Rising rain. Sediments (brown) may be filling inverted "basins" at the core-mantle boundary.

'Sand' Piles Up Atop Earth's Core

Dick writes about Earth and planetary science for Science magazine.

There's probably no stranger place in Earth's layered interior than where the molten iron core meets the bottom of the rocky mantle some 2900 kilometers down. Now one group of geophysicists suggests that the boundary is a topsy-turvy world that acts like a mirror image of the ocean--instead of sinking to the sea floor, detritus rises through the molten iron and collects on the core's roof.

Weird things happen at the core-mantle boundary. For instance, patches of the mantle tens of kilometers thick act as if they are partially melted, slowing the seismic waves that pass through them by up to 30%. Even stranger: Analyses of the subtle nodding of Earth's rotation axis suggest that some sort of layer at the boundary can magnetically couple the core and the mantle. But to do that, the layer would have to be more electrically conductive than partially molten rock is.

An explanation for such strange behavior might be a slow rain of precipitates floating to the boundary, say geophysicist Bruce Buffett of the University of British Columbia in Vancouver and colleagues. The precipitates are impurities in the liquid iron that could precipitate as the core cools, much like the salts that appear when the temperature of hot brine drops. The sandlike particles are less dense than the iron soup and would rise to the top of the core. Iron trapped in the sediment pile would provide enough conductivity to explain the nodding of Earth's rotation, Buffett and his colleagues suggest in the 17 November issue of Science, and the sediments should hold enough fluid to account for slowed seismic waves.

Other geophysicists are intrigued but cautious. "It's a very plausible idea," says David Stevenson of the California Institute of Technology in Pasadena. "It's also obviously a speculative thing." One problem, he says, is that mineral physicists know so little about the behavior of materials at core temperatures and pressures. But if true, the work could end up adding a layer to the "onion" model of Earth structure that now includes surficial sediments, crust, mantle, core, and inner core.

Posted in Earth