Decommissioning Nuclear Facilities

(Updated May 2022)

  • About 200 commercial, experimental or prototype reactors, over 500 research reactors, and several fuel cycle facilities have been retired from operation. Some of these have been fully dismantled.
  • Most parts of a nuclear power plant do not become radioactive, or are contaminated at only very low levels. Most of the metal can be recycled.
  • Proven techniques and equipment are available to dismantle nuclear facilities safely and these have now been well demonstrated in several parts of the world.
  • Decommissioning costs for nuclear power plants, including disposal of associated wastes, are high relative to other industrial plants but are reducing, and contribute only a small fraction of the total cost of electricity generation.

All power plants, coal, gas and nuclear, have a finite life beyond which it is not economically feasible to operate them. Generally speaking, early nuclear plants were designed for a life of about 30 years, though with refurbishment, some have proved capable of continuing well beyond this. Newer plants are designed for a 40 to 60 year operating life. At the end of the life of any power plant, it needs to be decommissioned, cleaned up and demolished so that the site is made available for other uses.

For nuclear plants, the term decommissioning includes all clean-up of radioactivity and progressive dismantling of the plant. This may start with the owner's decision to write it off or declare that it is permanently removed from operation. For practical purposes it includes defueling and removal of coolant, though NRC at least defines it as strictly beginning only after fuel and coolant are removed. It concludes with licence termination after decontamination is verified and wastes removed.

A Table showing shutdown reactors is at the end of this paper. About 25 of these have completed the decommissioning process.

Decommissioning options for nuclear plants

The International Atomic Energy Agency (IAEA) has defined three options for decommissioning, the definitions of which have been internationally adopted:

  • Immediate Dismantling (or Early Site Release/'Decon' in the USA): This option allows for the facility to be removed from regulatory control relatively soon after shutdown or termination of regulated activities. Final dismantling or decontamination activities can begin within a few months, depending on the facility. Following removal from regulatory control, the site is then available for re-use within a decade.
  • Safe Enclosure ('Safstor') or deferred dismantling: This option postpones the final removal of controls for a longer period, usually in the order of 40 to 60 years. The facility is placed into a safe storage configuration until the eventual dismantling and decontamination activities occur after resudual radioactivity has decayed. There is a risk in this case of regulatory change which could increase costs unpredictably.
  • Entombment (or 'Entomb'): This option entails placing the facility into a condition that will allow the remaining on-site radioactive material to remain on-site without ever removing it totally. This option usually involves reducing the size of the area where the radioactive material is located and then encasing the facility in a long-lived structure such as concrete, that will last for a period of time to ensure the remaining radioactivity is no longer of concern.

Each approach has its benefits and disadvantages. National policy determines which approach or combination of approaches is adopted or allowed. In the case of immediate dismantling (or early site release), responsibility for completion of decommissioning is not transferred to future generations. The experience and skills of operating staff can also be utilised during the decommissioning programme, which may be undertaken by the utility or handed over to a specialist company, with transfer of licence and accumulated funds. Alternatively, Safe Enclosure (or Safstor) allows significant reduction in residual radioactivity, thus reducing radiation hazard during the eventual dismantling. The expected improvements in mechanical techniques should also lead to a reduction in the hazard and also costs.

In the case of nuclear reactors, about 99% of the radioactivity is associated with the fuel which is removed following permanent shutdown. Apart from some surface contamination of plant, the remaining radioactivity comes from "activation products" in steel which has long been exposed to neutron irradiation, notably the reactor pressure vessel. Stable atoms are changed into different isotopes such as iron-55, iron-59 and zinc-65. Several are highly radioactive, emitting gamma rays. However, their half life is such (2.7 years, 45 days, 5.3 years, 245 days respectively) after 50 years from closedown their radioactivity is much diminished and the occupational risk to workers largely gone.

Decommissioning experience

Considerable experience has been gained in decommissioning various types of nuclear facility. About 200 commercial, experimental or prototype power reactors, as well as over 500 research reactors have been retired from operation. About 25 reactors have been fully dismantled.

(Ships and numerous submarines have also been decommissioned but are not included in this paper.)

European reactors

To decommission its retired gas-cooled reactors at the Chinon, Bugey and St Laurent nuclear power stations, EDF chose partial dismantling and postponed final dismantling and demolition for 50 years. As other reactors will continue to operate at those sites, monitoring and surveillance do not add to the cost.

A recycling plant for steel from dismantled nuclear facilities is at Marcoule, in France. This metal will contain some activation products, but it can be recycled for other nuclear plants.

Decommissioning has begun at 33 UK reactors, 26 of them early Magnox types with graphite moderators.* One of the first was Berkeley nuclear power station (2 x 138 MWe), closed for economic reasons in 1989 after 27 years of operation, where defuelling was completed in 1992. The cooling ponds were then drained, cleaned and filled in and the turbine hall was dismantled and demolished. The reactor buildings are in an extended Safstor period. Ultimately they too will be dismantled, leaving the site to be leveled and landscaped. The same pattern is being followed at other UK reactor sites.

* Costs for decommissioning gas-cooled reactors are much higher per unit of capacity than for light water reactors – at least five times for Magnox. This is due to the large volume of material and the need to dispose of a lot of graphite moderator. Decommissioning waste volumes per unit capacity for Magnox are ten times those for Western light water reactors. See also later subsection on Graphite.

Spain's Vandellos 1, a 480 MWe gas-graphite reactor, was closed down in 1990 after 18 years operation, due to a turbine fire which made the plant uneconomic to repair. In 2003 ENRESA concluded phase 2 of the reactor decommissioning and dismantling project, which allowed much of the site to be released. After 30 years Safstor, when activity levels have diminished by 95%, the remainder of the plant will be removed. The cost of the 63-month project was €93 million.

Eleven of Germany’s 19 shutdown units were subject to immediate dismantling. At Greifswald nuclear power station in the former East Germany, where five reactors had been operating, immediate dismantling was chosen. Similarly, the site of the 100 MWe Niederaichbach nuclear power plant in Bavaria was declared fit for unrestricted agricultural use in mid-1995. The 15 MWe Kahl experimental BWR was shut down in 1985, after 25 years of operation. After decontamination, the plant was completely dismantled and the site was cleared for unrestricted use. The 250 MWe Gundremmingen A unit was Germany's first commercial nuclear reactor, operating 1966-77. Decommissioning work started in 1983, and moved to the more contaminated parts in 1990, using underwater cutting techniques. This project demonstrated that decommissioning could be undertaken safely and economically without long delays, and recycling most of the metal.

Of the eight German units shut down in March 2011 for political reasons, most will be dismantled over about 15 years. The four operators had €38 billion set aside for decommissioning and waste disposal.

Japan's Tokai 1 reactor, a 160 MWe UK Magnox design, is being decommissioned after 32 years service to 1998. After 10 years storage, in Phase 2 (to 2011) the steam generators and turbines were removed, and in Phase 3 (to 2018) the reactor is being dismantled, the buildings demolished and the site left ready for re-use. The total cost will be JPY 93 billion (USD 1.04 billion) – 35 billion for dismantling and 58 billion for waste treatment which will include the graphite moderator (which escalates the cost significantly, see subsection below).

US reactors

Experience in the USA has varied, but about ten power reactors are using the Safstor approach, while about 20 – mostly single-unit plants – are using, or have used, Decon.

Early in 2017 a joint venture, Accelerated Decommissioning Partners, was set up under the leadership of NorthStar Group Services with Orano USA to acquire and decommission shutdown US nuclear reactor facilities and take over their used fuel. Another joint venture in the field is Comprehensive Decommissioning International (CDI), formed in 2018 between SNC-Lavalin and Holtec. Holtec Decommissioning International (HDI) functions as the licensed operator for Holtec-owned nuclear power plants and provides the licensee oversight of CDI. HDI additionally, manages the decommissioning trust fund and other owner interests, such as licensing strategy, insurance, land and government interface. EnergySolutions was already active in this kind of role.

Procedures are set by the Nuclear Regulatory Commission (NRC), and considerable experience has now been gained. A total of 32 power reactors have been closed and decommissioned. NRC requires that the operating licence of a closed reactor be terminated and decommissioning activities be completed within 60 years. Site release often excludes the onsite used fuel storage in an ISFSI (independent spent fuel storage installation), which usually remains, to await the Department of Energy taking away the used fuel (over which it has title) to a national repository sometime in the future.

Rancho Seco (single 913 MWe, PWR) was closed in 1989, and in 1995 NRC approved a Safstor plan for it. However, the utility subsequently decided upon incremental dismantling and this was completed in 2009, leaving about 3 ha still under NRC licence for waste storage. About 32 ha has been released for unrestricted use.

At multi-unit nuclear power stations, the choice has been to place the first closed unit into Safstor until the others end their operating lives, so that all can be decommissioned in sequence. This will optimise the use of staff and the specialised equipment required for cutting and remote operations, and achieve cost benefits.

Thus, after 14 years of comprehensive clean-up activities, including the removal of fuel, debris and water from the 1979 accident, Three Mile Island 2 was placed in post-defuelling monitored storage (Safstor) until unit 1 eventually closed down, so that both units could be dismantled together. EnergySolutions has a contract with FirstEnergy to decommission unit 2. Unit 1 is being placed in Safstor.

San Onofre 1 (436 MWe PWR), which closed in 1992, was put into Safstor until licences for units 2&3 expire in 2022-23. However, after NRC changes, dismantling was brought forward to 1999, so it became an active Decon project which was largely completed in 2008. A small amount of work remains to be completed with eventual dismantling of units 2&3 (each 1070 MWe PWR) on the site, which were shut down in May 2013. In April 2016 the California Public Utilities Commission approved $4.41 billion in decommissioning costs for them, from funds held in trust. Southern California Edison said that funds were then $3.37 billion. Used fuel will be removed from 2024 to an onsite ISFSI, and dismantling the plant is expected to be completed by 2030. SONGS Decommissioning Solutions – a joint venture between Aecom and EnergySolutions – was selected in December 2016 as the general contractor.

A US Decon project was the 60 MWe Shippingport reactor, which operated commercially from 1957 to 1982. It was used to demonstrate the safe and cost-effective dismantling of a commercial scale nuclear power plant and the early release of the site. Defuelling was completed in two years, and five years later the site was released for use without any restrictions. Because of its size, the pressure vessel could be removed and disposed of intact. For larger units, such components have to be cut up.

Immediate Decon also the option chosen for Fort St Vrain, a 330 MWe high temperature gas-cooled reactor which was closed in 1989. This took place on a fixed-price contract for US$ 195 million (hence costing less than 1 cent/kWh despite only a 16-year operating life) and the project proceeded on schedule to clear the site and relinquish its licence early in 1997 – the first US power reactor of this size to achieve this.

Shoreham BWR, on Long Island, generated very little power and never received a full operating licence, so the level of activation products was minimal. It was shut down in 1989 and became a Decon project, completed in 1994. The 59 MWe Pathfinder prototype BWR in South Dakota, shut down in 1967 after a very short life was also a Decon project, completed in 1992.

For Trojan (1180 MWe, PWR) in Oregon the dismantling was undertaken by the utility itself. The plant closed in 1993, steam generators were removed, transported and disposed of at Hanford in 1995, and the reactor vessel (with internals) was removed and transported to Hanford in 1999. Except for the used fuel storage, the site was released for unrestricted use in 2005. The cooling tower was demolished in 2006. This was a relatively low-cost operation – about $300 million.

Another US Decon project was Maine Yankee, a 860 MWe PWR plant which closed down in 1996 after 24 years operation. The containment structure was finally demolished in 2004 and except for 5 ha with the dry store for used fuel, the site was released for unrestricted public use in 2005, on budget (about $500 million) and on schedule.

Haddam Neck/Connecticut Yankee (560 MWe PWR) also shut down in 1996 after 29 years operation. Decommissioning work began in 1998 and demolition was concluded in 2006. The site was released for unrestricted public use in 2007, apart from 2 ha for dry cask used fuel storage. Residual contamination on the land is below the NRC's limit of 0.25 millisievert per year for maximum radiation dose.

In 2006 the site of 72 MWe Big Rock Point nuclear power plant in Michigan, shut down in 1997 after 35 years operation, was largely returned to greenfield status. In January 2007 most of the land was released for derestricted public use, though 43 hectares still has the dry cask storage facility where used fuel is stored pending transfer to the national repository. The total cost was $836 million.

Vermont Yankee (535 MWe BWR) was shut down in 2014 and is in Safstor and was expected to cost $577 million to decommission over many years. However, Entergy arranged to accelerate the process by transferring the licence and selling the whole site to a consortium of decommissioning companies headed by NorthStar Group Services, which expects to complete the process and sell the cleared site by 2030. Fuel was moved to dry storage in 2017. Areva Nuclear Materials, now Orano USA, is contracted to cut up and remove the reactor vessel and internals.

With Exelon's Zion 1&2 reactors (2 x 1098 MWe) closed down in 1998 and initially in Safstor, Exelon decided to accelerate the process and contracted a specialist company – EnergySolutions – to dismantle the plant, ship the waste to its disposal site in Utah, and return the site to greenfield status. To achieve this, in 2010 the plant's licence and decommissioning funds were transferred to EnergySolutions, which became owner and licensee, and the site should be returned to Exelon about 2022. In 12 months to January 2015 EnergySolutions transferred all 2226 used fuel assemblies to 61 Magnastor dry casks onsite – an ISFSI, where it will remain until taken to the future national repository. By September 2021 the site was clean and being vegetated. EnergySolutions had similarly decommissioned LaCrosse, a 50 MWe BWR, Big Rock Point, a 67 MWe BWR, and the small (20 MWt) SEFOR experimental reactor in Arkansas. It is signed up to decommission Three Mile Island 2, San Onofre (see above) and Fort Calhoun.

Exelon's Oyster Creek (619 MWe BWR) in New Jersey will be dismantled and cleaned up by Comprehensive Decommissioning International (CDI) – the SNC-Lavalin and Holtec JV – starting in 2019 and expected to take less than ten years. The licence was transferred from Exelon in July 2019. CDI will also be the decommissioning general contractor for the Pilgrim BWR shut down in 2019, and for Indian Point 1, 2&3 (1974, 2020, 2021) and Palisades from 2022.

Entergy’s Indian Point plant is expected to cost $2.3 billion to decommission and dismantle over about 12-15 years, by CDI after ownership passed to Holtec in April 2021. Unit 1 of 277 MWe, in Safstor since 1974, would cost $598 million, unit 2 $702 million and unit 3 $1 billion. Decommissioning trust funds for the three reactors amounted to $2.1 billion in 2019 so were expected to cover the work by the time unit 3 shut down in 2021. Only an ISFSI would remain onsite after 2032.

Duke’s Crystal River 3 (860 MWe) was expected to cost $1.18 billion (2013 dollars) to decommission via Safstor over 60 years, during which time the funds reserved for the purpose would accrue interest, thereby fully covering the cost, despite the fact that is was closed after only 35 years of operation. Immediate decommissioning (Decon) was then expected to cost $994 million, but the decommissioning fund would have had less time to grow sufficiently to cover it, and a $195 million impact on Florida ratepayers would have resulted.

Construction of an onsite dry cask storage facility commenced in 2016, and fuel transfer started in summer 2017 and completed in January 2018. The spent fuel is planned to remain at the onsite storage facility until 2036, when it would be moved to a federal facility. Safstor was expected to end with removal of the unit’s remaining components about 2070 and site restoration in 2074. In August 2020 the Florida Public Service commission approved Accelerated Decommissioning Partners completing demolition by 2027 and managing the ISFSI to 2038 for a fixed price of $540 million, fully covered by the existing trust funds of over $700 million (with balance plus any money recovered from suing the DOE returned to ratepayers). Duke will remain the owner to 2038, but Accelerated Decommissioning Partners became the licensed operator of the plant and ISFSI in October 2020.

However, in May 2019 Duke announced that it had contracted with Accelerated Decommissioning Partners (NorthStar and Orano joint venture) for decommissioning Crystal River by 2027, and that this would be covered by the decommissioning fund, which was worth $717 million in the first quarter of 2019. The contract structure transfers the risk of potential cost overruns and latent liabilities to the joint venture. Duke would remain the NRC-licensed owner of the plant, as well as the owner of the land and the trust fund. However, Accelerated Decommissioning Partners would become the plant operator and would take ownership of the associated dry cask storage facility as well as spent nuclear fuel onsite.

Omaha Public Power District’s Fort Calhoun plant (482 MWe) closed in 2016 after 43 years operation, since 2012 by Exelon. It is being decommissioned by EnergySolutions from 2019. Safstor was originally intended at a cost of $1.3 billion, but following an OPPD board decision in October 2018 and a contract with EnergySolutions it will now be cleaned up by mid-2020s with an expected saving of $200 million.

In summary, US plants with Decon completed are: Big Rock Point, Elk River, Fort St Vrain, Haddam Neck/Connecticut Yankee, La Crosse, Maine Yankee, Pathfinder, Rancho Seco, San Onofre 1, Saxton, Shippingport, Shoreham, Trojan, Zion 1&2 and Yankee Rowe. (Also Santa Susana Field Laboratory's Sodium Reactor Experiment, which is not included in power reactor tables.) Decon is in progress at Humboldt Bay 3, and will now be adopted at Fort Calhoun and Indian Point.

Yankee-Rowe-after-decommissioning-_12_12_2006_2_lg-(Yankee-Rowe).jpg

Yankee Rowe after decommissioning (Yankee Rowe)

US plants in Safstor, either in part or in full, include Crystal River 3, Dresden 1, Fermi 1, Indian Point 1, Kewaunee, Millstone 1, and Peach Bottom 1, as well as NS Savannah. Three Mile Island 2 is in post-defuelling monitored storage. San Onofre 2&3 will also enter Safstor when defuelled.

The only US plants subject to the Entomb option are small experimental ones: Bonus BWR in Puerto Rico, Piqua organic-moderated reactor in Ohio, Hallam graphite-moderated sodium-cooled reactor in Nebraska, and in 2015, EBR-II. No NRC-licensed plant has used this option.

In addition to the above is the first floating nuclear power plant, installed in a former liberty ship, and utilised at Panama 1967-76. The Sturgis had a 45 MWt/10 MWe (net) PWR which provided power to the Canal Zone. After it was defuelled in 1977, some 89 m3 of solid radioactive waste and 363 m3 of liquid waste was removed and the vessel placed in safstor at Fort Belvoir, Virginia. In 2014 CB&I was awarded a contract from the US Army to decommission and dismantle the MH-1A PWR. In 2015 the vessel was towed to Texas for dismantling, which was completed in 2019. About 700 tonnes of radioactive waste was removed and 270 tonnes of lead was recycled.

The US Army's SM-1A reactor at Fort Greely in Alaska was shut down in 1972 after ten years' operation. Most of the radioactive waste was then removed and primary components were sealed inside the containment with concrete to Safstor status. Further work is planned to 2022. It operated at 1.8 MWe and 10 MW thermal. The Army's SM-1 at Fort Belvoir, Virginia, was similar and shut down in 1973. It is in Safstor and further work is planned from 2020.

Further information on decommissioning in USA is available from the Nuclear Energy Institute, and from the NRC.

Russian reactors

Rostechnadzor oversees a major program of decommissioning old fuel cycle facilities, financed under the Federal target program on Nuclear and Radiation Safety. Six civil reactors are being decommissioned: three early LWGRs, the Melekess VK-50 prototype BWR, and two larger prototype VVER-440 units at Novovoronezh. Most were shut down 1981-90, the fuel removed, and they now await dismantling, which is staring to progress at Novovronezh.

Russia's Rosenergoatom has concluded that moving from the 'deferred' to 'immediate' dismantling approach to its nuclear reactors being decommissioned would bring a 20% cost saving. Immediate dismantling enables the maximum use of the residual life of the equipment and structures of the shutdown units, reduces maintenance costs, makes use of existing radioactive waste management facilities, and employs the skills of personnel. 

Graphite

A number of first-generation reactors had graphite blocks as moderator. This is high-quality reactor-grade material produced at about 3000°C which accumulates some radionuclide contamination while in service, particularly carbon-14 at levels which often means that it must be classified as intermediate-level waste. While it oxidizes slightly under neutron bombardment and also at high temperatures (to CO), it does not burn, but sublimes at 3652°C. There is no risk of oxidation under Safstor conditions.

A 2006 report commissioned by EPRI states: "The graphite moderators of retired gas-cooled nuclear reactors present a difficult challenge during demolition activities. As a result, utilities have not dismantled any moderators of CO2 cooled power reactors to date." However, it concludes that adequate information exists to enable the safe dismantling and processing of graphite moderators, and that the three main options for disposal of this graphite are oxidation to the gas phase and release as carbon dioxide (difficult), direct burial, or recycling into new products for the nuclear industry. In each case, opportunities exist for pre-processing to concentrate or remove radionuclides to enhance the safety of the chosen option. The radionuclide inventory of irradiated graphite is unusual in comparison with other nuclear wastes. Cobalt-60 and tritium are the principal isotopes of short-term importance, carbon-14 and chlorine-36 are dominant in the longer term.

Fast neutron reactors

Several sodium-cooled fast reactors have been decommissioned, but only a few have been dismantled. Germany’s KNK-2 at Karlsruhe was shut down in 1991 after 14 years' operation, with fuel removal and initial dismantling in 1993. High levels of activation products meant that much work was remote, and residual sodium meant that some cutting was done in a nitrogen atmosphere. Total cost is estimated as €364 million, with completion expected in 2020.

Other facilities

The French Atomic Energy Commission is decommissioning the UP1 reprocessing plant at Marcoule. This plant started up in 1958 and treated 18,600 tonnes of metal fuels from gas-cooled reactors (both defence and civil) to 1997. Progressive decontamination and dismantling of the plant and waste treatment will span 40 years and cost some EUR 5.6 billion, nearly half of this for treatment of the wastes stored on the site.

Areva is decommissioning the Eurodif enrichment plant at Marcoule since 2012. This involved over 2012-15 the decontamination with ClF3 gas to remove the residual uranium left inside, and extract it as UF6, then recovery of all produced chloride and fluoride gas before the opening of equipments and circuits. Then over 2016-25 the plant is being dismantled.

Areva completed the dismantling and clean-up of the MOX fuel fabrication plant and plutonium technology workshops (ATPu) at CEA’s Cadarache site in 2017. The small fuel fabrication plant had produced fuel for French fast reactors and was closed down in 2003; the ATPu was closed in 2008. Areva described the project as "one of the largest dismantling projects in the world."

The US Department of Energy in August 2010 awarded a $2.1 billion contract to a joint venture between Fluor Corp and Babcock and Wilcox (B&W, now BWXT) for decontamination and decommissioning of the huge (1500 ha) Portsmouth Gaseous Diffusion Plant (GDP) uranium enrichment site in Ohio, from March 2011. In March 2016 the contract was extended by 30 months.

The US Department of Energy in May 2017 awarded a $1.5 billion contract to Four Rivers Nuclear Partnership LLC for the continued deactivation and remediation (D&R) of facilities at the Paducah gaseous diffusion uranium enrichment plant in Kentucky and clean-up of the 1400 ha site. Four Rivers is a CH2M-led company with partners Fluor and BWX Technologies. The Paducah plant was commissioned in 1952 for defence purposes and was leased to USEC from 1993 to 2013. The D&R contract is valued at $750 million for five years, followed by three-year and two-year option periods together worth about $750 million. Fluor had a three-year $420 million DOE contract to clean up the Paducah site from 2014-17.

The US Department of Energy contractors in 2016 finished demolishing the Oak Ridge diffusion enrichment plant in Tennessee, which operated from the early 1940s to 1985, making 120 ha available for other uses.

Many nuclear submarines have been decommissioned over the last decade. In USA, after defuelling, the reactor compartments are cut out of the vessels and are transported inland to Hanford, where they are buried as low-level waste. Russia has decommissioned three nuclear powered icebreakers: Lenin, Sibir and Arktika.

Cost and finance

In most countries the operator or owner is responsible for the decommissioning costs.

The total cost of decommissioning is dependent on the sequence and timing of the various stages of the program. Deferment of a stage tends to reduce its cost, due to decreasing radioactivity, but this may be offset by increased storage and surveillance costs.

Even allowing for uncertainties in cost estimates and applicable discount rates, decommissioning contributes a small fraction of total electricity generation costs. In USA many utilities have revised their cost projections downwards in the light of experience.

Financing methods vary from country to country. Among the most common are:
Prepayment, where money is deposited in a separate account to cover decommissioning costs even before the plant begins operation. This may be done in a number of ways but the funds cannot be withdrawn other than for decommissioning purposes.
External sinking fund (Nuclear Power Levy): This is built up over the years from a percentage of the electricity rates charged to consumers. Proceeds are placed in a trust fund outside the utility's control. This is the main US system, where sufficient funds are set aside during the reactor's operatinig lifetime to cover the cost of decommissioning.
Surety fund, letter of credit, or insurance purchased by the utility to guarantee that decommissioning costs will be covered even if the utility defaults.

In the USA, utilities are collecting 0.1 to 0.2 cents/kWh to fund decommissioning. They must then report regularly to the NRC on the status of their decommissioning funds. About two-thirds of the total estimated cost of decommissioning all US nuclear power reactors has already been collected, leaving a liability of about $9 billion to be covered over the remaining operating lives of about 100 reactors (on the basis of an average of $320 million per unit). NRC data for the end of 2018 indicated that there was a combined total of $64.7 billion held in the decommissioning trust funds covering the 119 operational and retired US nuclear power reactors.

An OECD Nuclear Energy Agency survey published in 2016 reported US dollar (2013) costs in response to a wide survey. For US reactors the expected total decommissioning costs range from $544 to $821 million; for units over 1100 MWe the costs ranged from $0.46 to $0.73 million per MWe, for units half that size, costs ranged from $1.07 to $1.22 million per MWe. For Finland’s Loviisa (2 x 502 MWe) the estimate was €326 million. For a Swiss 1000 MWe PWR the detailed estimate amounts to CHF 663 million (€617 million). In Slovakia, a detailed case study showed a total cost of €1.14 billion to decommission Bohunice V1 (2 x 440 MWe) and dismantle it by 2025.

Recycling and reuse of materials from decommissioning

Significant volumes of concrete and steel arise from demolishing decommissioned nuclear facilities. Much of this can be recycled or reused in some way where allowed by national regulation, and unlike ‘wastes’, transferred to other countries. Clearance levels differ internationally, leaving scope for greater harmonisation, and some levels are unjustifiably conservative, e.g. 100 Bq/kg applying for many radionuclides. Public acceptance of recycling and reuse is often low.

There are several options for recycle and reuse:

  • Material which is essentially uncontaminated and unconditionally released.
  • Material that can be melted in a regulated environment followed by metal recycle for consumer products (conditional clearance).
  • Material with short half-life products that is melted and fabricated in a regulated environment and released for specific industrial applications (e.g. steel bridge).
  • Material that cannot be released from regulatory control but which may be recycled in the nuclear industry.

Recycling materials from decommissioned nuclear facilities is constrained by the level of radioactivity in them. This is also true for materials from elsewhere, such as gas plants, but the levels specified can be very different. For example, scrap steel from gas plants may be recycled if it has less than 500,000 Bq/kg (0.5 MBq/kg) radioactivity (the exemption level). This level however is one thousand times higher than the clearance level for recycled material (both steel and concrete) from the nuclear industry, where generally anything above 500 Bq/kg may not be cleared from regulatory control for recycling.*

* EC proposed clearance levels for specific nuclides in 1996 were 10 MBq/kg for Fe-55 and Ni-63. Those for the IAEA were 0.3 and 3 MBq/kg respectively. For mixtures of artificial radionuclides, the weighted sum of the nuclide specific activities or concentrations in the same matrix divided by the corresponding release limit must be applied. 

There is increasing concern about double standards developing in Europe which allow 30 times the dose rate from non-nuclear recycled materials than from those out of the nuclear industry. Norway and Holland are the only countries with consistent standards. Elsewhere, 0.3 to 1.0 mSv/yr individual dose constraint is applied to oil and gas recyclables, and 0.01 mSv/yr for release of materials with the same kind of radiation from the nuclear industry. The double standard means that the same radionuclide, at the same concentration, can either be sent to deep disposal or released for use in building materials, depending on where it comes from. The 0.3 mSv/yr dose limit is still only one-tenth of most natural background levels, and two orders of magnitude lower than those experienced naturally by many people, who suffer no apparent ill effects.

The main radionuclide in scrap from the oil and gas industry is radium-226, with a half-life of 1600 years as it decays to radon. Those in nuclear industry scrap are cobalt-60 and caesium-137, with much shorter half-lives. Application of a 0.3 mSv/yr dose limit results in a clearance level for Ra-226 of 500 Bq/kg, compared with 10 Bq/kg for nuclear material.

In 2011, 16 decommissioned steam generators from Bruce Power in Canada were to be shipped to Sweden for recycling. Although the Canadian Nuclear Safety Commission (CNSC) approved Bruce Power’s plans in 2011 and confirmed steam generator processing is an excellent example of responsible and safe nuclear waste management practices, this caused public controversy at the time, and following the Fukushima nuclear accident plans for this shipment were shelved. These steam generators were each 12m long and 2.5m diameter, with mass 100 t, and contained some 4g of radionuclides with about 340 GBq of activity. Exposure was 0.08 mSv/hr at one metre. They were classified as low-level waste (LLW). Studsvik in Sweden would recycle much of the metal and return about 10% of the overall volume as LLW for disposal in Ontario. The balance would be under 100 Bq/kg, which appeared to be the clearance level.

In 2012 five steam generators from UK plants were shipped to Studsvik in Sweden for recycling. Studsvik has also set up a plant in the UK, at Lillyhall in Cumbria, to recycle materials from nuclear facilities, and this became fully operational in 2013 after processing 2000 t of metal from numerous sites and recycling 96% of it. The balance went to a LLW repository.

To 2015, Studsvik had processed by melting 32,000 tonnes of carbon steel, 5200 t stainless steel, 2033 t aluminium, 1153 t lead, and 3896 t copper cables. All this could be released. EnergySolutions in the USA has recycled over 62,000 tonnes of metal, mostly ferrous.

Under the NEA CPD program (see below) a 2017 report on Recycling and Reuse of Materials Arising from the Decommissioning of Nuclear Facilities has been published.

International cooperation

The IAEA, the OECD's Nuclear Energy Agency and the Commission of the European Communities are among a number of organisations through which experience and knowledge about decommissioning is shared among technical communities in various countries.

In 1985, the OECD Nuclear Energy Agency launched the International Co-operative Program for the Exchange of Scientific and Technical Information Concerning Nuclear Installation Decommissioning Projects, now known as the Co-operative Program on Decommissioning (CPD). This international collaboration has produced a great deal of technical and financial information. Initially consisting of 10 decommissioning projects in eight countries, the programme has since grown to 70 projects (40 reactors and 30 fuel cycle facilities) in 14 NEA member countries, one non-member economy, and the European Commission. The current agreement runs to 2018.

In 2013, following the Fukushima accident, an International Research Institute for Nuclear Decommissioning (IRID) was established, based in Japan. As well as bringing together knowledge and experience from wrecked reactors, IRID will build up the knowledge base for routine decommissioning.

In January 2015 the IAEA launched an International Project on Decommissioning and Remediation of Damaged Nuclear Facilities, the DAROD project, to help increase nuclear safety under the agency’s Nuclear Safety Action Plan that was unanimously adopted by IAEA member states following the 2011 Fukushima accident. The project will run for three years, and involves 35 international experts from 19 IAEA member states.

The important areas where experience is being gained and shared are the assessment of the radioactive inventories, decontamination methods, cutting techniques, remote operation, radioactive waste management and health and safety. The aims are to minimise the radiological hazards to workers and to optimise the dismantling sequence and timing to reduce the total decommissioning cost.

Reasons for shutdown

Most decommissioned reactors were shut down because there was no longer any economic justification for running them. Practically all are relatively early-model designs, and about 45 are experimental or prototype power reactors. Three categories are listed here:

  1. Experimental, early commercial types and commercial units whose continued operation was no longer justified, usually for economic reasons. Most of these started up before 1980 and their short operating lifetime is not surprising for the first couple of decades of a major new technology. Many of these (marked with an asterisk in the Table) ran relatively full-term, for design lifetimes of 25-35 years or so (design lifetimes today are 40-60 years).
  2. Units which closed following an accident or serious incident (not necessarily to the reactor itself) which meant that repair was not economically justified.
  3. Units which were closed prematurely by political decision or due to regulatory impediment without clear or significant economic or technical justification.

In fact the distinctions are not always sharp, e.g. Chernobyl 2 was closed in 1991 after a turbine fire when it would have been politically impossible to repair and restart it, Rheinsberg was closed in 1990 though it was nearly at the end of its design lifetime – both these are in the 'political decision' category.

Reactors closed following damage in an accident or serious incident (12)

Location Reactor Type MWe net Years operable Shutdown reason
Germany Greifswald 5 VVER-440/V-213 408 0.5 11/1989 Partial core melt
  Gundremmingen A BWR 237 10 1/1977 Botched shutdown
Japan Fukushima Daiichi 1 BWR 439 40 3/2011 Core melt from cooling loss
  Fukushima Daiichi 2 BWR 760 37 3/2011 Core melt from cooling loss
  Fukushima Daiichi 3 BWR 760 35 3/2011 Core melt from cooling loss
  Fukushima Daiichi 4 BWR 760 32 3/2011 Damage from hydrogen explosion
  Monju Prot FNR 246 1 2016 Sodium leak
Slovakia Bohunice A1 Prot GCHWR 93 4 1977 Core damage from fuelling error
Spain Vandellos 1 GCR 480 18 mid-1990 Turbine fire
Switzerland St Lucens Exp GCHWR 6 3 1966 Core melt
Ukraine Chernobyl 4 RBMK LWGR 925 2 4/1986 Fire and meltdown
USA Three Mile Island 2 PWR 880 1 3/1979 Partial core melt

Reactors closed prematurely by political decision or consideration (46)

Location Reactors Type MWe net each Years operable each Shutdown
Armenia Metsamor 1 VVER-440/V-270 376 13 1989
Bulgaria Kozloduy 1-2 VVER-440/V-230 408 27, 28 12/2002
  Kozloduy 3-4 VVER-440/V-230 408 24, 26 12/2006
France Super Phenix FNR 1200 12 1999
Germany Greifswald 1-4 VVER-440/V-230 408 10, 12, 15, 16 1990
  Muelheim-Kaerlich PWR 1219 2 1988
  Rheinsberg VVER-70/V-210 62 24 1990
  Biblis A* PWR 1167 36 2011
  Biblis B* PWR 1240 34 2011
  Brunsbüttel* BWR 771 30 2007
  Grafenrheinfeld* PWR 1275 33 2015
  Gundremmingem B* BWR 1284 33 2017
  Isar 1* BWR 878 32 2011
  Krümmel BWR 1346 25 2009
  Neckarwestheim 1* PWR 785 35 2011
  Phillipsburg 1* BWR 890 31 2011
  Phillipsburg 2* PWR 1392 35 2019
  Unterweser PWR 1345 32 2011
  Grohnde PWR 1360 37 2021
  Brokdorf PWR 1410 35 2021
  Gundremmingen C BWR 1288 37 2021
Italy Caorso BWR 860 12 1986
  Latina GCR 153 24 1987
  Trino PWR 260 25 1987
Japan Fukushima Daiichi 5 BWR 760 33 2011
  Fukushima Daiichi 6 BWR 1067 32 2011
Lithuania Ignalina 1 RBMK LWGR 1185 21 2005
  Ignalina 2 RBMK LWGR 1185 22 2009
Slovakia Bohunice 1 VVER-440/V230 408 28 12/2006
  Bohunice 2 VVER-440/V230 408 28 12/2008
Sweden Barsebäck 1 BWR 600 24 11/1999
  Barsebäck 2 BWR 600 28 5/2005
Ukraine Chernobyl 1 RBMK LWGR 740 19 12/1997
  Chernobyl 2 RBMK LWGR 925 12 1991
  Chernobyl 3 RBMK LWGR 925 19 12/2000
USA Shoreham BWR 820 3 1989
  Three Mile Island 1 PWR 998 47 2020
  Indian Point 2 PWR 998 47 2020
  Duane Arnold BWR 601 45 2020
  Indian Point 3 PWR 1030 45 2021
  Palisades PWR 805 50 2022

Reactors closed having fulfilled their purpose or being no longer economic to run (139)

Location Reactor type MWe net each Start-up Years operable each Shutdown
Belgium BR-3 Prot PWR 10 1962 24 1987
Canada Douglas Point Prot PHWR 206 1967 17 1984
  Gentilly 1 Exp SGHWR 250 1971 6 1977
  Gentilly 2 * PHWR 635 1982 30 2012
  Rolphton NPD Prot PHWR 22 1962 25 1987
France Bugey 1 GCR 540 1972 22 1994
  Chinon A1 Prot GCR 70 1963 10 1973
  Chinon A2 GCR 180 1965 20 1985
  Chinon A3 * GCR 360 1965 25 1990
  Chooz A Prot PWR 305 1967 24 1991
  Brennilis EL-4 Exp GCHWR 70 1967 18 1985
  Fessenheim 1 * PWR 880 1977 43 2020
  Fessenheim 2 * PWR 880 1977 43 2020
  Marcoule G1 Prot GCR 2 1956 12 1968
  Marcoule G2 Prot GCR 39 1959 20 1980
  Marcoule G3 Prot GCR 40 1960 24 1984
  Phenix * FNR 130 1973 37 2010
  St Laurent A1 GCR 390 1969 21 1990
  St Laurent A2 GCR 465 1971 21 1992
Germany Juelich AVR Exp HTR 13 1968 21 1989
  Uentrop THTR Prot HTR 296 1985 3 1988
  Kalkar KNK 2 Prot FNR 17 1978 13 1991
  Kahl VAK Exp BWR 15 1961 24 1985
  MZFR Exp PHWR 52 1966 18 1984
  Groswelzheim Prot BWR 25 1969 2 1971
  Lingen Prot BWR 183 1968 10 1979
  Niederaichbach Exp GCHWR 100 1973 1 1974
  Obrigheim * PWR 340 1968 36 2005
  Stade * PWR 640 1972 31 2003
  Wuergassen BWR 640 1972 22 1994
Italy Garigliano BWR 150 1964 18 1982
Japan Fugen Prot ATR 148 1978 24 2003
  Genkai 1 * PWR 529 1975 40 2015
  Genkai 2 * PWR 529 1980 39 2019
  Hamaoka 1 BWR 515 1974 26 2001
  Hamaoka 2 BWR 806 1978 25 2004
  Ikata 1 * PWR 538 1977 39 2016
  Ikata 2 * PWR 538 1982 36 2018
  JPDR Prot BWR 12 1963 13 1976
  Mihama 1 * PWR 320 1970 45 2015
  Mihama 2 * PWR 470 1972 43 2015
  Ohi 1 * PWR 1120 1977 42 2019
  Ohi 2 * PWR 1120 1978 41 2019
  Onagawa 1 * BWR 498 1984 34 2018
  Shimane 1 BWR 439 1974 41 2015
  Tokai 1 * GCR 137 1965 33 1998
  Tsuruga 1 * BWR 340 1970 45 2015
Kazakhstan Aktau BN-350 Prot FNR 52 1973 27 1999
Netherlands Dodewaard * BWR 55 1968 28 1997
Pakistan Karachi 1 PHWR 90 1971 50 2021
Russia Bilibino 1 * LWGR 11 1974 45 2019
  Obninsk AM-1 Exp LWGR 5 1954 48 2002
  Beloyarsk 1 Prot LWGR 102 1964 19 1983
  Beloyarsk 2 Prot LWGR 146 1968 22 1990
  Kursk 1 RBMK 925 1976 45 2021
  Leningrad 2 * RBMK 925 1975 45 2020
  Melekess VK50 Prot BWR 50 1964 24 1988
  Novovoronezh 1 Prot VVER-440/V210 197 1964 23 1988
  Novovoronezh 2 Prot VVER-440/V365 336 1970 20 1990
  Novovoronezh 3 * Prot VVER-440/V179 385 1971 45 2016
South Korea Kori 1 * PWR 576 1977 40 2017
  Wolsong 1 * PHWR 661 1982 36 2019
Spain Garona * BWR 446 1971 42 2012
  Jose Cabrera * PWR 141 1968 38 2006
Sweden Agesta Prot HWR 10 1964 10 1974
  Oskarshamn 1 * BWR 473 1972 45 2017
  Oskarshamn 2 * BWR 638 1974 39 2013
  Ringhals 1 * PWR 881 1974 46 2020
  Ringhals 2 * PWR 910 1974 45 2019
Switzerland Muehleberg * BWR 373 1971 48 2019
UK Berkeley 1-2 * GCR 138 1962 26 1988-89
  Bradwell 1-2 * GCR 138 1962 39 2002
  Calder Hall 1-4 * GCR 49 1956-59 44-46 2003
  Chapelcross 1-4 * GCR 48 1959-60 44-45 2004
  Dungeness A 1-2 * GCR 225 1965 41 2006
  Dungeness B 1-2 * AGR 545 1983-85 36-38 2021
  Hinkley Point A 1-2 * GCR 235 1965 35 2000
  Hunterston A 1-2 * GCR 150 1964 25 1989-90
  Hunterston B 1 CGR 490 1967 54 2021
  Oldbury 1-2 * GCR 217 1967 44 2011-12
  Sizewell A 1-2 * GCR 210 1966 41 2006
  Trawsfynydd 1-2 * GCR 195 1965 26 1993
  Wylfa 1-2 * GCR 490 1971 44, 41 2015, 2012
  Windscale Prot AGR 24 1963 18 1981
  Dounreay DFR Exp FNR 11 1962 18 1977
  Dounreay PFR Prot FNR 234 1975 19 1994
  Winfrith Prot SGHWR 92 1968 23 1990
USA Big Rock Point* BWR 67 1962 35 1997
  BONUS Exp BWR 17 1964 4 1968
  CVTR Exp PHWR 17 1963 4 1967
  Crystal River PWR 860 1977 35 2013
  Dresden 1 BWR 197 1960 18 1978
  Duane Arnold * BWR 601 1975 45 2020
  Elk River BWR 22 1963 5 1968
  Enrico Fermi 1 Prot FNR 61 1966 6 1972
  Fort Calhoun * PWR 482 1973 43 2016
  Fort St. Vrain Prot HTR 330 1976 13 1989
  Haddam Neck/Connecticut Yankee* PWR 560 1967 29 1996
  Hallam Exp sodium cooled GR 75 1963 1 1964
  Humboldt Bay BWR 63 1963 13 1976
  Indian Point 1 PWR 257 1962 12 1974
  Indian Point 2 * PWR 998 1974 45 2020
  Indian Point 3 * PWR 1030 1976 44 2021
  Kewaunee* PWR 566 1974 39 2013
  La Crosse BWR 48 1968 19 1987
  Maine Yankee * PWR 860 1972 25 1997
  Millstone 1 BWR 641 1970 28 1998
  Oyster Creek * BWR 619 1969 49 2018
  Pathfinder Prot BWR 59 1966 1 1967
  Peach Bottom 1 Exp HTR 40 1967 7 1974
  Pilgrim BWR 677 1972 47 2019
  Piqua Exp Organic MR 12 1963 3 1966
  Rancho Seco 1 PWR 873 1974 15 1989
  San Onofre 1 * PWR 436 1967 25 1992
  San Onofre 2 * PWR 1070 1982 31 2013
  San Onofre 3 * PWR 1080 1983 30 2013
  Saxton Exp PWR 3 1967 5 1972
  Shippingport Prot PWR 60 1957 25 1982
  Three Mile Island 1 PWR 819 1974 45 2019
  Trojan PWR 1095 1975 17 1992
  Vallecitos Prot BWR 24 1957 6 1963
  Yankee NPS * PWR 167 1960 31 1991
  Zion 1 * PWR 1040 1973 25 1998
  Zion 2 * PWR 1040 1973 25 1998
  Sturgis FNPP PWR 10 1967 9 1976
Taiwan Chinshan 1 * BWR 604 1977 41 2018
  Chinshan 2 * BWR 604 1978 41 2019
  Kuosheng 1 * BWR 985 1981 40 2021

prot= prototype, exp= experimental, (total prototype + experimental = 44),  * = ran approx full-term


Notes & references

IMechE Conference Transactions, International Conference on Nuclear Decommissioning, ISBN 978-0852989555 (1995-7)
OECD Nuclear Energy Agency, Recycling and Reuse of Scrap Metals, A Report by a Task Group of the Co-operative Programme on Decommissioning (1996)
OECD Nuclear Energy Agency, Decommissioning Nuclear Power Plants: Policies, Strategies and Costs (2003)
OECD Nuclear Energy Agency, Decommissioning Funding: Ethics, Implementation, Uncertainties, A Status Report, NEA No. 5996 (2006)
OECD Nuclear Energy Agency, Costs of Decommissioning Nuclear Power Plants, NEA No. 7201 (2016)
OECD Nuclear Energy Agency, Recycling and Reuse of Materials Arising from the Decommissioning of Nuclear Facilities, NEA No. 7310 (2017)
Dennis Reisenweaver and Michele Laraia, Preparing for the End of the Line: Radioactive Residues from Nuclear Decommissioning, IAEA Bulletin, 42/3/2000 (September 2000)
Nuclear Energy Institute, Decommissioning of Nuclear Power Plants factsheet (August 2016)
Edward C. Doubleday, A Decommissioning Wrapup: Commercial Reactor Decommissioning Status in 2006, Radwaste Solutions, Volume 14, Number 2 (March/April 2007)
International Atomic Energy Agency, Power Reactor Information System (PRIS)
International Atomic Energy Agency, Policies and Strategies for the Decommissioning of Nuclear and Radiological Facilities, Nuclear Energy Series No. NW-G-2.1, STI/PUB/1525 (December 2011)
Graphite Decommissioning: Options for Graphite Treatment, Recycling, or Disposal, including a discussion of Safety-Related Issues, EPRI, 1013091 (March 2006)
Decommissioning in Germany, Nuclear Engineering International (27 March 2013)



Related information


You may also be interested in