Nuclear energy & greenhouse

Is going nuclear the answer?
Mark Diesendorf

Since 2000 the nuclear industry has mounted a massive international media and lobbying campaign to promote nuclear energy as a solution to the enhanced greenhouse effect. Nuclear energy, it is claimed, emits no carbon dioxide (CO2) and can be rapidly deployed to substitute for coal-fired power stations.

If these claims were true, then nuclear energy may deserve serious consideration. But, without checking the claims with independent scholars, a number of journalists, politicians and even academics have taken up the nuclear cause. This article presents evidence that these recently recruited supporters of nuclear energy have been premature in their support, because they have uncritically omitted to address two fundamental problems:
• CO2 emissions from the nuclear fuel chain; and
• inherent constraints on the speed of deployment of nuclear energy.

Before addressing these fundamental issues, this article sets out the key steps in the life-cycle of nuclear energy, enabling us to identify the stages where CO2 is omitted and incidentally where nuclear weapons explosives are produced. Although this article doesn't address the hazards of nuclear energy, apart from CO2 emissions, it must be pointed out that all the previously debated hazards are still present and are very real:
• nuclear weapons proliferation, including their use by terrorists;
• nuclear waste management and decommissioning nuclear power stations;
• rare but devastating accidents;
• poor economics (nuclear energy is even more expensive than wind power in the UK and USA).

Nuclear fuel chain

The source of energy in nuclear power stations is known as nuclear fission, the splitting of heavy atomic nuclei into smaller nuclei with the release of energy. An element that can undergo fission, such as uranium-235 (U-235) or plutonium-239 (Pu-239), is called 'fissile'. Nuclear power stations use the heat generated from the fission of U-235 and Pu-239 to boil water, to produce steam, to turn a steam turbine, to turn a generator. This may sound simple in theory, but in practice a complicated multi-stage process is required to produce nuclear electricity.

The nuclear fuel chain extends from the mining and milling of uranium, through the construction and operation of the power station, to the management of spent fuel and the decommissioning of the power station at the end of its operating life. Three of the steps - reprocessing, decommissioning and long-term waste management - are not at present carried out on a commercial basis, despite several attempts. The original intention was to have a nuclear fuel 'cycle', with uranium and plutonium being extracted from the spent fuel by reprocessing and fed back into the power station as additional fuel. Since this is not occurring in practice to a significant degree, it is more appropriate to describe the processes as a nuclear fuel 'chain'.

In general, uranium exists at low concentrations in ore bodies. When the ore is mined and milled, it releases long-lived low-level radiation into the environment. At this step it becomes 'yellowcake', U3O8. Next it is converted into the gas, uranium hexafluoride, UF6.

The element uranium exists naturally in two isotopes, which are chemically identical but differ slightly in atomic weight. The common isotope, U-238, is not fissile, but the rare isotope, U-235, is. For the most common type of nuclear power station, the light water reactor, uranium, in the form of uranium hexafluoride, is enriched to increase the concentration of U-235 from its naturally occurring level of 0.7 per cent to 3-5 per cent of the uranium. It is then fabricated into fuel pellets that are inserted into fuel rods and loaded into the reactors. If uranium is not loaded into the reactor, but rather is further enriched to a much higher concentration, it becomes suitable for use as a nuclear weapons explosive. The bomb that was exploded over Hiroshima used highly enriched uranium.

After a period in the reactor, much of the uranium has been converted into a highly radioactive waste mix. This spent fuel is removed and fresh fuel pellets are loaded. Most spent fuel is stored for decades under water to allow the radioactive elements with shorter half-lives to decay and to allow the temperature of the spent fuel to drop. Then, as mentioned above, the original idea was to put it through chemical reprocessing, in order to extract the unused U-235 and a new element, plutonium-239 (Pu-239), that is created as a by-product of the fission of uranium. Like U-235, Pu-239 is fissile - it can be used either as fuel in a nuclear power station or as a nuclear weapons explosive. The bomb that was exploded over Nagasaki used Pu-239. If inhaled, plutonium is very carcinogenic.

Some supporters of the nuclear industry claim falsely that only weapons-grade plutonium from special military reactors can be used as a nuclear explosive, and that the 'reactor-grade' plutonium extracted from spent fuel from nuclear power stations is too contaminated with a non-fissile isotope of plutonium to explode. This false notion has been contradicted by members of the former US Nuclear Regulatory Commission and the current US Department of Energy, as well as by a leading nuclear bomb designer, Dr Theodore Taylor.

Because the spent fuel is 'hot' both in terms radioactivity and temperature, it must be handled remotely, behind shielding. This is dangerous and difficult. In the USA three reprocessing plants were built at various times, but none was commercially viable and they have all been shut down permanently. The UK reprocessing plant at Sellafield (formerly Windscale) was shut down in April 2005, when it was discovered that 83,000 litres (equivalent to half an Olympic swimming pool) of highly radioactive liquid had been leaking from it unnoticed for the previous nine months. At the time of writing (April 2006) it is unclear whether Sellafield will ever be reopened. However, reprocessing is carried out to a limited extent, mainly in France. In practice very little plutonium is being 'recycled' in this way and vast quantities of nuclear wastes, containing plutonium and highly radioactive fission products such as strontium-90 and cesium-137, are in temporary storage in cooling ponds.

The long-term management of high-level radioactive wastes only exists in theory. The USA is building a waste repository at Yucca Mountain with an estimated life-cycle cost over hundreds of thousands of years of US$57 billion. But Yucca Mountain will have inadequate storage capacity even for the high-level wastes from the USA's existing nuclear power stations. If global nuclear power output were trebled, new waste repositories equivalent to Yucca Mountain would have to be built 'roughly every three or four years'.

For decades, the Australian SYNROC technology for long-term geological disposal has been described as 'promising'. It depends on spent fuel being reprocessed first. Yet, many nuclear experts now believe that reprocessing should not be done, because it makes plutonium much more accessible. SYNROC is not being used commercially.

Although several small nuclear power stations have been decommissioned at great cost, this has never been carried out for a full-size (1000 MW or more) nuclear power station. Costs could be as high as the original capital cost of the power station.

The use of nuclear energy grew rapidly during the 1960s, driven in part by the belief that it would soon become 'too cheap to meter' and by huge subsidies. Today, nuclear energy supplies 16 per cent of global electricity. France, Belgium and Sweden all generate half or more of their electricity from nuclear energy, while the USA and the UK 20 per cent and 11 per cent respectively. Australia has a small research reactor at Lucas Heights in Sydney that does not generate electricity. Concerns about hazards and unfavourable economics have stopped the growth of nuclear energy in all but one of the developed countries, Finland. In the USA there have been no new orders for nuclear power stations since 1978. However, in 2007 the US government plans to allocate over US$600 million towards nuclear energy, including R&D for a new generation of nuclear power stations. There is still growth in nuclear energy in developing countries.

Carbon dioxide emissions from the nuclear fuel cycle

The recent push for a revival of nuclear energy has been based on its claimed reduction of CO2 emissions where it substitutes for coal-fired power stations. But most of the energy inputs to the full life cycle of nuclear fuel come from fossil fuels and so are responsible for CO2 emissions at almost all stages of the nuclear fuel chain. The energy inputs and associated CO2 emissions of several of these steps of the nuclear fuel 'cycle' have been investigated by researchers who are independent of the nuclear industry: in 1991 by Nigel Mortimer, until recently Head of the Resources Research Unit at Sheffield Hallam University, UK; and independently in the 2000s by Jan Willem Storm Van Leeuwen, a senior consultant in energy systems, together with Philip Smith, a nuclear physicist, both of whom are based in Holland.1

They find that, especially for mining and milling and enrichment, the results depend sensitively on the grade of uranium used. Following Van Leeuwen and Smith, I define high-grade uranium ores to be those with at least 0.1 per cent uranium oxide (yellowcake, U3O). In simpler terms, for each tonne of ore mined, at least 1 kg of uranium can be extracted. For high-grade ores, such as most of those being mined in Australia, the energy inputs are indeed much less than the electricity generated. But, even for high-grade ores, there is a significant contribution from the construction of the nuclear power station, with the result that the station must operate for several years to generate its energy inputs. (For comparison, wind power requires only 3-7 months.) For a nuclear power station with lifetime of about 35 years, this is acceptable, although it introduces a limitation on the rate of growth of the nuclear industry, as discussed in the next section.

I define low-grade uranium ores to contain less than 0.02 per cent yellowcake, i.e. at least 5 times less concentrated than the high-grade ores. To obtain 2 kg of yellowcake, at least 10 tonnes of low-grade ore has to be mined. As you can imagine, this entails a huge increase in the fossil energy required for mining, milling and enrichment. Van Leeuwen and Smith find that the fossil energy consumption for these steps in the nuclear fuel chain becomes so large that nuclear energy emits the same or more CO2 than an equivalent combined cycle gas-fired power station. Furthermore, the quantity of known uranium reserves with ore grades richer than this critical level is very limited. The vast majority of the world's known uranium resources are low-grade. With the current contribution by nuclear energy of 16 per cent of the world's electricity production, the high-grade reserves would only last a few decades. If nuclear energy were to be expanded to contribute (say) half of the world's electricity, high-grade reserves would last less than a decade. No doubt more reserves of high-grade uranium ore will be discovered, perhaps even doubling current reserves, but even this would be insufficient for a sustainable substitute for coal.

The World Nuclear Association (WNA) has written a reply of sorts to Van Leeuwen and Smith. The main body of WNA's 'information brief' only considers high-grade uranium ore and focuses on the mechanism of enrichment. The appendix to this report does not engage with the main issues that Van Leeuwen and Smith raise, but rather obscures them. In particular, it quotes an unpublished report (only a brief summary is in the public domain) by the Swedish electricity utility Vattenfall that obtains different results from Van Leeuwen and Smith. Using unpublished reports is very poor science.

So, are there alternative future pathways for nuclear energy with lower CO2 emissions? Although there are vast quantities of uranium oxide in the Earth's crust, they exist at very low concentrations, typically 4 x 10-4 per cent, at which 1000 tonnes of ore would have to be mined to obtain 4 kg of uranium in the form of yellowcake. In this case, the energy inputs to extract uranium would be much greater than the energy outputs of the nuclear power station. Sea-water contains uranium at a concentration of about 2 x 10-7 per cent, meaning that 1 million tonnes of sea-water would have to be processed to extract just 2 kg of uranium.

A technically possible solution would be to switch to fast breeder reactors, which produce so much plutonium that in theory they can multiply the original uranium fuel by 50. Large-scale chemical reprocessing of spent fuel would be necessary to extract the plutonium and unused uranium, and this has its own hazards and costs, since spent fuel is intensely radioactive. We have already discussed the collapse of the 'commercial' reprocessing industry.

Fast breeders use liquid sodium as a coolant and so are more dangerous than ordinary nuclear reactors. So far, fast breeders have all been technical and economic failures. The largest was the French 1200 MW Superphoenix, a name that alludes to the mythical bird that burnt itself on a funeral pyre and then arose from the ashes to live again with renewed youth. Reality was rather different: Superphoenix commenced operation in 1985 as a 'commercial industrial prototype'. It operated only intermittently and very rarely at full power, experiencing leaks from its cooling system and several other accidents. It was shut down at the end of 1998 after costing an estimated total of about A$15 billion. At present there are no commercial scale fast breeders operating. The Russian 600 MW demonstration fast neutron reactor, Beloyarsk, is operating, but it has a history of accidents and does not seem to have ever operated as a breeder. The pro-nuclear MIT study does not expect that the breeder cycle will come into commercial operation during the next three decades.

Another theoretical solution to the shortage of high-grade uranium arises from estimates that there is about three times as much thorium in the Earth's crust as uranium. Thorium itself is not fissile, but it can be converted into U-233 which is fissile. But using thorium and U-233 would be a complicated system involving a type of breeder reactor, which takes us back to the problems outlined in the previous paragraph. India is attempting to develop such a system.

Following the disaster at Chernobyl, many energy policy professionals have also questioned the continued use of the ordinary 'burner' reactor. They propose that any future development of nuclear energy should be based on new types of reactors that are 'fail-safe' and proliferation-proof. Since the nuclear industry appears unwilling to make large, risky investments in the development of new technology, this would require huge quantities of government funding. This is all the more reason for scrutinising rigorously the claim that nuclear energy can safely and sustainably reduce CO2 emissions in the long term.

Another potential future technology is controlled nuclear fusion, in which light elements such as isotopes of hydrogen (i.e. deuterium and tritium) are combined to form heavier elements with the release of energy. This is the same type of nuclear reaction that occurs in the interior of our Sun and in hydrogen bomb explosions. For nearly half a century research has continued on the fundamental problem of creating a controlled nuclear fusion reaction and containing it in a laboratory. In stars, the very hot ionised gas or plasma undergoing the reaction is contained by the force of gravity. But, in the laboratory, high magnetic fields are used to contain the plasma. These systems are prone to many instabilities and so far scientists have been unsuccessful in achieving a controlled nuclear fusion reaction in which more usable energy is created than the energy input required to maintain the plasma.

Despite this fundamental shortcoming, several countries are combining their resources to build a US$5 billion experimental fusion reactor called ITER. The proponents claim that the reactor is technically ready to commence construction and could begin to operate in 2016 for a 10-year test period. It is not designed to generate electricity - that task would be reserved for the next phase, if ITER is successful. In theory, by scaling up from laboratory to large reactor, it may be possible to obtain a net energy gain. But it is also possible that the scale-up will introduce new kinds of instability. Clearly, nuclear fusion would not be a commercial prospect for at least 25 years, if ever. Even if it does eventually generate more electrical energy than the fossil energy inputs for containing the plasma, it may still require quite large fossil energy inputs for building the power station and preparing the initial fuel charge.

So, on the basis of present nuclear technology and the small existing high-grade uranium reserves, the potential contribution of nuclear power to the reduction of CO2 emissions is limited.

Inherent constraints on the speed of deployment

With growing evidence that global climate change may be accelerating, it is essential that the principal technologies for replacing coal-fired power stations can be implemented rapidly, without themselves causing a large increase in CO2 emissions.

Even in the unlikely event that vast new reserves of high-grade uranium ore are discovered, there still is an inherent limit on any rapid expansion of nuclear power stations as a solution to the enhanced greenhouse effect. This limit goes beyond the long planning and construction periods (8-10 years or more) experienced in building nuclear power stations. It results from the debt of fossil fuel inputs and associated CO2 emissions incurred during the construction of the station, even during the limited period when high-grade uranium reserves can be utilised. According to Van Leeuwen and Smith, the first 7-10 years of operation of the station must be devoted to paying back the energy and CO2 debt. For comparison, an independent UK study by AEA Technologies for the European ExternE study estimates that the construction energy alone (ignoring mining, milling, enrichment, etc.) of a 1000 MWe nuclear reactor is 40 x 1015 joules. Assuming a lifetime average capacity factor of 75 per cent, the energy payback period for construction is nearly two years of operation. It is emphasized that the energy payback period discussed here has nothing to do with money and everything to do with energy.

The nuclear industry and its supporters claim that Van Leeuwen and Smith have overestimated the energy payback period of nuclear power stations. Some claim that the construction energy alone can be paid back in just a few months. But independent analysis of the energy payback periods for the construction alone of coal-fired power stations gives about two years. Since nuclear power stations have in general similar materials inputs and poorer performance than coal stations, one would expect their energy payback periods for construction to be slightly greater than those of coal-fired power stations.

If nuclear power is to make a big contribution to the reduction of CO2 emissions within a few decades, many new nuclear power stations must be built globally every year. For simplicity, let's consider the hypothetical example in which a country or group of countries embarks upon a crash program to build one new nuclear power station per year until nuclear power has replaced half its coal-power in 20 years, and that the energy payback period of each nuclear power station is two years, which is optimistic. We assume that the power stations are identical with the same capacity, that high-grade uranium is available and that the energy inputs to stages other than the construction of the power station are negligible. We measure energy generation in units corresponding to the electricity generation by a single nuclear power station in one year. Then, in each of the first 20 years, there are two units of energy input and one unit of energy output. After Year 20, there are no energy inputs until high-grade uranium ore is used up. The figure above illustrates the cumulative energy inputs over 40 years, assuming there is sufficient high-grade uranium to last out this period.

In theory, it will take 40 years from the commencement of this rapid program of nuclear deployment to repay (in energy terms) the energy inputs to the program. At this time the lower curve overtakes the higher curve in the illustration. If, to first approximation, we assume that one year of CO2 inputs is equal to one year of CO2 emissions saved by the operation of the station, then the CO2 payback period is also 40 years. In practice, it is likely that high-grade uranium will be used up within the payback period and then the substantial energy and CO2 inputs from mining, milling and enrichment will replace those from construction and the CO2 debt may never be repaid.

Fortunately many improvements in efficiency of energy use and several renewable energy technologies have very short energy payback periods: e.g. wind power and solar PV modules based on thin films or slivers of silicon have payback periods of several months.


Nuclear energy is caught in a double bind. On one hand, if it is expanded slowly, it will run out of high-grade uranium ore within several decades. There are vast resources of low-grade uranium ore, but production of nuclear fuel from low-grade uranium ore produces quite high CO2 emissions. On the other hand, if nuclear energy is expanded rapidly, the CO2 emissions from construction could push the planet over the edge into irreversible climate change before the CO2 debt is repaid. Therefore, new nuclear power stations based on existing technologies cannot substitute significantly for CO2 emissions from coal-fired power stations.

To make matters worse, new nuclear electricity is more expensive than wind power and some forms of bioenergy in the UK, the USA and most other countries.

The suitability of nuclear energy as a means of reducing CO2 emissions could be re-examined when and if new nuclear technologies are introduced that are lower in life-cycle CO2 emissions, safer and less expensive.

Mark Diesendorf is a Senior Lecturer at the Institute of Environmental Studies at the University of New South Wales.


1 Van Leeuwen, Jan Willem Storm and Smith, Philip (2005), Can nuclear power provide energy for the future; would it solve the CO2-emission problem?