Analysts say the search for a suitable technology could become a $150 billion-plus market. But a big worry is that gases may leak from badly chosen underground sites, perhaps jolted open by an earthquake.

Such leaks could be deadly and would stoke climate change.

Part of the answer could be to petrify or liquefy gases like carbon dioxide -- emitted for example from power plants and factories run on coal, oil or natural gas -- if technical hurdles can be overcome and costs are not too high.

"If you can convert (the gases) to stone, and it's environmentally benign and permanent, then that's better," said Juerg Matter, a German scientist at Columbia University in New York who is working on a project in Iceland to turn carbon dioxide, the main greenhouse gas, to rock.

In theory, carbon dioxide reacts with porous basalt and turns into a mineral, but no one knows how long that takes. Matter and US, French and Icelandic experts plan to inject 50,000 tonnes of the gas into basalt in a test starting in 2009.

Other researchers say pumping the gas into sediments below the seabed at depths of around 3,000 metres (9,800 ft) would expose it to enough pressure to turn it into a viscous liquid, like honey or treacle.

"High pressure combined with low temperature results in liquid carbon dioxide that can in some cases be denser than sea water," said Kurt Zenz House, of Harvard University.

In the worst known example of a leak, a natural volcanic eruption of carbon dioxide from Lake Nyos in Cameroon in 1986 killed more than 1,700 people. Carbon dioxide is not toxic but can cause asphyxia in high concentrations.

HEATWAVES, STORMS

The UN Climate Panel said in a 2005 report that carbon storage could be one of the main ways to offset global warming, which could cause more powerful storms and heatwaves, and melt glaciers from the Himalayas to the Andes.

Carbon capture and storage could keep up to a third of all manmade carbon dioxide out of the atmosphere. But no commercial-scale power plant uses the technology yet and a lack of public funding plus legal concerns and safety worries are casting doubts over new projects.

Costs are a big problem: the Climate Panel estimates the penalty for emitting carbon dioxide would have to be stable at $25-$30 a tonne to make carbon storage viable -- pushing up the cost of everything from electricity to steel as the price of slowing climate change.

Carbon emissions' prices in a European Union market of more than 11,000 industrial sites are now above the trigger level, at 24.55 euros ($37.97) a tonne. But the volatile market has often been far below $25 since trading began in 2005.

Most plans by firms such as BP, Rio Tinto, E.ON, American Electric Power and StatoilHydro focus on capturing greenhouse gases and pumping them into porous rocks in shallower oil and gas reservoirs, or into disused mines or saline aquifers.

The longest-running commercial project, stripping carbon dioxide from natural gas at Norway's Sleipner field, began in 1986 and has pumped 10 million tonnes of carbon dioxide into sedimentary rocks about 800 metres (2,625 ft) below the seabed.

"There have been no leaks," said Tore Torp, project leader at operator StatoilHydro. The gas, under pressure, is in a "supercritical" state between liquid and gas and would stay put even in the event of a cataclysmic earthquake, he said.

The project is commercial because Norway, outside the EU, imposed taxes on carbon emissions in 1991. The natural gas at Sleipner has an unusually high percentage of carbon dioxide.

Torp said that the area was far from an earthquake zone.