Froth Flotation in Mineral Processing: The Complete Guide to Reagents, Machines, and Circuits

Froth flotation is the single most important concentration process in the mining industry. Most of the world’s copper, lead, zinc and molybdenum concentrates, and much of its sulfide nickel, phosphate and potash, are produced by froth flotation, and fine coal is cleaned by it. No other beneficiation technique treats such a wide range of ores at such large tonnages. Understanding how flotation works — the chemistry of bubble-particle attachment, the role of reagents, the engineering of cells and circuits — is therefore central to understanding mineral processing itself.

This guide covers froth flotation from first principles through to advanced circuit design. It draws on the foundational work of Fuerstenau, Somasundaran, and their colleagues in surface chemistry and process engineering, and it is intended for metallurgists, process engineers, plant operators, and students who want a rigorous, practical grounding in the subject. If you are new to flotation, start with our shorter froth flotation basics article.

What Is Froth Flotation and Why Does It Work?

Froth flotation exploits differences in the surface chemistry of mineral particles. When an ore is ground to liberation size and suspended in water as a slurry, air bubbles are introduced into the pulp. Mineral particles that are hydrophobic — those that repel water — attach to the rising air bubbles, are carried to the surface of the cell, and overflow into a concentrate launder. Particles that are hydrophilic — those that are wetted by water — remain suspended and report to the tailings stream.

Very few commercially important minerals are naturally hydrophobic. Molybdenite, graphite, talc, coal, elemental sulfur, and stibnite are notable exceptions. The vast majority of valuable minerals — copper sulfides, galena, sphalerite, iron oxides, phosphate minerals — are naturally hydrophilic and must be rendered hydrophobic by the addition of selective reagents called collectors. The genius of the flotation process is that collectors can be chosen and conditioned to coat one mineral selectively while leaving others untouched, allowing extraordinarily fine separations that would be impossible by gravity or magnetic methods alone.

The thermodynamic condition for flotation is that a particle-bubble contact must result in a negative free energy change. This relationship is expressed through the contact angle — the angle subtended by a bubble resting on a submerged solid, measured through the liquid phase. When the contact angle exceeds zero, particle-bubble attachment is thermodynamically favored. Practically speaking, contact angles above about 40–50° are necessary for stable flotation, and the role of collectors is precisely to raise that angle on the target mineral while leaving the gangue with contact angles near zero. A full treatment of the surface-thermodynamic principles underlying flotation can be found in resources such as the Wikipedia overview of froth flotation or in the classic SME reference texts published by the Society for Mining, Metallurgy, and Exploration.

Flotation involves three phases (solids, water and air) and depends on chemical and physical variables at the same time. A useful way to picture this is as a triangle of chemistry, ore and machine. Chemistry controls whether a surface becomes hydrophobic. The ore sets particle size and liberation. The machine sets air rate, bubble size and mixing. This is why flotation is often described as a physicochemical process.

Work of Adhesion, Induction Time and Recovery Mechanisms

The strength of the particle–bubble bond follows from a balance of the three surface tensions at the contact line. The work needed to pull a particle off a bubble, the work of adhesion, is W_s/a = γ_w/a(1 − cos θ), where γ_w/a is the surface tension of water (about 72 mJ/m²) and θ is the contact angle. A larger contact angle gives a stronger bond that better survives the turbulence of a flotation cell.

Attachment is also time-dependent. When a particle collides with a bubble, the thin water film between them must drain and rupture. The time this takes is the induction time. On a hydrophobic surface it is a few milliseconds, shorter than the contact time during a collision, so the particle attaches. On a hydrophilic surface it exceeds the contact time and the particle bounces off. Collectors work by shortening the induction time.

Material reaches the concentrate by three mechanisms:

• True flotation: selective attachment of hydrophobic particles to bubbles. This is the main mechanism and the only selective one.
• Entrainment: particles suspended in the water that bubbles carry into the froth. Valuable and gangue particles are carried alike, and fines are affected most.
• Entrapment: particles caught between attached particles in the froth.

Entrainment and entrapment lower concentrate grade. The froth phase counters them by letting entrained material drain back while attached particles are retained, and multi-stage circuits are used because one stage rarely reaches a saleable grade.

Flotation Reagents: The Chemistry That Makes Separation Possible

Reagent selection and dosage are the dominant variables that determine flotation selectivity and recovery in any given ore system. Reagents fall into several functional classes: collectors, frothers, extenders, activators, depressants, and dispersants. Each plays a distinct role, and their interactions can be complex.

Collectors

The collector’s job is to adsorb selectively on the target mineral and impart hydrophobicity to it. Every collector molecule contains at least two functional parts: a polar or ionic group that anchors the molecule to the mineral surface through electrostatic attraction, chemisorption, or covalent bonding, and a nonpolar hydrocarbon chain that points away from the surface into the slurry water, creating the hydrophobic patch that allows bubble attachment.

The most important collectors for sulfide mineral flotation are the alkyl xanthates (potassium or sodium salts of alkyl xanthic acid). Ethyl, propyl, butyl, amyl, and hexyl xanthates are all in commercial use; increasing chain length increases the hydrophobicity and collector strength but reduces selectivity. Xanthates adsorb on sulfide minerals through covalent bonding with surface metal ions, and in many systems the actual collector species on the surface is dixanthogen — the product of xanthate oxidation — rather than the xanthate anion itself. This electrochemical dimension of sulfide flotation chemistry, governed by the rest potential of the mineral relative to the reversible xanthate/dixanthogen potential (approximately +0.13 V at standard conditions), is one of the defining features that distinguishes sulfide flotation from oxide flotation. Other important sulfhydryl collectors include dithiophosphates, thiocarbamates, and mercaptobenzothiazole. Dithiophosphates are the second most common thiol collectors. They are usually used together with xanthate and are stable over a wider pH range. Dithiocarbamates are more reactive than xanthates. Reactivity falls from dithiocarbamates to xanthates to dithiophosphates, and selectivity rises in the same order.

For oxide minerals, silicates, and non-sulfide systems, fatty acid collectors such as oleic acid, stearic acid, and tall oil are widely used, particularly for phosphate ores, iron oxide ores (haematite, goethite), and calcite. Fatty acid collectors adsorb through electrostatic interaction and precipitation of insoluble metal soaps at the mineral surface. Selective adsorption depends heavily on pH and the point-of-zero charge of the mineral, which varies from about pH 2 for quartz to pH 9 for corundum. Cationic amine collectors — primarily dodecylamine and related long-chain aliphatic amines — are used for the reverse flotation of silica from iron ore concentrates, for potash (sylvite from halite), and for feldspar recovery.

Speciality collectors for difficult ores include hydroxamates and salicylaldoximes (oxime-type reagents), which chelate strongly with copper, iron, and manganese surface species. These find application in the flotation of oxidized copper ores such as chrysocolla and malachite, which are otherwise poorly recovered by conventional xanthate-based chemistry.

Collectors are classed by how they behave in water. Nonionizing collectors, such as kerosene and fuel oil, are practically insoluble and are used to boost naturally hydrophobic minerals such as coal and molybdenite. Ionizing collectors are anionic or cationic. The anionic types are sulfhydryl (thiol) collectors for sulfides and oxyhydryl collectors, such as fatty acids, for non-sulfides. The cationic types are mainly amines, also used for non-sulfides. Petroleum sulfonates have less collecting power than fatty acids but are more selective. They are used for barite, celestite, fluorite, apatite, chromite, kyanite, mica, cassiterite and scheelite. Hydroxamates are chelating collectors that form stable complexes with metal cations such as iron, manganese, copper and the rare earths. They are used for pyrochlore, hematite, bastnaesite and oxidized copper minerals, and have proved more effective than fatty acids for the rare-earth mineral bastnaesite.

Collector adsorption is described as chemisorption, in which a chemical bond forms with metal cations on the surface, or physisorption, which relies on electrostatic attraction to a charged surface. In physisorption, the surface charge measured as zeta potential determines which collector works at a given pH. Collectors are dosed sparingly. Too much collector costs money, floats unwanted minerals and can make the froth immobile or cause it to collapse, and a collector is harder to remove once adsorbed than to keep off the surface in the first place. Plants often add a selective collector at the head of a bank and a stronger one further down for slow floaters. Dosage compared per tonne of contained metal, rather than per tonne of ore, makes plants easier to compare. In Canadian practice it is typically about 1–4 kg per tonne of metal.

Frothers

Frothers serve a different purpose from collectors: their job is to stabilize the air bubbles and the froth layer at the top of the flotation cell long enough for mineral-laden bubbles to overflow into the concentrate launder — but not so long that the froth becomes unmanageably stable and drains poorly. Common frothers include:

• Methyl isobutyl carbinol (MIBC) — the most widely used frother globally; produces a relatively unstable, fast-draining froth and generally suits finer particles, while polyglycols suit coarser particles
• Pine oil and terpineol — traditional frothers derived from pine processing; still used in coal flotation
• Cresylic acid (xylenols) — used in base metal sulfide flotation, particularly in South Africa
• Polypropylene glycols and polyglycol ethers — increasingly popular because their molecular weight can be tailored to produce froth stability tuned to the specific application

Frother molecules adsorb at the bubble surface and stabilize the liquid film between bubbles by exploiting the Marangoni effect: if the film thins locally, the reduction in surfactant concentration at that point raises surface tension, pulling liquid and surfactant back into the thinned region. Proper frother dosage is critical — over-dosing creates an excessively stable, water-laden froth that entrains unwanted gangue particles and reduces concentrate grade.

Frothers have three main functions: they help form and preserve small bubbles, they slow the bubbles’ rise through the pulp, and they form the froth. Smaller, slower bubbles mean more bubble surface area and more collisions. Adding frother typically reduces bubble size from about 4 mm to about 1 mm and raises the gas fraction in the pulp, which together can increase bubble surface area by roughly an order of magnitude. MIBC and polyglycol ethers account for more than 80% of frother use in metallic ore flotation. A single frother rarely suits the whole particle size range, so blends are common. Froth stability depends strongly on the particles carried in the froth: MIBC makes little froth in clean water but supports a substantial froth in a plant.

Activators

Some minerals do not respond to a collector in their natural state but can be made to respond by pre-treating with an activator — a reagent that modifies the surface to provide adsorption sites for the collector. The classic example is the activation of sphalerite (ZnS) with copper sulfate: in its natural form sphalerite adsorbs xanthate only weakly and floats poorly, but when copper ions exchange for zinc ions at the surface, the sphalerite surface behaves like a copper sulfide and responds vigorously to xanthate collectors. Copper sulfate addition ahead of the zinc flotation stage is universal practice in lead-zinc differential flotation circuits. Similarly, calcium ion activation (from lime addition) is used in the fatty acid flotation of quartz and silicate gangue minerals at high pH values.

Depressants

Depressants suppress the flotation of unwanted minerals, either by blocking collector adsorption sites or by rendering a mineral hydrophilic. Their action is essential for achieving selectivity in complex ores where valuable minerals must be separated from one another (differential flotation) or from closely related gangue minerals. Important depressants include:

• Sodium cyanide — depresses pyrite, sphalerite, and pyrrhotite in copper-lead-zinc circuits by forming stable metal-cyanide complexes on mineral surfaces that block xanthate adsorption; widely used but toxicity demands careful handling
• Zinc sulfate — depresses sphalerite in lead circuits, often used in combination with sodium cyanide
• Lime (CaO / Ca(OH)₂) — raises pH to depress pyrite (most effective above pH 11–12) and activates or depresses various minerals depending on the system
• Starch, dextrin, and CMC (carboxymethyl cellulose) — depress naturally hydrophobic minerals such as talc, graphite, coal, and molybdenite by adsorbing on their surfaces and extending hydrophilic hydroxyl groups into solution
• Sodium silicate (water glass) — disperses slimes and depresses silicate gangue minerals in iron ore and phosphate flotation
• Sulfur dioxide and sulfites — reducing agents used to depress sphalerite and pyrite, and galena in copper–lead separation (copper–molybdenum separation uses sodium hydrosulfide or Nokes reagent instead)

The interplay between activators and depressants is at the heart of differential flotation — the sequential recovery of multiple valuable minerals from a single complex ore. A typical copper-lead-zinc circuit might float copper and lead together in a bulk sulfide float, depress lead with dichromate while floating copper, then activate sphalerite with copper sulfate and float zinc, and finally float iron sulfides or discard them as tailings. Each step is tuned by pH control, reagent selection, and dosage optimisation. More on multi-mineral circuits is discussed in our article on metallurgical balances and sampling in mineral processing.

The Chemistry of Flotation by Mineral Type

Naturally Hydrophobic Minerals

A handful of minerals are hydrophobic in their natural state because their crystal structure exposes cleavage planes built from non-polar bonds. Molybdenite (MoS₂) has a layered structure in which electrically neutral S-Mo-S sheets are held together by weak van der Waals forces; cleavage along the basal (0001) plane exposes a nonpolar surface with a contact angle of about 75°. Similarly, graphite, talc, coal, and elemental sulfur all float without collectors, though in practice frothers and sometimes light hydrocarbon oil extenders are added to boost recovery. Depression of these naturally hydrophobic gangue minerals in circuits where they are not wanted (particularly talc and graphite in copper-nickel ores) requires the use of polymer depressants such as starch or CMC.

Polar and Nonpolar Minerals

Minerals are classed as nonpolar or polar by their surface bonding. Nonpolar minerals, such as graphite, sulfur, molybdenite, diamond, coal and talc, have weak molecular bonds at the surface that do not bond with water. They are naturally hydrophobic, with contact angles between about 60° and 90°. Layered minerals such as molybdenite and talc become less floatable as they are ground finer, because the proportion of polar edge surface to nonpolar face increases.

Polar minerals have covalent or ionic surface bonds that react strongly with water. They are often divided into five groups of increasing polarity. Group 1 is the sulfides (galena, chalcopyrite, pyrite, sphalerite and others) and the native metals, which are only weakly polar. Polarity then increases through the sulfates, carbonates, phosphates and oxides–hydroxides to the silicates and quartz, which are strongly hydrophilic. Freshly broken quartz forms silanol (Si–OH) groups that hydrogen-bond with water. This classification helps predict flotation behaviour and guides the choice of reagents.

Sulfide Minerals

Sulfide mineral flotation is the cornerstone of the world’s copper, lead, zinc, nickel, gold, and silver production. Unlike oxide minerals, sulfides are semiconductors whose surfaces develop measurable electrochemical potentials (rest potentials) in aqueous solution, and the electrochemistry of xanthate oxidation on these surfaces largely governs how well they respond to xanthate collectors.

For minerals such as pyrite, arsenopyrite, chalcopyrite, and pyrrhotite — which have rest potentials above the reversible xanthate/dixanthogen potential of +0.13 V — the surface reaction produces dixanthogen (X₂) as the primary collector species. For minerals such as galena and bornite — which have rest potentials below this threshold — the collector species is the metal xanthate salt (e.g., lead xanthate on galena). This distinction has important practical consequences: it explains why the same xanthate dosage and pH produces different results on different sulfide minerals, and why pH control is a powerful lever for selective flotation.

Galena (PbS) is generally one of the easiest minerals to float; complete recovery can be achieved with ethyl xanthate over a wide pH range from 2 to 10. Chalcopyrite (CuFeS₂) requires somewhat more careful pH control and collector choice for efficient separation from pyrite. Sphalerite (ZnS) is inherently difficult without copper activation. Pyrite (FeS₂) is the most troublesome gangue mineral in base metal circuits because its high rest potential means it is collected by xanthate almost as readily as the target sulfides — separation relies on pH elevation (lime addition) and selective depression.

pH and Pulp Potential in Practice

With thiol collectors, raising the pH works against collector adsorption, and each sulfide is depressed at its own pH. For the common sulfides, pyrite is generally depressed first, around pH 10, and chalcopyrite last, around pH 12. The gap between the two is the operating window for selective flotation. The critical pH also depends on collector concentration, which is why collector is often kept at starvation levels.

Lime is the most common pH regulator. In sulfide plants its total cost is often about double that of the collector, so pH control is worth optimizing. Because lime adds calcium, which can depress valuable minerals, some plants use soda ash instead or alongside it; several Peñoles copper–lead–zinc concentrators improved results by replacing lime with soda ash. Lime also partly depresses galena, so lead circuits may control pH with soda ash and depress pyrite and sphalerite with cyanide. Where pH must be lowered, sulfuric acid or sulfur dioxide is used.

Pulp potential, the oxidizing or reducing condition of the slurry, is the second chemical control in sulfide systems. If the potential is too low, the collector does not adsorb. Copper–molybdenum separation uses this deliberately: sodium sulfide lowers the potential, the chalcopyrite is depressed, and the naturally hydrophobic molybdenite keeps floating. If the potential is too high, sulfide surfaces oxidize and become hydrophilic; some plants aerate the pulp before flotation to depress iron sulfides this way. Recovery therefore rises with potential, levels off, then falls. When diagnosing low recovery, it helps to establish where on that curve the plant is operating. The oxygen demand of the slurry and EDTA extraction of surface oxidation products are practical ways to assess this.

Oxide and Silicate Minerals

Oxide mineral flotation is governed by surface charge phenomena. All oxide minerals — quartz, hematite, goethite, corundum, rutile, cassiterite — develop pH-dependent surface charges through protonation and deprotonation of surface hydroxyl groups. The pH at which the net surface charge is zero is the point-of-zero charge (PZC): approximately pH 2 for quartz, pH 6 for rutile, pH 7 for hematite, and pH 9 for corundum. Below the PZC the surface is positively charged; above it, negatively charged.

Anionic collectors (fatty acids, sulfates, sulfonates) adsorb on positively charged surfaces through electrostatic attraction, while cationic collectors (amines) adsorb on negatively charged surfaces. Selectivity between oxide minerals is achieved by adjusting pH to give a large difference in surface charge between the target mineral and the gangue. In iron ore reverse flotation, for example, amine collectors float quartz from hematite at about pH 9.5–10.5. Both minerals are negatively charged at this pH, so selectivity comes from starch, which adsorbs on the iron oxides and keeps them from floating.

Phosphate and Soluble Salt Minerals

Phosphate rock (principally fluorapatite) is one of the most important industrial minerals processed by flotation, primarily in Florida, North Africa, and the Middle East. Fatty acid and tall oil collectors are used at neutral to alkaline pH to float apatite from silica and carbonate gangue. The system is complicated by the similarity in surface chemistry of apatite and calcite — both are “semi-soluble” salt-type minerals whose surface charge and collector adsorption depend strongly on the dissolved mineral species in solution. Sodium silicate is almost always used as a gangue depressant in these circuits. For more detail on the downstream handling of phosphate concentrates, see our overview of dewatering operations in mineral processing.

Flotation Machines: Engineering the Bubble-Particle Contact

A flotation machine must accomplish three things simultaneously: suspend solids in the pulp, disperse air into fine bubbles, and transport mineral-laden bubbles to the froth zone efficiently. The design of the machine profoundly affects bubble size distribution, energy consumption, and residence time — and therefore recovery and selectivity.

Mechanical Flotation Cells

Mechanical cells are the dominant machine type in the industry. Their defining feature is a mechanically driven impeller-stator assembly that simultaneously agitates the pulp, draws air down the hollow impeller shaft or through an external blower, and disperses it into fine bubbles. The impeller creates a turbulent zone in which bubble-particle collision and attachment occur; the quiescent froth zone above allows mineral-loaded bubbles to accumulate without being re-pulped by agitation.

Major manufacturers of mechanical cells include Metso (which merged with Outotec, formerly Outokumpu, in 2020), FLSmidth (which absorbed Dorr-Oliver and Wemco), and BGRIMM. Cell volumes have grown dramatically over the past 40 years — from 100 ft³ (2.8 m³) machines in the 1970s to tank cells of 600 m³ and more today — driven by the need to process ever-larger tonnages economically. Larger cells reduce civil costs and the number of units to maintain, but they demand more sophisticated impeller design to maintain uniform solids suspension and bubble distribution throughout the cell volume.

Cell arrangements in industrial practice are either cell-to-cell (with adjustable weirs between cells, allowing better level control and reduced back-mixing) or open-flow banks (where pulp flows freely from cell to cell). Open-flow banks are simpler and have lower capital cost; cell-to-cell arrangements give more flexibility in circuit operation. Banks of cells are arranged in parallel when feed rates are too high for a single line, and in series when the desired retention time exceeds what a single cell can provide.

Column Flotation Cells

Column flotation represents a fundamentally different approach to flotation machine design. Industrial columns are tall, narrow vessels — commonly up to about 13 metres high and up to about 3.5 metres in diameter, round or square — with no moving parts. Air is introduced at the base through a sparger (a porous distributor made of pierced rubber or filter cloth) which generates fine bubbles that rise counter-currently against the downward-flowing slurry feed. The ore feed enters the column at roughly two-thirds of its height, and concentrate overflows at the top while tailings are removed from the bottom.

The key operational feature that distinguishes columns from mechanical cells is the addition of wash water sprayed into the froth zone at the top of the column. Wash water displaces entrained slurry water from the froth, dramatically improving concentrate grade by reducing the carry-over of hydrophilic gangue particles. Maintaining a “positive bias” — a net downward flow of water through the column — is essential for stable operation and high selectivity. Column flotation is therefore particularly well suited to cleaning applications where high grade is the primary objective, and it has largely displaced mechanical cleaner cells in many copper and phosphate circuits since the 1980s.

Typical operating superficial velocities in columns range from 0.5 to 3.0 cm/s for gas, 0.7 to 2.0 cm/s for slurry feed, and 0.05 to 0.3 cm/s for wash water. Bubble size is a critical variable: smaller bubbles (0.5–2 mm) produced by fine-pored spargers give more surface area for collection and better recovery at the cost of reduced carrying capacity. The 911 Metallurgist website has a useful practical introduction to froth flotation that covers operational variables in column flotation in accessible terms.

Other Machine Designs

Beyond conventional mechanical cells and columns, a number of specialised machines occupy niches in the industry:

• Jameson cells — developed in the 1980s by Mount Isa Mines and the University of Newcastle (Australia); a plunging slurry jet entrains air in a vertical downcomer, where gas holdup can approach 60%, and the aerated pulp disengages in a shallow separation tank; widely used for base-metal cleaning and fine coal, and to recover organic from copper solvent-extraction electrolyte
• Microcel — a column-type machine with a micro-bubble generator designed for fine and ultrafine particle flotation, a growing challenge as ore grades decline and grinding becomes finer
• Eriez HydroFloat — a coarse-particle separator that combines an aerated, fluidized teeter bed with flotation; it treats pre-classified feed, typically coarser than about 250 µm, and is used for potash, phosphate, diamonds, spodumene and vermiculite
• Flash flotation cells — treat part of the grinding-circuit cyclone underflow to recover coarse, fast-floating, high-grade particles before they are over-ground, with the tailings returning to the mill; widely used for gold, which concentrates in the circulating load
• Imhoflot pneumatic cells — high-intensity pneumatic cells developed by Maelgwyn Mineral Services; used for coarse particle flotation in some circuits
• REFLUX Flotation Cell (RFC) — a recent development from the University of Newcastle that combines a downcomer, fluidization and a set of parallel inclined channels to give high throughput while keeping mineralized bubbles out of the tailings

Flotation Kinetics: Recovery as a Rate Process

Engineers treat flotation like a first-order chemical reaction, with bubbles and particles as the reactants and the bubble–particle aggregate as the product. The rate at which floatable material leaves the pulp is proportional to the amount remaining, and the proportionality constant is the flotation rate constant, k. It lumps together all the unmeasured variables and is the standard engineering measure of floatability. Recovery then depends on how material flows through the cell:

• Batch test or plug flow: R = 1 − e^−kt
• One fully mixed cell (a good approximation for mechanical cells and columns): R = kτ / (1 + kτ), where τ is the mean residence time
• A bank of N equal fully mixed cells with total residence time T: R = 1 − (1 + kT/N)^−N

These relationships explain why cells are arranged in banks. For 90% recovery, a single fully mixed cell needs kτ = 9, while plug flow needs only kτ = 2.3, so a single mixed cell needs about 3.9 times the volume. Dividing the same volume among several cells in series moves the bank toward plug flow. With kT = 4, for example, one cell recovers 80%, four cells about 94%, and the plug-flow limit is about 98%. A bank also separates fast- and slow-floating minerals more sharply than one large cell. Residence time distributions measured with tracer tests show how close a cell is to fully mixed and reveal short-circuiting.

Gas Dispersion and the Rate Constant

The rate constant combines contributions from the machine, the particles and the froth. The machine contribution is described by the bubble surface area flux, S_b = 6J_g/D₃₂, where J_g is the superficial gas velocity and D₃₂ is the Sauter mean bubble diameter. The pulp-zone rate constant rises roughly in proportion to S_b, and this relationship has been shown to hold from a 60 L pilot cell to a 100 m³ industrial cell. In most flotation cells D₃₂ is between about 0.5 and 2.5 mm.

There is a ceiling. Mechanical cells and columns hold at most about 15% gas by volume (gas holdup). Beyond that, large bubbles form, the pulp–froth interface is lost and the cell appears to boil. A commonly cited safe operating range is J_g of 1–2 cm/s, D₃₂ of 1–1.5 mm and S_b of 50–100 s⁻¹. Reactor/separator machines such as the Jameson cell get around the limit: gas holdup in a Jameson downcomer can approach 60%, and energy goes more directly into bubble–particle contact.

Particle Size, Liberation and Entrainment

Recovery plotted against particle size shows three regions. Intermediate sizes float fast and recover well. Fine particles float slowly, and their recovery keeps rising with flotation time. Coarse particles also recover poorly, but extra time helps them little. The boundaries vary with the mineral, and a strongly floatable mineral has a wider high-recovery size range than a weakly floatable one.

• Fine particles have too little inertia to cross the water streamlines around a rising bubble, so they collide with bubbles less often. There is no lower size limit to flotation, only slower kinetics. Because air rate and bubble size are capped by the gas-holdup limit, the practical remedy is more flotation time, and installed cell volume, including scavengers, is often set by the slow-floating fines. Higher energy input also helps fines, which is why some plants raise impeller speed down a bank.
• Coarse particles are lost mainly because they detach from bubbles in turbulence and drop back out of the froth. Poorer liberation also contributes. Stronger collector chemistry helps, and high energy input hurts. Staged collector addition along a bank targets slow-floating coarse particles after the fast floaters have been removed. At Mt Keith, splitting the feed into size fractions for separate flotation improved nickel recovery enough to repay the project within a month.
• Entrainment increases sharply below about 50 µm and is proportional to water recovery to the concentrate. The degree of entrainment is defined as the recovery of an entrained particle class divided by the recovery of water. Reducing entrainment means reducing water recovery, through cleaner stages, dilution of cleaner feed, deeper froths and froth washing.

Recovery-by-size-by-liberation data from automated mineralogy are increasingly used to diagnose flotation losses. See our guide to ore mineralogy and liberation.

Flotation Circuits: Roughers, Scavengers, and Cleaners

A flotation circuit is an assembly of flotation cells arranged and interconnected to maximize both the recovery of valuable mineral into concentrate and the grade of that concentrate. Recovery and grade are inherently in tension — increasing recovery by floating more aggressively invariably pulls in more gangue — and circuit design is fundamentally about finding the best trade-off between these two objectives for a given ore and economic context.

Most flotation circuits are structured around three functional stages:

1. Rougher Circuit

The rougher circuit receives new feed (conditioned slurry from the grinding circuit) together with any recycled products (scavenger concentrate and cleaner tailings). Its objective is to maximise recovery: reagent dosages are typically generous, and the resulting rougher concentrate will contain significant gangue. Rougher tailings are sent to the scavenger stage. Because the rougher must process the full plant feed, it is the largest cell bank in the circuit and the dominant consumer of flotation reagents.

2. Scavenger Circuit

The scavenger circuit treats rougher tailings, attempting to recover any valuable mineral that escaped the roughers. Because scavenger feed is already depleted in the target mineral, long retention times and high reagent additions are sometimes used. Scavenger concentrate — low grade but recovered — is typically recycled to the rougher feed or treated in a separate cleaning bank. Scavenger tailings are the final plant tailings and should contain no more valuable mineral than economic recovery allows. Understanding where valuable minerals report in the tailings stream is essential; our guide on tailings management discusses downstream storage and environmental obligations for these streams.

3. Cleaner Circuit

The cleaner circuit upgrades rougher concentrate to saleable grade. A single cleaning stage is rarely enough for complex ores, so most circuits include multiple cleaner banks — first cleaner, second cleaner (recleaner), and sometimes third cleaner — with each stage progressively increasing grade. Cleaner tailings are returned (with or without regrinding) to the rougher or scavenger circuit to recover any valuable mineral lost with the gangue. In circuits where cleaner tailings are a significant stream, a cleaner-scavenger bank is sometimes added specifically to process these.

Circuit Configuration and Sizing

The complexity of the flotation circuit scales with the complexity of the ore. A simple monomineralic ore with an amenable gangue might be processed in a rougher-cleaner circuit with two stages of cleaning. A polymetallic copper-lead-zinc sulphide ore requires a differential flotation flowsheet with multiple bulk and selective float stages, selective depression, and interleaved regrinding. The number of cells required depends on the daily tonnage processed, the desired flotation retention time, pulp density (percent solids), and specific gravity of the ore. For a given ore type, rule-of-thumb retention times range from 4 to 8 minutes in roughers to 12 to 16 minutes in scavengers for coarser, more competent particles.

Modern circuits increasingly incorporate regrinding between rougher and cleaner stages, using IsaMills, stirred mills, or ball mills to liberate locked particles in the rougher concentrate before cleaning. This is particularly valuable in circuits processing finely disseminated ores. Our guide to grinding mills in mineral processing explains the equipment options for regrinding applications.

Choosing a Circuit Layout

Recycle streams let a circuit separate minerals more sharply than any single bank. The price is a circulating load that needs extra cell volume, so the target circulating load should be set at the design stage. Two basic layouts dominate:

• Rougher–scavenger–cleaner (R-S-C): the conventional arrangement, with scavenger concentrate and cleaner tailings recycled to the rougher. It buffers feed changes, but the stages interact, which makes control harder. It suits easy separations, including those mainly against entrained gangue, such as talc, coal and bitumen flotation.
• Rougher–cleaner–cleaner/scavenger (R-C-CS): cleaner-scavenger tailings go to final tails instead of back to the rougher. The rougher is responsible for recovery and the cleaner block for grade, so each can be controlled independently. Analysis shows this layout gives better selectivity when the target and gangue minerals float at similar rates. Brunswick Mine adopted it in the late 1980s to improve sphalerite–pyrite selectivity, the Noranda group moved to it across its operations in the 1990s, and it is common in copper–molybdenum plants that use columns as cleaners and mechanical cells as cleaner-scavengers.

Recycled streams carry the hardest material to float, namely fines, coarse particles and locked particles, so it is usually worth regrinding them. Concentrate from the first rougher cells or a flash cell can look high-grade enough to send straight to final concentrate. However, it can carry fine entrained gangue that a cleaner would reject with little loss of value. Plants with two or more payable minerals either repeat a circuit for each mineral or float a bulk concentrate and separate it afterwards. Simulators are increasingly used in circuit design, but the final choice must be a circuit that is easy to operate.

Designing for Stable Operation

A large agitated tank between grinding and flotation smooths out swings in grade and tonnage and can double as a conditioner. Without adequate conditioning, the first cells of a bank act as extra conditioning volume and recovery suffers. Parallel banks let the plant keep running at reduced feed rates, for example when a mill is down for maintenance. Most rougher flotation runs at 25–40% solids by weight, with a range of about 8–55%: a denser pulp needs less cell volume and less reagent, while cleaner feeds are often diluted to reject entrained gangue.

Flotation Testing

Most laboratory testwork uses batch flotation cells with 500 g to 2 kg samples. Good practice includes the following:

• Keep all solids in suspension without breaking up the mineralized froth.
• Condition the pulp with reagents before the air is turned on.
• Hold the froth depth at about 2–5 cm and add frother in stages to control froth volume.
• Use a fairly dilute pulp, about 10% solids, so that bubbles are not overloaded.
• Use the plant’s own process water, because dissolved species affect flotation.
• Automate froth removal where possible, because hand scraping varies between operators.
• Allow for entrainment when fitting rate constants.

Concentrates collected at timed intervals give the recovery–time data used to fit rate constants for scale-up. Batch tests cannot recycle cleaner tailings the way a plant does. Where cleaning is critical, locked-cycle tests repeat the test, normally at least six times, adding each cycle’s recycle streams and recycled process water to the next cycle until the test approaches steady state. Locked-cycle tests are laborious and do not always reach equilibrium. Pilot plants then confirm performance on a continuous basis. Surface analysis techniques such as ToF-SIMS and XPS can diagnose why particles are or are not floating.

Industrial Applications of Froth Flotation

Copper

Copper sulfide ore flotation is the largest single application of froth flotation globally. Porphyry copper deposits — which make up the majority of world copper production — typically contain 0.3 to 1.0% Cu as chalcopyrite (CuFeS₂), bornite (Cu₅FeS₄), and sometimes chalcocite (Cu₂S), disseminated in a large volume of granite or diorite. After grinding to 75–150 µm (80% passing), the ore is conditioned with xanthate collector (potassium amyl xanthate or PAX is common) and MIBC frother at pH 10–11 (maintained by lime addition to depress pyrite). The rougher float concentrates copper to 5–15% Cu in a bulk sulfide concentrate; cleaning stages raise this to the 25–35% Cu range required for smelting.

In copper-molybdenum porphyry deposits, molybdenite and chalcopyrite are floated together in the bulk copper circuit and separated in a subsequent Cu-Mo differential float where copper is depressed using sodium hydrosulfide (NaHS) or Nokes reagent while molybdenite is selectively floated.

Lead-Zinc

Lead-zinc ore (typically galena + sphalerite + pyrite in a carbonate or silicate gangue) is processed by differential flotation. In the most common circuit design, galena is floated first using low xanthate dosage at mildly alkaline pH while sphalerite is depressed with zinc sulfate and sodium cyanide. After lead concentrate has been removed, copper sulfate is added to activate the sphalerite, and a higher xanthate dosage floats the zinc into a zinc concentrate. Pyrite and iron sulfides report to tailings after depression with lime at pH 11–12.

Gold and Silver

Gold and silver are recovered by flotation when the precious metals are associated with sulfide minerals (pyrite, arsenopyrite, pyrrhotite) or when refractory ores contain fine gold encapsulated within sulfide hosts. Flotation in these cases concentrates the sulfide carrier minerals — not gold directly — into a bulk sulfide concentrate that is subsequently processed by cyanide leaching, pressure oxidation, or bacterial oxidation (BIOX) to liberate and recover the gold.

Phosphate

Florida and North African phosphate rock (fluorapatite) is processed by fatty acid flotation to produce phosphate rock concentrate for fertiliser manufacture. Typical circuits use tall oil fatty acid collectors with fuel oil extenders at neutral to slightly alkaline pH, with sodium silicate as a silica depressant. The unique chemistry of apatite-calcite-dolomite systems — where all three minerals are semi-soluble and respond to fatty acid collectors — makes selective separation challenging, and considerable research has been directed at developing reagent combinations that preferentially depress carbonate gangue while collecting apatite.

Coal

Coal flotation exploits the natural partial hydrophobicity of coal relative to mineral matter (clay, pyrite, quartz). Collectors are typically light fuel oils or diesel; frothers are pine oil, MIBC, or polyglycol-based. The challenge is that coal’s hydrophobicity depends strongly on its rank (degree of coalification), ash content, and the presence of oxidized surfaces. Column flotation has gained significant market share in coal circuits because the wash water addition dramatically reduces ash content in the product froth.

Iron Ore

Iron ore flotation is used both in the direct flotation of iron oxides (haematite, magnetite) and in reverse flotation — where silica is floated away from iron oxide concentrates using amine collectors (the cationic route) or fatty acids with calcium activation (the anionic route), with starch depressing the iron oxides. Reverse flotation has become the dominant approach for upgrading magnetic concentrates from magnetite circuits because it allows very high iron grades (above 68% Fe) to be achieved with acceptable silica content. For a full treatment of the upstream concentration steps used in iron ore beneficiation, including magnetic separation, see our article on magnetic and electrostatic separation.

Flotation Circuit Optimisation and Process Control

Flotation circuit performance is highly sensitive to variations in ore mineralogy, grind size distribution, water chemistry, and reagent dosage. Modern plants use a combination of manual metallurgical sampling, online instrumentation, and advanced process control to maintain performance in the face of these disturbances.

Key controlled variables include:

• pH — monitored continuously in conditioning tanks and flotation cells; lime or acid dosing is used to maintain target pH in each stage
• Froth depth and froth velocity — controlled through cell level (adjustable dart valves) and froth-scraping speed; froth cameras and froth analysis software are increasingly used to classify froth texture and infer metallurgical state
• Air flow rate — varied to control gas hold-up and bubble size; excessive air destroys froth stability; insufficient air reduces collection zone efficiency
• Reagent dosage — traditionally fixed on a g/t-of-feed basis; modern systems use on-stream analyzers and real-time grade sensors, backed by laboratory automated mineralogy (QEMSCAN, MLA), to adjust dosage in response to actual ore type
• Grind size (P80) — controlled in the grinding circuit upstream; coarser grind reduces energy consumption but may leave valuable minerals locked; finer grind risks generating slimes that impair flotation selectivity

Effective regulatory control comes before any of this: stable flows, densities, levels and froth depths are a prerequisite for meeting grade and recovery targets, and they do not depend on online assays. The most important single variable is pulp level. It is usually measured by a float with an ultrasonic sensor; conductivity probes and differential-pressure cells are also used. Level is controlled by a PI loop that adjusts the dart or pinch valve on the bank tailings, with feed-forward from upstream flows so that banks do not disturb one another.

Few flotation plants can run unattended for long. Practitioners have found that about 30% of control valves were in poor mechanical condition, 85% of PID controllers were poorly tuned, and only about 25% of control loops performed better in automatic than in manual. Instrument installation and maintenance, sound control-strategy design and loop tuning need to be fixed before advanced control can deliver. Measuring gas dispersion (bubble size distribution, gas velocity, gas holdup and bubble surface area flux) helps compare cells and find poorly performing units.

Advanced process control (APC) systems — model-predictive control or expert systems — are now in operation at many large flotation plants. These systems integrate signals from multiple sensors (pH, conductivity, level, airflow, froth cameras, online assay) to make coordinated reagent and operating adjustments that stabilize grade and recovery over time. The economic case for APC in flotation is well established: recovery improvements of 0.5–2.0 percentage points are commonly reported, which at copper plants processing millions of tonnes per year translate to substantial revenue gains.

Sampling and mass balancing of flotation circuits is an important tool for identifying performance bottlenecks, verifying reagent effectiveness, and benchmarking against design targets. Our guide on sampling and mass balancing in mineral processing provides a practical framework for this work.

Environmental Considerations in Flotation

Flotation reagents are chemicals, and their management has environmental implications that plant designers and operators must take seriously. Several reagents are acutely toxic or environmentally persistent:

• Sodium cyanide — highly toxic to aquatic life; subject to strict discharge limits under national regulations, with the voluntary International Cyanide Management Code setting good practice; must be destroyed in tailings water by SO₂/air or hydrogen peroxide before disposal
• Xanthates — generally biodegradable but can exert oxygen demand in receiving waters; products of xanthate decomposition (carbon disulfide, carbonyl sulfide) are volatile and odorous
• Frothers and petroleum-derived extenders — generally of low acute toxicity but can affect aquatic organisms at elevated concentrations; water management and circuit water recycling reduce discharge loads
• Process water chemistry — closed-loop water systems in modern flotation plants recycle process water rather than discharging to the environment; accumulated ions (sulfate, heavy metals, dissolved reagents) in recirculated water can affect flotation performance and must be managed

The development of green reagents — biodegradable collectors and frothers derived from renewable sources — is an active area of research, driven by tightening environmental regulations and ESG (environmental, social, and governance) commitments in the mining industry. Fatty acids derived from plant oils, bio-based frothers, and enzyme-treated starch depressants are examples of technologies gaining commercial traction. For more on the environmental obligations associated with the downstream disposal of tailings from flotation circuits, see our comprehensive guide to tailings management and safe storage.

Summary: Why Flotation Remains Indispensable

Froth flotation has been the dominant mineral concentration technology for over a century, and it will remain so for the foreseeable future. No other method can separate mineral particles at the fine sizes — 10 to 200 µm — at which most ore bodies must be ground to achieve liberation, across the enormous tonnage scales of modern mining, with the degree of selectivity that flotation routinely delivers. Its flexibility — the ability to tune chemistry, machine design, and circuit configuration to almost any ore type — is unmatched.

At the same time, flotation faces real challenges: declining ore grades requiring finer grinding and processing of more waste; more complex polymetallic ores that resist simple reagent schemes; increasing water and reagent costs; and mounting pressure to reduce the environmental footprint of mining. Addressing these challenges requires deeper understanding of surface chemistry, better sensor technology, smarter control systems, and continued innovation in machine design. The fundamental science laid out by Fuerstenau, Somasundaran, and their colleagues — contact angle thermodynamics, electrical double layer theory, electrochemical models of sulfide collector interaction — remains the bedrock on which all of this applied work rests.

For teams working on the upstream preparation of feed to flotation circuits, our articles on comminution and size reduction, hydrocyclone classification, and ore mineralogy and liberation provide essential context for understanding how grind size and liberation state affect flotation performance.

Plants that use froth flotation

These Global Circuits profiles describe operating and planned plants that use froth flotation. Each summary is based on the operation’s public technical report.

• Quellaveco (Peru, operating): A copper-molybdenum concentrator commissioned in 2022.
• Aripuanã (Brazil): 2.2 Mt/a sulphide flotation to separate copper, lead and zinc concentrates.
• Don David (Mexico, operating): Differential copper, lead and zinc flotation, then sulphur flotation.
• Lola Graphite (Guinea, project): Flotation to a graphite concentrate.
• Tamarack (United States, project): Bulk nickel-copper rougher flotation, cleaning and Cu/Ni separation.
• Woodlawn (Australia): Fresh ore and reclaimed tailings, fine grinding and differential flotation.
• Loma Larga (Ecuador, project): Separate gold-copper and pyrite-gold flotation concentrates.
• Boumadine (Morocco, project): Flotation to lead, zinc and pyrite concentrates.

Browse all profiles tagged Flotation in Global Circuits.

Sources and Further Reading

• Wills, B.A. and Finch, J.A. (2016). Wills’ Mineral Processing Technology, 8th edition. Butterworth-Heinemann/Elsevier, Oxford.
• Fuerstenau, M.C., Jameson, G.J. and Yoon, R.-H. (Eds.) (2007). Froth Flotation: A Century of Innovation. SME, Littleton, CO.
• Rao, S.R. and Leja, J. (2004). Surface Chemistry of Froth Flotation, 2nd edition. Kluwer Academic/Plenum Publishers, New York.
• Bulatovic, S.M. (2007). Handbook of Flotation Reagents: Chemistry, Theory and Practice. Volume 1: Flotation of Sulfide Ores. Elsevier, Amsterdam.
• Nguyen, A.V. and Schulze, H.J. (Eds.) (2004). Colloidal Science of Flotation. Surfactant Science Series, vol. 118. Marcel Dekker/CRC Press, New York.
• Finch, J.A. and Dobby, G.S. (1990). Column Flotation. Pergamon Press, Oxford.
• Lynch, A.J. et al. (2010). History of Flotation. Spectrum Series 18. AusIMM, Carlton, Victoria.

Online resources

• Froth flotation, Wikipedia
• Froth Flotation 101, 911Metallurgist
• Flotation, Metso (mechanical flotation cells and columns)
• Jameson Cell, Glencore Technology
• HydroFloat coarse particle flotation, Eriez

Equipment sizes and industry figures are given as reported in these sources and may have changed since. Online resources were accessed in October 2026.