Refueling and propellant depots
DraftEvery mission flown has departed Earth with all the propellant it will ever have. The rocket equation then applies to the whole stack at once, and its exponential penalty falls on the entire mission rather than on each leg separately.
Why single-launch architecture is expensive
From The rocket equation, propellant mass grows exponentially with delta-v. A mission that must carry departure, insertion, descent, ascent and return propellant from the surface of Earth pays that exponential on the sum.
Refuelling breaks the sum into pieces. A vehicle that departs LEO with a full tank, having been filled by tankers launched separately, only needs to carry the delta-v for its next leg. The mass that must be launched in one piece falls dramatically, and it becomes possible to fly missions no single launch vehicle could support.
This is not a marginal improvement. It is the difference between a lunar architecture that fits on one heavy-lift vehicle and one that does not, which is why Starship HLS depends on it. See Artemis and human lunar return.
Architectures
| Approach | Description | Trade |
|---|---|---|
| Direct transfer | Tanker meets the target vehicle and transfers immediately | No storage problem; requires precise launch timing |
| Depot | A persistent storage facility that tankers fill and vehicles draw from | Decouples tanker and mission schedules; must store propellant for months |
| Aggregation | Multiple tankers fill one vehicle in sequence with no separate depot | Simpler than a depot; the receiving vehicle must tolerate a long fill campaign |
The distinction matters because storage duration drives everything. Direct transfer avoids long-term storage entirely at the cost of requiring the tanker and the mission vehicle to be ready simultaneously. A depot accepts the storage problem in exchange for schedule flexibility.
Location follows the same logic.
| Location | Advantage | Disadvantage |
|---|---|---|
| Low Earth orbit | Cheapest to reach from the ground | Deepest in the gravity well; departing costs the most |
| Earth-Moon or | Cheap to reach from anywhere in cislunar space | Expensive to supply from Earth |
| Lunar orbit, NRHO | Close to the surface; compatible with Gateway | Small market unless surface activity is substantial |
| Lunar surface | Propellant made locally from polar ice | Requires a mature ISRU industry first |
The boil-off problem
Cryogenic propellants are the ones worth storing, because they have the best performance, and they are the hardest to store.
| Propellant | Boiling point | Storage difficulty |
|---|---|---|
| Liquid hydrogen | 20 K | Extreme; also permeates and embrittles |
| Liquid methane | 112 K | Moderate; close enough to oxygen to share insulation |
| Liquid oxygen | 90 K | Moderate |
| Storable hypergolics | Ambient | None, but poor performance and high toxicity |
Any heat leaking into a cryogenic tank boils propellant, which must then be vented or the tank will burst. Loss rates of a few per cent per month are typical for passive systems, which means a depot filled over six months has lost a substantial fraction before it is used.
Mitigations run in a sequence of increasing cost:
- Multilayer insulation and sun shields
- Passive, cheap, and insufficient alone for long durations.
- Vapour cooling
- Route the boil-off gas through the tank shielding on its way to being vented, using its cooling capacity before losing it.
- Active cryocoolers, zero boil-off
- Refrigerate the tank to reject the heat leak entirely. Costs power and mass, and the coolers must run for the depot life.
- Choose methane over hydrogen
- A 112 K boiling point is far more manageable than 20 K, and methane and oxygen can share a thermal design. This is a significant part of why new vehicles are methalox.
Zero boil-off is the enabling technology for long-duration depots, and it has been demonstrated on the ground and at small scale in flight, not at depot scale.
Transfer in microgravity
Moving propellant between two vehicles in orbit is harder than it sounds, for the reason set out in Onboard propulsion: there is no bottom to the tank.
| Problem | Approach |
|---|---|
| Liquid position is undefined | Settle with a small acceleration burn, or use capillary propellant management devices |
| Gas ingestion into the transfer line | Liquid acquisition devices at the outlet; sensors to detect vapour |
| Tank chilldown | The receiving tank is warm and will flash-boil the first propellant in; it must be chilled first, which consumes propellant |
| Pressure control | Both tanks must be managed as propellant leaves one and enters the other |
| Gauging | Knowing how much has actually transferred, to a few per cent at best |
Settling with a low-thrust burn is the conceptually simplest solution and is what most concepts assume. It requires the vehicles to be docked and accelerated together for the duration of the transfer, consuming propellant to move propellant.
Tank chilldown is the underappreciated cost. Filling a warm tank with a cryogen boils a substantial quantity, and that quantity is lost. For a multi-tanker campaign this happens repeatedly.
What depots unlock
The reusable tug is the item with the clearest near-term economics. A tug that moves payloads from LEO to GEO or to lunar orbit and then returns to be refuelled replaces an expended upper stage on every flight. That only works if refuelling exists, and it changes the cost structure of every high-energy mission.
The ISRU branch is the long-term case. Propellant made from lunar polar ice or Martian atmosphere never has to be lifted out of Earth’s gravity well, which removes the dominant cost entirely. See Lunar surface systems and Mars habitats.
Status
Orbital cryogenic transfer has not been demonstrated at operational scale. It is on the critical path for Artemis, which makes it the most consequential unproven capability in current human spaceflight, and several demonstration missions are in development.
The technical steps are individually plausible and collectively unproven. That combination is what makes schedule estimates for anything depending on it unreliable.
References
- Chato, D. J. “Cryogenic Fluid Transfer for Exploration”, Cryogenics, 48:206, 2008.
- Notardonato, W. U. et al. “Zero boil-off methods for large-scale liquid hydrogen tanks using integrated refrigeration and storage”, IOP Conference Series: Materials Science and Engineering, 278:012012, 2017.
- Chandler, F. et al. “Propellant Depots for Earth Orbit and Lunar Exploration”, AIAA 2007-6081.
- Salvail, P. et al. NASA studies on in-space cryogenic propellant transfer demonstration missions.
- Zegler, F. and Kutter, B. “Evolving to a Depot-Based Space Transportation Architecture”, AIAA 2010-8638.