Lunar regolith is roughly 43% oxygen by mass, locked in metal oxides (SiO₂, FeO, Al₂O₃…). A molten-regolith electrolysis cell melts the dust and splits those oxides into free oxygen gas and reduced metal once enough current — and enough power — flows through it:
2 MOₓ + energy → 2 M + x O₂
rate(kg/s) = min(stockpile/dt, P_used / E_specific)
Power comes from the solar array, which only delivers
P = A·η·1361 W/m²·cos(θ_sun) — it swings from zero at "night" to a peak at local noon, so the reactor can only ever draw
min(requested, available). Asking for more than the array can supply, or running well under the cell's rated power, both drop conversion efficiency below its peak.
- Collection rate — how fast the rover scoops fresh regolith into the hopper; the reactor can't process faster than the stockpile is replenished.
- Reactor power — requested electrolysis power; actual draw is capped by live solar output.
- O₂ ≈ 20 MJ/kg (oxide electrolysis), H₂O ≈ 3 MJ/kg (ice sublimation — far cheaper, but only ~5% of shadowed regolith is ice), Metal ≈ 8 MJ/kg (reduced Fe/Si from oxides).
This is the trade-off every real ISRU plant design faces: bigger solar arrays and batteries cost mass and money, so mission planners size the reactor to the power budget, not the other way round.