Copper Smelting Process
and Workflow Explained

A smelter does not process ore. It processes concentrate, the upgraded product of mining and froth flotation, typically grading 20-30% copper. The rest is iron, sulfur, silica, and a variable cocktail of minor elements. Chalcopyrite (CuFeS₂) is the dominant mineral in most commercial concentrates.

Two compositional ratios define how a smelter must operate. The sulfur-to-copper ratio determines how much autogenous heat the smelting furnace generates and how much SO₂ the gas train handles. The iron-to-silica ratio controls slag chemistry. These are not theoretical concerns. A smelter designed around a 28% Cu concentrate with 1.5 Fe/SiO₂ will struggle operationally if it suddenly receives 22% Cu material at 2.2 Fe/SiO₂. The furnace runs cold (less autogenous heat per unit copper), the slag becomes iron-rich and prone to magnetite precipitation, and the acid plant sees higher SO₂ throughput per tonne of copper produced.

The global concentrate quality trend is negative. ICSG data and CRU Group analyses have tracked rising arsenic, bismuth, and antimony levels in traded concentrates over the past fifteen years as mines deplete the cleaner upper portions of major porphyry deposits and push into more complex ore zones. Enargite-bearing ores (Cu₃AsS₄), common at high elevations in deposits across the central Andes, are a particular problem. Arsenic reports to every output stream of the smelter. In flue dust it complicates recycling. In sulfuric acid it creates off-spec product that fertilizer plants reject. In electrolyte it forms floating slimes that ruin cathode surfaces. In anode slime it contaminates precious metals refining. The penalty clauses in concentrate purchase contracts for arsenic above 0.2% are severe, and above 0.5% many smelters refuse the material entirely. The Japanese smelting industry (Saganoseki, Toyo, Naoshima, Onahama) has historically been the most willing to process complex concentrates, partly because decades of experience with diverse Pacific Rim ore sources built the engineering capability, and partly because Japan's absence of domestic copper mines means its smelters cannot afford to be selective about feed. Even so, Japanese smelters have invested heavily in dedicated arsenic treatment circuits, including scorodite (FeAsO₄·2H₂O) stabilization for permanent disposal, and these circuits are not cheap.

Concentrate arrives at 8-10% moisture. Flash dryers or rotary dryers reduce this to below 0.5%. There is nothing chemically interesting about this step, but it causes more unplanned downtime than its simplicity would suggest. Wet concentrate sticks to chutes, plugs transfer points, freezes in stockpiles during cold weather, and segregates on conveyor belts. Port-based smelters receiving seaborne concentrate deal with salt contamination from ocean spray during shipping, which introduces chloride into the system and accelerates corrosion in downstream gas handling equipment. Smelter availability studies, when they disaggregate downtime by cause, often show materials handling and drying accounting for 3-5% of total lost operating hours.

The smelting furnace performs the primary separation. Copper sulfide and iron sulfide enter together. Copper leaves in the matte (a molten sulfide phase, denser, collecting at the bottom). Iron leaves in the slag (a molten silicate phase, lighter, floating on top). The separation exploits a thermodynamic fact: oxygen attacks iron sulfide before copper sulfide because the Fe-S bond is weaker. FeO formed by this oxidation reacts with silica flux to make fayalite slag (Fe₂SiO₄). Copper concentrates in the remaining sulfide matte.

Outokumpu flash smelting is the most analyzed and best-documented technology. Dried concentrate is injected with oxygen-enriched air into a vertical reaction shaft. Particles ignite in suspension, react while falling, and settle into a molten bath below. The sulfur oxidation is exothermic and provides most of the process heat. In a well-run flash furnace at 60% matte grade with oxygen enrichment above 60%, external fuel consumption is negligible. The furnace feeds itself.

This is the standard textbook account, and it describes the technology used at Aurubis Hamburg, most Freeport-McMoRan smelters, and the majority of plants built before 2005. But the global picture has shifted. Chinese smelter construction since 2005, which now accounts for roughly 40% of world primary copper smelting capacity according to Wood Mackenzie estimates, has favored different technologies. The bottom-blown oxygen smelting furnace (SKS or BBS design, developed from Chinese lead smelting experience at the Shuikoushan facility and scaled up by ENFI Engineering) and the Ausmelt/Isasmelt top-submerged-lance (TSL) furnace have captured a large share of new Chinese installations. The reasons were pragmatic: lower capital cost, shorter construction timelines, and a domestic engineering base familiar with these designs. Whether these furnaces match flash smelting in energy efficiency or SO₂ capture performance at equivalent scale is debated within the industry. Proponents argue the gap has closed. Skeptics, particularly in the Japanese and European engineering establishments, point to higher refractory consumption rates, less stable thermal profiles, and greater difficulty achieving consistent high matte grades in TSL and bottom-blown furnaces compared to mature flash smelting operations. This is a genuine technical disagreement with commercial stakes, not a settled question.

The copper content of the matte, expressed as weight percent, is the single most important operating parameter. At 55% Cu, the matte contains substantial FeS alongside Cu₂S, meaning the converting step must oxidize and remove a large quantity of iron. At 65% Cu, nearly all iron has already been driven into the slag during smelting, so converting is faster but the smelting furnace must operate at higher oxygen potential. Higher oxygen potential pushes more copper into the slag as dissolved Cu₂O. The crossover point, where the copper lost to slag exceeds the savings in converting, depends on slag treatment efficiency. A smelter with a well-designed slag flotation circuit (recovering 80-90% of entrained copper from cooled, milled slag) can tolerate higher slag copper and therefore run higher matte grades profitably. A smelter without flotation, relying solely on electric slag cleaning with lower recovery, cannot push matte grade as aggressively without bleeding copper into the discard pile.

Flash furnaces operate with blast enrichment between 40% and 75%+ O₂. High enrichment reduces gas volume (smaller gas trains, smaller acid plant absorbers), concentrates SO₂ in the off-gas (improving acid plant conversion efficiency), and intensifies reaction kinetics. The constraint is thermal management. Reducing nitrogen ballast removes a heat sink from the reaction shaft, risking localized overheating, accelerated refractory erosion, and in extreme cases, damage to the copper water-cooling jackets embedded in the shaft walls. Oxygen plant power consumption is substantial. At large smelters, the cryogenic air separation unit draws 15-20% of total site electricity. The cost of oxygen is therefore directly tied to the local electricity price, which means the optimal enrichment level at a smelter in Chile (relatively expensive grid power) is different from one in Zambia or DR Congo (where power costs and availability fluctuate for entirely different reasons).

Smelting slag at 1-2% Cu represents a large copper inventory at production scale. At a smelter producing 300,000 tonnes per year of cathode copper, slag copper losses at 1.2% versus 0.6% in discarded slag represent roughly 3,000-5,000 tonnes of metal per year. At $9,000/tonne copper price, this is $27-45 million annually.

Electric furnace slag cleaning holds molten slag in a quiescent pool, allowing matte droplets to settle by gravity. Stokes' law governs settling velocity: larger droplets in less viscous slag settle faster. Retention time is typically 4-8 hours. Electrical energy input (through submerged graphite electrodes) maintains temperature and provides mild reduction to convert some dissolved Cu₂O back to metallic or sulfide copper, improving recovery. The discharged cleaned slag targets below 0.7% Cu.

Slag flotation is the alternative. Slag is slow-cooled (in pots or granulated, though pot cooling is preferred because it grows larger copper sulfide grains), crushed, milled, and floated in a conventional flotation circuit. Recovery is good but the circuit requires significant capital (grinding mills, flotation cells, thickeners, tailings dam capacity) and operating cost.

Some smelters use both. The slag cleaning furnace handles the initial copper recovery, and the slag from the cleaning furnace goes to flotation. This is expensive in capital but achieves the lowest overall copper losses.

Converting removes iron and sulfur from matte in two stages. The slag blow oxidizes FeS: the iron is fluxed with silica into converter slag, which is recycled to the smelting furnace because it contains 4-8% Cu. When all iron is gone, the remaining bath is white metal (Cu₂S). The copper blow then oxidizes the sulfur: Cu₂S + O₂ → 2Cu + SO₂. The product is blister copper, 98.5-99.5% Cu.

Peirce-Smith converters have done this job since 1909. They are batch vessels, about 4-5 meters in diameter and 10-13 meters long, lined with magnesia-chrome refractory brick. Molten matte is ladled in through the converter mouth while the vessel is rotated to a charging position. Air is blown through a row of tuyeres along the bottom. Slag is skimmed by rotating the vessel to pour from the mouth. The cycle (charge matte, blow slag, skim slag, charge more matte, finish copper blow, pour blister) takes several hours.

The batch nature creates real problems. Every time the converter is rotated for charging or skimming, the tuyere air supply is interrupted and uncontrolled SO₂ escapes from the mouth. These fugitive emissions are the primary source of ambient SO₂ exceedances at most conventional copper smelters, and they are difficult to eliminate because the operations that cause them (tilting a 300-tonne vessel full of 1200°C liquid) cannot be enclosed in a sealed gas collection hood without creating severe safety hazards. Secondary hooding systems partially capture the fugitives, but capture efficiency is typically 85-95%, not 99%+.

Continuous converting alternatives exist. The Kennecott-Outotec flash converting furnace processes solid high-grade matte (granulated and dried) in a flash reaction shaft similar to the smelting furnace, producing blister copper directly. The Mitsubishi process uses a top-blown lance furnace receiving molten matte continuously from the smelting furnace via a heated launder. Both achieve near-complete SO₂ capture because they operate as sealed, continuous systems.

Adoption of continuous converting has been slower than environmental arguments would predict. Flash converting requires granulating, stockpiling, drying, and re-injecting the matte as a solid feed, adding infrastructure and eliminating the thermal advantage of handling matte in the molten state. The Mitsubishi process requires the smelting furnace, converting furnace, and slag cleaning furnace to operate as a tightly coupled train, which means a shutdown in any one unit stops the entire line. Peirce-Smith converters, despite their emissions problems and thermal inefficiency, accept variable matte chemistry, tolerate cold scrap additions, recover quickly from interruptions, and can be taken offline individually for maintenance while other converters continue operating. This operational resilience, difficult to quantify in a project feasibility study but intensely felt during actual operations, explains the technology's persistence. Claiming that Peirce-Smith converters survive only because of industry conservatism misreads the situation. They survive because they solve a real operational problem (flexibility under variable conditions) that continuous alternatives do not solve as well, while creating a different problem (fugitive emissions) that is addressed through secondary capture systems at considerable but manageable cost.

Blister copper is charged to an anode furnace. Air oxidation removes residual S, Fe, Pb, Zn, and other impurities as volatile or slaggable oxides. A thin oxide slag is skimmed. The melt now contains roughly 0.6-0.8% dissolved oxygen as Cu₂O, which must be removed before casting because it makes the copper brittle and causes porosity in the anode.

Reduction uses natural gas, LPG, or in some older operations, ammonia-cracked forming gas. Hydrocarbons decompose at the melt surface, generating CO and H₂ that reduce Cu₂O to Cu. The endpoint is tough pitch, approximately 0.1-0.15% oxygen. The traditional endpoint test is the sample button: a small quantity of copper is cast into a mold, allowed to solidify, and the surface is inspected. A flat surface means the oxygen content is correct. A domed or blistered surface means excess sulfur or insufficient reduction. A concave surface means over-reduction. There are now continuous dissolved-oxygen sensors (based on EMF measurement with zirconia electrolyte probes) that provide real-time readings, but the sample button remains in use at most operations because operators trust the integrated signal it provides.

Anode casting is the transfer point where pyrometallurgical quality translates into electrochemical performance. Copper at 1100-1150°C is poured through a launder system onto a Hazelett-type continuous casting wheel or a conventional rotary wheel with individual molds. Anode weight targets are typically 350-380 kg. Dimensional tolerance is tight: thickness variation beyond ±3 mm causes current distribution problems in the tankhouse, leading to uneven dissolution, rough cathode deposits, and short circuits that require manual intervention by tankhouse operators.

Anodes are hung in electrolytic cells, typically 40-60 anodes per cell, interleaved with cathode blanks. The electrolyte is acidified copper sulfate solution, maintained at about 45 g/L Cu²⁺ and 170-190 g/L H₂SO₄, circulating at controlled temperature (60-65°C). Current density runs at 280-330 A/m². A refining cycle takes 14-28 days depending on anode weight and current density.

Copper dissolves at the anode: Cu → Cu²⁺ + 2e⁻. Copper deposits at the cathode: Cu²⁺ + 2e⁻ → Cu. The net result is copper transfer from impure anode to pure cathode at 99.99% purity (LME Grade A, as registered on the London Metal Exchange brand list).

Elements less noble than copper (Fe, Ni, Co, Zn) dissolve into the electrolyte but their deposition potential is more negative than the cathode operating potential, so they stay in solution. Elements more noble (Au, Ag, Se, Te, Pt, Pd) are not oxidized at the anode potential and fall to the cell bottom as anode slime, a black or dark gray mud.

Electrolyte purification is continuous. A bleed stream is withdrawn, copper is recovered in liberator cells (electrowinning at high current density from the impure electrolyte), and nickel is crystallized out as nickel sulfate hexahydrate through evaporative crystallization. Arsenic, if present at high levels, may be removed by solvent extraction or by precipitation as a copper arsenate compound. The purified, copper-depleted electrolyte is returned to the circuit. Nickel sulfate is sold, primarily into the stainless steel and (increasingly) battery precursor markets. At smelters processing high-nickel concentrates, nickel sulfate revenue is a meaningful line item.

Anode slime is collected when cells are cleaned between cycles. At a large smelter (300,000+ tpa cathode), slime production may be 2,000-5,000 tonnes per year. This slime goes to a precious metals refinery, which is usually on-site or nearby. The refinery process involves decopperizing (acid leach or roast-leach to remove residual copper), then smelting and cupellation to produce doré (a gold-silver alloy), then electrolytic parting to separate gold and silver to 99.99% purity. Selenium is recovered from the decopperizing leach liquor or from roaster off-gas and refined to commercial grade (99.5%+ Se). Tellurium follows a similar path.

The commercial value of anode slimes shapes concentrate purchasing decisions in ways that are not obvious from a purely metallurgical perspective. When a smelter evaluates a potential concentrate supply contract, the gold and silver credits in the concentrate feed through to anode slime revenue after applying a chain of metallurgical recovery factors: smelting distribution (what fraction of Au/Ag in concentrate reports to matte versus slag), converting distribution, fire refining distribution, and electrorefining collection efficiency. Each smelter has its own empirical recovery data for this chain, accumulated over years of operation and considered commercially sensitive. A smelter achieving 97% payable gold recovery across the entire chain has a real financial advantage over a competitor at 93%, and this difference directly affects the TC/RC terms it can offer to mines. During the annual benchmark TC/RC negotiations (typically conducted between October and January for the following year's contracts, with Freeport-McMoRan and Chinese smelters historically setting the pace), precious metals terms are negotiated alongside copper treatment charges, and for concentrates from gold-rich deposits like those in Indonesia and Papua New Guinea, the precious metals economics can dominate the negotiation.

SO₂ gas is generated during smelting and converting. Modern gas handling systems capture it, clean it of dust and volatile impurities (As₂O₃, SeO₂, Hg, fluorides), and route it to a double-absorption contact acid plant. SO₂ is oxidized to SO₃ over V₂O₅ catalyst at 420-440°C, then absorbed in 98% H₂SO₄. Conversion rates in double-absorption plants exceed 99.7%.

A smelter producing 300,000 tpa of copper cathode generates roughly 800,000-1,000,000 tpa of sulfuric acid, depending on concentrate sulfur content. This volume of acid must go somewhere. In regions adjacent to phosphate rock deposits (Morocco, Florida, Saudi Arabia), acid is readily absorbed by fertilizer production. In landlocked locations or regions with surplus acid, disposal becomes a cost rather than a revenue source.

The Chinese situation deserves specific discussion because it illustrates how the acid market constrains copper production. Between 2005 and 2020, Chinese copper smelting capacity approximately tripled, according to ICSG and CRU data. Sulfuric acid production grew proportionally. Chinese domestic acid demand, driven largely by phosphate and ammonium sulfate fertilizer production, grew as well but not fast enough to absorb all the smelter acid. During periods of acid oversupply (2014-2015 was a notable trough), spot acid prices in eastern China dropped to levels where smelters were paying to dispose of acid. This created a situation where the binding constraint on copper cathode production was not the copper market, not the concentrate market, but the acid market. Some smelters curtailed throughput specifically because they could not move acid. The Yanggu Xiangguang and Tongling Nonferrous operations, among the largest in China, had to negotiate acid offtake agreements that effectively subsidized acid disposal. This coupling between a base metal and an industrial chemical has no real equivalent in zinc, lead, or nickel smelting and is one of the structurally distinctive features of the copper smelting business.

Custom smelters charge mines a treatment charge (TC, $/tonne of dry concentrate) and a refining charge (RC, ¢/lb of payable copper) to process concentrate. The mine retains metal ownership. The smelter collects tolling fees plus premia or penalties for impurity levels, plus retained precious metals fractions beyond the payable percentages.

Annual benchmark TC/RCs have varied widely. In 2004, benchmarks were around $45/4.5¢. By 2006 they had risen above $90/9.0¢ as new mine supply from Escondida and Grasberg expansions filled the market. They compressed again during 2012-2015 as Chinese smelter capacity grew faster than concentrate supply. In 2024, spot TC/RCs collapsed to the range of $10-20/1-2¢ as a combination of mine disruptions (Cobre Panama closure, First Quantum's problems, Codelco production shortfalls) tightened concentrate supply against a backdrop of still-expanding Chinese smelting capacity. This was a crisis for custom smelters: at TC/RCs below $30/3¢, most smelters outside China cannot cover operating costs from tolling income alone and depend on precious metals credits and acid revenue to survive.

This produces a dynamic that confuses people outside the industry: the copper price can be at historical highs while smelter margins are at historical lows. The copper price rewards the mine. The TC/RC rewards the smelter. They move on different supply-demand balances. A world with strong copper demand, high copper prices, but insufficient concentrate supply (because mine investment lagged) is a world where mines capture most of the value and smelters get squeezed. This has been the prevailing situation for much of 2023-2025.

The recycle loops bind the smelter into a single system. Converter slag (4-8% Cu) returns to the smelting furnace. Anode scrap (the undissolved remnants of electrorefining anodes, typically 12-18% of original anode weight) goes back to the anode furnace. Electrolyte bleed goes to copper recovery and nickel crystallization. Flue dust from the gas train, enriched in volatile elements (As, Bi, Pb, Cd, Zn), is either recycled to the furnace (if arsenic levels permit) or processed hydrometallurgically to recover copper and stabilize arsenic. Precious metals refinery residues loop back through their own sub-circuit.

A disturbance at any node propagates. If the concentrate blend changes, slag chemistry shifts, magnetite risk changes, converter cycle time adjusts, anode impurity profile shifts, electrolyte management requirements change, and precious metals refinery feed composition changes. All simultaneously. There is no step that is independent of the others.

The difference between a smelter that runs at 92% availability and one that runs at 96% is almost never about the furnace or the converter or the tankhouse in isolation. It is about how well the transitions and material transfers between these units are managed. The furnace-to-converter matte transfer, the converter-to-anode-furnace blister transfer, the anode-furnace-to-casting-wheel copper transfer, the casting-to-tankhouse anode logistics. Each interface is a potential bottleneck, a potential source of temperature loss, a potential scheduling conflict. Smelters that invest in launder heating, covered transfer ladles, automated casting wheels, and real-time production scheduling systems tend to gain more throughput per dollar spent than those that invest in incrementally larger furnaces or faster converters. This is an unsexy observation, and it is ignored in most engineering feasibility studies, which focus on unit operation capacity rather than inter-unit transfer efficiency. Operational people know this. Project engineers frequently do not.