Six sister metals arrive together in the same rare rock, in trace amounts, so alike that separating them is the whole task — and turning that rock into pure, separated metal can take half a year.
There is almost certainly some within a few metres of you right now.
The whole journey
Turning platinum-group ore into finished metal runs as a left-to-right line with a loop: rock is crushed, floated, smelted and refined into six pure metals — and spent car catalysts feed back into the refining end, because a scrapped converter and a fresh lump of ore end up in the same place.
The same refinery that separates metal from fresh ore also separates it from your last car's exhaust. Recycling isn't a side-story here — it's the same machine, fed from the other end.
↓ Tap any stage to explore ↓
Two very different sources end up at the same refinery — one extraordinary body of rock, and the metal already in circulation.
Roughly 95% of the world's known PGM reserves lie in a single geological formation — the Bushveld Complex in South Africa (USGS; IPA) — a layered sheet of ancient rock whose thin metal-bearing seams, or reefs, are followed underground for kilometres. Some shafts run more than two kilometres deep, where the rock is hot enough that the mine has to be cooled for people to work in it. The second source is the metal already in circulation: because PGMs are used rather than used up, a scrapped car or a spent catalyst is itself an ore — a far richer one than any rock. One honesty clause worth keeping: the reef figures describe South Africa, which dominates platinum but not palladium. South African and Zimbabwean rock is richer in platinum, while Russian and North American ore — mined mostly as a nickel by-product — is richer in palladium.
The Bushveld Complex, South Africa: about 95% of world reserves, just 2–6 grams of PGM per tonne, mined up to two kilometres deep. PGMs travel with nickel and copper, and often ride out as their by-product.
Spent autocatalysts, industrial catalysts and electronic scrap — far richer than any rock, and re-entering the process at the refining end without a single new hole being dug.
A PGM atom is never really consumed — only borrowed. The same platinum can pass through a catalytic converter, be recovered, and go back into a new one, indefinitely. What limits recycling isn't the metal; it's getting the scrap to the refinery.
There are two routes to a finished ounce, and they meet at the end: the primary route makes new metal from ore, the secondary recovers it from scrap. Both finish at the same refinery, and their outputs are indistinguishable.
First the ore is crushed and milled to a powder, freeing the tiny metal-bearing grains. Flotation then floats those grains off as a concentrate — stirred into water with air blown through, the metal-rich grains cling to bubbles and rise as a froth that is skimmed. The concentrate is smelted in an electric furnace at more than 1,500 °C into a dense liquid "matte." Air is blown through the matte in a converter. Oxygen oxidises the iron and sulphur — the iron into slag, the sulphur into sulphur dioxide, which the smelter captures rather than vents.
The base metals — nickel, copper and cobalt — are removed first; only then are the six PGMs separated from one another. That last step is the genuinely hard one, and it is why the whole thing can take up to six months: the six metals are chemically so alike that no single reaction sorts them, so they are dissolved and coaxed apart one at a time by solvent-extraction, distillation and ion-exchange. The mining moves mountains of rock, but the refining is where the skill and the time go.
This describes the South African route — underground, deep and electricity-hungry. Russia and North America mine PGMs mainly as a by-product of nickel and copper, so there the metal rides out on another industry's back.
The recycling route skips the mine and starts with something that already contains the metal — usually a spent catalytic converter cut from a scrapped car. Consumer parts are collected and pre-processed; industrial catalysts, which never leave the factory, skip even that.
From there the material is smelted into a matte or dissolved into solution and sent through the same separation and refining steps. Done well, this recovers more than 95% of the PGM in the scrap, and the recovered metal is identical to freshly mined metal — it can be recycled again and again with almost no loss.
Platinum and palladium are the main event — together the large majority of what's dug up, and they pay for the mine. The other four ride along in trace amounts. There will never be an "iridium mine."
The route produces six metals, not one, and they are not equals. Platinum and palladium pay for the mine; rhodium, ruthenium, iridium and osmium are genuine by-products, recovered from the same ore because they are there. What they share is a family resemblance — dense, corrosion-proof, heat-proof, and quietly brilliant as catalysts — and each leans on a different one of those traits for its main job.
Autocatalysts, jewellery, and chemical & fuel-cell catalysts.
Autocatalysts for petrol engines, and electronics.
Hard-disk layers, electrical contacts, and as a hardener.
Crucibles, spark plugs, and hydrogen electrolysers.
Hard alloys and fine instrument pivots — the densest of all.
Two countries, two metals. South Africa supplies about seven in ten ounces of the world's mined platinum; Russia leads palladium. When one of them stumbles, the whole basket moves.
The world mines only about 450 tonnes of all six metals a year (IPA / BGS) — orders of magnitude less than common metals, and the reason a little has to go such a long way. Supply is also unusually concentrated: most platinum comes from South Africa, most palladium from Russia, and the two together set the price.
A mined ounce of PGM is among the most carbon-heavy metals on Earth to make — yet the same metals spend their working lives cutting pollution. Recycling is where that paradox resolves.
A mined ounce of PGM is one of the most carbon-heavy metals on Earth to produce — not because the chemistry is dirty, but because the ore is so dilute and the mines so deep that vast amounts of rock have to be moved and lifted, and the South African grid that powers it is roughly four-fifths coal. Yet the same metals spend their working lives cutting pollution: a converter's few grams neutralise more than two tonnes of exhaust pollutants over a car's life. The metal that costs so much carbon to make then saves far more in use — and recycling is where that paradox resolves.
Most PGMs today go into vehicle exhausts — autocatalysts take roughly 45% of platinum and 85–90% of palladium and rhodium. That is also the biggest question mark. A battery-electric car has no exhaust and needs no catalyst, so as electric vehicles spread, the industry's largest single market softens, and carmakers are already trimming the metal loaded into each converter. The counter-current is hydrogen: the same catalytic talent that cleans an exhaust also runs a fuel cell (platinum) and splits water in an electrolyser (platinum and iridium). Whether hydrogen grows fast enough to replace the demand electric cars erode is the open question hanging over the whole sector — a genuine question, not a settled forecast.
Finally, supply sits in few hands. With most platinum from South Africa and most palladium from Russia, PGMs carry a concentration risk that recycling can soften but not remove — and many South African mines are now running at or below break-even, which caps how much new metal comes out of the ground whatever the price does. Recycling recovers 20–30% of the metal the market needs, and the technology can reach 95% — but the weak link is collection, not refining. The metal that never reaches a recycler is lost not in the furnace but in the scrapyard.
Close the collection gap — the biggest lever the industry has, because the metal is already above ground.
>30% of car-catalyst PGM is lost before it reaches a refinerThrifting and substitution keep cutting the grams per converter, without losing the clean-up.
Less metal, same jobHydrogen — fuel cells (Pt) and electrolysers (Pt + Ir) — as the possible replacement for autocatalyst demand.
The open question