Based on chemical analysis of slag found on the back side of a wing leading edge fragment, investigators now believe the breach that destroyed the shuttle Columbia occurred at or very near the lower inboard corner of reinforced carbon carbon panel No. 8, very close to where a so-called T-seal was mounted between RCC panels 7 and 8.

But a member of the Columbia Accident Investigation Board said today it's unlikely Columbia began its Feb. 1 re-entry with a large hole in RCC panel No. 8. It's more likely the disaster began with a smaller breach, possibly a gap caused by a missing T-seal, or part of the T-seal between RCC panels 7 and 8.

"If it is, for example, a T-seal, you're talking about a narrow slit which eventually is going to have to start growing in size," said James Hallock, one of the CAIB members trying to pin down the exact failure scenario. "And how does it grow? Maybe it grows by having the bottom part of panel 8 break off and that's when you're now talking about a large amount of this hot gas getting in there and things happening rather quickly after that.

"With this very small hole, things would slowly transpire and take place over a period of time. And the timing is very critical. The comment we constantly keep saying to each other is gee, this craft made it (all the way) to eastern Texas. If we had an eight-inch hole (in RCC panel 8) out over the Pacific, I'm not sure we're going to make it to Texas.

"So we have to have something that has to evolve with time," he said. "So that's the part where I'm saying, I don't think we had a big eight-inch hole there initially, I think we had something that then grew with time."

The day after Columbia's launch, military radars detected an object separating from the space shuttle. More than two dozen shuttle components have now been tested at Wright Patterson Air Force Base to compare their radar "signatures" with the flight day two mystery object. Only two components match: Part of a T-seal with support hardware attached and a large section of an RCC panel.

While the bottom half of RCC panel 8 is missing, portions of the upper section of the U-shaped panel have been recovered. Investigators have found small globules of slag-like material on the inner surface of the recovered fragments. Those globules apparently were blown there by a plume of super-heated air rushing into a breach, melting insulation and support hardware and splattering molten material back onto the inner surface of the panel.

"By taking X-rays of this thing, they've been able to find a lot of very small globular shapes, spherical shapes," Hallock said. "And when they analyze them, they all turn out to be Inconel (or a related material). Why is that important? These globular shapes we're finding only behind Number 8, which is quite interesting. We're looking at doing both a chemical analysis as well as trying to understand the pattern of the deposition of all this slag. It's very interesting.

"If you look at just the chemical analysis and just the slag itself, particularly behind panel 8, it's very much talking about perhaps a breach right down in the lower part of panel 8, right near 7. Why do I say that? When you look at the layers of all the material, the molten layers that have formed, the slag itself, the lowest layer of this thing contains Inconel. Now Inconel is stuff that comes from the spanner beam, and foil insulation and the fittings.

"So it's beginning to tell us that indeed, when this so-called breach happened, that it was the fittings that were hit," Hallock said. "The reason we're saying it's more toward (panel 7) is that over on (the outboard) side we have a lot of other mounting hardware and it's all stainless steel and we did not find any stainless steel in the slag in behind panel No. 8. So that's sort of telling us a little bit about direction."

Inconel melts at about 3,200 degrees Fahrenheit.

"When you look at all of that, particularly around panel 8, it really does line up very nicely with a breach in the lower left hand corner of it," Hallock said. "The part of it that we're still trying to decide is that THE event or is that something that came along later? Because it's fairly clear from what we've seen, the bottom pieces of all these RCC panels in this neighborhood all came off. So was this before or after the initial event?

"For the most part, I'd say maybe 90 percent of us think that probably was the place where it came in but I can't really put it down as saying absolutely it is at this point. I hope to, but I don't know if I'll get to that stage or not."

CAIB Chairman Harold Gehman said the board plans to begin relocating in the Washington, DC, area beginning the week of June 2 and to hold a public hearing there on June 12. In the near term, investigators are gearing up to begin critical tests with a mockup of Columbia's left wing leading edge in an attempt to find out whether the impact of external tank foam insulation could have broken a T-seal or cracked an RCC panel during Columbia's launching.

A suitcase-size piece of debris broke away from Columbia's external tank 81 seconds after liftoff Jan. 16 and slammed into the left wing leading edge at more than 500 mph. Hallock said today an accelerometer on board the shuttle actually detected the impact. But investigators have not yet been able to prove whether the foam strike caused the breach that doomed the shuttle during re-entry.

"The issue that's the hard part in all of this thing is the cause and effect," Hallock said. "There's no question I can prove foam hit the shuttle and I think we can prove it hit underneath the left wing. ... Not only do we have a lot of movies that show that, we even have an accelerometer that was bouncing around, like plus or minus 1 G, and then suddenly went to 2 Gs at precisely the same time it looks like it hit. That I can do.

"Now, did it do damage? That is the hard part. How do you show whether it did damage then or not? If you didn't, and I don't know this, then where do I turn? So all of these things play together in a very complex type of thing. Right now, what we're trying to do, because we're looking at the calendar ... we're going to have to make some calls here."

The leading edge simulator was shipped to the Southwest Research Institute in San Antonio, Texas, Monday. A Fiberglass panel substitutes for RCC panel No. 5 while RCC-6, taken from the shuttle Discovery, has flown 30 times and is considered a close match to the panels in place aboard Columbia. Using a nitrogen gas canon, SRI engineers plan to fire foam debris at the Fiberglass panel starting May 28 to collect critical engineering data.

If all goes well, foam will be fired at the flight article - RCC panel 6 - starting June 7. Whether the foam impacts actually break anything remains to be seen. But Gehman said in the end, it doesn't matter.

RCC test section. Photo: CAIB

"The board's report is not going to be scenario dependent," he said today. "By that I mean, unlike the Rogers report on the Challenger, in which they had a specific event with a specific cause that was directly related to the accident and then they went right at how to fix that, we are conducting a much broader review of NASA here.

"Not only are we going to address the foam shedding issue, but we're going to address a dozen other issues that we're concerned about. We're going to tell them to fix the foam shedding, absolutely. But we're actually looking more broadly than that. Maybe the fact that we can't prove that the foam actually broke a hole in the orbiter, it may actually be a good thing in the long run because it really is causing us to look much, much broader at contributing factors in this accident, down to and including climate and atmosphere and leadership and management and safety programs and things like that which I believe will give a much, much deeper and broader report in the end.

"So it may be the fact we don't have a cause and effect that hits us in the head like a 2-by-4 may actually be a blessing in disguise. Obviously, we find that the process of rationalizing away the impact of foam hitting the orbiter over a period of all these years is not one that we're going to be able to live with."

Gehman said it "doesn't bother me in the least" if the CAIB fails to define a root cause of the Columbia disaster. And he pointed out the current best-fit scenario could quickly change.

"You heard Jim Hallock just now talking about T-seals and lower halves of RCC panels," he said. "I would point out that the entire lower half of RCC panel No. 9 is missing. One of the reasons we're not talking about the slag on RCC panel No. 9 is because there isn't any RCC panel No. 9. Suppose tomorrow we wake up and farmer Brown finds all of panel No. (9) out on his field? Well, there goes that theory. We can't write a report that's that scenario dependent because somebody will find the offending piece.

"On the other hand, I do believe we can do NASA and the shuttle program a world of good if we take a very broad and complex view of this and go after multiple causes and multiple flaws and shore them all up. I'm quite confident with that approach. It's harder and it may be a little bit of a challenge for us to write the report in ways people understand the point we're trying to make here. But I believe it actually (will be) a better report."

Gehman said the board hopes to have its final report finished before Congress takes its summer recess in August.