Nasa is moving away from the theory that foam debris seen to strike Columbia's left wing on launch may have been the factor responsible for the orbiter's loss.

Ron Dittemore, the US space agency's shuttle programme manager, expressed the view on Wednesday that the reason for the disaster probably lay elsewhere.

We're looking somewhere else; was there another event that escaped detection Ron Dittemore, shuttle programme manager

"It doesn't make sense to us that a piece of debris could be the root cause of the loss of Columbia and its crew," he told a technical press conference at the Johnson Space Center in Houston. "There's got to be another reason."

Video taken of the launch shows a large chunk of insulation foam peel away from the orbiter's external fuel tank and smash into the vehicle's left wing.

It has been suggested that the impact may have dislodged heat-resistant tiles, exposing Columbia to catastrophic temperatures on re-entry.

But Mr Dittemore said it was evident from enhancement of the video that the wing suffered "no gross, large areas of damage".

No ice

Nasa became aware of the foam incident very soon after blast-off and asked engineers to model what had happened.

SHUTTLE BREAK-UP

Re-entered atmosphere at 12,500 mph (20,000 km/h) Disintegrated 40 miles (65 kilometres) above the Earth Debris scattered over Texas and Louisiana - reports now being checked of sightings in California and Arizona

Click here for a step-by-step guide to the disaster

The engineers estimated the foam block to weigh about 2.6 pounds (1.2 kilograms) and to measure 20 inches by 16 by 6 (50 centimetres by 40 by 15). They calculated it hit the wing with a speed of about 750 feet per second (228 metres per second).

But even when these values were altered and the angle of strike was varied, there was no scenario in which significant damage would be done to the vehicle, Mr Dittemore said.

"We doubled the local air velocity, we were conservative on the weight of the material and we were using a tool that we knew overestimated damage," he added.

And he dismissed the idea that water or even ice might have entered the insulation foam, increasing its potential to do damage on impact.

"Foam does not absorb moisture," Mr Dittemore said. "It is essentially waterproof."

Losing control

Nonetheless, Nasa specialists are now carrying out experiments on the foam and the shuttle tiles to check all their assumptions about the incident are correct.

"It's hard for us to believe as engineers that this particular piece of foam debris shedding from the tank represented a safety of flight issue.

Atlantis was the next shuttle scheduled to fly

"So, we're looking somewhere else; was there another event that escaped detection. As I've said before, we're trying to find the missing link."

Columbia's sensor data have already revealed abnormal heating in the orbiter's left-wheel well and the mid section of its upper fuselage.

It is clear, also, that the vehicle was experiencing increased drag on its left side, possibly because it had sustained damage to, or even lost, a large area of tiles. Two of its yaw jets were firing to try to correct the instability.

Precisely how Columbia lost control may come out of an analysis of the final 32 seconds of data transmitted to Earth from the shuttle's sensors. But Mr Dittemore said the error rate in the data was high and investigators were currently having difficulty making sense of the information.