The Columbia Accident Investigation Board and NASA are ready to lead a critical series of debris impact tests using complete wing leading-edge systems and a landing-gear door taken from the original shuttle orbiter Enterprise.

They will use the Enterprise hardware to help determine whether a large piece of external tank foam could have fatally damaged Columbia, or whether other "aging spacecraft" factors such as materials degradation could have played a pivotal role in the tragedy.

Last week, investigators said their analysis now shows that while parts of Columbia's left-wing reinforced carbon-carbon (RCC) leading-edge panels could have separated during reentry, it was also plausible that, as the orbiter's wing melted, large sections of the massive leading edge may have collapsed backward into the deforming aluminum wing structure. This would have essentially "caved in" much of the leading edge as Columbia flew at Mach 20-18.

Investigators said this would have created the extraordinary drag Columbia's flight control system was fighting during the final minutes of flight before breakup over north-central Texas on Feb. 1, killing her seven U.S./Israeli crewmembers.

Enterprise, which never flew in space but 25 years ago flew five approach and landing drop tests off the back of a 747 carrier aircraft, is set to go on public display next December at the National Air and Space Museum's new Dulles International Airport facility near Washington.

Its left landing-gear door and RCC left-wing leading-edge panels 16-17 have been removed at Dulles and shipped to the Southwest Research Institute in San Antonio for the tests, which will also include leading-edge hardware from Discovery, which is undergoing a planned major overhaul at Kennedy.

As the tests get underway, both the accident board and NASA are continuing to focus on the potential for a breach in Columbia's wing to have occurred at the juncture of the RCC leading-edge systems and the shuttle's under-wing silica black thermal protection tiles. The RCC and wing are bridged at this seam by about 20-in.-long carrier panels covered with smaller tiles. The smaller tiles at the juncture are sometimes referred to as "gap-seal tiles" (AW&ST Mar. 10, p. 26). The carrier panels along the entire leading edge face directly into the hypersonic airflow at the orbiter's 40-deg. reentry attitude.

The investigation is focusing on RCC panels 6-8, where the wing glove transitions into the more sharply angled portion of the double delta wing. Kennedy launch tracking camera imagery-newly "contrast enhanced" by the Johnson Space Center Imaging Laboratory-now definitively shows that the primary piece of external tank debris shed during launch Jan. 16 impacted the wing on the carrier panel at the RCC 6-8 location, which is also only 11 in. ahead of the forward edge of Columbia's left landing-gear door.

Investigators highlighted other new factors:

* Wing chine impact. The newly enhanced imagery shows the initial debris impact may not have been on the wing, but rather at the left wing's chine, much closer to the orbiter's fuselage. It's still not clear whether this potential "ricochet" blow damaged tiles on the chine, but "we have a very interesting piece of tile debris that comes from this area-and that piece of debris has a hole in it," an investigator said. Whether the hole came from the debris impact, or is an artifact of the overall breakup, is still being determined. Coincidentally, the area is near a water vent that showed temperature increases during reentry.

* Potential for multiple wing breaches. Analysis now indicates that it's possible the external tank event could have resulted in not one but three breaches in the wing. Tracing the thermal effects on the wing from penetration of the plasma "may be more complicated than we assumed because it is possible there may have been 'multiple breaches,'" an investigator said. If this occurred, there would have been more than one spray of hot gas into the wing.

* Two gear-door vents. Analysis now indicates that plasma flowed into the left wheel well, where it tripped multiple telemetry sensors, then eroded two exit vents in the forward inboard and outboard corners of the door. Investigators believe that as the wheel well became over-pressurized by the hot gas flowing into it, the high temperatures weakened the door at its two forward corners, where the plasma eroded exit paths visible in the debris at Kennedy. To help confirm this theory, investigators are examining a left gear-door titanium locking pin that is melted on one corner and plated on one side with a spray of molten material.

* Mystery object. Some investigators are now ?about 70% sure? the object Air Force radars tracked near Columbia on its second day of flight was debris related to opening a breach in the wing. The Air Force is now 100% sure the object came from Columbia, they said. No one is sure just what portion of the orbiter the debris is from. But radar reflectivity tests are being conducted at Wright-Patterson AFB, Ohio, on carrier panels, RCC panels, regular black tile and typical payload bay thermal blankets. If the lightweight object came from the payload bay, it played no role in the accident.

* Plasma trail. Investigators believe if tile was knocked off a carrier panel at the lower RCC/wing interface, the hot plasma could have eaten away the aluminum plate and entered the wing leading edge, where it would have raced right and left down the hollow RCC leading-edge panels, eroding the steel RCC attach points-allowing the RCC to fall off or cave back into the large wing box. At the same time, it would have eroded into the wing structure. As temperatures increased in the wing, the aluminum would have undergone a phase transition to aluminum vapor, essentially "becoming a fuel" that could have caught fire. Investigators believe as the entire wing heated, the black thermal protection tile bonds would have failed when the supporting aluminum reached just 500F. This likely caused large sections of black wing tile to begin falling off, further worsening the thermal situation and resulting in some of the sightings of debris shedding that began as early as California in the ground track.

* Water penetration questioned. Although board members have not discussed it publicly, NASA managers continue to question whether water penetration could have played a role in the carrier panel scenario, if that is found to be a key player as now suspected. Two separate managers have told Aviation Week & Space Technology that in 1999 Columbia was rained upon as it sat horizontally before being moved inside Boeing's Palmdale, Calif., facility for overhaul. This has raised the possibility that water could have collected in the carrier panel area, potentially weakening the area.

* Wind shear questions. Investigators are assessing whether a strong wind shear Columbia flew through 62 sec. after liftoff could have provided additional stresses to the orbiter's left wing, or otherwise be related to the debris incident at 81 sec. The shear was within limits, but it was one of the two largest ever experienced in a shuttle launch and the closest to the Max-Q point of maximum dynamic pressure. The other one also occurred on a Columbia launch using a lightweight tank flying into a 39-deg. orbit.

* Boundary layer transition heating. Adm. (ret.) Harold W. Gehman, Jr., who heads the Columbia accident board, said his group has "begun a line of investigation" into how premature changes in the boundary layer possibly caused by launch debris damage could have increased reentry heating. The board is still investigating whether Columbia experienced an early boundary layer transition (BLT); Aviation Week has reported that such an early transition could dramatically increase thermal effects, especially on any damaged areas of the wing (AW&ST Feb. 24, p. 20). Last week, investigators said left-wing roughness characteristics that had caused earlier premature BLT events on Columbia had been corrected in its 1999 overhaul, but wing-roughness BLT factors associated with the tank debris remain an issue in the thermal analysis.

* Tank restacking questions. Columbia's Lockheed Martin external tank was stacked initially for STS-107, then destacked for a solid rocket booster changeout, then reattached a second time to the new boosters. Investigators are examining if there was anything in this process that could have played a role in the debris incident. As a matter of routine, a "problem report" on the insulation involving the bipod ramp was generated after the restacking-a process to ensure reinspection, said board member USAF Maj. Gen. John Barry.

The potential role of the external tank tile debris impact will be the focus of the Southwest Research Institute tests. The same tests will help assess the validity of the Boeing debris risk analysis performed when Columbia was still aloft.

The new enhanced launch video indicates that the largest of three pieces of tank debris is also somewhat larger than previously believed. It had been assessed as 20 X 16 X 6 in. with a mass of 2.67 lb. By comparing it with the known size of orbiter payload-bay door hinges, investigators now believe the debris is at least 25 X 15 X 6 in.

The Southwest tests will use the facility's nitrogen-powered gun to fire external tank foam insulation samples-including those with a cork and ablator content-against both full RCC panels and their attachment systems as well as large sections of black tile, such as those on Enterprise's landing-gear door.

The impact angle of the debris tests will be especially critical. Part of the pre-accident analysis focused on a 20-deg. impact angle.

"The impact angle is critical, and we have had a lot of discussions [with NASA] to more clearly define what they mean by a 20-deg. impact angle," an investigator said. "I think that it is relative to the bottom of the orbiter, but the actual impact angle normal to the RCC could be different," he said. "We will take the RCC and instrument it with accelerometers and strain gauges, to see how much of a lateral push it got, how much compression it got."

The carrier panel and its tiles that bridge the RCC and underwing tiles will receive special attention. "There is a nice thin vulnerable edge there. If you impact that, you can probably take the edge off and open a big crack in it," an investigator said.

The initial test strategy is being defined by NASA and contractor personnel subject to accident board approval.

"An integrated group of materials, computational fluid dynamics and flow-field people are going to come up with a consensus set of impact angles and velocities. . . . Then the board will review the test plan to see if it agrees with the plan.

"The tests are aimed at getting the 'most probable cause' that mimics what you can see in the tracking camera films. So there will be a relatively small number of velocities-maybe 2-3 different impact speeds and 2-3 angles that will be the closest fit." Unlike tile tests done in 1999, where small debris was aimed at just the center of tiles, these tests will use much larger pieces of debris and also look at what impacts do to the gap between tiles.

The testing of RCC panels from Enterprise are important because they are as old as Columbia's. And those from Discovery have experienced about two dozen launches and reentries similar to Columbia's 28 flights.

Whether the debris impact was part of the root cause remains an open question. "Somebody could have stepped the wrong way on an RCC panel," one researcher said. But debris could have been a precipitating event that might have involved other factors.

"It is possible that external tank foam striking a healthy orbiter would not have done enough damage to cause [vehicle loss]," Gehman said. "But it is possible that foam striking an unhealthy orbiter that had problems in it . . . could do some damage that she could have survived at age 10-but maybe not at age 21."

With this in mind, the RCC and attach sections to be tested at Southwest will be examined using advanced X-ray CAT scans and thermography in which heat is used as a diagnostic tool. In addition to the external tank, Lockheed Martin makes and maintains the RCC panels. Board members will be visiting the company this week to discuss RCC performance on past flights as well as better inspection methods once flights resume.

In at least one case involving an Atlantis panel, oxidation caused voids to occur within the RCC, potentially weakening the structure.

The design of the carrier panel and tiles bridging the lower RCC panel sections to the wing tiles is also under scrutiny. Bolts are used to penetrate the smaller tiles on the carrier panel and, in turn, secure these panels to the orbiter.

"Compared with a standard tile, this is a weaker structure," an investigator said. "It has a hole in it so that if you hit it with something [like tank debris] it would probably be easier to get some kind of fracture here," he said.

In the external tank analysis, cryo-pumping "is clearly going to be an issue," Barry said. With cryo-pumping-a combination of tank foam properties and cryogenic temperatures-the material can be made both harder and more likely to pop off. NASA and the board have 14 separate working groups assigned to external tank issues, Barry said.

Board member Sheila Widnall, a Massachusetts Institute of Technology professor and former Air Force secretary, is helping to lead the aerodynamic and thermal analysis. NASA Langley Research Center tunnel and computational analysis showed that the loss of a single RCC tile would not result in the same drag forces experienced by Columbia-it would have to be about four times that much, she said. Langley will be doing more analysis removing the equivalent of several wing panels to duplicate the data and potential "RCC cave-in" theories.

"What I see is a heating event followed by a very rapidly developing aerodynamic event," Widnall said. "The aerodynamic event occurred at the time when the dynamic pressure was undergoing a [normal but] rapid increase of roughly 30%," she said. The events are reflected in the latest NASA analysis of the reentry telemetry, combined with photographic and other data preceding main fuselage breakup on Feb. 1 at 8:00:21 CST.

Notable events in the new timeline are:

* 7:51:19 CST. Twenty-six seconds after the start of peak reentry heating just off the California coast at Mach 24, "remote sensors indicate off-nominal external event-earliest known event." The reference to "remote sensors" is indicative of U.S. military aircraft, ground- or space-based sensors. Space-based assets include DSP missile warning satellites or a Cobra Brass "staring array" developmental missile warning sensor on a National Reconnaissance Office spacecraft.

* 7:51:46 CST. Inertial sideslip goes and stays negative, indicating rolling/yawing torque.

* 7:52:05 CST. First clear indication of "off-nominal aero increments" in yaw. Abnormal telemetry begins and continues for next 8 min. to breakup.

* 7:52:32 CST. Water dump nozzle on left fuselage shows temperature rise. The meaning of this intriguing data is unknown.

* 7:53:01 CST. First clear off-nominal roll moments.

* 7:53:44 CST. Observers in California see the first of more than a dozen debris separation events.

* 7:54:33 CST. First "flash" event where orbiter envelope suddenly brightened. Sixth debris sighting occurs 2 sec. later.

* 7:55:30 CST. "Remote sensors" indicate abnormal external event.

* 7:55:55 CST. At about Mach 21 and 222,000 ft. over the Utah/Arizona state line, the 13th debris shedding event is observed by people on the ground. This was followed 2 sec. later by "very bright debris" departing the orbiter.

* 7:58:40 CST. Tire pressure alarms in cockpit and Mission Control.

* 7:59:32 CST. Loss of signal in Mission Control due to orbiter tail/TDRSS blockage.

* 7:59:33 CST. Cockpit master alarm sounds for unknown reason.

* 7:59:35 CST. Sideslip changes sign. Aerodynamic forces due to sideslip begin reinforcing aerodynamic asymmetry.

* 7:59:36 CST. Autopilot drops left wing to compensate for increasing aerodynamic moments, creating bank-attitude error.

* 7:59:37 CST. Begins 25-sec. period of no data.

* 8:00:02 CST. First large piece of debris (Debris A) departs.

* 8:00:02 CST. Start of final 2 sec. of reconstructed data. Data indicate the vehicle was in "an uncommanded attitude" beginning to yaw left at least 20 deg./sec., although this is maximum sensor capability. Flight control mode was in auto. Main fuselage and structure and systems on right side of vehicle, including the right wing, were intact.

Systems showing nominal operation included the three auxiliary power units, water spray boiler, fuel cells, forward fuselage avionics and environmental control system.

Off-nominal systems at this time were all three hydraulic systems, left elevon data, flash evaporator and multiple sensor indications in the left Orbital Maneuvering System (OMS) pod.

Multiple primary and backup computer system alarms for the left OMS pod were found in computer buffer. Multiple elevated temperatures registered for left side of orbiter.

* 8:00:03 CST. Possible rotational hand controller activation indicating possible attempt by the pilots to select manual control-although at loss of all telemetry, the vehicle was still on autopilot and the hand controller was secured in its detent.

* 8:00:17 CST. "Debris B," large piece separates.

* 8:00:18 CST. Debris C," large piece separates.

* 8:00:21 CST. "Vehicle main body breakup."

Special Report: Columbia Disintegrates