Seabed yields mark of nearby supernova.

A stellar explosion rocked Earth's neighborhood about 5 million years ago, according to German researchers who have uprooted the first direct evidence of a supernova recent and so close.

Such an explosion would have shone 10 times as brightly as the full moon for months, and a star remnant that appeared the size of 20 moons would have smoldered for millennia. Had the supernova been much closer to Earth, its energy would have caused mass extinction, the researchers say. "What was over the years only a subject of speculation--namely, that a stellar explosion could happen very close to the solar system--now seems to be a reality," says study coauthor Wolfgang Hillebrandt of the Max Planck Institute for Astrophysics in Garching, Germany.

Hillebrandt's team sampled and dated three thin layers of a deep-sea sediment called a ferromanganese crust. The researchers sought particles of iron-60, a radioactive iron isotope produced abundantly in supernovas but having few other sources in the solar system. The group used a method known as accelerator mass spectrometry to isolate iron-60 from elements and molecules of different masses. A gas-filled magnet then teased iron-60 apart from nickel-60, another isotope of the same mass.

The handful of iron-60 ions that emerged was enough to point to a supernova, the group says. The newest layer in the sample contained 14 iron-60 ions; the middle layer, 7 of the ions; and the oldest layer, only 2, the team reports in the July 5 PHYSICAL REVIEW LETTERS. The researchers used the layers' known ages to take into account radioactive decay. They conclude that the quantity of iron-60--especially in the 4- to 6-million-year-old middle layer--indicates a massive stellar explosion, probably a type II supernova, about 5 million years ago.

Although there are other explanations for iron-60 on Earth, such as cosmic ray bombardment, the group says that none accounts for the samples' high concentration. "Everything hangs together on just those few ions," notes physicist Louis Brown of the Carnegie Institution of Washington (D.C.). Nonetheless, he says, "my impression would be that these people know what they're doing."

The German researchers say that after the stellar explosion, gaseous iron-60 condensed on dust particles, probably from inside the star. Hitching a ride on these particles, the iron-60 had enough velocity to pierce the solar wind and reach Earth. From the amount of iron-60 in the samples, the supernova must have been within about 90 light-years, they calculate.

"It's not exact, but it's a pretty close window on when and where such an event happened, so in that sense I think it's quite exciting," Hillebrandt says. He adds that further tests must confirm the results. His team is searching for iron-60 elsewhere and for other supernova products, such as plutonium-244.

Astrophysicist Stanford E. Woosley of the University of California, Santa Cruz says that the group's data are compelling but additional confirmation is essential, given the importance of the discovery.

"If this holds up, it's so astonishing that it's going to spur a tremendous amount of research," says astronomer Donald D. Clayton of Clemson (S.C.) University. "The scientific tentacles of this are very deep."