How Does Gold Flotation Work? Step-by-Step Process, Reagents and Machines Explained

An ore grading 3 to 4 grams of gold per tonne can be profitable in one plant and marginal in the next, and the difference is usually not the assay itself. It is where the gold sits in the rock. Free, visible grains respond to gravity concentration and dissolve quickly in leaching circuits. Gold locked inside pyrite and arsenopyrite, often as inclusions only a few microns across, stays out of reach of both. That gap is precisely the problem flotation was developed to solve, and it is why flotation sits at the center of most sulfide-hosted gold operations around the world.

The short answer to how gold flotation works: finely ground ore is mixed with water and reagents so that gold-bearing particles become water-repellent, then air is dispersed through the pulp as millions of fine bubbles. The water-repellent particles attach to the bubbles, ride them upward into a mineralized froth on the cell surface, and are skimmed off as concentrate, while water-attracting waste minerals sink and discharge as tailings. Every other detail in the plant, from grind size to reagent chemistry to circuit layout, exists to make that one separation selective, fast, and economical.

This guide walks through the principle in plain terms, the process step by step, the reagents and machines involved, the operating variables that move recovery numbers, the ore types where flotation earns its place, and the problems operators meet most often. A FAQ at the end answers the questions plant managers ask before committing to a flowsheet.

The Principle: Why Gold-Bearing Particles Attach to Bubbles and Waste Does Not

Every particle in a flotation cell presents its surface to water. Hydrophilic surfaces, such as quartz and most silicates, bond readily with water molecules, so water spreads over them and they sink. Hydrophobic surfaces repel water; when a bubble touches such a surface, the water film between them drains away and the bubble spreads across the particle and holds on. The measure used in the laboratory is the contact angle: the larger the angle at the point where particle, bubble and water meet, the stronger the particle clings to the bubble. Flotation is, at bottom, the industrial-scale management of contact angles.

Native gold is naturally somewhat hydrophobic, which is why panning works. In most orebodies, however, gold does not present itself as clean, liberated metal. It occurs as tiny inclusions inside pyrite and arsenopyrite, as attachments along mineral grain boundaries, or alloyed with silver and copper within sulfide hosts. A flotation circuit designed for such ore is therefore designed to float sulfide minerals, and the gold rides along with its carrier. When engineers talk about gold flotation, they are in practice talking about recovering pyrite, arsenopyrite and, in many deposits, copper sulfides, with the precious metal distributed through them.

Collectors create the hydrophobicity deliberately. Xanthates, of which sodium isopropyl xanthate and potassium amyl xanthate are workhorse examples, are heteropolar molecules: one end adsorbs onto a sulfide surface, while the other is a hydrocarbon chain that faces the water. A particle coated in this way behaves like a tiny wax bead in water. Dithiophosphates and thionocarbamates serve as secondary or alternative collectors where selectivity matters, for example when copper minerals must be floated away from pyrite in a copper-gold ore.

Frothers make the air side of the equation work. Bubbles in pure water coalesce and burst within seconds; frothers such as MIBC or polyglycol-type reagents let a cell hold a swarm of small, stable bubbles and support a load-bearing froth layer that survives the trip from the rotor to the launder.

For a particle to actually be recovered, three things must happen in sequence: bubble and particle must collide, the thin water film between them must drain and rupture, and the attached pair must survive turbulence all the way into the froth. Large, dense particles detach far more easily than small ones, which is one reason very coarse gold is usually routed to gravity equipment rather than asked to float.

The Gold Flotation Process, Step by Step

The sequence below shows the journey from run-of-mine ore to concentrate and tailings. In a real plant the circuit loops continuously: scavenger froth returns to roughing, cleaner tails return to regrinding, and process water circulates throughout. The order of operations, however, does not change.

A simplified linear view of what is, in practice, a counter-current circuit with internal recirculation loops.

Step 1: Crushing and Grinding, Where Liberation Sets the Ceiling

No downstream machine can recover a particle that was never freed. Grinding therefore determines the upper limit of flotation recovery before the pulp ever reaches a cell. Sulfide-hosted gold ores are commonly ground to a P80 somewhere between roughly 45 and 75 microns; refractory ores with very fine gold inclusions are ground finer still. A coarser grind saves energy but leaves composite particles that neither float cleanly nor report cleanly to tailings. Over-grinding creates the opposite problem: particles below about ten microns collide inefficiently with bubbles, float slowly, and consume disproportionate reagent. Balancing these pressures is a core task of flowsheet design, and it is settled in testwork long before a plant is built.

Grid-type discharge keeps the pulp level inside the mill low and the passage of material brisk, which suits the fine wet grinding stage that precedes flotation and helps limit the over-grinding that damages downstream response.

Step 2: Sizing and Pulp Preparation

Closed-circuit grinding with hydrocyclone classification delivers a fine overflow to flotation while sending oversize back to the mill. Rougher feed is normally conditioned at roughly 30 to 40 percent solids by weight, thick enough for high tonnage per cell but thin enough that viscosity does not smother bubble movement. Water chemistry deserves more respect than it usually receives: high salinity, hardness or organic contamination changes bubble behavior and froth texture, and plants that recycle process water must account for what accumulates in it.

Step 3: Reagent Addition and Conditioning

Conditioning is where the chemistry is loaded into the pulp. Lime raises pH into the target window, typically mildly alkaline for gold circuits. Collector is added in stages, part into the mill or cyclone feed where fresh mineral surfaces keep appearing, and the balance into agitated conditioning tanks ahead of the cells, because collector needs minutes of contact to adsorb before aeration begins. Frother follows, either conditioned with the pulp or added directly to the cell feed. Reagents are prepared as solutions of fixed strength so that dosing pumps can deliver repeatable rates hour after hour; sloppy reagent preparation is one of the quietest ways a sound circuit underperforms.

Step 4: Aeration, Bubble Formation and Froth

Inside the cell, a rotating rotor paired with a stationary stator shears the incoming air, drawn in by the rotor itself in self-aspirating machines or supplied by a blower in forced-air machines, into a swarm of fine bubbles, typically from a few tenths of a millimetre up to about two millimetres across. Hydrophobic particles colliding with these bubbles attach and are carried upward. At the surface, a froth layer builds, drains and is pushed by paddles or overflow into the concentrate launder. The depth of that layer is a live control handle: deeper froths drain entrained gangue and raise grade, while shallower froths pull more mass and raise recovery.

Step 5: Cleaning, Regrinding and Concentrate Production

Rougher concentrate is rarely sellable as it leaves the first bank. It carries entrained gangue and composite particles, so it is re-floated in two to five cleaner stages, with wash water sprayed over the froth in the final cells to displace mechanically trapped slimes. Middling streams, meaning cleaner tails and scavenger concentrates, are usually reground to liberate locked particles before re-entering the circuit. Final gold concentrates commonly grade in the tens of grams per tonne, though the number depends heavily on the ore and on whether the concentrate feeds a smelter, an oxidation plant, or an on-site leach circuit.

Step 6: Dewatering, Tailings and Water Return

Concentrate is thickened and filtered for transport; tailings are thickened to recover water and reduce the volume reporting to the storage facility. In gold plants this final step is more than housekeeping. Well-thickened tailings and recycled process water cut operating cost and reduce environmental load, and central-driven thickeners sized for the full tailings tonnage are standard equipment in any complete flotation plant.

Rougher, Scavenger and Cleaner: How the Circuit Fits Together

A gold flotation plant is arranged around three duties. Roughers take fresh feed and aim at speed: recover the fast-floating fraction early, at a mass pull that typically runs from single digits up to around twenty percent depending on how sulfide-rich the ore is. Scavengers treat the rougher tailings; their froth, lower in grade, is returned to the head of the circuit so that slow-floating or weakly attached particles get a second and third chance. Cleaners take the combined rougher concentrate and push it up to saleable grade, discarding gangue in their tailings back into recirculation.

The logic is counter-current: grade increases stage by stage toward the final concentrate, while recovery is protected by recycling every marginal stream. Between cleaner stages, regrinding of middlings is common, particularly for refractory ores, where the first pass frees gold from pyrite but the particles need finer grinding to separate cleanly from composite gangue.

Circuits get rebuilt around this logic at every scale. The renovation and expansion of a gold mine concentrator in the Solomon Islands followed the same pattern: existing capacity was upgraded and the concentrator circuit rebuilt around updated equipment to lift both throughput and recovery. Projects of this kind succeed or fail on the fundamentals described above, namely liberation, reagent practice and cell duty, rather than on any single machine.

For a plant owner, the practical takeaway is that each duty is sized independently. Rougher volume is set by feed tonnage and the flotation kinetics of the ore; cleaner volume by the mass pull it must upgrade; scavenger volume by how much value still hides in the rougher tails.

Reagents Used in Gold Flotation

Reagent selection decides what floats and what stays behind. The goal in a gold circuit sounds simple, collect the sulfides that carry gold and leave everything else, but each ore presents its own complications: carbonaceous matter that steals dissolved gold, talc and clays that slime the froth, or copper minerals that must be separated later. The main reagent families and their roles are summarized below.

Order of addition matters as much as the list itself. Modifiers set the pH window first, because collector adsorption depends on the surface chemistry it meets. Collector enters early, part with the grinding circuit and part in conditioning, and frother goes in last. Stage addition beats dumping the full dose at one point: fresh mineral surfaces keep appearing as particles grind and circulate, and they need fresh collector to respond.

The pH window deserves emphasis. Gold-bearing pyrite floats well in mildly alkaline pulp, commonly between about pH 8 and 9.5. Push pH above roughly 10.5 and pyrite itself begins to depress, which is exactly what a copper separation circuit wants and exactly what a bulk gold circuit must avoid. Lime is cheap and effective; soda ash appears where the calcium ions from lime cause trouble elsewhere in the plant.

Two special cases are worth knowing. In ores with carbonaceous matter, depressants and blinding agents are used to passivate the carbon, which otherwise adsorbs recovered gold and creates the well-known preg-robbing problem. In copper-gold separation, some plants use depressant packages that include cyanide salts; in gold plants these demand strict control, because cyanide also dissolves gold. Where arsenopyrite carries the gold, activation with copper sulfate is sometimes applied to strengthen its response, provided testwork supports it.

Flotation Machines: Matching the Cell to the Duty

Mechanically agitated cells dominate gold flotation. A rotor-stator pair does three jobs at once: keeps solids in suspension, shears air into fine bubbles, and provides the turbulence that brings particles and bubbles together. Machine selection is therefore about how air gets into the pulp and how large the tanks are built.

Self-aspirating machines draw air in through a hollow shaft or an air duct around the shaft, powered by the rotor alone. They need no blower, their aeration roughly self-regulates with pulp level, and they simplify utilities, which makes them attractive in smaller and mid-size plants and in roughing duties where steady aeration suffices. Forced-air machines take air from a blower through a duct; the operator controls air rate per cell independently of agitation, which gives finer control and, at very large tank volumes, better energy efficiency. Large forced-air cells are frequently paired in series so that one cell can draw the following cell's feed without additional pumping, and this configuration is the standard answer for high-tonnage sections.

Sizing comes down to residence time. Total rougher residence commonly falls anywhere from under ten minutes to more than thirty depending on how fast the ore floats; banks are built from enough cells in series to approach plug flow within the bank, because a single large tank lets fresh feed short-circuit straight to the launder. Cleaners run shorter times with wash water on the froth, and scavengers mirror the roughers they follow.

Whatever the model, the practical checklist for a buyer is the same: aeration control per cell, rotor and stator wear life, froth paddle and launder design matched to the froth the ore actually produces, and maintenance access that allows a worn rotor to be changed without stopping the whole bank.

The Operating Variables That Move Recovery

Typical starting windows for sulfide-hosted gold ores. Every ore demands its own testwork; treat these figures as orientation, not targets.

Grind size is the loudest of these variables because it works twice: it sets liberation, and it sets collision behavior in the cells. A grind that is too coarse sends composite particles to tailings; a grind that is too fine smothers the froth with slow-floating slimes. Recovery as a function of grind typically rises, peaks and falls, and the plant wants to live at the peak, which testwork locates before design and daily assays confirm in operation.

Collector dosage follows a law of diminishing returns. The first tens of grams per tonne buy most of the achievable response; further additions buy progressively less recovery while eroding selectivity, pulling barren pyrite into the froth and inflating the reagent account. Testwork identifies the knee of that curve, and disciplined dosing keeps the plant on it.

pH control is the quiet variable. A drift of one pH unit toward high alkalinity can visibly flatten pyrite response, and operators who chase recovery with extra collector while the real culprit is over-liming are treating the symptom instead of the cause.

Conditioning time and pulp density travel together. Thick, cold, viscous pulp slows reagent adsorption and bubble movement; thin pulp raises pumping and capacity costs. Both must be held stable for the cells downstream to behave predictably from shift to shift.

Aeration rate and froth depth are the operator's fast handles. More air and shallower froth raise mass pull; less air and deeper froth raise grade. Automated control loops increasingly hold these steady, but a competent operator reading froth texture still catches a drifting circuit before the assay laboratory does.

Water chemistry rounds out the list. Recycled water carries dissolved salts and residual reagents that accumulate shift after shift; the froth changes character, and the plant must either bleed and treat a portion of the water or adjust the reagent scheme to match.

Which Gold Ores Respond Well to Flotation

Free-milling ores in which gold rides in pyrite or arsenopyrite are flotation's home ground; recoveries to concentrate commonly run high, and the product can be sold, oxidized, or leached on site. Ores in which gold is locked in solid solution within arsenopyrite, the classic refractory case, still pass through flotation, but the concentrate then feeds an oxidation step such as roasting, pressure oxidation or biological oxidation before leaching. In copper-gold ores, flotation produces a bulk copper-gold concentrate sold on copper grade with gold paid as a credit, or a separate circuit splits the two metals.

There are also ores where flotation is not the first choice: coarse free gold belongs to gravity concentration, simple oxidized ores with little sulfide belong to direct leaching, and some carbonaceous ores resist flotation unless the carbon is first mitigated. The comparison below summarizes where each main route wins.

A practical screening question helps at the evaluation stage: if a significant share of the gold reports to sulfide intervals in mineralogical work, flotation deserves a place in the flowsheet; if the gold is free and coarse, gravity and leaching may carry the load on their own.

For a wider view of how these decisions play out on real projects, the overview of mineral processing methods and case studies for gold ore is a useful next read.

Common Problems in Gold Flotation and Practical Fixes

Most flotation losses in gold plants trace back to a short list of recurring causes. The list below pairs each with the fix that plants actually apply on the ground.

1. Slimes coating coarser particles. Ultrafine clay and oxidation slimes blanket sulfide surfaces and depress response. Fixes: deslime ahead of flotation where the mineralogy allows, add a dispersant, and hold pulp density in the workable range.
2. Coarse gold escaping to tailings. Particles too large or too dense to float reliably leave through the tailings launder. Fixes: gravity concentration ahead of flotation, and a flotation cell installed on the grinding circuit discharge to catch liberated coarse values early, before they have another chance to be lost.
3. Over-liming. pH pushed above the pyrite window quietly depresses the carrier mineral. Fix: meter lime against a measured pH target rather than habit, and watch pyrite recovery whenever the dose is raised.
4. Froth instability. Froth that collapses carries nothing; froth that is too stiff carries gangue. Fixes: check frother dose and water contamination first, since wood chips, oils and upstream organics are frequent offenders, then adjust aeration and froth depth.
5. Oxidized or weathered surfaces. Sulfides exposed to prolonged oxidation respond poorly to collectors. Fixes: longer conditioning, a stronger or blended collector, and where applicable a sulfidizing treatment; severely oxidized material may justify re-evaluating the whole route.
6. Reagent dosing drift. Dosing pumps set once and forgotten let the plant wander as the ore changes. Fix: tie dosing to feed rate and assay feedback, and audit solution preparation strength on a regular schedule.

None of these problems are exotic, and all of them respond to attention, instrumentation and a disciplined operating routine.

From Flowsheet to Working Plant: What the Equipment List Looks Like

A flotation plant is a chain, and the cells are only one link. A complete flowsheet built around gold flotation typically includes the following equipment groups, each sized to the tonnage and the ore's flotation kinetics.

• Crushing and screening ahead of the mill
• A wet grinding circuit, commonly a grid-type ball mill in closed circuit with classification
• Agitated conditioning tanks plus reagent preparation and dosing systems
• Rougher, scavenger and cleaner flotation banks
• Regrinding equipment where middlings require further liberation
• Thickeners for concentrate and tailings, with clarified water returned to the plant
• Filtration for the final concentrate product

The engineering path to that list runs through testwork: batch flotation tests on drill core to screen reagent schemes, locked-cycle tests to approach steady-state performance, and, where the stakes justify it, pilot campaigns. Skipping these steps is how plants end up grinding to the wrong size or floating at the wrong pH, and corrections made after commissioning cost far more than the testwork would have.

For buyers, commissioning support matters as much as the hardware itself. A manufacturer that supplies the full chain, from grinding and conditioning through flotation and thickening, carries responsibility for the interfaces between stages, and the interfaces are where flowsheet drawings and real plants tend to disagree. Installations across gold, copper, zinc, molybdenum, potash, phosphate and lithium projects point to one common lesson: the flowsheet is proven in testwork, but it is delivered in the details of installation and commissioning.

Frequently Asked Questions About Gold Flotation

Does flotation recover all the gold in an ore?

No. Very coarse gold floats poorly and is better captured by gravity concentration; extremely fine gold locked in silicate gangue may not float at all; and gold held in solid solution inside sulfides floats only together with its host mineral. A well-designed plant accounts for this by combining methods rather than asking flotation to do everything.

Why choose flotation instead of simply leaching the whole ore?

When gold is locked inside sulfides, leaching solutions cannot reach it, so whole-ore treatment recovers little. Flotation concentrates the gold-bearing sulfides into a small fraction that can be oxidized or leached efficiently. Even in free-milling ores, floating sulfides first can cut the tonnage reporting to the leach circuit and reduce reagent consumption.

What grind size does gold flotation need?

Sulfide-hosted gold ores commonly start at a P80 between roughly 45 and 75 microns, and refractory ores go finer. The correct answer for a specific deposit comes from mineralogical analysis and flotation testwork, not from a handbook.

How many flotation cells does a plant need?

The number follows from the required residence time and the cell volume available, not from a fixed rule. Rougher banks are built from multiple cells in series to approach plug flow; scavenger and cleaner banks mirror the duties they serve. Residence-time requirements for the specific ore come from kinetic testwork.

What grade does gold concentrate reach?

Gold concentrates commonly grade in the tens of grams per tonne, with the exact figure set by the ore and by downstream requirements. Payability depends on the buyer's terms, on penalty elements such as arsenic, and on whether the concentrate is sold to a smelter or processed on site. Copper-gold concentrates are quoted on copper grade with gold credited separately.

Can flotation treat oxidized gold ores?

Usually not economically on its own. Oxidized ores carry little floatable sulfide, and oxidized mineral surfaces respond weakly to collectors. Where an ore body transitions from oxidized material to sulfide mineralization at depth, plants often float only the sulfide zone and route the oxide material to leaching.