Engineers for the first time today fired a large chunk of foam insulation at a Fiberglass wing leading edge panel, knocking a so-called T-seal out of place and leaving a long gap between two panels. Such a gap on a real shuttle wing leading edge would provide an entry point for deadly super-heated gas during the descent from orbit. Today's test results, assuming no problems are found, mark the first concrete evidence supporting the widely held theory that a foam impact during the shuttle Columbia's launch doomed the orbiter and its seven-person crew by creating a breach in the left wing's leading edge.

"Investigators are always a little cautious and there were no raised eyebrows, but it certainly is an interesting initial result," said a spokesman for the Columbia Accident Investigation Board.

The tests are being conducted at the Southwest Research Institute in San Antonio, Texas, using a nitrogen gas canon capable of firing large pieces of foam insulation at velocities of more than 500 mph. Earlier test runs focused on firing foam chunks at panels of heat-shield tiles like those on the belly of the shuttle. Today's test was the first using a large mockup of a wing leading edge.

The goal was to collect engineering data to ensure the validity of future tests. Investigators eventually plan to fire foam at an actual shuttle reinforced carbon carbon panel taken from the shuttle Discovery that has a flight history similar to panels that were aboard Columbia.

For today's test, a 1.67-pound chunk of foam with a volume of 1,200 cubic inches was fired at a Fiberglass panel in RCC position No. 6. The foam was fired at a velocity of 779 feet per second, or 531 mph, striking the Fiberglass panel at an angle of 20 degrees. The actual angle of impact in the Columbia foam strike is believed to have been less than that.

The leading edge panels are attached to the wing spar with so-called T-seals. Each edge of an RCC panel fits into a T-seal that is bolted to the front of the spar. The result is an alternating pattern of RCC panel, T-seal, RCC panel, T-seal and so on.

In today's test, the the foam strike caused the T-seal between RCC panels 6 and 7 to lift and pull away toward the outboard panel 7, the CAIB spokesman said, leaving an opening approximately 22 inches long. The width of the resulting gap varied from the thickness of a dime to more than a quarter inch.

A suitcase-size chunk of foam insulation broke away from Columbia's external fuel tank 81 seconds after launch, slamming into the left wing around RCC panels 7 and 8 at more than 500 mph. Engineers have long suspected the impact caused damage that left the shuttle with an open breach when the crew began re-entry Feb. 1. Today's test is the first to provide concrete evidence such an impact could actually displace a T-seal.

But the CAIB spokesman pointed out that Fiberglass is 2.5 times more resilient to impacts than the carbon composite material making up a real RCC panel. And the angle of impact was different, imparting more energy than a lower angle-of-incidence impact. How those variables are balanced in upcoming tests aiming at an actual RCC panel will go a long way toward determining whether or not a foam strike could have doomed Columbia.

Interestingly, the day after launch a piece of debris drifted away from Columbia that was detected by Air Force radar systems. Subsequent tests show the debris could have been a T-seal or, perhaps, a large section of an RCC panel. It is possible, some investigators believe, that the foam strike weakened a T-seal or cracked an RCC panel and that a piece of the damaged component finally broke free the day after launch. The Southwest Research Institute tests may provide the hard data needed to support, if not prove, that theory or some variation.

Computer modeling software developed and refined in the wake of the Columbia accident predicted damage similear to what was observed after today's test. Engineers plan to begin shooting foam at an actual RCC panel around June 10. More information on today's test will be provided as it becomes available.