Subscribe to our newsletter
Weekly updates with new Fifty Two stories
Early in the acclaimed HBO mini-series Chernobyl, a deputy engineer named Anatoly Sitnikov tells his boss, Nikolai Fomin, that there’s been an explosion in the core of the nuclear reactor.
“Are you stupid?” asks Fomin, grim and disbelieving. “Sitnikov, you are a nuclear engineer. So am I. Now please tell me how an RBMK reactor core explodes.” Fomin is willing to admit the possibility of a “meltdown,” but is convinced a nuclear reactor can’t explode. Sitnikov flails, but insists he’s seen rubble and noticed graphite in it. The core of the reactor is made of graphite blocks.
Within three months of the Chernobyl explosion, 28 firemen and emergency clean-up workers died of acute radiation sickness. Many more workers suffered radiation-induced cataracts. By 2005, more than 6,000 cases of thyroid cancer were found among children and adolescents living in Belarus, Ukraine and other affected areas.
Parts of Chernobyl were filmed at an actual nuclear reactor, a partially decommissioned plant named Ignalina, in the town of Visaginas in Lithuania. Decommissioning goes beyond simply ceasing the commercial operations of a reactor. It involves a thorough clean-up of radioactivity at the site, possibly preparing it for alternate use, including but not limited to filming a mini-series.
Ignalina’s two 1500-megawatt reactors were of the RBMK class, too: the reaktor bolshoy moshchnosti kanalnyy, or the ‘high-power channel-type reactor.’ A legacy of the erstwhile Soviet Union’s nuclear power programme, it’s the oldest commercial reactor design still in operation. There was a time when Ignalina, which began operations in 1983, generated power for 80% of Lithuania’s electricity requirements. But the similarity in its design with the Chernobyl reactors compelled Lithuania into decommissioning Ignalina. In fact, decommissioning it was one of the conditions Lithuania had to fulfil to join the European Union in 2004.
uilding nuclear plants is expensive and technically challenging for every nation in the world. So is phasing them out. I first started thinking about India’s own plans for decommissioning its nuclear plants when Japan signed a nuclear power cooperation agreement with India, despite domestic opposition in Japan. It was November 2016, and The New York Times reported that the deal came as a “lifeline” for a Japanese nuclear industry that had been “floundering” since 2011.
A decade ago, on 11 March 2011, an earthquake followed by a tsunami badly affected the cooling functions of three reactors of Fukushima Daiichi nuclear power plant, causing the reactor cores to overheat. There was a meltdown of nuclear fuel, and a subsequent release of radioactive material in the environment.
That was what the International Atomic Energy Agency  designates a Level 7 event: a “major release of radioactive material with widespread health and environmental effects requiring implementation of planned and extended countermeasures.” It was the first, and as of this writing, the last Level 7 event after Chernobyl.
Nuclear energy is widely considered a safe, clean and cheap way to generate electricity. Its supporters have long argued that environmental and human rights activists tend to be alarmist about grotesque, but incredibly rare accidents. India’s Department of Atomic Energy claims that nuclear power is as safe as air travel, a popular line of argument among pro-nuclear thinkers: it’s statistically a safe bet.
But theoretically, an increase in the number of nuclear plants increases the chances of an accident taking place. A nuclear plant’s components release radiation into the atmosphere even when it isn’t in operation. India’s nuclear establishment says that the risk of human harm from radiation is negligible compared to the damage caused by air pollution, related to the use of fossil fuels. Studies have challenged this view. For instance, investigations by news publication DNA a decade ago revealed an increased incidence of a rare form of bone marrow cancer in the villages around the Kalpakkam nuclear site in Tamil Nadu. As for hazardous radioactive waste, there’s even less clarity: how and where is nuclear waste dumped?
There’s also the matter of nuclear plants ageing. I am writing this from Mumbai, just 100 kilometres away from the Tarapur plant, site of Tarapur reactors 1, 2, 3 and 4. Reactors 1 and 2 are among the oldest nuclear power plants in the world: they were India’s first commercial reactors, built in 1969. Their lifespan was initially planned to be 40 years, after which they were to be decommissioned. The Tarapur reactors are built along similar lines as the Fukushima ones.  Most reactors of their age around the world are already out of operation.
I spoke to India’s nuclear experts and filed several Right to Information (RTI) requests with government bodies to understand how India plans to manage its ageing nuclear reactors, especially because those which have reached their expiry date can pose serious public and environmental safety hazards. The government already has a decommissioning fund, which has accumulated over ₹2000 crore. The source of the funds is a decommissioning levy, which is collected on every unit of energy sold from nuclear power plants.
The fund’s purpose is to ensure that old plants which have reached their expiry date are fully retired and decommissioned. But what does the existence of a fund tell us? Do we know if this is enough to decommission nuclear reactors, a process which is enormously costly? Does the government have the competency, technology and the intent to decommission reactors like Tarapur 1 and 2? Here’s what I found.
n 2019, there were 443 operational reactors in 30 countries. 22 of these are in India. Taken together, they supply 3.2% of the nation’s electricity. If that sounds very low, it is. It also appears that very little of our future requirements will be met by nuclear energy. That’s not even going into how consistently the government has fallen behind on its nuclear energy targets. Presently, India’s nuclear power output is 6,780 MW. We were supposed to be generating 8,000 MW as far back as 1980.
The estimates have grown wilder over the years. In the early 2000s, a prediction of 20,000 MW was made for 2020. The latest estimate, from last month, proposes 22,480 MW of nuclear power by 2031.
From the 1970s, nuclear plants sprang up in Rajasthan, Tamil Nadu, Uttar Pradesh, Karnataka and Gujarat. These reactors are operated by the Nuclear Power Corporation of India Limited (NPCIL), a public sector undertaking wholly owned by the government. A regulatory board—the Atomic Energy Regulatory Board or AERB—is tasked with developing safety policies, guides and standards for nuclear and radiation facilities.
After the disaster in Japan, both the NPCIL and AERB conducted assessments of the consequences of Fukushima-like events in India. Predominantly, these studies established that the safety of India’s nuclear power plants is assured. 
“How does the nuclear establishment arrive at such decisions? The life of a reactor cannot be prolonged forever.”
But there’s more. The AERB reviewed the question of how safe Indian nuclear reactors would be from tsunamis, cyclones and floods, and nuclear station black-outs: occurrences of the total loss of AC electrical power to a nuclear plant, from both on- and off-site sources. In Fukushima, an earthquake first cut off electricity supply from the grid to the plant. The subsequent tsunami disabled emergency diesel generators. This meant that critical cooling functions did not kick in.
The AERB’s sub-committees uncovered several concerns during the review, especially in the case of the Tarapur reactors. The experts M.V. Ramana and Ashwin Kumar have observed that these are of a design very similar to the Fukushima Daiichi I reactor. During station blackout,  Tarapur 1 and 2 could only maintain cooling for the reactor cores for about eight hours.
The AERB assessment found that if a tsunami occurs, Tarapur 1 and 2 would have access to only the most basic safety measures  to keep the core cooled, and manage containment. The IAEA indicates that plants should have severe accident management guidelines: what to do when there’s a partial or total melting of fuel in a reactor core for unforeseen reasons. In its 2018 annual report, the AERB declared that it had reviewed and instituted accident management guidelines for all operating power plants. The report did not address the elephant in the room: the age of the Tarapur reactors.
The Tarapur reactors 1 and 2 completed their original planned lifespan of 40 years in 2009. But they continue to operate well beyond. According to the AERB, the operating licences of Tarapur 1 and 2 are, subject to extension, valid until March this year. In November last year, I filed an RTI inquiring about plans for the future of Tarapur 1 and 2. The NPCIL said there are no plans to decommission them. I asked if the organisation possesses the technology needed to decommission plants. The NPCIL said it does, but provided no details whatsoever.
“How does the nuclear establishment arrive at such decisions?” S.P. Udayakumar, founder of the People’s Movement Against Nuclear Energy (PMANE), asked. “The life of a reactor cannot be prolonged forever, but Tarapur 1 and 2 is continuing after refurbishments. There is no popular participation, accountability or transparency.”
PMANE is at the forefront of protests against a nuclear plant at Kudankulam in Tamil Nadu, India’s biggest.  Activists there have raised questions about the quality of components used in the reactor, and highlighted concerns about the impact of radioactivity on marine life in the area. That the east coast is vulnerable to tsunamis is central to the argument that Kudankulam is hardly the ideal choice to locate a nuclear power plant. A tsunami alert in Tamil Nadu in 2012, linked to an earthquake in Indonesia, lent weight to their case.
Recent studies have revealed that India’s west coast isn’t as immune to tsunamis as previously thought. In 2020, research led by geologist C.P. Rajendran concluded that the Makran subduction zone, located within the Arabian Sea, is susceptible to a high-intensity earthquake followed by a tsunami event. If such a thing happened and affected plants on India’s western coast, in Maharashtra and Karnataka, the consequences could be catastrophic.
n India, fears about nuclear safety have always existed uneasily with the ambition to embrace one of the most significant scientific developments of modern history. Producing nuclear energy to power the nation-building effort was part of Jawaharlal Nehru’s scientific vision. In Dr. Homi Bhabha, founding director of the Tata Institute of Fundamental Research, he had the perfect man for the job. The Department of Atomic Energy (DAE) was established in 1954. It had two goals: producing more nuclear-powered electricity for India, and developing nuclear technology.
Since then, several elements have been added to the institutional matrix that is India’s nuclear establishment. The Atomic Energy Commission was set up in 1958. The Commission’s task was to formulate DAE policies for consideration and approval by the prime minister. The NPCIL was created in 1987 as a public company to design, construct and operate nuclear power plants. It fell under the administrative control of the DAE.
Before this, in 1983, the AERB had been established as a ‘watchdog’ body—it was meant to regulate and enforce safety norms for all nuclear activities, including those of the DAE. But there was a catch. The AERB was set up by notification under Section 27 of the Atomic Energy Act, 1962, which allows the central government to delegate powers and duties to a subordinate authority. The implication of this is that its independence is compromised.
AERB’s founding notification makes it clear that the “Board shall be responsible to the Atomic Energy Commission.” In a 2002 article, A. Gopalakrishan, former chairman of the AERB who has written extensively about the regulatory framework, noted: “The AERB chairman reports to the Atomic Energy Commission, which is also headed by the secretary of the DAE who has ultimate responsibility for the DAE installations.”  This continues to be true in 2021.
The DAE, in its turn, has been at pains to underplay its supervision of the AERB. It has assured the Parliament more than once that the AERB is “functionally independent” and “did not come under the Department of Atomic Energy,” even while acknowledging it reports to the Atomic Energy Commission. 
Other experts have also remained unconvinced. “Had we had a genuinely independent, professional regulatory body instead of the AERB, an objective evaluation of the increasing risks of an aging nuclear power plant would have been made,” the retired bureaucrat E.A.S Sarma, former secretary in the Union power ministry, wrote to me in an email. “A decommissioning policy consistent with the public interest would have been formulated.”
This situation puts India at odds with its promises to the global community. It was one of the first signatories of the Convention on Nuclear Security in 1994, which requires a country’s nuclear regulator to be effectively separate from any other body or organisation which promotes or utilises nuclear energy. Gopalakrishnan has written that the AERB’s “subservience” to the DAE “clearly violates the article of the Convention to which India is a signatory.”
The Fukushima disaster finally prodded the DAE to propose the Nuclear Safety Regulatory Authority Bill, which proposed to dissolve the AERB and establish a more independent regulator. A Parliamentary Standing Committee examined it in detail and made several important suggestions in 2012. “But no further action was taken, once public memory of the perils of Fukushima faded,” Sarma explained.
n 2012, the office of the Comptroller and Auditor General of India, then held by Vinod Rai, conducted an audit. It found that all 20 nuclear power plants functioning at the time were operating without decommissioning plans. This was in contravention of international requirements, as well as India’s own standards.
The AERB had published a manual in 1998 which required already operational power plants to submit their preliminary decommissioning plans over the following five years. Newer plants were to do the same in order to get construction or operation licences. Not a single older plant had submitted its plans. Newer units had actually received operating licences “without the AERB insisting upon submission of decommissioning plans,” the CAG audit report said.
The AERB protested. Its manual was advisory, it contended; nothing in it was supposed to be mandatory or recommendatory. The audit concluded that “the reply confirms that AERB does not have an adequate mandate in respect of decommissioning of NPPs (nuclear power plants), research reactors and other nuclear fuel cycle facilities.”
The IAEA was also unimpressed. When it conducted a review in India in 2015, the agency judged that the AERB wasn’t a sufficiently independent regulator. That was just one major point of concern. The second had to do with the government’s strategy to manage radioactive waste from nuclear plants: the IAEA made it clear that it doubted India’s commitment to effectively handle radioactive waste.
aypersons often think of nuclear refuse as the poisonous, immortal detritus of human ambition, destined to kill everything it touches, capable of destruction long after the age of humans is over. The immediate realities are more prosaic. In India’s 3-stage nuclear programme, the government aims to minimise nuclear waste because it adopts the closed fuel cycle.
Briefly, this means spent fuel from one stage is used in the next. The stages were designed to optimise the use of India’s abundant thorium but limited uranium reserves. In the first stage, uranium-fuelled pressurised heavy water reactors, or PHWRs, produce electricity. The spent fuel of this stage contains Pu-239, an isotope of plutonium that is fissile in nature. In other words, its atoms are capable of splitting and sustaining the chain reaction that produces energy.
“It is true that external events triggered the disaster. But in Fukushima, the design flaws and the age of the reactor became important.”
The second stage is the fast breeder stage, in which thorium comes into the picture. Thorium is itself not capable of fission, so it must combine with fissile material to generate energy, typically Pu-239. Thorium and Pu-239 combine to form an isotope of uranium called U-233. In the final breeder stage, U-233 will serve as fuel to run thorium-based reactors. The second stage of India’s nuclear power programme has not yet been operationalised, though it has been in the works for some time now.  The country’s first fast breeder reactor at Kalpakkam has been plagued by a history of delays.
The three stage-programme may minimise nuclear waste by using spent nuclear fuel. But questions about handling daily and long-term waste, transportation and storage have not been addressed. This was a major issue in the 2012 protests against the Kudankulam plant, where the operators built an on-site facility as an “interim measure” for storing spent fuel.
While greenlighting the Kudankulam plant in 2013, the Supreme Court stressed that it was of “utmost importance” that the operator find a place for a permanent Deep Geological Repository or DGR: an excavated, underground facility expressly designed and constructed for the permanent disposal of high-level radioactive waste. NPCIL asked for five years’ time to construct a facility. As of 2020, the Indian government’s public position is that the country doesn’t need a DGR at all, because the closed fuel cycle generates “very less” radioactive waste. 
he Indian nuclear establishment has a strategy for spent nuclear fuel. But plant safety involves much larger questions. “When nuclear waste is removed from the reactors, the nuclear spent fuel rods, the most dangerous part of the reactor, are removed,” S.P. Udayakumar told me. “The remaining part of the reactor is not as virulent as the spent fuel rods. But the structure continues to be hazardous and radioactive. Therefore, reactors need to be decommissioned at some point.” Decommissioning plans must go well beyond the removal of spent nuclear fuel from the reactor. They will have to involve shutting down plants and clearing the premises of radioactivity altogether.
Environmentalist and social activist G. Sundarrajan said India possesses neither the resources nor the technology for decommissioning. “During the protests against the Kudankulam reactor, we had requested that the Supreme Court direct the DAE to set up a decommissioning authority of India. But nothing has come of it.”
ollowing the deeply critical 2012 audit, the AERB had asked all nuclear power plants to submit decommissioning plans. Subsequently, it reviewed and approved ‘conceptual decommissioning’ plans for all operating power plants.  In an RTI response, AERB said that these plans would be reviewed periodically. Nuclear plants under construction also need to submit conceptual decommissioning plans to AERB as part of their initial licensing.
As of now, there is no precedent to say how these plans will work out in practice. A few years ago, when questioned about capacities in the Rajya Sabha, the United Progressive Alliance government said that India could and did decommission nuclear plants: it had successfully done so with Purnima and Zerlina reactors at Mumbai’s Bhabha Atomic Research Centre.
What the government did not reveal  was that Purnima and Zerlina were experimental research reactors, not full-scale electricity-generating plants. “You don’t need elaborate arrangements to decommission them,” Udayakumar said. The physicist Suvrat Raju, a member of the Coalition for Nuclear Disarmament and Peace, agreed: “It’s like saying they decommissioned a toy reactor. To date, India has not decommissioned a commercial nuclear power plant, or a reactor with any significant electricity output.”
In RTI responses, the NPCIL and the AERB both confirmed to me that India has never decommissioned an electricity-generating nuclear power plant. In spite of a sea change in government and nearly a decade following the questions raised in the Rajya Sabha, the plan to do so doesn’t exist.
In 2019, Jitendra Singh, minister of state for atomic energy, said as much in the Lok Sabha. “The design of nuclear plants is inherently robust and they can be operated beyond their initial economic life based on systemic life assessment studies and undertaking necessary life extension measures.” His statement mentioned “eventual decommissioning” in accordance with AERB guidelines, without specifying when. In other words, the Indian government is confident about the inherently robust design of its nuclear power plants: enough to prolong their lives without a definite end.
Japanese authorities had much the same confidence in the Fukushima reactors before 2011. In the lead-up to that disaster, regulatory failure, negligence by operators and a mix of ignorance and arrogance in the nuclear establishment all played their part. We know this because the committees set up in the aftermath of the meltdown identified all these problems. The chairman of one official inquiry commission called it a “profoundly man-made disaster.”
“People had already pointed to the flaws in the Fukushima reactors, but they were still operating,” Raju told me. “It is true that external events triggered the disaster. But in this accident, the design flaws and the age of the reactor became important.”
An aside: Decommissioning is different from a shutdown, but the distinction is important. India can do and has done the latter. In 2004, authorities shut down RAPS-1 in Rajasthan, a nuclear plant that has been running since 1973, because of a “chequered history of operation,” and problems with its components and equipment.  The DAE said that fuel has been removed from its reactor, and it is now in “a safe shutdown condition” while authorities study options for its refurbishment.
n one regard, India has long been preparing for decommissioning: it’s been collecting the funds. But neither the NPCIL nor the AERB are able to publicly provide figures to RTI questions about how much it will cost India to decommission a plant. “The actual cost of decommissioning an aged nuclear power plant can far exceed the unit cost element factored into the tariff structure at present,” Sarma told me. “This implies a hidden subsidy for electricity from nuclear power plants.”
NPCIL has, for years, collected a decommissioning levy from electricity distribution companies on behalf of the government. India has had a decommissioning fund since 1988, when the DAE notified that it would sell power from nuclear plants at 1.25 paisa per kilowatt hour of energy. In 1991, the Department revised its charges to 2 paise per kilowatt hour. The charges have never been revised after that. We discovered as much during the 2012 CAG audit, and it was confirmed by the NPCIL in a response to my RTI request.
As of March 2020, India’s decommissioning fund had accumulated ₹2,349 crore; of this, ₹84.21 crore was collected over the financial year 2019-20, and ₹68.43 crore was collected in 2018-19. But the public doesn’t yet know how exactly this money is going to be used.
In 2013, the scholar M.V. Ramana wrote in an academic paper that there’s “no reliable estimate of how much decommissioning a reactor will cost.” In other countries, decommissioning has “invariably cost much more than expected.” However, as of this writing, there are no special legal arrangements in place to fund decommissioning, a fact pointed out by both the CAG audit and the IAEA.
There’s no doubt that this is going to be an extremely expensive and time-consuming proposition. Conservative estimates peg the cost of decommissioning at anywhere between 10 percent to 25 percent of the cost of constructing a nuclear plant; the actual costs, as was noted in the Bulletin of Atomic Scientists in 2014, could be considerably higher.
“Some radioactive waste has a life of three days, 50 days or several thousand years. Where and how will it be stored?”
And there is the problem of radioactive waste generated from the decommissioning. In an RTI response, the NPCIL said that both short-term and long-term radioactive waste will be disposed of in Near Surface Disposal Facilities. But the IAEA recommends near surface facilities only for the disposal of low levels of solid radioactive waste. NPCIL also stated that it has opted to delay decommissioning to allow time for natural radioactive decay to occur, ostensibly to reduce the quantity of radioactive waste.
“NPCIL hopes that the integrity of the near surface disposal facility will remain intact, without leakages or theft,” Suvrat Raju told me. “But that is not a good solution for long-term waste.”
“The waste could be radioactive for hundreds or thousands of years depending on the isotope that is formed,” Sundarrajan explained. “Some have a life of three days, 50 days or several thousand years. Where and how will it be stored?”
Raju said that a long-term solution, like a geological repository, will need to last tens of thousands of years. “But nobody has found something that would be stable that long.”
“Decommissioning was rarely considered in the reactor design,” 2020’s World Nuclear Industry Status Report says. The costs were “usually discounted away, and thus, subsequently, largely ignored. However, as an increasing number of nuclear facilities either reach the end of their operational lifetimes or are already closed, the challenges of reactor decommissioning are coming to the fore.”
The 2018 edition of the report detailed how cumbersome the process of decommissioning is. There’s the onerous task of removing spent fuel from the reactor, but this is only a first stage. In the second stage, highly contaminated or activated parts such as the reactor pressure vessel and its internal parts need to be dealt with. The final stage of decommissioning requires removal of operating systems and decontamination of buildings. This stage ideally ends with the demolition of buildings and preparing the site for unrestricted use, or for use with some restrictions.
Countries that have kicked off the decommissioning process understand that there is a huge gap between theory and practice. Sweden, for instance, discovered that there was a need for waste technicians who could identify and categorise radioactive material. They also found that environmental engineers need to be consulted for know-how on preventing the contamination of the plant’s immediate surroundings. 
According to S.P. Udayakumar, decommissioning is a complicated process involving elaborate steps. “You need huge amounts of space, huge amounts of steel and concrete, lots of conventional energy,” he said.  “And then, you cannot lock up the place and throw away the keys. The structure has to be monitored for emissions from the reactor and it has to be foolproof for hundreds of years because radioactivity within the decommissioned nuclear power plant can last up to even 48,000 years.”
“Not just India, no other country has found a good way to deal with long-term nuclear waste. This is one of the problems of nuclear waste––it is an indefinite time problem,” Raju told me. As of 2020, globally, there are only 20 fully decommissioned reactors. The average duration of the decommissioning process ranges from anything between 6-42 years. The average time taken is 20 years. The only countries to have completed the decommissioning process are the United States (14 reactors), Germany (5), and Japan (1). Countries such as Canada, France, Russia  and the UK have not completely decommissioned even a single reactor.
The problem may not just be one of bureaucratic myopia. There’s systemic apathy as well. “There is, of course, an issue of kicking the can down the road,” Raju said. “But it doesn’t have to do with individual bureaucrats not being able to look past their careers. It is a deeper problem within the system. We see it in other cases as well, such as the climate crisis.” Gopalakrishnan agreed. He told me that it is hard to get the authorities to act because it is a future problem. But, in the case of reactors like Tarapur and RAPS-I, the future is already here.
Urvashi Sarkar is an independent journalist based in Mumbai.