The ore dressing process, from run-of-mine ore to concentrate
Every concentrator runs the same four-stage chain: crush and screen, grind and classify, separate, then dewater. This guide walks the ore dressing process end to end - what each stage decides, and the signals that tell you it's being done wrong.
Discuss your project
Overview
Strip away the vendor language and the ore dressing process is one long argument with physics. The valuable mineral arrives locked inside barren rock, and every stage between the mine and the concentrate shed exists to unlock it, sort it, and hand it over dry enough to ship. You'll see the same skeleton in almost every concentrator - crushing and screening, grinding and classification, separation, then dewatering and tailings handling - whether the feed is copper, iron, lithium or fluorite. Some engineers call it mineral dressing. Others say beneficiation process, or ore beneficiation. Same chain, same decisions.
What changes from plant to plant isn't the skeleton; it's which branch the separation stage takes, and how each stage's mistakes surface downstream. So this page walks the chain end to end and asks two things of every stage: what does this step actually decide, and what tells you it's being done wrong?
The market sets the finish line
One orientation point before the stages. You don't get to invent the end specification - concentrate targets come from smelters, refiners and commodity benchmarks. Iron ore pricing references spot prices for imported fines at 62% iron content delivered at Tianjin Port, as reported in the U.S. Geological Survey's Mineral Commodity Summaries. Fluorspar splits into acid grade above 97% CaF₂ and metallurgical grade at 97% or below, per the same USGS series. A beneficiation flowsheet is really a negotiation between the grade your buyer demands and the recovery you're prepared to trade away to reach it. Push grade up and recovery drops. The curve is unforgiving, and where you choose to sit on it is a commercial call as much as a technical one.
Every stage below serves that negotiation. Keep it in view and the engineering choices stop looking arbitrary.
Crushing and screening: the size the whole plant inherits
Run-of-mine ore comes in at whatever size the blast delivered. Multi-stage crushing, usually in closed circuit with screens, brings it down to a size the grinding circuit can accept economically. That's the decision this stage owns: the transfer point between cheap breakage and expensive breakage. Crushing costs far less per unit of size reduction than grinding does, so a circuit that leaves the crushers underworked quietly shifts the burden onto the most power-hungry section of the plant, and the operating cost consequences of that shift run for the life of the mine.
The signals here are rarely subtle. Screens blind, oversize recirculates, and the circuit chokes at a throughput well below its nameplate. Feed to the mills surges, so the grinding circuit alternately starves and floods. If your mills look undersized on paper while the crushers loaf along below capacity, the problem usually started back here. Screening deserves the same scrutiny as crushing itself: it protects the crushers from material that's already fine enough and defines the top size everything downstream must live with.
Grinding and classification: liberation, paid for in kilowatts
Grinding is where mineral particles are finally freed - liberated - from the rock that hosts them, and it's usually the largest single power consumer in a concentrator. This stage decides your liberation size and, through classification, the actual size distribution the separation stage receives. Mills do the breaking; classifiers such as hydrocyclones or spiral classifiers close the loop, cutting the stream at a target fineness and sending coarse material back for another pass.
Grind too coarse and composite particles report to the wrong product, so recovery leaks into tailings and no reagent scheme will buy it back. Grind too fine and you've paid power to make slimes: ultra-fine particles that float badly, settle slowly and drag reagent consumption upward. The tell-tale is recovery that refuses to improve as you grind finer, or a flotation bank holding stable grade while the tailings assay creeps up week after week.
Fineness targets are ore-specific and belong to testwork. In a copper flotation test programme for a plant in Kazakhstan, as recorded in our project brochures, closed-circuit tests at a grind of 50% passing 200 mesh, using one rougher, two scavenger and three cleaner stages, produced a copper concentrate grading 16.23% Cu at 66.23% recovery from a 3.50% mass yield. Those numbers describe that ore under those conditions and nothing else. Your ore will write its own numbers, which is exactly why the test campaign comes before the flowsheet.
Separation: the stage where the ore picks the method
Everything upstream is preparation. Separation is where value and waste finally part company, and the method isn't a preference - it follows from mineralogy.
- Flotation works on differences in surface chemistry, using reagents and air bubbles to lift target minerals out of the pulp. It carries most sulphide ores of copper, lead and zinc, and a long list of non-metallic minerals, and it's often the only practical answer for fine-grained, complex ores.
- Gravity separation exploits density contrast and needs no reagents, which keeps operating cost and environmental load down. Gold, tungsten, tin and zircon-titanium sands are its classic territory.
- Magnetic separation sorts on magnetic susceptibility and anchors most iron and manganese circuits, plus cleaning duties in quartz and feldspar processing.
- Leaching dissolves the metal instead of concentrating the mineral - the standard route when gold is too finely disseminated for physical methods alone.
Real flowsheets mix these branches. A 3,500 t/d iron operation in Mongolia recorded in our project brochures delivers concentrate at 65% Fe or better at a yield of roughly 31% - a magnetic-circuit outcome. A 3,000 t/d gold plant in Guinea, run under our EPC+M+O scope and recorded in the same brochures, holds overall recovery around 93% from feed grading 1.2 g/t - a leach-route outcome. And in test programmes for a 2 Mt/a spodumene project in Zimbabwe, flotation couldn't separate the petalite cleanly, so the recommended flowsheet led with dense-medium separation instead, keeping gravity recovery for the tantalum-niobium and holding flotation back for the mica and spodumene steps further down the line.
The wrong-route signals: reagent dosages that keep climbing while grade stalls, gravity circuits recovering less each month as the orebody's character drifts, a concentrate that only meets specification when recovery is sacrificed well past plan. Those aren't operating problems; they're flowsheet problems wearing an operator's uniform, and they trace back to decisions made before the plant was built.
Dewatering and tailings: the quiet stage that closes the loop
Concentrate leaves the separation circuit as slurry, and nobody buys slurry. Thickeners settle the solids and return clarified water to the plant; filters take the underflow down to a moisture level that shipping and smelting will accept. On the other side of the plant, tailings need thickening, transport and a storage facility engineered to outlast the mine itself. This stage decides your water balance, your moisture penalties at the port and a large share of your closure liability.
Warning signs: a persistently cloudy thickener overflow sending dirty water back into the process, filter cakes too wet to stack or load, flocculant consumption drifting upward without a change in feed. Water chemistry circles back on you too - recycled water carries residual reagents that can shift flotation behaviour upstream, and if you've never connected a flotation upset to the return-water line, it's worth a look. We cover the storage side separately in our guide to tailings storage facilities.
Why the route comes from testwork
Reading a chain like this, it's tempting to pattern-match: iron means magnetic, gold means leach, copper means flotation. Resist that. The branch points above are settled by mineralogy, liberation size, water chemistry and market specification working together, and the only honest way to settle them for your deposit is a structured test programme on your actual sample. Xinhai's research teams run roughly 200 mineral test programmes a year across what the company reports as more than 70 ore types, working from a CNAS-accredited laboratory backed by an industrial-scale pilot base. The output that matters isn't the assay pile - it's a flowsheet you can defend to a lender. We've written about how that works in practice in our guides to metallurgical testwork and flowsheet development.
The chain keeps rewarding attention after commissioning, too. At the Zimbabwe spodumene operation mentioned above, our operating team lifted lithium recovery from 59.5% to 69% while holding equipment availability at 95.7% and running mill throughput 7.6% above design, as recorded in our project brochures. The improvement played out across the same four stages, with test data in hand.
If you're holding drill core, assay reports or an old feasibility study and wondering which branch your ore belongs on, that's a question our engineering and EPC+M+O team answers for a living, with the manufacturing capability to follow the design through. Send the mineral, the country, the capacity and the project stage. The right first answer may be a proposal, or it may be a request for a better sample. Both move you forward.