Captain America: "How come [this shield is] not standard issue?"
Howard Stark: "That's the rarest metal on Earth. What you're holding there, that's all we've got."
RTG materials are special
In an RTG, you need some very special traits.
- You need a half-life that is long enough to be useful. The power output will taper off, but you want a long enough half life that it tapers down only a little in the duration of the mission. Iodine-131 is plentiful, but it's gone in a month or two.
- The half-life can't be too long. Decay events are what make the heat, and a long-half-life element like Uranium-235 will not make decay events often enough to make any useful heat.
- The element must decay with the right kind of radiation - a kind trivially blocked. That is for human-safety reasons: a gamma emitter is Right Out. Alpha decay is ideal, as it's stopped by a few centimeters of ordinary air, or the skin - but must not be ingested.
- The material must be anti-proliferation. Meaning that if it were captured by terrorists (or fell out of the sky right into their laps during a failed launch), they must not be able to make an atomic bomb. It cannot be fissile and cannot be a feedstock easily made fissile.
- The material must be possible to separate from its origin and byproduct materials. For instance elements can be separated chemically, but adjacent isotopes of a heavy element are inseparable at an industrial scale. Enrichment of uranium is only possible because they are 3 apart: U-235 and U-238, and still requires state-level commitment (hence America's never-ending sparring match with Iran to prevent them from developing that ability).
- The process that makes it can't quickly over-make it. For instance if you get the material by neutron bombardment, but that also converts the target material into a useless, inseparable material, that doesn't work at all.
Plutonium 238 ticks all the boxes... but...
Plutonium-238 has an 87 year half-life (just right) and is an alpha emitter. As an alpha emitter, it would make a poor 'dirty bomb' because it could only threaten humans as dust they breathed during the initial moments of the attack.
So let's make a bomb? Our buddy has a fission reactor, let's bombard it with neutrons for 30 days to make fissile Pu-239. *Great, we've done that, and now we have a slug of mostly Pu-238 with a little Pu-239. Let's do isotopic separation of -- Oh, snap. Even the United States can't do that.
Okay then, load it back in the reactor for a year and convert the kaboodle to Pu-239. That's working, except the Pu239 is also absorbing neutrons and becoming Pu240 and Pu241... which are too radioactive and will cause a bomb to misfire. So the fact that it already is plutonium actually helps the nonproliferation issue.
... but it's really hard to make.
As far as how to make Pu238, there are several processes to make it from spent reactor fuel. However, these are special processes, not just a "load it into a common power reactor with the normal fuel rods" situation*. Indeed, both the US and the Russians had shut down their Pu238 production, because their lines were manufacturing it using waste materials from bomb core production, which had ceased under the various START talks. Several countries had to start their own Pu238 manufacture operations, specifically for making RTGs.
Suffice it to say, Pu238 is rare, hard to make, and expensive.
* Canada's efforts to make it in a reactor is a special exception, because their reactors are. Unlike conventional BWR and PWR/VVER designs, the CANDU is specifically designed to allow fuel rod changes while the reactor is underway. This is a trait it shares with the Soviet RBMK and weapons reactors such as the Hanford "B", because hey, they were all after the same thing: the ability to achieve criticality on natural uranium, and remove rods in 30-90 days - perfect for breeding weapons grade plutonium. Canada wanted to have the ability in its back pocket, and Russia wanted to be able to scale up their existing ability quickly.