Space Shuttle Columbia disaster

Columbia prior to launch. The circled area on the external tank (ET) is the left bipod foam ramp, and the circled area on the orbiter is the location that was damaged.

The Space Shuttle was a partially reusable spacecraft operated by the U.S. National Aeronautics and Space Administration (NASA).[1]^: 5, 195 It flew in space for the first time in April 1981,[2]^: III-24 and was used to conduct in-orbit research,[2]^: III-188 and deploy commercial, military, and scientific payloads.[2]^{: III-66, 68, 148} At launch, it consisted of the orbiter, which contained the crew and payload, the external tank (ET), and the two solid rocket boosters (SRBs).[3]^: 363 The orbiter was a reusable, winged vehicle that launched vertically and landed as a glider.[2]^: II-1 Five operational orbiters were built during the Space Shuttle program.[1]^: 5 Columbia was the first space-rated orbiter constructed, following the atmospheric test vehicle Enterprise. The orbiter contained the crew compartment, where the crew predominantly lived and worked throughout a mission.[2]^: II-5 Three Space Shuttle main engines (SSMEs) were mounted at the aft end of the orbiter and provided thrust during launch.[3]^: II-170 Once in space, the crew maneuvered using the two smaller, aft-mounted Orbital Maneuvering System (OMS) engines.[3]^: II-79

The orbiter was protected from heat during reentry by the thermal protection system (TPS), a thermal soaking protective layer around the orbiter. In contrast with previous US spacecraft, which had used ablative heat shields, the reusability of the orbiter required a multi-use heat shield.[4]^: 72–73 During reentry, the TPS experienced temperatures up to 1,600 °C (3,000 °F), but had to keep the orbiter vehicle's aluminum skin temperature below 180 °C (350 °F). The TPS primarily consisted of four sub-systems. The nose cone and leading edges of the wings experienced temperatures above 1,300 °C (2,300 °F), and were protected by reinforced carbon-carbon material (RCC). Thicker RCC was developed and installed in 1998 to prevent damage from micrometeoroid and orbital debris.[2]^{: II-112–113} The entire underside of the orbiter vehicle, as well as the other hottest surfaces, were protected with black high-temperature reusable surface insulation. Areas on the upper parts of the orbiter vehicle were covered with white low-temperature reusable surface insulation, which provided protection at temperatures below 650 °C (1,200 °F). The payload bay doors and parts of the upper wing surfaces were covered with reusable felt surface insulation, as the temperature there remained below 370 °C (700 °F).[3]^: 395

Two solid rocket boosters (SRBs) were connected to the ET, and burned for the first two minutes of flight.[3]^: II-222 The SRBs separated from the ET once they had expended their fuel and fell into the Atlantic Ocean under a parachute.[3]^: II-289 NASA retrieval teams recovered the SRBs and returned them to the Kennedy Space Center, where they were disassembled and their components were reused on future flights.[3]^: II-292

When the Space Shuttle launched, the orbiter and SRBs were connected to the ET, which held the fuel for the SSMEs.[3]^: II-222 The ET consisted of a tank for liquid hydrogen (LH2), stored at −253 °C (−423 °F) and a smaller tank for liquid oxygen (LOX), stored at −183 °C (−297 °F). It was covered in insulating foam to keep the liquids cold and prevent ice forming on the tank's exterior. The orbiter connected to the ET via two umbilicals near its bottom and a bipod near its top section.[5]^: 50–51 After its fuel had been expended, the ET separated from the orbiter and reentered the atmosphere, where it would break apart during reentry and its pieces would land in the Indian or Pacific Ocean.[3]^: II-238

Close-up of the left bipod foam ramp that broke off and damaged the orbiter wing

During the design process of the Space Shuttle, a requirement of the ET was that it would not release any debris that could potentially damage the orbiter and its TPS. The integrity of the TPS components was necessary for the survival of the crew during reentry, and the tiles and panels were only built to withstand relatively minor impacts. On STS-1, the first flight of the Space Shuttle, the orbiter Columbia was damaged during its launch from a foam strike. Foam strikes occurred regularly during Space Shuttle launches; of the 79 missions with available imagery during launch, foam strikes occurred on 65 of them.[5]^{: 121–122}

The bipod connected the ET near the top to the front underside of the orbiter via two struts with a ramp at the tank end of each strut; the ramps were covered in foam to prevent ice from forming that could damage the orbiter. The foam on each bipod ramp was approximately 30 by 14 by 12 inches (76 by 36 by 30 cm), and was carved by hand from the original foam application.[6] Bipod ramp foam from the left strut had been observed falling off the ET on six flights prior to STS-107, and had created some of the largest foam strikes that the orbiter experienced. The first bipod ramp foam strike occurred during STS-7; the orbiter's TPS was repaired after the mission but no changes were made to address the cause of the bipod foam loss.[5]^: 123 After bipod foam loss on STS-32, NASA engineers, under the assumption that the foam loss was due to pressure buildup within the insulation, added vent holes to the foam to allow gas to escape. After a bipod foam strike damaged the TPS on STS-50, internal NASA investigations concluded it was an "accepted flight risk" and that it should not be treated as a flight safety issue. Bipod foam loss occurred on STS-52 and STS-62, but neither event was noticed until the investigation following Columbia's destruction.[5]^: 124

During STS-112, which flew in October 2002, a 4-by-5-by-12-inch (10 by 13 by 30 cm) chunk of bipod ramp foam broke away from the ET bipod ramp and hit the SRB-ET attachment ring near the bottom of the left SRB, creating a dent 4 inches (10 cm) wide and 3 inches (8 cm) deep.[5]^: 124 Following the mission, the Program Requirements Control Board declined to categorize the bipod ramp foam loss as an in-flight anomaly. The foam loss was briefed at the STS-113 Flight Readiness Brief, but the Program Requirements Control Board decided that the ET was safe to fly.[5]^: 125

A debris strike from the ablative material on the right SRB caused significant damage to Atlantis during the STS-27 launch on December 2, 1988. On the second day of the flight the crew inspected the damage using a camera on the remote manipulator system. The debris strike removed a tile; the exposed orbiter skin was a reinforced section, and a burn-through might have occurred had the damage been in a different location. After the mission, the NASA Program Requirements Control Board designated the issue as an in-flight anomaly that was corrected with the planned improvement for the SRB ablator.[5]^: 127

The crew of STS-107. From left to right: Brown, Husband, Clark, Chawla, Anderson, McCool, Ramon.

For STS-107, Columbia carried the SpaceHab Research Double Module, the Orbital Acceleration Research Experiment, and an Extended Duration Orbiter pallet.[5]^: 30 The mission passed its pre-launch certifications and reviews, and began with the launch. The mission was originally scheduled to launch on January 11, 2001, but it was delayed thirteen times, until its launch on January 16, 2003.[5]^: 28

The seven-member crew of STS-107 were selected in July 2000.[5]^: 28 The mission was commanded by Rick Husband, who was a colonel in the U.S. Air Force and a test pilot. He had previously flown on STS-96.[7] The mission's pilot was William McCool, a U.S. Navy commander who was on his first spaceflight.[8] The payload commander was Michael Anderson, a U.S. Air Force lieutenant colonel who had previously flown on STS-89.[9] Kalpana Chawla served as the flight engineer; she had previously flown on STS-87.[10] David Brown and Laurel Clark, both Navy captains, flew as the mission specialists on their first spaceflights.[11][12] Ilan Ramon, a colonel in the Israeli Air Force and the first Israeli astronaut, flew as a payload specialist.[13][5]^: 29

Footage of the debris strike at T+81.9 seconds

Columbia launched from the Kennedy Space Center Launch Complex 39A (LC-39A) at 10:39:00 am. At T+81.7 seconds, a piece of foam approximately 21 to 27 inches (53 to 69 cm) long and 12 to 18 inches (30 to 46 cm) wide broke off from the left bipod on the ET. At T+81.9 seconds, the foam struck the reinforced carbon-carbon (RCC) panels on Columbia's left wing at relative velocity of 625 to 840 feet per second (426 to 573 mph; 686 to 922 km/h).[5]^: 34 The foam's low ballistic coefficient caused it to lose speed immediately after separating from the ET, and the orbiter ran into the slower foam.[5]^: 60 The mission or ground crew did not notice the debris strike at the time.[5]^: 140 The SRBs separated from the ET at T+2 minutes and 7 seconds, followed by the ET's separation from the orbiter at T+8 minutes 30 seconds.[5]^: 35 The ET separation was photographed by Anderson and recorded by Brown, but they did not record the bipod with missing foam.[5]^: 148 At T+43 minutes, Columbia completed its orbital insertion as planned.[5]^: 35

STS-107 crew posing for a group photo in space

After Columbia entered orbit, the NASA Intercenter Photo Working Group conducted a routine review of videos of the launch. The group's analysts did not notice the debris strike until the second day of the mission. None of the cameras that recorded the launch had a clear view of the debris striking the wing, leaving the group unable to determine the level of damage sustained by the orbiter. The group's chair contacted Wayne Hale, the Shuttle Program Manager for Launch Integration, to request on-orbit pictures of Columbia's wing to assess its damage. After receiving notification of the debris strike, engineers at NASA, United Space Alliance, and Boeing created the Debris Assessment Team and began working to determine the damage to the orbiter.[5]^{: 140, 143} Intercenter Photo Working Group believed that the orbiter's RCC tiles were possibly damaged; NASA program managers were less concerned over the danger caused by the debris strike.[5]^: 141

Boeing analysts attempted to model the damage caused to the orbiter's TPS from the foam strike. The software models predicted damage that was deeper than the thickness of the TPS tiles, indicating that the orbiter's aluminum skin would be unprotected in that area. The Debris Assessment Team dismissed this conclusion as inaccurate, due to previous instances of predictions of damage greater than the actual damage. Further modeling specific to the RCC panels used software calibrated to predict damage caused by falling ice. The software predicted only one of 15 scenarios that ice would cause damage, leading the Debris Assessment Team to conclude there was minimal damage due to the lower density of foam to ice.[5]^{: 143–145}

The crew's view of re-entry

To assess the possible damage to Columbia's wing, members of the Debris Assessment Team made multiple requests to get imagery of the orbiter from the Department of Defense (DoD). Imagery requests were channeled through both the DoD Manned Space Flight Support Office and the Johnson Space Center Engineering Directorate.[5]^{: 150–151} Hale coordinated the request through a DoD representative at KSC. The request was relayed to the U.S. Strategic Command (USSTRATCOM), which began identifying imaging assets that could observe the orbiter. The imagery request was soon rescinded by NASA Mission Management Team Chair Linda Ham after she investigated the origin of it. She had consulted with Flight Director Phil Engelauf and members of the Mission Management Team, who stated that they did not have a requirement for imagery of Columbia. Ham did not consult with the Debris Assessment Team, and cancelled the imagery request on the basis that it had not been made through official channels.[5]^{: 152–153} Maneuvering the orbiter to allow its left wing to be imaged would have interrupted ongoing science operations, and Ham dismissed the DoD imaging capabilities as insufficient to assess damage to the orbiter.[5]^{: 153–154} Following the rejection of their imagery request, the Debris Assessment Team did not make further requests for the orbiter to be imaged.[5]^: 157

Throughout the flight, members of the Mission Management Team were less concerned than the Debris Assessment Team about the potential risk of a debris strike. The loss of bipod foam on STS-107 was compared to previous foam strike events, none of which caused the loss of an orbiter or crew. Ham, scheduled to work as an integration manager for STS-114, was concerned with the potential delays from a foam loss event.[5]^{: 147–148} Mission management also downplayed the risk of the debris strike in communications with the crew.[5]^: 161 On January 23, flight director Steve Stich sent an e-mail to Husband and McCool to tell them about the foam strike and inform them there was no cause for concern about damage to the TPS, as foam strikes had occurred on previous flights.[5]^: 159

During ascent at approximately 80 seconds, photo analysis shows that some debris from the area of the -Y ET Bipod Attach Point came loose and subsequently impacted the orbiter left wing, in the area of transition from Chine to Main Wing, creating a shower of smaller particles. The impact appears to be totally on the lower surface and no particles are seen to traverse over the upper surface of the wing. Experts have reviewed the high speed photography and there is no concern for RCC or tile damage. We have seen this same phenomenon on several other flights and there is absolutely no concern for entry.[5]^: 159

The crew were also sent a fifteen-second video of the debris strike in preparation for a press conference, but were reassured that there were no safety concerns.[5]^: 161

On January 26, the Debris Assessment Team concluded that there were no safety concerns from the debris strike. The team's report was critical of the Mission Management Team for asserting that there were no safety concerns before the Debris Assessment Team's investigation had been completed.[5]^: 164 On January 29, William Readdy, the Associate Administrator for Space Flight, agreed to DoD imaging of the orbiter, but on the condition that it would not interfere with flight operations; ultimately, the orbiter was not imaged by the DoD during the flight. At a Mission Management Team on January 31, the day before Columbia reentered the atmosphere, the Launch Integration Office voiced Ham's intention to review on-board footage to view the missing foam, but concerns of crew safety were not discussed.[5]^: 166

Columbia was scheduled to reenter the atmosphere and land on February 1, 2003. At 3:30 am EST the Entry Flight Control Team started its shift at the Mission Control Center.[5]^: 38 Onboard the orbiter, the crew stowed loose items and prepared their equipment for reentry.[14]^: 1.5

At 45 minutes before the deorbit burn, Husband and McCool began working through the entry checklist.[14]^: 1.6 At 8:10 am the Capsule Communicator (CAPCOM), Charlie Hobaugh,[15] informed the crew that they were approved to conduct the deorbit burn. At 8:15:30 the crew successfully executed the deorbit burn, which lasted 2 minutes and 38 seconds. At 8:44:09 Columbia reentered the atmosphere at an altitude of 400,000 feet (120 km), a point named entry interface. The damage to the TPS on the orbiter's left wing allowed for hot air to enter and begin melting the aluminum structure.[5]^: 9 Four and a half minutes after entry interface, a sensor began recording greater-than-normal amounts of strain on the left wing; the sensor's data was recorded to internal storage and not transmitted to the crew or ground controllers.[5]^: 38 The orbiter began to turn (yaw) to the left as a result of the increased drag on the left wing, but this was not noticed by the crew or mission control because of corrections from the orbiter's flight control system.[14]^: 1.8 This was followed by sensors in the left wheel well reporting a rise in temperature.[14]^: 1.10

Columbia at about 08:57. Debris is visible coming from the left wing (bottom). The image was taken at Starfire Optical Range at Kirtland Air Force Base.

At 8:53:46 am, Columbia crossed over the California coast; it was traveling at Mach 23 at an altitude of 231,600 feet (70.6 km), and the temperature of its wings' leading edges was estimated to be 2,800 °F (1,540 °C).[5]^: 38 Soon after it entered California airspace, the orbiter shed several pieces of debris, events observed on the ground as sudden increases in brightness of the air around the orbiter. The MMACS officer reported that the hydraulic sensors in the left wing had readings below the sensors' minimum detection thresholds at 8:54:24 am. Columbia continued its reentry and traveled over Utah, Arizona, New Mexico, and Texas, where observers would report seeing signs of debris being shed.[5]^: 39

At 8:58:03, the orbiter's aileron trim changed from the predicted values because of the increasing drag caused by the damage to the left wing. At 8:58:21, the orbiter shed a TPS tile that would later land in Littlefield, Texas; it would become the westernmost piece of recovered debris.[14]^: 1.12 The crew first received an indication of a problem at 8:58:39, when the Backup Flight Software monitor began displaying fault messages for a loss of pressure in the tires of the left landing gear. The pilot and commander then received indications that the status of the left landing gear was unknown, as different sensors reported the gear was down and locked or in the stowed position.[14]^: 1.13 The drag of the left wing continued to yaw the orbiter to the left until it could no longer be corrected using aileron trim. The orbiter's Reaction Control System (RCS) thrusters began firing continuously to correct its orientation.[14]^: 1.14

The loss of signal (LOS) from Columbia occurred at 8:59:32. Mission control stopped receiving information from the orbiter at this time, and Husband's last radio call of "Roger, uh ..." was cut off mid-transmission.[5]^: 39[14]^: 1.14 One of the channels in the flight control software was bypassed as the result of a failed wire, and a Master Alarm began sounding on the flight deck.[14]^: 1.15 Loss of control of the orbiter is estimated to have begun several seconds later with a loss of hydraulic pressure and an uncontrolled pitch-up maneuver.[14]^: 1.16 The orbiter began flying along a ballistic trajectory, which was significantly steeper and had more drag than the previous gliding trajectory.[14]^: 1.17 The orbiter, while still traveling faster than Mach 15, entered into a flat spin of 30° to 40° per second. The acceleration that the crew was experiencing increased from approximately 0.8 g to 3 g, which would have likely caused dizziness and disorientation, but not incapacitation.[14]^: 1.18 The autopilot was switched to manual control and reset to automatic mode at 9:00:03; this would have required the input of either Husband or McCool, indicating that they were still conscious and able to perform functions at the time. All hydraulic pressure was lost, and McCool's final switch configurations indicate that he had tried to restore the hydraulic systems at some time after 9:00:05.[14]^: 1.20

At 9:00:18, the orbiter began a catastrophic breakup, and all on-board data recording soon ceased.[14]^: 1.20 Ground observers noted a sudden increase in debris being shed, and all on-board systems lost power. By 9:00:25, the orbiter's fore and aft sections had separated from one another.[14]^: 1.21 The sudden jerk caused the crew compartment to collide with the interior wall of the fuselage, resulting in a depressurization of the crew compartment by 9:00:35.[14]^: 1.22 The pieces of the orbiter continued to break apart into smaller pieces, and within a minute after breakup were too small to be detected by ground-based videos. By 9:35, all debris and crew remains were estimated to have impacted the ground.[14]^: 1.77

The loss of signal occurred at a time when the Flight Control Team expected brief communication outages as the orbiter stopped communication via the west tracking and data relay satellite (TDRS). Personnel in Mission Control were unaware of the in-flight break-up, and continued to try to reestablish contact with the orbiter.[5]^: 43 At approximately 9:06, when Columbia would have been conducting its final maneuvers to land, a Mission Control member received a phone call concerning news coverage of the orbiter breaking up. This information was passed on to the Entry Flight Director, LeRoy Cain, who initiated contingency procedures.[5]^: 44 At KSC, where Columbia had been expected to land at 9:16, NASA Associate Administrator and former astronaut William Readdy also began contingency procedures after the orbiter did not land as scheduled.[16]^: 5

During reentry, all seven of the STS-107 crew members were killed, but the exact time of their deaths could not be determined. The level of acceleration that they experienced during crew module breakup was not lethal.[5]^: 77 The first lethal event the crew experienced was the depressurization of the crew module. The rate and exact time of depressurization could not be determined, but occurred no later than 9:00:59. The remains of the crew members indicated they all experienced depressurization. The astronauts' helmets have a visor that, when closed, can temporarily protect the crew member from depressurization. Some of the crew members had not closed their visors, and one was not wearing a helmet; this would indicate that depressurization occurred quickly before they could take protective measures.[14]^: 1.24[17]^: 103

During and after the breakup of the crew module, the crew, either unconscious or dead, experienced rotation on all three axes. The astronauts' shoulder harnesses were unable to prevent trauma to their upper bodies, as the inertia reel system failed to retract sufficiently to secure them, leaving them only restrained by their lap belts. The helmets were not conformal to the crew members' heads, allowing head injuries to occur inside of the helmet. The neck ring of the helmet may have also acted as a fulcrum that caused spine and neck injuries. The physical trauma to the astronauts, who could not brace to prevent such injuries, also could have resulted in their deaths.[14]^: 1.25[17]^{: 103–105}

The astronauts also likely suffered from significant thermal trauma. Hot gas entered the disintegrating crew module, burning the crew members, whose bodies were still somewhat protected by their ACES suits. Once the crew module fell apart, the astronauts were violently exposed to windblast and a possible shock wave, which stripped their suits from their bodies. The crews' remains were exposed to hot gas and molten metal as they fell away from the orbiter.[17]^{: 106–108}

After separation from the crew module, the bodies of the crew members entered an environment with almost no oxygen, very low atmospheric pressure, and both high temperatures caused by deceleration, and extremely low ambient temperatures.[17]^: 93 Their bodies hit the ground with lethal force.[14]^: 1.29

After looking at sensor data, the CAIB considered damage to the left wing as a likely culprit for Columbia's destruction. It investigated that recovered debris and noted the difference in heat damage between the two wings. RCC panels from the left wing were found in the western portion of the debris field, indicating that it was shed first before the rest of the orbiter disintegrated.[5]^: 73–74 X-ray and chemical analysis was conducted on the RCC panels, revealing the highest levels of slag deposits to be in the left wing tiles.[5]^: 75–76 Impact testing was conducted at the Southwest Research Institute, using a nitrogen-powered gun to fire a projectile made of the same material as the ET bipod foam. Panels taken from Enterprise, Discovery , and Atlantis were used to determine the projectiles' effect on RCC panels.[5]^: 79–80 A test on RCC panel 8, taken from Atlantis, was the most consistent with the damage observed on Columbia, indicating it was the damaged panel that led to the in-flight breakup.[5]^: 82–83

The CAIB was critical of NASA organizational culture, and compared its current state to that of NASA leading up to the Challenger disaster.[5]^: 99 It concluded that NASA was experiencing budget constraints while still expecting to keep a high level of launches and operations.[5]^: 100 Program operating costs were lowered by 21% from 1991 to 1994,[5]^: 107 despite a planned increase in the yearly flight rate for assembly of the International Space Station.[5]^: 114 Despite a history of foam strike events, NASA management did not consider the potential risk to the astronauts as a safety-of-flight issue.[5]^: 126 The CAIB found that a lack of a safety program led to the lack of concern over foam strikes.[5]^: 177 The board determined that NASA lacked the appropriate communication and integration channels to allow problems to be discussed and effectively routed and addressed.[5]^: 187 This risk was further compounded by pressure to adhere to a launch schedule for construction of the ISS.[5]^: 198

In its report, the CAIB discussed potential options that could have saved Columbia's crew.[33] They determined that the mission could have originally been extended for up to 30 days, after which the lithium hydroxide canisters that were used to remove carbon dioxide would have run out.[5]^: 173 On STS-107, Columbia was carrying the Extended Duration Orbiter, which increased its supply of oxygen and hydrogen.[34]^: 398 To maximize the mission duration, non-essential systems would have been powered down,[34]^: 399 and animals in the Spacehab module would have been euthanized.[34]^: 397

When STS-107 launched, Atlantis was undergoing preparation for the STS-114 launch on March 1, 2003. Had NASA management decided to launch a rescue mission, an expedited process could have begun to launch it as a rescue vehicle. Some pre-launch tests would have been eliminated to allow it to launch on time. Atlantis would have launched with additional equipment for EVAs, and launched with a minimum required crew. It would have rendezvoused with Columbia, and the STS-107 crew would have conducted EVAs to transfer to Atlantis. Columbia would have been remotely deorbited; as Mission Control would have been unable to remotely land it, it would have been disposed of in the Pacific Ocean.[34]^{: 400–404}

The CAIB also investigated the possibility of on-orbit repair of the left wing. Although there were no materials or adhesives onboard Columbia that could have survived reentry, the board researched the effectiveness of stuffing materials from the orbiter, crew cabin, or water into the RCC hole. They determined that the best option would have been to harvest tiles from other places on the orbiter, shape them, and then stuff them into the RCC hole. Given the difficulty of on-orbit repair and the risk of further damaging the RCC tiles, the CAIB determined that the likelihood of a successful on-orbit repair would have been low.[34]^{: 405–406}

The Space Shuttle program was suspended after the loss of Columbia.[35] The further construction of the International Space Station (ISS) was delayed, as the Space Shuttle had been scheduled for seven missions to the ISS in 2003 and 2004 to complete its construction.[36] To prevent future foam strikes, the ET was redesigned to remove foam from the bipod. Instead, electric heaters were installed to prevent ice building up in the bipod due to the cold liquid oxygen in its feedlines.[37] Additional heaters were also installed along the liquid oxygen line, which ran from the base of the tank to its interstage section.[38] NASA also improved its ground imaging capabilities at Kennedy Space Center to better observe and monitor potential issues that occur during launch. The existing cameras at LC-39A, LC-39B, and along the coast were upgraded, and nine new camera sites were added. The camera on the belly of the orbiter was changed from a film camera to a digital camera to allow images of the ET to be viewed on the ground soon after launch.[37] The Orbiter Boom Sensor System, a camera on the end of the Canadarm, was added to allow the crew to inspect the orbiter for any tile damage once they reached orbit.[39][40] Each of the orbiter's wings was equipped with 22 temperature sensors to detect any breaches during reentry and with 66 accelerometers to detect an impact. Post-landing inspection procedures were updated to include technicians examining the RCC panels using flash thermography.[39]

As well as the updates to the orbiter, NASA prepared contingency plans in the event that a mission would be unable to safely land. The plan involved the stranded mission docking with the ISS, on which the crew would inspect and attempt to repair the damaged orbiter. If they were unsuccessful, they would remain aboard the ISS and wait for a rescue.[41]^: 81 The rescue mission, designated STS-3xx, would be activated, and would use the next-in-line hardware for the orbiter, ET, and SRBs. The expected time to launch would be 35 days, as that was the requirement to prepare launch facilities.[41]^: 89–91 Before the arrival of the rescue mission, the stranded crew would power up the damaged orbiter, which would be remotely controlled as it was undocked and deorbited, and its debris would land in the Pacific Ocean.[41]^: 62 The minimal crew would launch, dock with the ISS, where it would spend a day transferring astronauts and equipment before undocking and landing.[41]^: 89–91

The STS-114 ET losing the largest piece of foam

The first Return to Flight mission, STS-114, began with the launch of Discovery on July 26, 2005, at 10:39 am (EDT).[42] Sixteen pieces of foam from the ET were dislodged during the launch that were large enough to be considered significant by NASA investigators,[43]^: 7 including one piece that was approximately 36 by 11 inches (91 by 28 cm).[43]^: 18 Post-launch investigations did not find any indications of damage from the foam loss, but ET video did reveal that a small piece of TPS tile from the nose landing gear fell off during launch.[44] Upon reaching orbit the crew inspected Discovery with the Orbiter Boom Sensor System. On July 29 Discovery rendezvoused with the ISS and, before docking, performed the first rendezvous pitch maneuver to allow the crew aboard the ISS to observe and photograph the orbiter's belly.[45] The next day, astronauts Soichi Noguchi and Stephen Robinson performed the first of three spacewalks. They tested a tile repair tool, the Emittance Wash Applicator, on intentionally damaged TPS tiles that had been brought in the payload bay.[46]^: 62 On August 3 the same astronauts performed the third EVA of the mission, during which Robinson stood on the ISS's Canadarm2 and went to Discovery's belly to remove two gap fillers between tiles that had begun to protrude.[46]^: 63[47] After a delay due to bad weather at KSC, the decision was made to land at Edwards AFB. Discovery successfully landed at 8:11 am (EDT) on August 9.[42][45] Had Discovery been unable to safely land, the crew would have remained on the ISS until Atlantis was flown to rescue them.[48] As a result of the foam loss, NASA grounded the Space Shuttle fleet again.[35][44]

To address the problem of foam loss for the second Return to Flight mission (STS-121), NASA engineers removed the foam ramp from the protuberance air load (PAL) on the ET, which was the source of the largest piece of debris on STS-114.[49] The launch was postponed from its scheduled launch of July 1, 2006, and again on July 2 due to inclement weather at KSC.[50] On July 3 a piece of foam approximately 3 by .25 inches (7.62 by 0.64 cm) and weighing 0.0057 pounds (2.6 g) broke off from the ET.[51][52] The mission still launched as scheduled at 2:38 pm (EDT) on July 4.[53] After reaching orbit, Discovery performed post-launch inspections of its TPS and docked with the ISS on July 6.[50] The orbiter carried a 28 feet (8.5 m) remote control orbiter in-flight maintenance cable that could connect the flight deck systems to the avionics system in the mid-deck; it would allow the spacecraft to be landed remotely, to include controlling the landing gear and deploying the parachute.[54][55] On July 12 astronauts Piers Sellers and Michael Fossum performed an EVA to test the NonOxide Adhesive eXperiment (NOAX), which applied protective sealant to samples of damaged TPS tiles.[46]^: 70 Discovery undocked from the ISS on July 14 and safely landed at 9:14 am on July 17 at KSC.[50] Had the crew been stranded in orbit, NASA planned to launch Atlantis to rescue them from the ISS.[56][57]

In January 2004 President Bush announced the Vision for Space Exploration, calling for the Space Shuttle fleet to complete the ISS and be retired by 2010, to be replaced by a newly developed Crew Exploration Vehicle for travel to the Moon and Mars.[58] In 2004, NASA Administrator Sean O'Keefe canceled a planned servicing of the Hubble Space Telescope and decided that future missions would all rendezvous with the ISS to ensure the safety of the crew.[59] In 2006, his successor, Michael Griffin, decided to have one more servicing mission to the telescope, STS-125, which flew in May 2009.[60][61] The retirement of the Space Shuttle was delayed until 2011,[62][63] after which no further crewed spacecraft were launched from the United States until 2020 when SpaceX's Crew Dragon Demo-2 mission successfully carried NASA astronauts Doug Hurley and Robert Behnken to the ISS.[64]