The debris detective

On 1 February 2003 a bright light over Texas and a shower of sparks indicated one of manned spaceflight’s greatest disasters: the break-up of the Space Shuttle Columbia as it re-entered the atmosphere. While the US, India and Israel mourned its astronauts, NASA had to pick up the pieces - over 85,000 of them. It was these fragments that would tell the story of the orbiter’s final moments. And the lessons learned from this exhaustive investigation should hopefully ensure that history isn’t repeated. Indeed the camera attached to the robot arm used to inspect Discovery’s heat shield earlier this month was developed as a direct result of this process. The debris field from the Columbia disaster was the largest in history. Over 16,000 volunteers combed an area 1,000km long and 15km wide, gathering pieces ranging from a few metres to a few millimetres across. These were brought back to a hangar at the Kennedy Space Centre, where specialists attempted to reconstruct the layout of the shuttle. From this they would then try to determine the sequence of events that had culminated in that blazing trail. Steve McDaniels, chief of NASA’s Failure Analysis and Materials Evaluation Branch, explained that the ultimate cause of the disaster - insulating foam falling from the external fuel tank on lift-off and striking the left wing of the orbiter - was apparent quite early. ‘While they were in orbit the question arose of what had happened. They reviewed the video and saw the strike but couldn’t visualise where it was.’ It was therefore down to the reconstruction team to make sense of the debris. McDaniels, a University of Florida educated materials scientist, spent four years investigating helicopter and plane crashes for the US Navy before transferring to NASA to do more specialised failure analysis work. But his experience was scant preparation for the Columbia analysis. The orbiter broke up at an altitude of over 185,000m, travelling at almost 18 times the speed of sound, so the temperatures and stresses it experienced were beyond anything seen in any civil or military plane crash. Very little debris had previously been recovered from man-made objects that had broken up on re-entry. ‘We were told that once you’d worked on a piece for a couple of hours, you were the world expert on this type of re-entry damage,’ said McDaniels. ‘There were just a handful of pieces of hardware that had come through the atmosphere before and survived.’ The team had some information available. The first signs of malfunction came from sensor readings in the left landing gear bay. ‘We looked at the debris from the left-hand wheelbay, and could see evidence of plasma and slumping [melting] of the insulating tiles, but when you started to look at it, the path of the damage didn’t make sense. It seemed to be coming in the wrong direction.’ Taking a different tack, the team looked at the pattern of the recovered debris. A large amount of material from the right wing had been found, but there was a great deal missing from the left. Most of the aluminium panels were missing completely, burned away in re-entry. Even more interesting were three missing reinforced carbon-carbon (RCC) panels from the leading edge of the wing. ‘And what was there was showing peculiar damage,’ McDaniels commented. Another clue was where the debris was found. Because the Shuttle was travelling eastwards, debris found further west had been shed earlier and had therefore been affected by failing systems first. Despite the unusual nature of the damage to the debris, McDaniels’ team knew how to tackle their task: ‘We may not have known where the path would lead, but you look for patterns, distributions and trends, and then you have photo documentation, X-ray crystallography, scanning electron microscopy, microprobes, stereomicroscopy.’ One of the most important questions was how hot the orbiter was when it broke up. The key to this was, once again, which structures were missing, and the traces found on the surviving fragments. ‘We saw a lot of aluminium melting, so that indicated it had got to 1,000^o C. And we knew that Inconel [a nickel-based high-temperature alloy] had melted, and that goes at around 2,500^oC. Then when we saw Cerochrome [a ceramic insulator] traces, we knew it had gone above 3,200^o C. That told us that the damage had to be caused by plasma.’ The melted minerals tended to be deposited on surviving materials in distinctive patterns. Cerochrome and one type of Inconel, from a foil used to wrap the ceramic insulator, were found in globules towards the front of the wing. A different type of Inconel, used to hold parts of the wing structure together, was found further back, in spheroids. More insulation remains were found on top of the wing, indicating it had been ‘blown’ over the wing. And a remaining sooty covering was a mixture of most of the metallic elements in the wing structure. But what was missing was any trace of melted stainless steel, which is used for the wing spars of the Shuttle. ‘That confirmed that the breach had been a long way from the spars,’ McDaniels said. Exploring the layers of melted and resolidified debris proved to be similar to analysing geological core samples, McDaniels said. ‘We used scanning electron microscopy to give us dot-maps for X-ray analysis - that told us whether a particular artefact was a piece of Inconel, but not what sort of Inconel it was. Then we went to a microprobe, which can differentiate the ratios of elements like iron to nickel, and find trace amounts of alloying elements like niobium. We also used electron spectroscopy for chemical analysis, which allows you to sputter down through a surface and get to the subsurface. It really is just like dipping down through a core sample. The closer you get to the substrate, the earlier you are in the sequence of events.’ The analysis eventually determined the course of events. The initial damage was to the underside of the 8th RCC panel on the leading-edge of Columbia’s left wing. As the Shuttle began re-entry the atmospheric gases, pressed and heated by the decelerating orbiter, entered the breach as a plasma and started melting the Cerochrome insulation, which was wrapped in Inconel foil. Once this happened the wire bundles inside the wheelbay burned through, giving the misleading readings. The structural components of the wing then gave way, and finally integrity was lost, leading to loss of control and the collapse of the wing. ‘All the pieces had a story to tell,’ McDaniels said. ‘We just had to decide how they fitted into the whole picture.’ The results are now being used both to make the Shuttle safer for subsequent flights, and in the design of the Shuttle’s replacement orbiter. Vibroacoustic sensors and additional temperature sensors will give advance warning of abnormal occurrences, and improved camera systems, mounted on booms, allow inspection of previously inaccessible parts of the Shuttle. The fuelling system also now includes exterior heaters to limit the formation of ice as the liquid oxygen fuel is pumped into the external tank. Interestingly, the rarity of the Columbia material has led NASA to lend universities pieces of the shuttle to help train the next generation of spaceflight designers. ‘The first priority was to make the systems safer, but we also wanted to reach out to educational institutions to encourage students to pursue science and technology,’ McDaniels said. ‘Also, the university researchers have been very helpful in coming up with ideas for what kind of research we might want to pursue - they’ve asked questions we hadn’t thought of, and come up with new ways of finding the answers.’ For the future orbiter McDaniels said, ‘We’re looking at which materials and designs worked and what didn’t, whether the RCC panels need reinforcement, what changes we could make to the exterior and interior. We’re building on what went before, just as we built on Mercury, Gemini, Apollo, Soyuz and the Space Station.’