By the YELI Technical Team · Updated August 2026 · 7 min read
Short answer: To detect blended or adulterated activated carbon, test the delivered lot, not the sample. Twelve checks separate genuine coconut shell gold carbon from blended or under-spec material: hardness (96–99% genuine, 90–94% for palm blends, ASTM D3802), platelet content <1.5%, apparent density 0.45–0.52 g/cm³, ash, optional third-party K/R validation where required, moisture against the declared as-packed basis, mesh distribution, batch-to-batch drift, regeneration-fines trend, documentation, and your right to verify. The definitive check is a side-by-side lab test of the delivered lot against the approved sample.
The oldest trick in the activated carbon trade: send a perfect sample, then ship a cheaper blended lot. Coconut shell carbon is the premium feedstock, so some suppliers cut it with palm shell, apricot shell or coal-based material to lower cost — while keeping the paperwork identical. If the carbon in your circuit suddenly produces more fines, or recovery drifts, blended feedstock should be on your checklist. Here is how to catch it.
Why Blending Is Hard to Catch on Paper
A supplier can label the lot "coconut shell, iodine 1050" and the COA may even be technically true for the blended average. The problem is performance: palm shell and other feedstocks are softer, so they generate more fines in your circuit — and fines carry gold out. The COA won't tell you this; only testing the actual lot will.
Check 1: Hardness (The Most Telling Test)
Pure coconut shell carbon typically tests 95%+ hardness. Palm shell, apricot shell and wood-based material are softer. Run an abrasion test on a sample from the delivered lot. If hardness comes back below the COA, or significantly below 95%, you have strong evidence of blending — because hardness is the first thing blending degrades.
Hardness is measured by the ball-pan method (ASTM D3802): a 100 g sample is tumbled with steel balls for a fixed time, and the fraction retained on the reference sieve is reported as percent hardness. The numbers separate feedstocks cleanly. Coconut-shell gold-grade carbon typically tests 96–99%, while palm-shell blends typically sit at 90–94%. On a lot labeled pure coconut shell, a result below 95% is strong evidence of blending — and hardness is the first test to flag it.
Check 2: Iodine vs. Actual Performance
Blended carbon can still show a decent iodine number — iodine measures surface area, which some blends preserve. But real gold loading depends on pore structure and hardness. If the delivered lot has a good iodine number yet loads gold poorly or produces fines, the pore structure is wrong for the job, which points to non-coconut feedstock.
Check 3: Ash and Apparent Density
Different feedstocks carry different ash profiles. Coconut shell is naturally low-ash; coal-based carbon is higher; palm shell sits between. Compare the delivered lot's ash and density against the COA and against a known coconut shell reference. A jump in ash is a red flag.
Check 4: Visual and Sieve Inspection
Blended lots often show visible differences: inconsistent particle color, shape or gloss, and a wider mesh distribution than a single feedstock would give. Spread a sample on a white tray — pure coconut shell looks relatively uniform; a blend shows mixed particle types. Run a sieve analysis to check the distribution matches the certified cut.
The 12 Checks at a Glance
Checks 1–4 above catch most blended lots. These are the twelve checks we run in full — with the method, the value a genuine lot shows, and what missing it costs you.
| # | Check | Genuine coconut shell gold carbon | Blended / under-spec material | How to verify | What it costs you if missed |
|---|---|---|---|---|---|
| 1 | Hardness (ASTM D3802, ball-pan) | 96–99% | 90–94% (palm/apricot blends), lower for coal | Abrasion test on the delivered lot | More fines → adsorbed gold leaves with the tailings |
| 2 | Platelet content | <1.5% | Unactivated shell wall left in; visibly higher | Microscope, or a supplier test report that states its method | Platelets fracture in the train → gold loss |
| 3 | Apparent density (ASTM D2854 / GB/T 12496.1) | 0.45–0.52 g/cm³ | Higher = coal or mineral filler; lower = under-activated or wood | Graduated cylinder — doable on site | You pay for weight, not for adsorption capacity |
| 4 | Ash (ASTM D2866, 650 °C) | Use the declared grade limit and dry/as-packed basis; there is no universal ash cutoff | Higher — coal, mineral matter, re-activated carbon build-up | Muffle furnace, or an outside lab | Pores blocked, surface turns hydrophilic, loading drops |
| 5 | Optional third-party K/R check | Only if lab performance is required; independent laboratory, GB/T 32992—2016 | Good iodine, poor gold loading — wrong pore geometry | Third-party K-value test | A third-party report can check gold-specific performance, but is not part of the standard COA |
| 6 | Regeneration-fines trend | Follow the agreed thermal-cycle protocol and compare against the approved sample | Lower — carbon never fills within one hour of contact | Same standard, same laboratory | Gold pushed to the next tank and out to the tails |
| 7 | Moisture as packed (ASTM D2867) | Match the grade-specific limit and as-packed basis on the COA; there is no universal 5% authenticity cutoff | A result above the agreed as-packed limit, or a basis mismatch, needs correction | Dry at 110–115 °C, weigh before and after | Direct money loss, plus every other test is skewed |
| 8 | Mesh distribution (ASTM D2862) | Certified 6×12 cut with a stated tolerance | Wider distribution, over- and under-size fractions | Standard sieve stack | Oversize passes screens, undersize leaves with the pulp |
| 9 | Batch-to-batch drift | Three consecutive lots inside a stated band | Same spec sheet, different measured behaviour | Retained samples, one laboratory, side by side | Production swings you cannot attribute to the ore |
| 10 | Regeneration life | 5–8 thermal cycles per charge | Collapses in 2–3 cycles | Track hardness and iodine after each cycle | Change-out frequency silently doubles |
| 11 | Documentation | Batch COA with batch number, named laboratory, test method | One COA reused across shipments, no batch number, no method | Match the COA batch and date against the container | No traceability when it goes wrong |
| 12 | Right to verify | Welcomes pre-shipment inspection and arrival re-test; keeps retained samples | Refuses inspection, no retained sample, no split sample | Put it in the contract | You carry all the risk with no remedy |
The Definitive Check: Side-by-Side Lab Test
If you suspect blending, do not argue on paper. Take a sample from the delivered lot and a reference sample of known pure coconut shell carbon, and run them together through the full battery:
| Test | Method | What it catches |
|---|---|---|
| Sieve analysis | ASTM D2862 | Mesh cut matches the certified 6×12 distribution |
| Moisture | ASTM D2867 (dried at 110–115 °C) | Removes the variable that skews every other result |
| Iodine number | ASTM D4607 (mg/g) | Surface area — blends can still score high here |
| Hardness | ASTM D3802 (ball-pan) | The feedstock fingerprint — coconut vs. palm band |
| Ash & apparent density | ASTM D2866 (ash at 650 °C) | Feedstock fingerprints that blending shifts |
Then the definitive gold test: a bottle-roll on your own ore and solution — same pulp density, same contact time — fire-assay both carbons and compare gold loading side by side. The numbers side by side tell the truth quickly. One rule: a COA must state the test method — a number without a method cannot be compared. Any confident supplier welcomes this test — and we do.
Verify Your Carbon Against Ours
YELI Carbon produces pure coconut shell activated carbon in Longyan, Fujian — Philippine and Indonesian coconut shell feedstock, hardness 98%+ (98% available), iodine 1000-1300, COA with every batch. Send us a sample of what you are using now and we will run iodine, hardness and ash against ours, side by side.
Run a Comparison Test →Why "Cheap" Carbon Is Cheap — Three Legitimate Reasons, Five That Are Not
A lower price is not automatically a red flag. Scale, off-season feedstock and honestly labelled coal-based grades all cost less for legitimate reasons. What matters is why the number is low — so ask, and make the answer part of the deal.
| Reason for a lower price | Legitimate? | Typical discount | What you are actually getting |
|---|---|---|---|
| Larger production scale, off-season shell purchase | ✅ | 5–15% | Same material, thinner margin |
| Coal-based carbon (a different product, correctly labelled) | ✅ | 30–50% | Not a coconut product — read the label, then check the fit |
| Stock lot, older batch, shorter remaining shelf life | ✅ if disclosed | 10–20% | Usable, but plan for a faster change-out |
| Blending palm or apricot shell into a "coconut" lot | ❌ | 20–35% | Softer carbon, more fines, gold out with the tails |
| Coal blended into a "coconut" lot | ❌ | 25–40% | Lower hardness, higher ash, wrong pore structure |
| Selling regenerated carbon as virgin | ❌ | 30–50% | High ash, low iodine, unpredictable kinetics |
| Shipping wet (moisture above spec) | ❌ | 5–12% — and it is a price rise | You pay carbon price for water |
| Oversize / undersize substitution against the certified cut | ❌ | 5–15% | Screen losses and pulp carry-over |
| No third-party testing, no retained sample, no batch COA | ❌ | 3–10% | You are the quality department, with no recourse |
The Maths of a Discount
A worked model, not a quote. Assumptions: head grade 2.0 g/t; gold USD 2,600/oz; carbon make-up 60 g per tonne of ore; carbon USD 2,400/t.
| Scenario | Per tonne of ore |
|---|---|
| Carbon at $2,400/t, 60 g/t make-up | $0.14 |
| Same carbon bought $400/t cheaper | $0.12 — you saved 2.4 US cents per tonne of ore |
| Losing 1% of recovery (0.02 g/t at 2.0 g/t head grade) | $1.67 |
The whole argument is in two lines: a USD 400 per tonne discount saves about 2.4 cents per tonne of ore. One percentage point of recovery costs about 1.67 dollars. That is a gap of roughly 70× — so price-led sourcing optimises the smallest line in the cost sheet and gambles the largest one.
First 24 Hours with a Delivered Lot
- Photograph the container seal, the bags, and the batch number printed on the COA.
- Pull a composite sample across at least 10 bags — never from the top layer alone.
- Weigh and dry a moisture sample, and record the result as packed.
- Run a sieve check against the certified cut (ASTM D2862).
- Run a density check in a graduated cylinder — site-doable, and it catches coal filler fast.
- Split the composite: one part retained on site, one part to an independent laboratory for hardness, iodine, ash, K and R.
- Label the retained sample with batch number and date. Every future dispute rests on it.
Contract Clauses That Do the Work
- Batch COA with batch number, named laboratory and test method, per shipment
- Retained samples held by both parties for a defined period
- Right to inspect pre-shipment (SGS / Intertek / BV) and to re-test on arrival
- Defined rejection thresholds: hardness, moisture, mesh distribution, ash
- Remedy named in advance — replacement, price adjustment or return, and who pays freight
The Bottom Line
Blended carbon is hard to catch on a COA but easy to catch in a lab. Test hardness, ash and sieve distribution on the delivered lot, and run a side-by-side comparison when in doubt. It takes a day and saves months of hidden gold loss. For the full supplier verification process, see our verification guide.
FAQ: Detecting Adulterated Activated Carbon
Test the delivered lot, not the sample. Hardness is the most telling: pure coconut shell tests 95%+, while palm shell and other feedstocks are softer, so an abrasion test exposes blending. Also check ash (different feedstocks carry different ash profiles) and sieve distribution, plus visual inspection for inconsistent particle types.
Palm shell and other non-coconut feedstocks are softer, so they generate more fines in your CIP/CIL circuit — and fines carry adsorbed gold out of the circuit. A blended lot can still show a decent iodine number while performing poorly, because pore structure and hardness are wrong for gold duty.
A COA describes what the supplier claims. The definitive check is testing the actual delivered lot: hardness, ash, sieve distribution, and a side-by-side comparison against a known pure coconut shell reference. A confident supplier welcomes this.
Order a small trial lot, test it independently (iodine, hardness, ash), and run a side-by-side comparison with a reference. Also verify the factory with a live video tour or visit. These three checks catch most quality and blending issues before you commit to a container.
Start with hardness (ASTM D3802): coconut-shell gold-grade carbon typically tests 96–99%, while palm-shell blends typically sit at 90–94%. Then check ash and apparent density, which shift when a second feedstock is blended in. The definitive check is a side-by-side gold loading test on your own ore and solution, with both carbons fire-assayed after the same contact time.
Legitimate reasons exist: larger production scale, off-season shell purchase, honestly labelled coal-based grades, or a disclosed stock lot. The illegitimate ones are blending (palm or apricot shell, or coal) into a lot labelled coconut, selling regenerated carbon as virgin, shipping above spec moisture, substituting mesh sizes, and skipping third-party testing. Ask which one applies — a supplier who cannot answer has answered.
Yes. Coal-based carbon is a different product and can be 30–50% cheaper when it is correctly labelled; the question is whether it suits your circuit. Scale and off-season feedstock can take 5–15% off coconut shell grades. Stock lots can save 10–20% if the age is disclosed. What is never legitimate is a coconut-shell price that only works because part of the lot is not coconut shell.
Gold-grade coconut shell carbon typically tests 96–99% by the ball-pan method (ASTM D3802), and we ship at 98%+. Palm-shell blends typically land at 90–94%. On a lot labelled pure coconut shell, a delivered-lot result below 95% is strong evidence of blending, because hardness is the first property blending degrades.
Directly. Carbon is sold by weight, so moisture above spec means you are buying water at carbon prices — a lot at 12% moisture against a 5% specification carries roughly 7% of its weight as water. It also skews every other test, which is why moisture is measured first and reported as packed. Dry at 110–115 °C (ASTM D2867) and weigh before and after.
You can, and the contract should say so before the container ships. Standard practice: pre-shipment inspection by SGS, Intertek or Bureau Veritas at the supplier's cost, retained samples held by both parties, and arrival re-testing at an agreed laboratory. Define in advance which results trigger rejection or a price adjustment, and who pays return freight.
Five clauses do most of the work: a batch COA per shipment stating batch number, named laboratory and test method; retained samples held by both parties; the right to inspect pre-shipment and re-test on arrival; defined rejection thresholds for hardness, moisture, mesh distribution and ash; and a spelled-out remedy (replacement, price adjustment or return, with freight responsibility named).
Yes. Send us a sample of what you are using now and we will run iodine, hardness and ash against ours and send you both sets of numbers — including the ones that do not favour us. We are a coconut shell manufacturer in Longyan, Fujian, so we would rather be measured than trusted.