An Ore Roasting Furnace is a high-temperature pyrometallurgical reactor that heats sulphide or carbonate ores in a controlled-air atmosphere to drive off volatile species and convert metal sulphides into oxides. It solves the problem that most sulphide concentrates cannot be smelted directly — the sulphur must come out first as SO₂. Ore is fed continuously, held at 500–900 °C with regulated oxygen, and the resulting calcine drops out for downstream smelting or leaching. Modern fluidized-bed roasters at operations like Glencore's Altonorte process 3,000+ t/day of copper concentrate this way.
The Ore Roasting Furnace in Action
Roasting sits between concentration and smelting. You take a flotation concentrate — say 28% copper, 30% sulphur, balance iron and gangue — and you need to get the sulphur down to a level the smelter can handle without choking on SO₂. The furnace does this by holding the ore at a temperature where the sulphide reaction with oxygen runs hard but stays below the melting point of the calcine. For chalcopyrite-style feeds that window sits around 650–750 °C. Push past 800 °C and you start sintering particles into clinker that bridges the hearth or defluidizes the bed. Run below 550 °C and the reaction kinetics fall off a cliff and you get under-roasted product with residual sulphide that fouls downstream leach circuits.
The two dominant designs are the multiple-hearth roaster (MHR) and the fluidized-bed roaster (FBR). In a multiple hearth, ore enters at the top and rabble arms — slow-rotating water-cooled arms with angled teeth — plough the bed across each hearth, dropping it through alternating inner and outer drop holes so it cascades down through 6 to 12 hearths. Air enters at the bottom and rises counter-current. In a fluidized bed, the ore sits in a 1–2 m deep bed suspended on upward air flow through a constriction plate; the whole bed behaves like a boiling liquid, gas-solid contact is exceptional, and reaction rates are 5 to 10 times higher than a hearth design.
What goes wrong? Three things, mostly. Tramp oversize feed plugs the drop holes on a multiple hearth — anything above 6 mm should not be in the feed, period. On a fluidized bed, off-spec air distribution through a worn constriction plate creates dead zones that sinter into a brick of fused calcine, and once that starts the bed has to come out cold. And in both designs, autogenous operation — running on the heat of the sulphide reaction itself with no auxiliary fuel — depends on feed sulphur staying above roughly 22%. Drop below that and the burners come back on, fuel costs blow out, and the off-gas SO₂ concentration drops below what the acid plant downstream needs to make 98% sulphuric acid.
Key Components
- Refractory-Lined Shell: The pressure boundary and thermal envelope. For a 6 m diameter fluidized-bed roaster you're looking at 250–350 mm of high-alumina brick backed by 100 mm of insulating castable, with shell skin temperature held below 200 °C. Hot face refractory typically goes 4–6 years between major rebuilds.
- Air Distribution System: On a fluidized bed this is the constriction plate or tuyere grid that creates uniform upward gas velocity of 0.6–1.2 m/s. Hole pitch and diameter must be matched to bed density; a 10% variation in open area across the plate creates dead zones that sinter.
- Rabble Arms (Multiple Hearth): Water-cooled steel arms rotating at 0.5–1.5 RPM with cast tooth ploughs angled to move ore radially. Cooling water inlet must hold below 40 °C to prevent arm sag — a sagged arm scrapes the hearth refractory and you'll be replacing brick within weeks.
- Feed System: Screw feeders or rotary valves metering 50–500 t/h of concentrate within ±2% mass flow. Feed moisture must stay below 8% — wetter than that and the feed balls up at the inlet and you get plugging instead of dispersion.
- Off-Gas Handling: Cyclones plus electrostatic precipitator capturing entrained dust, then routed to a sulphuric acid plant. Off-gas SO₂ concentration of 8–14% is the sweet spot for double-contact acid plants; below 4% the plant cannot autothermal.
- Calcine Discharge: Overflow weir or underflow drawoff at 600–700 °C, dropped to a cooler — usually a fluid-bed cooler or rotary cooler — before transfer to smelter or leach. Discharge temperature drift of more than ±25 °C means the bed is not at steady state and calcine quality is off-spec.
Who Uses the Ore Roasting Furnace
Roasting still earns its keep wherever the feed is a sulphide or refractory ore that cannot be cleanly leached or smelted as-is. The driving question is always the same: do you need to remove sulphur, oxidize arsenic, decompose carbonates, or expose locked gold? Each answer points to a different roaster type and operating window.
- Copper Smelting: Glencore's Altonorte operation in Antofagasta, Chile runs a fluidized-bed roaster ahead of the smelter, processing copper concentrate to lower the sulphur load on the flash furnace and feed SO₂ to a 3,400 t/day acid plant.
- Zinc Production: Nyrstar's Hobart smelter uses Lurgi-type fluidized-bed roasters to convert zinc sulphide concentrate to zinc oxide calcine for downstream sulphuric acid leaching — autogenous operation at around 950 °C.
- Gold Recovery: Newmont's Carlin operation in Nevada uses circulating fluidized-bed roasters on refractory sulphide ore to oxidize sulphides and arsenopyrite, exposing gold for cyanide leaching that would otherwise recover under 30%.
- Molybdenum Processing: Climax Molybdenum's Fort Madison roaster runs multiple-hearth furnaces to convert MoS₂ concentrate to MoO₃ technical-grade oxide for steel alloy and catalyst production.
- Nickel Refining: Vale's Long Harbour processing facility uses partial roasting on pentlandite concentrates to control the sulphur-to-metal ratio fed to the matte smelter.
- Cobalt/Mixed Sulphides: Sherritt's Fort Saskatchewan refinery uses fluid-bed roasting on mixed nickel-cobalt sulphides ahead of pressure leaching, targeting 0.5–1.0% residual sulphide for optimal leach kinetics.
The Formula Behind the Ore Roasting Furnace
The single most useful sizing calculation for a roaster is the autogenous heat balance — does the sulphide oxidation reaction generate enough heat to hold operating temperature without auxiliary fuel? At the low end of typical concentrate sulphur content (around 18–20%) the answer is no, and you'll burn natural gas or oil to make up the deficit. At the nominal range (28–34% S, typical chalcopyrite or pyrite-rich concentrate) the reaction is comfortably exothermic and the issue flips to removing excess heat with cooling coils. At the high end (above 38% S, pyrite-heavy feeds) you have so much heat available that bed temperature runs away unless you dilute with air or recycle calcine. The formula below estimates net heat available per tonne of concentrate.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Qnet | Net heat available per tonne of concentrate after losses and sensible heating | MJ/t | BTU/short ton |
| mS | Mass of sulphur oxidized per tonne of concentrate | kg/t | lb/short ton |
| ΔHS | Heat of reaction for S → SO₂ (≈ 9.28 MJ/kg S) | MJ/kg | BTU/lb |
| mFe | Mass of iron oxidized to Fe₂O₃ per tonne of concentrate | kg/t | lb/short ton |
| ΔHFe | Heat of reaction for 2Fe → Fe₂O₃ (≈ 7.37 MJ/kg Fe) | MJ/kg | BTU/lb |
| Qlosses | Radiation and shell losses, typically 8–15% of gross heat | MJ/t | BTU/short ton |
| Cp | Average specific heat of feed | kJ/kg·K | BTU/lb·°F |
| ΔT | Temperature rise from feed to operating | K | °F |
Worked Example: Ore Roasting Furnace in a copper concentrate fluidized-bed roaster
A regional copper concentrator in Sonora, Mexico is sizing a fluidized-bed roaster to handle 800 t/day of chalcopyrite concentrate at 30% S, 28% Cu, 25% Fe, balance gangue. Feed enters at 25 °C and the bed runs at 680 °C. The team needs to confirm the unit will operate autogenously before committing to the burner-down design.
Given
- mS = 300 kg/t
- mFe = 250 kg/t (assume 60% oxidizes to Fe₂O₃ → 150 kg/t)
- ΔHS = 9.28 MJ/kg
- ΔHFe = 7.37 MJ/kg
- Cp = 0.75 kJ/kg·K
- ΔT = 655 K
- Qlosses = 12% of gross —
Solution
Step 1 — calculate gross reaction heat from sulphur oxidation at the nominal 30% S feed:
Step 2 — add iron oxidation heat (150 kg/t to Fe₂O₃):
Step 3 — subtract sensible heat to bring 1 tonne of feed from 25 °C to 680 °C, and subtract radiation/shell losses at 12% of gross:
That's a strongly positive heat balance at nominal conditions — the bed will need cooling coils to dump excess heat, not burners to add it. Now compare across the operating range. At the low end of typical sulphide concentrates, 20% S and 18% Fe, gross heat falls to roughly 2,650 MJ/t and Qnet drops to about 1,850 MJ/t — still autogenous but with thinner margin and no room for moisture spikes in the feed. At the high end, 38% S pyrite-heavy feeds, gross heat climbs above 4,800 MJ/t and Qnet exceeds 3,700 MJ/t. At that level cooling coil duty becomes the limiting factor; if the coils can't pull enough heat the bed climbs through 750 °C and you start sintering chalcopyrite into clinker that defluidizes the bed.
Result
Q<sub>net</sub> ≈ 2,932 MJ/t at the nominal 30% S design point — comfortably autogenous with surplus heat to manage. In practice this means the operator runs with cooling coils active, burners on standby for startup only, and off-gas SO₂ near 11% which is right in the band an acid plant wants. Across the typical operating range, the 20% S low-end gives roughly 1,850 MJ/t (autogenous but tight, watch feed moisture) while the 38% S high-end approaches 3,700 MJ/t and pushes cooling-coil duty into the design limit. If the actual measured Q<sub>net</sub> on commissioning is 20% below predicted, look first at feed moisture above 8% pulling latent heat out of the bed, second at iron oxidation conversion below 50% (insufficient residence time — check superficial velocity has not been pushed past 1.2 m/s and short-circuited the bed), and third at calcine carry-over to the cyclones carrying unreacted sulphur with it, which shows up as SO₂ readings dropping below 8% on a healthy-looking feed.
Choosing the Ore Roasting Furnace: Pros and Cons
The roaster choice is a question of throughput, feed flexibility, and downstream integration. Multiple-hearth roasters are the older technology but still earn their place on small-tonnage refractory feeds, while fluidized beds dominate large copper and zinc operations. Rotary kilns serve a niche where ore is coarse or feed chemistry varies widely.
| Property | Fluidized-Bed Roaster | Multiple-Hearth Roaster | Rotary Kiln Roaster |
|---|---|---|---|
| Throughput per unit | 500–3,500 t/day | 50–400 t/day | 100–800 t/day |
| Reaction rate vs hearth baseline | 5–10× faster | 1× (baseline) | 1.5–3× faster |
| Feed particle size limit | ≤ 2 mm, dewatered concentrate | ≤ 6 mm, tolerates lumps | ≤ 25 mm, very forgiving |
| Off-gas SO₂ concentration | 8–14% (acid-plant ready) | 4–8% (often too lean) | 2–6% (usually too lean for acid) |
| Capital cost (relative) | High | Medium | Medium-High |
| Major refractory rebuild interval | 4–6 years | 2–4 years (rabble arm wear) | 3–5 years |
| Best application fit | High-tonnage Cu/Zn sulphide concentrate | Refractory gold, Mo, small-tonnage | Variable feed, carbonate decomposition |
| Turndown ratio | 2:1 (narrow) | 4:1 (wide) | 3:1 |
Frequently Asked Questions About Ore Roasting Furnace
Reported sulphur on a daily composite assay does not equal the sulphur the bed actually saw at any given hour. The most common cause is a flotation upset upstream that briefly dropped concentrate grade to 18–22% S for several hours — long enough to cool the bed and force burners on, but invisible in the daily average.
Check the on-stream XRF or hourly grab samples, not the daily composite. Also check feed moisture: every 1% above 6% moisture pulls roughly 25 MJ/t of latent heat out of the bed, and four points of moisture excursion can wipe out your autogenous margin without changing the sulphur number at all.
At 200 t/day you're in the overlap zone where both work, and the deciding factor is usually arsenic content and downstream gas handling. If your feed is arsenopyrite-bearing and you need precise temperature staging to fix arsenic as stable Fe-As-O species rather than volatilizing it as As₂O₃, the multiple-hearth gives you independent temperature control on each hearth — that's why Newmont kept multiple-hearths on parts of the Carlin circuit even when fluid beds were available.
If arsenic is low and you want maximum gold liberation per unit footprint, a circulating fluid bed wins on capital and on reaction completeness. Below 100 t/day, multiple-hearth almost always wins. Above 500 t/day, fluid bed almost always wins.
Bed temperature being correct only tells you the average is right. Leach recovery problems point to under-roasted particles — residual sulphide locked inside larger feed grains that the leach acid can't reach. Two real causes: short-circuiting in the bed (gas channelling around dense regions, leaving pockets of unreacted feed) or feed top-size creeping above 2 mm because a screen broke upstream.
Sample the calcine and run a sulphide-sulphur assay separate from total sulphur. If sulphide-S is above 1.5% you have an under-roast problem regardless of what the temperature controllers say. Pull a screen analysis on the feed — anything on the +2 mm screen is your culprit.
Defluidization usually starts as a local sintering event — a dead zone above a plugged tuyere or a hot spot from a feed surge — that fuses particles into agglomerates too heavy for the gas velocity to suspend. Once those agglomerates collect, bed pressure drop spikes, gas distribution gets worse, and the failure cascades.
The early warning is a slow rise in bed differential pressure at constant gas flow. Catch it there and you can sometimes recover by dropping bed level, increasing superficial velocity 10–15% temporarily, and pulling feed for 20 minutes. Miss that window and the bed solidifies and you're cold-dumping through the discharge port — usually a 3–5 day outage.
The textbook heat balance assumes nameplate combustion air temperature, usually 25 °C. In a Sonora plant in summer that's fine. In a Sudbury or Trail winter operation pulling −20 °C ambient air, you're spending an extra 80–120 MJ/t just heating combustion air to bed temperature, and that's enough to flip a marginally autogenous design into a fuel-burning one.
The fix is air preheat — typically a recuperator on the off-gas stream lifting combustion air to 250–400 °C. Most northern roasters have this; if yours doesn't and you're fighting fuel costs every winter, it's the first thing to look at.
Either you're roasting less sulphur than you think, or you're diluting the off-gas with excess air. Both are bad for the acid plant downstream — below 8% SO₂ a double-contact acid plant struggles to run autothermal and you start burning sulphur or molten sulphur to make up.
Check the air-to-feed ratio first. Stoichiometric air for a 30% S concentrate is roughly 2.4 Nm³/kg; if you're running at 3.5 Nm³/kg because someone opened a damper to control bed temperature, you're diluting SO₂ by 40%. The right answer is to use cooling coils to control temperature, not excess air. Second check is feed grade — if concentrate slipped to 22% S you'll see exactly this off-gas pattern even with correct stoichiometry.
References & Further Reading
- Wikipedia contributors. Roasting (metallurgy). Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
