Activated carbon is produced by heating coconut shell or other carbon-rich material until it develops a very large internal pore surface, several hundred square metres per gram. Dissolved gold cyanide adsorbs onto that surface, and screens retain the carbon granules in each tank while the slurry flows on. Carbon performance depends on two properties: capacity, the amount of gold a tonne of carbon can hold at equilibrium, and activity, the rate at which it takes gold up. Each pass through the plant fouls the carbon with lime, organic matter and other contaminants, so it is acid washed and regenerated in a kiln, and testing shows whether that regeneration is effective.
Carbon management has a large influence on soluble gold losses in CIL and CIP plants. This article describes the standard laboratory tests for carbon, the equations behind them, and the mass balance that links carbon movement to loading in the plant. Our article on CIL and CIP circuits covers the circuits themselves.
Capacity and the adsorption isotherm
In an isotherm test, a fixed mass of carbon is added to bottles of gold cyanide solution at different starting concentrations, the bottles are rolled until no further gold is taken up (typically 24 hours or more), and the gold remaining in solution is measured. Gold on carbon is calculated by difference:
gold on carbon (g/t) = (starting gold − final gold in solution, mg/L) × solution volume (L) ÷ carbon mass (kg)
The results almost always follow the Freundlich equation:
where
- gold loading on carbon at equilibrium (g/t carbon)
- gold left in solution at equilibrium (mg/L)
- ,
- Freundlich constants fitted to the test data; K is the loading at 1 mg/L
Taking logarithms gives a straight line, so K and n are obtained from a straight-line fit on log-log axes.
For the worked data, the slope is 1/n = 0.44 and K = 3,970 g/t. At equilibrium with a feed solution of 1.6 mg/L gold, the carbon would load to about 4,900 g/t, but at equilibrium with 0.02 mg/L, typical of the final tanks, only to about 700 g/t. This difference is the reason carbon is moved counter-current to the slurry, so that it is loaded in the early tanks where gold in solution is highest.
Isotherms depend on solution chemistry. Higher free cyanide, higher pH and competing metal cyanides such as copper reduce gold loading, while calcium increases it. Isotherms should therefore be run in site solution, not only in clean laboratory solution.
Activity and the kinetic test
A kinetic test contacts carbon with a large volume of gold solution, so that the gold concentration changes very little, and measures the loading at intervals, for example 15 and 30 minutes and 1, 2 and 4 hours. Fleming and Nicol's empirical model describes the early stage of loading well:
where
- gold loading on the carbon after time t (g/t)
- gold in solution (mg/L), held roughly constant in the test
- contact time (h)
- ,
- rate constants fitted to the test (Fleming and Nicol model)
| Time (h) | 0.25 | 0.5 | 1 | 2 | 4 |
|---|---|---|---|---|---|
| Gold on carbon (g/t), solution held at about 1 mg/L | 180 | 265 | 390 | 560 | 830 |
A straight-line fit of log loading against log time gives n = 0.55 (the slope) and k = 387 at 1 mg/L (the loading at 1 hour). The model is used to compare carbons and in CIL and CIP circuit models, and is valid for the early stage of loading, well before the carbon approaches equilibrium.
Monitoring regeneration through carbon activity
Plants track carbon activity by running a standard test, under fixed conditions set by the laboratory, on samples of regenerated carbon and of fresh reference carbon. Activity is usually reported as relative activity:
relative activity (%) = loading of plant carbon in the standard test ÷ loading of fresh carbon in the same test × 100
In the example, the regenerated carbon loaded 390 g/t in one hour and the fresh carbon 520 g/t, a relative activity of 75%. Declining activity indicates inadequate acid washing, a kiln operating too cool or with too short a residence time, or contamination by organics such as flotation reagents and oils. Because each laboratory's standard test differs, results should only be compared within one method.
Plant carbon balance
Gold removed from solution in the adsorption tanks reports to the carbon that is harvested. This balance links the plant's figures:
where
- solution flow through the adsorption tanks (m³/h)
- ,
- gold in solution entering the first and leaving the last carbon tank (g/m³ = mg/L)
- carbon moved forward and harvested (t/d)
- ,
- gold on loaded carbon and on carbon returned after elution (g/t)
Worked example: 900 m³/h of solution enters the carbon tanks at 1.60 mg/L gold and leaves at 0.02 mg/L. Gold adsorbed = 900 × (1.60 − 0.02) = 1,422 g/h, or 34.1 kg/d. With 8 t/d of carbon harvested and returned after elution at 50 g/t, the loaded carbon grade = 50 + 34,128 ÷ 8 = 4,320 g/t.
That is about 88% of the 4,900 g/t equilibrium loading with the feed solution, which is a reasonable approach to equilibrium. Moving carbon faster lowers the loading, which requires more elution capacity but improves adsorption in the later tanks; moving it more slowly raises the loading but risks higher gold in the final tank. Gold on carbon in the tanks is also a large component of gold in circuit, so carbon profiles feed directly into reconciliation.
Recommended practice
A carbon profile, measuring gold on carbon and gold in solution in every tank, should be run at least weekly, because it shows where adsorption is falling short. Carbon fines and losses should be tracked, since carbon broken by agitation passes through the interstage screens and carries loaded gold to tailings. New carbon supplies should be checked for activity, hardness and size before use.
Glossary
- Activated carbon
- Highly porous carbon, usually from coconut shells, that adsorbs gold cyanide from solution.
- Adsorption
- Molecules sticking to a surface; here, gold cyanide sticking to the pores of carbon.
- Isotherm
- The relationship between gold in solution and gold on carbon at equilibrium, at constant temperature.
- Freundlich equation
- An empirical isotherm, q = K·C^(1/n), which is a straight line on log-log axes.
- Carbon activity
- How fast carbon loads gold, compared with fresh carbon in a standard test.
- Elution
- Stripping gold off loaded carbon with hot caustic cyanide solution or water.
- Regeneration
- Heating eluted carbon in a kiln to burn off organic fouling and restore activity.
- Carbon profile
- A survey of gold on carbon and gold in solution in every adsorption tank.
Frequently asked questions
What is a good carbon activity?
There is no universal figure, because each laboratory's standard test differs. Relative activity against fresh carbon should be tracked over time, and a declining trend is the warning sign.
Why does gold loading on carbon fall when cyanide is high?
Free cyanide and other metal cyanides compete with gold cyanide for adsorption sites and keep more gold in solution, which lowers the loading at a given gold concentration.
How is gold loading on carbon measured?
A carbon sample is washed, dried and either fire assayed or ashed and digested. Gold on carbon is reported in g/t on a dry carbon basis.
What does n represent in the k and n model?
It describes how quickly the loading rate slows with time. It is usually below 1, because the most accessible adsorption sites fill first.
Discuss your plant with a metallurgist
Tiegel is a Perth metallurgy company. We help with testwork, plant optimisation, metal accounting and investigations, and we build the Tiegel Suite software.
Talk to a metallurgistReferences and further reading
- Fleming, C.A. and Nicol, M.J. (1984). The absorption of gold cyanide onto activated carbon. III. Factors influencing the rate of loading and the equilibrium capacity. Journal of the South African Institute of Mining and Metallurgy, 84. Link
- Marsden, J.O. and House, C.I. (2006). The Chemistry of Gold Extraction, 2nd edition. Society for Mining, Metallurgy and Exploration.
- Fleming, C.A., Mezei, A., Bourricaudy, E., Canizares, M. and Ashbury, M. (2011). Factors influencing the rate of gold cyanide leaching and adsorption on activated carbon, and their impact on the design of CIL and CIP circuits. Minerals Engineering, 24(6), 484–494. Link
Worked examples and charts marked illustrative use invented figures, not data from any operation.