A specification limit does not decide by itself whether a result passes. ASTM E29-22 defines two conformance methods — absolute and rounding — and ASTM’s own worked example shows a result of 1.8 % against a 2 % minimum elongation specification failing under the absolute method and passing under the rounding method. E29 only governs if the specification explicitly cites it and states which method applies and, for rounding, the rounding interval. Most activated carbon datasheets cite nothing of the kind. That is why a buyer and a seller can look at the same certificate, do no arithmetic wrong, and reach opposite verdicts.
This article works through what each line commits a supplier to — minima and maxima, tolerances, typical values, units, capacity indices, hardness, density — and ends with what happened when we ran the same test across all 46 of our own published sheets.
The short version
- ASTM E29-22 gives two conformance methods. A 1.8 % result against a 2 % minimum fails the absolute method and passes the rounding method at a 1 % rounding interval. E29 binds only when the specification cites it.
- No standards body defines “typical value” or “guaranteed value.” The boilerplate about typical values not being specification limits is industry practice, not a standardised definition. Get the contractual lines named in writing.
- 1 g/cm³ = 1 g/mL = 1000 g/L = 1000 kg/m³ (NIST SP 811). So 0.5 g/cc and 500 g/L are the same number, and “0.5 g/L” for a granular carbon is physically impossible.
- Every capacity-type ASTM method in this family carries the same disclaimer — the value is a quality-control and lot-comparability index, not a service prediction. D3467, D4607, D5742, D3802 and D6646 all say a version of it.
- ASTM D3802 states that ball-pan hardness does not measure in-service resistance to degradation. It establishes lot comparability, and most specifications sit at 97–98 on a 100 scale.
Eight specification lines, annotated. Three are defects of a kind that appear on datasheets across the industry — a capacity figure with no method, an unreadable tolerance, and a regulatory-sounding claim with no protocol.
What does a “minimum” on a datasheet actually commit a supplier to?
A minimum commits the supplier to a floor — but only as far as the stated test method, the stated conditions and the conformance rule take it. Three things must be present before a minimum is enforceable in any practical sense:
- A named test method with its designation and revision. “Iodine number min 1000 mg/g” is weaker than “Iodine number, ASTM D4607-14(2021), min 1000 mg/g.”
- The conditions, where the method allows variants. ASTM D2867 for moisture offers three procedures — oven drying as the reference method, a moisture balance, and xylene extraction. Three procedures, one line on a datasheet.
- A conformance rule. This is ASTM E29’s territory, and almost nobody specifies it.
The E29 problem, in full
ASTM E29-22, Standard Practice for Using Significant Digits in Test Data to Determine Conformance with Specifications, exists to give committees a uniform way of indicating which digits are significant when setting specified maximum and minimum values. It offers two methods:
| Method | How conformance is judged | Effect |
|---|---|---|
| Absolute | The result as measured must fall within the stated limits | The limit is the limit |
| Rounding | The result is first rounded to the decimal place of the limit, then compared | Effectively widens the limit by half a rounding interval |
ASTM’s published worked example uses a 2 % minimum elongation specification. A result of 1.8 % fails the absolute method. Rounded to the nearest whole percent at a 1 % rounding interval, 1.8 becomes 2, and it passes. One number, one specification, two defensible verdicts.
The trap is that E29 applies only if the specification explicitly invokes it and says which method governs and, for rounding, what the interval is. Absent that, a purchaser applying the absolute method and a supplier applying the rounding method are both behaving reasonably and will disagree. If a number matters commercially — a mercury loading, a breakthrough time, an ash ceiling driving a downstream catalyst warranty — the purchase order should say which conformance method applies. One sentence removes a whole category of dispute.
What is a “typical value” and what is a “guaranteed value”?
They are industry usage, not defined terms. No standards body publishes a definition of either. The familiar boilerplate — that typical values are for information and not to be construed as specification limits — is a convention suppliers adopted, not a standardised meaning you can cite. We will not assert one, because none exists to assert.
What is citable is a standards body acknowledging what single-point datasheet values are for. ISO 10350-1, the plastics standard for comparable single-point data, exists because users find that available data cannot always be readily compared across materials, especially when it comes from different sources. Its stated purpose is to present single-point data for the “preliminary selection of materials.” That is a committee saying on the record that datasheet numbers are a selection and comparison tool, not an acceptance criterion.
Working definitions that hold up commercially:
| Datasheet form | What it usually means | What you can hold the supplier to |
|---|---|---|
| min / max with named method | An acceptance limit | The limit, subject to the conformance rule |
| value ± tolerance (e.g. 450 ± 30 g/L) | A target band around a production centre | The band, if the method is named |
| Bare number, no qualifier (e.g. “1000 m²/g”) | A nominal or historical figure | Very little — ask whether it is min, typical, or a range |
| “Typical” | Historical central tendency | Nothing, unless the contract says otherwise |
| “Guaranteed” | A commercial promise | Whatever the contract says — the word has no standard meaning |
| Range (e.g. 1000–1500 m²/g) | Both ends are limits | Both ends, if the method is named |
The practical move is to ask the supplier, in writing, which lines are contractual acceptance criteria and which are informational, and to have that answer travel with the purchase order. A supplier that publishes its specifications openly should have no difficulty answering.
Units: 0.5 g/cc and 500 g/L are the same number
They are. NIST Special Publication 811, the US guide to the SI, gives the identity directly: 1 g/cm³ = 1 g/mL = 1000 g/L = 1000 kg/m³. So a carbon at 0.50 g/cc is at 500 g/L is at 500 kg/m³. Nothing has changed but the dress.
This matters because carbon datasheets mix all four forms, sometimes within one product family. Ours does: SURSORB PAC states apparent density in g/L, SURSORB PAG in kg/m³, while SURSORB MAG states bulk density in g/mL and SURSORB SA in g/cc — the same physical quantity in four notations. A buyer building a comparison spreadsheet will normalise them or produce nonsense.
Two consequences worth stating plainly:
- “0.5 g/L” is physically impossible for a granular activated carbon. It is three orders of magnitude out. So is “500 g/cc.” If you see either, it is a typographical error, and the sheet has not been checked.
- A dimensionally incoherent tolerance is a defect, not a style. A bulk density written as
0.600±50 g/mlcannot be read: 0.600 g/mL plus or minus 50 of what? If the intent was ± 0.050 g/mL, the sheet should say so.
For scale, ANSI/AWWA B604 reports apparent density in both lb/ft³ and g/cc, putting bituminous granular carbon at 27–41 lb/ft³, equal to 0.43–0.66 g/cc; multiply g/cm³ by 62.43 to get lb/ft³. Scholz’s 2023 review puts activated carbon apparent density generally at 0.6–0.8 g/cm³. Any number far outside that neighbourhood deserves a question before a purchase order.
Apparent density: what ASTM D2854 says it is for
ASTM D2854-09(2025) covers apparent density of activated carbon — granular material, minimum 90 % larger than 80 mesh. Its significance statement is unusually blunt about why the number exists: determination of packed density “is essential when designing vessels to hold the material” and for ordering purposes when procuring material.
That is the honest use of an apparent density line: it tells you how many kilograms fill a vessel, and therefore what a change-out weighs and costs. It is not a quality indicator. As covered in what CTC activity actually measures, density typically falls as activation increases — across SURSORB PAC-900, PAC-1050 and PAC-1250 it falls from 480 ± 30 to 380 ± 30 g/L while CTC activity rises from minimum 50 % to minimum 75 %. Lower density there means more pore volume, not a worse carbon.
The procedure matters more than people expect. Practitioner guidance describes carbon delivered slowly into an appropriately sized graduated cylinder through a funnel at 0.75–1.0 mL/s, and notes that twelve particles or more across the bed diameter avoids cylinder-wall effects. Fill faster, or use too narrow a cylinder, and the number moves. If two laboratories disagree on density, check the fill rate first.
Ball-pan hardness: what ASTM D3802 does and does not claim
ASTM D3802-23 measures ball-pan hardness of activated carbon, for granular carbons 90 % of which are larger than 80 mesh (180 µm). The test set is fifteen ½ in (12.7 mm) steel balls and fifteen ⅜ in (9.5 mm) steel balls in a hardened brass pan of about 8 in (204 mm), run 30 minutes in a Ro-Tap sieve shaker, then sieved. The hardness number is:
H = 100 × B / A, where A is the mass loaded before shaking and B is the mass retained on the hardness test sieve.
The scale ends at 100, and most specifications call for 97 to 98; Scholz’s review notes values are typically above 95. Our own range sits in that band — SURSORB Ultra at ball pan hardness minimum 99, SURSORB SA at minimum 98, SURSORB PAC-1250 at minimum 96, SURSORB PAG at minimum 85.
Now the part that gets over-read. ASTM states that the method “does not actually measure in-service resistance to degradation” and that it can be used to establish “the comparability of lots ostensibly of the same grade.” That is the entire claim. A hardness of 98 rather than 96 does not entitle anyone to predict fines generation, head loss development or bed life in a specific vessel; we have found no citable basis for that extrapolation.
There is a further wrinkle that matters for water-treatment purchasing: AWWA B604 does not apply D3802 universally. It prescribes the Ro-Tap test for bituminous- and coconut-based carbons and the stirring abrasion test for lignite- and wood-based carbons. So two carbons quoting “abrasion resistance” under the same AWWA standard may have been tested by two different methods, depending on precursor. Comparing those numbers across precursors is not comparing like with like.
The disclaimer that every capacity method carries
This is the most useful single fact in the whole specification business, and it is fully verifiable from public ASTM material: every capacity-type test method in this family states that its result is a quality-control or comparability index, not a prediction of service performance.
| Method | Designation | What it measures | ASTM’s stated limitation |
|---|---|---|---|
| CTC activity | D3467-04(2020) | Mass gain at CCl4 saturation | Does not necessarily measure effectiveness on other adsorbates or at other conditions |
| Iodine number | D4607-14(2021) | mg iodine/g at 0.02 N residual | Does not necessarily indicate capacity for other species |
| Butane activity | D5742-16(2023) | Mass gain ratio for butane | Positioned as a QC alternative to the CCl4 test |
| Ball-pan hardness | D3802-23 | Attrition under steel balls | Does not measure in-service resistance to degradation |
| H2S breakthrough | D6646-03(2022) | Accelerated breakthrough capacity | Does not duplicate conditions an adsorber meets in practical service |
D6646 is the sharpest of them, because it is most often quoted as if it were a service capacity. The word “Accelerated” is in the official title. The method passes a humidified air stream containing 1 % by volume H2S — 10,000 ppmv — through a bed until 50 ppm breakthrough at the outlet, at a superficial velocity of 4.77 cm/s and a contact time of 4.8 s, which puts bed length at about 23 cm. The result is reported as g H2S per cm³ of carbon at 99.5 % removal efficiency. ASTM then says, in effect, do not read that as service life: the laboratory setup favours carbons with faster removal kinetics, and the mass transfer zone in the 23 cm column is proportionally much larger than in a typical industrial bed. A number generated at 10,000 ppmv is not a number for a 200 ppmv digester gas stream.
Moisture and ash: small lines, real consequences
ASTM D2867-23 covers moisture. Three procedures: oven drying (the reference method), moisture balance, and xylene extraction. The xylene route exists because certain activated carbons can ignite at temperatures as low as 150 °C — a real hazard when someone dries a sample harder to get a better number. ASTM also notes that moisture in carbon packed in typical shipping containers will usually increase during transport and storage. So a moisture result at the receiving dock is not a challenge to the mill’s certificate; it is expected, and the specification should say where the sample is drawn.
Moisture is also the quietest commercial line on a sheet. If you buy by mass at 5 % maximum moisture and it lands at 5 %, you have bought 5 % water. Across a 20-tonne order that is a tonne.
ASTM D2866-11(2025) covers total ash content: the reference method is a muffle furnace at 650 °C ± 25 °C, with thermogravimetric analysis permitted as an alternative and the muffle furnace as referee. ASTM’s significance statement notes that the amount and composition of the ash may influence the capabilities and certain desired properties of the carbon. High ash is not automatically bad — an inorganic impregnant raises it by construction. SURSORB MAG, a magnesium-oxide impregnated grade, is specified at ash maximum 15 %, which is consistent with the impregnant rather than with a dirty carbon. Reading a 15 % ash line as a defect without asking why it is 15 % is a common misread.
Surface area: name the standard, and know its limits
If a datasheet says “surface area (BET) 1000 m²/g,” ask which standard. ISO 9277:2022 is the BET standard for gas adsorption. ASTM D6556 is frequently miscited as the ASTM BET method for activated carbon — it is not; it is a carbon black method, Volume 09.01, Subcommittee D24.21, outside the D28 activated carbon committee entirely.
ISO 9277 states that BET theory is applicable to Type II and Type IV isotherms, handling microporous Type I solids — which is what activated carbon is — as a special case in an annex. The IUPAC 2015 technical report is blunter: for a Type I isotherm the BET area “represents an apparent surface area,” best regarded as a useful adsorbent “fingerprint.” Treat it as an identity check between lots, not as an accessible area.
The questions to send back with an RFQ
For each line that matters commercially:
- Which test method, at which designation and revision, produced this number?
- Is this line an acceptance limit or informational?
- Which ASTM E29 conformance method applies, and at what rounding interval?
- For capacity figures: inlet concentration, temperature, humidity, superficial velocity, contact time and breakthrough endpoint.
- For density: which unit, and which method?
- Will the certificate of analysis carry the measured result, or the word “conforms”?
The last one is decisive and is covered in what a certificate of analysis should contain. A datasheet is a selection document. A certificate of analysis is the document that tells you what you actually received.
What happened when we ran this on our own 46 sheets
We audited all 46 published Suracsh specifications against the six questions above. Six lines across five grades did not meet them, and all six were sent back to the issuing factory for correction. The audit and its outcome are set out below, because a checklist a supplier is unwilling to run on itself is marketing rather than guidance.
Four were clerical: a transcription error in the moisture and ash lines on SURSORB SA, one in the hardness line on SURSORB Neo, and a malformed density tolerance carried through from source data in the form 0.600±50 g/ml — the unreadable construction set out earlier in this article. A fifth was a qualifier conflict on the surface area of SURSORB BAL and the alumina fraction of SURSORB BLEND, where the sheet was inconsistent about whether 300 m²/g was a minimum or a nominal value. As the table above shows, those are not the same commitment, so the sheet now has to pick one.
The two that mattered most were not errors at all. They were numbers that could not be defended under this article’s own standard, which is a harder thing to act on.
SURSORB CAG carried an H₂S adsorption figure of 0.12 g/cc with no test method and no conditions. By the argument set out here that number is uninterpretable — without the inlet concentration, humidity, superficial velocity and breakthrough endpoint behind it, it cannot be compared with any other supplier’s H₂S figure and cannot be used to size a bed. A figure that cannot be interpreted should not be on a specification sheet, so it is being restated against ASTM D6646 with its conditions attached, or withdrawn. There is no third option that survives the question “measured how?”
SURSORB NGC carried a methyl iodide removal efficiency of min 99.0 % with no protocol. For a nuclear-grade media that is the wrong way round: the protocol is the specification. The sheet is being reissued with the qualification protocol named, and in the meantime the product page carries an explicit instruction to request it before specifying against a regulatory requirement.
Correction log
| Line | Grade | Correction required | Reissued sheet |
|---|---|---|---|
| Moisture, ash | SURSORB SA | Correct value and qualifier | Pending |
| Ball pan hardness | SURSORB Neo | Correct value, confirm the test method | Pending |
| Bulk density tolerance format | Multiple | Tolerance stated in the same unit as the value | Pending |
| Surface area qualifier | SURSORB BAL, BLEND | Resolve to min 300 or 300 typical | Pending |
| H₂S adsorption | SURSORB CAG | Restate against ASTM D6646 with conditions, or withdraw | Pending |
| Methyl iodide efficiency | SURSORB NGC | Name the qualification protocol | Pending |
Pending means the correction has been raised and the reissued sheet is not yet in hand. This table is updated as each one lands, so the gap between finding a defect and fixing it is visible too.
Two things follow from publishing every specification openly rather than behind a request form. The first is that anyone can check us against the standard we have just set out. The second is that when something is wrong, it is visible, and the only available response is to fix it and say so. We would rather have that constraint than the alternative.
The full specification for every grade named here is published in HTML at suracsh.us, and every datasheet downloads without a form or an NDA.
Sources
- ASTM E29-22 — Using Significant Digits in Test Data to Determine Conformance with Specifications
- ASTM — Absolute or Rounding: Conformity to Specification
- ISO 10350-1 — Plastics: acquisition and presentation of comparable single-point data (preview)
- NIST Special Publication 811 — Guide for the Use of the International System of Units
- ASTM D2854 — Apparent Density of Activated Carbon
- ASTM D3802 — Ball-Pan Hardness of Activated Carbon
- Gilson — ASTM D3802 ball-pan hardness test set
- Testing Tips for Activated Carbon Users, Water Conditioning & Purification, 2013
- ANSI/AWWA B604-18 — Granular Activated Carbon, preview
- ANSI/AWWA B604-12 — Granular Activated Carbon, preview
- ASTM D3467 — Carbon Tetrachloride Activity of Activated Carbon
- ASTM D4607 — Iodine Number of Activated Carbon
- ASTM D5742 — Butane Activity of Activated Carbon
- ASTM D6646 — Accelerated Hydrogen Sulfide Breakthrough Capacity of Activated Carbon
- ASTM D6556 — Carbon Black: Total and External Surface Area by Nitrogen Adsorption
- Scholz, Technologies 2023, 11(6), 153
- ASTM D2867 — Moisture in Activated Carbon
- ASTM D2866 — Total Ash Content of Activated Carbon
- ISO 9277:2022 — Specific surface area of solids by gas adsorption, BET method
- Thommes et al., Physisorption of gases (IUPAC Technical Report), Pure Appl. Chem. 2015
Questions
What is the difference between a minimum, a typical and a guaranteed value on a carbon datasheet?
A minimum or maximum with a named test method is an acceptance limit the supplier can be held to. A typical value is a historical central tendency with no commitment attached. No standards body defines typical or guaranteed — those are industry usage, so ask the supplier in writing which lines are contractual.
Does 0.5 g/cc mean the same thing as 500 g/L?
Yes. Per NIST SP 811, 1 g/cm3 equals 1 g/mL equals 1000 g/L equals 1000 kg/m3. So 0.5 g/cc, 0.5 g/mL, 500 g/L and 500 kg/m3 are one number in four dresses. A granular activated carbon quoted at 0.5 g/L would be physically impossible.
What does ball pan hardness measure?
ASTM D3802 tumbles a carbon sample with fifteen half-inch and fifteen three-eighths-inch steel balls in a brass pan for 30 minutes in a Ro-Tap shaker, then sieves it. Hardness is 100 times the mass retained divided by the mass loaded. ASTM states it does not measure in-service resistance to degradation.
Can a buyer and a seller both be right about whether a result passes?
Yes. ASTM E29-22 defines an absolute method and a rounding method for judging conformance. Its own example: against a 2 percent minimum, a result of 1.8 percent fails the absolute method and passes the rounding method. E29 only governs if the specification cites it and says which method applies.