Flotation is the third major recovery route covered in this set, alongside gravity recovery and the leaching methods described on tailings gold recovery methods, and it works on a genuinely different principle from either. Rather than separating by density or dissolving gold into solution, flotation exploits surface chemistry, making sulphide minerals selectively attach to air bubbles and float to the surface as a froth while the bulk of the gangue sinks and reports to the tailings stream. It is the sub-head for this branch of the recovery methods set, and every flotation-specific node below it, from flowsheet sequencing to refractory pretreatment, builds on the mechanism described here first.
How does froth flotation actually separate sulphides from gangue?
Froth flotation separates sulphides from gangue by making the sulphide particle surfaces hydrophobic, water-repelling, so they attach to rising air bubbles and float, while gangue particles stay hydrophilic and remain suspended in the pulp. Collectors, commonly xanthates such as PAX (potassium amyl xanthate) or SIBX (sodium isobutyl xanthate), are the reagents that coat sulphide surfaces and confer that hydrophobicity; without a collector, most sulphide particles would not float on their own. Frothers, MIBC (methyl isobutyl carbinol) being the most widely used, stabilize the air bubbles themselves, keeping the froth layer intact long enough at the cell surface to be skimmed off mechanically before it collapses back into the pulp.
Getting flotation to work well depends on more than just adding these two reagent classes. Pulp pH is controlled with lime or soda ash to activate the collector's affinity for the target sulphide while suppressing unwanted minerals from floating alongside it, and conditioning time, the period during which reagents are mixed into the pulp before it enters the flotation cells, needs to be long enough for the collector to actually coat particle surfaces before air is introduced. A flotation circuit that skips or shortens conditioning typically floats poorly regardless of how well the reagent dosages themselves are calculated.
Why is flotation relevant to gold specifically, if gold itself isn't floated?
Flotation matters to gold recovery because gold in many ores and tailings is physically associated with sulphide minerals, most commonly pyrite and arsenopyrite, rather than existing as fully liberated free particles. Floating the sulphide host mineral concentrates the gold that travels attached to or locked inside it, producing a small, sulphide-rich, gold-rich concentrate from a much larger tonnage of barren gangue. This is an indirect recovery route: flotation itself never touches the gold chemically, it simply rides along with its sulphide host to the froth layer, which is why a flotation concentrate on its own is not a saleable gold product; it still needs a downstream stage, typically leaching, to actually extract the gold from the sulphide matrix it was concentrated in.
This distinction matters for how a technical reader should interpret a flotation recovery figure. A high sulphide recovery to the froth does not automatically mean a high gold recovery, since gold deportment within the sulphide, whether it rides on the mineral surface, is included within the crystal lattice, or sits as sub-microscopic inclusions, determines how completely a given sulphide recovery translates into gold recovery. Deportment testwork, covered elsewhere in this site's metallurgy section, is what actually answers that question for a specific dump rather than assuming a fixed ratio.
What particle size range does flotation actually work within?
Flotation has its own particle size window, distinct from gravity's, and performance drops off at both ends of it. Particles that are too coarse are too heavy for an air bubble to lift regardless of how hydrophobic their surface is, which is why flotation feed is typically ground finer than a gravity circuit's feed would need to be. Particles that are too fine, meanwhile, have so little mass relative to their surface area that they can fail to attach to a bubble with enough force to stay attached through to the froth layer, or they can simply follow water into the froth indiscriminately regardless of surface chemistry, a phenomenon plant metallurgists call entrainment, which lowers concentrate grade by carrying unwanted fine gangue along for the ride rather than a genuine selective float.
This means flotation circuit design has to balance grinding fine enough to liberate the sulphide from gangue against grinding so fine that the resulting ultrafine fraction starts working against selective flotation. That balance is set from the same particle size and liberation testwork referenced throughout this recovery methods set, not assumed from a generic grind target borrowed from a different ore type.
When does flotation beat leaching directly on tailings?
Flotation beats direct leaching on tailings where the feed is sulphide-rich, cyanide-hungry, or carries gold in a form leaching alone struggles to reach efficiently. Base-metal sulphides consume cyanide in their own side reactions without yielding any gold in return, so a feed with meaningful sulphide content can be reagent-expensive to leach directly at full tonnage. Floating that sulphide fraction out first, then leaching only the resulting concentrate, redirects the bulk of the cyanide demand onto a small fraction of the original mass rather than the whole stream, which is exactly the flowsheet logic explored further on sulphide flotation before leaching.
Flotation is far less relevant, and often not worth its added capital and reagent cost, on oxidized, low-sulphide tailings where gold is already free-milling and directly leachable. A large share of Ghana's weathered legacy tailings fall into this category, which is why flotation is presented here as a conditional, testwork-driven addition to a flowsheet rather than a default stage every project should build.
The economic case for adding a flotation stage rests on comparing its capital and reagent cost against the cyanide savings and recovery improvement it delivers on the specific feed in question. A flotation circuit adds cells, reagent dosing systems, and a concentrate thickening and regrind stage of its own, all real capital that a low-sulphide dump's leach-only economics may not need to carry. Diagnostic leach testwork, described in the metallurgy and testwork section of this site, is what actually quantifies whether a specific dump's sulphide content and cyanide consumption profile justify that added investment.
Where does flotation sit in a tailings retreatment flowsheet?
Flotation sits after the gravity stage and before the leach circuit in a flowsheet that uses all three methods, since gravity cheaply captures coarse, free gold first, flotation then concentrates whatever sulphide-associated gold remains in the gravity tailings, and the leach circuit finally dissolves gold from both the flotation concentrate and any material bypassing flotation entirely. That sequencing mirrors the overall plant logic described on gold processing plants for tailings, where each stage is sized for the fraction of the feed it is actually needed for rather than the whole tonnage.
Does flotation work on tailings the same way it works on fresh ore?
Flotation on tailings does not always behave identically to flotation on fresh ore, because the sulphide surfaces in weathered tailings have often oxidized to some degree during years or decades of storage, and an oxidized sulphide surface responds differently to standard collector chemistry than a fresh, unoxidized one. Reagent conditioning, collector selection, and sometimes a sulphidizing pre-treatment step to restore a more floatable surface may need adjusting specifically for tailings feed rather than simply copying a fresh-ore reagent scheme. This is a genuine, testable metallurgical variable, not an assumption that should be applied uniformly across every dump regardless of its oxidation history.
Locked-cycle laboratory testwork, running representative tailings samples through a full simulated flotation circuit including reagent recycle and middlings recirculation rather than a single open-circuit batch test, is the standard way this behavior gets characterized before a plant is designed. Skipping this step and assuming a fresh-ore reagent scheme will transfer directly to weathered tailings is a common and avoidable source of underperforming flotation circuits on retreatment projects.
How this connects to the next stage in the flowsheet
A flotation concentrate is only an intermediate product, and what happens to it next, how it gets leached and how much of its cyanide demand it actually removes from the rest of the plant, is worked through in full on sulphide flotation before leaching.
Why oxide versus sulphide composition matters before flotation is even considered
Whether a specific tailings dump has enough sulphide content to justify a flotation stage at all comes down to its oxide-versus-sulphide composition, a mineralogical distinction that should be established through composition testwork before flotation capital is committed rather than assumed from the dump's age or location alone.
Reagent handling and safety considerations
Xanthate collectors and frother reagents require proper storage, dosing control, and worker handling protocols, since xanthates in particular can decompose and generate carbon disulfide under poor storage conditions. Sound reagent management here is a routine part of responsible plant operation, the same standard applied to cyanide handling elsewhere in this recovery methods set.
Why flotation concentrate mass matters to overall project planning
A well-run flotation stage typically produces a concentrate representing only a small fraction of the original feed mass, which is precisely the point: it lets a project size its downstream leach circuit for that reduced tonnage rather than the full plant throughput. Understanding that mass reduction ratio for a specific dump, established through the same testwork program that sets reagent dosages, is what actually lets a project size its leach tankage correctly rather than over- or under-building it.