RTGs, or more fully Radioisotope Thermoelectric Generators, have been a key feature for long-distance interplanetary missions that require constant power beyond the practical limits of solar power and the lifespan of batteries and chemical power supplies.
The radioactive material contained within the RTG gives a constant supply of heat, part of which is used to heat the spacecraft and keep the electronic components functioning when lack of insolation can cause temperatures to fall well below 200K (even as low as 50K for the surface of Pluto). The rest uses the Seebeck effect to produce the relatively low but constant wattages needed to power those components.
The future of RTGs in space is constrained by political problems. Plutonium is commonly used by NASA, but supplies are constrained in their availability and the infrastructure required is expensive. Alternatives exist in the form of Thorium, Uranium, and alternative designs that require less Plutonium, but each requires large investment in new production facilities or research in improving the efficiency of the electricity generation process.
Huge advances in solar technology have covered much of the growing gap, but there are some applications for which solar is just not appropriate. Deep space missions beyond the orbit of Saturn where sunlight is sparse, lander missions where the body may be out of sunlight for large periods of time, and long-duration missions in areas where exposure to dust or debris is a risk.
Hopefully the political issues surrounding the use of nuclear material in space will diminish over time, particularly as technology allows the use of more abundant material, and until then compromises are going to have to be made. However a long term strategy for powering missions without solar is going to be essential not just for long-range scientific missions but also as a backup for human-led exploration - an area where putting eggs in a single basket is simply not acceptable.
The National Space Policy of the United States mandates that NASA can use nuclear power sources if and only if they “significantly enhance” any given space mission. “The rules say that we can only use a nuclear power system if other options can’t get the job done,” says Curt Niebur, NASA’s Europa program scientist. “And for a multiple flyby mission of Europa a solar power system can get the job done. So we chose solar.” A careful reading reveals a troubling flaw in this logic, because no consensus exists on how far solar power can be pushed. Lacking clearly defined boundaries beyond which solar becomes unreasonable, and relying in part on relatively unproven technology, mission planners could soon find themselves on a slippery slope, sliding into uncertainties that inadvertently preclude the use of nuclear power for practically all space missions.