GeaSpirit
EN

Recovery Technologies

Tailings reprocessing, mineral separation and secondary resource recovery

This page answers the question that comes after “there is something in this waste”: how would anyone actually get it out? It is written for people who will judge the answer — process engineers, metallurgists, geologists, regulators and the owners of the material. There is no universal technology for mine waste, and the reason is not that nobody has invented one yet.

A recovery route is decided by the material, not by the element in the assay: mineralogy, particle size, liberation, mineral deportment, moisture, density contrast, magnetic and electrical response, surface chemistry, and what the historical plant already did to it. A technology that works beautifully on one waste body is useless on the one next to it.

mine waste AS DEPOSITED characterisation SIZE · MINERALOGY LIBERATION physical response DENSITY MAGNETIC ELECTRICAL OPTICAL / SENSOR physical preconcentration barren fraction REJECT · OTHER USE selective wet or chemical treatment ONLY THE ENRICHED FRACTION recovered material water and reagent recycle THE ORDER IS THE ARGUMENT — EVERY STEP DECIDES WHAT THE NEXT ONE EVEN SEES
The general shape of the problem. Characterisation comes first because every later step is decided by it, and the physical stage comes before the chemical one wherever it can — not because dry is virtuous, but because treating less mass is cheaper in every currency at once.

Liberation, before anything else

This is the single idea that decides whether any physical separation can work, and it is the one most often skipped. A separator does not act on elements. It acts on particles, and it can only sort a particle by a property that particle actually has.

A · LIBERATED the valuable mineral is a grain of its own. A separator can act on it: it has its own density, its own magnetic response, its own surface. B · LOCKED the same mineral is enclosed in gangue. The separator sees ONE particle with the average property of the composite — and sorts it as gangue. An element being present is not the same as a mineral being available.
The same mineral, twice. On the left it is a grain of its own, with its own density, magnetic response and surface — a separator can act on it. On the right it is enclosed in gangue, and the separator sees one composite particle with the average property of the whole thing. It sorts it as waste, correctly, by its own logic.
This matters most in exactly the material we care about.

A flotation tailing contains what the plant failed to recover — and one of the commonest reasons a mineral is not recovered is that it was never liberated in the first place. So the material most likely to still hold something is also the material where physical preconcentration is least likely to work. Assuming otherwise is how a project rejects seventy per cent of its mass and seventy per cent of its metal in the same step.

Breaking rock, and where the energy goesphysical

Comminution consumes the majority of the energy a concentrator uses and a few per cent of all the electricity generated on the planet. Almost none of it goes into breaking anything: most is lost as heat and noise, and a large part of what remains breaks rock that did not need breaking. For waste reprocessing this is the difference between a project and a rounding error, because the metal is dilute and the tonnage is enormous.

Two facts drive every decision here. Grinding finer liberates more mineral and always costs more energy. And material that has already been through a mill once — which is what tailings are — does not need the coarse stages at all.

ENERGY PER TONNE RISES FASTER THAN FINENESS 020 40kWh/t 100 mm10 mm 1 mm100 µm 10 µm crushers HPGR tumbling mills stirred / ultrafine mills tailings already start here
The energy curve, and why reprocessing starts halfway along it. A tailing has already paid for every stage to the left of the dashed line. That is a real, permanent advantage of waste over fresh ore — and it is separate from, and much more reliable than, any claim about grade.
High-pressure grinding rollscommercial Two counter-rotating rolls crush a compressed bed rather than single particles. More efficient than a tumbling mill and it induces micro-cracks along grain boundaries, which makes the downstream leach faster.
Stirred / ultrafine millscommercial Fine media agitated by an impeller. The only economic way below about 20 µm, and the enabling step for locked or finely disseminated minerals. Applied to a concentrate, not to the whole feed.
High-voltage pulse fragmentationpilot Microsecond electrical discharges through the rock break it preferentially along mineral boundaries, liberating at a much coarser size. Genuinely selective; throughput and electrode life are the open questions.
Microwave pre-treatmentpilot Some minerals absorb microwaves and their neighbours do not; the differential expansion cracks the boundary between them. Reduces the grinding work needed afterwards.
Grinding finer is the default answer and usually the wrong one for waste.

It is the one intervention guaranteed to improve liberation, which makes it the reflex. But energy per tonne rises steeply as size falls, fine particles float and settle badly, and the residue becomes a slurry that needs a dam — reintroducing the exact liability the project was meant to remove. Grinding the whole feed finer to chase a low-grade metal is how reprocessing studies die. Grinding a small pre-concentrate finer is how they survive.

Wet or dry?

Both, and the choice is not ideological. Water is not a defect — it is what keeps fine particles apart so that a separator can act on them individually.

WET — THE PARTICLE IS CARRIED BY A FLUID flotation · gravity in slurry · leaching bioleaching · wet magnetic · classification HANDLES FINES water keeps particles apart, so a separator still acts on individual grains at 10 µm. COSTS pumping · thickening · filtration · reagents effluent · and the residue still has to go somewhere afterwards. DRY — THE PARTICLE IS ACTED ON DIRECTLY screening · air classification · dry gravity dry magnetic · electrostatic · sensor sorting STRUGGLES WITH FINES below roughly 20 µm, surface forces exceed the field: grains stick together and the separator sorts agglomerates, each carrying the average property of what is inside it. AND TAILINGS ARE USUALLY FINE which is why "just do it dry" is not an answer.
Two families, and the honest trade. Wet processing handles fines and brings pumping, thickening, filtration, reagents and effluent with it. Dry processing avoids all of that and runs into physics below roughly twenty micrometres, where surface forces exceed the separating field. Most tailings are fine.

Flotation

Reagents make the surface of one mineral water-repellent; air bubbles carry it to a froth and everything else stays behind. It remains the workhorse of sulphide processing and is actively studied for reprocessing.

FROTH SLURRY feed + reagents collector · frother · depressant concentrate tailings Air is blown through. A mineral whose surface the collector has made water-repellent attaches to a bubble and rises. Everything else stays in the slurry.
A flotation cell, conceptually. The selectivity is entirely in the surface chemistry — which is also why a surface that has weathered for fifty years in a dam is not the surface the original circuit was designed around.
Historical flotation does not mean reflotation is pointless.

Material is left behind for reasons that have nothing to do with the method being wrong: poor liberation, the reagent chemistry available at the time, the equipment of the period, a different target mineral, oxidation, or an economic cut-off that has since moved. Decades later the same family of technology can work. Historical use is a reason for caution, not a reason to rule a route out.

Gravity separation

Denser particles settle, are thrown or are carried differently. Simple, cheap in energy terms, and completely dependent on liberation.

FEED dense — settles early light — carried further A density difference between MINERALS does nothing until the valuable mineral is a grain of its own.
Density does the sorting. A density contrast that exists between two minerals means nothing until the valuable one is a grain of its own.

Magnetic separation

Wet or dry, at low, high or high-gradient field. Intensity is a dial rather than a switch, and it reaches progressively weaker responses.

DRUM FEED magnetic — held to the drum non-magnetic — thrown clear Intensity is a dial, not a switch: low, high and high-gradient fields reach progressively weaker responses.
Response, not composition. Iron in a chemical analysis is not a magnetic mineral. The phase decides — and the phase is not in any assay.

Electrostatic and triboelectric separation

Dry, and for that reason perennially attractive. Particles acquire opposite charges by contact and an electric field pulls them apart.

TRIBOCHARGING +− fraction Afraction B Particles charge by contact, then an electric field pulls them apart. It is dry, and it is the reason the fine-particle problem matters: an agglomerate carries the average charge of what is inside it.
Where the fine-particle problem bites hardest. Below roughly twenty micrometres, particles agglomerate, and an agglomerate carries the average charge of what is inside it. The separator then sorts clumps rather than minerals.

Sensor sorting

Each particle is measured — by X-ray transmission, near-infrared, optical or other sensing — and ejected mechanically.

SENSOR ejectedpasses Each particle is measured and mechanically ejected. It needs particles large enough to measure one at a time — which is why it suits waste rock and stockpiles far better than fine tailings.
It needs particles it can see one at a time. Excellent for waste rock, stockpiles and preconcentration ahead of grinding; far less useful on fine tailings.

Leaching and hydrometallurgy

Dissolve the target selectively, then recover it from solution — by precipitation, solvent extraction, ion exchange or membranes. Acid, alkaline, cyanide and chloride systems all belong to this family, as does bioleaching, where microorganisms do the mobilising.

solid + lixiviant selective dissolution solid–liquid separation purification metal recovery residual solid Solubility belongs to the MINERAL PHASE, not to the element. The same element in two minerals behaves differently.
Selective dissolution. Solubility belongs to the mineral phase, not to the element: the same element in two different minerals behaves completely differently.

Bioleaching can lower some chemical inputs, but it is slow, strongly material-dependent and sensitive to temperature and pH. It is not green by definition, and it is not fast.

Coarse particle recoverysurface chemistry

Conventional flotation loses particles above roughly 150 µm: a grain that heavy tears off the bubble in a turbulent cell. The industry answered that for a century by grinding finer, which is where most of the energy in a concentrator goes. Fluidised-bed flotation removes the turbulence instead — the particle sits in a quiescent rising water column, so a bubble that attaches is not shaken loose — and recovers composite grains three to five times coarser.commercial

For waste this matters more than it does for fresh ore. A coarse reject can be dewatered by screening alone and stacked dry, so a plant that recovers coarse produces a residue that does not need a dam. The trade is selectivity: coarse composites carry gangue with them, so the product is a pre-concentrate, not a final concentrate.

RECOVERY vs PARTICLE SIZE 0%75% 100% 1045 150400 850 µm conventional cell fluidised bed the coarse window a dam-free residue needs Both curves fall away at the fine end for the same reason — a grain too small to carry a bubble — and the
Why coarse recovery changes the residue, not just the yield. Recovering at 400 µm instead of 100 leaves a reject coarse enough to dewater by screening. The curves are schematic: the crossover size is a property of the specific mineral and its liberation, not a constant, and no facility in this warehouse has the size-by-size data that would place it.

Bio-processingbiological

Certain bacteria and archaea make a living oxidising iron and sulphur. In doing so they regenerate ferric iron and acid, and ferric iron dissolves sulphide minerals. The organisms do not attack the metal — they maintain the chemistry that does, continuously and without a reagent bill.

Heap bioleachingcommercial Irrigated heaps of low-grade sulphide, worked at ambient temperature over months to years. Operated at full scale for copper and for nickel–zinc. The natural fit for waste, because a dump is already a heap.
Tank biooxidationcommercial Stirred, aerated, temperature-controlled. Used to crack refractory gold ores where the gold is locked inside arsenopyrite or pyrite — the bacteria open the sulphide, and conventional leaching then reaches the gold.
Sulphidogenic bioreactorsdemonstrated Sulphate-reducing bacteria turn drainage into hydrogen sulphide, which precipitates metals as sulphides that can be sold. Turns a liability stream into a product stream.
Phytomining / agrominingpilot Hyperaccumulator plants concentrate nickel or cobalt into biomass that is then burned to a bio-ore. Slow and land-hungry, but it needs no plant and can run on ground too poor for anything else.
THE IRON CYCLE — THE BACTERIA NEVER TOUCH THE METAL bacteria oxidise Fe²⁺ → Fe³⁺, and S → H₂SO₄ sulphide mineral CuFeS₂ · ZnS · FeAsS Fe³⁺ attacks Fe²⁺ regenerated metal in solution to SX / IX air + water No reagent is consumed to do the oxidising: the organisms rebuild the oxidant as fast as the mineral spends it. What they cost is TIME — months to years, not hours.
Why bioleaching is cheap and slow at the same time. The cost that disappears is the oxidant; the cost that appears is residence time and the area to hold the material while it reacts. For a dump that has already sat for eighty years, time is the one input that is genuinely free.
Bioleaching is not a universal solvent.

It works on sulphides. Oxide, silicate and carbonate hosts are largely untouched by it, and a tailing that has weathered in air for decades may have lost the very sulphide the organisms need. Acid-consuming gangue — carbonate especially — can neutralise the system faster than the bacteria can acidify it. Whether a given dump still holds leachable sulphide is a mineralogical question, and it is exactly the question no assay answers.

Choosing the lixiviantchemical

Leaching is not one technology. The solvent decides the cost, the permit, the residue and half the public objection, and the last twenty years have produced real alternatives to the two reagents the industry was built on — cyanide for gold, sulphuric acid for copper.

LixiviantTargetsConditionsWhy it is chosenWhat it costsStatus
Sulphuric acidCu, U, Ni, Zn oxidesambient, pH <2 cheap, understood, works in a heap consumed by carbonate gangue; leaves an acid residue commercial
CyanideAu, Agambient, pH 10–11 unmatched selectivity for gold at very low grade toxicity, permitting, destruction circuit, social licence commercial
ThiosulfateAuambient, mildly alkaline no cyanide; handles carbonaceous ore that robs cyanide reagent loss and recycle complexity commercial
GlycineAu, Cu, Ni, Co, Znambient–60 °C, alkaline an amino acid: benign, recyclable, does not attack silicate gangue slower; needs a recovery step that returns the glycine demonstrated
ChlorideCu (chalcopyrite), Au, PGMambient–elevated, saline attacks chalcopyrite, which acid heaps famously do not corrosion; materials of construction demonstrated
Ammonia / ammoniacalCu, Ni, Coambient, alkaline ignores carbonate gangue that would eat acid volatility, containment, odour commercial
Deep eutectic solventsbase & precious metalselevated tunable selectivity, very low volatility reagent cost and recycle unproven at scale laboratory

Glycine deserves the attention it is getting. It is alkaline, so it does not dissolve the silicate and carbonate gangue that makes acid leaching of low-grade waste uneconomic — the reagent goes to the metal instead of to the rock. It is recoverable and re-usable, non-toxic and biodegradable, and it will take copper, gold, nickel, cobalt and zinc. That combination is unusual. It is also slower than the reagents it would replace, and the economics turn on how completely the glycine is recovered rather than on how fast the metal dissolves.

SULPHURIC ACID ON A CARBONATE-BEARING WASTE consumed by carbonate and silicate gangue neutralised to metal GLYCINE, ALKALINE, ON THE SAME MATERIAL losses recycled to metal The bar is the reagent, not the metal. Acid does not fail on a carbonate-rich waste because it cannot dissolve the target — it fails because the rock spends it first, and at 0.2% copper there is a great deal more rock than copper. An alkaline lixiviant simply is not offered that reaction. Proportions are illustrative of the mechanism. The real split is set by the acid-consuming mineral content of the specific material, which is a measurement — and not one GeaSpirit holds for any facility.
The reagent budget, which is what actually decides a low-grade project. Recovery percentages get quoted; reagent consumption per tonne of feed is what closes or kills the economics of waste.

Getting the metal out of the solutionsolution recovery

Leaching produces a dilute, dirty liquor, not a product. Every heap-leach operation on Earth is followed by a second plant whose only job is to concentrate and purify what the first one dissolved — and for a reprocessing project this half is often the larger risk, because waste liquors carry impurities that fresh ore liquors do not.

SX–EW — THE ORGANIC IS A FERRY, NOT A CONSUMABLE leach liquordilute · impure EXTRACTIONmetal → organic STRIPPINGorganic → acid ELECTROWINNINGcathode metal loaded organic strippedorganic back raffinate back to the heap 99.99% Cu The organic phase is loaded on one side and unloaded on the other, then returned. Selectivity lives in the extractant molecule: it is chosen to bind ONE metal and reject the iron, aluminium and manganese that dominate a waste liquor by mass. For waste specifically: impurities the original ore never had can poison the organic, and that is a testwork question, not a literature one.
Solvent extraction and electrowinning. The step most often left out of a reprocessing concept, and the one that turns a dissolved metal into something saleable. Ion exchange and resin-in-pulp do the same job with a solid resin instead of an organic liquid, and suit low concentrations and slurries where SX struggles.
Solvent extraction (SX)commercial An organic extractant selectively binds one metal, then releases it into clean acid. The backbone of copper hydrometallurgy.
Ion exchange / RIPcommercial A solid resin does the same. Tolerates slurry and very dilute liquor — the usual condition when leaching low-grade waste. Standard for uranium and gold.
Sulphide precipitationcommercial Add sulphide, drop the metal out as a solid. Crude but robust, and the natural partner to a sulphidogenic bioreactor.
Electrowinning / electrorefiningcommercial Current plates the metal onto a cathode at sale-ready purity. Needs a concentrated, clean electrolyte, which is why it is always the last step, never the first.
Selective adsorbentspilot Engineered sorbents pull one ion from a complex brine. Developed hard for lithium; the same chemistry is being pointed at cobalt, nickel and rare earths.
Membranes / nanofiltrationpilot Separates by size and charge with no reagent at all. Attractive for dilute streams, limited by fouling on real mine water.

Thermal treatment, and what it destroys

Roasting and smelting change the phase an element sits in. That is the point — and it is also permanent. A carrier mineral may simply not exist any more in the material a roaster produced, which constrains what can ever be done with that residue.

This is where knowing the historical flowsheet stops being background and becomes a constraint. But it applies to the stream that went through the stage, not to every waste body at the same mine: a plant that roasted a flotation concentrate left the flotation tailings untouched.

Water and residue management

wet treatment thickening filtration dry stack / paste recovered water returns to the circuit Water management and metal recovery are related problems and not the same problem.
A closed loop. Thickening, filtration, paste and dry stacking are how a wet circuit stops being a water problem. Related to recovery, and not the same problem.

Mineral deportment — the number that matters more than the assay

THE ASSAY SAYS — 40 mg/kg OF THE ELEMENT one number for the whole sample DEPORTMENT SAYS — WHERE IT ACTUALLY IS 5% mineral A 80% mineral B 15% unidentified fines Two facilities can report the same concentration and need completely different processing routes, because the route acts on the MINERAL, never on the element. If 80% sits in a mineral no separator can reach at this grain size, a high assay means very little. This is measured by automated mineralogy on physical material. No register publishes it.
Where the element actually sits. Two facilities can report an identical concentration and need completely different routes, because every route acts on the mineral and never on the element.
GeaSpirit R&D concept — a hypothesis to test, not a technology

Selective hybrid recovery

Rather than chemically treating an entire waste mass to recover a small fraction of it, first establish whether cheap physical separation can reject a substantial barren fraction, and treat only what is left.

mine waste dry / low-water preconcentration IF LIBERATED bulk barren fraction smaller enriched fraction selective wet or chemical treatment recovered ONLY IF the physical stage does not reject the valuable mineral with the barren fraction — in a flotation tailing the mineral is often locked, which is exactly why it was left behind in the first place.
Dry first, wet only on what survives. The economic case is that less mass means less water, fewer reagents, less energy downstream and smaller equipment.

The condition it depends on, and it is not a detail: the physical stage must not reject the valuable mineral along with the barren fraction. In a flotation tailing the mineral is often locked — which is precisely why it was left behind — and locked grains follow the gangue. Whether this works for any given waste body is a question for a separation test on real material, not for a model.

We publish no mass-rejection figure, no recovery percentage, no water saving and no cost, because none has been measured.

Value that is not a metalphysical & chemical

A waste facility can pay for its own removal without anyone recovering a metal from it. These routes are unglamorous, they move enormous tonnages, and for a low-grade dump near a market they are frequently the only ones that close.

Aggregate and construction fillcommercial Coarse waste rock is often a competent aggregate. Value per tonne is very low, so the distance to the buyer decides everything: haulage beats grade.
Supplementary cementitious materialcommercial Some tailings and slags are pozzolanic and replace part of the clinker in cement. The buyer is a cement plant with a hard carbon target, which is what makes this one move.
Geopolymers and alkali-activated bindersdemonstrated Aluminosilicate tailings activated into a binder with a fraction of the emissions of Portland cement. Consumes the fine fraction nobody else wants.
Carbon mineralisationdemonstrated Magnesium silicate tailings — ultramafic, so nickel, diamond and chromite operations — react with CO₂ to form stable carbonate. Permanent storage in material that is already crushed, already on surface, already owned.
Paste backfillcommercial Return the tailings underground as structural fill. Removes surface liability and buys ground support at the same time.
Soil manufacturepilot Blended with organic amendment into a growing medium for closure. Depends entirely on the material not being acid-generating or metal-mobile.
TWO DIFFERENT BUSINESSES, OFTEN ON THE SAME DUMP value€/t tonnage moved metal recovery high €/t · small mass bulk valorisation low €/t · enormous mass carbon credit price is policy, not ore The same facility can sit in more than one of these. A project that only asks "what metal is in it?" never sees the other two.
Why "is there metal in it?" is the wrong opening question. Bulk routes are decided by distance to a buyer and by the material's physical properties, not by grade — so a dump with no economic metal at all can still have a route, and a rich one far from anywhere may not.

Recovery is not the only objective

A route recovering 95% at very high cost can be worse than one recovering 75% with a far smaller processing burden. The eventual measure has to weigh metal recovered against mass processed, energy, water, reagents, residue generated, capital and operating cost, and what the permits allow — and in the end express it per kilogram of material actually recovered.

We do not calculate that today, for any asset, because it needs test data that does not exist yet.

Who actually does what in this industry

Three different businesses get spoken about as if they were one, and confusing them is how a reprocessing idea ends up addressed to the wrong company.

RoleWhat it sellsHow it earnsExamples
Equipment & flowsheet suppliers Engineering, machines, testwork, automation, service. Crushing, grinding, classification, flotation, thickening, filtration, tailings handling — the whole plant. Selling and servicing the plant. They do not usually take ownership of a waste body or of the metal recovered from it. Metso, FLS (formerly FLSmidth)
Retreatment operators Nothing. They own or lease the residue and sell the metal. Waste in, metal out, at enormous scale and very low unit cost. DRDGOLD, Pan African Resources
Selection layer Which waste bodies are worth anyone's attention, on what evidence, and what should be measured next. Deciding where the other two should be pointed. where GeaSpirit sits

This page exists because of the third row, not the first. GeaSpirit does not design plants and does not intend to. Understanding what happens after screening is what makes a screening judgement worth anything — recommending a facility without knowing whether any route could plausibly treat it is how a shortlist becomes a liability.

The economics: cost per tonne is the adversary

A primary mine can carry an expensive flowsheet because the feed is rich. A waste body usually holds a small amount of value spread through millions of tonnes, and there the arithmetic runs the other way: nothing can be spent per tonne that the tonne does not return. Everything below is a way of spending it.

RegrindingThe single largest energy cost, applied to the largest possible mass. The reflex answer to poor liberation, and the commonest way a reprocessing study dies.
ReagentsA consumption of a few kilograms per tonne is invisible on a spreadsheet and enormous across a hundred million tonnes.
Stage countEvery unit operation is capital, maintenance, pumps, pipework and people. Eight stages is a plant; three is a business.
WaterSourcing it, moving it, treating it, and the residue it leaves — a slurry that needs a dam reinstates the liability the project set out to remove.
HaulageLow-value material does not travel. For bulk routes the distance to the buyer decides the project outright.
FinenessMaterial that is already very fine cannot be dewatered cheaply and floats and settles badly. It is the hardest starting point there is.

The operators who have made this work did not find an exotic technology. They combined enormous throughput, a conventional and well-understood circuit, mechanisation, short pumping distances and a low reagent bill — publicly citing the absence of regrinding and low reagent consumption among the reasons a low-grade retreatment plant stands up at all. The lesson is not that reprocessing needs better chemistry. It is that it needs fewer things happening per tonne.

WHAT THE TONNE RETURNS, AND WHAT IT COSTS TO TREAT € / t very low gradehigher grade value of the tonne complex flowsheet grind · float · leach · SX · EW simple flowsheet screen · magnetic or gravity · product a simple route pays here… …a complex one only from about here Cost is nearly flat with grade; value is not. So the question a waste body has to answer is not "how much metal", but "how few things must happen to it".
The crossing points, not the curves, are the decision. Both lines are schematic — their heights are properties of a specific material, a specific site and a specific year. What is not schematic is their shape: treatment cost barely falls as grade falls, which is why the cheap route reaches down much further than the sophisticated one.

The question worth asking

Two waste bodies can assay the same and be worth entirely different things.

100 ppm · hard Ultrafine, the element locked inside a refractory host. Needs fine grinding, acid, heat and a solution circuit. High grade, and probably not a project.
70 ppm · easy Coarser, the element carried in a liberated mineral with a strong physical contrast. Screen, separate, sell. Less metal, and a far better prospect.

So the useful question is not can a better technology be invented? — that costs years and millions, and Metso and FLS are already doing it. It is:

can a waste body be found where a simple technology that already exists is enough?

That is a search problem over evidence, which is what GeaSpirit is. The asset worth finding is not the richest one. It is the intersection of large tonnage, a valuable enough element, a known carrier mineral, decent liberation, a strong physical contrast to separate on, little regrinding, few reagents, recyclable or no water, and a buyer within haulage distance. Miss any one of those and grade will not rescue it.

And the honest limit.

Liberation and mineral contrast are measurements, not inferences. A carrier mineral reported at the source deposit tells you what to test for; it does not tell you the grain is free, or that it will answer to a magnet in this material. Until that is measured, the correct value for how hard a residue would be to treat is unknown — and GeaSpirit records it as unknown rather than as promising.

The selection map

Nothing on this page is chosen from a menu. The route follows from what the material is, and the diagram below is the order in which the questions have to be answered. Read left to right: each branch is a measurement, not a preference.

EACH BRANCH IS A MEASUREMENT Grain size ofthe host mineral Is it a sulphide?or oxide / silicate Density or magneticcontrast? Acid-consuminggangue? coarse and liberated sensor sorting · coarse flotation locked and fine → ultrafine grind first sulphide flotation · bioleaching · POX oxide / silicate → leach, not float contrast present gravity · magnetic · dense medium no contrast → chemical route only carbonate-rich alkaline lixiviant · glycine · ammonia low carbonate → acid is viable ROUTE FAMILY not a flowsheet — a family of flowsheets worth testing on real material Every box on the left is something GeaSpirit either holds as measured evidence or records as absent. None of them is inferred from an assay.
Route families follow from mineralogy, not from grade. This is why the Evidence Passport treats mineralogy as a pillar in its own right rather than as supporting detail: without it, an assay cannot even narrow the choice.

Maturity, stated plainly

A page that lists a laboratory result beside a technology treating a hundred thousand tonnes a day, in the same typeface, is misleading by layout alone. Every route named above carries one of these four labels, and they mean what they say.

LabelWhat it meansWhat you may conclude
commercial operating at production scale, on real feed, for years engineering risk is manageable; the question is whether it suits this material
demonstrated proven at industrial or large pilot scale, limited number of installations viable, but expect to fund the learning curve
pilot continuous operation beyond the bench, not yet at production scale promising; do not build a business case on it alone
laboratory batch results under controlled conditions interesting; treat published recoveries as an upper bound

The labels describe the technology, never a facility. A commercial technology applied to an untested material is still an untested proposition.

Can one process do everything?

It is the right question to ask and the answer is no — not for a reason of engineering effort, but for one of physics. What can be recovered from a material is governed by deportment: which mineral actually hosts the metal, how big those grains are, and what they are locked inside. Two dumps assaying an identical 0.3% copper need opposite flowsheets if one carries it as chalcocite at 200 µm and the other as chalcopyrite locked in silicate at 15 µm. No solvent, no cell and no sorter changes that, because the difference is in the rock and not in the machine.

So a universal process is not available. A universal sequence is — and that is where the real gain of the last decade has been.

the architecture worth pursuing

Reject early · concentrate coarse · leach small

Every expensive step in a plant scales with the mass that reaches it, and every low-grade waste project dies because too much mass reaches the expensive steps. The winning move is not a better reactor. It is arranging for 90–98% of the feed to never see one.

Stage 1 — reject. Sensor sorting at coarse size, on a conveyor, no water. Throw away what demonstrably carries nothing. Cheap per tonne precisely because it is crude.

Stage 2 — pre-concentrate coarse. Coarse-particle flotation or enhanced gravity, at a size that dewaters by screening. The reject leaves as a dry stack rather than as a slurry needing a dam.

Stage 3 — grind only the concentrate. Ultrafine milling applied to 2–10% of the original mass instead of all of it. This single reordering is what makes the energy arithmetic survive.

Stage 4 — leach small, benign, and recycle the reagent. A small, concentrated stream is where an alkaline lixiviant such as glycine becomes affordable, and where recovering the reagent is worth the equipment.

Stage 5 — recover from solution. Ion exchange or SX, then electrowinning.

Every one of those five stages is commercially available today. Nothing here is invented. What is genuinely modern is the order, and the discipline of sizing the plant around rejection rather than around throughput. It is polyvalent in the way that matters: the sequence holds for copper, gold, zinc, cobalt and rare earths, while the settings inside each stage are tuned per site. It is modular, so it can be skid-mounted and moved between dumps that are each too small to justify a fixed plant — which is the actual economic shape of mine waste.

MASS THROUGH THE CIRCUIT — THE BAND IS THE FEED 100%feed ~40% ~12% ~4% ~1% 1 · SORTdry · coarseno water 2 · PRE-CONCcoarse floator gravity 3 · FINE GRINDonly what gotthis far 4 · LEACHbenign lixiviantrecycled 5 · RECOVERIX / SX → EWcathode rejected material — coarse, dewatered by screening, stackable dry, no dam required The point is not that any single box is new. It is that the costly boxes are the small ones, and that the residue leaving the circuit is coarse and dry rather than a slurry — so the plant does not create the liability it was built to remove. Percentages illustrate the principle. Real rejection rates are a property of the material and are established by testwork, not assumed.
Where the mass goes, which is where the money goes. Conventional plants are sized on throughput. A waste circuit has to be sized on how much it can credibly throw away in the first ten metres.
what GeaSpirit could genuinely contribute

The bottleneck is selection, not metallurgy

Reprocessing projects rarely fail because the flowsheet was wrong. They fail earlier: money is spent on testwork at a site that was never a candidate, because the only cheap information anyone had was an assay — and an assay cannot distinguish the two dumps in the first paragraph of this section.

The gap is a cheap, standardised, deportment-first screen: automated mineralogy on a correctly taken sample, a sequential extraction to show what is actually soluble, and a size-by-size assay to show where the metal sits. Together they identify the family of flowsheet a material belongs to, before anyone commissions a pilot. That is small money against a testwork programme, and it is the measurement that decides whether the testwork is worth doing at all.

GeaSpirit does not build plants and does not intend to. What it can do is establish, from public evidence, which facilities are worth that screen — and say honestly when the remote evidence is exhausted and the next real step requires touching the material. That is a selection instrument, and selection is the step the sector currently does worst.

What this page will not claim.

That GeaSpirit has invented a unit operation; that any of these routes is proven on any specific facility in the warehouse; or that a percentage recovery can be predicted from anything short of testwork on real material. Every number drawn on this page is illustrative of a mechanism. None of them is a measurement of a place.

What GeaSpirit does with this — and what it refuses to do

GeaSpirit does not select a process for a waste body. Choosing a flowsheet remotely would need particle-size distribution, liberation, mineral deportment, magnetic and electrical response and bench recovery data, and none of that is published by any register anywhere: it comes from automated mineralogy on physical material.

What the evidence can do is eliminate. So each route on each facility carries a state, privately, and none of them is a recommendation:

OpenNothing currently known rules it out. Not recommended, not proven, not economic.
DisfavouredReal evidence raises a real concern — and a concern is not a proof of impossibility.
ExcludedDocumented physical or chemical incompatibility, on this stream. A high bar, rarely met.
UnknownThe route turns on a property nobody has measured here. Today this is most of them.
TailingsIntelligence EvidenceCompass EvidencePassport RecoveryConstraints Next BestTest physicalcharacterisation Recovery Router NOT BUILT — NEEDS TESTS WHAT WE KNOW → WHAT IS STILL PLAUSIBLE → WHAT TO MEASURE → AND ONLY THEN, WHAT TO DO The last box is dashed on purpose: choosing a route needs data that comes only from physical material.
Where a route decision would sit, if we had the data. The last box is dashed because it does not exist: comparing routes needs measurements that only physical material can give.
Evidence PassportWhat is known about one facility, what is inferred, and what is missing.
Recovery ConstraintsWhich routes the evidence still leaves plausible for that facility — private, and never a recommendation.
Next Best TestThe smallest measurement that would move the decision.

What has to be measured before any route can be chosen

Particle size distributionDecides which separators can physically act on the material at all.
MineralogyWhich minerals are present, in what proportions.
LiberationWhether the valuable mineral is a grain of its own or locked in gangue.
Mineral deportmentWhich mineral carries the element, and at what grain size.
MoistureGoverns whether any dry route is workable.
Bulk densityTurns a volume into a mass — and it is almost never measured.
Magnetic susceptibilityResponse, as opposed to iron content.
Surface chemistryWhat decades of weathering did to flotation behaviour.
Element concentrationSolid phase. A leachate value is an environmental result, not a grade.
RepresentativityWhether the samples can speak for the whole body, or only for the spot they came from.
Historical processWhat the plant targeted, and what it left behind.
Bench recoveryThe only thing that settles recoverability. Nothing on this list substitutes for it.

Three situations, and why they are not the same

Coarse waste rock

Particles large enough to be measured individually. Sensor sorting and physical preconcentration are genuinely plausible, and the question is grade and tonnage.

Fine flotation tailings

Fine, and probably poorly liberated — that is often why the material is there. No route can be chosen without mineralogy and liberation data, and a dry preconcentration assumption is exactly the wrong instinct.

Residue after thermal treatment

The mineral phases have changed. The historical process becomes the decisive evidence for what remains possible, and it applies to that stream alone.

Sources

Where a limit depends on the material — and most do — this page says so rather than quoting a universal number. No figure here is presented as applying to any particular waste body.