Thorium as a nuclear fuel: myths and realities
Thorium was identified for the first time in 1828 by the Swedish chemist Jöns Jacob Berzelius, but the announcement of this discovery was made only 1 year later at the Academy of Sciences of Sweden.
Author: Dominique Greneche
This study is intended to provide a ‘fresh look’ of the thorium nuclear fuel cycle which is considered by some experts as a complement or even a substitute of the classical uranium nuclear fuel cycle for the operation of nuclear reactors. The authors first examine the incentives that led to numerous studies in the past and that still justify the continuation of some work on this topic in the world. For that purpose, they describe and explain the main physical and nuclear properties of thorium as well as those of the fissile isotope that is generated by the thorium, uranium-233, since this is its major advantage. They also give a summary of the available data on terrestrial thorium reserves. Then, they present an overview of the history of the development of thorium cycle including the acquired experience of its use in nuclear reactors. On these bases, they explain the various problems that may arise using the thorium cycle on an industrial scale, from the front end of the fuel cycle to the final stage of spent fuel reprocessing and recycling of materials and vis-à-vis the waste associated with this cycle. Finally, they discuss some generic issues related to implementation of the thorium cycle, in particular the issues of non-proliferation and economy. A conclusion is a summary of the advantages and disadvantages of thorium cycle and its development prospect.
Thorium was identified for the first time in 1828 by the Swedish chemist Jöns Jacob Berzelius, but the announcement of this discovery was made only 1 year later at the Academy of Sciences of Sweden. The name of this element is derived from the Norse god of thunder ‘Thor’. Its radioactive nature was established by the German Carl Schmidt, 2 months before Marie Curie rediscovered this property in March 1898. However, the radioactive half-life of its unique isotope thorium-232 (Th232) is extremely long as it is worth about 14.1 billion years, making it virtually stable, hence his presence on Earth today.
Historically, one of the major non-nuclear applications of thorium was its use in form of oxide, ThO 2, incandescent Mantles, thanks to the very high melting temperature of this oxide which is the highest of all known oxides (3390°C). However, its use for various non-nuclear applications is increasingly rare today because of its radioactive nature. It is soon after the discovery of fission that thorium aroused new interest when it was determined that the atomic nucleus of Th232 could give birth to a fissile isotope of uranium, the uranium-233 (U233), after some nuclear transformations induced by a neutron capture. It seems that the first scientists to have glimpsed this possibility are British physicists Bretscher and Feather in 1940, on the basis of the new theory of the atomic nucleus of Bohr and Wheeler. However, at the end of that same year, Glenn Seaborg entrusted to a young student of his team, John Gofman, experiments of irradiation on thorium, using his famous cyclotron. He isolated then in March 1941, so practically at the same time as the discovery of plutonium by Seaborg himself, a tiny quantity of protactinium-233 (Pa233) in which he (with Seaborg) detected an increasing alpha radiation (as the Pa233 decreased) who signed the presence of U233. More importantly, Seaborg and his team estimated roughly that this new isotope of uranium thus created had a half-life in the range of 100,000 years, and so it was relatively stable. Reporting these events in 1976, Seaborg wrote:
‘Of special importance is our demonstration through these results that U233 is sufficiently long-lived to be practical source of nuclear energy should it be found to be fissionable with slow neutrons and should methods for its large-scale production be developed’.
The fissile nature of the U233 was clearly experimentally established 2 February 1942 and confirmed during the following days, from a quantity of a few micrograms to U233. This time, it was a kind of birth of the thorium fuel cycle.
Therefore, using thorium as a nuclear fuel is not a new idea and was considered as an option for the fuel of nuclear reactors since the birth of nuclear energy. Indeed, the ‘New Pile Committee’ created in April 1944 in the USA to explore a variety of reactor concept did recommend that ‘more work should be done on the nuclear development of thorium because of its greater availability’ (greater than uranium) and the Committee also suggested experiments to develop reactors that would convert thorium to U233. It is worthwhile to mention that members of this Committee included the most eminent physicist and chemists working in the area of nuclear reactors, and in particular three Nobel Prizes in physics. Among them, the initial idea of ‘thorium converters’ was submitted in early 1944 by Eugène Wigner (Nobel Prize in 1963) as an option of the making of the atomic bomb, using U233 as a fissile material. The proposal of using thorium for nuclear reactors was then taken again by Phil Morrison, another distinguished scientist, in the frame of works of the New Pile Committee.
However, thorium cannot be a real alternative to uranium, at least in the short term, because unlike the latter, it has no ‘fissile’ isotope with slow neutrons. It allows simply generating U233 by neutron capture in a nuclear reactor (via Th233, which half-life is 22 min, and Pa233, which half-life is 27 days). This artificial uranium isotope thus created, which is relatively stable (half-life of 160,000 years), is an excellent fissile isotope for slow neutron reactors. It is even better than uranium-235 (U235) or plutonium-239 (Pu239), and this is why the thorium cycle has always been significantly studied worldwide.
Reasons for considering thorium as a fuel for nuclear reactors
A chain reaction can obviously not be sustained with only the Th232 (unlike natural uranium containing 0.711% fissile U235). Therefore, a fissile material (U235, plutonium, U233) must be added to thorium to make a fuel for a nuclear reactor. In this case only, the thorium material can generate U233 which such as plutonium can be ‘burned’ in situ or recycled. In fact, it is first necessary that sufficient quantities of fissile material is available for this generation process of U233 from thorium can be initiated and developed at a large-scale and provided that the residual U233 content in spent fuel is recycled (as is done in France for plutonium in the uranium cycle). However, the process can be accelerated if one can design reactors called ‘breeder’ in which are produced globally more fissile material (U233) that is consumed during a cycle irradiation in the reactor. It is only on this condition that one could imagine an almost total replacement uranium cycle by thorium cycle in the long term.
The average number of fission neutrons produced by the absorption of a thermal neutron (called ‘eta’ factor, usually denoted η) is typically 2.27 U233 in a standard neutron spectrum compared with 2.07 for U235 and 2.11 for Pu239. This is a major advantage of the thorium cycle: the high value of η makes U233 the best fissile isotope in the range of thermal neutrons from all existing fissile isotopes. With such a value, it is theoretically possible to achieve BREEDING in today's thermal reactors with a THORIUM–U233 cycle, which is NOT POSSIBLE with STANDARD URANIUM–PLUTONIUM fuel cycle. In effect, the breeding process (that is to say, the ability to produce an excess of fissile nucleus compared with those which are consumed to produce power) is possible only if η− 1 is >1 (or η − 2 > 0). However, in nuclear reactor cores, one loses neutrons in various sterile captures or leakages and it is therefore necessary that the margin to 1 of (η − 1) be sufficient. For slow neutrons, this margin is not sufficient for U235 (and Pu239), but it is more comfortable for the U233. We note in passing that the situation is quite different for fast neutrons where Pu239 is most suitable for enabling the fast neutron breeding process [which also explains that plutonium is the preferred fuel for fast neutron reactors (FNRs)].
It is only after the spent fuel is unloaded from the reactor that one can recover the residual U233 formed in the reactor for a possible recycling. This is the so-called ‘reprocessing–recycling’ process, similar to that which can be implemented for the uranium–plutonium cycle. Thorium, which consumption in reactors is very low, can also be recovered and recycled in reactors.
However, this recycling operation of U233 raises difficult technical problems related to the unavoidable presence of a certain amount of the isotope uranium-232 (U232) mixed with U233 [mainly formed by ( n, 2 n) reaction on U233]. This is because the U232 is a radioactive isotope with a half-life of 72 years, which some descendants are very strong gamma emitters (and thus noxious).
Natural abundance and reserves of thorium
Natural thorium is a relatively abundant element on Earth, with an average concentration of 7.2 ppm in the Earth's crust. This value is significantly higher than for uranium (2.5–3 ppm), which results in the very long half-life of Th232 compared with U238 (4.5 billion years). However, this does not mean that the exploitable reserves of thorium are two or three times higher than those of uranium, as many people say in the literature. In fact, because of its use so far very limited to specific applications mentioned above, there has never been a comprehensive survey of the thorium so that reliable estimates of world reserves of thorium are not currently available. The well-known ‘Red Book’ of the IAEA – OECD  updated periodically included fairly detailed thorium resources until its 1981 edition. There is also a more recent assessment by the US ‘Geological survey’  that ‘reserves’ world are estimated at 1.4 million tons of ThO 2.
From these various estimates, one can evaluate that quantities of thorium that could be extracted from the ground in reasonable commercial conditions, certainly amount to several million tons and are probably at least the same order of magnitude as those of uranium. In any event, if a closed thorium cycle was deployed a day at a large industrial scale, thorium reserves are not in fact a real problem because, as the U238 is a fertile isotope, the transmuted portion of which can be easily recycled. Thus, a thorium cycle deployed with recycling of U233 would be able to support the development of nuclear energy for hundreds or even thousands of years in a breeding mode.
Past developments of thorium fuel cycle
Feedback experience of thorium utilisation in nuclear reactors
During the early development of nuclear energy for civilian purposes, in the 1950s and 1960s, a large number of ways to use the thorium were studied, not only in the United States and the former USSR, but also in Europe and, to some extent, in Asia . To illustrate the interest aroused by the thorium in the early years of the development of nuclear energy, we can mention that the first international symposium dedicated to the thorium cycle took place in 1962 in the United States.
The initial impetus for studies on thorium cycle was essentially blown through the development of molten salt reactors (MSRs) in the late 1950s in the United States. These reactors are indeed particularly suitable for thorium cycle thanks to the opportunity they offer to reprocess the fuel ‘online’, which opens the way for continuous extraction of Pa233, giving rise to the U233 free of U232, and recycling of the U233 always ‘online’. This potential, combined with other favourable properties of MSRs allows to reach breeding with a ‘thermal’ reactor, which is an attractive prospect in many ways (safety, economy, technology, sustainability), at least at the conceptual stage.
The most remarkable, however, is that the first commercial light water reactors (LWRs) developed in the United States initially operated with a thorium-based fuel. These are the Elk River reactor boiling water reactor (BWR), which started in 1963 just after the prototype BWR Dresden (started in 1960) and the Peach Bottom reactor pressurised water which started in 1967 just after the prototype of Shippingport (started in 1957). It is also remarkable that the demonstration of the possibility of LWR breeder was carried out precisely at the Shippingport reactor in the 1970s and 1980s using a U233/thorium cycle [ 3 ]: the conversion factor (CF) reached was 1.0139 [ 4 ]. This was the only power reactor that used U233 as the fissile material. It was thus demonstrated experimentally that it is possible to produce more U233 than it was consumed in an LWR. However, this performance was achieved at the cost of a sophisticated design of the reactor core, which would be difficult to extrapolate to commercial reactors, and to the detriment of other performances. From that time, many attempts to use thorium fuel in research reactors and power has been made, and significant experience has been accumulated on this type of fuel.
In the frame of this paper, it is not possible to relate more on all this experience and we refer the reader who wishes to deepen this aspect to in [3–5].
Analysis of incentives and hindrances in the historical development of the thorium fuel cycle
In the early 1980s, several factors have much dimmed the enthusiasm for alternative cycles uranium. First, interest in the nuclear option has weakened significantly, particularly in the United States, where public support for nuclear power has declined considerably after the accident at the Three Mile Island in April 1979. This mistrust was then intensified in Europe after the Chernobyl disaster 7 years later. This is also from the early 1980s, and in conjunction with these events, that the price of uranium has fallen to very low levels, so that the search for an alternative fuel to uranium became virtually without interest, even if also the use of thorium fuel had some other benefits (which are discussed in this introduction). A third event that occurred in the late 1970s was the banning of commercial processing of irradiated fuel by US Presidents Ford and Carter administrations for political reasons related to the risk of nuclear proliferation from fissile material separated during these operations. Following the same logic, the use of enriched beyond 20% uranium in civilian reactors was almost forbidden in the world. These decisions obviously penalised much the possible implementation of a thorium cycle for two reasons. First, the promoters of the thorium cycle had to give up the reference cycle based on the use of HEU by replacing it with uranium enriched to 20% maximum (uranium called ‘MEU’ which stands for ‘mid enriched uranium’), which many complicating the implementation of the thorium cycle and reduces its overall performances. Second, the ban on reprocessing the spent fuel implied to deny the possibility of recovering the U233 content in spent fuel for recycling, at least in the United States.
In the last decade however, there has been a renewed interest in thorium-based fuels. This seems to have been initially motivated by the development of a fuel cycle deemed ‘proliferation-resistant’ developed in the late 1990s by an American team led by Alvin Radkowsky . Nevertheless, the significance of this concept in terms of non-proliferation is questionable (it is not possible to further elaborate here on this non-proliferation aspects of the thorium cycle, which is dealt with by the author in some other specialised publications).
Another factor that boosted (timidly) the interest in the thorium cycle is the impetus for the development of nuclear energy in the world during the years 2000–2010, though mitigated to some extent by the accident in Fukushima 2011. In the wake of the expected growth of nuclear power worldwide, new questions have emerged here and there on the long-term availability of cheap uranium. Finally, new strengths emerged in favour of the thorium fuel cycle including the low production of plutonium (for those who consider that plutonium is a burden) or the ability to ‘burn’ excess plutonium in thermal reactors by implementing a thorium–plutonium cycle. Such concepts can also be imagined in very innovative reactors such as MSRs, accelerator driven systems (reactors also called ‘hybrid’) and even blankets in fusion-based energy systems .
All these reasons make that there is currently a resurgence of interest in thorium in several academic and R&D institutions, but also from certain industrial reactor designers and/or fuel vendors. One particular case is represented by India, which is known as a strong proponent of the thorium cycle since the early 1950's. This is because this country has almost no domestic uranium resource and a very limited access to imports of uranium for political reasons, but has large thorium resources identified in their soil and an ambitious nuclear programme. One can add that in Japan, the HTTR reactor could be used in the future with thorium, as well as the HTR-10 in China.
Present status of the thorium fuel cycle development
Studies on the thorium cycle continues today in several countries such as the United States, Russia, China, Canada, Sweden, Norway, Japan, France and especially India. The European Union is also active on the subject, but at a modest level. In fact, many of these works are carried out in the broader context of research on MSRs already mentioned above. Nevertheless, most of these programmes are mainly restricted to academic works which have limited use of experiences and that generates virtually no technological development, with the exception of India.
Nuclear material management from thorium fuels
The one through thorium cycle (that is to say, without recycling the residual fissile materials) would require of course an external supply of available fissile material, that is, to say either the U235 (in the form of enriched uranium at an appropriate level) or plutonium (with adequate isotopic composition). In this case, one should seek the highest possible combustion rate (accompanied by an increase in the split of the core) to produce and ‘burn’ the maximum in situ U233 generated by the thorium, knowing nevertheless that the overall benefit of this process in terms of consumption of existing fissile material remains limited. However, one should especially aim the highest possible CFs. In this regard, the CFs achieved with a thorium cycle are always higher than those obtained with conventional uranium–plutonium cycle. This is, for example, the case heavy water reactors for which CF can reach values close to 0.8 and even up to 0.9, so close to 1 for some concepts. We must also mention MSRs with which it is theoretically possible to achieve iso-generation or even breeding (CF > 1), thanks in particular to the online extraction of Pa233 and of certain capturing fission products. This is a really attractive option for the implementation of a thorium cycle as it opens the way for a system of thermal breeder reactors (having in particular a significant lower fissile inventory than FNRs) in which the raw fissile material becomes practically inexhaustible.
Numerous studies have been devoted to the use of thorium in thermal reactors as various combinations of fuel cycles are possible by combining different types of reactors, acting as symbiotic systems. It turns out that thorium can be mixed with four types of fissile material:
- Highly enriched uranium (>90% U235), which gives a cycle referred to as ‘Th-HEU’. It was the reference fuel for HTR reactors in the 70s in the USA and Germany, but today its use seems totally excluded given the severe restrictions related to nuclear proliferation risks because HEU can be used directly and through a relatively simple way to make an atomic bomb.
- Medium enriched uranium (20% U235), which gives a cycle referred to as ‘Th-MEU’, already mentioned above. This leads to the cycle sometimes called ‘denatured’ in the sense that uranium cannot be used directly in the manufacture of an atomic bomb. Such an option might be called ‘last resort’ because it leads to a complicated materials management insofar as it combines all the constraints associated with the materials management of uranium and thorium cycles.
- Plutonium, whatever its isotopic composition, simply called cycle ‘Th–Pu’. This cycle is sometimes viewed as the best option for the use of thorium as it offers an interesting alternative to the standard MOX fuel for the recycling of plutonium produced by standard uranium-based fuels.
- U233, where it is available in large quantities after processing of spent thorium-based fuel. This is the ‘Th–U233’ cycle.
The results of these studies show that the use of thorium in ‘conventional’ thermal neutrons reactors allow a global economy of the use of uranium from a few tens of per cent to a maximum of about 80%, when equilibrium of the system is achieved. The results depend on a lot of combinations of reactor types considered and recycling options used in these studies.
Regarding the use of thorium in FNRs, several studies have also been carried out especially in Russia for the BN-800 reactor, but also in France and Europe. They have demonstrated the ability to achieve iso-generation of fissile material or even breeding with a Th–U233 cycle. However, the performance achieved in this area are lower than those obtained with a uranium–plutonium cycle, mainly for reasons related to neutronic best qualities of plutonium from U233 for fast neutron spectra. Another reason is that the fission cross-section of thorium itself for fast neutrons is much smaller than that of U238 (about three times lower). The use of the thorium cycle in FNRs is thus not very attractive, though some proponents of thorium cycle see some benefits of such use, in particular because of reactivity sodium void effect much less positive than that of uranium–plutonium cores.
Overall, if the thorium would be heavily used in thermal neutron reactors with a closed cycle (that is to say, with recycling of U233), it is hoped that the overall resources of fissile material could be increased a factor of 2 or maybe more in the long term, provided that there are sufficient amounts of natural uranium to fuel the reactor park over a long period. Let us recall, however, that if the FNRs were developed intensively with ‘classical’, uranium–plutonium cycle, the energy potential of natural uranium resources would be multiplied by a factor of 50–100. In this case, the use of thorium cycle as a complement to this system would increase again this already enormous energy potential by a factor of 2 or more, depending on the thorium resources available.
Back end of the thorium fuel cycle
To take full advantage of the assets of the thorium cycle, it is necessary to reprocess the spent fuel to extract the residual U233 and then to recycle it in nuclear reactors. At the same time, this option allows recycling thorium itself which consumption in reactors cores is very low.
Experience on the reprocessing of thorium-based fuel is very limited, but it is not zero. In fact, it has existed since the late 1940s in the United States and some other countries such as India have also conducted experiments in this field, but on very small amounts of fuel.
Research and development works as well as small-scale experiments have shown that the treatment of thorium-based fuels is technically less simple to perform than the uranium-based fuels mainly due to the dissolution step. Thorium is actually much more difficult to dissolve than uranium, either in the form of metal or oxide. Historically, it is the Oak Ridge National Laboratory in the United States, which has developed a hydrometallurgical process called Thorex for thorium-based fuel reprocessing, derived from the Purex process used today for the reprocessing of uranium-based fuels. One can find in Section 6.2 of [ 3 ], a good description of this process based on the addition of fluoride in the acid dissolution solution. To overcome these drawbacks, alternative non-aqueous processes to Thorex process have been studied in the past, as the process of volatilisation of fluorides or electro-refining processes, but these studies have not produced very good results and they were interrupted prematurely.
Once the dissolution performed, the steps of chemical separation and purification of materials for thorium cycle should not be fundamentally different from that of a uranium–plutonium cycle. Only the management of an additional element, the thorium, could possibly add some difficulties without making the process too complicated.
U233 recycling issue
As already mentioned, U233 is inevitably mixed with U232 which descendants such as Tl208 are strong gamma emitters (alone, Tl208 contributes to 85% of the total dose of radiation emitted after 2 years by the descendants of the U232). This requires handling the U233 remotely behind shielding after a few months, which greatly complicates the processes of the fabrication of fuels with U233 in terms of technology. Though it is perfectly possible with the techniques now available, it is clear that this leads to significant additional manufacturing costs for these fuels.
Final disposal of radioactive waste
The real difference between the uranium and thorium fuel cycles actually comes from the quantities of ‘Minor Actinides’ generated in each case. Indeed, in the case of the uranium cycle, significant amounts of three long-lived transuranic elements are generated: neptunium, americium and curium. The trouble is that these radioelements are particularly radiotoxic alpha emitters and, even if they are in a small amount, they contribute to almost all of the global radiotoxic inventory (GRI) of ultimate radioactive waste beyond a few 100 years (excluding plutonium which is supposed to be recycled in nuclear reactors).
With U233, almost none of these minor actinides are produced. Indeed, americium and curium mainly come from plutonium while the main isotope de neptunium, Np237, mainly comes from the U235. However, a Th/U233 fuel produces other long-lived radionuclides (excluding isotopes of thorium and uranium which are supposedly to be recycled). The chief of these is the minor actinide protactinium-231, basically formed from reactions (n, 2 n) on Th232. This isotope of protactinium (half-life 32,760 years) is therefore involved in the long-term GRI significantly. Nevertheless in a Th–U233 fuel, one can find also some minor actinides associated with the uranium–plutonium cycle such as Pu238 (half-life 88 years) and Np237 already mentioned (half-life 2.144 million years), but in very small amounts. In fact, we have seen that there are various ways to deploy thorium cycle using different fissile material (MEU, plutonium, U233) in different types of reactors (light water, heavy water, fast neutron, molten salts etc.), often in ‘symbiotic’ systems which involve different types of reactors at the same time, with a proportion which varies with time. This makes comparisons difficult in terms of GRI for each scenario. However, all studies on this subject show that nuclear fuel cycles based on thorium lead to GRIs well below those cycles uranium (see, for example, the study in ).
This overview shows that thorium offers interesting perspectives, especially in terms of natural uranium savings (if U233 is recycled), but also with regard to the reduction of total final waste radiotoxic inventory. Thorium fuels present also attractive features in terms of behaviour under irradiation and neutronic behaviour in the reactor. However, despite the existence of examples of use of thorium reactor in the past, industrial experience in this cycle is still very limited and almost non-existent on the back end of the cycle (reprocessing and recycling).
Most of the basic knowledge on thorium cycle is thus available, but it is clear that the deployment of this cycle on a large scale would require much R&D, especially in the field of reprocessing and fabrication of U233-based fuels, as well as heavy industrial investments. It is unlikely that in the near future, the conditions are met to justify the initiation of such effort for a majority of countries. However, in few decades, the emergence of new constraints could change the current situation and lead to industrial deployment of fuel cycles based on thorium. To this regard, one of the drivers could be the opportunity offered by these fuel cycles to approach or even to reach the iso-generation of fissile material in some types of thermal reactors. Another incentive would be the recycling of plutonium from LWR MOX fuel in a symbiotic system of nuclear reactors.
In that perspective, the thorium cycle deserves certainly further research and considerations. This report will provide an update and a comprehensive overview of worldwide works on this important topic.
- AIEA – OCDE/AEN: ‘Uranium 2009: resources, production and demand (so called Red Book)’.
- Cordier D. J.: ‘U.S. geological survey, mineral commodity summaries’, January 2012, p. 168.
- ‘Thorium fuel cycle-potential benefits and challenges, international atomic energy’. IAEA-TECDOC-1450, Vienna, May 2005.
- ‘Thorium based fuel options for the generation of electricity: developments in the 1990s’. IAEA TECDOC-1155, Vienna, May 2000.
- World Nuclear Association (WNA), November 2004. Available at http://www.world-nuclear.org/info/inf62.htm.
- Galperin A. Reichert P. Radkowsky A.: ‘Thorium fuel for light water reactors – reducing proliferation potential of nuclear power fuel cycle’, .
- Gruppelaar H. Shapira J. P.: ‘.