As promised, and despite the Amos-6 accident, Elon Musk unveiled SpaceX’s plans to colonize Mars this week at the International Astronautical Congress. The presentation was rich on substance and technical detail (the full talk can be found here), and the community will certainly spend many hours poring over those details in the weeks and months to come. Today, we just want to examine a small, but vitally important part of the overall architecture – in-situ resource utilization, or ISRU for short.
Specifically, we will discuss using the Sabatier reaction (along with a few other processes) to create rocket fuel (methane and oxygen) on Mars. In one of his slides, he gives a diagram for this :
It’s perhaps worth going in depth as to exactly why SpaceX are so motivated to use Martian resources rather than bringing fuel with them from Earth, an approach which might superficially seem simpler and more reliable. After all, if the fuel plant goes wrong, the astronauts are all stranded there for at least as long at it takes to send a replacement, manufacture the rest of the fuel and then return to Earth, a problem which bringing fuel with you avoids completely.
The reason for using ISRU, however, is the sheer quantity of fuel you would have to bring with you for a return journey. Using this numbers given on side 36 (below) each ship weighs 150 tonnes empty and can carry 450 tonnes of cargo with it to Mars (assuming it has been refueled in orbit an appropriate number of times first). Using this information, the specific impulse of the Raptor engines and the orbital parameters of Earth and Mars, you can calculate how much fuel you would need to bring. We’ve done this calculation here, and to keep it manageable, we made several idealistic assumptions about the behavior of the rocket. Specifically, we considered the amount of fuel needed for a Hohmann transfer burn (the most efficient maneuver possible, in most situations) starting from rest at the surface of Mars (accounting for the Oberth effect) and leaving the spacecraft on a collision course with Earth. A few of the assumptions are thus:
No propellant boiloff.
An ideal transfer window.
A spacecraft with negligible crew and cargo mass, compared to the ship’s own mass.
No gravity losses, pressure losses or drag losses on Mars.
Perfect (no fuel needed) areocapture on Earth arrival.
All of these are of course unrealistic to an extent, but the calculation is about giving a back-of-the-envelope lower bound, not a precise figure.
The maths works out to give 930 tonnes of propellant (so, in practice, this means you need at least 930 tonnes, and probably significantly more than that), or at least two and a half cargo ships worth of propellant for each one ship you want to return. A few things arise from this.
Firstly, it’s clearly impossible to reuse most of your spaceships like this. If you have to send seven ships out to bring two home again, obviously most of your fleet will be single-use. One of the things Musk emphasized when he was talking about reducing the cost of a trip to Mars was reusing every element of the architecture, and that’s clearly not possible when you need to bring that much fuel.
Secondly, even if you planned to use ISRU in the longer term, and just brought fuel with you as a backup for the initial missions, you would still need huge increase in the number of ships needed for the same crew. Bringing fuel with you means not only four times as many ships, but four times as many refueling trips, four times as many first-stage reuses, and four times as many opportunities for something to go horribly wrong. All of this would add hugely to both the expense and time needed to start the colony, running completely against the principles on which the architecture was designed.
Finally, the cost associated with making 930 tonnes of propellant (186 of methane, the rest of oxygen) on Mars is not as great as it might seem. If you divide the energy needed to make this from carbon dioxide and water by the power produced by the spaceship’s solar panels (200 kW, as per 19:45 in the talk), you get a time to produce this much fuel of about 1.6 years. Obviously, the solar panels will be needed for other things as well, and they won’t produce power all of the time, but if a few more are kept in storage to be deployed after landing in addition to the ones used in flight, producing this same amount of propellant seems feasible at least from an energy perspective. Of course this doesn’t consider the details of how to extract large quantities of Martian ice or perform the Sabatier reaction on an industrial scale on the surface of another planet, but that is a separate discussion to be had.
Overall, it’s good to see that that even simplistic physical calculations strongly justify the use of ISRU. The idea of this is not new – the Mars Society have been talking about it for decades – but this is the first time that a team of engineers might finally put the idea into action. We certainly live in exciting times.
Your author has, in the past, appeared on the Mars Society’s podcast.