Processing & metallurgy

SpecificationFigures below are from the AMIT open-courseware mining course at the University of Alaska Fairbanks, a university teaching resource rather than a standard or a vendor specification. Jaw crusher: run-of-mine feed up to 1.5 m, product minus 300 mm for transport or minus 200 mm for SAG feed, reduction ratio 4:1 to 9:1, typically preferred for feed rates of 900 tph or less, power 2.5 to 400 kW (the course table's upper capacity figure of 13,000 tph is not credible for a jaw crusher, is inconsistent with both its own 900 tph preference and its 400 kW power ceiling, and is omitted here). Gyratory crusher: feed 1 to 1.5 m, product 10 to 15 cm, reduction ratio 3:1 to 10:1, capacity 300 to 9,100 tph, power up to 750 kW, designed for direct truck dump from trucks up to 300 tons. Standard-head cone (secondary): reduction ratio 6:1 to 8:1, closed side setting 3.2 to 38 mm, capacity 260 to 1,750 tph, power 25 to 500 kW. Short-head cone (tertiary): reduction ratio up to 10:1, closed side setting 3.2 to 19 mm. Fine-crushing cone types: feed under 50 mm, product 6 to 9 mm, reduction ratio around 8:1. US metal mining crushing energy is about 1,900 Btu per ton against 57,900 Btu per ton for grinding (US Department of Energy, Mining Industry Energy Bandwidth Study, exhibit 24).

Where it falls downCompressive crushers choke on wet, clayey or sticky feed and on tramp metal. Reduction ratio per stage is limited, so three stages are needed to reach mill feed size, and each stage needs screening and surge capacity. Product size is controlled by the closed side setting, which opens as liners wear, so product size drifts through a liner campaign. Crushers cannot make a product finer than a few millimetres economically — below roughly 5 to 20 mm, grinding takes over.

SpecificationThe page cited covers ball mills and ball-mill circuits only, so no SAG, AG or rod mill figures are taken from it. Ball mills take a maximum feed around 2.5 cm; overflow-discharge mills are best suited to fine grinding to 75-106 microns while grate or diaphragm discharge mills are typically used for grinds to 150-250 microns. The ball-mill figures here come from the AMIT open-courseware mining course at the University of Alaska Fairbanks, a university teaching resource rather than a standard. Ball mill length-to-diameter ratio is normally 1.4 to 1.8; optimal speed is 70-80 percent of critical speed, where critical speed in rpm equals 42.3 divided by the square root of (mill diameter minus ball diameter); optimal ball charge is 35-45 percent of mill volume; commercial ball sizes run 10-150 mm, with steel ball bulk density 4,650 kg per cubic metre. Specific energy is estimated from the Bond equation E = Wi x (10/root P80 - 10/root F80) in kWh/t. Gold circuits commonly grind to 80 percent passing 75 microns for non-refractory ore. Grinding is about 40 percent of total US mining industry energy — roughly 494 TBtu/yr of 1,246 TBtu/yr (US Department of Energy, Mining Industry Energy Bandwidth Study).

Where it falls downTumbling mills are grossly energy-inefficient — most of the input becomes heat, noise and liner wear. Overgrinding produces slimes that are lost in flotation and gravity and that carry water into tailings. SAG mills are sensitive to feed size distribution and ore competency: a change in the coarse end of the feed can drop throughput sharply, and critical-size pebbles accumulate and must be crushed and returned. Steel media and liner wear contaminate the pulp with iron, which can consume reagents and depress flotation. Grinding is where the ore's hardness variability bites: a circuit sized on a composite work index will underperform on hard domains.

SpecificationAn industrial HPGR of 800 mm roll diameter by 400 mm width, fed minus 12 mm pre-screened material, achieved a specific energy consumption of 4.57 kWh/t at a fresh feed rate of 27.08 t/h and 107.51 t/h total throughput including circulating load, at an optimal specific pressure of 3.5-4.0 MPa (efficiency gains diminish above 4.0 MPa), producing a minus 1 mm product suitable for tower-mill feed; open-circuit P80 fell from 9.14 mm to 3.40 mm at 4.0 MPa, and a laboratory unit at 4.0 MPa used 2.02 kWh/t net. The same study reported a Bond ball work index of 16.06 kWh/t for the ore, and cited reported energy savings exceeding 45 percent for HPGR-stirred mill circuits against conventional blasting/SAG/ball mill circuits. Vertical stirred mills are restricted to a maximum feed size around 6 mm and a maximum media size around 30 mm, and a frequently quoted 1,119 kW unit rating that corresponds to one particular mill size rather than to a limit of the technology.

Where it falls downHPGR rolls wear rapidly on abrasive ore and are sensitive to tramp metal and to feed segregation; performance depends strongly on feed moisture and top size, and a circulating load of the order of 230-337 percent was reported across the tested feed range. Screening efficiency ahead of the rolls controls the circuit: in the cited work it fell from 86.45 percent at optimal conditions to 60.52 percent at maximum feed rate. Stirred mills cannot accept coarse feed at all, so they only substitute for the fine end of a circuit, and fine grinding increases slime generation and downstream water and reagent demand.

SpecificationHydrocyclone figures below are from the AMIT open-courseware mining course at the University of Alaska Fairbanks, a university teaching resource rather than a standard. Classifying cyclones are routinely applied below 300 microns. In the worked design example a 150 micron target separation size at 95 percent passing in the overflow corresponds to a corrected d50 of 110 microns, so the target separation size and the d50 are not the same number. Apex diameter is 10-35 percent of cyclone diameter and vortex finder 20-45 percent. Solids concentrations as high as 50 percent by weight are achievable without roping. Partition-curve sharpness is measured by the imperfection I = (d75 - d25) / 2 d50, with typical alpha values between 2 and 4. The page cited covers classifying cyclones only; no screening specifications are taken from it.

Where it falls downHydrocyclones bypass a fraction of the fines straight to the underflow with the water, so a cyclone is a poor classifier for fines and returns already-fine material to the mill, wasting grinding energy. Because the split is by settling rate, dense minerals report preferentially to underflow regardless of size, which is why gold and galena are chronically overground in cyclone circuits. Screens blind on near-size, wet and clay-rich material, and fine screening capacity per unit area falls steeply as aperture shrinks. Roping — a thick, unfanned underflow discharge — collapses classification completely.

SpecificationFigures here are from the AMIT open-courseware mining course at the University of Alaska Fairbanks, a university teaching resource rather than a standard. Feasibility is judged by the concentration criterion, CC = (SG of heavy mineral minus SG of fluid) divided by (SG of light mineral minus SG of fluid): CC above 2.5 means separation is easy down to 75 microns; 1.75-2.5 is possible down to 150 microns; 1.5-1.75 down to 1.7 mm; 1.25-1.5 down to 6.35 mm; below 1.25 separation is impossible at any size. Jigs stratify a bed by pulsed fluid flow; hydraulic jigs handle coarse material and dry jigs only close-sized material. Teeter-bed (autogenous dense-medium) separators run at 1.0-2.0 tons per hour per square foot and tolerate about a 3:1 size range. Centrifugal concentrators apply roughly 60 g in fluidised-bed designs and up to 300 g in flowing-film designs, at capacities up to about 100 tph.

Where it falls downThe concentration criterion is an absolute barrier: below 1.25 no gravity device will work at any size, which rules out most sulphide-from-silicate separations. Every gravity device needs a narrow feed size range — a wide size distribution lets fine dense particles behave like coarse light ones, so classification ahead of the device is mandatory. Recovery falls away below about 75 microns for all conventional devices, and slimes must be removed or they blanket the separation. Dry jigs perform recognisably worse than hydraulic jigs because air contributes negligible density, and they tolerate only a low content of near-gravity material.

SpecificationFigures here are from the AMIT open-courseware mining course at the University of Alaska Fairbanks, a university teaching resource rather than a standard, and describe coal preparation practice. Medium is a magnetite suspension (magnetite density 5,170 kg per cubic metre, required purity greater than 97 percent magnetics, fineness graded — a US Grade B is 95 percent passing 44 microns, with superfine around 31 microns and ultrafine around 26 microns). The underflow-to-overflow density differential is held below 0.40 relative density units. Dense medium vessels treat coarse material (plus 1/4 inch, above 6.35 mm) and are sized at 25 tons per hour per foot of overflow weir length with 260 gpm of medium per foot of weir. Dense medium cyclones treat between 3 inch and 100 mesh, are installed about 10 degrees from horizontal, operate at 9 cyclone diameters of head (10-20 psi inlet pressure) with a medium-to-solids ratio of about 5:1, and range from about 75 tons per hour at 20 inch diameter to about 416 tons per hour at 40 inch diameter. Sharpness is expressed as the probable error Ep, approximately 0.033 times separation density divided by particle size in mm. Medium is recovered by wet low-intensity magnetic drums (around 750 gauss; high-gradient rare-earth units around 21,000 gauss / 2.1 tesla), with total magnetite losses of 0.50-3.0 kg per tonne of feed.

Where it falls downSharpness collapses as particle size falls — dense medium cyclone application to 16 by 100 mesh material is limited by magnetite loss, and there is no practical dense medium route for fine material. Medium viscosity above about 20 centipoise degrades efficiency, and viscosity rises with magnetite concentration, contamination and fineness, so medium must be continuously cleaned and its contamination held below about 10 percent in both product streams. Density gradients inside the vessel trap near-density particles and cause surging, particularly with top size above 15 mm. Magnetite is a consumable and a cost, and its loss is dominated by adhesion on the drain-and-rinse screen.

SpecificationSeparation is principally based on differences in surface hydrophobicity, with particle size and density also significant. Reagent classes are collectors (which attach to the mineral surface and make it hydrophobic), frothers (which stabilise the bubble film), activators and depressants (which switch a mineral's floatability on or off). In copper practice the collector cited is sodium xanthate, with methylisobutyl carbinol as frother, producing a concentrate of approximately 30 percent copper; US copper recovery is quoted as 87.5 percent into concentrate (US Department of Energy, Energy and Environmental Profile of the U.S. Mining Industry: Copper). In lead-zinc practice, pH is set with lime, cyanide is used as a sphalerite depressant during galena flotation, and zinc is then activated and floated to a separate zinc concentrate. Rougher residence times, from the AMIT open-courseware mining course at the University of Alaska Fairbanks (a teaching resource rather than a standard): about 3-5 minutes for coal, 6-8 minutes for lead, 13-16 minutes for copper. Pulp density in roughing is about 4-8 percent solids for coal, 25-35 percent for lead and 32-42 percent for copper. Circuits are arranged as roughers, then scavengers, then cleaners, with roughers given maximum residence time and cleaners run for grade at lower residence time.

Where it falls downFlotation has a hard particle-size window. Below roughly 10-20 microns, bubble-particle collision probability becomes too low and fines are lost to tailings; above roughly 150-200 microns for sulphides, the bubble cannot carry the particle and it detaches — both are why the grind target is a compromise, not an optimum. It is highly sensitive to water chemistry, dissolved ions, clay slimes, oxidised mineral surfaces and iron from grinding media. Oxidised ores generally do not float without sulphidisation. Reagent suites are ore-specific and must be developed by test work; a suite that works on one domain can fail on the next. Talc, graphite and other naturally floatable gangue report to concentrate and dilute grade.

SpecificationLow-intensity magnetic separators operate at flux densities of about 2,000 gauss (approximately 0.2 tesla) and recover ferromagnetic material — chiefly magnetite — and highly paramagnetic material. High-intensity separators reach up to 20,000 gauss (about 2 tesla) for paramagnetic minerals, and superconducting systems reach 60,000 gauss (6 tesla). Dry high-intensity separation is effective down to about minus 100 plus 75 micrometres; below that a wet high-intensity separator is used, and below about 500 microns dry low-intensity separation is normally replaced by wet. Ilmenite, rutile and leucoxene behave as conductors; zircon, sillimanite, garnet, monazite and quartz behave as non-conductors.

Where it falls downMagnetic separation only works where there is a susceptibility contrast — it is useless on a suite of similarly paramagnetic silicates, and the cited review notes that magnetic concentration may not be satisfactorily efficient for particles finer than about 100 micrometres, because magnetic force on a particle falls faster than the competing drag and gravity forces. Middlings containing intergrown magnetic and non-magnetic phases report to whichever product the field strength favours, so liberation governs the result. Electrostatic separation requires a dry feed, because residual surface moisture short-circuits the surface charge, so feed is normally dried and heated; the optimum temperature is feed-specific and is established by test work rather than fixed by any general threshold. It is also restricted to a narrow size band and to single-layer feed presentation, which caps throughput per unit.

SpecificationPublished heap-leach ranges across commodities (2019 review in Sustainability): gold recovery 30-95 percent, with reported cycle times of 40 to 60 days; copper recovery reported across individual cases from about 53 to 77 percent, with sulphuric acid consumption of 25-75 kg per tonne of ore, one case reaching 73 percent in 140 days at 25 degrees C; uranium about 70 percent in one to six months; agglomeration of crushed ore with lime, Portland cement or fly ash before stacking to preserve permeability. Design practice for gold tank leaching (SAIMM, 1999) uses 6 to 12 leach reactors in series, 20-40 hours total residence, about 50 percent solids, 200-250 ppm cyanide in the first vessel falling to a minimum around 120 ppm in the last, at a grind of typically 80 percent passing 75 microns. In-situ recovery of uranium injects a lixiviant of water with oxygen or hydrogen peroxide plus sodium carbonate or carbon dioxide, recovers it through production wells and strips the uranium by ion exchange to yellowcake. Pressure oxidation of a pyrite-arsenopyrite concentrate (Processes, 2023, the source cited in this category's numbers list): 200 degrees C, 7 bar oxygen, 1 hour (typical POX residence for such concentrates is 30-60 minutes), giving 97-99 percent sulphide sulphur oxidation and 96-97 percent subsequent gold recovery. Stirred-tank bio-oxidation: 40 degrees C, 15 percent pulp density, three reactors in series at 2 days each, giving 43 percent sulphide oxidation and 68 percent gold recovery at 2 days, 74 percent and 82 percent at 4 days, and 79 percent and 88 percent at 6 days.

Where it falls downHeap leaching trades recovery for cost: reported gold heap-leach recoveries run as low as 30 percent, so a large fraction of the contained gold and silver may never be extracted. Permeability governs everything — clay, fines migration, over-compaction and poor agglomeration create channelling and dead zones no amount of extra cycle time will fix. Acid-consuming gangue (carbonates, some silicates) can make copper heap leaching uneconomic regardless of copper grade. Cyanide leaching fails on refractory ore where gold is locked in sulphide or associated with carbonaceous material, which is what forces the POX or bio-oxidation route. Bio-oxidation is slow, temperature- and pH-sensitive, and poisoned by chloride and some metals; pressure oxidation needs expensive acid-resistant pressure vessels and produces an acidic residue requiring neutralisation. In-situ recovery is restricted to permeable, confined, saturated host rock and carries an aquifer-restoration obligation.

SpecificationCIP design practice: carbon at 10-25 grams per litre of pulp (0.5 to 1.2 percent by volume), typically 6 to 8 adsorption tanks with a mean pulp residence time of about an hour per tank, interstage screens of 0.6-0.8 mm aperture, loaded carbon values on operating plants ranging from 300 to 20,000 grams of gold per tonne of carbon, gold recovery from solution of 90-99 percent depending on feed tenor, and adsorption tails solution of 0.01 to 0.005 grams per cubic metre regarded as practically achievable. Elution at 90-120 degrees C in 0.1 to 2 percent by mass sodium cyanide and 0.1 to 2 percent sodium hydroxide, targeting eluted carbon below 40 grams per tonne, followed by regeneration in a rotary kiln at 650-750 degrees C in a steam atmosphere to limit oxidation. Copper electrowinning (SAIMM Journal volume 109 number 6 page 343, 2009, not the carbon-in-pulp paper cited for this entry): current densities vary between 200 and about 375 A per square metre, with 250-300 A per square metre typical after solvent extraction; about 2 to 2.3 V per cell; current efficiency as high as 95 percent after solvent extraction but as low as 65 percent for direct electrowinning; electrolyte copper tenor 25-50 g/L; 1.54 g of sulphuric acid is regenerated for every gram of copper plated, from the stoichiometry CuSO4 + H2O giving Cu + H2SO4 + half O2, which is 98.08 g of acid per 63.55 g of copper (the cited electrowinning paper prints 1.98 g per gram, which does not follow from that reaction); optimum electrolyte temperature 45-55 degrees C for conductivity but held below 40 degrees C where solvent extraction is upstream; a cathode deposition cycle longer than eight days is not recommended. US Department of Energy figures (Energy and Environmental Profile of the U.S. Mining Industry: Copper) put hydrometallurgical copper processing at 1,490 Btu per ton against 484,000 Btu per ton for pyrometallurgical processing.

Where it falls downCarbon adsorption is defeated by preg-robbing carbonaceous ore, where natural organic carbon in the ore adsorbs the gold before the activated carbon can — CIL exists precisely to compete with it, and where it fails a blanking or oxidation step is needed. Carbon attritions and is lost, carrying gold with it, and fouls with organics, calcium carbonate and silica, which is why acid washing and thermal regeneration are mandatory. Solvent extraction requires a solution clean enough and of sufficient copper tenor to be worth treating, is sensitive to entrainment losses of organic, to crud formation at the interface, and to manganese and chloride in the leach solution. Electrowinning current efficiency falls sharply with iron in the electrolyte, which cycles between oxidation states and consumes current; short circuits between electrodes waste current and damage cathode quality; and lead anodes contaminate cathode with lead unless controlled.

SpecificationCopper (EPA AP-42 Section 12.3, a 1990s document): concentrate fed to the smelter grades 15 to 35 percent copper; roasting is carried out at about 650 degrees C (1,200 degrees F) and removes 20-50 percent of the sulphur as sulphur dioxide; reverberatory smelting temperature may exceed 1,500 degrees C (2,730 degrees F) — reverberatory furnaces are documented in AP-42 but are historical, not current US primary practice — while flash furnaces run at approximately 1,000 degrees C (1,830 degrees F); matte ranges from 35 to 65 percent copper with 45 percent the most common; converting yields blister copper at 98.5 to 99.5 percent copper; the anode furnace is maintained at 1,100 degrees C (2,010 degrees F); electrorefined cathode is 99.95 to 99.96 percent pure. Off-gas sulphur dioxide strengths: fluidised-bed roasters 10-12 percent, flash furnaces 10-70 percent, converters 4-7 percent. Conventional copper smelting is quoted at 30.92 to nearly 40 million Btu per ton of cathode copper recovered, with conventional smelting recovery of 98.7 percent (US Department of Energy, Energy and Environmental Profile of the U.S. Mining Industry: Copper; neither figure is in AP-42). Zinc roasting and electrolytic zinc figures have been removed from this entry: AP-42 Section 12.3 covers primary copper smelting only.

Where it falls downSmelters are capital-intensive, scale-dependent and inflexible — they need a steady concentrate supply of predictable chemistry. Penalty elements in concentrate (arsenic, antimony, bismuth, fluorine, mercury, chlorine) are charged against the seller and can make a concentrate unsaleable regardless of copper grade. Sulphur dioxide must be captured and converted to acid, which ties the smelter to an acid market; converter gas at 4-7 percent SO2 is the weakest and most difficult stream. Electrorefining is slow, works only on an anode already close to specification, and generates an anode slime that must itself be processed to recover gold, silver, selenium and tellurium. Roasting a gold concentrate mobilises arsenic and mercury and requires dedicated gas cleaning.

SpecificationConventional slurry tailings are typically about 20 to 40 percent solids by weight and up to 60 percent. High-density thickened and paste tailings are typically 60 to 75 percent solids by weight; high-rate and high-capacity thickeners produce conventional thickened underflow at 50-65 percent solids, high-compression thickeners produce underflow at high yield stress and high solids, and deep-cone paste thickeners reach up to 70 percent solids. High-density thickened and paste tailings have shear yield stresses from 40 Pa to 200 Pa and therefore require positive-displacement pumps. Filtered tailings are typically dewatered to more than 80 percent solids and often close to optimum moisture content, and give the highest water recovery during processing. In concentrate handling the same sequence is used: concentrate is thickened by settling, pumped out, dewatered by a filter press and dried.

Where it falls downDewatering rate is governed by particle size and clay content — a fine, clay-rich or high-slimes tailing settles slowly and holds water regardless of equipment, which is the practical ceiling on thickener underflow density. Flocculant is a continuing chemical cost and its dose must be retuned as ore changes. Paste and filtered tailings demand far more energy and capital than slurry, and filtered stacks require external water management because they hold little water on the facility. Thickener upsets — bed level loss, rat-holing, rake torque — propagate straight back into the plant water balance.