Researchers have known for years that carbon is stored in the Earth's mantle, a layer of plasticky rock that lies beneath the planet's fragile crust.

Exactly how much is down there is unknown. Most estimates, drawn from analyses of gases emerging from the mantle, say the store is many times more than all the carbon in the Earth's atmosphere, soil and sea combined.

The worry is that if just a part of this gigantic reservoir is quickly released as carbon dioxide (CO2), that could create a runaway greenhouse effect. The CO2-soaked atmosphere would store up heat from the Sun, shrivelling plant life and destroying species along the food chain.

"The (mantle) reservoir is just gigantic compared with anything that we have on the Earth's surface," says Hans Keppler, a professor at the Institute of Sciences at Germany's University of Tuebingen.

Reporting in Thursday's issue of Nature, Keppler and his colleagues conducted an ambitious experiment aimed at finding whether mantle rock is a stable storage for CO2.

Most of the rock in the Earth's upper mantle is a crystalline silicate called olivine.

In a lab chamber, Keppler's team replicated the fiery heat and intense pressures, of 1,200 C (2,200 F) and 3.5 gigapascals, which are likely to exist in the deeper parts of the upper mantle.

They used these conditions to create olivine crystals from raw ingredients of magnesium oxide and silicon dioxide and expose them to carbon and water.

The carbon turned out to be almost completely insoluble in olivine: just a tiny amount, between 0.1 and one parts per million by weight, was absorbed into the rock.

So if the carbon is not in the olivine, that leaves only one major source, Keppler says.

"If you cannot store the carbon in the olivine, then the only plausible place for storing it are carbonates," he told AFP.

Carbonate rocks have a much lower melting point than olivine, which is able to absorb the punishing furnace-like heat radiating from the Earth's core and still not melt.

Heated to a molten state, carbonates are capable of squeezing through cracks in the olivine, rising up towards the surface and absorbing the free carbon as they go.

They can pick up so much that as much as 10 or 20 percent of their mass is carbon.

The risk, says Keppler, is that this carbonate reservoir could suddenly be breached in the event of a major volcanic eruption.

"Once the carbonate comes up to the surface, as soon as it is below [a pressure of] 20 or 30 kilobars, which corresponds to a depth of 40 or 60 kilometres (25 to 38 miles) in the mantle, as soon as it comes up beyond this depth, it will decompose and release carbon dioxide."

The nightmare is of a gigantic spewing out of CO2, imperilling life on the surface.

"There has been some evidence that something like this has happened in the past. There is a very good correlation with (CO2) flooding that coincides with several mass extinction events -- some massive, sudden change of carbon dioxide in the atmosphere," Keppler believes.

One of these events, he says, occurred around 245 million years ago, at the end of the Permian era, which was the largest extinction in the Earth's history: fossil evidence shows as many as 96 percent of all marine species were lost and more than three quarters of vertebrate species on land.

The other -- possibly a cluster of smaller events -- was at the end of the Triassic period around 208 million years ago, when around half of the world's species suddenly died out.

That event essentially handed rule of the planet to the dinosaurs. Their domination ended, some 65 million years ago, by a mass extinction apparently inflicted by an asteroid impact, kicking up dust that caused climate change.

TERRA.WIRE