Shuttle Tile Repair Kit
New Scientist magazine (London)
15th November 2003

Next to a giant vacuum chamber for testing spaceships is a semicircle of tables covered with hands-on demos. Samples of rubbery pink “goop” are being handed out as souvenirs. It feels like a school excursion to a science museum, but the scene is actually a special media workshop to showcase NASA’s efforts to get the shuttle back to flight.

The speakers include top engineers and project managers, and all strike a cautious tone. Because of technical hitches, NASA has already put back the scheduled return-to-flight date from March 2004 to “no earlier than next September 12th”. By then, or within a reasonable time period thereafter, NASA must solve the problems that lead to the space shuttle Columbia’s catastrophe on February 1st. If it fails, the three surviving shuttles, now grounded, may never fly again.

NASA’s main challenges concern the shuttle’s heat shield, or, in NASA jargon, its thermal protection system’, or TPS, the external structure intended to protect the spaceship and its crew from the searing heat of high-speed entry into the atmosphere. As the independent Columbia Accident Investigation Board found in August, NASA needs to implement two strategies before the grounded fleet can return to space. Firstly, the organisation needs to fix the flaws in management oversight that opened the step-by-step path to space disaster. And secondly, it needs to develop a way to detect and fix any future damage to the TPS while a shuttle is in flight. That way, if something goes wrong, and there again is a mortal wound to a shuttle’s heat shield, the damage will be detected in time for astronauts to make repairs that will be good enough for their safe return to Earth. At least, that’s the new plan.

Columbia was lost due to a small wound inflected on its left wing during launch on January 16. Insulation from the belly-mounted ‘external tank’ flaked off, as had happened on four fifths of the previous 110 shuttle missions. This time the piece, larger than most, hit a section of the wing more vulnerable than most, its leading edge. The material there is a carbon composite material called “reinforced carbon-carbon” or “RCC”, and it is much more brittle than the more famous “shuttle tiles” that cover the bottom and sides of the spaceship. A hole around 30 centimetres wide was punched into panel number 8.

Sixteen days later, as the vehicle slammed into the upper atmosphere, superheated plasma jetted through the gap and gradually melted away the aluminum structure. But if only Mission Control had known there was a hole, they could have thrown together a jury-rigged spacewalk patch plan. That way an astronaut might fill the gap, reduce the plasma flow, and extend the shuttle’s survival long enough for it to get low and slow enough that the crew could survive a bailout into the upper atmosphere.

For any future occurrence of damage to either tiles or the RCC panels, NASA now knows it has to build workable repair tools, and be ready and able to use them. It needs a reliable plan for inspecting the entire outer surface of the shuttle, and a stable way for spacewalking astronauts to reach damaged areas to execute the necessary repairs.

The gadgets are the easiest part, since clever space engineers relish any opportunity to live up to their popular folklore as “rocket scientists”. And NASA has a large body of test data on damage to tiles, collected almost a quarter century ago, the last time the organisation found it worthwhile to explore tile repair technologies. Before the first space shuttle mission in 1981, NASA made serious efforts to develop an in-flight tile repair capability.

Back then, NASA was worried about entire tiles falling off, since their ‘space glue’ wasn’t holding them in place. So they were worried there would be big gaps. One thing they did NOT worry about – since their design precluded it – was impact damage from things falling off the tank. In any case, well before the first flight NASA chemists had solved the adhesive problem and the urgent motivation for a repair kit vanished.

As part of the previous effort, engineers developed an applicator gun for injecting a fast-setting silicone-based repair substance into cavities in damaged tiles. This substance, a silicone-based paste called MA-25S, is the “goop” that was handed out to the media in September. The mix includes iron oxide, which makes it a salmon colour. Space-suited astronauts tested the paste on zero-G aeroplane flights, but the tests revealed significant problems with the spray-on materials foaming uncontrollably, creating bulging “repairs” that would have created damaging air turbulence during entry into the atmosphere. Also, NASA’s lack of experience with spacewalks led to fears that clumsy astronauts would damage the tiles more than the original problem they would be sent outside to fix. Testing also showed MA-25S was sticky in the wrong way — it clung to the applicator tools and pulled off the often-dusty cavities where tiles were missing. “The material stuck better to the tools than to the tiles,” spacewalk tools expert Dana Weigel told reporters.

NASA had tested tools to apply the ‘goop’, and found that in a spacesuit a clumsy repairman might do more harm than good to other tiles. They subjected the ‘goop’ to space conditions such as vacuum and fiery heat, and found that it foamed up excessively, creating a charred bulge that induced dangerous turbulence during entry. The MA-25S turned out to be sticky on the wrong side, clinging to the astronaut’s tools while easily pulling loose from the dusty tile cavities.

Nevertheless, astronauts’ future repair kit will include the same MA-25S that was picked in 1980. Most of what killed [the effort] in 1980 was the operational aspect,” says materials scientist Michael Fowler, “but we’re a lot smarter on EVA [extravehicular activity, or space walks] now.” Fowler is leading a team who have changed the way the foam is processed in order to reduce the air and water contamination that had caused the excess foaming in the 1980 tests. “We do the mixing in vacuum now,” he explains. As an astronaut works on the job, two compounds from separate tanks are continuously mixed and extruded through the applicator gun. There is no time for contamination, and the paste does not foam.

Stickiness was fixed serendipitously. A new type of widely-used DIY paintbrush and roller, made from polyurethane foam, was one of several candidates for testing, and engineers were astonished to find that for some reason the goop wouldn’t stick to this material. At the demonstration, reporters were invited to manipulate the goop with these tools and see for themselves. With considerable effort, some dubious hacks succeeded in getting some of the goop to adhere to the tools, but only in small amounts.

Once in a tile cavity, an arc jet at a temperature of 1260 degrees centigrade for 15 minutes approximates the heat load of entry. The cured paste chars but expands only slightly, and provides full thermal protection.

While it will certainly be comforting to have such a tile fixer, flight experience has shown that the thermal protection system is indeed already remarkably resistant to damage to tiles. On flight after flight they have been chipped, cracked, gouged, even knocked off entirely, and the surrounding “acreage” of tiles has held up well against the heating. This persistent survivability contributed to lulling NASA engineers into the belief that such impacts really weren’t that big a deal.

At the media workshop, project manager Hill admitted, “Holes the size you see here in tiles, we probably wouldn’t even need to repair.” He admitted that a certain degree of tile damage had been found to be acceptable. But what really threatened the shuttle’s survival was any induced surface irregularities that seriously altered the hypersonic air flow during atmospheric entry, causing skin temperatures to more than triple. The goop would be critical, Hill explained, to smooth out such damage and allow air to flow in clean layers.

This is because the superheated air surrounding an entering spacecraft is created by shock wave compression, not the commonly misidentified ‘friction’ at the vehicle’s surface. Therefore the hottest air is not in direct contact with the skin, unless the layered flow of air is disturbed by the roughness of surface damage. Hill described how he imagined a space repairman smoothing a damaged tile edge by laying “a bead of material” along it. They will be able to do it with the tool kit that his team is well on its way to developing. “My confidence is relatively high that we are very close to declaring victory regarding the tiles,” he says.

But all these efforts are only half the job, and the easy half at that. The far more fragile and less-repairable component of the shuttle’s thermal protection system is the RCC panels that line the wing leading edges and the shuttle nose, where the highest temperatures are experienced. The panels are made by alternately soaking a rayon cloth in hydrocarbon fluid and then oxidising and burning off the hydrogen. When this process is repeated three or four times, the result is a strong, layered graphite composite. On Columbia, it was in one of the two dozen RCC panels on the left wing that flying debris punched a hole. Techniques being developed to repair such material are “much less mature”, NASA admits. “The reinforced carbon composite is a challenge,” says Steve Paulos, NASA’s space shuttle engineering manager.

Panels were impact-tested in the early 1980s with hypervelocity micrometeoroid impacts in mind, and held up even with tiny holes punched in them. But astonishingly, NASA had never actually tested RCC panels against insulation impacts. The RCC panels feel metallic and can be hit with a hand, or even a hammer, without apparent damage. Compared to “soft” insulating foam, these were seen as extremely strong. Each RCC panel is a U-cross-section segment of material that is 80 to 120 centimetres long, and about 1 centimetre thick. Unlike the shuttle tiles, which are attached directly to the spaceship’s aluminum skin, the RCC panel is hollow behind the surface. It wasn’t until months after Columbia was lost that NASA, under orders from the independent Columbia Accident Investigation Board, actually conducted foam-on-RCC tests. Using a gas gun that propelled a hunk of foam at the speed observed during Columbia’s launch, the test punched a 30 centimetre wide hole in an RCC panel. This was the ‘smoking gun’ of the catastrophe’s cause, and glaring proof of NASA’s oversight.

Holes of that size can’t be fixed with spray-on goop. Something entirely different will be needed. And NASA has six of its centers and a dozen commercial contractors developing and testing a wide variety of ideas. It is also lining up government, commercial, and university testing facilities, including the Italian Aerospace Research Centre’s ‘plasma wind tunnel’ in Capua, Italy. This will be used to generate the extreme heat equivalent to atmospheric entry, to test the different proposed techniques.

There are three “key challenges” for an RCC repair capability. First, the repair has to remain attached to the RCC during the tremendous heating. Second, it has to be smooth – “As low as a tenth of an inch, even zero sometimes”, says Hill. Lastly, the repair should prevent cracks in the RCC from spreading during entry stresses extending beyond the boundary of the repair.

The most unsophisticated proposal is to insert a rubber bladder through the hole into the cavity behind the broken panel, and then inject foam into the bladder until it expands to fill the entire cavity. Sounds crude, but definitely an improvement over the emergency repair technique NASA later decided it would have tried on the Columbia mission if the damage had been detected. In that case, astronauts would have put a flexible container into the hole and then filled it with water, letting it freeze solid before trying to return to Earth.

The American rocket company Thiokol has offered a patch system that it already uses on the engine bells, the cones at the base of the Delta-4 rocket, which are also made of carbon-carbon. A rigid patch, between one and ten centimetres across can be attached to a molybdenum bolt with an umbrella-like locking device at the other end that can be inserted smoothly, point-first, into the hole, but cannot move back the other way, so the patch is held on. “It’s a different heat environment then we see in entry,” Hill admits, “but it is also a severe environment.” The temperature of air at the inside of the bells when the rocket is firing is similar to that at the shuttle wing on re-entry, but the air flows over the surface at a lower rate, so the bells heat up less.

Another technique already being tested is a flexible carbon-carbon sheet that can be cut to fit over an entire RCC panel. The so-called “overwrap” has an inner adhesive layer, and once exposed to vacuum, cures rigid within 24 hours. Perhaps combined with the bladder in the inside cavity, this appears to be a promising candidate.

But in terms of schedules for developing actual tools based on one of more of these repair materials, Hill was clear there was a long way to go: “No predictions yet,” he told reporters. The need to test, select, and develop a repair technique for damaged RCC panels was one of the three reasons NASA officials gave in early October for officially delaying the next shuttle mission.

And one additional feature of a repair procedure remains to be developed: moving the spacesuited repairman to the location underneath the shuttle where the damage has been detected. NASA considered everything from jet backpacks (not stable enough) to long deployable booms (which sway too much) to stick-on pads (potential damage to the rest of the tiles) and finally settled on one idea that even Hill admits sounded like a “goofy idea” when first proposed. Don’t move the astronaut, move the shuttle.

The technique exploits a new policy for future shuttle missions: all future shuttles will be docking at the International Space Station, apart from missions to service the Hubble Space Telescope, which will need to be in a different orbit. The space station provides both an observation platform from which to survey all sides of the approaching shuttle for damage, and a storehouse to keep the repair tools and supplies. It also can provide a work platform for any repairs that prove necessary.
Once a repair operation has been chosen, the procedure will work like this. The shuttle, docked to one end of the space station, grapples a hook on the station with its own robot arm. It then unfastens its docking mechanism while remaining attached by the arm. Over a long period – up to four hours, NASA’s Paul Hill estimates, the motorised shoulder, elbow, and wrist of the robot arm will turn the shuttle’s bottom side towards the station.

Two spacesuited astronauts then go out the shuttle airlock with their repair tools, and move to one edge of the payload bay. There they meet up with the ‘hand’ at the end of the robot arm installed on the space station. While one remains back to observe and pass equipment, the actual repairman anchors himself and a tool caddy to the end of the station arm, which then moves him away from the shuttle payload bay and down under the belly where the repair is needed. After the repair, the process is undone in reverse. “We expect this to be a long day,” Hill says, “but it is well within our capabilities”. Neither the shuttle’s nor the ISS robot arms were ever designed for such an application. But their designers did build them to be mechanically robust, with greater strength and flexibility than were strictly required. That intuitive decision to design for the unexpected makes the repair operation feasible.

And this goes to the heart of the need for a cultural revolution, the part of the repair strategy that will be far harder to implement. NASA is still relatively inexperienced in space, and the first rule of any adventure is to expect the unexpected and to arm yourself appropriately. Even if NASA succeeds in developing the precise materials and gadgets it would need to avoid another Columbia, such gadgets are most useful only if their introduction is accompanied by a change in the ‘gung-ho’ mindset that led officials at the time to not even consider the potential for damage. So far, that repair plan remains obscure.