The Bushveld Igneous Complex in South Africa is the largest layered intrusion on Earth. It holds an estimated 70% of the world's platinum-group elements — roughly 65,000 tonnes of platinum alone. The standard model is clean: a mantle plume rose from the deep earth 2.06 billion years ago, carrying PGE from the core-mantle boundary, and deposited them in the Merensky Reef and UG2 chromitite.
The osmium isotopes say otherwise.
What the Numbers Actually Say
Osmium has seven stable isotopes. Two matter here: 187Os (produced by radioactive decay of 187Re with a half-life of 42 billion years) and 188Os (non-radiogenic, primordial). The ratio 187Os/188Os tells you where the osmium came from.
At 2 billion years ago, the primitive upper mantle had an initial 187Os/188Os of approximately 0.110 — what you'd get from a mantle that had been slowly accumulating radiogenic 187Os since Earth's formation. Chondritic meteorites — the undifferentiated building blocks of the solar system — had an initial ratio of approximately 0.141 at 2 Ga. Continental crust, rich in 187Re, produces ratios above 0.5.
When you measure Bushveld's Os isotopes, here's what you find:
| Deposit | Age (Ma) | N | Mean Osᵢ | ±σ |
|---|---|---|---|---|
| Bushveld | 2060 | 15 | 0.1450 | 0.005 |
| Vredefort (Wits) | 2023 | 7 | 0.1438 | 0.003 |
| Sudbury (all) | 1850 | 17 | 0.3613 | 0.224 |
| Sudbury (cleanest 3) | 1850 | 3 | 0.1565 | — |
| Chondrite at 2 Ga | — | — | 0.1408 | — |
| Mantle at 2 Ga | — | — | ≈0.110 | — |
The Convergence Nobody Connected
The published literature treats each deposit in isolation:
Bushveld. Schoenberg et al. (1999) and McCandless et al. (1999) measured 187Os/188Os ratios in the Merensky Reef and UG2 chromitite and found them too high for a pure mantle source. Their explanation: "crustal assimilation" — the magma picked up radiogenic Os from continental crust on its way up. To get from 0.110 to 0.145 requires 10–20% crustal contamination. It's a plausible fix for one deposit.
Sudbury. Morgan et al. (2002) and Walker et al. (1991) showed the Sudbury Igneous Complex carries a clear meteoritic Os signature. This is uncontroversial — Sudbury is an impact structure. The PGE came from the impactor. The cleanest samples (footwall breccia, least contaminated by crustal target rock) converge at Osᵢ ≈ 0.147–0.157.
Vredefort. Also an impact structure. Also sits on the Kaapvaal Craton, 40 million years after Bushveld. The Witwatersrand reefs carry PGE with Osᵢ ≈ 0.144. Barely examined in the context of a common source.
Here's the problem: contamination is an ad-hoc fix for Bushveld, but Vredefort and Sudbury — which are definitely impact-related — show the same signal. The "crustal assimilation" explanation works for one deposit in isolation. It doesn't work for three deposits across two separate cratons converging on the same non-mantle, non-crustal value.
If it's contamination, it's the same contamination in the same proportion in three different locations, in different host rocks, on different continents, 210 million years apart.
Or they all got their PGE from the same source.
What This Means
The Bushveld plume model has a problem it cannot solve by invoking contamination. The osmium doesn't look like mantle osmium. It looks like a meteorite. And the same meteoritic fingerprint appears in two impact structures — one of which (Vredefort) formed on the same craton within 40 million years of Bushveld, the other (Sudbury) on a different craton 210 million years later.
The data is sitting in the published literature. Schoenberg published the Bushveld numbers. Morgan published the Sudbury numbers. Reimold and Gibson published the Vredefort context. What nobody did was overlay them — and when you do, the mantle plume becomes an explanation in search of a dataset it doesn't match.