The trade
Exhaust Velocity Sets Propellant; Thrust per Kilowatt Sets Time
The corridor fixes the Δv. Specific impulse (Isp) fixes how much propellant that Δv takes, through the rocket equation.
Chemical engines carry their energy in the propellant, so they deliver high thrust but cap out near 465 s. Electric engines take energy from a separate power source and reach thousands of seconds, but their thrust is limited by how many kilowatts the spacecraft carries per kilogram, its specific power.
Nuclear thermal sits between: high thrust and about twice the Isp of hydrogen chemical, ground-tested and never flown.
The classes
12 Ways to Change Velocity
Isp, thrust class and the hardware that has flown each one.
| Class | Propellant | Isp | Thrust | Flown on | Rated |
|---|---|---|---|---|---|
| Storable hypergolic | MMH/NTO | 310-320 s | High | Apollo lunar module; Chang'e landers; Tianwen-1 | Observed |
| Kerosene | LOX/RP-1 | 300-348 s | High | Falcon 9 Merlin; Long March 5, 6 and 7 YF-100 | Observed |
| Methane | LOX/CH4 | 330-380 s | High | Starship Raptor; LandSpace Zhuque-2, first methane rocket to orbit, 2023 | Observed |
| Hydrogen | LOX/LH2 | 420-465 s | High | Centaur RL10; Long March 5 YF-77 and YF-75D | Observed |
| Solar electric, Hall | Xenon, krypton, argon | 1,500-2,500 s | Very low | Starlink; NASA Psyche; Tiangong | Observed |
| Solar electric, gridded ion | Xenon | 3,000-4,000 s | Very low | NASA Dawn to Vesta and Ceres; JAXA Hayabusa2 | Observed |
| Nuclear thermal | Hydrogen | 850-900 s | High | NERVA and Kiwi ground tests, 1959-1972; never flown | Projected |
| Nuclear electric | Xenon, argon, lithium | 2,000-10,000 s | Low | Never flown; NASA Prometheus cancelled 2005 | Modeled |
| Solar sail | None | No propellant | Microscopic | JAXA IKAROS, 2010; LightSail 2, 2019 | Observed |
| Aerobraking | Planetary atmosphere | No propellant | - | Mars Global Surveyor, 1997; Mars Reconnaissance Orbiter, 2006 | Observed |
| Aerocapture | Planetary atmosphere | No propellant | - | Never flown; assumed at Titan and Neptune | Modeled |
| Mass driver | Electricity | No propellant | Launch only | Not built off Earth; see The Moon | Modeled |
Reactor programs, including SR-1 Freedom, DRACO and the China-Russia lunar reactor, are on Space Nuclear. Gear ratios at each Isp are on Gear Ratio.
Specific power
Transit Past Mars Waits on a Megawatt Reactor
Electric propulsion has flown at 20 kW; the outer corridors assume a reactor 50 to 500 times larger.
| Spacecraft | Power | Thruster | Rated |
|---|---|---|---|
| Dawn · NASA, 2007-2018 | 10 kW solar at 1 AU | NSTAR gridded ion, 3,100 s | Observed |
| Psyche · NASA, launched 2023 | About 20 kW solar at 1 AU | SPT-140 Hall thrusters | Observed |
| Tianhe core module · CMSA | Station bus | Four Hall thrusters for orbit keeping | Observed |
| Prometheus / JIMO · NASA, cancelled 2005 | 200 kWe reactor, design only | Gridded ion | Modeled |
| Crewed Mars NEP concepts | 1-10 MWe reactor | Hall or magnetoplasmadynamic | Modeled |
Sunlight at Jupiter is about 4% of what it is at Earth, so solar-electric power falls off past the Belt and outer-system NEP needs a reactor. Every transit time past Mars in the Corridor Index assumes megawatt-class NEP; until one flies, those times are placeholders.
China
Chinese Engines in Every Flown Chemical Class
Hypergolic landers, kerosene and hydrogen launchers, the first methane rocket to orbit, and Hall thrusters on Tiangong.
| Hardware | Class | Data point | Rated |
|---|---|---|---|
| YF-100 · AALPT | Kerosene | Long March 5, 6 and 7 boosters; YF-100K for Long March 10 | Observed |
| YF-77 · AALPT | Hydrogen | Long March 5 core stage | Observed |
| Chang'e lander engine · CAST | Storable hypergolic | 7,500 N variable thrust; flown on Chang'e-3 to Chang'e-6 | Observed |
| TQ-12 · LandSpace | Methane | Zhuque-2 reached orbit July 2023 | Observed |
| Tianhe Hall thrusters · CMSA | Solar electric | Orbit keeping on Tiangong since 2021 | Observed |
Contested
Deliberately Not Claimed
The load-bearing uncertainties, tagged with the confidence they actually earn.
Aerocapture has never flown. The Titan and Triton corridors assume it; a failed pass costs the vehicle, so the first flight carries more risk than the Δv saving shows.
Isp ranges hide vacuum against sea level. Raptor and Merlin each have a sea-level and a vacuum variant; corridor math uses vacuum values.
Specific power for NEP is unproven. No space reactor has produced more than a few kilowatts of electricity in flight; SNAP-10A in 1965 made about 0.5 kWe.
Record
Provenance and Data Sources
- Confidence
- Observed for flown, fielded or measured - Building for hardware and programs in development - Projected for announced plans - Modeled for original analysis
- Status date
- Q4 2026
- Hardware data
- Published engine Isp; NASA Dawn, Psyche and Prometheus; JAXA IKAROS and Hayabusa2; NERVA test program; SNAP-10A
- China
- AALPT YF-100 and YF-77; CAST Chang'e landers; LandSpace Zhuque-2; CMSA Tianhe
- Weakest link
- Megawatt-class NEP