SpaceX reusable launch system development program

Grasshopper rocket performing a 325 meter flight followed by a soft propulsive landing in an attempt to develop technologies for a reusable launch vehicle.

SpaceX used a set of experimental technology-demonstrator, suborbital reusable launch vehicles (RLV) to begin flight testing their reusable booster technologies in 2012. Two versions of the prototype reusable test rockets were built—the 106-foot (32 m) tall Grasshopper (formerly designated as Grasshopper v1.0) and the 160-foot (49 m) tall Falcon 9 Reusable Development Vehicle, or F9R Dev1—formerly known as Grasshopper v1.1[74]—as well as a capsule prototype for testing propulsive landings of the Dragon crew and cargo capsule for the Falcon 9—DragonFly.[74] Grasshopper was built in 2011–2012 for low-altitude, low-velocity hover testing that began in September 2012 and concluded in October 2013 after eight test flights.[16][17][74] The second prototype vehicle design, F9R Dev1, was built on the much larger Falcon 9 v1.1 booster stage was used to further extend the low-altitude flight testing envelope on a vehicle that better matched the actual flight hardware, and made five test flights in 2014.[74][94][95] The low-altitude, low-speed flights of the test vehicle rockets and capsule were conducted at the SpaceX Rocket Test Facility in McGregor, Texas[16][17][74]

SpaceX indicated in November 2018 that they considered testing a heavily modified Falcon 9 second stage that would look like a "mini-BFR Ship" and be used for atmospheric reentry testing of a number of technologies needed for the full-scale spaceship, including an ultra-light heat shield and high-Mach control surfaces,[96][97] but two weeks later, Musk dismissed the approach in favor of using a full-diameter BFR instead.[98]

Grasshopper, the company's first VTVL test vehicle, consisted of a Falcon 9 v1.0 first-stage tank, a single Merlin-1D engine, and four permanently attached steel landing legs. It stood 106 feet (32 m) tall.[17] SpaceX built a 0.5-acre (0.20 ha) concrete launch facility at its Rocket Development and Test Facility in McGregor, Texas to support the Grasshopper flight test program.[99] Grasshopper was also known as Grasshopper version 1.0, or Grasshopper v1.0, prior to 2014 during the time the followon Grasshopper-class test vehicles were being built.

In addition to three test flights in 2012, five additional tests were successfully flown by the end of October 2013—including the fourth test overall in March 2013—in which Grasshopper doubled its highest leap to rise to 80.1 meters (263 ft) with a 34-second flight.[100] In the seventh test, in August 2013, the vehicle flew to 250 meters (820 ft) during a 60-second flight and executed a 100-meter (330 ft) lateral maneuver before returning to the pad.[101] Grasshopper made its eighth and final test flight on October 7, 2013, flying to 744 meters (2,441 ft) before making its eighth successful landing.[102] The Grasshopper test vehicle is now retired.[103]

As early as October 2012, SpaceX discussed development of a second-generation Grasshopper test vehicle, which was to have lighter landing legs that fold up on the side of the rocket, a different engine bay, and would be nearly 50% longer than the first Grasshopper vehicle.[95] In March 2013, SpaceX announced that the larger Grasshopper-class suborbital flight vehicle would be constructed out of the Falcon 9 v1.1 first-stage tank that was used for qualification testing at the SpaceX Rocket Development and Test Facility in early 2013. It was rebuilt as the F9R Dev1 with extensible landing legs. Five test flights occurred in 2014.[74]

The second VTVL flight test vehicle—F9R Dev1, built on the much longer Falcon 9 v1.1 first-stage tank, with retractable landing legs—made its first test flight on April 17, 2014.[74][92] F9R Dev1 was used for low-altitude test flights in the McGregor, Texas area—projected maximum altitude below 3,000 meters (10,000 ft)[74]—with a total of five test flights, all made during 2014. This vehicle self-destructed as a safety measure during its fifth test flight on August 22, 2014.[104]

By April 2014, a third flight test vehicle—F9R Dev2—was being built and was planned to be flown at the high-altitude test range available at Spaceport America in New Mexico where it was expected to be flown at altitudes up to 91,000 meters (300,000 ft)-plus.[74] It was never flown as SpaceX moved the high-altitude testing program to its controlled-descent testing of used boosters following their use on a paid orbital launch and ascent.

DragonFly was a prototype test article for a propulsively landed version of the SpaceX Dragon capsule, a suborbital reusable launch vehicle (RLV), intended for low-altitude flight testing. As of May 2014 it was planned to undergo a test program in Texas at the McGregor Rocket Test Facility, during 2014–2015.[32][105]

The DragonFly test vehicle is powered by eight SuperDraco engines, arranged in a redundant pattern to support fault-tolerance in the propulsion system design.[106] SuperDracos utilize a storable propellant mixture of monomethyl hydrazine (MMH) fuel and nitrogen tetroxide oxidizer (NTO), the same propellants used in the much smaller Draco thrusters used for attitude control and maneuvering on the first-generation Dragon spacecraft.[105] While SuperDraco engines are capable of 73,000 newtons (16,400 lbf) of thrust, during use on DragonFly flight test vehicle each will be throttled to less than 68,170 newtons (15,325 lbf) to maintain vehicle stability.[105]

A test flight program of thirty flights was proposed in 2013–2014, including two propulsive assist (parachutes plus thrusters) and two propulsive landing (no parachutes) on flights dropped from a helicopter at an altitude of approximately 10,000 feet (3,000 m). The other 26 test flights were projected to take off from a pad: eight to be propulsive assist hops (landing with parachutes plus thrusters) and 18 to be full propulsive hops, similar to the Grasshopper and F9R Dev booster stage test flights.[105][106] As of 2014, the DragonFly test program was not expected to start until after the completion of the F9R Dev1 booster testing at the McGregor facility.[106]

CRS-6 booster landing attempt

In an arrangement highly unusual for launch vehicles, SpaceX began in 2013 using some first stages of the Falcon 9 v1.1 rockets for propulsive-return controlled-descent flight tests after they completed the boost phase of an orbital flight. Since the advent of spaceflight in 1957, launch vehicle boosters would ordinarily just be discarded after setting their payloads on their way. The over-water tests started by SpaceX took place in the Pacific and Atlantic oceans south of Vandenberg Air Force Base and east of Cape Canaveral Air Force Station. The first flight test occurred on September 29, 2013, after the second stage with the CASSIOPE and nanosat payloads separated from the booster. These descent and simulated landing tests continued over the next two years, with the second flight test taking place on April 18, 2014,[26][29][90] two more test in 2014, and four subsequent tests conducted in 2015.[107] SpaceX continued to make iterative and incremental changes to the booster design, as well as the specific reusable technologies, descent profile and propellant margins, on some 2016-2018 Falcon 9 and Falcon Heavy flights to tweak the design and operational parameters. Many of these descent and landing tests were tested on active orbital spaceflight missions for SpaceX customers as the booster reentered the atmosphere and attempted recoverable landings.

Following analysis of the flight test data from the first booster-controlled descent in September 2013, SpaceX announced it had successfully tested a large amount of new technology on the flight, and that coupled with the technology advancements made on the Grasshopper low-altitude landing demonstrator, they were ready to test a full recovery of the booster stage. The first flight test was successful; SpaceX said it was "able to successfully transition from vacuum through hypersonic, through supersonic, through transonic, and light the engines all the way and control the stage all the way through [the atmosphere]".[25] Musk said, "the next attempt to recovery [sic] the Falcon 9 first stage will be on the fourth flight of the upgraded rocket. This would be [the] third commercial Dragon cargo flight to ISS."[27]

This second flight test took place during the April 2014 Dragon flight to the ISS. SpaceX attached landing legs to the first stage, decelerated it over the ocean and attempted a simulated landing over the water, following the ignition of the second stage on the third cargo resupply mission contracted to NASA. The first stage was successfully slowed down enough for a soft landing over the Atlantic Ocean.[29] SpaceX announced in February 2014 the intent to continue the tests to land the first-stage booster in the ocean until precision control from hypersonic all the way through subsonic regimes has been proven.[90] Five additional controlled-descent tests were conducted in the remainder of 2014 through April 2015, including two attempts to land on a floating landing platform—a SpaceX-built Autonomous Spaceport Drone Ship—on the Atlantic Ocean east of the launch site, both of which brought the vehicle to the landing platform, but neither of which resulted in a successful landing.

Falcon 9 Flight 20's first stage landing viewed from a helicopter, December 22, 2015.

During the 2015 launch hiatus, SpaceX requested regulatory approval from the FAA to attempt returning their next flight to Cape Canaveral instead of targeting a floating platform in the ocean. The goal was to land the booster vertically at the leased Landing Zone 1 facility—the former Launch Complex 13 where SpaceX had recently built a large rocket landing pad.[108] The FAA approved the safety plan for the ground landing on December 18, 2015.[109] The first stage landed successfully on target at 20:38 local time on December 21 (01:38 UTC on December 22).[110][107]

First stage booster B1019 never flew again after the flight.[111] Rather, the rocket was moved a few miles north to the SpaceX hangar facilities at Launch pad 39A, recently refurbished by SpaceX at the adjacent Kennedy Space Center, where it was inspected before being used on January 15, 2016, to conduct a static fire test on its original launchpad, Launch Complex 40.[112] This test aimed to assess the health of the recovered booster and the capability of this rocket design to fly repeatedly in the future.[113][107] The tests delivered good overall results except for one of the outer engines experiencing thrust fluctuations.[113] Elon Musk reported that this may have been due to debris ingestion.[114] The booster was then retired to the SpaceX facility in Hawthorne, California.

First stage of Falcon 9 Flight 21 descending over the floating landing platform, January 17, 2016, immediately prior to a soft touchdown followed by deflagration of the rocket after a landing leg failed to latch, causing the rocket to tip over.

Falcon 9 Flight 21 launched the Jason-3 satellite on January 17, 2016, and attempted to land on the floating platform Just Read the Instructions,[115] located for the first time about 200 miles (320 km) out in the Pacific Ocean. Approximately 9 minutes into the flight, the live video feed from the drone ship went down due to the losing its lock on the uplink satellite. The vehicle landed smoothly onto the vessel but one of the four landing legs failed to lock properly, reportedly due to ice from the heavy pre-launch fog preventing a lockout collet from latching.[116] Consequently the booster fell over shortly after touchdown and was destroyed in a deflagration upon impact with the pad.[117][118]

Flight 22 was carrying a heavy payload of 5,271 kilograms (12,000 lb) to geostationary transfer orbit (GTO). This was heavier than previously advertised maximum lift capacity to GTO being made possible by going slightly subsynchronous. Following delays caused by failure of Flight 19, SpaceX agreed to provide extra thrust to the SES-9 satellite to take it supersynchronous.[119] As a result of these factors, there was little propellant left to execute a full reentry and landing test with normal margins. Consequently the Falcon 9 first stage followed a ballistic trajectory after separation and re-entered the atmosphere at high velocity, making it less likely to land successfully.[120][119] The atmospheric re-entry and controlled descent were successful despite the higher aerodynamical constraints on the first stage due to extra speed. However the rocket was moving too fast and was destroyed when it collided with the drone ship. SpaceX collected valuable data on the extended flight envelope required to recover boosters from GTO missions.

First stage of Falcon 9 Flight 23 landed on autonomous droneship

Starting in January 2015, SpaceX positioned stable floating platforms a few hundred miles off the coast along the rocket trajectory; those transformed barges were called autonomous spaceport drone ships.[121] On April 8, 2016, Falcon 9 Flight 23, the third flight of the full-thrust version, delivered the SpaceX CRS-8 cargo on its way to the International Space Station while the first stage conducted a boostback and re-entry maneuver over the Atlantic ocean. Nine minutes after liftoff, the booster landed vertically on the drone ship Of Course I Still Love You, 300 km from the Florida coastline, achieving a long-sought-after milestone for the SpaceX reusability development program.[122]

A second successful drone ship landing occurred on May 6, 2016, with the next flight which launched JCSAT-14 to GTO. This second landing at sea was more difficult than the previous one because the booster at separation was traveling about 8,350 km/h (5,190 mph) compared to 6,650 km/h (4,130 mph) on the CRS-8 launch to low Earth orbit.[123] Pursuing their experiments to test the limits of the flight envelope, SpaceX opted for a shorter landing burn with three engines instead of the single-engine burns seen in earlier attempts; this approach consumes less fuel by leaving the stage in free fall as long as possible and decelerating more sharply, thereby minimizing the amount of energy expended to counter gravity.[124] Elon Musk indicated this first stage may not be flown again instead being used as a life leader for ground tests to confirm others are good.[125]

A third successful landing followed on 27 May, again following deceleration from the high speed required for a GTO launch. The landing crushed a "crush core" in one leg, leading to a notable tilt to the stage as it stood on the drone ship.[62]

Over the subsequent missions, landing of the first stage gradually became a routine procedure, and since January 2017 SpaceX ceased to refer to their landing attempts as "experimental". Low-energy missions to the ISS fly back to the launch site and land at LZ-1, whereas more demanding satellite missions land on drone ships a few hundred miles downrange. Occasional missions with heavy payloads, such as EchoStar 23, do not attempt to land, flying in expendable configuration without fins and legs.

During 2016 and 2017, SpaceX has recovered a number of first stages to both land and drone ships, helping them optimize the procedures needed to re-use the boosters rapidly. In January 2016 Elon Musk estimated the likelihood of success at 70 percent for all landing attempts in 2016, hopefully rising to 90 percent in 2017; he also cautioned that we should expect "a few more RUDs" (Rapid Unscheduled Disassembly, Musk's acronym to denote destruction of the vehicle on impact).[126] Musk's prediction was vindicated, as 5 out of 8 flown boosters (63%) were recovered in 2016, and 14 out of 14 (100%) in 2017. Three GTO missions for heavy payloads (EchoStar 23 in March 2017, Inmarsat-5 F4 in May 2017 and Intelsat 35e in July 2017) were flown in an expendable configuration, not equipped for landing. One booster which could have been recovered was intentionally flown without legs and left to sink after a soft touchdown in the ocean (booster B1036 for the Iridium NEXT 31–40 mission in December 2017).

As of 6 August 2018, SpaceX had recovered 21 first-stage boosters from previous missions, of which six were recovered twice, yielding a total 27 landings. In 2017, SpaceX flew a total of 5 missions out of 20 with re-used boosters (25%). In total, 14 boosters have been re-flown as of August 2018.

On July 28, 2016, the first stage from the JCSAT-2B mission was successfully test-fired for a full duration at the SpaceX McGregor facility.[127] The first reuse attempt occurred on 30 March 2017[128] with the launch of SES-10,[129] resulting in a successful flight and second landing of the B1021 first stage recovered from the CRS-8 mission of April 2016.[130] Another reflight succeeded in June 2017 with BulgariaSat-1 riding the B1029 booster from the January 2017 Iridium NEXT mission.[131] Booster B1031 flew the CRS-10 mission to the ISS in February 2017 and helped loft communications satellite SES-11 to geostationary orbit in October 2017. Boosters B1035 and B1036 were flown twice each for the same customer, B1035 for NASA missions CRS-11 and CRS-13 in June and December 2017, and B1036 for two batches of 10 Iridium NEXT satellites, also in June and December 2017. B1032 was re-used for GovSat-1 in January 2018 after NROL-76 in May 2017.

SpaceX spent four months refurbishing the first booster to be re-used, B1021, and launched it again after approximately one year.[132] The second booster to be flown again, B1029, was refurbished in "only a couple of months"[4] and re-launched after five months.[131] Elon Musk has stated a goal to turn around a first stage within 24 hours.[133] Musk remains convinced that this long-term goal can be met by SpaceX rocket technology,[134] but has not stated that the goal would be achieved with the Falcon 9 design.

Boosters B1019 and B1021 were retired and put on display. B1029 was also retired after the BulgariaSat-1 mission. B1023, B1025, B1031 and B1035 were recovered a second time, while B1032 and B1036 were deliberately sunk at sea after a soft ocean touchdown.[135]

By mid-2019, having reflown any single booster only three times to date, SpaceX indicated that they plan to use a single booster at least five times by the end of 2019.[136] No booster achieved this timeline, but B1048 flew four times and two more (B1046 and B1049) made a fourth flight in January 2020. In March 2020, SpaceX first flew a booster (B1048) for the fifth time.[137]

Falcon Heavy test flight had no contracted customer, and in order to limit the cost on such a test flight, SpaceX targeted to have reused side-boosters. Boosters B1023 and B1025 that had been flown as a Falcon 9 configuration, were reconfigured and used as side boosters on the first flight of Falcon Heavy in February 2018, and then both landed side-by-side at almost the same time on the ground pads. Later Falcon Heavy flights used either new boosters, or side-boosters previously flown on a Falcon Heavy. SpaceX has been unable to recover the central core in any of the three Falcon Heavy, but managed to recover all the six-side boosters.[138]

With a streak of 19 successful recovery attempts of the first stage from 2016 through to early 2018, SpaceX has focused on rapid reusability of first stage boosters. Block 3 and Block 4 proved economically feasible to be flown twice, as 11 such boosters have been reflown in 2017 and 2018. Block 5 has been designed with multiple reuses in mind, up to 10 reuses with minimal inspection and up to 100 uses with refurbishment.[139] New aggressive reentry profiles were experimented with expendable Block 3 and Block 4 boosters in early 2018, to test out the limitations on the range of recoverable launch margins that are potential for future Block 5.[140] On 9 May 2021, B1051 became the first booster to be launched and landed for the tenth time, achieving one of SpaceX's milestone goals for reuse.[141] The current reuse record of 16 flights is jointly held by B1058 and B1060.

Payload fairings have traditionally been expendable, where they have either burned up in the atmosphere or were destroyed upon impacting the ocean. As early as mid-2015, Musk hinted that SpaceX might be working on fairing reusability, following the discovery of wreckage of an unidentified Falcon 9 launch vehicle section off the coast of The Bahamas, and was subsequently confirmed by SpaceX to be a component of a payload fairing that had washed ashore.[142] By April 2016, SpaceX had publicly announced Falcon 9 fairing recovery as an objective.[42] The cost of the fairing is about $6 million per launch, which accounts for approximately ten percent of the overall launch costs.[143] Musk said in 2017: "Imagine if you had $6 million in cash in a pallet flying through the air, and it was going to smash into the ocean. Would you try to recover that? Yes, yes you would."[144]

In March 2017, as part of the SES-10 mission, SpaceX for the first time performed a controlled landing of the payload fairing and successfully recovered a fairing half, aided by attitude-control thrusters and a steerable parachute, helping it glide towards a gentle touchdown on water.[3][42] The company announced intent to land the fairings eventually on a dry flexible structure, jokingly described by Musk as a "floating bouncy-castle", with the aim of full fairing reuse.[80] With successive tests and refinements on several flights, intact fairing recovery was stated as an objective for 2017, with reflight of a recovered fairing planned in 2018.[43]

The "bouncy castle" meme was in fact a net strung between large arms of a fast platform supply vessel named Mr. Steven (now GO Ms. Tree). The recovery vessel is equipped with dynamic positioning systems, and was tested after the launch of the Paz satellite from Vandenberg Air Force Base in 2017.[145][146] This mission was also the first to use a version 2 fairing, explicitly designed to "improve survivability for post-launch recovery attempts, and to be reusable on future missions".[147] This recovery attempt was not fully successful; the fairing missed the boat by a few hundred meters but landed intact in the water[148] before being recovered and taken back to port.[146] As of August 2018, all four attempts by SpaceX to land a fairing on a recovery ship had failed, despite fitting Mr. Steven with larger nets before the July 2018 attempt.[149][150]

In October 2018, at least two fairing recovery tests were performed, involving Mr. Steven and a helicopter, which would drop a fairing half from the height of about 3300 meters. The outcome of the tests was unclear.[151]

In April 2019, during the second Falcon Heavy mission, recovery boat Go Searcher fished the fairing halves out of the sea and it was announced the fairings would be used on a Starlink mission.[152] These fairings were reused in a Starlink mission on 11 November 2019.[153]

In June 2019, following the third Falcon Heavy launch, the first successful fairing catch was made. Images posted to Twitter hours after launch showed one half of the fairing nestled in the net of the recovery vessel GO Ms. Tree.[154]

By late 2020, payload fairings were being regularly recovered by SpaceX, with SpaceX dispatching two custom-modified recovery ships—Ms. Tree and Ms. Chief—to collect the fairings on most launches from their Florida launch site. By this time, SpaceX was also regularly reflying recovered fairings on launches, usually on their own flights where Starlink satellites are the primary or only payload. As of August 2020 however, successful net landings were not yet routine, with less than half of the fairings of the previous three months being caught in the nets, but most still recovered anyway after a soft landing in the ocean.

By April 2021, SpaceX had abandoned the experimental program to attempt recovery of dry payload fairings under parachute descent in a net on a fast ship. SpaceX decided to operationalize "wet recovery" of fairings on future Falcon 9 flights, having found that they can clean, refurbish, and reuse such fairings more economically.[155] SpaceX released Miss Tree and Miss Chief from their contracts and purchased two ships for fairing recovery operations as well as for towing and supporting droneships on the east coast. These two ships were named in honour of Demo-2 astronauts Doug Hurley and Bob Behnken as Doug[156] and Bob. The earlier names of the ships Bob and Doug were Ella G and Ingrid respectively. Currently, Doug is operating at Port Canaveral while Bob is at Tampa undergoing construction.

By 26 May 2021, SpaceX had launched 40 flights that reflew at least one previously-flown fairing half, and one fairing had flown on five different flights, having been recovered and cleaned four previous times.[144]

Despite early public statements that SpaceX would endeavor to make the Falcon 9 second-stage reusable as well, by late 2014, they determined that the mass needed for a re-entry heat shield, landing engines, and other equipment to support recovery of the second stage as well as the diversion of development resources from other company objectives was at that time prohibitive, and indefinitely suspended their second-stage reusability plans for the Falcon rockets.[157][158] However, in July 2017[43] they indicated that they might do experimental tests on recovering one or more second-stages in order to learn more about reusability to inform their Starship development process,[159] and in May 2018 provided additional details about how they might carry out some of that testing.[160]

The Starship is planned to replace all existing SpaceX launch and space vehicles after the mid-2020s: Falcon 9, Falcon Heavy and the Dragon spacecraft, aimed initially at the Earth-orbit launch market but with capability to support long-duration spaceflight in the cislunar and Mars mission environments.[161] Both stages will be fully reusable. The integrated second-stage-with-spaceship design has not been used in previous launch vehicles.[161]

SpaceX's Dragon capsules have been gradually improved for reuse. Structural elements and internal components are being refurbished between flights, while the heat shield is replaced for each new mission. The last newly built Dragon cargo capsule first flew in July 2017; all subsequent ISS resupply missions were conducted with refurbished capsules,[162] some capsules made a third flight.[163][164] Dragon's trunk section cannot be reused, as it is designed to burn up in the atmosphere after completing its mission.[165]

The SpaceX Dragon 2 capsule is reused as well. Initially it was planned to use new capsules for all crewed NASA missions[166] but experience with the demonstration missions led to NASA and SpaceX agreeing on reuse starting from Crew-2.[167][168]

Artist depiction by SpaceX of Big Falcon Rocket in flight

The first reference by SpaceX of a rocket concept with Starship lifting capabilities was in 2005. In a student conference, Musk briefly mentioned a theoretical heavy‑lift launch vehicle code-named BFR, later known as the Falcon XX.[176] It would be powered by a larger version of the Merlin engine, called Merlin 2, and feature a lifting capability of 140 metric tons (310,000 lb) to low Earth orbit.[177] In 2012, in a public discussion about a conceptual Mars colonization program, Musk described the Mars Colonial Transporter. He envisioned it as a reusable super heavy-lift launch vehicle that could deliver approximately 150 to 200 metric tons (330,000 to 440,000 lb) to low Earth orbit. The Mars Colonial Transporter might be powered by Raptors, consuming liquid methane and liquid oxygen.[178]

In September 2016, at the 67th International Astronautical Congress, Musk announced the Interplanetary Transport System (ITS), a conceptual reusable rocket conceived to launch humans to Mars and other destinations in the Solar System. The ITS was to be 122 m (400 ft) tall, 12 m (39 ft) wide, and capable of lifting 300 metric tons (660,000 lb) to low Earth orbit.[179] Both stages were to be made from carbon composites. The first stage or booster was to be powered by 42 Raptors, and the second stage by nine Raptors.[180] Once refueled while in Earth orbit, the spacecraft could accelerate to Mars.[181] When an Interplanetary Spaceship enters the atmosphere, it cools itself via transpiration and controls the spacecraft's descent by moving its delta wings and split flaps.[182] At the following Congress, Musk announced a replacement rocket called the Big Falcon Rocket or informally Big Fucking Rocket. The Big Falcon Rocket is 106 m (348 ft) tall and 9 m (30 ft) wide.[183] In that conference, he talked about a possible suborbital transportation feature and termed it Earth to Earth.[184]

In November 2018, the present names were first used: Super Heavy for the booster, Starship for the spacecraft, and Starship system or just Starship for the whole vehicle.[185] Around that time, Musk announced a redesigned spacecraft concept with three aft flaps and two forward flaps.[186] In January 2019, Musk announced that Starship would be made from stainless steel and stated that this might be stronger than an equivalent carbon composite in a wide range of temperatures.[187] In March, Musk tweeted that SpaceX opted for a heat shield composed of hexagonal ceramic tiles instead of transpiration.[188] In October, the Starship spacecraft's engine configuration was changed to three Raptors optimized for atmospheric pressure and three optimized for space.[189] The number of rear fins was reduced from three to two and placed at the heat shield's edges.[190]

SN8 launch at the Boca Chica launch site, firing three Raptors

On 27 August 2019, a simplified test article named Starhopper hopped 150 m (490 ft) high.[191] Unveiled in a SpaceX event in September 2019, Starship Mk1 (Mark 1) was the first full‑scale Starship test article to be built. The Mk2 in Florida was constructed five months later.[192] Neither flew: Mk1 was destroyed during a cryogenic proof test and Mk2 was scrapped.[193] In early 2020, SpaceX changed Mk3's name to SN1 (serial number 1).[194] During a cryogenic proof test on 28 February 2021, a fault in SN1's bottom tank caused it to crumble. On 8 March 2020, SN2 stripped-down test tank completed its only cryogenic proof test.[195] On 3 April 2020, during SN3's cryogenic proof test, a valve leaked the liquid nitrogen inside its lower tank, causing the vessel to depressurize and collapse.[196] After SN4's fifth successful static fire test on 29 May 2020, the quick disconnect fuel line caused it to explode.[197] On 15 June 2020, Musk tweeted that new prototypes would be made from SAE 304L instead of 301 stainless steel.[172] On 4 August 2020, SN5 completed a 150 m (490 ft) hop using a single Raptor, the first full-scale test article to complete a flight test intact.[198] On 24 August 2020, SN6 replicated SN5's flight path successfully.[199] SN7 was not completed, but as of October 2021, its tanks were salvaged for various experiments.[200]

SN8 was the first complete test article.[201] In October and November 2020, SN8 underwent four static fire tests; the first, second, and fourth were successful, but the third caused an engine shutdown. According to Musk, the force from the engine destroyed parts of the launch pad sending some pieces of it into the engine.[202] On 9 December 2020, SN8 performed the first flight by a Starship, reaching an altitude of 12.5 km (7.8 mi). During landing, its methane header tank did not provide sufficient fuel to the Raptors, reducing thrust from one engine. The test article exploded on impact.[203] On 2 February 2021, SN9 flew 10 km (6.2 mi) high.[204] While descending, one of its engines did not function and burst on landing at an angle.[205] On 3 March 2021, SN10 repeated SN9's flight path, then hard landed and destroyed itself in an explosion 8 minutes later.[206]

The first Super Heavy booster named BN1 (booster number 1) finished construction on 8 March 2021, but it had not received engines.[207] On 30 March 2021, SN11 exploded in midair without a confirmed explanation because of the dense fog at the launch site.[208] A possible explanation is that an engine might have burned the test article's avionics and could have caused a hard start on the engine's turbopump.[209] After the launch, SpaceX skipped SN12, SN13, SN14, and BN2, and incorporated obsolete test articles' improvements to SN15 instead.[210] On 5 May 2021, the test article flew the same flight path as previous test articles and soft landed successfully.[211][212] On 20 July 2021, BN3 fired its engines for the only time.[213] As of October 2021, SN15, SN16, and BN3 had been retired and displayed.[211][214]

As of October 2021, skipping over SN16, SN17, SN18, and SN19, SN20 along with BN4[lower-alpha 1] are targeted to an orbital flight scheduled for late 2021. However, a NASA public document that outlines NASA's plan to observe Starship's reentry inferred that the launch might happen near March 2022.[170] However, the article did not specify that the observed flight would be the first test flight of the Starship, so an approximate launch date is still unknown. The Federal Aviation Administration (FAA) has not approved the environmental impact statement draft released on 19 September 2021.[216][217] BN4 is expected to separate about three minutes into the orbital flight and splashdown in the Gulf of Mexico, approximately 30 km (19 mi) from the shoreline. The rocket's SN20 second stage ground track will then traverse the middle of the Straits of Florida between Key West and Cuba, a range safety consideration to assure any debris from a malfunction falls away from populated areas. Once over the Atlantic Ocean, SN20 is then expected to accelerate close to orbital speed and splashdown around ninety minutes later 100 km (60 mi) northwest of Kauai, Hawaii.[218][219]

• Pictures of Starship test articles
• 
  Starship Mk1's nose cone near construction tents
• 
  Starhopper in construction
• 
  A crane lifting Starship SN5
• 
  Starship SN7's tank
• 
  Starship SN9 on the launchpad
• 
  A worker is examining Starship SN20's ceramic tiles