Return to Master Table of Contents
 

What's the best guess as to what caused the breakup?

While it's still too early to tell, between the information confirmed by NASA officials during the press conferences and some of the "revelations" made by the news media in the days following the loss of Columbia, a "best guess" isn't as easy as it was when it first appeared to have been caused by External Tank Insulation Foam damaging some tiles and/or the wheel well hatches. Keep in mind that as the investigation continues, facts and findings are subject to change. As of this writing, quite a number of facts, guesses and serious speculations have been made public by NASA, the CAIB, and the news media. Some have been discounted since first being reported, and some that have been discounted have been revisited in light of new data resulting from the investigation.

What appears below are summaries of the more credible pieces of information, arranged somewhat in chronological order. Following the summaries, the reader will find a general "best guess", as derived from the most credible data available so far:

As of this writing, the following facts and serious speculations have been made public:

• First on the list is the External Tank Foam strike. The most visible piece of evidence seen so far, NASA ground-based ascent tracking cameras captured a sizeable fragment of material believed to be 20 inches and 2 1/2 pounds (0.5 m ~ 1 kg) by in size - either External Tank insulation or condensational ice - breaking loose from the Shuttle's external fuel tank and impacting the leading edge and/or the underside of the Shuttle. The point of impact appears to be along the left wing, either on the leading edge where the wing meets the fuselage and/or the underside of the left wing near the wheel well - which is just below where the wing meets the fuselage. The current belief is that the impact of the foam may have damaged one or any combination of the following areas:

• The leading edge of the left wing.
• One or more of the protective tiles on the underside of the wing.
• The seals along the edge of the wheel well hatches that protect the landing gear during reentry.

A failure in any one of these elements in the Thermal Protection System could lead to catastrophic failure during reentry. A combination of failures would increase such chances of catastrophe by several orders of magnitude depending on the nature and locations of the failures.
 

It should be noted that as of the 2/5/03 press conference, NASA officials were downplaying the possibility that the debris that impacted the underside of Columbia was ice, but was more than likely a chunk of foam insulation. The determining factor at this time appears to be the nature of the way the debris disintegrated after impact as shown on the footage taken by the ground cameras during ascent. At the same time, Shuttle program manager Ron Dittemore was backing off somewhat on the possibility that any sort of impact damage from either ice or foam insulation striking the underside of Columbia was responsible for the mishap. Dittemore also stated that investigators are now asking if there was "another event that escaped our attention'' that might have caused Columbia to break up just minutes before the end of its 16-day mission, killing all seven astronauts.

Note also that 02/21/03, NASA released a new set of internal e-mails which continued concern about potential tile damage. Included in the e-mails was a post-launch analysis by Boeing engineers which shows that as many as three pieces of debris may have sheared off from the External Tank and impacted the orbiter during ascent. Based on their analysis, the Boeing team concluded Columbia could safely land despite potential damage to the shuttle's fragile heat-shield tiles.

• In a very important revelation, AvLeak broke a story on 2/7/03 revealing that high-resolution images taken from a ground-based Air Force tracking camera in Nevada showed what appeared to be serious structural damage to the inboard leading edge of Columbia's left wing. These images were taken as Columbia flew overhead ~60 seconds before the vehicle broke up over the Dallas-Fort Worth Metroplex area in Texas. One of these images was finally released to the public during the afternoon NASA Accident Review Briefing on 2/7/03. The raw image, along with some enhanced versions, can be found at the following URLs:
   

  Raw image from TV Capture Card	Image with Color Data discarded & Contrast Enhancement
  Image with Color Data discarded, Contrast Enhanced, and Black/White Inverted:	Same as Previous, with Annotations

One caveat on these images: the "stars" you see are actually artifacts from TV interference that most capture cards experience, and some of the "blockyness" may also be due to the same. We can probably expect cleaner images of this as time passes. In fact, here they are:
 

Raw Image (Courtesy CAIB)	Contrast Enhanced with Black & White Inverted	Contrast Enhanced with Black & White Inverted, with Annotations

The image provided by the Air Force came from a high-resolution digital videotape taken by the Starfire Optical Range (SOR) at Kirtland AFB near Albuquerque. The imagery is infrared, and according to sources inside NASA there's about a minute of imagery captured at 25 frames per second during the time frame when the temperature sensors on the left wing failed.

This site has some info about the SOR:

http://www.fas.org/spp/military/program/track/starfire.htm

According to AvLeak, sources close to the investigation state that the images are currently under analysis at the Johnson Space Center in Houston, and reportedly show a "jagged edge on the left inboard wing structure near where the wing begins to intersect the fuselage,", and is believed to be consistent with telemetry. The images also reportedly show the orbiter's right aft yaw thrusters firing in an apparent attempt to correct the vehicle's attitude that was being adversely affected by the added drag caused by the left wing damage. They also reportedly show that Columbia's fuselage and right wing appear normal, and that unlike the damaged and jagged left wing section, the right wing appears smooth along its entire length.

The ragged edge on the left leading edge would most likely indicate that either a small structural breach, such as a crack, dent or tear, occurred from the impact with the External Tank insulation foam during ascent, or that a small portion of the leading edge fell off at that location. Either would have allowed reentry heat at 3000° F (1650° C) to flow thru the wing infrastructure and erode the wing from the inside. This could explain why the sensors near the landing gear bay began showing the staggered failures in the pattern that that was recorded prior to Columbia's breakup.

• It should also be noted that the leading edge where the jagged shape was observed in the images reportedly transitions from black thermal protection tiles to a much different mechanical system made of reinforced carbon-carbon material that is bolted on, rather than glued on as the tiles are. This implies that in addition to the possible failure of black tile at the point where the wing joins the fuselage, a catastrophic failure involving the attachment mechanisms for the leading edge sections may have occurred. While this failure may or may not have been related to the debris impact during ascent, it does appear to account for the recorded rise in temperature along the fuselage adjacent to where the left wing connects. The leading edge is designed to protect against this, and a damaged or especially a missing section would compromise this ability dramatically.

The actual front structure of a shuttle wing has a flat edge. To provide a "blade edge" aerodynamic shape and as part of the heat protection system, each wing is fitted with 22 U-shaped Reinforced Carbon-Carbon (RCC) leading-edge structures. The carbon material in the leading edge, as well as the Shuttle's nose cap, is designed to protect the shuttle from temperatures above 2,300F during reentry. The U-shaped RCC sections are attached to the wing "with a series of floating joints to reduce loading on the panels due to wing deflections. These joints are referred to as "tee seals", and allow for lateral motion and thermal expansion differences between the carbon sections and sections of the orbiter wing that remain much cooler during reentry.

Naturally, any breach or compromise of this leading-edge material would have a high probability of producing a catastrophic failure of the wing, which is what appears to have happened. From this point, excessive heating of the wing infrastructure through the crack, combined with atmospheric stress, appears to have resulted in catastrophic structural failure to the left wing, which based on the footage taken by amateur photographers and the news photographers of ABC affiliate WFAA appears to have then sheared off. Again, as reported, before, during and after the wing failed, what appears to be RCS thruster firings in an attempt to retain control over the Shuttle, are visible. These are followed almost immediately thereafter by a brief brightening of the main body of the Shuttle, followed by the apparent breakup of the Shuttle into multiple fragments.

• On 2/8/03, CBS News broke a report stating that an unspecified military radar system may have detected an object separating from Columbia on 1/17/03, about 24 hours after launch from KSC. NASA sources have confirmed they are looking at the radar data, which apparently shows a small object suddenly separating from the shuttle at about five meters per second, or roughly 11.2 mph. At the time of this writing, NASA officials have not determined the source and/or nature of this object, but have acknowledged the following potential sources:
   
  • The object is possibly a chunk of ice created by a routine waste water dump. The waste water port is on the left side of the orbiter, and in most cases when waste water is dumped overboard it usually turns instantly into a rapidly expanding cloud of ice crystals more commonly known as the "Constellation Urion". In addition, large plugs of ice have developed on the water dump nozzles as well, the first occurrence of which produced the "Cosmic Space Turd Scare" during an early 80's shuttle flight.
    
    At the time of this writing, no recorded routine water dumps on any previous mission have resulted in a radar "signature" even remotely similar to the one observed. On the other hand, although an ice plug having been blown off during an otherwise routine water dump would be a more likely possibility of having shown up on radar as the object moving away from the orbiter. Neither of these frozen waste sources have been confirmed or denied by NASA, although both are currently under investigation.
     
  • It's also possible some piece of non-critical hardware, such as something from the cargo bay, was somehow released or ejected from the Shuttle without the crew's knowledge. Although NASA officials have declined to speculate on the nature and/or source of such an ejection, the possibility has existed on practically every single manned space flight. In fact, during Ed White's EVA on Gemini IV, a thermal insulation glove floated outside the capsule and went off on an orbital trip of its own.

  One thing to consider is that regardless of which of these radar sources turns out to be the most likely suspect, there is at this time no evidence that the object is the cause, or even related to the cause, of the loss of Columbia. As with any investigation into a mishap such as this, all factors must be considered and evaluated.
   

• On 2/9/03, MSNBC broke another story regarding the RCC on the leading edge of the left wing. According to veteran space shuttle engineer Ray Erikson, Undetected corrosion of the leading edge spars” on the left wing may have already weakened the structure enough to allow the small additional damage from the debris to cause the later catastrophic failure.
  
  In this suggested scenario, Erikson theorizes that the loss of several tiles just behind the left wing leading edge led to significant heating that, while not catastrophic by itself, was conducted into the support structure for the leading edge panels. Despite its necessity in airframe design due to weight considerations, aluminum is notorious for weakening rapidly at even relatively small temperature increases. As stated previously in this section, the most severe entry heating for a Shuttle is on the nose and along the leading edge of the wings - especially where the wings join to the fuselage. At this point, reinforced carbon-carbon (RCC) panels are bolted on with shuttle tiles arrayed downwind of them.

The problem, according to Erikson, is that the bolts were composed of a material that reacted badly with the aluminum mounting structure under certain conditions. The bolts themselves were made of Inconel 718 alloy and of A286 corrosion-resistant steel (CRES). The, and by themselves they were not affected by the salt water. But the support structure, fabricated from 2024 aluminum honeycomb, did corrode when salt spray from the nearby ocean was dissolved in rainwater and seeped into the structure while Columbia was on the launch pad. When exposed to salt water - such as that which permeates the air at KSC year-round - the metal in the bolts and the metal in the mounting structure behaved not unlike a weak chemical battery. And like a battery that's been left in a flashlight too long, corrosion set in and led to a gradual decomposition of the aluminum support structure.

Such corrosion was reportedly detected as early as 1983, according to Erikson. Corrosion was detected around the attach points of the panels, and was discovered during careful post-flight interior inspections of the wings. Aluminum splints were glued to the corroded areas, a process that was repeated for several years while Erikson was working at the launch site until he left NASA in 1985.

When combined with any pre-existing corrosion, Erickson theorizes a bolt could have broken free and allowed one leading edge panel to separate. Without the leading edge to provide both smooth air flow over the wing edge and distribute heat accordingly, extensive damage and eventual vehicle loss was, according to Erikson, inevitable.

It should also be noted here that Erikson has also stressed that the source of the corrosion was unique to Columbia and that the wing designs of the remaining three Orbiters - Discovery, Atlantis, and Endeavour - had been improved to the point that they are immune to this particular hazard and hence need, no modifications to fly safely in the near future.

On a side note, it has been speculated that should this corrosion problem in fact be proven to not only exist, but to have also contributed to the loss of Columbia, this would add a further point of order against upgrading Enterprise to active status. Both Columbia and Enterprise used the same aluminum-based parts in their leading wing edges.

• Note: before quoting this next subsection, make sure you read it completely, as it's included only as an example of how facts and reality can take some interesting shavings by Occam's Razor.
   
• On 2/5/03, this FAQ first reported that according to a first draft of the timeline of events according to NASA, Columbia sensors reported a Left Main Gear down indication as the orbiter flew over West Texas. Since that report, NASA has confirmed the existence of a short exchange of internal e-mails that occurred on 1/28/03 between officials at JSC and engineers at the Langley Research Center (LaRC) regarding potential outcomes in the unlikely event that the landing gear door or wheel well was breached during the orbiter’s reentry into Earth’s atmosphere.
  
  These e-mails can be found in their entirety at:

http://www.nasa.gov/columbia/COL_landgear_email_030212.html

The summation of these e-mails show that the concern was focused on the possibility of the landing gear *not* deploying due to damage from plasma entering a "a substantial breach of the wheel well."

On 2/13/03, the Orlando Sentinel apparently received a copy of the same rough draft of the timeline as reported in this FAQ, and reported the timeline entry regarding the gear down reading, and oddly claimed exclusivity on the report. Since then, several news services have pointed out the Sentinel's error, and this incident has prompted a review of the data both by NASA and those involved in compiling the FAQ.

This has led to the following addition to this section dealing with the "best guess".

The theory that's been proposed by those who see the gear down indicator as significant is that the plasma entering whatever breach in Columbia's Thermal Protection System - whether thru a section of  the underside exposed by a dislodged tile, a breach in the leading edge RCC, or a crack in the seals surrounding the landing gear wheel well hatches - set off one or more of the pyros designed to assist in deploying the landing gear if normal procedures fail. In this theory, the following events occur:

1. Plasma at 3000°F (1650°C) enters the wheel well.
    
2. The plasma ignites one or more of the pyros. At this point, one of the following results occur:
    
   1. The wheel well hatches are blown away, accounting for the increase in drag along the left side of the Orbiter.
       
   2. The left landing gear deploys through the wheel well hatches, accounting for the increase in drag along the left side of the Orbiter.
       
   3. The plasma subjects one or more of the tires to heating intense enough to cause the inflation gas inside to expand at an explosive rate, producing an explosive "blow-out" that caused massive damage to the wheel well hatches and produced the results listed in a).
       
   4. An as yet undetermined combination/synthesis of a) b) and c).

Since this theory was first postulated, it has come under some scrutiny thru peer review. Those who are very familiar with the workings of the Shuttle - both amateur and professionally, including some who either work for NASA, or have contracted with them in the past - have asked that the following observations be included in the FAQ with regards to the landing gear having been deployed accidentally:

• First off, there are pyros in the landing gear assembly, but they cannot go off unless charged. You'd need at least two independent failures - in fact, probably more like three to five - to inadvertently deploy the gear. The flight controllers would have seen the charge and discharge voltage signatures. If the pyro had fired, the Left Main Gear would have deployed, which would have generated a half dozen or so telemetry signals. At this time, it does not appear that the pyros fired as no telemetry was received by MCC.
   
• Under normal circumstances, the landing gear is hydraulically deployed. There is a latch that holds the gear in place, and when the landing gear Arm and Down buttons are pushed, various relays close and a valve opens, letting hydraulic fluid enter a line that causes the latch to open. The gear then free-falls down. The gear and the door are physically connected through linkages, so you can't have (for example) the gear drop and not come out because the door got stuck - as the gear drops, the linkages cause the door to open. Again, at this time, based on telemetry the door did not open or fall off.
   
• For whatever reason, if the gear is not hydraulically deployed when the buttons are pushed, there's a proximity sensor within the wheel well that determines whether the gear is still up after one second. If it is, then commands are sent via the proximity sensor electronics to fire the pyro, which basically blows the latch off and allows the gear to drop. There are two pyros on each main gear for redundancy. In addition, because the Nose Gear is lighter and gets more dynamic pressure from being at the nose, there's an extra pyro that always fires to help it deploy, even when the gear is hydraulically deployed.

For more info on the gear, check out the following link:

http://spaceflight.nasa.gov/shuttle/reference/shutref/orbiter/lgear/overview.html

These are the pyros in the wheel well that Ron Dittemore was referring to in the various press conferences since the loss of Columbia. There are 7 in all (3 nose, 2 left, 2 right). There are others that will blow the side hatch (5 total), jettison the RMS (even though the RMS was not installed, it is not known at this time whether the pyros were still onboard), jettison the Ku-band antenna, and vent the cabin if preparing for a bailout. Each is probably redundant as well. So you're looking at somewhere in the neighborhood of 15-20 total pyrotechnic devices.

Based on current data available, it is now believed that the gear down indication at 7:59:05am was an anomaly probably caused by the damage to the sensor lines due to the plasma penetrating the wing thru the still-undetermined point of breach in the TPS. However, since that same damage may have caused the other sensors to fail so as to provide false readings that the gear was deployed, this possible deployment, while highly unlikely, cannot be 100% ruled out at this time. If anything, it may be far more likely that a deployment would be a symptom and not a cause of the increased drag and subsequent apparent failure of the left wing infrastructure. Ergo, a landing gear deployment in this fashion would be essentially the same thing as the trunk popping open on a car after the car has hit a tree,  and your tool kit and spare tire winding up scattered all over the road.
 

Sounds good, right? However, as warned before, much of what you read under this topic will be speculation based on available facts, and is subject to change, correction, and even nullification. In this case, the theory was ironically put to rest by NASA about five minutes after it was written up and inserted in the FAQ.

As of 6:35pm on 2/13/03, NASA officially discounted the possibility of a gear deployment. Based on current data available, it is now believed that the gear down indication at 7:59:05am on 2/1/03 was an anomaly, probably caused by the damage to the sensor lines due to the plasma penetrating the wing thru the still-undetermined point of breach in the TPS. All other indicators and drag information do not support the the possibility of unexpected gear deployment at this time.

• On a related note with regards to the landing gear theory, on 2/18/03, NASA announced that one of the main landing gear struts was confirmed as having been among the first delivery of Columbia debris to KSC. However, at this time it is still not known whether this was the left or right main strut.
   
• On 2/24/03, MSNBC reported that new analysis of the garbled last 32 seconds of telemetry from Columbia appears to show that the Shuttle lost its struggle to keep its nose pointed ahead, and began a flat spin to the left. The airstream, according to experts, was probably not powerful enough to tear the vehicle apart immediately, and indicates that Columbia could have turned through at least one full tumble in about 20 seconds while the cabin remained intact and pressurized. Following the breakup of the vehicle, the cabin fell for tens of seconds before it was crushed by the heat and deceleration. In this scenario, odds are high that the crew survived up to a minute after the spaceship began tumbling out of control and breaking up.

  The most persuasive evidence for this scenario comes from the reconstruction of the last 32 seconds of corrupted data received from Columbia. Although the signals were unreadable by Mission Control in real time because of massive “data dropouts”, the data was recorded at a ground station in White Sands, NM, and analysts have have had some success in retrieving useful data from the corruption. According to official NASA sources, the reconstructed data shows that for several seconds after loss of voice contact, the flight continued much as it had up to that point, except that the RCS thrusters in the tail section had been activated by the orbiter's autopilot in an apparent attempt to counter growing drag on the left wing. After a few seconds of garbled data, all communications with Columbia, both voice and data, went dead for about 20 seconds. This was followed by a brief burst of corrupted data three or four seconds in duration, followed by the total and final loss of all contact.

  To many experts, including knowledgeable space historians, enthusiasts, and even aerospace engineers working and/or retired from NASA and its contracting agencies, have interpreted this sequence as possible evidence that Columbia had been forced into a flat spin to the left, which would have turned the orbiter's antenna away from the angle required to uplink to the Tracking Data & Relay Satellite (TDRS). As a result, the final burst of data could have resulted from the uplink being briefly reestablished when the Shuttle completed a full rotation and the antenna was momentarily turned to face the satellite. Although there are numerous antennas to allow for the Shuttle to maintain constant contact with the TDRS, it is believed that the computer systems designed to switch between these antennas simply could not cope with the fast spin rate.

  The reconstructed data has also revealed that in those final few seconds, Columbia was a spaceship that was mortally wounded, but still working. Power was still being generated by the fuel cells under the payload bay, and signals were being received all the way from the back end of the craft. In addition, although the crew cabin was still pressurized, and the four main computers were still functioning, several other main systems were failing. The three redundant hydraulic pressure generators that help control the Shuttle’s aerodynamic surfaces were still functioning, but hydraulic pressure on the left side of the Shuttle was zero in all three main lines. At the same time, the tail section RCS thrusters were reporting a massive leakage of propellant.

  In addition to the this flat-spin-by-increased-drag theory, other experts have proposed that instead, the left wing was either totally torn off, or was bent up against the side of the orbiter. In any one of these three wing failure scenarios, the failure would have caused the vehicle to roll to the left until Columbia would have turned its tail end into the wind. Parts of the tail would have torn off first, accompanied by one or more of the bell- shaped rocket nozzles of the SSME's, the OMS pods, and, according to some sources, the drag chute package. Damage in this section of the orbiter would explain the propellant leak alarms, which most likely would have been followed by a detonation resulting from the uncontrolled mixture of the hypergolic fuels. Interpretation of the videotapes of the disintegration over DFW does show flashes of light that could indicate the detonation of propellant in the leaking tanks.

  Note that this particular reconstruction of the loss of Columbia has two caveats:
   

  1. It contrasts sharply with most preliminary assessments that maintain the Shuttle disintegrated suddenly and totally.
      
  2. This scenario has *not* been confirmed by either senior officials at NASA, or by members of the Columbia Accident Investigation Board (CAIB). It is, however, reportedly shared by a growing number of space experts, both inside and outside of NASA, some of whom have discussed their views on condition of anonymity.

• On 2/28/03, following a new analysis of data from Columbia's liftoff and a review of earlier studies, NASA investigators officially narrowed the likely area where debris struck shuttle Columbia during launch to RCC panels 6 thru 9 on the leading edge of the Shuttle's left wing. These U-shaped RCC panels - made of a Reinforced Carbon-Carbon (RCC) material designed to withstand temperatures greater than 2,300 degrees - lie in and outside a slight curve in the leading edge where the wing arches away from the fuselage. The panels are bolted to corrugated aluminum supports to form the leading edge. Panels numbering 1 thru 5 are located between the orbiter's fuselage and the curve in the wing. Panels numbering 6 through 22 are located from the curve to the wing's trailing edge.

  In the course of Columbia debris recovery, search teams have found 11 spar fittings used to attach panels 1 thru 5 to the left wing. Spars are steel brackets that help mount the panels to the spar. Four fittings are used for each panel. According to NASA sources, if the foam insulation struck between panels 6 through 9 and caused a breach there, engineers would expect to find the greatest heat damage to panel 5's spar fittings, the ones closest to the impact area.

  However, the actual state of the fittings hints differently. Some fittings located closer to the fuselage show significantly more damage from exposure to intense heat. Per a summary from the Space Shuttle Vehicle Engineering Office to the Columbia Mishap Response Team, "Initial visual interpretation indicates that the fittings from panel 2 were significantly more heat affected than the fittings from the other panels [1-5]." Engineers are studying what could account for the finding. It's still unclear whether the damage to the spar fittings could have been caused by a breach during re-entry or after the wing's breakup.
   

• In addition to the spar fitting report, NASA officials also acknowledged that debris teams retrieved a section of Columbia's left wing from an area adjacent to the left landing-gear door. What astonished investigators originally was the fact that the section's protective heat tiles appear melted in some areas, indicating exposure to extreme heat, yet the undersides of the tiles - the side normally exposed to reentry heating - show almost no damage. Although this has not been officially determined to be the cause of Columbia's demise, according to investigators it is believed to be evidence that a breach of some sort occurred on the wing, allowing superheated plasma to compromise the integrity of the wing from within, resulting in a catastrophic failure of the wing's infrastructure.
   
• On 3/4/03, the CAIB acknowledged during a press briefing that one major focus of the investigation involves examining burn patterns on debris from Columbia’s left side in an effort to “follow the heat.” These burn patterns include a spray of dark discoloration on tiles, believed to be congealed molten aluminum, as well as signs of extreme damage to the left-wing tires.  However, the board stressed that despite earlier reports by NASA officials regarding the narrowing of the focus to five specific RCC panels on the Shuttle's left wing, the investigation has yet identified the actual location of the breach.
  
  “Until we have located that breach, every postulated cause of the accident is just a theory,” said CAIB member Roger Tetrault. “What we have to do is follow the heat,” he said.
  
  Tetrault also cited several pieces of the puzzle that appeared to point toward a catastrophic failure on the shuttle’s left side:
   

  Tires: Two photos of recovered tires were shown during the briefing, one of a right-side tire, two from the left side.
  When the two sets are compared side-by-side, according to Tetrault, the two left-side tires look as if “they’ve gone through extreme trauma. The threads are basically pulled apart, and then [there was] heat damage at some later point." At the same time, however, Tetrault noted that telemetry received from Columbia just prior to the final loss of signal indicated that the wheels were still intact. “it is possible," he stated, "that the tires on the left side blew very late in the event.”
  	Left Tire	Right Tire

   

  • Molten aluminum: With photos of recovered tiles for examples, Tetrault explained that black deposits found on tiles located near the left wheel well were believed to have originated from a spray of molten aluminum from the Shuttle’s infrastructure. He also acknowledged earlier reports regarding the steel-aluminum slag on the inside of the RCC panels from the shuttle wing’s leading edge, but stressed that the origin of the slag was still unknown. “I don’t know exactly whether that is coming from the event or whether it’s coming from re-entry heating,” Tetrault said. “We’ve got some work to do in trying to refine that.
     
  • Left-wing hydraulics: In the final piece of the puzzle noted by the CAIB so far, Tetrault noted that there was a hole in a hydraulic actuator that controlled the left wing’s elevon, with no evidence of significant overheating. Again, investigators still had to determine whether the damage was done before or after the shuttle’s breakup.
     
• On 03/07/03, CBS News correspondent Bill Harwood reported that NASA engineers who are working on correlating telemetry from Columbia's left wing and the plasma flow patterns found in wing debris are becoming more and more convinced that the eroding flow of plasma most likely entered the wing from a breach at or near the leading edge area, close to the ship's fuselage, then worked its way into the left main landing gear wheel well. These sources, as well as other sources within NASA, say wreckage from the underbelly of the Shuttle just adjacent to the left landing gear door indicates superheated gasses from inside the wheel well may have spewed out around the inboard edge of the door as the catastrophe unfolded. This is the reverse of what was previously theorized shortly after the loss of Columbia, in which it was suspected that a breach may have occurred in the wheel well hatch seals and allowed the plasma flow to enter the wheel well.
  
  The scenario in question goes something like this:
   
  • At some point in the mission - most likely during reentry, although an in-orbit failure has still not been ruled out - the forward edge of the left wing just behind the RCC panels was breached. This allowed superheated plasma generated during reentry to burn its way into the unpressurized infrastructure of the left wing, in front of the left main landing gear wheel well.
     
  • As Columbia proceeded through a right bank maneuver designed to bleed off energy, the inrush of plasma flowed directly onto the lower part of the forward side of the wheel well near where it merges with the aluminum skin of the lower wing. This heats the aluminum to the melting point and produces some of the molten aluminum deposits found on some of the recovered debris.
     
  • Then, as Columbia banked to the left for another energy-reduction maneuver, the plasma flow direction most likely shifted in a direction that burned through electrical and sensor data cables routed around the upper front face of the wheel well. It is at this point when molten aluminum and stainless steel gets sprayed along the insides of the remaining RCC panels.
     
  • As Columbia returned to a straight course, the plasma burned thru the wheel well itself, probably damaging and possibly causing one or both tires to expand, weaken and later explode, then bouncing off to another part of the well, and eventually burning through the seals around the main landing gear door.
     
  • The resulting breaches would have easily compromised the lifting body dynamics of the Shuttle, and the smooth flow of air over the wing would have been disrupted, quite possibly generating a hot vortex of plasma that would have most likely spun back along the left side of the shuttle's fuselage, above the left wing. Upper surface tiles and blankets may have been ripped away, showing up as debris shedding events to observers below. Such a vortex could easily explain the elevated temperatures detected by sensors well above the wing on the side of the fuselage.

  One source within NASA confirms that some debris from the left wheel well shows signs that the flow of plasma came from within the wheel well and not from without. "My reading when I look at the tile around the door, that one piece we've been talking about, it looks to me that the plume is coming out," the source said. "What's happening is it's burned through the seals and it's coming out, not going in." Other NASA sources have said that a similar analysis was presented at a JSC Mishap Response Team meeting on 3/07/03, although another board source stressed other explanations are possible and reemphasized that some investigators have not yet ruled out a breach in the door seal area. Another source close to the CAIB outlined a similar scenario as well, but stressed that other explanations are possible and that board members have not yet drawn any conclusions about where the breach occurred or what might have caused it in the first place. The source also noted that several lines of evidence point to the possibility of a breach toward the front of the wing that burned its way deeper inside toward the wheel well as Columbia crossed above California and the southwest.
  
  Complicating the investigation is the fact that engineers have not yet conclusively determined whether the damage resulting from atmospheric heating that has been found in the recovered debris occurred before, during or even after Columbia's disintegration over the DFW Metroplex. "I don't think anything right now is THE leading scenario because we're looking at a group of scenarios that, based on the evidence, are probably more likely than other scenarios," the unnamed source said, adding that initial scenarios in aircraft disasters frequently get discarded as an investigation proceeds. Some have theorized that the molten aluminum and steel deposits found in unexpected places amongst the debris may have resulted from reentry heating that occurred after Columbia broke apart, and may still have no real bearing on what caused the mishap in the first place.
  
  Investigators are trying to determine the location of the breach and the behavior of the presumed internal plume of plasma by making various assumptions and then determining which scenarios best fit the pattern of sensor failures and elevated temperatures that were recorded during the shuttle's final minutes. Such scenarios also predict certain types of heat-related damage that can then be checked against actual wreckage as it is recovered. That work is far from complete. But the source familiar with the board's investigation outlined one such scenario to provide a sense of how engineers are working through the problem.
   

• On 3/13/03, a contrast-enhanced version of a view taken of the underside of Columbia during ascent was released by the CAIB. The unenhanced version of the video indicated the possibility of foam debris from the ship's External Tank may have struck at least two different areas on the orbiter's left side, one well in front of the left wing. However, when playing the clip over and over again in a loop, a subtle brightening can be seen as the debris fell past the left side of Columbia's fuselage near the extreme forward part of the wing known as the chine area.
  
  As described by a source inside the investigation who spoke with reporters on the condition of anonymity, the debris "tumbles and then I want to point out a subtlety to you. Watch the contrast against this black line [in the chine area] and you'll see what appears to be an apparent contact way up here. We're not sure if it actually contacts it or not. It may just be the light foam against the dark background. ... But you can see it against the dark."
  
  The source did stress a caveat regarding this footage. Regardless of whether the chine area suffered an actual debris impact, the brightening may simply be the result of lighting effects as the debris tumbled toward the wing. In addition, the source also stressed that no conclusions have been drawn based on the footage, and that scenarios based on as such may be incorrect. "Think of what kind of disaster we'd have if we jump to a conclusion that fits all the facts and come up with the answer and it's the wrong answer," the source said. "So we have to be prepared to really look at multiple answers and not jump to a conclusion."
   
• With regards to the plasma plume that is suspected to have entered the infrastructure of the wing thru the still-unidentified breach, the anonymous source explained what he claims is the current opinion of the investigative team. "You've got a gas there that is at 10,000 degrees. Because it's so rarified, the transfer of the heat from that creates a temperature in (the wing) of about 2,500, 2,800 degrees. The transfer mechanism in here is what they're struggling with. It's not strictly convection, it's not strictly radiation. You're dealing with a hundred thousand times lower pressure than you have on the surface."
  
  The plasma plume may well have moved about as Columbia went through standard entry maneuvers and caused the breach to worsen. The source observed, "If you think of a fire hose that you're not holding and it's moving all around, we may find that's more what we're dealing with. As it hits a hole that has different surfaces, the surfaces are going to be directing it in different directions. I think when (Columbia is) in the right bank, it may very well be forcing some of this air ... laterally down the spar and then when you go into the left bank, it actually moves up. It goes from one side to the other and you might find this thing changing over periods of time depending on the hole size as it eats away, what's in its way and those kind of things. So it may be much more complicated than just trying to point a finger and say it came in this hole and it went right through the wing here."
   
• The same anonymous source within the CAIB also noted on 3/13/03 that despite the destructive nature of plasma entering a breach in the wing, a single source for the loss of Columbia may not be the case. "One RCC doesn't do it and a couple of tiles won't do it," the source observed. "So we have to look for things that may increase in size over periods of time. One of the things we're looking at is if one panel comes off, it's likely others will come off because you've got hot air ... melting pieces and you would have a cascading kind of event. People have talked about unzippering of tiles. We're talking about unzippering of RCCs."
  
  Alternatively, a small but growing breach in the leading edge area could cause surrounding RCC panels to partially melt and collapse inward. So far, recovered RCC panel fragments and support hardware don't provide definitive clues as to which scenario may be more accurate. Based on the evidence gathered, as of 3/13/03 it appears likely that a plume of hot gas entered the wing just forward of the landing gear wheel well, burning through the forward skin of the wing and spraying aluminum slag back on the inner surfaces of nearby RCC panels, and ultimately working its way into the main landing gear wheel well.
  
  Damage found in recovered debris shows some indications the left landing gear door remained in place until the wing itself began breaking up late in Columbia's re-entry and that a plume of super-heated air entering the wing from the breach near the leading edge worked its way into the wheel well and then out through seals around the landing gear door. In fact, burn damage on recovered debris just inboard of the door appears to indicate a plume of hot gas spewed out of the well's forward inboard corner in the direction of the shuttle's centerline. Recently recovered debris indicates the possibility of a similar jet from the outboard forward corner of the door.
  
  "There is a similar one, but much smaller in nature, on the other side," the source noted. "We have a piece on the outboard side that has a very small indication of the same effect. You can't have that effect if the door is gone. You have to have something that's holding the pressure in and allowing the vent hole to occur. Again, we're coming back to the theory that if the door is lost, the door is lost very, very late in the event.
   

  "We also have another debris item," the source added, "a door hook roller uplock. It's not the hook itself, it's the pin, uplock pin, the door latches to. And it appears to be eaten away on one corner, which may be a key to us for the direction of flow. "We believe you're getting heat coming into the well, heat and air coming into the well, building up temperature on the inside and pressure on the inside. What's happening is then that temperature and pressure is allowing the aluminum corners to get a little bit soft and the pressure pushes up the corners a little bit and that's where the events are coming from. That also indicates, again, that the door is still on." Adding credence to this theory, numerous tiles have been recovered that originated from the door. Had a major breach occurred in the door area fewer tiles would have been recovered, and based on what was observed in the ascent video footage there are no obvious signs of tile damage before or after the suspected foam impacts. According to several experts on sci.space.shuttle and sci.space.history, even with the low resolution of the footage one would expect to see some difference in color on the underside of Columbia if the black surface layer of many tiles were in fact removed by such an impact. Even with the enhanced video, no obvious permanent change in the color or contrast of the underside can be discerned. "No item would you rest a case on individually," the source added. "But when you put all these together, it appears to give you a story that we're really dealing with something which is farther up from the wheel well and all these things about the wheel well and tile fractures is probably not the story."

  
  

• With regards to previously announced impact testing of Shuttle components as part of the investigative efforts, investigators at the Southwest Research Institute verified on 3/13/03 that plans to begin a series of tests that could shed light on just what happened when the foam impact occurred would take place over "the next few weeks." Pieces of foam insulation roughly the same size as the chunks seen falling away from Columbia's external tank will be fired into various wing components with a powerful "chicken gun" normally used to shoot debris into jet engines and through aircraft cockpit glass.
  
  According to SRI investigators, one RCC panel and the associated support hardware will be taken from Discovery to serve as a target, as well as two RCC panels and a landing gear door from Enterprise. Foam will be fired at these parts as well as surplus TPS tiles to determine the effects of impacts with broad areas of tile on the underside of the wing.
  
  Although similar to tests conducted by the SRI in 1999, this series of tests will use large pieces of foam and specifically include impacts on tile edges and other areas. According to SRI investigators, the tests will not be exhaustive, but "they will, I think, provide a piece of information that's going to be critical to determining what kind of damage you can do, what kind of damage you can't do. This isn't going to just be a piece of carbon carbon on plywood," one SRI investigator said. "They're actually taking the structure and instrumenting it with accelerometers and strain gauges to figure out how much of a lateral push it gets, how much compression it gets and then do the before and after." According to the other SRI investigator, "I don't see where foam is going to hit an RCC and break the (RCC). I can see more likely where it might break some of the support structure underneath it, a T-seal potentially, or shear a bolt or something. But inherently, this is a very weak area."
   
• on 4/11/03 NASA investigators announced that the second of two suspect carrier plates had been identified and found to be relatively intact. As a result, accident investigators are now reassessing their preliminary finding that the loss of a carrier plate close to the fuselage while Columbia was in its second day in orbit was the source of the breach in the leading edge of the Shuttle's left wing.
  
  Prior to this announcement, experts believed the foam struck the leading edge of the wing under panels five and six, possibly causing a breach that allowed superheated gas to invade the wing. Based on their analysis of NASA launch videos, as well as radar imagery that detected an object floating away from Columbia shortly after completing a small positioning adjustment. The analysis, conducted at Wright Patterson Air Force Base near Dayton, OH, suggested the object was most likely carrier panel #5 or #6, based on the radar cross-signature.
  
  However, both of these panels have been recovered and identified by search teams. As a result, the radar analysis has been extended to determine whether the drifting object might have been something more than just a carrier plate. Current theories for the object have now been expanded to include the following possibilities:
   
  • A carrier plate with three or four silica tiles attached.
     
  • A half piece of a carbon-composite panel.
     
  • A U-shaped composite seal that normally separates each carbon panel on the wing.

  As carrier plates are still missing from panels 12, 13, 14, 19, 20 and 21 on the left wing, NASA investigators have not ruled out the possibility that the "mystery object" might still be a carrier plate, it would have to be one much closer to the tip of the left wing than to the fuselage.
   

•  

Since this FAQ first presented its "best guess" as to what caused the loss of Columbia and her crew, NASA has focused its own "best guess" on what has become known as "Scenario #2". The scenario, one of a list of 10 scenarios for the catastrophe, appears to be supported by the data retrieved from the OEX Recorder.:

During ascent, at least one section, and as many as three, of the External Tank Insulation Foam, broke loose from the tank from or near the bipod ramp - an aerodynamic ramp of shaped foam just in front of one of the two struts that hold the Shuttle's nose to the tank. The foam fell back at roughly 500 mph in the slipstream between the orbiter and tank, and impacted the leading edge of the left wing in a two-foot-wide area centered on RCC Panel #8 and/or the underside of the orbiter outboard of the hatch for the wheel well containing the left landing gear. This impact is likely to have damaged parts of at least one of the RCC Panels between #7 and #9 and/or dislodged one or more T-Seals between the RCC Panels and/or one or more components of the Thermal Protection System, possibly including an as-yet undetermined combination of the Tiles, one or two so-called "carrier panels" that form a flush surface between the RCC panels and/or a section of the leading edge RCC structures near the joint where the wing meets the fuselage, and, although extremely unlikely considering the location of the impact footprint, possibly the seals on the wheel well hatch. Any one or combination of damages to the TPS could have compromised the integrity of the TPS to a critical and fatal level.

• During orbital operations, several military radar system apparently detected an object separating from Columbia on 1/17/03, about 24 hours after launch from KSC. NASA sources have confirmed the validity of the radar data, which apparently shows a small object suddenly separating from the shuttle at about five meters per second, or roughly 11.2 mph. On 4/8/03, the US DOD via USSTRATCOM issued the official 2-line orbital elements of this mystery object. Following a suggestion by MSNBC Space Expert James Oberg to assign official orbital elements to the object, a single elset was issued:

1 27713U 03003B 03018.89361124 .04158089 28060-1 63330-2 0 15
2 27713 39.0177 214.2209 0013015 7.8234 352.0943 16.06862990 17

The object reportedly decayed on 1/20/03 over the South Pacific.

The elset is available to registered users of the NASA/GSFC Orbital Information Group Web Site:

http://oig1.gsfc.nasa.gov/scripts/foxweb.exe/app01

At the time of this writing, while NASA officials have not determined the source and/or nature of this object, the consensus is that it originated from Columbia, and was not a foreign object which collided with Columbia. It has also been determined that  based on initial comparisons between the mission logs of STS-107 and the time the object was first observed, the object does not appear to be associated with a waste water dump, a process which on some previous Shuttle flights has produced large chunks of flash-frozen ice of the size that registered on the tracking radars. Based on analysis of radar data by NASA and CAIB investigators, the most likely suspects for the "mystery object" are either a chunk of RCC panel #8 and/or an adjacent T-seal.

• During re-entry, following a series of aerobraking maneuvers, Columbia experienced an abnormal increase in atmospheric drag along the left wing. This drag increase appears to occur shortly after amateur observers in California capture both still and video images of one or more fragments detaching and falling behind Columbia's re-entry trail. In an attempt to compensate for the drag, the vehicle's autopilot activated the RCS thrusters to correct Columbia's course. At some point during this course correction process, an as yet undetermined section of the Shuttle's left wing - either part of the RCC on the leading edge and/or a section of the Thermal Protection System on the underside of the left wing near the landing gear wheel well or possibly the seals of the landing gear hatches - sheared loose from the wing, due most likely to stresses placed on the Orbiter from aerodynamic stress. This caused one or more breaches in the surface of the left wing, allowing the 3000° F (1650° C) plasma created by the shock wave of friction-compressed air that had built up a few feet in front of Columbia to penetrate into the infrastructure of the left wing in front of the left wheel well.
   
• Once breached, the heated plasma then proceeded to cut thru the wing like a blowtorch until the infrastructure became too unstable to withstand the increased aerodynamic drag resulting from the increased physical deterioration of the surface of the wing, both from within and without the wing. The plume eventually burned through the outboard wall of the wheel well and/or its forward face, dramatically raising temperatures in the well. Based on analysis of a titanium pin used to mount the left landing gear door to the outboard side of the wheel well, the amount of melting the pin experienced indicates it was subjected to temperatures as high as 3,400 F. Other debris from an area just inboard of the forward corner of the door shows signs of hot gas outflow. Engineers believe the plume of hot gas entering the wheel well likely exited around the corners of the heat-softened landing gear door. The breach continued to worsen as the plasma plume burned through wiring from sensors located toward the rear of the wing. Lost data from those sensors were among the first indications in mission control that anything was amiss.
   
• As the left wing continued to deteriorate, Columbia's flight control system struggled to keep the Shuttle on proper course. By adjusting the ship's roll trim, the onboard computers attempted to compensate for the increasing aerodynamic drag and eventually had to resort to firing four right-side RCS yaw thrusters. With 16 minutes left in the re-entry phase, the left wing failed either by shearing loose from the orbiter, or bending it back and/or inwards towards the fuselage. This sent Columbia into a tumble beginning with a flat spin to the left that exposed the rear section of the orbiter to aerodynamic forces that may have ripped portions off the tail section, causing a catastrophic integrity compromise of the fuel lines and possibly the tanks associated with the propellant systems onboard Columbia. The orbiter apparently survived long enough to make at least one complete revolution in the direction of the flat spin to briefly reestablish a data uplink with the Tracking Data and Relay Satellite, after which all contact was completely lost. Very shortly afterwards, no more than a minute later at best, Columbia disintegrated at Mach 18 over the Dallas-Fort Worth Metroplex at an altitude of ~200,000ft (~60,000 m), spreading across a debris path from DFW to as far east as Fort Polk, LA.

Once again, this scenario is speculation based on the meager available data available, and should not be assumed to be an official scientific analysis in any way, shape or form. It is merely an attempt to speculate what might have happened based on what data is available.