The Blast Furnace
Iron ore is iron bound to oxygen. Coke burns to carbon monoxide at the tuyeres, and the gas strips the oxygen from the ore on its way up; liquid iron gathers in the hearth.
Scroll
Oxygen in the Ore
About 5 % of the Earth’s crust is iron, but nearly all of it is bound to oxygen. The ore smelted today is mostly hematite (Fe₂O₃), sometimes with a little magnetite (Fe₃O₄).
To get the iron, the oxygen has to come off. A blast furnace does it with carbon: coke burns to carbon monoxide low in the furnace, and the carbon monoxide rises, takes oxygen from the ore and becomes carbon dioxide.
In 2025 the world’s blast furnaces made about 1,271 million tonnes of pig iron, and the world made 1,849 million tonnes of crude steel. By worldsteel’s count, about 71 % of steel today takes the blast-furnace route: the furnace makes liquid iron, and an oxygen furnace turns it into steel.
About 5 % of the crust is iron (USGS) · hematite Fe₂O₃, magnetite Fe₃O₄ · world pig iron 2025: about 1,271 Mt (worldsteel)
In at the Top
Ore (mostly sintered into lumps or rolled into pellets) and coke go in at the top in turn and lie in the furnace in alternate layers. Coke is fuel and reducing agent, and also the column’s skeleton: without it to carry the load, the furnace could not run.
The top is sealed: the furnace runs 0.25–2.5 bar above the air outside. Most large furnaces now have a bell-less top: the burden falls from a hopper onto a rotating chute, and changing the chute’s angle lays it at different radii of the stock surface. On one thyssenkrupp furnace the chute is about 3 m long and turns once every 7.5 seconds.
The EU average in 2004: for each tonne of hot metal, about 1.6 t of iron-bearing burden (1,088 kg of sinter, 358 kg of pellets, 180 kg of lump ore) and 359 kg of coke went in at the top, and 162 kg of coal was blown in at the tuyeres.
Ore and coke in alternate layers · the bell-less top lays them with a rotating chute · EU 2004 average per tonne of hot metal: 359 kg coke, 162 kg coal (BREF)
Fire at the Tuyeres
A ring of tuyeres runs round the top of the hearth, up to 42 on a large furnace. The hot blast drives into the coke and blows out a cavity where the coke swirls as it burns: the raceway. Temperatures here can exceed 2,000 °C.
The coke burns first to carbon dioxide, but it is surrounded by glowing coke, and the carbon turns the carbon dioxide straight back into carbon monoxide (C + CO₂ → 2 CO). At 1,000 °C, gas in equilibrium with coke is more than 99 % carbon monoxide (computed here from thermodynamic data). What leaves the raceway is carbon monoxide and the air’s nitrogen.
Tuyeres are water-cooled copper. The two faces of the cut fall just between two tuyeres: they show the raceways from the side.
Up to 42 tuyeres · the raceway can exceed 2,000 °C (BREF) · the gas leaving it: CO and N₂
Oxygen Off, Step by Step
Carbon monoxide doesn’t take the oxygen all at once. From about 500 °C, hematite turns into magnetite (Fe₂O₃ → Fe₃O₄), an easy step; magnetite then becomes wüstite (FeO), and wüstite finally iron. At every step, a molecule of carbon monoxide takes one oxygen atom and leaves as carbon dioxide.
The last step is the hard one: at 900 °C, the gas must be at least about 68 % carbon monoxide before wüstite turns to iron (computed here from thermodynamic data). So the gas leaving the top still carries a lot of carbon monoxide: only about half has become carbon dioxide.
In the middle of the shaft, the burden stays at 950–1,000 °C for a stretch, the thermal reserve zone, where the gas reduces wüstite to iron. Above about 1,100 °C, what oxide is left reacts with the coke itself (FeO + C → Fe + CO): direct reduction, which takes in a lot of heat. The glowing lump on the stand is a schematic, and moves far faster than a real one.
Fe₂O₃ → Fe₃O₄ → FeO → Fe · at 900 °C, reducing FeO takes about 68 % CO (computed) · thermal reserve zone 950–1,000 °C (Nippon Steel 2006)
Down Against Up
The burden goes down, the gas goes up. Coming down, the burden gets hotter; going up, the gas gets cooler: over 2,000 °C in front of the tuyeres, only 100–230 °C at the top. Heat and oxygen change hands all the way.
A lump of ore takes five to eight hours from the top to the hearth. The white streaks in the section are a schematic: they show the gas going up, not its real speed or path.
The gas leaving the top is about 20–28 % carbon monoxide, 17–25 % carbon dioxide and 50–55 % nitrogen, and a cubic metre of it gives 2.7–4.0 MJ of heat. It goes through a dust catcher, is washed, and then fires the stoves.
The burden takes 5–8 hours to come down · top gas 100–230 °C · CO 20–28 %, CO₂ 17–25 %, N₂ 50–55 % (BREF)
Where the Ore Softens
At about 1,100 °C the ore layers begin to soften, stick and melt. In the furnace this band takes the shape of an inverted V, or a V or a W, depending on how the burden is laid and the furnace is run: the cohesive zone.
Its shape, thickness and position decide where the gas goes. The coke is still solid here; below the cohesive zone, drops of iron and slag trickle down through the coke into the hearth.
It was seen only after furnaces were quenched and cut open: in the 1970s, four experimental and nine production furnaces were dissected in Japan, and the layers of ore and coke kept their charged order until the ore melted.
Softening and melting from about 1,100 °C · an inverted V, a V or a W · seen when quenched furnaces were dissected in the 1970s (Zhang 2016)
Slag Floats on Iron
Besides the iron oxides, ore carries gangue such as silica and alumina, and coke leaves ash. Limestone (CaCO₃) turns to lime (CaO) when heated and joins them as slag, mostly lime, silica, alumina and magnesia.
Today most of the lime goes into the sinter first: European sinter averages a basicity of 1.8, and the limestone charged into the furnace itself averages only about 26 kg per tonne of hot metal.
Each tonne of hot metal brings 175–350 kg of slag. Slag is far lighter than liquid iron (about 2,400 against 6,700 kg/m³) and floats on it. About three-quarters of the EU’s furnace slag is granulated in water and can stand in for part of the clinker in cement.
Slag: CaO 38–42 %, SiO₂ 33–37 % · 175–350 kg per tonne of hot metal · granulated slag goes into cement (BREF)
Opening the Taphole
Iron and slag gather in the hearth and are let out from time to time. A drill opens a taphole in the hearth wall, and iron and slag run out together, the slag usually a little after the iron. A furnace has one to four tapholes.
The iron comes out at about 1,440–1,500 °C, with about 4 % carbon. At the skimmer, the slag floating on top is held back and turned into the slag runner; the iron passes underneath into a ladle or a torpedo car.
When the cast is over, a mud gun drives a heat-resistant clay into the taphole and plugs it.
Hot metal at about 1,440–1,500 °C with about 4 % carbon · the skimmer parts slag from iron · the mud gun plugs the hole (BREF, Metinvest)
The Stoves
The hot blast comes from stoves beside the furnace, three or four to each. Each is a tower filled with checker bricks: first it burns the furnace’s own gas until its dome is at about 1,100–1,500 °C and the heat is stored in the bricks; then the fire is cut off and cold air is blown through the other way, leaving as a blast of 900–1,350 °C.
The stoves take turns: one on blast, the others on gas. There are two on the stand; the nearer one is cut open to its combustion chamber and checker bricks.
Six large furnaces listed by Nippon Steel in 2006 blew at 1,138–1,262 °C, with 2.8–8.3 percentage points of extra oxygen in the blast.
Three or four stoves to a furnace · dome 1,100–1,500 °C · blast 900–1,350 °C (BREF)
From Charcoal to Coke
In 1709 Abraham Darby smelted iron with coke instead of charcoal in the old furnace at Coalbrookdale, England. Others had tried before him; he made it pay. The furnace still stands, part of the Museum of Iron.
In 1828 James Beaumont Neilson patented the hot blast (No. 5701). By Percy’s account of 1864, the Clyde Ironworks took about 8.1 t of coal for a ton of pig iron on cold blast in 1829, and about 5.2 t in 1830 with the blast heated to about 150 °C.
In 1857 Edward Alfred Cowper patented the regenerative hot-blast stove. His paper of 1860 reports the first pair, at the Ormesby works, heating the blast to about 700 °C, the two stoves changing over every two hours. Stoves still work this way.
1709, Darby: coke · 1828, Neilson: hot blast · 1857, Cowper: the regenerative stove
The Carbon in Every Tonne
The furnace takes oxygen away with carbon, and most of the carbon ends as carbon dioxide. By the IEA’s estimate, iron and steel emitted about 2.6 billion tonnes of carbon dioxide directly in 2019, about 7 % of the energy system’s emissions.
worldsteel’s figures (2024, including indirect emissions such as bought electricity): about 1.92 t of carbon dioxide for the average tonne of steel; about 2.34 t by blast furnace and oxygen furnace, about 1.47 t by gas-based direct reduction and electric furnace, about 0.69 t from scrap in an electric furnace.
Another way is to take the oxygen with hydrogen instead of carbon, leaving water. Sweden’s HYBRIT made about 100 t of hydrogen-reduced sponge iron in its pilot plant in 2021, and its three companies say the method removes about nine-tenths of steelmaking’s emissions. The industrial demonstration plant is not built: in 2024 LKAB put it back to the end of 2027, probably 2028.
Iron and steel’s direct emissions in 2019: about 2.6 Gt CO₂, about 7 % of the energy system (IEA) · 1.92 t CO₂ per average tonne of steel (worldsteel 2024)
Shell, Brick and Staves
The furnace comes apart into the top gear, the steel shell, the cooling staves, the brick lining, the hearth’s carbon blocks, the tuyeres with the bustle pipe, and the foundation. In 2013 POSCO’s Gwangyang No. 1 was enlarged to 6,000 m³, then the world’s largest; Shagang’s No. 4, at 5,800 m³, was designed for 14,000 t of hot metal a day. By Nippon Steel’s report of 2006, its most recently relined furnaces had run about 15 years between relinings.
By Metinvest’s account, the shell is steel plate up to 35 mm thick. Inside it, a ring of water-cooled staves, copper in the shaft of some large furnaces. Inside those, the refractory, up to 1.5 m thick at the hearth, whose floor is carbon blocks under a layer of ceramic.
The part under the pointer lifts out alone and is named; left alone, the parts lift out in turn.
Top gear · shell · staves · brick lining · carbon hearth · tuyeres and bustle pipe · foundation
Hot Blast, Hot Flame
Reading one: the raceway’s adiabatic flame temperature, from an energy balance on the NIST-JANAF thermodynamic tables: dry blast, coke entering at 1,500 °C, no coal injected, no moisture, no heat lost. Reading two: the heat each cubic metre of blast brings in.
On this page, 100 °C more blast makes the flame about 79 °C hotter; one percentage point more oxygen, about 45 °C. Real furnaces inject coal and add moisture and run cooler: a Nippon Steel model’s base case gives 2,111 °C at 1,100 °C blast.
The raceways in front of the tuyeres brighten and dim with the reading.
Readings: adiabatic flame temperature · heat per cubic metre of blast · computed here from NIST-JANAF
Try thisRaise the hot blast from 900 °C to 1,250 °C: how much hotter does the flame at the tuyeres get? And with 5 % more oxygen?
What the Model Leaves Out
The stand is a model of the principle, not any one furnace: the furnace is drawn after a large one of about 5,000 m³ (1:80); the stoves, the dust catcher and the cast house are smaller than real and closer together; a real furnace has three or four stoves, here two; platforms, piping and the cast house’s building are much reduced. In the section, the layers, the cohesive zone, the raceways, the deadman and the iron are placed and shaped as a schematic: the layers and grains are drawn thick, and they sink more than a hundred times faster than real ones; the white streaks only show the gas going up. The glowing lump is a schematic. The lab’s flame temperature is an ideal energy balance (dry blast, no coal, coke at 1,500 °C, no heat lost); real furnaces run cooler.
Easily Misread
Carbon monoxide doesn’t act only at the top: it reduces all the way up from the tuyeres. Nor does it take all the oxygen: where it is hot enough, the coke takes oxygen directly, and that costs heat. Not all the coke burns: it carries the column and goes down to the hearth. Slag isn’t waste: most of it becomes building material. The 2.34 t of carbon dioxide per tonne of blast-furnace steel is worldsteel’s average on its own basis, and plants differ widely. Hydrogen reduction changes only what takes the oxygen; HYBRIT’s “nine-tenths” is the companies’ own figure, and the industrial plant is not built.
Common questions · 5
- How does a blast furnace get iron out of ore?
To get the iron, the oxygen has to come off.
01Oxygen in the Ore ↑- What is the raceway of a blast furnace?
The hot blast drives into the coke and blows out a cavity where the coke swirls as it burns: the raceway.
03Fire at the Tuyeres ↑- How is iron ore reduced step by step in a blast furnace?
Carbon monoxide doesn’t take the oxygen all at once.
04Oxygen Off, Step by Step ↑- What is the cohesive zone of a blast furnace?
At about 1,100 °C the ore layers begin to soften, stick and melt.
06Where the Ore Softens ↑- How do hot-blast stoves heat the air?
The hot blast comes from stoves beside the furnace, three or four to each.
09The Stoves ↑
Sources
- European Commission JRC, Best Available Techniques reference document for Iron and Steel Production (2013): zones and temperatures, stoves, burden and coke rate, gas, slag, casting, lining
- Nippon Steel Technical Report No. 94 (2006): thermal reserve zone, blast temperature and oxygen, coke and coal rates, gas utilisation, large furnaces’ volumes and campaigns
- M. Naito, K. Takeda, Y. Matsui (2015): Ironmaking Technology for the Last 100 Years, ISIJ International 55, 7 (the cohesive zone’s shape, bell-less charging, Japan’s large furnaces)
- W. Zhang (2016): A Review on the Dissection of Quenched Blast Furnaces, Processes 4, 36 (the zones seen in dissected furnaces)
- Z. Zhang (2023): The Characteristics and Reduction of Wustite, IntechOpen (the order and temperatures of reduction, the burden’s residence time)
- M. Roche et al. (2020): Drainage Model of Multi-taphole Blast Furnaces, Metall. Mater. Trans. B 51 (tapholes, the densities of iron and slag)
- NIST-JANAF Thermochemical Tables (4th ed., 1998): the enthalpies and Gibbs energies behind this page’s flame temperature and reduction equilibria
- thyssenkrupp Steel press release (20 Jan 2023): fifty years of the bell-less top, Duisburg blast furnace 4 in 1972, the rotating chute
- Metinvest: Iron and slabs (hot metal with 4.0–4.3 % carbon, a steel shell up to 35 mm thick)
- Danieli Corus news (2024, 2026): Shagang blast furnace 4, 5,800 m³, designed for 14,000 t of hot metal a day
- POSCO company history: Gwangyang blast furnace 1, 6,000 m³ inner volume, 2013
- worldsteel: World Steel in Figures 2026 (crude steel, pig iron and DRI output in 2025, BOF and EAF shares, CO₂ per tonne of steel)
- IEA (2020): Iron and Steel Technology Roadmap (the sector’s direct emissions and share)
- U.S. Geological Survey (2026): Mineral Commodity Summaries, Iron Ore; Statistical Compendium, Iron ore (iron in the crust)
- HYBRIT (SSAB, LKAB, Vattenfall) press releases (2021) and LKAB releases (2024, 2026): hydrogen-reduced sponge iron, the demonstration plant’s progress
- UNESCO: Ironbridge Gorge (World Heritage site 371)
- IMechE: The Old Furnace, Coalbrookdale (Engineering Heritage)
- J. Percy (1864): Metallurgy: Iron and Steel (Neilson’s patent and the Clyde Ironworks’ coal)
- E. A. Cowper (1860): On hot-blast stoves, Proc. IMechE (the first pair of stoves, at Ormesby)
Data
The raceway’s flame temperature and the gas that reduces each iron oxide are computed from the NIST-JANAF Thermochemical Tables (M. W. Chase, 4th edition, 1998). Furnace practice and numbers are from the EU’s Best Available Techniques reference document for iron and steel production (2013), the Nippon Steel Technical Report No. 94 (2006), M. Naito, K. Takeda and Y. Matsui (ISIJ International, 2015), W. Zhang (Processes, 2016) and others listed below; production from worldsteel’s World Steel in Figures 2026 and the USGS; emissions from the IEA (2020) and worldsteel; the history from J. Percy, Metallurgy: Iron and Steel (1864), E. A. Cowper (Proc. IMechE, 1860), UNESCO and the IMechE. The furnace, the stoves and the cast house are a model of the principle, not any one plant. Material presets, parts and the softbox studio of the site’s model kit are ported from AirsupHQ/machine-skill (MIT, © 2026 Air Supply Technologies Inc.).
The text, pictures and models on this page are licensed CC BY-NC 4.0: free to use and adapt for teaching and other non-commercial purposes, with credit to world.lapla.si. Third-party material listed above keeps its own licence.
Cite this plate
Laplace’s Mini World, Plate 57, “The Blast Furnace”, https://world.lapla.si/en/blastfurnace/, CC BY-NC 4.0, updated 2026-10-29.
Follow the question
Plate 54Nitrogen from the AirAir is nearly four-fifths nitrogen, yet crops can’t use it: the nitrogen molecule’s triple bond is too strong. The Haber–Bosch process presses nitrogen and hydrogen to 100–250 bar, heats them to 350–550 °C and joins them on an iron catalyst to make ammonia. Cold, the equilibrium favours ammonia but the reaction crawls; hot, it runs fast but yields little, so plants settle on a compromise. Each pass converts only a fifth to a third, and the rest goes round again. About 70 % of the ammonia made today becomes fertiliser.
Plate C01Watt’s Steam EngineAn iron vessel standing in cold water saved most of the coal; two spinning balls taught the engine its own pace. Built in Birmingham in 1788, it stands at rest in London today.
