Salt of the earth

A design of nuclear reactor not built for more than 40 years is being scrutinised by scientists. They believe it could be economically viable when uranium becomes scarce.

The molten salt reactor (MSR) does what it says on the tin. It uses molten salt at atmospheric pressure to cool the reactor. The radioactive fuel is also contained within the molten liquid. So, unlike most contemporary reactors, it has no pressurised gas coolant and no solid fuel rods. It could also be used to breed fuel and, at temperatures above 850ºC, to make hydrogen.

It has been done before — once. Oak Ridge National Laboratory, Tennessee, US, built the first MSR in 1964. The small experimental reactor operated until 1969 but the design was never commercialised. 'People believed it was more economically interesting to use other designs,' said Jan Leen Kloosterman, associate professor in the faculty of applied sciences at the Technical University of Delft in the Netherlands.



Changing economics

But economics are changing and uranium is getting more expensive. The demand for the precious ore is expected to rise as more nuclear reactors are built. The International Atomic Energy Agency predicts that 130 new nuclear power stations will be built in the next 15 years.

This is one reason why representatives from 11 nations, including the UK, founded the Generation IV Forum (GIF). The GIF has identified six 'fourth generation' reactor concepts that could be economically viable but avoid the use of highly fissile material and so reduce the risk of nuclear weapon proliferation. MSR is one of the chosen concepts. So now groups in France, Germany, the Czech republic, the US and the Netherlands are investigating its potential.

'The main benefit is that it can use thorium as a fuel, which is much more abundant than uranium,' said Kloosterman, who has been funded by the Dutch government to hire a PhD student to study the nuclear physics of an MSR.

The principle is straightforward. The nuclear fuel is dissolved in the molten fluoride salt coolant, which reaches criticality by flowing into a graphite core that also serves as moderator. A variant relies on ceramic fuel being dispersed in a graphite matrix with the molten salt providing low-pressure, high-temperature cooling.

One additional benefit is the ease with which the fuel could be replenished without the need to shut down the process. As the fluoride salt is used both to remove heat from the core and to circulate the fuel, a fraction of the salt stream could be diverted to extract the fission products and add fresh fuel. This means there would be no need to stop the process for refuelling, which could reduce the operation costs considerably.



Safety first

Kloosterman also believes MSR is a safer process. Put simply, the fuel is already in molten form so there is no possibility of meltdown. Moreover, it can be self-regulating. 'If the temperature increases, then the molten salt will expand, which will decrease the density of the fuel within the reactor core and so any chain reaction will die,' said Kloosterman. 'That, of course, depends on the design of the core,' he added.

The calculation of all physics aspects of an MSR is a challenging task never undertaken before. The Delft team believes that by applying present-day computing power, it should be possible to calculate simultaneously the spatial distribution of the neutron flux, the power, the salt density, the precursor atom density and the temperature during steady state reactor operation and during transients.

Meanwhile, groups elsewhere are studying the materials that would have to be developed for an MSR. The Oak Ridge experiment ran at up to 705ºC but future designs may operate at 900ºC, so that hydrogen could be produced thermo-chemically as a secondary benefit.

Molten fuel-carrying salt at these elevated temperatures can create an extremely hostile environment. The presence of any water vapour would form hydrofluoric acid, which is highly corrosive. At Oak Ridge, Hastelloy-N, a proprietory super alloy whose main alloying ingredient is nickel, did the job; but materials such as carbon composites may be developed for future MSRs.

One thing that is certain is that commercial MSRs are a long way off. The theoretical experiments may be accelerated as computing power increases but that is rarely possible with physical research and development. 'Things go very slowly in the nuclear world,' said Kloosterman. 'It takes a few years to irradiate materials and the results can point to the need for more tests.'

If Kloosterman's team, and successors, are successful with their research, the first MSR since 1969 could be up and running by 2036.