Launching spacecraft with radioisotopic thermoelectric generators onboard runs the huge risk of contaminating the area around the launch site with radioactive material if the launch vehicle fails.
It is true that launch vehicles do not enjoy the highest reliability record. However, with this in mind, the radioactive material is stored accordingly. A multiple barrier approach is taken to prevent any uncontrolled release even in the harshest conditions an RTG might experience, including events such as the destruction of the Space Shuttle Challenger or uncontrolled reentry during misaimed flyby. The first step is to manufacture the fuel in a ceramic form, such as plutonium dioxide, which is heat and chemical resistant and does not aerosolise readily. This helps to ensure that the fuel will remain in unbreathable chunks in the event of dispersal. Since plutonium-238 is a pure alpha emitter, it can only be threatening when inhaled or ingested.
Figure 1- Blowout schematic of an RTG aboard the Cassini spacecraft. (Image courtesy NASA JPL)
The second step is to clad the fuel in iridium. Iridium is strong and heat tolerant so it makes a good protective layer. Outside of that, the clad fuel is placed inside a protective graphite shell, which is also heat resistant as well as resistant to impacts. Further layers of carbon fibre protection are then placed around that.
The resilience of this technology was proven during Apollo 13. All Apollo mission from Apollo 12 onwards carried RTGs, the SNAP-27, to power the Apollo Lunar Surface Experiments Package. The SNAP-27 was stowed aboard the lunar module for the astronauts to deploy when they landed. Of course, Apollo 13 never made it to the lunar surface after the oxygen tank explosion during the trans-lunar coast. The result was that the LM, Aquarius, was used to support the astronauts during the lunar flyby and return to Earth. In preparation for final reentry, the astronauts returned to the Command Module, Odyssey, and Aquarius was jettisoned. It burnt up in the atmosphere with the SNAP-27 still onboard. The protection kept the fuel intact and it is now resting safely contained at the bottom of the Peruvian Trench.
On spacecraft like Cassini, the safety in independent redundancy approach is also taken by splitting the fuel into dozens of separately contained modules. This significantly reduces the possibility of a large release of plutonium because it is incredibly unlikely that all the individual modules, all of which would be subject to slightly different conditions in the event of an accident, will fail.
Further to that, even if the highly unlikely chain of events leading to failure of all layers of protection were to occur, the wide dispersal would dilute the radiotoxicity to insignificant levels. Much of it would remain in the upper atmosphere away from people for a long period of time, during it would be able to decay further. Once settled, the insolubility of plutonium dioxide would prevent it from being absorbed into the food chain, meaning very little would get to the population.
The environmental impact statement for Cassini concluded that even in the worst case scenario events, which was inaadvertent steep reentry during swingby breaching protection and dispersing vaporised plutonium (it was not said that this dispersal would occur, merely that they calculated worst case scenario effects in the highly unlikely event that it did), the resultant health effects would amount to a 0.0005% increase in cancer incidence over a 50 year period, a calculation that is statistically undetectable. For any dispersal accidents on launch, no health effects were expected at all. These calculations used the conservative assumption of the LNT.
When Galileo was deorbited into the atmosphere of Jupiter, the plutonium could have reacted with the hydrogen in the atmosphere causing the planet to blow up.
Apparently this was a real argument. Astronomer Phil Plait deals with it.
Why use RTGs at all when solar panels or batteries will do?
Because they will not do. Batteries are prohibitively heavy for the missions in question and, given the quantity required and lack of protection, the risk from dispersal of the hazardous chemical material used to make the batteries far outweighs the risk from the dispersal of plutonium. Solar panels would also not work for missions into the outer solar system because of the falloff in solar intensity. Earth is 1 astronomical unit from the Sun. Jupiter is 5AU from the Sun. This means that the Sun is 25 times less intense. The increased size of the solar panels to compensate for this would be prohibitive. There is also an application for surface missions, such as Mars Exploration Rovers, where radioisotopic heating units, tiny radiological heat sources, are used to keep the spacecraft within operating temperatures during the frigid night on that planet.
There are proposals to use full fission reactors in space. This would be even more dangerous.
The handy thing about fission reactors over RTGs is that they can be turned on and off. This means that the actinide fuel is all that will be inside the core until the operators decide to bring the reactor online. The fuel of a fission reactor is likely to be an isotopic composition of uranium. This would be significantly less radioactive than the plutonium-238 of an RTG. Since the reactor would not be powered up until the spacecraft was safely on course for the outer solar system, the hazard from this material is significantly lower. Suitable protection will also of course be used, including manufacturing the fuel as a ceramic and using layers of carbon to isolate and protect the fuel from stress. Hazardous fission products will not be produced until the spacecraft is clear of Earth with no threat of return.
Nuclear electric and nuclear thermal rockets would spread fallout throughout space.
NERs use a fission reactor to provide electrical power to an ion drive. NTRs use a fission reactor to superheat an inert exhaust gas. Both have the advantage of providing a great deal of momentum from very little weight of fuel. However, none of them discharge radioactive material into space. It also bears noting in any event that space is a very large amount of real estate, which is already permeated by ionising radiation mostly from the Sun.
Some have suggesting using various types of nuclear propulsion for launch spacecraft from Earth, one of which actually used nuclear explosives.
It has been accepted for a long time that given the present state of technology, nuclear propulsion methods would only be used for on-orbit operations. The Orion concept popularised during the early fifties involved launching a large spacecraft on the blast of a series of nuclear explosions. It is not particularly viable, firstly because it throws non-proliferation out of the window, but also because it depended on the invention of a pure fusion explosive to eliminate fallout. Despite its creativity, no-one today, apart from a few idealists, is seriously entertaining the Orion Project.
Using nuclear power in space is a slippery slope to war in space.
This is conspiracy theorist paranoia and another shallow attempt to prejudice the public against nuclear power by forcibly associating it with something bad. Organisations like the Global Network Against War and Nuclear Power in Space may protest the launch of space based weapons, which violate the 1967 Outer Space Treaty in any event, but obstructing science because they cannot tell the difference between a scientific spacecraft and a military one, is not a noble sentiment.