For a satellite weight is even more critical than for an aircraft. Your answer implies that for a satellite the weight to energy ratio favors nuclear but for an aircraft it favors chemical power. You could probably expand a little bit more on why it isn't the same in both cases. Either way nice answer.

Kasper - for a satellite it's not weight to energy so much, as the lack of alternatives...

Satellites use isotope generators, that is atomic piles with energy conversion based on thermocouples, not full fledged reactors and turbines. And btw I believe flying/orbiting nuclear reactors are not permitted by international laws.

Yep - it's a very different model. Works well in that low power space.

"if you remain under the sea for months or even years at a time, diesel or anything that uses oxygen is not suitable" In particular, human crew use oxygen and wouldn't be suitable for staying under the sea for years at a time. :-)

Lol - that is an issue, but is manageable using scrubbers.

@mins Yeah, I know. The actual physical limit on sub crews staying underwater is food, as I recall.

Also of note: a major benefit of nuclear power is am aircraft that can stay aloft effectively indefinitely. This brings up the challenge of finding a crew that can do the same.

The reason RTGs can be so much lighter is that they don't employ any chain reaction, just decay of individual atoms. Therefore you can use arbitrarily small amounts of fuel (power will just go down linearly) and don't need any of the tricky control-rod / neutron-breaking / criticality-risk stuff. The flip side is that you need fuel that is by itself much more radioactive even without the chain-reaction boost, i.e. isotopes with short half-life, but these don't occur naturally, so they need to be bred in conventional reactors. Hence RTGs are not viable for large-scale power generation.

...In fact, even top-notch missions like Juno don't carry RTGs anymore because the Plutonium-238 is too scarce. We certainly don't want to waste it on plane propulsion!

we know that nuclear reactors can be miniaturized. world-nuclear.org/information-library/nuclear-fuel-cycle/…

The Soviet Union's RORSAT spy satellites were powered by a NaK cooled nuclear reactor. The USA's SNAP-10A was an experimental satellite nuclear reactor, it wasn't used in production. There are a few other US programs linked from the SNAP-10A article on Wikipedia.

@Timpanus you're ignoring maintenance, which is a significant limiting factor.

@David Richerby: But if you have ample power from some non-combustion source, like your nuclear reactor, you can get oxygen by electrolizing water.

@jamesqf I was trying to be cute about oxygen. As I said in a later comment, the actual problem is food.

@David Richerby: I don't know about that (seriously, not cute). I'd think you could stock food for a longer period than you could oxygen.

@jamesqf - as I mentioned, scrubbing and hydrolysis are sufficient for oxygen. David is correct about food being the problem