It is the most abundant metal in the Earth's crust, yet it clings so tightly to oxygen that no one could unlock it until the 1880s. Freeing it still takes one of the most electricity-hungry processes in all of industry — and the world now makes about 74 million tonnes of it a year.
Light enough that you stop noticing it is metal at all.
For most of history, no one could get at it. Aluminium was only isolated as a metal in the 1820s, and prising it loose from oxygen was so hard that the metal was worth more than gold: France put aluminium bars on show beside the crown jewels, and Emperor Napoleon III is said to have saved his aluminium cutlery for his most honoured guests. In 1884 a small pyramid of it — then about as costly as silver — was chosen to cap the Washington Monument. Then, in 1886, two 22-year-olds working independently — Charles Martin Hall in Ohio and Paul Héroult in France — found they could split alumina apart with electricity. Within a generation the price fell from hundreds of dollars a pound to pennies. Aluminium had not become any more common — electricity had.
The whole journey reads as a change of colour: a red rock is mined, washed into a white powder, then split apart with electricity into a silver metal — bauxite, to alumina, to aluminium. Tap any stage to look closer.
↓ tap any stage to explore ↓
Aluminium is everywhere in the ground, but never as metal — it is always bonded to oxygen. Making it is really about prising those two apart, and oxygen does not let go easily.
Aluminium makes up roughly 8% of the Earth's crust — more than any other metal — yet you will never find a nugget of it. It bonds to oxygen so fiercely that it took chemists until the 1880s to work out how to separate the two, and the method they landed on is still the one used today. Getting to finished metal takes two very different transformations back to back: a chemical wash that pulls a pure white powder out of red rock, and then a torrent of electricity that tears that powder apart. Three ingredients make it possible.
The ore. A reddish rock rich in aluminium hydroxide minerals — gibbsite Al(OH)3 and boehmite AlO(OH). Its rusty colour comes from iron oxide; silica and titania ride along as impurities. Chief sources: Guinea, Australia, China and Brazil.
Hot, concentrated sodium hydroxide (NaOH) is the solvent that does the washing. Under heat and pressure it dissolves the aluminium out of crushed bauxite, leaving the iron oxides and other insoluble minerals behind as a residue known as red mud.
Alumina melts at 2,072°C — far too hot to handle directly. So instead of melting it, the process dissolves it in molten cryolite (Na3AlF6), a salt bath that is itself liquid at about 960°C. Electrolysis then happens in that bath, not in molten alumina. Carbon blocks carry the current in, and are slowly eaten away as they do.
Making aluminium from ore is one of industry's most energy-hungry jobs. Making it from old aluminium is one of its cheapest — the same metal, for a twentieth of the energy.
Every piece of aluminium takes one of two routes. The primary route starts from bauxite and runs it through two plants — a refinery that washes out pure alumina, then a smelter that splits that alumina with electricity. It is the only way to make brand-new metal, and it is enormously power-hungry. The secondary route skips all of that and simply re-melts scrap. Because aluminium can be recycled without any loss of quality, remelting needs only about 5% of the energy of the primary route — which is why recycled metal now makes up the majority of supply in places like the United States, even as fresh smelting still dominates worldwide.
Then they converge. Whichever road it came down, molten aluminium ends up in the same holding furnace, where alloying elements are stirred in to fix its recipe. From there it is cast into ingots, billets and slabs, then rolled into sheet and foil, extruded into window frames and tubes, or cast into finished shapes. One metal, two beginnings, a single finish.
Pure aluminium is too soft to be much use. Stir in a pinch of another metal and it transforms — and which metal you choose decides everything about the family it belongs to.
On its own, aluminium is soft and easily bent. Its usefulness comes from alloying — dissolving a little copper, magnesium, silicon, manganese or zinc into it. The industry sorts these into numbered families, where the first digit names the main added metal.
Strong as some steels — the alloy that flies. Aircraft skins and structures.
Easy to shape and draw. The beverage-can alloy, also roofing and cookware.
Shrugs off seawater. Boat hulls, pressure vessels and marine structures.
The workhorse — easy to extrude into long shapes. Window and bike frames, railings.
Bauxite is mined where it lies, but the metal is made where the electricity is cheap — and increasingly, that means one country.
Aluminium's supply chain funnels down at every step: about 440 million tonnes of bauxite become 150 million tonnes of white alumina, which become roughly 74 million tonnes of finished metal. And where each of those steps happens is not the same map — the ore comes from the tropics, but the smelting has migrated to wherever power is cheapest.
The concentration problem, in one number. China smelts about 45 million tonnes of the world's 74 — some 61% of all primary aluminium — despite mining only a fifth of the bauxite it needs. It runs right up against its own self-imposed ceiling: a 45 Mt cap introduced in 2017 to curb overcapacity and emissions, which national capacity finally hit in 2024. No other major metal is made so overwhelmingly in one place.
Aluminium has no single carbon footprint — it depends almost entirely on where the electricity comes from. The same metal can carry 16 tonnes of CO₂ or 4, decided only by whether its smelter runs on coal or hydropower. Recycling collapses it to about half a tonne. That is why more than 90% of the industry's emissions come from primary production, and why cleaning up aluminium is, first and last, an electricity story.
A tonne of aluminium can be smelted fresh from ore or remelted from scrap, and the two routes have different geographies. Primary smelting is extraordinarily concentrated in China; recycling — though also China-led — is spread more widely across scrap-rich economies. Japan, Germany and South Korea lead the tier just behind those shown.
China leads both routes and India appears in each, so the two maps aren't opposites — what sets them apart is everything below the top. Primary smelting migrates to wherever power is cheapest, which is why Russia and Canada rank so high, and why China alone makes about 61% of the world's fresh metal; bauxite and alumina then travel oceans by ship to reach those smelters. Recycling instead follows the scrap, taking root in mature, metal-rich economies: the United States now smelts very little, yet more than three-quarters of its aluminium supply comes from remelting rather than smelting.
Most of aluminium's carbon is just the electricity it drinks — clean the grid and that shrinks. But a stubborn slice is baked into the chemistry itself, and no amount of clean power removes it.
Aluminium's emissions split into two very different problems. The larger part is simply the electricity that electrolysis devours — clean that power and it falls away. The smaller, tougher part comes from the reaction itself: the carbon anodes are consumed and turned into CO2, and that happens no matter how green the grid is. Solving aluminium means tackling both.
In a normal cell, the carbon anodes do two jobs: they carry the current, and they chemically soak up the oxygen freed from the alumina. That second job turns them into carbon dioxide — the anodes are literally eaten away and must be replaced.
Those carbon anodes account for almost a sixth of the greenhouse gases behind new aluminium — the part that a clean grid alone can never remove. Replacing them is a chemistry change, not a wiring change.
In November 2025, ELYSIS — a joint venture of Alcoa and Rio Tinto — started up the first commercial-size inert-anode cell, a 450-kiloampere pot slotted into an existing potline at Rio Tinto's Alma smelter in Quebec. It makes aluminium with no direct carbon emissions from smelting, releasing oxygen instead of CO2.
The pilot version has already made metal used in some Apple laptops, Audi wheels and Michelob cans. The next step is a multi-year test to prove the anodes survive full industrial conditions, with the aim of maturing the technology by the end of the decade.
It is aluminium's equivalent of hydrogen steelmaking: not a cleaner fuel poured into the old process, but a genuinely different reaction.
Aluminium's problem was never scarcity — it's the most abundant metal in the ground. It's energy. The metal drinks electricity to be born, so its footprint rises and falls with the power behind the plug: coal-fired smelting emits around sixteen tonnes of CO2 per tonne of metal, hydropower roughly a quarter of that, and recycling barely half a tonne. A clean grid does most of the work, inert anodes now tackle the stubborn sixth baked into the chemistry, and recycling quietly carries the rest. As of 2025 all three are real — what decides how fast they spread is where the world chooses to plug its smelters in.