pair that do not sit together sensibly, and a residue that ages far more than the grade should. Where the value at risk justifies it, asphaltene content by ASTM D6560 or a four-fraction separation by ASTM D4124 will show a composition that does not look like a paving grade — but neither is a specification line, and neither gives you a contractual right to reject unless you wrote it into the contract first.

If solubility is absent from a COA, that absence is itself a finding. Ask for it before shipment, not after.

Flash point — ASTM D92 / AASHTO T 48 / EN ISO 2592, and ASTM D3143 for cutbacks

The sample is heated at a controlled rate and a test flame is passed over the surface until the vapours ignite briefly. The Cleveland open cup, ASTM D92, is used for paving and oxidized grades. Cutbacks must be tested by the Tag open-cup method written for them, ASTM D3143, because they contain volatile solvent and flash far below a paving grade. ASTM D1310 is the general-purpose Tag open-cup method for liquids and appears on some documentation, but the designation cutback specifications actually name is D3143.

What it tells you, and what it does not. A flash point is an apparatus-defined number that sets the hard ceiling for storage and heating. It is the figure your insurer, terminal and tank farm operator will ask for, and it is what the heating regime on site has to be built around. It is not a performance property. It correlates with nothing about how the pavement will behave, and a flash point comfortably above the minimum tells you only that the material is safe to heat, not that it is any good. Its diagnostic value is entirely negative: a low flash point on a paving grade means volatile material is present that should not be.

Seeing ASTM D92 quoted on a cutback certificate is a genuine red flag, because it indicates the wrong method was used on a product whose safety case depends on getting this right. What a flash point permits in practice is covered on the heating temperature guide.

Water content — ASTM D95 / AASHTO T 55 / IS 1211

Water is distilled out of the sample with a solvent, condensed and measured volumetrically. Water in hot bitumen is dangerous rather than merely undesirable: it flashes to steam and can cause violent foaming and boil-over from a tank or a drum. It is also a quantity issue, because you are invoiced for it by weight. Note that EN 1428 is written for bituminous emulsions, where water is the continuous phase by design, and is not the counterpart method for paving grades.

Density and specific gravity — ASTM D70 / AASHTO T 228 / EN 15326 / IS 1202

Measured with a pycnometer. Paving bitumen typically falls between 1.01 and 1.06 g/cm³, which is common industry experience rather than a requirement of ASTM D946. This is the value that converts your cargo between tonnes and cubic metres, so it belongs on every certificate for a bulk shipment.

The reference temperature is not a detail. ASTM D70 is run at 15.6 °C or 25 °C, EN 15326 at 25 °C and IS 1202 at 27 °C. A specific gravity quoted at 27 °C is a slightly different number from the same material at 15.6 °C, and if you feed the wrong one into a volume-to-tonnage calculation the error lands directly on the invoice. Always take the reference temperature with the value. The conversion arithmetic, including correction of a hot loading volume to a base temperature, is on the tonnage and volume page.

Ash and inorganic residue — ASTM D482

The sample is burned and the inorganic residue weighed. For oxidized and industrial grades this is often more informative than solubility, because it directly quantifies the non-combustible content. It appears routinely on industrial bitumen and enamel specifications and rarely on paving grade certificates.

Spot test — AASHTO T 102

A drop of bitumen dissolved in a specified solvent is placed on filter paper. A uniform brown stain is reported negative; a dark centre ring is positive and indicates the bitumen has been over-cracked during processing. Many Middle East export specification sheets carry this line with a required result of negative.

Be clear about its status. The spot test is a historic method, it is subjective in the reading, and it does not appear in ASTM D946 or in EN 12591. It survives on export documentation as a legacy line rather than as a modern specification requirement. Where your contract carries it, a positive result is a rejection point and you can act on it. Where your contract does not, a missing spot test is not an omission worth chasing, and the ageing residue lines will tell you far more about the same underlying question.

Wax content — EN 12606-1 and EN 12606-2

Wax is a paraffinic fraction inherited from the crude, not created by the refinery, and it damages a pavement at both ends of the temperature range: it crystallises and embrittles the binder in winter and thins it in summer. The reason it deserves its own line is that neither penetration nor softening point can see it. A high-wax binder can return a textbook result on both and still be rejected on arrival.

Two EN methods exist and they measure different things. EN 12606-1 distils the sample and crystallises the wax out of the distillate. EN 12606-2 extracts it. They do not produce the same number on the same material, so a wax limit is only meaningful when the specification names the method. Chinese practice uses a distillation method under the JTG E20 test series and sets class limits under JTG F40; ASTM D946 and EN 12591 set no wax requirement at all, which is why an export certificate written to those standards normally carries no wax figure. The limits, the method and what they permit are set out on the wax content page.

A grade such as PG 64-22 states that the binder meets its high-temperature criteria at an average seven-day maximum pavement temperature of 64 °C and its low-temperature criteria at a minimum pavement temperature of −22 °C. Both halves are earned through testing, not assigned by inspection.

The four instruments and the criteria they produce

• Rotational viscometer — ASTM D4402 / AASHTO T 316. Maximum 3.0 Pa·s at 135 °C. A pumpability and workability requirement rather than a performance one, but it is the first thing to check on any PG claim, because a binder that cannot be pumped cannot be laid.
• Dynamic shear rheometer — ASTM D7175 / AASHTO T 315. A thin disc of binder is oscillated between two plates at a controlled temperature and frequency, reporting the complex shear modulus G* and the phase angle δ. The rutting parameter G*/sin δ must be at least 1.00 kPa on the binder as supplied and at least 2.20 kPa on RTFOT residue. The fatigue parameter G*·sin δ must be no more than 5000 kPa on PAV residue.
• Bending beam rheometer — ASTM D6648 / AASHTO T 313. A small beam of PAV-aged binder is loaded at low temperature and its deflection tracked. Two values are read at 60 seconds: creep stiffness S, maximum 300 MPa, and the m-value, minimum 0.300. Together they say the binder is neither too stiff nor too slow to relax stress when the pavement contracts.
• Direct tension test — ASTM D6723 / AASHTO T 314. An alternative low-temperature criterion, with a minimum failure strain of 1.0 %, used where the BBR creep stiffness is marginal rather than clearly passing.

Two things about this suite that are easy to miss

The BBR is not run at the grade temperature. It is run 10 °C warmer than the low grade temperature. That is deliberate: because of time-temperature superposition, a 60-second loading at 10 °C above the design temperature stands in for a much longer loading at the design temperature itself. A laboratory that runs the BBR at −22 °C for a PG 64-22 has not followed the specification, and the result will be conservative in a way that is not comparable with anyone else's.

Traffic is now measured rather than assumed. AASHTO M320 handles heavy or slow traffic by bumping the grade a step. The multiple stress creep recovery test, ASTM D7405 / AASHTO T 350, measures non-recoverable creep compliance directly, and AASHTO M332 grades a binder by traffic level from that measurement instead. Where a specification is written in M332 terms, a plain M320 report does not satisfy it. The full sequence, the test temperatures by grade and what changes for modified binders are on the performance grading guide; that material is not repeated here.

What this means for a buyer

You cannot convert a penetration grade into a PG grade on paper, and no correlation table can do it for you. A 60/70 binder from one source may test as PG 64-16 and from another as PG 64-22, because penetration at 25 °C carries no information about behaviour at −22 °C — the two temperatures are 47 degrees apart and the relationship between them is a property of the individual crude and refining route, not of the grade name.

So if a project specification is written in PG, ask for the actual DSR and BBR test report, with the material basis stated for each criterion. A supplier asserting a PG grade without that report is asserting something they have not measured. The converse is also true and less often said: a PG report does not tell you the penetration grade either, and a contract written in one system cannot be satisfied by a certificate written in the other.

Repeatability and reproducibility are two different numbers

Measurement is not exact, and standards say so explicitly. Two statistics describe how inexact a given method is.

• Repeatability (r) is the difference expected between two results obtained by the same operator, on the same apparatus, in the same laboratory, on the same material, within a short interval. It is the tightest case: everything is held constant except the measurement itself.
• Reproducibility (R) is the difference expected between two results obtained by different operators, on different apparatus, in different laboratories, on the same material. Everything that can legitimately vary between two competent laboratories is allowed to vary.

Both are stated as limits at 95 % probability: the difference between two properly obtained results should exceed the stated value in only about one case in twenty. The arithmetic behind them is standard across ASTM and ISO: r and R are approximately 2.8 times the corresponding standard deviation, the factor being 1.96 multiplied by the square root of two. The terminology and the calculation are set out in ASTM E177; the interlaboratory studies that produce the values follow ASTM E691, with ASTM C670 governing how a precision statement for a construction materials test is written. EN and ISO methods take the equivalent statistics from ISO 5725.

Reproducibility is always larger than repeatability, and for bitumen tests it is often several times larger. That single fact resolves most quality arguments in this trade. A difference between your laboratory and the supplier's laboratory that would be alarming if it appeared twice within one laboratory can be entirely normal between two.

Why the same material gives different numbers

None of the sources of variation below is misconduct. All of them are inside the tolerance a method allows.

• Sample preparation: heating history, how long the tin sat in the oven, whether it had been reheated before.
• Bath control: a bath a few tenths of a degree off changes a penetration reading, and the operator cannot see it happening.
• Thermometer and apparatus calibration, which produces a systematic offset that repeats on every sample a laboratory runs and therefore looks like consistency rather than error.
• Operator judgement: the moment the ring-and-ball specimen is judged to touch the plate, the point at which a ductility thread is called broken, the reading of a spot test stain.
• Specimen geometry: film depth in an ageing test, briquette trimming in a ductility test, plate gap in a DSR.

The decision rule for a buyer

When your result and the supplier's result differ, work through this before you write the word discrepancy into an email.

• Confirm you ran the same test. Same designation, same edition, same temperature, and above all the same material basis — as supplied, short-term residue or PAV residue. Most apparent disagreements are two different tests.
• Read the precision clause of that method. It is at the end of the standard and it gives r and R for the property, often varying with the level of the result and sometimes with the type of material.
• Compare the difference with R, not with the specification band. If the difference between two laboratories is within the published reproducibility, the two results are consistent with a single material and neither laboratory is wrong. Reacting to that as bad faith damages a relationship over nothing.
• If the difference exceeds R, you have a real difference — either in the material or in one laboratory's execution. That is when a third laboratory, or a retest on the sealed retained sample, becomes worth paying for.
• Use the published resolution practice. ASTM D3244 is the standard practice for using test data to determine conformance with specifications. It takes the supplier's result and the receiver's result, weights them using the precision statement of the method itself, and produces an assigned test value against which conformance is judged. It converts a stand-off into a calculation, and it is the clause to name in a contract before you need it. A counterparty who refuses to accept a published resolution practice is telling you what they expect a retest to show.
• Know how the limit is applied. ASTM E29 is the practice for using significant digits to determine conformance, and it defines two approaches. Under the rounding method, an observed value is rounded to the same number of decimal places as the limit before it is compared, so a solubility of 98.96 % rounds to 99.0 % and conforms to a 99.0 % minimum. Under the absolute method, it does not. The specification is supposed to state which applies. If yours does not, that is a sentence worth adding, because it decides marginal cases.

This page deliberately prints no repeatability or reproducibility values. They differ by test, by the level of the result, by the type of binder and by the edition of the standard, and a figure copied from an older revision is worse than no figure at all. Read them from the current edition of the method you are actually working to. That is the only place they are authoritative.

Which tests are worth commissioning independently, and which are not

Independent testing is cheap relative to a container load and decisive in a way that no amount of document analysis can be. It is not, however, uniformly worth doing, and commissioning the wrong tests wastes money and time without improving your position.

Commission on any first order, and on any offer priced materially below the market: penetration, softening point, solubility, flash point and density. All five are fast, all five are available in any competent petroleum testing laboratory rather than a specialist binder laboratory, and between them they catch grade substitution, mineral extension and solvent contamination — which is most of what actually goes wrong. This is the set with the best ratio of information to cost, and it is the set to write into a pre-shipment inspection scope as standard.

Commission where the value or the design justifies it: the short-term ageing procedure with residue penetration and residue ductility, and absolute viscosity at 60 °C where the cargo is a viscosity grade. These need a laboratory with the right oven and viscometers, and the ageing test alone occupies most of a working day before anything is measured. They are worth it because retained penetration is the best single predictor of premature cracking, and because it is the line most often absent from a weak certificate.

Commission only when the specification is written in performance grade terms, or when a PG claim is being made: DSR, PAV and BBR. These require an accredited binder laboratory, the pressure ageing vessel alone runs for 20 hours before any measurement begins, and a full grading sequence on one sample is a multi-day job. Do not commission them to "check" a penetration-graded cargo — the results will not map onto the specification you are buying against, and you will have paid for information you cannot act on.

Rarely worth commissioning independently: ductility, because it is the most preparation-sensitive routine test and a disagreement is as likely to be the briquette as the binder; the spot test, because it is subjective and absent from ASTM D946 and EN 12591; wax content, unless the buying specification carries the line and names the method; and asphaltene or four-fraction composition work, which is informative to a chemist and gives you no contractual right to act unless you wrote one.

Two rules govern all of this. Do not commission a test whose result you have no contractual right to act on — agree the specification, the laboratory, the retest right and the resolution practice first, and commission the test second. And test the right sample: a result from a sample your supplier selected and sent you is a result about a sample your supplier selected. The test that counts is on a sealed sample drawn at the load point by an inspector you appointed, with a retained portion held by both parties.

• Avoiding fraud in bitumen trade — the penetration index cross-check worked through with numbers, the signatures of a fabricated certificate, and where ASTM D3244 sits in a claim
• Adhesion and anti-stripping — the one property no certificate can predict, because it belongs to the binder and the aggregate together
• Wax content and the Chinese specifications — a limit that decides acceptance in one market and does not exist in ASTM D946 or EN 12591 at all
• Heating temperature guide — what a flash point actually permits in a tank, and the temperatures at which ageing starts to run away
• Tonnage and volume conversions — why the reference temperature on a density line lands directly on the invoice
• Bitumen 60/70 and Bitumen VG-30 — the same tests worked through as complete specification tables on the two most traded grades