Lost World

By Carl Zimmer|Wednesday, July 01, 1992
Geologists have their own Atlantis: the Earth’s earliest crust. Over the eons, colliding plates have drowned the planet’s primordial crust, plunging it down into the hot mantle whence it came. The oldest rocks known to have escaped this recycling are 3.96 billion years old. By then Earth had been evolving for 600 million years.

Now two Harvard geochemists claim they have found a clear record of Earth’s first crust--or at least of when it formed. They say it formed within 100 million years after the planet itself coalesced (along with the rest of the solar system) from a swirling nebula of gas and dust. The evidence isn’t an actual rock; it’s a chemical footprint the crust left behind.

In the primordial Earth, as in the solar nebula, chemicals were smoothly blended together. The formation of the crust changed that. Then as now, hot rock in the mantle rose slowly toward the surface, decompressed, and began to melt. But the rock was only hot enough for a few percent of its minerals to reach their melting point. It was this select fraction of the mantle that squirted through volcanic fissures, crystallized, and formed crust--a process that continues today.

The process gives crust and mantle different chemical makeups. Compared with the primordial, well-mixed Earth--whose composition is known from the analysis of ancient meteorites--the crust has a higher proportion of some elements, while the mantle has a lower one. The rare earths samarium and neodymium are especially prone to flee the mantle for the crust--and neodymium more so than samarium.

As it happens, samarium has two radioactive isotopes that decay into isotopes of neodymium at a steady rate. That makes them good clocks. By measuring the ratio of a parent isotope to its daughter, you can tell when a rock solidified and trapped both parent and daughter. Several years ago Stein Jacobsen of Harvard used one of these clocks, based on the extremely slow decay of samarium 147 to neodymium 143, to date rocks from Isua, on the west coast of Greenland. He found that the rocks had solidified from hot mantle 3.8 billion years ago, when Earth was 760 million years old.

The surprise, though, was a clue that Earth might have had a solid crust much earlier than that. The clue was in the Isua rocks’ ratio of neodymium 143 to neodymium 144, a stable isotope that is not the decay product of anything. Compared with meteorites--that is, compared with the primordial mantle--the Isua rocks’ ratio was high. Apparently they did not emerge from primordial, virgin mantle that had never given birth to rocks before. Instead, Jacobsen’s results suggested the Isua rocks formed from a part of the mantle that had already surrendered some neodymium to an even earlier piece of crust. Later they acquired more neodymium 143 from the decay of samarium, but no more neodymium 144; hence their high isotope ratio.

Now Jacobsen and his colleague Charles Harper say they have pinned down the age of this pre-Isuan crust. This time they looked at the second, much faster samarium-neodymium clock: the decay of samarium 146 to neodymium 142. Samarium 146 has a half-life of just 103 million years, and it is--was--extremely rare. By the time Earth was half a billion years old, it was virtually extinct, and since then there has been no source of new neodymium 142. If Earth was still chemically well mixed at that time--in other words, if it had no crust--then every rock in the crust today should contain the same ratio of neodymium 142 to normal, stable 144.

That’s not what Jacobsen and Harper found. The Isua rocks contained a hair (but a statistically significant hair) more neodymium 142 than other rocks do. From the size of the anomaly, the researchers calculate that crust must have started forming before Earth was even 100 million years old.

Today crust forms mainly at mid-ocean ridges, where the crustal plates spread apart; it is recycled into the mantle where plates collide. That’s probably not how the first crust formed, says Harper. If plate tectonics had been active from day one, our anomaly probably wouldn’t have survived, he says. Instead, Earth may have looked a lot like Mars or the moon, with crust forming slowly at isolated volcanoes, and meteorites continually pulverizing its surface.
Comment on this article