RS-25: The Space Shuttle Engine That Now Powers the SLS Moon Rocket
The RS-25 is the most tested, most flown, most sophisticated liquid hydrogen rocket engine ever built. It flew on all 135 Space Shuttle missions, accumulated over a million seconds of test-fire time, and now sits at the base of NASA’s Space Launch System — the rocket designed to send astronauts back to the Moon. There is just one bitter irony: an engine originally built to be reusable is now thrown into the Atlantic Ocean after a single use.
But that irony barely scratches the surface of the RS-25’s story. This is an engine that nearly killed the Space Shuttle program before it ever flew, survived a development so troubled that Congress held hearings titled “The Engine That Couldn’t,” went on to compile one of the most remarkable reliability records in spaceflight history, and is now being manufactured new — at $100 million or more per copy — for a rocket that cannot bring them back.
This guide covers the engineering, the history, the performance numbers, and the strange second life of the most capable rocket engine America has ever built.
What Makes the RS-25 Special
The RS-25 is a staged combustion, liquid hydrogen / liquid oxygen (hydrolox) engine. If those words do not mean much to you yet, here is the short version: it squeezes every possible joule of energy out of its propellants in a way that most rocket engines cannot match.
In a “staged combustion” cycle, the propellants are burned twice. First, a small amount is combusted in a preburner to drive the turbopumps that feed propellant to the main engine. Then those hot, fuel-rich exhaust gases from the preburner are injected into the main combustion chamber, where they burn again with the remaining oxygen. Nothing is wasted. In simpler engine cycles (like the “gas generator” cycle used by the Merlin engine), the turbopump exhaust is dumped overboard, wasting 2-3% of the propellant.
The result is an engine with a vacuum specific impulse (Isp) of 452 seconds — the highest of any hydrogen engine at its thrust class. To put that in perspective, the Merlin engine achieves 311 seconds in vacuum. The RS-25 wrings about 45% more impulse out of every kilogram of propellant. That is the difference between reaching orbit and falling short.
The Numbers That Matter
At full power, the RS-25 produces 2,279 kilonewtons (512,000 pounds-force) of thrust in vacuum. Its combustion chamber pressure is 206 atmospheres — roughly the pressure you would feel at the bottom of a 2-kilometer-deep ocean. The hydrogen turbopump spins at 37,000 RPM, generating 71,000 horsepower from a unit the size of an automobile engine. The oxygen turbopump adds another 28,000 horsepower. Combined, the two turbopumps in a single RS-25 produce roughly the horsepower of two aircraft carrier propulsion plants.
The engine weighs 3,527 kilograms — about the weight of a large SUV. It stands 4.3 meters tall with a nozzle exit diameter of 2.4 meters. The nozzle expansion ratio is 77.5:1, a carefully chosen compromise that lets the engine operate from sea level all the way to vacuum. Most engines are optimized for one or the other; the RS-25 had to do both because the Shuttle had no expendable first stage.
The Development Nightmare
The RS-25 was developed by Rocketdyne (later acquired by Aerojet Rocketdyne, now part of L3Harris) starting in the early 1970s. NASA wanted an engine that was reusable for 55 flights, throttleable from 65% to 109% of rated power, and capable of operating at combustion pressures far beyond anything previously attempted. It was, by a wide margin, the most ambitious rocket engine specification ever written.
The development was brutal. The high chamber pressures caused resonant combustion instabilities that cracked and destroyed engine components. The hydrogen turbopump bearings failed repeatedly under the extreme RPM and cryogenic temperatures. The sheet-metal nozzle tubes that carry cooling hydrogen cracked from thermal cycling. Between 1975 and 1978, engine after engine was destroyed on the test stand.
“The Engine That Couldn’t”
By 1977, the situation was dire enough to attract Congressional attention. The Senate held hearings with the blunt title “The Engine That Couldn’t.” Rocketdyne engineers were questioned about whether the RS-25 would ever work, and whether the entire Shuttle program was at risk because of a single component.
The core problem was that Rocketdyne was attempting multiple engineering firsts simultaneously: the highest chamber pressure of any production engine, the first full-flow staged combustion cycle at this scale, the first reusable engine rated for dozens of flights, and extreme throttling range. Each of these alone would have been a challenging development program. Attempting all four in one engine, on a fixed schedule tied to the Shuttle program, was nearly impossible.
What saved the program was dogged, iterative engineering. The combustion instability was tamed through careful injector redesign — changing the pattern and sizing of the holes through which hydrogen and oxygen spray into the chamber. The turbopump bearings were redesigned with better materials and improved cooling. The nozzle tube cracks were addressed by changing the brazing process and adjusting the tube geometry. Each fix solved one problem, often revealing the next one.
The first Shuttle flight, STS-1, launched on April 12, 1981 — roughly four years behind the original schedule. The three RS-25 engines performed flawlessly.
The 109% Power Level Mystery
If you have ever watched a Shuttle launch and heard the call “throttle up” followed by “engines at 109 percent,” you might have wondered: how can an engine run at more than 100%? Does that mean it is exceeding its design limits?
No. The naming convention is a historical artifact. When the RS-25 was first designed, its rated thrust was set at a specific value — call it the “100% Rated Power Level” (RPL). That is what the engine was nominally designed to produce. During development, engineers realized the engine could safely operate at higher thrust levels, so they certified it for additional power.
Rather than redefining 100%, they simply added percentage points above it. “104%” meant 104% of the original rated power level. “109%” — the level used for most Shuttle missions — meant 109% of that original baseline. In absolute terms, 109% RPL corresponds to 2,279 kN of vacuum thrust.
NASA later certified the engines for “111%” RPL for emergency abort scenarios, where squeezing out an extra 2% thrust could mean the difference between reaching orbit and ditching in the ocean. Some engines were even tested to 113% on the ground, though this was never used in flight.
The whole percentage system is confusing and NASA knows it. For the SLS program, they have quietly redefined the baseline: what was “109%” on the Shuttle is now called “100%” on SLS. Same thrust, less confusing terminology.
135 Missions, 1 In-Flight Shutdown
Across all 135 Space Shuttle missions from 1981 to 2011, the RS-25 compiled an extraordinary reliability record. The only in-flight engine shutdown occurred on STS-51F in July 1985, and the story is worth telling because it reveals how robust the system actually was.
About five and a half minutes into flight, a temperature sensor on the center engine (Engine #1) sent a reading indicating the turbopump was overheating. The engine’s onboard controller automatically shut the engine down to prevent damage. The Shuttle immediately transitioned to an “Abort to Orbit” — a contingency mode where the remaining two engines burn longer to reach a lower-than-planned but still functional orbit.
Here is the thing: the sensor was wrong. Post-flight analysis determined that the temperature was actually normal — the sensor itself had failed, sending a false high reading. The engine was performing perfectly. A faulty nickel wire in a thermocouple cost the mission its planned orbit altitude.
It got even more dramatic. Shortly after the center engine shutdown, a sensor on one of the remaining two engines also started giving anomalous readings. Had the automatic shutdown system triggered again, the Shuttle would have lost a second engine — and with only one remaining, it would have had to attempt a transatlantic abort landing, one of the most dangerous contingency procedures ever devised. Flight controller Jenny Howard overrode the automatic shutdown system by sending a command to inhibit further automatic shutdowns, keeping the second engine running. The mission reached orbit successfully.
After STS-51F, NASA modified the shutdown procedures to require confirmation from multiple sensors before automatically killing an engine. One faulty sensor would never again be able to shut down an otherwise healthy engine during flight.
From Reusable to Expendable: The SLS Paradox
The RS-25 was designed, from its very first specification document, to be reusable. Each engine was built to fly 55 missions with maintenance between flights. After a Shuttle mission, the three RS-25s were removed from the orbiter, inspected, refurbished as needed, and reinstalled for the next flight. Some individual engines flew 20 or more missions.
Then the Space Shuttle retired in 2011, and NASA chose the RS-25 to power the core stage of the Space Launch System. There was just one problem: SLS is expendable. After each launch, the core stage — along with its four RS-25 engines — drops into the ocean and sinks.
This means engines that were meticulously designed for reuse, built with the precision of a Swiss watch, and originally costing around $40 million each (in 1980s dollars) are now being used once and discarded. At current production costs exceeding $100 million per engine, each SLS launch throws away over $400 million in engines alone.
Why Not Recover Them?
It is a reasonable question. If SpaceX can land and reuse a Merlin-powered booster, why can’t NASA recover the SLS core stage and its RS-25s?
The short answer is that SLS was never designed for it. The core stage has no landing legs, no grid fins, no re-entry thermal protection, and no ability to restart its engines after cutoff. Adding those capabilities would require a fundamental redesign of the vehicle — effectively building a new rocket. By the time SLS was designed in 2011, the architecture was locked: expendable core stage, reusable Orion capsule, expendable solid boosters.
There is also a programmatic reality. SLS exists partly to maintain the workforce and supply chain that built the Space Shuttle. Building new RS-25s keeps production lines running and engineers employed. Recovering and reusing them would reduce the number of new engines needed, which would undermine one of the program’s political justifications. This is not conspiracy — it is openly acknowledged in congressional testimony.
The New RS-25: Cheaper (Relatively Speaking)
The 16 leftover RS-25 engines from the Shuttle program were enough for the first four SLS missions. After that, NASA needs new production engines. Aerojet Rocketdyne (now L3Harris) has been contracted to build new RS-25s with design modifications aimed at reducing cost and improving manufacturability.
The new engines, designated RS-25E (“E” for expendable), incorporate changes like a simplified nozzle manufacturing process, reduced inspection requirements (since they only need to work once), and additive manufacturing (3D printing) for some components. The goal is to bring the per-engine cost down from the Shuttle-era equivalent of $100M+ to something lower, though the exact target and achieved costs are subjects of ongoing debate.
One notable change: the new RS-25s are rated for 111% RPL as their standard operating power level — a thrust that was reserved for abort scenarios on the Shuttle. The SLS needs all the thrust it can get, and since the engines only fire once, there is less concern about the cumulative wear that higher power levels would cause over multiple flights.
Engine Spec Comparison
| Parameter | RS-25 (SSME) | Merlin 1D | Raptor 2 | RD-180 |
|---|---|---|---|---|
| Manufacturer | Aerojet Rocketdyne / L3Harris | SpaceX | SpaceX | NPO Energomash |
| Cycle | Staged combustion | Gas generator | Full-flow staged combustion | Staged combustion (O₂-rich) |
| Propellants | LOX / LH₂ | LOX / RP-1 | LOX / CH₄ | LOX / RP-1 |
| Thrust (sea level) | 1,860 kN | 845 kN | ~2,256 kN | 3,830 kN |
| Thrust (vacuum) | 2,279 kN | 981 kN | ~2,550 kN | 4,152 kN |
| Isp (sea level) | 366 s | 282 s | ~327 s | 311 s |
| Isp (vacuum) | 452 s | 311 s | ~350 s | 338 s |
| Chamber Pressure | 206 atm | 97 atm | ~300 atm | 257 atm |
| Engine Mass | 3,527 kg | 470 kg | ~1,600 kg | 5,480 kg |
| Thrust-to-Weight Ratio | ~66:1 | ~183:1 | ~163:1 | ~78:1 |
| Throttle Range | 67–111% | ~40–100% | ~40–100% | 47–100% |
| Nozzle Expansion Ratio | 77.5:1 | 16:1 | ~33:1 | 36.9:1 |
| Reusability (design) | 55 flights (Shuttle); 1 (SLS) | ~20+ flights | ~100+ flights (goal) | 1 flight |
| Approx. Unit Cost | $100M+ | ~$1M (est.) | ~$1M (est. at scale) | ~$10M |
Why the RS-25 Has the Highest Isp in Its Class
The RS-25’s 452-second vacuum Isp is the highest of any operational hydrogen-oxygen engine at its thrust class. Only the RL-10 (used on the Centaur upper stage) matches or slightly exceeds it at 465.5 seconds — but the RL-10 produces just 110 kN of thrust compared to the RS-25’s 2,279 kN. At its thrust level, the RS-25 stands alone.
Three factors drive that efficiency. First, hydrogen is the lightest element — its exhaust has the lowest molecular weight of any chemical propellant combination, which means higher exhaust velocity. Second, the staged combustion cycle wastes almost no propellant — virtually everything goes through the combustion chamber. Third, the 77.5:1 nozzle expansion ratio is extremely generous, extracting maximum velocity from the exhaust gases as they expand.
The trade-off for all that efficiency is the bulk of hydrogen itself. Because liquid hydrogen has very low density (only 70 kg/m³, about fourteen times less dense than liquid oxygen), hydrogen-fueled rockets need enormous tanks. The Shuttle’s external tank was 47 meters tall and 8.4 meters in diameter, and most of that volume was hydrogen. This is why methane (used in Raptor) is increasingly popular — it offers a good middle ground between hydrogen’s efficiency and kerosene’s density.
The RS-25 on SLS
The Space Launch System uses four RS-25 engines on its core stage, compared to the Shuttle’s three. All four fire at liftoff alongside two massive five-segment solid rocket boosters (evolved from the Shuttle’s four-segment SRBs). The core stage burns for roughly eight minutes, consuming about 730,000 kilograms of liquid hydrogen and liquid oxygen before engine cutoff.
On SLS, the RS-25 engines operate at 109% RPL for the initial Block 1 configuration and will operate at 111% RPL for later Block 2 missions. The engines are fixed at the base of the core stage and gimbal for steering — the same principle as on the Shuttle, though the flight control algorithms are entirely new.
Artemis I, the first SLS mission, launched on November 16, 2022. Its four RS-25 engines — all Shuttle veterans, with serial numbers E2045, E2056, E2058, and E2060 — performed flawlessly during the eight-minute core stage burn. Engine E2045 alone had previously flown on 12 Shuttle missions. After Artemis I, all four engines sank to the bottom of the Atlantic Ocean. Decades of spaceflight heritage, gone in eight minutes.
Legacy and Context
The RS-25 occupies a unique position in rocket engine history. It is not the most powerful engine — the RD-170 produces roughly eight times more thrust. It is not the most efficient — the RL-10 edges it out in specific impulse. It is not the cheapest, the lightest, or the most producible.
What the RS-25 is, uniquely, is the most operationally proven high-performance engine in history. No other engine has accumulated more test time, flown more missions, or demonstrated more reliability at pressures above 200 atmospheres. Its development story is a cautionary tale about the cost of extreme ambition, but its operational record is a testament to what that ambition can ultimately achieve.
Whether it makes sense to build new RS-25s at $100M+ per unit for an expendable rocket — when SpaceX is building Raptors for a fraction of that cost and reusing them — is a question that engineers, policymakers, and taxpayers will be debating for years. But the engine itself, as a piece of engineering, remains extraordinary. Fifty years after its first test firing, nothing else in Western rocketry quite matches it.
Frequently Asked Questions
Why did NASA choose the RS-25 for SLS instead of developing a new engine?
NASA had 16 flight-ready RS-25 engines left over from the Shuttle program, plus extensive tooling, documentation, and workforce expertise. Using existing engines reduced technical risk and allowed SLS development to proceed faster than designing a new engine from scratch. The decision was also driven by Congressional direction to use Shuttle-derived hardware.
Could SpaceX’s Raptor replace the RS-25 on SLS?
Theoretically, yes — Raptor produces comparable thrust and uses a more advanced full-flow staged combustion cycle. Practically, no. The SLS core stage is designed around hydrogen propellant, and Raptor burns methane. Switching fuels would require redesigning the entire core stage — tanks, plumbing, thermal systems, everything. It would be a new rocket, not a modified SLS.
How many RS-25 engines still exist from the Shuttle era?
NASA had 16 flight-ready engines after the Shuttle retired. Four were used on Artemis I (2022) and four on Artemis II. The remaining Shuttle-era engines will be used on Artemis III and IV. After that, newly manufactured RS-25E engines will power subsequent missions.
What does “full-flow staged combustion” mean, and is Raptor better than the RS-25 because of it?
In full-flow staged combustion (used by Raptor), both fuel and oxidizer are fully gasified in separate preburners before entering the main combustion chamber. The RS-25 uses a fuel-rich staged combustion cycle — only the hydrogen is fully gasified. Full-flow is theoretically more efficient and puts less stress on turbopumps, but Raptor’s lower Isp (350 vs. 452 seconds) reflects its methane propellant, not an inherently less capable cycle.
What was the actual failure rate of the RS-25 during Shuttle missions?
One in-flight shutdown out of 405 engine-flights (135 missions × 3 engines) — a 99.75% in-flight reliability rate. And that single shutdown was caused by a faulty sensor, not an actual engine malfunction. If you count only genuine engine failures, the in-flight reliability was 100%. Several pad aborts occurred due to engine-related issues during startup, but no catastrophic engine failures ever occurred during flight.
Is the RS-25 the most expensive rocket engine ever built?
On a per-unit basis, yes — at $100M+ for a new RS-25E, it is the most expensive production rocket engine in the world. The Soviet RD-170 and its derivatives were also extremely expensive but exact figures are not publicly available for comparison. For context, SpaceX reportedly builds Raptor engines for around $1 million each at scale, meaning you could build roughly 100 Raptors for the price of one RS-25.