© St. Petersburg Times, published February 4, 2003
NEW ORLEANS -- The foam insulation that peeled off as the Columbia lifted off Jan. 16, striking the shuttle's left wing, has been a headache for NASA officials for years.
The foam, sprayed onto the external fuel tank that carries the shuttle into space, has a history of coming unglued and damaging the sensitive tiles that protect the craft from burning up as it re-enters Earth's atmosphere.
NASA tried to fix the problem, but never believed it posed a risk to astronauts. Instead, it was seen as a money problem: Fixing damaged tiles is expensive.
But on Monday NASA officials called the flying foam the leading suspect in Columbia's demise.
"We're making the assumption that the external tank was the root cause of the accident," said Ron Dittemore, the shuttle's program manager. "It is a drastic assumption and it's sobering, but I think that's what we need to do."
It's not unusual for shuttle tiles to be lost or damaged in flight. Every time a shuttle goes into orbit, at least 40 of the thousands of tiles that cover the wings and belly of the craft are damaged or lost.
A 1995 paper by NASA scientists estimated that 90 percent of all tile damage on the shuttle belly resulted from the foam "debonding" during liftoff and smacking into the craft.
The solution: a laser to inspect the insulation for weak spots.
But in nine or 10 shuttle missions in 1997 and 1998, the foam popped off "like snow flurries," damaging about 100 of the tiles on one shuttle, said NASA engineer Greg Katnik, who inspects shuttles after they land.
"It never caused damage that we considered a flight safety issue, but it was something we were anxious to cure," Katnik said Monday.
Fixing the problem took time because the fuel tanks, constructed by Lockheed Martin Space Systems at NASA's Michoud Assembly Facility in New Orleans, are built years in advance of each mission and shipped by barge to Cape Canaveral.
"All these tanks were already fabricated and sitting on the line ready to go," Katnik said.
So technicians tried stopgap measures -- shaving the insulation down to lessen what could peel off and poking tiny holes in the foam to let air pockets escape. But the foam kept peeling off.
Then, Katnik said, scientists tried tweaking the formula. But as recently as October, a piece of insulation peeled off Atlantis and hit the solid rocket boosters. The foam apparently did no damage.
NASA has tried sticking pieces of the insulation in a Tennessee wind tunnel and onto the fuselage of an F-15 fighter jet, trying to mimic the conditions of shuttle liftoff. In 1999 they even hired researchers in Texas to fire chunks of the foam at shuttle tiles using a compressed-gas gun.
But their pursuit of the answer was always focused on cutting the expense and time lost in repairing the damage after each flight, not on whether the foam could cause a crash of the shuttle's orbiter craft.
"We've seen it hit the tiles before and pulverize, but we've always seen the orbiter come back," Katnik said.
The foam is as light as Styrofoam, but Dr. Hans Mark, a former deputy administrator of NASA, said speed of the shuttle is more important than mass.
"If it's moving fast enough, it can cause trouble, no question about it," said Mark, who left NASA in 1984 and is now a professor of aerospace engineering at the University of Texas in Austin. "Any material that's moving fast enough can cause damage."
The insulation's purpose: protect the tiles, not damage them.
The foam is sprayed on the outside of the 154-foot external fuel tank. The liquid hydrogen fuel inside the tank is kept at 423 degrees below zero. Without insulation, ice would form on its aluminum skin, break off during launch and severely damage tiles.
The foam insulation also protects the volatile fuel inside the tank from the enormous heat generated by the shuttle's rockets and the friction of the air during a launch that reaches four times the speed of sound.
The foam insulation is one of the last of a half-million parts and compounds that goes into each external fuel tank. The work is done in a World War II-era defense plant owned by NASA but operated by Lockheed Martin, which has built the external fuel tanks here since the shuttle program's infancy in 1973.
The sprawling Michoud plant covers a milelong stretch of two-lane road in a flat industrial strip in east New Orleans, near a thriving Vietnamese community and a Folger's coffee plant.
The tank is covered by 10 types of insulation made by six different companies.
The majority of the tank is covered with a polyurethane foam made by North Carolina Foam Industries in Mount Airy, N.C., that also makes insulation for homes and taxidermy molds for mounting deer and other game.
However, Lockheed officials said the foam that peeled away from Columbia was made by another manufacturer, whom they would not name.
The foam usually arrives in semiliquid form in two separate parts. The two compounds are mixed in the spray gun as they are applied. A metal primer is applied first to help it adhere. As it hardens, the foam cures to a mustard brown after about 24 hours.
NASA has touted the foam as a great example of the space program's spinoff technology. Besides roofing insulation, the foam has replaced the plaster used to produce molds for artificial limbs.
But the foam has been a problem from the start. During a 1981 test-loading of fuel in Columbia's external tank, the super-cold fuel caused shrinkage in the adhesive. During subsequent ground tests, technicians wrapped a cargo net around the tank to keep foam from shaking off.
In 1991, technicians scrambled to replace a four-inch-square of foam that came unglued before launch.
A series of missions in the mid 1990s saw small pieces pop off because heat from the launch created air pockets under the insulation, Katnik said.
"It was like watching popcorn," he said. "It would pop off a little piece here and a moment later there'd be another piece over there."
The problem appeared to peak with a 1997 launch of Columbia. Katnik's crew discovered about 100 damaged tiles.
The angle and location of the damage suggested the foam was at fault, but there were no pictures from liftoff of any foam chunks coming loose.
Finally, though, they found pictures that confirmed the foam had come unglued. They were photos taken by Columbia's crew.
-Times staff writer Chuck Murphy and researchers Kitty Bennett and Cathy Wos contributed to this story.
The shuttle has had numerous problems with foam insulation coming off the external tank and various fixes have been tried. Here are some examples:
FEBRUARY 1981: First test loading fuel in Columbia's external tank causes shrinkage in adhesive of insulation on tank. Cargo net wrapped around tank to prevent insulation from flying off.
JUNE 1991: Columbia launches after technicians replace 4-inch-square of insulation.
JUNE 1995: NASA estimates 90 percent of tile damage on the shuttle's underside caused by insulation debris.
DECEMBER 1997: Technicians find about 100 damaged tiles on Columbia, conclude cause was loss of insulation unseen by launch cameras.
MAY 1998: Despite new insulation precautions, abnormal amount of tile damage discovered.
JANUARY 1999: Six panels of shuttle insulation mounted to fighter plane to test stress of shuttle takeoff. No erosion of insulation noted.
MARCH 1999: Wind tunnel tests of insulation determines it flies loose because of speed and heat on ascending aircraft. Marshall Space Center "tweaks" formula.
MAY 1999: Discovery launch delayed after hailstorm gouges 150 holes in insulation.
OCTOBER 2002: Atlantis loses piece of insulation that strikes rear section of one solid rocket booster. Analysis indicates no harm, flight proceeds.
JANUARY 2003: As Columbia lifts off, chunk of insulation breaks loose from external rocket and strikes left wing.