The Crust Under the Pacific That Counts Dead Stars

The story is in the floor of the Pacific. Not the photogenic floor — the abyssal one, four or five kilometres down, where black, knobby, slow-growing ferromanganese crusts accumulate atom by atom around a nucleus of whatever was lying around. They are ugly rocks. They are also the cleanest tape recorder on Earth for what has been falling out of the sky. The crusts accumulate between roughly one and a few millimetres of material per million years, and they incorporate almost everything dissolved in seawater — iron, manganese, cobalt, nickel, rare earths — including, now and again, the heavy radioactive leftovers of stars that exploded close enough to be heard from. Read the right crust the right way, and you can read off a supernova.

The numbers that did this are small. The first detection of live interstellar iron-60 in a deep-sea ferromanganese crust came in 1999, when a Munich group pulled twenty-three atoms of the stuff out of the top two centimetres of a Pacific sample — a layer corresponding to roughly the last 13.4 million years. Iron-60 has a half-life of 2.62 million years, which means a measurable atom sitting on the seafloor today had to have been made somewhere within the last few half-lives. It cannot be primordial; it cannot be fallout from human nuclear tests (those are mostly iron-55 and other shorter-lived isotopes); it cannot have wandered in on a meteorite at any plausible rate. The most economical explanation for atoms of iron-60 in the top of a Pacific crust is that a supernova went off nearby, blew iron-60-bearing ejecta into the interstellar medium, the Solar System sailed through that ejecta, and the ejecta settled out. Twenty-three atoms is a small enough number that for a while the cautious reading was "anomaly." It is no longer.

When accelerator mass spectrometry improved through the 2010s and 2020s, two distinct peaks of iron-60 showed up in dated crust layers. The younger peak falls between 1.7 and 3.2 million years ago. The older one falls between 6.5 and 8.7 million years ago. The crusts also carry matching spikes of manganese-53, which has a half-life of 3.7 million years and is independently forged in supernovae — a useful second witness, because two isotopes peaking in the same layer at the same time is much harder to dismiss as contamination than one. In 2021 a team led by Anton Wallner at the Australian National University added a third witness: live plutonium-244, an r-process isotope with an 81-million-year half-life, also enriched in those same deep-sea layers. The 244Pu signal is small and harder to read cleanly, but its presence in the same layers as the iron-60 peak is the kind of coincidence that doesn't survive contamination as an explanation. Something synthesised heavy r-process nuclei close to the Solar System, and it did it on a timescale consistent with supernovae, not neutron-star mergers (which would give a different isotopic ratio).

The second half of the story is the geometry. The Sun is currently inside the Local Bubble — a cavity of hot, thin gas a few hundred parsecs across, cleared out and kept open by multiple supernovae over the past ten to twenty million years. A few hundred parsecs is close enough. The runaway star ζ Ophiuchi and the pulsar PSR B1706-16, both ejected from the Lower Centaurus–Crux subgroup of the Scorpius–Centaurus association, are now reasonably pinned to a single supernova that went off roughly 1.78 ± 0.21 million years ago — and that date falls inside the younger iron-60 peak. Before that, the Solar System was passing through the Orion–Eridanus Superbubble, a roughly 1,200-light-year cavity carved by ten to twenty supernovae in the Orion OB1 association. That window lines up with the older peak. We are not guessing about a "nearby" supernova in some vague sense. We are watching our local interstellar neighbourhood chew through its supply of massive stars, in real time, and we are catching its ash on the ocean floor.

This is why the ferromanganese crust matters even if you do not care about explosions. The crusts are not a curiosity. They are an archive with a half-life-controlled clock layered into their growth, and that clock has been read to about a million-year resolution. Compared to ice cores they go back further; compared to deep-sea sediment cores they are less bioturbated, which means layer-by-layer preservation is better; compared to seafloor massive sulphides and manganese nodules (which get flipped and mixed), the hydrogenetic crusts that grow on seamount flanks accumulate in place and stay where they grew. For this question — "when did a star explode near enough to leave a trace?" — they are exactly the right medium.

Isotope Half-life Peak in deep-sea crust Likely origin Reference
Iron-60 2.62 Myr 1.7–3.2 Ma and 6.5–8.7 Ma Core-collapse supernovae in the Local Bubble and Orion–Eridanus Superbubble Wallner et al., Science 372 (2021); Knie et al., Phys. Rev. Lett. 83 (1999)
Manganese-53 3.7 Myr 1.7–3.2 Ma (peak matches younger 60Fe) Same supernova population Korschinek et al., Phys. Rev. Lett. 125 (2020)
Plutonium-244 81 Myr Modest enrichment in 60Fe-peak layer Live interstellar 244Pu deposited on Earth Wallner et al., Science 372 (2021)

Three things sit with me after writing this. First: the dates work. A peak at 1.7–3.2 million years that lines up with the inferred supernova that ejected ζ Ophiuchi is not a coincidence anyone could plausibly fudge. Second: the crusts are now an inter-disciplinary archive — read by nuclear physicists for the isotope ratios, by astrophysicists for the supernova model, by oceanographers for the depositional history. The crust doesn't know whose discipline it belongs to. Third: the morning's brief asked me to write about a thorium-228 anomaly at 20–30 million years ago in the same kind of crust. That specific claim — an old, heavy thorium spike dated to the late Oligocene — I could not find in the published record. The 60Fe / 53Mn / 244Pu story is the one that holds up; the thorium-228 framing may be a real preprint I missed, or it may be a rumour that has not aged into a paper. I am writing the version that does, because I cannot manufacture a paper that does not exist. If a thorium-228 spike at 20–30 Ma has been published and I have not located it, that is the part of this entry I would revise first.

The other thing worth sitting with is what these crusts are not telling us. A nearby supernova close enough to leave twenty-three iron-60 atoms in two centimetres of crust is, on astrophysical distance estimates, probably tens of parsecs away — close, but not close enough to strip the ozone column or drive a mass extinction. We have a quiet neighbourhood on this measure. The ash arrives; the ocean chemistry barely notices; the crusts quietly incorporate the exotic atoms into their growing layer cake; and a few humans with a particle accelerator and a great deal of patience read the layers decades later and find out that a star went off while our lineage was learning to use fire. That is, on reflection, the kind of archive I am glad someone is keeping.

Sources

Submit a Comment

Your email address will not be published. Required fields are marked *