Since today is Monday, let’s talk about spacecraft power supplies.
Right now, spacecraft pretty much all use solar power, because it is highly reliable in space. This was not always the case; some early spacecraft just used batteries, and the Apollo missions used hydrogen fuel cells to produce electricity en route. And, of course the Voyager probes and a few other deep space missions have used radioisotope thermal generators (RTGs) for a long lasting nuclear-based power supply that will continue to work in areas of the solar system (and beyond) where the sun is too dim for solar.
Left to Right: Apollo Hydrogen Fuel Cell, Voyager Radioisotope Thermal Generator, ISS Solar Array
When you are picking a power supply, you are looking for a few things: Power density (how many watts can you get for each kilogram of power supply mass, measured in W/kg), longevity (how long before it runs out of fuel, or fails due to space radiation, micrometeoroid impacts, mechanical failure, etc. measured in weeks, months, or years) and mission suitability (i.e. don’t rely on solar panels out past Jupiter, because you’ll have less than 2% of the sunlight to work with).
The space community has long desired a type of power supply that would provide lots of power density, last a long time, and work anywhere in the solar system. The most obvious solution is a nuclear reactor, but it may not actually be the best option in the near term. Let’s talk about why.
First, note that a nuclear reactor is a bit different from an RTG; both use radioactive fuel (typically plutonium for RTGs and highly enriched uranium for reactors), but an RTG is typically sub-critical, which in super basic terms means there is enough radioactive material to get hot, but not enough to go boom. RTGs use the heat from radioactive decay to generate electricity via thermoelectric effect; this is a simple design with no moving parts, but it is pretty low efficiency, only about 2-7% of the heat ends up as electrical power, and when you account for the mass of the shielding, the thermocouples, etc… it ends up with a power density of 5 W/kg on the high end. Compare that to solar panels, some of which currently get 150 W/kg in LEO, and you’d think RTGs would never be used. But, fly those same solar panels out past Jupiter, and you only get 3 W/kg from solar; that’s why RGTs go on most deep space probes, and don’t go on most other things.
Nuclear reactors, though, are a whole different beast. First, they run super-critical, instead of subcritical. This means that there is enough radioactive material to go boom. Nuclear fission weapons work by triggering a chain reaction where decaying radioactive atoms emit neutrons that hit other atoms and split them apart, which releases more neutrons that go hit other atoms, and so on. The uncontrolled chain reaction releases a ton of energy very quickly and you get a mushroom cloud. Nuclear fission reactors operate using the same chain reaction, but they control it, mostly by putting space and mediating materials between the radioactive fuel pieces to slow and reflect some of the neutrons, so that there are always enough neutrons to keep the chain reaction going, but never enough to make it explode. If that sounds like you are balancing a power plant on a nuclear knife edge… well, yeah, that’s pretty much exactly how it works.
This is a pressurized water reactor. The rods sticking out are fuel rods, and the space between is filled with water to mediate the reaction. Fortunately, water is one of the best mediating materials, and the process works such that the density of the water changes the rate of the reaction: more dense water, faster reaction, less dense water, slower reaction. This is perfect for the knife-edge balancing act because as the reaction accelerates, the water gets hotter, which will reduce its density, and naturally slow the reaction down. So, it’s not quite as scary as it otherwise would be; there’s a bit of margin for error in operation.
Nuclear reactors, because they are super-critical instead of sub-critical, can generate much more power than RTGs from the same mass of fuel, and they can operate on electrical generation processes other than thermoelectric effect, offering a much higher power density; some old NASA designs were estimating 50 W/kg; some newer reactors may be able to hit 500 W/kg, though that is unproven so far. Also, because they don’t rely on the sun, nuclear reactors will offer the same power no matter where you are in the solar system. And, while they do need to be fueled (unlike solar arrays), nuclear reactors consume their fuel very slowly. Some designs on earth need refueling only once every 30 years or so; the target in space will probably be 10-15 years without refueling before power output drops below 80% of the output at launch.
So, what’s not to love? Well… two things. First, liquid mediated reactors (like the one with water, above) don’t work so well in space because they usually rely on gravity-fed flow of the cooling water, and in space the reactor will be in free fall, so gravity won’t help much. Similarly, the mediation strategy relies on density changes in the mediating liquid to produce the margin for error that makes many designs safe to operate, but density distributions in microgravity don’t work the same way that they do under gravity (no up, so the less dense stuff doesn’t float). This may not be a problem, and there are workarounds anyway that we use for different designs on Earth, but it probably does mean a unique reactor design, and designs for nuclear reactors are costly and very time consuming because nobody wants them to go boom.
Which takes us to the second challenge: Geopolitics. Nuclear reactors that run efficiently and have high power output for their mass rely on highly enriched fuels. Highly enriched fuels can also make nuclear bombs. As a result it is a very sensitive political issue when anybody says “We are going to enrich a bunch of uranium, put it on a spaceship, and fly it into an orbit that will pass right over many different countries about 16 times per day.” How do those countries know you aren’t building a weapon? What if the spaceship crashes on their country? What if the reactor fails and melts down and spews toxic nuclear fallout all over the upper atmosphere of the planet? These considerations are real, and cause quite a lot of international political opposition to any plan to put nuclear reactors in space.
Meanwhile, space solar array developers are not sitting still either. Some arrays already offer 150 W/kg, but I’ve talked to people who are actively working on 1000+ W/kg solar cells, based on LEO performance. If they succeed, those solar panels will outperform nuclear even as far out as the moons of Jupiter until those new reactor designs come online and start exceeding 50 W/kg. That’s enough of the solar system to include most of the most interesting destinations, including Mars and the entire asteroid belt. And, no fuel, no enrichment, no geopolitical opposition, likely a much faster and cheaper development timeline.
This is why I say that in the near term nuclear looks obvious, but really may not be the best option. In the long term, as those high-power, low mass reactor designs are completed, this picture may change, and as we master fusion power it almost certainly will change. But, for the next twenty years, my money is on solar power in space, even for large and distant missions like resupplying Mars colonies or mining asteroids. On the ground on the Moon or on Mars, nuclear will definitely be the best option most of the time (on Earth too, as SMR companies start to actually build things). But in space, solar will win.
Okay, that’s great, but why do we care so much about power density? Well, the things we can do in space are directly limited by how much delta-v we can carry. And, how much delta-v we carry is directly dependent on how efficiently we can operate our propulsion systems (ISP is the term for fuel efficiency on spaceships). And, the best propulsion systems for fuel efficiency are ion thrusters, which consume more electricity the more efficiently they consume fuel. More electricity means a larger, heavier power supply. So, there’s a break even point, where increasing your fuel efficiency adds more mass to your power supply than it removes from your fuel supply for the same performance, and being more fuel efficient is actually worse. But, the lighter your power supply is per watt, the higher that break-even efficiency is, and the more stuff we can do in space.
And that’s the goal: Do more stuff in space, by inventing more technology that makes it economically beneficial.