The test was the latest and most crucial in a series of firing experiments meant to simulate what accident investigators believe happened when foam struck the shuttle's left wing during liftoff.
Nearly 100 observers, including two shuttle astronauts, watched under sunny skies as the brief countdown ended with the word "zero" and the loud pop of a nitrogen-pressurized gun.
The 1½-pound piece of foam, shot at 525 mph, cracked the reinforced carbon panel and seal and knocked both out of alignment, creating a gap of less than one-tenth of an inch between them. The crack in the panel was at least 3 inches long.
"We demonstrated for the first time that foam at the speed of the accident can actually break" reinforced carbon wing pieces, said NASA executive Scott Hubbard, the member of the Columbia Accident Investigation Board in charge of the testing.
"To me, that's a step forward, maybe even a significant step forward in our knowledge and we need to complete the test series ... to understand the whole story."
Hubbard said more analysis would be needed to show that the damage would have allowed hot atmospheric gases to enter the wing during re-entry, as investigators believe happened to Columbia.
A suitcase-size piece of foam broke off the shuttle's big external fuel tank during the January liftoff. Investigators suspect it damaged the leading edge of the left wing enough to cause the ship's destruction. Seven astronauts died when the shuttle broke up over Texas on Feb. 1.
Friday's outdoor test was conducted at the independent Southwest Research Institute. To recreate the conditions at Columbia's launch, the foam was fired through the 35-foot barrel of a gun normally used to shoot debris at airplane parts. Most of the key pieces tested — slanted at a 20-degree angle — were taken from another shuttle, Discovery.
Twelve high-speed cameras documented the experiment, six of them inside the wing, six of them outside. Some of the footage was later played back in slow motion. It showed the 22-inch-long piece of foam skidding across the panel and shattering — which is also what happened to the chunk that hit Columbia.
On close examination, the crack in the panel was visible to the naked eye.
"If such a crack had been found on an inspection, you would not fly with it. You would not take a piece that is this damaged into space," Hubbard said.
The crack in the seal, also made of reinforced carbon, was not noticed until later.
The test, originally planned for Thursday, was delayed twice, first by thunderstorms and then by a brief electrical problem Friday.
The investigation board plans to complete its report by the end of July, and some elements of the latest draft outline were reported in Friday's Orlando Sentinel. Among the board's concerns were poor risk management, questionable policy decisions and constant budget battles.
Board spokeswoman Laura Brown declined to provide any copies of the draft, and emphasized that the outline was "a work in progress" and probably would change. She said the draft, dated May 23, was already the sixth revision and stressed that it had no findings or recommendations.
Friday's test was the first in which foam was shot at the panels and seals that form the leading edge of shuttle wings.
Last week, a similar-size piece of foam was fired at a wing replica made up of fiberglass panels and seals taken from the never-launched shuttle prototype Enterprise. The parts that took the brunt of the impact were deformed by the foam.
But reinforced carbon is more brittle than fiberglass. And Hubbard predicted before the test that the foam might even shatter the reinforced carbon.