By the YELI Technical Team · Updated August 2026 · 6 min read
Ask a gold plant manager what spec they check first on a carbon datasheet and most will say iodine value. Ask them what spec has cost them the most money over a year, and a surprising number will say hardness. It is the quiet spec — the one that does not appear in marketing numbers but shows up in your make-up carbon purchases month after month.
What Carbon Hardness Actually Measures
Hardness (or abrasion resistance) is measured by tumbling a sample of carbon in a standardized test (ASTM D3802) and measuring how much survives as intact granules versus how much breaks into fines. It is expressed as a percentage: a 97% hardness carbon loses about 3% of its mass to fines in the test. The version most labs run is the ball-pan method — a 100 g sample is tumbled with steel balls for a fixed time, and the fraction retained on the reference sieve is reported as the hardness percentage. Coconut shell gold-grade carbon typically tests 96–99%, while many coal-based grades sit at 90–95%.
In a gold circuit that number is not academic. Carbon is moved continuously — screened into the tanks, pumped between vessels, stripped in the elution column, roasted in the regeneration kiln. Every one of those steps is a small fight against the carbon's mechanical strength.
Why Fines Are Expensive in Gold Plants
When carbon breaks into fines, three bad things happen at once:
- Gold leaves with the fines. Fines are too small to be caught by interstage screens, so the gold adsorbed on them escapes into the tailings. That is real, unrecovered gold.
- You buy more carbon. Every tonne of fines is a tonne of carbon you have to replace. Make-up (fresh) carbon in CIP plants commonly runs 20–100 g per tonne of ore; softer carbon pushes consumption to the top of that band.
- The circuit degrades. Fines blind screens, thicken slurry, and can interfere with elution column flow — pushing up maintenance and downtime.
The Real Cost of a Few Percentage Points
A plant processing 2,000 tonnes of ore per day might cycle 30–40 tonnes of carbon through its circuit in a year. If the carbon loses an extra 2% to attrition, that is roughly an extra tonne of make-up carbon per year — and at gold-grade GAC prices of $2,000+/tonne, that is real money before you count the gold lost in the fines.
On a larger operation the arithmetic scales fast: a 2% hardness difference on a 100-tonne annual carbon throughput is two tonnes of carbon — and with it, the gold those two tonnes were carrying when they broke up.
Abrasion is checked the same way in plant practice: agitate a sample in water for a fixed time and weigh the fines below 0.6 mm — lower fines means better abrasion resistance, and that number tracks closely with the make-up tonnage you will actually buy.
Iodine tells you how well the carbon can work. Hardness tells you how long it will keep working. A gold plant needs both — but only one of them silently taxes your budget every month.
Coconut Shell vs Coal-Based: Why Hardness Differs
Coconut shell carbon is naturally harder than most coal-based grades because of its dense, uniform structure. Typical gold-grade coconut GAC tests at 95–98% hardness, while many coal-based grades sit lower. That is the main reason most gold plants standardize on coconut shell despite its higher price per tonne — over the life of the circuit, lower make-up cost and less gold-in-fines usually more than pay back the premium.
How to Verify Hardness Before You Buy
- Ask for the COA of the actual lot, not a brochure figure. Check that hardness is reported for the exact batch you are buying.
- Run the abrasion test yourself. The ASTM D3802 test is simple and any competent lab can run it on a sample from the delivered lot.
- Order a trial bag first. Run it through your own screens and, if possible, a regeneration cycle. Real circuits reveal real hardness.
- Ask what the supplier's plant does differently. Activation temperature, wash quality and final screening all affect the hardness that reaches your tanks.
If you want to verify a delivered lot yourself, the logic is short: sample correctly (split the bag so the sample represents the whole lot, not the top of it), run the ball-pan hardness test (ASTM D3802), compare the result against the COA value for that batch, and ask for the test method on the COA — a hardness number without a stated method cannot be compared.
Buying Carbon You Can Trust
Our gold-grade coconut shell GAC is produced in our own plant in Longyan, China — hardness ≥98%, iodine 1000–1300, COA with every batch, and we welcome third-party lab testing of any lot.
See Gold CIP/CIL Grade Specs →FAQ: Carbon Hardness in Gold Recovery
For gold CIP/CIL, hardness should be at least 95%, ideally 96–98%+. Coconut shell grades from YELI Carbon test at 95–98%+. Hardness is what keeps carbon from grinding into fines — and fines are how gold leaves the circuit, so this number directly controls your make-up cost and gold losses.
The standard method is an abrasion test: a sample is agitated under controlled conditions (commonly with a ball or attrition drum), then re-screened. The percentage of material that stays on the screen is the hardness value. Any competent lab can run it on a sample from the delivered lot — ask for the number on the COA, then verify it on the actual batch.
In the lab, 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 the hardness percentage. Coconut shell gold-grade carbon typically tests 96–99%, while coal-based grades sit at 90–95%. Whatever figure you are quoted, the COA must state the test method — a hardness number without a method cannot be compared.
Gold carbon is screened, pumped, stripped in elution columns and roasted in regeneration kilns — every transfer grinds a little off the surface. Soft carbon becomes fines that escape the screens carrying adsorbed gold, blind the elution column and force more frequent make-up purchases. A 2% hardness difference can mean nearly a tonne of make-up carbon a year in a 2,000 tpd plant.
Yes. Coconut shell carbon routinely tests at 95–98% hardness, while coal-based grades typically sit at 90–95% or lower. That gap widens after weeks of elution and kiln cycles, which is one of the main reasons most gold plants standardize on coconut shell despite its higher per-ton price.