Accident investigators say the radar signature of an object seen floating away from Columbia in space is consistent with it being a piece of the left wing.
By Dr David Whitehouse
BBC News Online science editor
It is thought the orbiter's destruction came about because superheated gases experienced during re-entry were able to leak into the airframe on that side of the vehicle.
The problem started in Columbia's left wing
Radar tests now show the object seen drifting away from the orbiter on the second day of its mission could have been a leading-edge carrier panel.
This connects the edges of the U-shaped reinforced carbon panels that make up the wing's leading edge and heat-resistant tiles stuck to the wing's lower surface.
However, engineers are being cautious about apportioning blame to any shuttle component at this stage in the investigation.
There has been much speculation about the nature of the object seen to move away from Columbia early in its flight and detected on radar.
To determine what it might have been, the US space agency (Nasa) sent 29 leading-edge components and tiles from the space shuttle to the Wright Patterson Air Force Base in Ohio for testing.
A leading-edge carrier panel was the only item found to have a radar signature consistent with the mystery object.
Although photo analysis shows that during launch debris from Columbia's external fuel tank struck the shuttle's left-wing leading-edge around carrier panel six, engineers say it does not necessarily mean the thermal breach started at that location.
However, they agree the loss of a single carrier panel at the wing's leading edge could have allowed enough heat to enter the wing to cause the temperature increases seen in Columbia's telemetry.
Detailed video analysis of Columbia's launch carried out by the US National Imagery and Mapping Agency indicates the shuttle was struck once on launch, by a piece of external tank foam measuring 61 centimetres by 38 cm by 13 cm, and weighing about 0.9 kilograms.
The velocity of impact was 702 km/h.
Tests are planned at the Southwest Research Institute, Texas, in the next week or so, to propel foam debris into shuttle panels to see what damage can result.