Critics of nuclear power argue that all reactors are inherently dangerous, and point to nuclear accidents such as Three Mile Island, Chernobyl and Fukushima as evidence of the danger. The primary worry about nuclear power stems from the release of radioactive fission products, and other radioactive materials that are produced inside reactor cores. Among those materials Plutonium which is produced by a nuclear process which occurs when Uranium-235 and U-238 fail to fission following the absorption of neutrons inside reactor cores. Tritium a hydrogen isotope is also viewed with concern, although in practice tritium is viewed as so safe that it is used to illuminate the hands and dials of some wrist watches. The Nuclear Regulatory Commission (NRC) explains,

Tritium (H-3) is a weakly radioactive isotope of the element hydrogen that occurs both naturally and during the operation of nuclear power plants. Tritium has a half-life of 12.3 years and emits a weak beta particle. The most common form of tritium is in water, since tritium and normal hydrogen react with oxygen in the same way to form water. Tritium replaces one of the stable hydrogens in the water molecule, H2O, and creates tritiated water, which is colorless and odorless.

Tritium can be found in self-luminescent devices, such as exit signs in buildings, aircraft dials, gauges, luminous paints, and wristwatches. It is also used in life science research and in studies investigating the safety of potential new drugs.

In fact, if there were any significant danger from tritium, the NRC would outlaw using it in wrist watches. Just how dangerous is tritium? Recently a very small tritium leak from the Vermont Yankee power and was the subject of a big todo in Vermont. Nuclear critics insisted that the tritium represented a huge danger to the people of Vermont. The blog Minor Heresies suggested,

Tritium causes all the usual radiological effects: cancer, genetic defects, cell death, birth defects, and loss of fertility. . . .
My conclusion from all this is that the present tritium leak at Vermont Yankee is no small thing. The material is dangerous at low concentrations, persistent in the human body, impossible to filter, and hard to contain. The leak is limited to the area in and around the plant for now, but I can’t imagine the isolation and cleanup is going to be easy.
Well not exactly. Not in the amount we are talking about.
The silliness of the Vermont Yankee tritium scare was captured by a former Naval Officer who served in nuclear powered submarines, Rod Adams, who is currently a well regarded blogger on the uses of nuclear energy. Adams noted,
Based on reading a number of different articles and checking through the tables provided by the Vermont Department of Health, the fluid that was leaking into the ground contained tritium at a concentration of approximately 2.5 million picocuries per liter. That is equal to 2.5 x 10^-6 curies per liter. The rate that it was leaving the pipe was roughly 100 gallons (370 liters) per day. If the leak had been going on for a year before being detected and stopped, the total quantity of fluid that left the pipe would equal 138,000 liters. The total activity released would be 0.35 curies.

If a single person consumed every drop of that water, their whole body radiation dose would equal roughly 30 rem. According to a 1977 UNSCEAR study, the LD-50 (lethal dose for 50% of the population receiving the exposure) for tritium in adult rats was determined to be 1000 Rad. For the kind of low energy beta emissions that are produced by tritium, a rem is equal to a Rad. A dose of 30 rem received over a 1 year period would be unlikely to cause any immediate health effects, though it might add an additional risk of developing cancer sometime during the person's life. The magnitude of that risk could be computed using the conservative linear, no-threshold dose assumption.

Of course, a person who tried to drink 378 liters per day for a year would have problems more immediate the possibility of increasing their lifetime risk of cancer.

I also was asked to put this discharge into some kind of perspective, so I decided to compare it to the allowable and measured releases from a well operated and safe CANDU reactor in Ontario. Pickering B has a Derived Release Limit (DRL) for tritium of 490,000 terabecquerels each year. That is 4.9 x 10^17 Bq or 13 million curies.
Critics of nuclear power would insist that if we had a release of 13 million curies of tritium in the United States, we would have a crop of two headed babies. However, in Canada where 13 million curies annual releases of tritium are the norm, two headed babies are not being born.

On the other hand no one doubts that the escape of plutonium from a reactor can be a dangerous matter. Yet unlike tritium which is less dangerous than salt, but likely to escape from a reactor, Plutonium os very dangerous, but very unlikely to escape from a reactor. Nuclear critics love to recited how dangerous plutonium is. For example, journalist David McNeill stated
We might also cite the example of MOX fuel and plutonium, a substance so toxic “that a teaspoon-sized cube of it would suffice to kill 10 million people,”
In fact the Guardian reported,
In a possible sign that the contamination is more widespread than previously thought, a university researcher said at the weekend a small amount of plutonium had been identified a mile from the front gate of the Fukushima plant.

It is the first time plutonium thought to have originated from the complex has been detected in soil outside its grounds.
But was the escaped plutonium dangerous? The Guardian reportsed,
Masayoshi Yamamoto, a professor at Kanazawa University, said the level of plutonium in the sample was lower than average levels observed in Japan after nuclear weapons tests conducted overseas.
In conventional Light Water Reactors plutonium is produced inside fuel pellets. The fuel pellet is a ceramic and in almost every case the fuel plutonium will remain there. The only way the plutonium might escape the fuel pellet, would require that the reactor core overheat to such an extent that the ceramic fuel pellets start to melt. Once the plutonium escapes the fuel pellet, it faces a further barrier, the pressure vessel, which contains the reactor core inside a thick steel wall. If the plutonium managed to get past the wall of the pressure vessel, it would face one or more cement barriers, and then the forces of gravity as it pulled the plutonium to the outside grounds of the nuclear power plant. So how did the plutonium manage to travel a mile away from the Fukushima reactors? The answer is probably because it was ejected from the reactor building by a hydrogen explosion. More likely the plutonium was contained inside the spent fuel pellets that were housed in a pool above the Fukushima reactors. It is far from satisfactory that any plutonium managed to escape from the beyond the grounds of the Fukushima reactors, but in fact the amount that escaped was so tiny that it could do no harm.

So is there anyway, to insure that no plutonium ever escapes from a reactor core? Yes there is, in fact no plutonium can escape from a reactor if plutonium is not produced inside the core. But how is that possible? First while a lot of plutonium is produced in uranium fuel cycle reactors, less than 10% of that amount is produced in the thorium fuel cycle. A 1 GW LFTR would produce about 40 Pounds of Plutonium a year. If the goal is to minimize plutonium production this can be easily done. If the goal is to destroy plutonium, the presence of thorium in a reactor core facilitates the burning of plutonium. Finally if the goal is to produce no plutonium, then the use of fluid fuel thorium breeders (LFTRs) is highly recommended, because Neptunium-237, a plutonium predecessor isotope can be cleaned from a molten salt coolant before it can be converted from neptunium into plutonium by absorbing a neutron. Cleaning NP-237 from molten salt fluid is a relatively easy and low cost procedure. Once out of the LFTR core the neptunium can be destroyed in a burner reactor.
Thus if preventing the escape of plutonium from a reactor core is a major nuclear safety goal, designing molten salt thorium fuel cycle reactors that feature neptunium cleaning from core salts, would prevent the production of plutonium. If there is no plutonium production there can be no escape of plutonium.
Thus we have a choice of safety approaches to plutonium management, with the possibility of complete elimination of plutonium from waste stream a real possibility if it was desirable to do so. Total burn of plutonium would be yet another option.

A further approach to Plutonium safety issues would involve the use of underground reactor placement. If the reactor core and all radioactive fluids were kept underground. In Thorium fueled underground power plant based on molten salt technology, Ralph Moir and Edward Teller, Nuclear Technology 151 334-339 (2005), the authors explain,
An important feature of our proposal is to locate every- thing that is radioactive at least 10 m underground—where all fissions occur—while the electric generators are located in the open, being fed by hot, nonradioactive liquids. The reactor’s heat-producing core is constructed to operate with a minimum of human interaction and limited fuel additions for decades. . . .

Under- grounding will preclude the possibility of radioactive contam- ination in case of airplane disasters. A combination of 10 m of concrete and soil is enough mass to stop most objects. It would eliminate tornado hazards and, most particularly, contribute to defense against terrorist activities. In case of accidents, under- grounding, in addition to the usual containment structures, en- hances containment of radioactive material. The 10-m figure is a compromise between safety and plant construction ex- pense. We anticipate the cost to construct underground with only 10 m of overburden using the berm technique will add ,10% to the cost.
Moir and Teller note the safety advantage of underground placement,
A fourth safety measure is locating the reactor underground, which itself is one extra “gravity barrier” aiding confinement. A leakage of material would have to move against gravity for 10 m before reaching the atmosphere.
Plutonium is very heavy. In order for plutonium to overcome the "gravity barrier" some force would have to transport it to the surface. That force cannot be an explosion, because there is nothing in the reactor or its fluid salts that can explode. Nor can it be a fire, because no fire is possible. Thus the plutonium is trapped underground. In "Migration Paths for Oklo Reactor Products and Applications to the Problem of Geological Storage of Nuclear Wastes," G. A. Gowan repotted that plutonium along with many fission products including Zr, Nb, Ru, Pd, Ag, Te, Bi, and the rare earths, were immobile the the Oklo natural reactors core areas over a period of time of well over a billion years.

D.G Brookins found,
The actinides . . . were all retained in the host pitchblende.
In fact it appears that less than 10% of the actinides present when the original Oklo deposit was laid down had been lost to natural causes over a nearly two billion year period of time. Thus we can have a high degree of certainty that plutonium would be contained in underground reactor chambers, following the very unlikely event of the release of plutonium carrying salt inside the underground reactor chamber.

As noted what holds true for Plutonium also holds true for many fission products. The Oklo Reactors "natural experiment," that an underground reactor accident would not lead the the release of most fission products and actinides in a reactor or in a nuclear cool down storage. Neptunium would be a potential exception and for that reason it should be removed from long term storage of nuclear wast and disposed of by nuclear burning.

The fission products that are likely to escape in the event of a nuclear accident are well known and their behavior is well understood. They are noble radioactive gases, and volatile fission products. In fact the noble gases appear to pose little danger, and while volatile fission products are more dangerous, that danger can easily be mitigated. However, in Molten Salt Reactors it is easy to prevent the escape of fission products from the core fluid simply be removing them by simple and well understood processes. Reactor researcher, David LeBlanc states,
The volatile fission products such as the noble gases and noble metals come out of the salt as produced. Noble gases simply bubble out and are stored outside the reactor loop. Noble and semi noble metals will plate out on metal surfaces and can be collected by replaceable high surface area metal sponges within the loop.
Ralph Moir and Edward Teller note,
The molten salt reactor that operated in the 1960s had a big advantage in the removal of many fission products without much effort. Gases ~Kr and Xe! simply bubble off aided by helium gas bubbling, where these gases are separated from the helium and stored in sealed tanks to decay. Noble and semi noble metals precipitated. In the planned reactor, the old method of removing the gases may be repeated.
And what of Tritium? Tritium can be completely eliminated from MSRs by elimination of Tritium predecessor isotopes from the salt formula, or if that is not considered desirable for other reasons, it can simply be trapped in coolant salts, or by venting it in the off gas system and then trapping it in sodium fluoroborate.