Aluminium Nearly Closes the Loop. Sorting Is What Stops It.
Most coverage of the circular economy, including ours, ends up explaining why a promising number does not hold up. So it is worth spending time on the material where the numbers largely do.
Aluminium is the strongest case circularity has. The energy argument is real, the recovery rates are genuinely high, and unlike paper or most plastics the metal does not inherently degrade when reprocessed. It still fails to close the loop — but for a reason that turns out to be far more interesting than “people do not recycle enough.”
The numbers that hold up
Recycling aluminium rather than smelting it from ore saves roughly 95% of the energy. The International Aluminium Institute puts primary production at 186 GJ per tonne from mine to cast house, against 8.3 GJ per tonne for recycled metal — a 95.5% reduction. Regional datasets for Europe and North America work out slightly lower, around 93–95%, so treat 95% as the top of a narrow range rather than a precise constant.
The emissions gap follows from the energy gap. The IAI puts secondary aluminium at roughly 0.5–1.0 tonnes of CO2e per tonne of metal. Primary production is an order of magnitude worse, and in some places much worse than that.
Collection works, too. Around 71% of aluminium cans were recycled back into new aluminium products globally in 2019, against roughly 34% for PET and 40% for glass bottles. The global Recycling Efficiency Rate across all aluminium sits near 76%, rising to about 81% in Europe.
A note on sourcing: most of these figures come from the International Aluminium Institute and national aluminium associations, which are industry bodies with an interest in the answer. We would normally discount accordingly. The energy claim in particular, though, is independently plausible from physics rather than advocacy: primary production requires electrolysis to break an extremely stable oxide bond, while recycling requires melting metal. Those are not comparable energy demands, and no accounting choice closes that gap.
One distinction worth getting right
You will frequently see the claim that around 75% of all aluminium ever produced is still in use today. It is well supported and it is genuinely remarkable.
It is also not a recycling rate, and it is routinely quoted as though it were.
That figure is a measure of stock — how much of the cumulative metal ever smelted remains in buildings, vehicles, cables and cookware. The corresponding flow measure, how much aluminium entering the system each year comes from scrap, is the 76% Recycling Efficiency Rate. Both are strong. They answer different questions, and conflating them overstates the case.
(The total tonnage behind the 75% is itself disputed, with sources giving either one or 1.5 billion tonnes ever produced. The share is consistent; the denominator is not. Best avoided.)
Why primary aluminium is so costly
The 95% saving only matters because the thing being avoided is so expensive in energy terms.
Aluminium smelting consumes roughly 3% of global electricity, and about 7% of China’s. More than 80% of Chinese smelting capacity runs on coal, much of it from captive coal plants built specifically to feed smelters — over 80 GW still in operation, responsible for around three-quarters of the greenhouse gas emissions from smelting electricity. Wood Mackenzie puts the average Chinese smelter at about 12.4 tonnes of CO2e per tonne of aluminium from power alone.
Set that against 0.5–1.0 tonnes for recycled metal and the stakes are obvious. Every tonne of demand met from scrap instead of a coal-fired smelter is one of the largest single-decision emissions reductions available in materials.
Which makes the remaining obstacle worth understanding precisely.
Aluminium is not one material
Here is the thing that breaks the loop, and it is not a collection failure.
“Aluminium” is a family of dozens of alloys with different compositions and incompatible uses. The practical division:
- Wrought alloys are rolled or extruded — drink cans, car body panels, window frames, foil. They require low impurity levels.
- Cast alloys are poured into moulds — engine blocks, gearbox housings, structural castings. They tolerate far more contamination.
Collected scrap is a blend of both, plus whatever came attached. Once alloying and contaminant elements are mixed into molten metal, separating them again is metallurgically very difficult and economically mostly impractical. Iron is the worst offender: it forms brittle intermetallic phases that degrade mechanical performance, and it accumulates with every cycle.
So mixed scrap flows downhill. It becomes cast alloy, because cast alloy will accept it. Producing wrought alloy from mixed scrap requires diluting it with large volumes of primary metal to get impurities back under tolerance — which reintroduces exactly the smelting the recycling was meant to avoid.
One frequently cited estimate puts around 6.1 Mt of wrought scrap per year being downgraded into cast products, with 8.0 Mt of primary aluminium used for dilution. That figure dates from 2012 and should be read as illustrative of the mechanism rather than current.
This is not downcycling in the plastics sense, where the polymer itself degrades. The aluminium atoms are fine. What degrades is the information — the knowledge of which alloy this metal was, lost the moment it entered a mixed bale.
The sink is closing
Inefficiency alone would be tolerable. The problem is that the destination for downgraded scrap is shrinking while the supply of it grows.
Vehicle lightweighting and electrification are pushing wrought demand up. Cast demand — historically driven by internal combustion engine blocks and transmission housings — is expected to stagnate or decline. Electric drivetrains simply need fewer large castings.
Follow that through. If mixed scrap can only become cast alloy, and cast demand is flat or falling while scrap volumes rise, then at some point the scrap collected exceeds what the market can absorb. Researchers modelling global aluminium flows have warned of exactly this: a scrap surplus with no suitable application.
That inverts the usual circular economy problem. For most materials the constraint is too little recycled content available. For aluminium, the emerging constraint may be too much of the wrong kind.
There is a quantity dimension as well, pulling the other way: aluminium products last decades, so scrap returns slowly relative to how fast smelting capacity was built. One modelling scenario has recycled metal rising from about 25% of production in 2024 to 60% by 2050. Both constraints can coexist — a shortage of scrap for total demand, and a surplus of scrap that cannot meet the specification demanded.
What would actually fix it
The fix is not more collection. It is sorting by alloy rather than by metal — keeping wrought and cast streams separate, and ideally distinguishing alloy families within them, so that scrap retains the identity that determines what it can become.
Modelling of Chinese vehicle recycling suggests that alloy-aware processing could cut the primary ingot required for vehicles by 10–37% between 2019 and 2050. That is a large reduction obtained purely by not mixing things.
Laboratory work is attacking the problem from the other direction too. A 2026 Nature Communications paper reports producing wrought alloys directly from end-of-life vehicle scrap without sorting or primary dilution, by designing around the contaminants rather than removing them. It is a lab-scale result with somewhat reduced elongation versus conventional alloys, and should be treated as promising rather than proven.
Policy is moving as well, if slowly. China has capped primary capacity at 45 Mt since 2017 and hit it in 2025, and brought aluminium into its national emissions trading scheme the same year. Both raise the relative cost of smelting against recycling.
Where this connects
Aluminium is an ESPR priority product group, with a delegated act indicatively due in 2027. That matters more here than for most materials.
The constraint on aluminium circularity is, precisely, lost information about composition. A Digital Product Passport recording alloy specification, carried with the component to end of life, addresses that constraint directly. It is the clearest case we have found of a transparency instrument that would do concrete physical work rather than merely document a claim — provided, as ever, that what it records is verified rather than self-declared.
The useful lesson
Aluminium demonstrates that circularity is achievable. The energy case is overwhelming, the collection infrastructure functions, and the material genuinely does not wear out.
It also demonstrates that “recyclable” is a property of a system, not of a material. The same metal is infinitely recyclable in a stream that knows what it is, and a one-way trip to engine blocks in a stream that does not. The atoms are identical. The difference is entirely whether anyone kept track.
For a sector that mostly talks about collection rates, that is the more useful place to put the next decade of effort: not gathering more material, but losing less of what we already know about it.
Sources: International Aluminium Institute, energy saving and recycling data; Transition Asia, The Chinese Aluminium Sector; Forecasting global aluminium flows, Waste Management; Nature Communications.