The Four Stages of Mineral Processing Explained: From Crushing to Dewatering

A concentrator treating 10,000 tonnes of ore a day often ships less than 400 tonnes of product. The other 9,600-plus tonnes are rock and water that must be broken, sorted, separated and drained in a fixed sequence - and that sequence is what plant designers call the four stages of mineral processing. Miss one link in the chain and the whole plant pays for it: a mill that grinds the wrong size starves the flotation cells, and a thickener that settles poorly sends water costs climbing for years.

The short answer first. The four stages of mineral processing are: (1) comminution, which covers crushing and grinding; (2) sizing and classification; (3) concentration; and (4) dewatering. Every flowsheet - copper porphyry, zinc sulfide, phosphate rock, potash brine or lithium-bearing ore - is built from these four stages arranged, looped and tuned around one goal: liberate the valuable minerals first, separate them second, and get the water out last.

The order is not negotiable, because each stage hands a specific product to the next. Flotation cannot recover particles that are still locked inside rock, so size reduction must come first. Grinding cannot hold a stable target size without classification closing the loop, so sizing sits between grinding and separation. And a wet concentrate is neither sellable nor cheap to move, so dewatering always closes the flowsheet. When a plant underperforms, experienced metallurgists diagnose the handoffs between stages before they touch the reagent bench.

The Four Stages of Mineral Processing at a Glance

Walk any well-run concentrator from the primary crusher to the concentrate shed and you will pass through the same four stages in the same order. The machines change with the ore - a potash plant on the Thai coast and a zinc plant in Inner Mongolia share the same logic even though their equipment differs. The strip below shows the typical journey of a particle of ore.

Two loops hide inside that line. Classification sends oversized particles back to the mill, so a grinding circuit commonly recycles 150-350% of its new feed as circulating load. Concentration sends middlings back for re-treatment, so a flotation circuit reprocesses cleaner and scavenger tailings several times before they escape. A flowsheet is therefore less a straight line than a controlled set of loops feeding a straight line - and the loops are where most design mistakes hide.

You will see this work called mineral processing, ore dressing or beneficiation - the terms describe the same discipline. What varies is emphasis. A placer gold operation may run only screening, gravity separation and dewatering, skipping grinding entirely because nature already liberated the gold. A refractory gold ore may add leaching after the four stages. Treat the four-stage framework as the backbone, then add or remove limbs to match the ore in front of you.

Stage 1: Comminution - Crushing and Grinding

Comminution exists for one reason: liberation. Ore leaves the mine as pieces of rock in which the valuable mineral is physically locked inside waste. No separator on the market can pick a crystal of chalcopyrite out of a 20-centimetre lump of porphyry. Stage one breaks that lump until the value minerals stand alone, and everything downstream - recovery, reagent cost, thickener sizing - depends on how well this is done.

Why liberation sets the target

Liberation size is a property of the ore, not of the equipment. In many porphyry copper ores, chalcopyrite grains measure roughly 50 to 150 microns across, which is why flotation feed is typically ground to a P80 - the screen aperture that 80% of the material passes - between about 45 and 150 microns. Grind coarser and particles stay locked, so recovery falls. Grind finer than the ore actually needs and you create slimes below about 10 microns, which float slowly, consume extra reagent and settle badly in the thickener. The entire economics of the plant sits between those two failure modes, which is why laboratory grind-recovery testwork is the first item on any serious project plan.

Grinding is the single largest power consumer in most plants, and industry estimates commonly attribute a few percent of global electricity generation to comminution across all mines. That cost explains a rule every mill superintendent knows: crush as much as possible, grind as little as necessary. Breaking rock in a crusher costs a fraction of achieving the same size reduction in a mill.

How crushing and grinding divide the work

Open-pit ore can arrive in blocks close to a metre across. Crushing takes it down in stages: a primary jaw or gyratory crusher to roughly 150-250 mm, a secondary cone crusher to 25-75 mm, and often a tertiary pass to 6-19 mm. Each machine achieves a reduction ratio of roughly 3:1 to 6:1 per pass, which is why plants chain crushers together instead of asking one machine to do everything.

Below roughly 10-12 mm, grinding takes over. The ore is ground in water, and the classic workhorse for first-stage grinding is the wet grid type ball mill. Its lattice discharge plate forces pulp out of the drum at a controlled level, letting coarse particles leave before they are ground past the target. For regrinding and finer second-stage duties, overflow discharge mills hold the pulp longer instead. Choosing between the two is not a catalogue decision; it follows directly from your ore's grind target and work index. If you want a deeper walkthrough of this machine - media charging, liner choice, common faults and maintenance routines - our guide to operating and maintaining a wet grid type ball mill covers the details.

Specifications to pin down before you buy

• Ore hardness: ask for the Bond work index in kWh/t. Soft ores such as limestone sit near 8-12, most porphyry copper and zinc ores near 12-16, and hard taconite-type ores can exceed 20.
• Feed and product size: the maximum feed lump the mill must accept, and the P80 the circuit must deliver, stated in writing.
• Throughput and duty: tonnes per hour for first-stage grinding versus regrind, because the same ore may need different machines for each.
• Steel consumption: grinding media and liner grams per tonne, a quiet cost that in abrasive ores rivals electricity.
• Power per tonne: quote kWh/t at the target grind, not just installed motor kilowatts.

A supplier who answers those five numbers from testwork on your actual sample is worth more than one who quotes a catalogue model twice the size at half the price. Oversized mills do not only waste capital; an underloaded mill breaks media and liners faster and can even grind coarser than a correctly loaded smaller one.

Stage 2: Sizing and Classification

Conclusion first: classification is cheap insurance for everything downstream. The stage itself recovers no metal and sells nothing. What it does is decide, particle by particle, which sizes move forward and which return for more work. When flotation results drift, the first place experienced metallurgists look is usually the cyclone overflow, not the reagent shelf.

Screens, spiral classifiers and hydrocyclones

Two families of equipment do this job. Screens work dry and by aperture, from grizzly screens protecting primary crushers down to fine screens near 0.1 mm. Classification machines work wet and by particle behaviour in water, which makes them the natural partner for ball mills. The spiral classifier uses an inclined trough and a rotating screw: coarse sands sink, the screw rakes them back uphill to the mill feed, and fines overflow to the next stage. It is simple, tolerant of feed surges and easy to maintain, which is why it remains common in mid-size circuits and two-stage grinding flowsheets where cuts are coarser and feed is moderate.

The hydrocyclone has no moving parts at all. Slurry is pumped in tangentially; centrifugal force throws coarse particles to the wall and out through the apex, while fines spiral out of the vortex finder at the top. By changing diameter, pressure, apex and vortex finder sizes, an operator can place the separation cut anywhere from about 5 to more than 100 microns. Cyclone underflow returns to the mill as circulating load - commonly 150-350% of new feed - and overflow becomes flotation feed at a stable P80.

Classification also protects liberation economics. Every particle ground below the target wastes power and becomes a slime that is hard to recover and slow to dewater. Keeping the loop tight - correct cut size, consistent density, wear parts replaced before they distort the cut - is one of the least glamorous and most profitable disciplines in a concentrator. Ask any supplier how their classifier or cyclone package holds its cut under wear, not just on commissioning day; the answer separates equipment vendors from genuine process partners.

Stage 3: Concentration - Separating Value Minerals from Gangue

Concentration is where the money appears. Feed enters at a grade often measured in fractions of a percent and leaves as a product worth shipping: copper concentrate commonly at 20-30% Cu, zinc concentrate above 45% Zn, with recovery deciding whether the mine's business plan actually works. All four stages matter, but this is the one where chemistry, physics and machine design meet in the most visible way.

Froth flotation: the industry workhorse

Flotation exploits a simple surface property: with the right reagents, value mineral particles refuse to be wetted and attach to air bubbles instead, riding them up into a froth that is skimmed off as concentrate. Reagents fall into a handful of roles - collectors make target surfaces hydrophobic, frothers stabilise the bubble blanket, lime or other modifiers set the pH, and depressants or activators switch individual minerals on and off. Before any of it works, the slurry must meet the chemistry properly, which is the job of conditioning: an agent mixing tank or a high-concentration mixing tank gives reagents several minutes of contact time so the first cell receives treated pulp rather than a chemical surprise.

A production circuit runs this as a loop of banks. Rougher cells make a fast, bulk recovery; scavenger cells squeeze the remainder from rougher tailings; cleaner cells re-clean the rougher concentrate to reach grade, with their tailings recycled to the head. Grade and recovery trade against each other at every valve, and the circuit layout - not the operator's reflexes - determines where that trade settles.

Machine choice follows particle size and tonnage. Self-priming SF, JJF and BF type machines draw their own air and slurry and suit medium-size plants that want fewer pumps and simpler layouts. For large cells and coarse or dense feeds, XCF and KYF tank machines run on a common blower: KYF cells deliver strong air dispersion at low power, while XCF cells handle slurry intake between banks so the whole line flows horizontally without auxiliary pumps. Coarse-grain duties call for CLF type machines that keep heavy particles in suspension.

Retrofits show the leverage clearly. In one 3,000-tonne-per-day zinc flotation upgrade in Yunnan, replacing the old banks with larger tank machines and tightening classification gave the metallurgists a stable P80 to work with - and recovery followed. Projects like that zinc flotation technical reform rarely change the ore; they change how consistently the four stages hand off to each other.

When flotation is not the answer: gravity, magnetic and electrostatic separation

Flotation dominates sulfides and many industrial minerals, but three other families earn their place where physics suits them better. Gravity separation exploits density differences in jigs, shaking tables and spiral concentrators; it is cheap per tonne and excels at coarse gold, cassiterite, chromite and heavy mineral sands, but loses efficiency as particles become fine. Magnetic separation is the backbone of magnetite iron ore - drum separators routinely recover magnetite above 90% - and high-intensity versions extend the method to hematite and ilmenite. Electrostatic separation splits conductors from non-conductors in dry beach-sand circuits, separating rutile from zircon where both are non-magnetic.

Stage 4: Dewatering - Thickeners, Filters and Water Recovery

Dewatering looks like housekeeping and behaves like a profit centre. A flotation feed at 30% solids carries more than two tonnes of water for every tonne of ore. If that water leaves with the concentrate and the tailings, the plant must pump, treat and pay for replacements; if it is recovered and returned, the same water serves the circuit for years. In arid mining regions, water recovery is not a sustainability line item - it is the operating licence.

Thickening returns most of the water first

The thickener does the heavy lifting. Diluted slurry enters a large tank, flocculant binds fine particles into fast-settling clusters, and clear water overflows the rim for reuse while rakes push the settled bed to a central discharge. A conventional thickener with a central drive - the configuration used across most mid-size concentrators - typically receives feed at 20-35% solids and delivers underflow at 50-70% solids, returning the overflow as process water. Sizing a thickener is an exercise in settling testwork: the area per tonne per day depends on how fast your flocculated solids actually settle, not on catalogue optimism.

The same logic extends to tailings. Thickening tailings to 55-70% solids before they leave the plant shrinks the tailings facility, cuts seepage risk and keeps water in the circuit instead of buried under a pond. For many new projects, the tailings thickener is now designed at the same time as the concentrate thickener, because both draw on the same settling data.

Filtration and drying: ready for the road

Filtration finishes the concentrate. Vacuum disc and drum filters commonly bring copper or zinc concentrates into the 8-15% moisture band, pressure filters go lower where specifications are strict, and ceramic disc filters shine on fine, high-value concentrates. Most shipped concentrates travel under moisture limits in the 8-12% range set by transport safety rules, so the filter is the last quality gate before the product is sold. Thermal drying is added only where filters cannot reach the specification, because fuel is dearer than filter cloth.

One often-missed point: dewatering performance reaches back upstream. Recycled water carries dissolved reagents and salts, and as the loop closes, water chemistry changes and can alter flotation selectivity. Plants that monitor return-water quality alongside grade and recovery avoid the slow drift that otherwise appears months after commissioning.

How the Four Stages Balance Each Other: A Mass Balance Example

Conclusion first: the four stages are not four separate purchases; they are one mass balance split into four jobs. The fastest way to see it is simple arithmetic on a mid-size copper plant.

Feed of 10,000 tonnes a day at 1.0% copper carries 100 tonnes of contained metal. At 90% recovery, 90 tonnes report to the concentrate. If the concentrate grades 25% copper, daily output is 360 tonnes - and roughly 9,640 tonnes of tailings leave the plant still holding the remaining 10 tonnes of copper. Every one of the four stages shows up in those numbers: comminution cost and grind size set how close to 90% recovery the plant can get, classification stability decides whether that recovery holds for the whole shift, concentration equipment sets the 25% grade, and dewatering decides how much of the more than 23,000 tonnes of water used daily at 30% solids comes back.

Read the interactions in both directions. A finer grind raises liberation and potential recovery but multiplies energy cost and slimes, which then burden both flotation and thickening. A coarser grind saves power and lets the thickeners run faster but leaks value into the tailings. Higher dilution upstream smooths flotation but loads the thickener; thicker underflows ease dewatering but risk coarse particles escaping to the cells. No stage can be tuned in isolation, which is why experienced designers settle the mass balance on paper before a single tank is welded.

A Buyer's Checklist for Equipment at Every Stage

Most procurement disappointments in mineral processing trace back to the same handful of gaps: testwork done on someone else's ore, energy quoted as motor size, and service terms left vague until the first breakdown. The list below turns the four stages into questions you can put to any supplier.

1. Run the assays and mineralogy first. Know which minerals carry value, their grain sizes and how they intergrow with gangue, because liberation size decides the whole flowsheet.
2. Demand testwork on your sample. Bench-scale grinding, settling and flotation tests on your ore - not on a "similar" deposit - are the cheapest risk reduction you can buy.
3. Match machine class to duty. Grid discharge for first-stage grinding and overflow for regrind; spiral classifiers or hydrocyclones to suit your cut; self-priming or blown cell types to suit tonnage and particle size.
4. Compare energy per tonne, not kilowatts. Ask for kWh/t at the target grind and a lifecycle power estimate, since grinding alone can take half the plant's electricity.
5. Price the wear parts. Media, liners, cyclone components, filter cloths and rake teeth are recurring costs; confirm their consumption rates and lead times in writing.
6. Check the water loop. A thickener quoted without settling testwork is a guess; ask for the settling area per tonne per day behind the offer.
7. Put installation and commissioning in the contract. A supplier with real field experience will commit to supervision, commissioning support and operator training, not just a delivery schedule.
8. Ask for reference plants with the same mineral family. A vendor who has commissioned similar ore at similar tonnage will design around problems you have not met yet.

Two of these deserve emphasis. Testwork on your actual ore is non-negotiable, because every rule of thumb in this article bends with mineralogy. And commissioning support separates equipment suppliers from plant builders: the first weeks of a new concentrator set its operating habits for years, and an experienced commissioning crew prevents bad habits from forming.

From Flowsheet to Working Plant: Why the Stages Are Best Planned Together

A flowsheet drawn by one team and equipment bought from several others tends to meet for the first time on site, where interface problems are expensive: a mill discharge that does not suit the classifier feed, a thickener sized for someone else's slurry, a control system that speaks three dialects. Engineering-procurement-construction delivery, usually shortened to EPC, exists to remove those seams. One party carries the mass balance from testwork through commissioning, so the four stages are designed against each other rather than assembled after the fact.

That model suits manufacturers who build the equipment themselves. Our factory in Zhuji, Zhejiang - about 70 kilometres from Hangzhou, with rail and highway links to port - has spent decades on wet-process enrichment and mixing equipment, backs its designs with an ISO 9001 certified quality system and more than 200 production and testing machines, and delivers full mineral processing EPC projects. Recent work spans the entire four-stage chain: XCF-KYF16 flotation machines, 3,000 by 3,000 mixing tanks and a 9-metre thickener shipped together for a potash plant in Thailand; spodumene and lepidolite concentrators in Sichuan and Zimbabwe; zinc, molybdenum and copper circuits across Inner Mongolia, Shaanxi and Guangxi; and salt-lake potash projects in Qinghai and Laos.

The practical benefit for a plant owner is single accountability. When the mill, the classifier, the flotation cells and the thickener come from one engineering scope, throughput guarantees are negotiated once, spare parts share a lead time, and the commissioning crew has seen the flowsheet before they land. That does not make a project risk-free - ores are always full of surprises - but it removes the self-inflicted risks, which are the ones that cost the most.

Frequently Asked Questions

If you are planning a new concentrator or rebuilding an old one, bring the four stages into the discussion early - ideally with a head assay, a target throughput and, where available, a laboratory report in hand. Those three documents let an equipment engineer sketch a mass balance, propose the right machine class for each stage and flag the risks that actually matter for your ore, instead of quoting from a brochure.

Our engineers work with copper, zinc, gold, phosphate, potash and lithium ores from first testwork to commissioned plant, and we are equally glad to review an existing flowsheet as to design a new one. Send your ore details, throughput target and grind requirement, and we will come back with a stage-by-stage proposal you can hold against any other offer.