This post is part of the series on space topics.
This post is not the last word on this topic. The usual caveats apply. I’m curious if you have strong opinions on different fuel mixes.
A growing Mars base has a prodigious need for power.
I’ve previously written two posts on powering the lunar base. The first explores solar power towers and thermal storage. The second reveals that beaming power to the lunar base region with microwaves from Earth is by far the cheapest, most flexible option – for operations on the Earth facing side of the Moon.
This, incidentally, is the only business case I know of for using microwaves to move power at scale between Earth and space. In particular, beaming power the other way as envisioned with space-based solar power doesn’t make financial sense.
The natural business case for using microwaves between space and Earth is via low Earth orbit comsats, as we’ve seen with Starlink, which I wrote about in 2019 and in all likelihood used to post this piece!
I’ve previously written a couple of books on Mars-related topics. The first concerns primarily the transport problem, addressed by Starship. The second focuses on Mars industrialization, and has a section on electrical power as well as an early exploration of an off-world synthetic hydrocarbon supply chain which ultimately inspired my start up company Terraform Industries (we’re hiring) and will be further examined in this post. Robert Zubrin has also written a number of books examining the Mars settlement question.
AI-generated conceptual image of Lunar power infrastructure – a phased microwave array on Earth.
While the Earth-facing side of the Moon can obtain practically infinite quantities of very cheap power if we beam it up from Earth, which is greatly preferable to attempting to engineer some sort of solar or thermal system which can cope with the Moon’s 28-day day-night cycle, this approach won’t work on Mars. It’s much too far away and its ~24 hour day means that half the time, any given base will face away from Earth anyway. Unlike the Moon, which is most efficiently served logistically as a kind of annex of Earth, Mars will have to localize power production and ultimately a good fraction of its material supply chain.
Diagram showing product consumption by weight and cost, the better to understand Mars manufacturing differential advantage.
Let’s begin by asking: What is power for?
While electricity on Earth is essentially post-scarcity (it comes out of the wall on demand) it is easy for us to take it for granted. In my house we consume an average of 2 kW and pay ~$100 month for the privilege, less than it costs to feed the family most days. But 2 kW – the amount of power you can get from just two outlets in the USA (or one in the civilized world) is equivalent to the food needed to feed 15 or 20 active grown adults. In other words, even in sunny California, we consume far more energy in the form of electricity (and gasoline) than we do in the form of food.
Yet most conventional representations of living on Mars are preoccupied with hard scrabble subsistence agriculture. This is fair – growing food on Mars will be extremely difficult relative to Earth – but it’s also extremely hard to make power there, and we’re going to need a lot more of it – much more than 2 kW per household. The per capita power consumption of an aircraft carrier, for example, is closer to 40 kW, and that’s not including airplane fuel or the manufacturing required to build and maintain its systems, which occurs on shore. Meanwhile, the surface of Mars is significantly more hostile than the surface of the ocean. If we want our nascent Mars base and its inhabitants to succeed, they’re going to have to achieve insanely high levels of productivity. Electrical and energy over-abundance are table stakes to ensure none of our vital human capital is being wasted dealing with shortages or rationing.
Mark Watney attempting to close his industrial production on the agriculture level.
In the US, per capita energy consumption is about 10 kW, with about 1.5 kW being used as electricity and 6.5 kW ending up as waste heat. If we want to close the entire Mars industrial stack with just a million people, on the surface of Mars, it is safe to assume at least a 10x increase in per capita productivity and energy consumption, so let’s baseline 100 kW per person.
This might still be too low as it neglects the fact that about 65% of US energy is derived from fossil fuel sources which don’t exist on Mars and, if synthesized, would carry a 3x energy consumption overhead. This is important since the major use for hydrocarbon chemical fuels on Mars is refueling Starships to fly people back to Earth, and they’re extremely thirsty! For example, if every Starship is capable of flying 100 people back to Earth, and every person flies back to Earth after a 2 year stay on the surface, then just fueling the ships will consume roughly 10 kW per capita. If mass margins limit return flights to 10 people per Starship, the per capita fuel energy cost rises to 100 kW!
Here, I’ll provide a rough sketch of the Mars base energy process to help develop intuitions, before adding more rigor.
The Mars base needs extremely large quantities of the cheapest possible energy. This could be nuclear, solar, batteries, fusion, whatever. The cost is a mix of development costs on Earth, transport costs to Mars (let’s baseline $1000/kg, assuming the Starship is expended on arrival), deployment and operating costs, depreciation and amortization.
For example, if we need a gigawatt of energy (10,000 people at 100 kW each) and space reactors weigh 150 T/MW, we’ll need to salami 150,000 T of reactors between 1500 Starship flights, costing $150b just to deliver the reactor. This doesn’t include development and manufacturing costs on Earth, or installation labor costs on Mars.
A 1 GW nuclear reactor on Mars would be much bigger than this.
Solar isn’t much better. If we baseline off Starlink but spring for higher efficiency solar panels, we can get down to about 7 T/MW. But fixed solar is only 25% capacity factor, there are seasonal variations, and Mars is 1.5x further from the sun. So per MW of continuous power, it’s closer to 80 T/MW – requiring 800 Starship flights at $80b for the power plant, and that’s not including 24 GWh of batteries required for multi-day back up. Baselining 300 kWh/T (which is cutting edge), we’d need another 80,000 T (800 Starships, $80b shipping cost) to transport the batteries, which is no better than the best case scenario for nuclear.
Solar as far as the eye can see.
Much of the power will be used for heat, which is both much cheaper to store or to extract from a nuclear reactor, so it’s possible we can do slightly better than this rough analysis, but also possible that it could be worse.
In my opinion, the major advantage of solar is that you don’t need to get regulatory approval to launch 1500 Starships containing enriched uranium. It’s also possible that at some point we will be able to localize production of some of the heavier and easier-to-manufacture components of the power system. For reference, the 1 GW power system referenced above would be adequate to serve 10,000 Martians consuming 100 kW each. This is large compared to an Antarctic base but tiny compared to the staff of even a moderately sized factory.
Power at the base is relatively easy to transmit with wires, and it will be used for life support (ECLSS), heating, cooling, computing, manufacturing, etc. For transportation within the relatively compact base, rechargeable electric vehicles are much more sensible than synthetic fuel-powered combustion.
We can enclose vast tracts of land in transparent hyperbolic isotensoid galleries and vaults, and built factories and houses within.
A decent chunk of the generated power must also be used to synthesize methane and oxygen to refuel the Starships to transport people and cargo back to Earth. I think it’s fair to assume that most Starships will carry cargo one way to Mars, but some fraction of them (1%?) will ultimately return to Earth.
As presently configured, each Starship contains 240 T of liquid CH4 (natural gas) and 960 T of liquid O2 (oxygen). CH4’s energy density is 55.5 MJ/kg (that’s a lot!) and the process of obtaining water and CO2 from the Martian environment, synthesizing, liquifying and storing this fuel adds considerable parasitic energy demand. Let’s baseline it at 250 MJ/kg-CH4 generated on Mars, including its share of oxygen. The 240 T of CH4 we need therefore costs 60 TJ, or 17 GWh of energy. That’s why an explosion on the launch pad would be so spectacular – it’s an insanely enormous quantity of stored energy!
Methane, the major component of natural gas, is also a fabulously versatile and useful chemical precursor for manufacturing, and roughly half of the natural gas in the US is used for purposes other than power generation. Broadly speaking, anything containing carbon needs a source of carbon to be made, including plastics, paints, epoxies, fertilizers, medicines, composites, pigments, etc etc. The Sabatier units that produce fuel for launching Starships on Mars will also be the upstream suppliers of chemical fuel for the entirety of Mars’ local industry.
Methane as a fuel has its advantages. It’s relatively easy to make, highly energetic, and non toxic. On the other hand, it has handling difficulties. It’s a gas under both human conditions and Mars surface conditions, so must either be pressurized or liquified at cryogenic temperatures for storage. Being volatile, any leakage into oxygen-containing areas, including any human habitation areas, presents an explosion hazard. It also burns extremely hot with pure oxygen, which is a feature in a rocket and bug for less energy intensive applications.
Are there other fuels which are, on balance, easier to make, handle, and use?
What do we need them for? By far, most of the fuels being made on Mars are for Starships – it’s just far, far more energetically demanding to fly to another planet than to do nearly anything else!
In addition to Starship rocket fuel and fixed industrial processes or chemical synthesis, which could equally well begin with CO2, carbon monoxide (CO), or methanol (CH3OH), what other uses do we have for chemical fuels?
Energy storage
Could we use, eg, stored Starship methane fuel to power the base overnight, and thereby avoid importing 80,000 T of batteries? The short answer is no. The end-to-end process efficiency for making methane is about 20%. Converting that fuel back into energy, even with the most efficient combined cycle gas turbines, is about 55% efficient. In all, a 10-1 energy hit for a single cycle. On top of this, the gas synthesis plant needs to operate 24 hours per day (or be 4x the size) which needs its own energy storage. For fixed energy storage, batteries or thermal storage (for heat applications) have much lower cost due to their simpler processes, lower weight, and markedly higher end-to-end efficiency.
A big lake of liquid water is a useful thermal sink for storing heat.
Mobility
We could run rovers, mining trucks, and other long range wheeled vehicles with batteries, but their energy density is far worse than chemical fuels. For applications constrained by weight or range, fuels have a much higher energy density.
In order of decreasing performance, these applications are rockets, aircraft, rovers and other vehicles, and space suits.
We can build truly gigantic trucks for Mars’ low gravity.
Above, we estimated battery specific energy at 300 Wh/kg, or 1 MJ/kg. While methane’s specific energy is 55 MJ/kg, mobile applications would also have to carry oxygen to burn the fuel, reducing the specific energy to just 10 MJ/kg. It turns out that for almost all fuel/oxidizer combinations, ~10 MJ/kg is achievable. That means that a fuel/ox rover can travel 10x further for the same energy mass fraction, accessing 100x the area. A transport network that uses depots to stage fuel, or remote fuel plants, can achieve even greater levels of mobility.
The wings could be even longer.
While rockets and aircraft on Mars can readily use cryogenic methane and oxygen as fuels, a rugged backcountry exploration outfit may prefer, where possible, to use fuels that are easier to handle and to move from vehicle to vehicle. Instead of being gasses, perhaps we can find fuels that are liquid, pourable, and pumpable, but won’t freeze in ambient Mars conditions, or boil inside a hab creating an explosive atmosphere? Could we also use it in tiny engines or fuel cells to power space suit life support?
Remote solar powered self-service fueling station on Mars.
Let’s go through the list
The simplest chemical fuel available on Mars is CO/O2. Carbon monoxide can be generated from ambient CO2 via a variety of pathways. On the Perseverance rover, the MOXIE experiment split CO2 into CO and O2 using zirconia solid oxide electrolysis. Burned with oxygen, it has about half the energy density as the hydrocarbons. Like oxygen, it’s liquid at cryogenic temperatures, and unlike oxygen, it can suffocate you. But it’s quite easy to make compared to more exotic fuels. Solid oxide reduction of CO2 to CO is perhaps 20% efficient – much lower than batteries – but the product has about 5x higher energy density (5 MJ/kg compared to 1 MJ/kg) than batteries. So for applications that require higher energy density storage, there’s an equilibrium which favors consuming ~5x more energy at the fuel production plant in exchange for 5x more energy storage on our vehicle.
Let’s examine the alkanes. Basic hydrocarbon chains. As they get longer, their boiling and melting points increase. Unfortunately, these chemicals either boil at hab temperatures (300 K) on the Mars surface, so an open container couldn’t be moved through an airlock without sealing it in a strong container or first cooling it down. Or they freeze at Martian surface temperatures. They’re also quite hard to make (except for methane), requiring the low yield Fischer-Tropsch process to synthesize. At longer lengths, they can also be toxic, carcinogenic, etc.
The exception to this is paraffin wax, the often unwanted byproduct of FT synthesis, which is a stable, non-toxic solid – and the main constituent of crayons. I’ve never heard of a paraffin fuel cell but it can be melted and burned in an engine, turbine, or thermal conversion cell. It could be stored as rods, tapes, or spools and deployed similar to an FDM 3D printer.
To get better thermal properties, we can introduce an OH group to the alkanes, creating alcohols. The OH group is polar, making them stickier and increasing their boiling and melting points.
Methanol is relatively easy to produce via high pressure high temperature catalysis, liquid at room temperature albeit very volatile, and about 10x as toxic as ethanol, the active ingredient in alcohol drinks. Still, in the class of pourable fuels for use on Mars, it’s an attractive choice.
Ethanol is also volatile and intoxicating, while propanol and butanol have higher boiling points making them less volatile and less likely to instantly boil if a warm container is depressurized outdoors. They’re also similarly toxic to methanol.
We can put another OH group on for even more stickiness, giving us the diols or glycols, depending on nomenclature preferences. Used on Earth as solvents and antifreeze, ethane-diol is also quite toxic. Propane-diol and butane-diol, which I discussed briefly in my first book, are technically liquid in all temperature and pressure ranges we’re likely to encounter on Mars, albeit with honey-like viscosity at the lower temperatures. They’re also relatively difficult to synthesize.
Having exhausted the obvious carbon-based fuels, we’ll now briefly examine some more exotic possibilities.
Silane (SiH4) is an incredibly toxic gas used in the chip industry, with the useful feature that it can burn in CO2, thus not requiring additional oxidizer. Like magnesium, which can also burn in CO2, some of the combustion products are solids, which foul up most sorts of engines. Also, any chemical reaction which uses CO2 as the oxidizer (including the Sabatier reaction, which consumes hydrogen) does not exactly break records for enthalpy. CO2 is not an entropically favorable oxygen carrier molecule!
CO2 itself can be liquified and/or compressed, albeit with more moving parts and much lower energy density than lithium ion batteries. Liquid CO2, which can be stored stably under pressure at ambient temperatures on Mars, actually requires energy to gassify, so could be used as a storable refrigerant, albeit not a very good one.
There’s always nitrogen-based fuels. Hydrazine is relatively easy to synthesize compared to larger molecules, it’s liquid at the right conditions, it can be used as a monoprop or burned with oxygen or other oxidizers, and it’s moderately toxic and corrosive. Hydrazine and nitrogen chemistry will almost certainly be performed on Mars but I am doubtful it will become the frontier fuel of choice.
That leaves the storable solid fuels, which contain their own oxidizer. Ammonium nitrate and ammonium perchlorate are good Google search terms to end up on some list, with the latter used in modern solid rockets. They’re relatively hard to make, mildly toxic, and famously safe (sarcasm). It is a fun exercise to imagine the hypothetical series of historical accidents necessary for ammonium perchlorate to become the Mars chemical fuel of choice.
So where does that leave us? Some combination of carbon monoxide, methane, methanol and paraffin, depending on preference, with a healthy side of high performance batteries.
Obligatory Sankey Diagram
Putting this together, here’s a potential energy flow diagram for a Mars city with 10,000 people, flying 100 Starships back to Earth every launch window, driving huge rovers long distances for mining operations, standing up enormous manufacturing operations, and constantly expanding their city. Per capita electricity production is ~200 kW. Almost half the total energy supply is consumed by direct chemical synthesis of fuels, which I’ve broken out separate to other industrial activity.
Does this sort of thing sound interesting to you in more than a casual blog reading kind of way?
Making valuable fuels for millions of grateful customers is not just some futuristic Mars fever dream, it’s happening right now. Three years ago, I founded Terraform Industries to bring the Mars synthetic hydrocarbon supply chain back home to Earth, to create unconditional energy abundance, to solve the climate carbon problem, to do something with our crazy surplus of solar panels, and to build factories in real life.
We’re hiring!
We like to work with ambitious, technically brilliant people. We’re always on the lookout for talent across the technical hardware spectrum, in particular mechanical engineers. Open job listings. Send a one pager to hiring@terraformindustries.com.
