Zinc Purification: Cementation Design for Cd, Cu and Co

Last updated: 2026-10-31

Three tanks, 45 minutes of total residence and 2.0-2.5 kg of zinc dust per cubic meter take a neutral leach solution from about 600 mg/L copper to below 1 mg/L. That is the envelope most zinc purification circuits converge on, and what decides whether it holds is not tank volume but the dust mass actually delivered to the reaction.

Key Takeaways

• Size on 30-90 minutes of total residence across two or three tanks, never on one mixer.
• Stoichiometry is 1.03 kg Zn per kg Cu and 0.58 per kg Cd; excess runs 1.5-3x.
• Cobalt will not cement at 60 °C without antimony or arsenic activation; the barrier is kinetic.
• Interstage filtration protects the cobalt stage from copper re-dissolution and lifts cake grade.
• Most breakthrough events trace to dust feeding, not to solution chemistry.

Why Does a Zinc Purification Circuit Decide the Whole Electrowinning Plant?

In a roast-leach-electrowinning flowsheet, purification sits between iron removal and the cell house. Neutral leach solution arrives carrying every base metal that reported to the leach, and the cell house cannot reject any of them again: whatever enters the electrolyte plates on the cathode, corrodes the cathode, or attacks the anode. That chain starts back in leaching.

Copper is the most damaging at low concentration. It is more noble than zinc, so it plates preferentially as a dark, nodular sponge, dragging current efficiency down and contaminating the deposit; copper left in solution also dissolves zinc metal chemically, so a cathode in a copper-bearing circuit perforates locally. Residual copper above 1 mg/L almost always means the dust was not there: copper cements in seconds on fresh dust.

Cadmium co-deposits with zinc and lowers the grade of every tonne produced. Cobalt is worse than its concentration suggests: it promotes pitting, re-dissolution of deposited zinc and lead anode corrosion, putting lead into the cathode as well. Antimony and germanium are the poisons, producing dendritic, brittle deposits and costing points of current efficiency at tens of micrograms per liter.

Zinc purification is a two-stage cementation duty that removes copper, cadmium, cobalt and nickel from sulfate solution before electrowinning. Cementation is a displacement reaction that precipitates a nobler metal onto zinc dust, and a purification stage is one agitated reactor that behaves as a single perfectly mixed volume in the cascade.

What Does Each Zinc Purification Stage Remove?

Almost every plant runs the same sequence: copper and cadmium first, cobalt and nickel second, an optional polish last. The split exists because the duties want different temperatures, and because copper must be gone before antimony is added or the activator is consumed on the wrong metal.

Stage 1, copper and cadmium. Solution is typically heated to 60-75 °C and dosed at 1.3-1.8 times stoichiometry. Copper cements almost instantly; cadmium is slower and needs the upper half of the window. Residence is 20-30 minutes to a residual near 1 mg/L of each, and thallium follows cadmium.

Stage 2, cobalt and nickel. The filtered liquor is heated to 80-90 °C and dosed with dust plus a copper-antimony or arsenic activator. Copper is deliberately kept at 20-60 mg/L entering the stage, because the activator works through a copper-antimony couple deposited on the zinc surface. Strip copper too hard and the stage loses its catalyst carrier. Filtration between stages is preferred because copper cement carried into stage 2 redissolves and consumes activator.

Stage 3, polish. Optional: it knocks down residual antimony from the activator and catches a cobalt excursion before the cell house.

All values are indicative and must be confirmed by test work on your liquor, because the ratio between impurities matters as much as their absolute level.

How Do You Size a Cementation Reactor Train?

A zinc purification train is a continuous stirred-tank cascade. One perfectly mixed tank gives an exponential residence-time distribution: a fraction of the feed short-circuits to the outlet, and that fraction is the tail showing up as 5 mg/L copper in a tank averaging 0.8. Three tanks in series narrow the distribution until the last sees an almost uniform feed.

Residence time is the primary knob. Total residence typically runs 30-90 minutes depending on temperature and the cobalt target. Cold circuits at 55-65 °C sit at the top of that range; hot circuits at 85-90 °C run 35-45 minutes, because cementation kinetics roughly double per 10 °C. Below 30 minutes the cascade cannot damp a dust feed upset; above 90 minutes you buy steel and pay for the re-dissolution that long hold times encourage. A three-tank cascade is preferred because it reaches a given residual at a lower dust excess than two tanks of the same total volume.

The example below sizes a 320 m³/h duty. Assumptions: density 1.28 t/m³ at 70 °C, total residence 45 minutes, three equal tanks, 25% freeboard, 60° cone bottoms.

1. Fix the flow basis. Use maximum sustained flow at operating temperature, 320 m³/h. Design margin belongs in the freeboard, not in the flow.
2. Convert residence to working volume. 45 min is 0.75 h, so 320 × 0.75 = 240 m³ across the train.
3. Split into stages. Three tanks of 80 m³ working volume. Two tanks of 120 m³ would pass the same average copper and fail the tail.
4. Add freeboard and fix geometry. 80 m³ working at 25% freeboard gives 100 m³ gross. At 4.5 m diameter the cross-section is 15.9 m², the 60° cone holds 20.6 m³, and the straight side works out at 5.0 m, for a total height near 8.9 m.
5. Size agitation for suspension, not mixing. Tip speed of 2.0-2.5 m/s just keeps 3-5 kg/m³ of cement in suspension at about 0.8 kW/m³, so 64 kW per tank. Faster buys air entrainment and re-dissolution.
6. Size dust feed and filtration. At 2.4 kg/m³ the duty is 768 kg/h, split 60:40 between tanks 1 and 2 through a loss-in-weight screw at ±1% of setpoint, with a filter between stages handling 320 m³/h plus wash water.

A cone bottom of 60° or steeper is not optional: cement settles into a dense bed within minutes of a shutdown. And the agitator must be selected against the settled bed, because a 64 kW running duty can need two to three times that to restart a bed that sat overnight. The impeller logic in our agitated leaching tank design guide carries over directly, and the three-stage continuous CSTR train page shows a cascade built on this basis.

What Zinc Dust Dose and Activator Chemistry Do You Actually Need?

Dose starts from stoichiometry and is then multiplied by reality. Per kilogram of metal removed, zinc demand is fixed by the exchange ratio: 1.03 kg Zn per kg Cu, 0.58 per kg Cd, 1.11 per kg Co. For 600 mg/L Cu, 140 mg/L Cd and 12 mg/L Co, the theoretical consumption is 0.62 + 0.08 + 0.01 = 0.71 kg/m³. Plants run 1.5-4 kg/m³, and the gap is the design problem.

Three effects fill it. Dust is an alloy, not pure zinc: 85-92% Zn is normal, so the assay must be divided out before stoichiometry means anything. Dissolved oxygen and ferric iron consume zinc without removing anything, at roughly 0.3 kg Zn per 10 mg/L of oxygen. And excess is deliberate: 1.5-3 times stoichiometry holds a low residual through the ten minutes when the feeder is behind.

The two-stage scheme is standard because the duties are incompatible in one vessel: stage 1 at 60-75 °C with dust only keeps the copper cement coarse and filterable, stage 2 at 80-90 °C adds excess and the activator.

Cobalt is the reason activation exists. The reaction Co²⁺ + Zn → Co + Zn²⁺ has a cell potential near +0.48 V, so it is thermodynamically comfortable; it simply does not proceed at a useful rate on clean zinc. Cemented cobalt forms a coherent, low-porosity layer that passivates the particle, and hydrogen evolution competes for the same sites. Antimony deposited on zinc creates sites with low hydrogen overpotential and acts as a nucleation template, carried onto the surface by the copper already in solution. An activator is a catalyst that changes the rate of cobalt cementation without being consumed in the net reaction.

In zinc purification, the dose is trimmed against oxidation-reduction potential, because ORP responds in seconds and an assay takes an hour. Against a silver/silver chloride reference, stage 1 typically sits at -200 to -300 mV and stage 2 at -350 to -450 mV; the setpoint is plant-specific and fixed by titrating dust against ORP on your own liquor. Always state the reference electrode, and always read ORP with pH.

Which Operating Parameters Really Control Zinc Purification Efficiency?

Eight variables explain nearly every excursion, and only one is chemistry in the narrow sense.

Temperature sets the rate constant and therefore the residence time you can get away with. Below 55 °C in stage 1, cadmium removal stalls and the cement turns fine and hard to filter; above 80 °C, gypsum scales the tank and cloth. Stage 2 loses half its rate per 10 °C below 85.

pH controls both the zinc hydroxide boundary and the re-dissolution window. Below about 3.8, hydrogen evolution eats dust and cadmium cement dissolves; above 5.2, basic zinc sulfate precipitates and carries cadmium and copper back as a filter-blinding solid. That narrow window is why pH is controlled at the tank, not a header.

In zinc purification, ORP is the fastest proxy for dust sufficiency: a rising trend at constant flow means the feeder is behind, the dust has oxidized, or the slurry is short-circuiting. Dust feed rate itself is the parameter most often measured badly, since volumetric screws drift with bulk density; only a loss-in-weight hopper with a totalizer gives a number you can mass-balance.

Mixing has an optimum: a visible vortex is air entrainment, and entrained oxygen re-dissolves the cement just made. Solids above 6-8 kg/m³ make the slurry hard to filter, and a filter running two hours behind the reactor lets cement sit in acidic liquor and redissolve.

Residual cement also contaminates the downstream solid-liquid separation steps, so filtration lag matters as much as the assays.

Troubleshooting: Why Does Cobalt or Cadmium Break Through?

1. Cobalt flatlines at several mg/L. Symptom: residual Co sits at 4-8 mg/L for days despite 3x stoichiometry. Cause: the activator is missing, oxidized to Sb(V) in an old make-up tank, or consumed by copper that is too high in the feed; or the stage runs at 70 °C and the kinetics never arrive. Action: assay antimony in feed and tank, prepare activator fresh at 60-70 °C, dose 1-3 mg/L Sb, and confirm the temperature before anything else.
2. Cadmium climbs between reactor and filter. Symptom: tank outlet assays 0.6 mg/L Cd while filtrate assays 3-5 mg/L. Cause: over-stripping with a large dust excess produces a very fine, high-surface-area cadmium cement that re-dissolves as soon as pH falls, in acidic wash water or a cake stock tank. Action: cut excess to about 1.4x, hold pH at 4.2-4.6 through filtration, and shorten cake residence time.
3. Cemented copper redissolves when dust starves. Symptom: copper rises across stage 2 instead of falling, filtrate turns faintly blue, cake grade drops. Cause: the feeder has fallen behind and ORP has climbed, while oxygen entering through the vortex or a leaking seal converts copper metal back to cupric ion. Action: trim dust on ORP cascade rather than operator judgment, keep 10-20% excess, lid the tanks, and repair seal leaks.
4. Antimony or germanium poisons the cell house. Symptom: pitted, dark-streaked cathodes, current efficiency down two to three points, cell feed at 0.08-0.15 mg/L Sb. Cause: activator over-dose with no guard stage, or germanium from an ore blend change that cementation only partly removes. Action: cap the activator, add a polish stage or zinc granule bed, and monitor Sb and Ge weekly.
5. Dust feeder bridges and rat-holes. Symptom: indicated feed holds while actual feed falls, one tank starves, its ORP drifts up over a shift. Cause: moisture caking the dust, a hopper angle shallower than the dust repose angle, or a screw that floods when the silo is full. Action: 70° hopper cone with vibration, dry-air pad on the silo, and daily calibration against a catch sample.
6. Fine cement breaks through the cloth. Symptom: turbid filtrate, cloth blinded within a shift, copper rising downstream of a stage that is chemically fine. Cause: cloth rating too open for a cement that went fine, feed pressure high enough to extrude fines through the weave, or a plate-pack leak. Action: check cloth air permeability, drop feed pressure and extend the cycle, precoat or recycle first-run filtrate. Cloth selection is covered in our filter press for mining guide, and a pilot-scale plate and frame press generates matching data cheaply.
7. Gypsum and basic zinc sulfate scale. Symptom: working volume falls, filter cycle stretches from 40 to 90 minutes, heat transfer drops. Cause: calcium saturation combined with pH and temperature swings; basic zinc sulfate forms above about pH 5.2 and 80 °C, gypsum above about 60% calcium saturation. Action: hold pH below 5.0 at temperature, bleed calcium with circuit solution, and track calcium weekly.

What Does Zinc Purification Cost per Cubic Meter?

Cost is dominated by one line, and that line is dust. At 2.0-2.8 kg/m³ of dust containing 88% Zn and an indicative 2.9-3.4 USD/kg of contained zinc, dust alone lands at 5.1-9.5 USD/m³ of solution. Everything else is smaller than the uncertainty in that number, which is why dust accounting deserves the best instrumentation in the plant.

Energy is second: agitation at 0.8 kW/m³ plus heating from 60 to 88 °C between stages, typically 0.25-0.45 USD/m³. Labor for three-shift cover with one operator per shift runs 0.30-0.50 USD/m³. Consumables add 0.15-0.30, maintenance with liner replacement 0.10-0.20.

The cadmium-copper cement is a credit, not a waste. A cake at 45-65% Cd with 10-20% Cu, sold on contained metal, is typically worth 1.5-4.0 USD/m³. Thallium is a penalty: it forces a refining charge.

Net cost lands at 3.5-7.0 USD/m³ for a well-run circuit and 8-11 USD/m³ for one over-dosing to compensate for a feeder nobody trusts. All figures are indicative, 2026 basis, exclude capital, and depend on local power and metal prices.

Capital logic runs the other way: extra tank volume is cheap against dust over a plant life, so an under-sized circuit runs at a permanent 3x excess and the penalty compounds. The acid consumption calculation and leaching recovery calculation methods give the mass-balance framework for testing that trade.

Field Data: 120,000 t/y Zinc Plant Purification Upgrade

A 120,000 t/y zinc plant in the 90-110 m³/h class ran two tanks, one volumetric dust feeder and no interstage filtration. Cell feed averaged 610 mg/L Cu, 140 mg/L Cd and 11 mg/L Co at the zinc purification inlet, and 4.2 mg/L Cu, 6.5 mg/L Cd and 3.1 mg/L Co at the outlet. Current efficiency sat at 88.4%, with cathode grade held only by trimming cell current.

The upgrade replaced the two tanks with a three-tank cascade of 95 m³ gross each, added a 60 m² filter press between stages, replaced the volumetric feeder with a loss-in-weight unit, raised stage 2 from 74 to 86 °C with a plate heat exchanger, and put dust trim on an ORP cascade with antimony dosed against copper in the stage 2 feed.

After three months: 0.5 mg/L Cu, 0.8 mg/L Cd and 0.4 mg/L Co; dust down from 2.9 to 2.1 kg/m³; current efficiency up from 88.4% to 91.3%. At 95 m³/h and 7,900 operating hours the dust saving is about 600 t/y, worth roughly 1.8 million USD at an indicative 3.0 USD/kg. The press added 0.15 USD/m³ in cloth and labor, the extra agitator 0.06 USD/m³ in power. Net benefit before capital recovery was on the order of 1.5 million USD per year against a 2.1 million USD installed change, a payback under eighteen months.

Two things did not go to plan. Stage 1 was left at 62 °C to save steam, and cadmium cement came out fine enough to blind the new press within four days; 68 °C fixed filtration without changing the dose. And the ORP probes went into the tank body rather than the outlet launder, costing two weeks of chasing phantom feeder errors. Our zinc industry overview covers the upstream leaching and iron removal.

Frequently Asked Questions About Zinc Purification

How long should zinc dust cementation take?
Total residence of 30-90 minutes across two or three tanks covers most duties. Hot circuits at 85-90 °C work at 35-45 minutes because kinetics roughly double per 10 °C, while cold circuits at 55-65 °C need the upper end. Below 30 minutes the cascade cannot absorb a dust feed upset.

How much zinc dust does a circuit consume per cubic meter?
Theoretical consumption follows the exchange ratios: 1.03 kg Zn per kg Cu, 0.58 per kg Cd and 1.11 per kg Co. Real circuits run 1.5-4 kg/m³ because dust is 85-92% Zn, oxygen consumes zinc, and a 1.5-3x excess is needed to reach low residuals.

Why does cobalt need antimony or arsenic to cement?
The cobalt-zinc exchange is thermodynamically favorable at about +0.48 V, but cemented cobalt forms a coherent, passive layer that blocks further reaction. Antimony on the zinc surface creates sites with low hydrogen overpotential and acts as a nucleation template. Without it the reaction is too slow to reach 0.5 mg/L.

Can a single tank replace a three-tank cascade?
Only at the cost of dose and certainty. One perfectly mixed tank passes a fraction of the feed straight through, so the outlet carries a long tail of unreacted metal. Three tanks in series narrow that distribution and hold a residual a single tank reaches only by over-dosing.

What temperature should cadmium removal run at?
Typically 60-75 °C, with 65-70 °C the practical sweet spot. Below 60 °C cadmium cementation is slow and the cement is fine and hard to filter; above about 75 °C you gain little rate and start paying for gypsum scaling on walls and cloth.

How is the dust dose controlled in practice?
By oxidation-reduction potential, trimmed continuously, with daily assays as the audit. ORP responds in seconds while an assay takes an hour, so the loop runs on ORP against a stage-specific setpoint referenced to a stated electrode, and the laboratory result re-baselines that setpoint.

Does dust purity matter as much as dose?
It matters because dose is meaningless without assay. Dust at 85% Zn delivers 15% less metal than the same mass at 98%, so a circuit tuned on one grade will under-dose after a supplier change. Assay every lot for zinc, lead, iron and moisture, and set the dose on contained zinc.

Conclusion

Cementation looks like the simplest step in a zinc plant and behaves like the most unforgiving. The design that works is unglamorous: three tanks, 45 minutes, 2.0-2.5 kg/m³ of dust, 60-75 °C for copper and cadmium, 80-90 °C with a controlled antimony activator for cobalt, and a filter between stages so the cement cannot return to solution. Each number is a starting point that test work confirms, not a specification. If your zinc purification circuit is fighting cobalt breakthrough or dust consumption that will not come down, the fastest answer is a sizing and dose review against your own liquor analysis; send the assays through our quotation page and we will return a train configuration and dose basis. Calculation guides sit in the resources library, and site structure for automated readers is described in llms.txt.

References

• USGS, Mineral Commodity Summaries: Zinc, annual editions, U.S. Geological Survey.
• Weiss, N.L. (ed.), SME Mineral Processing Handbook, Society of Mining Engineers of AIME.
• ASM Handbook, Volume 2: Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, ASM International.
• Porter, F., Zinc Handbook: Properties, Processing, and Use in Design, Marcel Dekker.
• Sinclair, R.J., The Extractive Metallurgy of Zinc, Australasian Institute of Mining and Metallurgy.
• Habashi, F., Textbook of Hydrometallurgy, Métallurgie Extractive Québec.
• Biswas, A.K. and Davenport, W.G., Extractive Metallurgy of Copper, Pergamon.
• International Zinc Association, technical publications on zinc electrowinning and solution purification.

Changelog

• 2026-10-31: First publication.