WHEN news spread in December 2007 that an ageing nuclear reactor in Canada might shut down for much longer than its scheduled two weeks, the world caught its breath. The reactor, at in Ontario, is the world’s biggest supplier of radioactive isotopes for medical use, and diagnostic tests for cancer and heart disease were put on hold while radiologists scrambled to find alternative supplies. It was called a crisis. All the while, lay people couldn’t help but wonder: did no one foresee this? Did no one think that this half-century-old reactor might someday need to be replaced?
As it happens, not only did someone think about it, they designed and built its successors right next door. Maple 1 and Maple 2 are two brand-new reactors, constructed at a combined cost of over C$350 million ($330 million) specifically to produce medical radioisotopes. A single Maple reactor can supply the world’s total current needs; the second one is a back-up to keep the supply flowing during routine repairs.
But the sad truth is that the Maples have never been officially switched on, and the chances are they never will be. This has led to a furious row over who is to blame for this costly and embarrassing debacle. Many in the nuclear industry point the finger at Canada’s nuclear regulator. The regulator’s view is that the reactors’ manufacturer failed to deliver a crucial safety feature that it had promised would underpin the design.
Advertisement
Others blame the Canadian government for killing off the project before crucial technical questions had been resolved; in May 2008 it announced, to everyone’s surprise, that the Maples were being shuttered for good. “It makes absolutely no sense to me,” says , an engineer at the University of Waterloo, Ontario, who gave evidence to a parliamentary committee now looking into the affair. It’s time the Canadian government reversed its decision, says Nathwani: “If the will was there, the Maples could be brought back in six to 18 months, with just one phone call from the prime minister.”
Radioisotopes have a vital role to play in modern medicine. They are used in almost 40 million medical procedures each year, mostly for treating and diagnosing diseases such as cancer and heart disease. Over 80 per cent of the diagnostic procedures rely on technetium-99m, a short-lived isotope that is produced by bombarding uranium-235 with neutrons inside a reactor (see diagram).
Tell the safety story
In principle any old reactor can do the job, but getting the isotopes out without causing leaks or messing up the normal function of the reactor can be tricky. This generally means it is safer and more efficient to use purpose-built designs with slots that allow material to be slipped in and out of the core through protective shielding.
Today almost all of the world’s supply of medical isotopes comes from just five reactors (see map). The National Research Universal reactor at Chalk River is the oldest and most important. Owned and operated by , it produces 40 per cent of the world’s technetium-99m and 75 per cent of all cobalt-60, which is used for treating cancer. In the early 1990s it became clear that, due to growing demand, a new source for isotopes had to be found. So in 1996, Ottawa-based isotope wholesaler hired AECL to design, construct and run two new reactors and a processing facility. AECL’s engineers proposed a tiny 10-megawatt design that was dubbed “Maple” – the Multipurpose Applied Physics Lattice Experiment. The project to build two of these reactors was budgeted at C$145 million ($137 million).
So where did it all go wrong? For a nuclear reactor to be licensed in Canada its operators have to provide a “safety story” for the . That story sets out all the possible things that could go wrong and how the proposed design will cope with each of them. The Maples story included two separate shutdown systems. One was a set of shut-off rods. Another was a system for rapidly draining the heavy water from around the core. Both systems would extinguish the fission reaction by rapidly reducing the quantity of neutrons zipping around inside the reactor.
As another crucial part of the safety story, AECL also promised the reactors would have something called a “negative power coefficient of reactivity (PCR)” – basically a negative feedback loop which ensures that as the reactor’s power output goes up, it becomes ever harder to squeeze still more power out of it. Negative PCR works like wind resistance does on your car: the harder you put your foot down, the more the wind impedes any further increase in your speed (see “It pays to be negative”).
In itself, a positive PCR is no barrier to getting a licence. In fact, the workhorse in the AECL stable, the CANDU reactor, can under certain conditions have a positive PCR and it’s not a problem.
The first hint of serious trouble for the Maples came in June 2003, when they had already been completed – behind schedule and with costs almost double the original budget – and were undergoing their commissioning tests. To everyone’s surprise, as they powered up, instead of showing a negative PCR, the coefficient was positive. The analysis had predicted a PCR of -0.84, but it was measured at +0.28. “It was recognised that this was unexpected and undesired,” says Fred Boyd, a nuclear physicist who has worked for AECL, the regulatory agency and as a policy-maker. “All the people involved were quite concerned.”
Commissioning was halted as the regulator asked AECL to resolve the discrepancy. For almost three years, AECL went through its calculations, but nothing changed: if its model of how the reactor functioned was correct, the PCR should be negative. It even hired consultants from the US’s Brookhaven and Idaho national laboratories, and from the Argentinian engineering company INVAP to go over the physics. “Nobody came up with a result that was significantly different from the original results,” the parliamentary committee was told by Harold Smith, who not only helped to design the Maples but was also their commissioning supervisor. “From this we concluded there must be an unmodelled effect taking place.” The regulator required that AECL come up with an explanation of what was causing the discrepancy, but it couldn’t.
It was all the more puzzling because by then a 30-megawatt reactor called , with the same basic design as the Maples, had been built in South Korea and was working just fine. Smith became convinced that the problem with the Maples lay in some detail of the way they were constructed. The South Korean reactor, for instance, had issues early on with its fuel containers bowing; stiffening them had solved the problem. Smith asked for permission to run more tests on the Maples to see if something about their components was causing the positive PCR, and he began a series of three tests in spring 2008. But before they could be completed, the government, which owns AECL, pulled the plug.
Promises, promises
According to AECL, uncertainty about the cost of solving the PCR problem was a key factor in the project being scrapped. “I can confirm there are significant technical and regulatory hurdles that require, in the best-case scenario, at least five to six years of intensive research and analysis before we can even consider bringing the Maple reactors on line,” the Maples project director, Jean-Pierre Labrie, wrote in an last year in the Toronto-based newspaper the National Post. He calls the positive PCR a “potentially insurmountable hurdle”.
But many in the nuclear industry vehemently disagree, and a number argue that the positive PCR is not in itself a big deal. Smith, for instance, believes it can be rectified. “The positive power coefficient of reactivity is not a mystery,” he told the parliamentary committee. “It is not an unsolvable engineering problem. It is a small thermal mechanical effect in a prototype design that requires a simple engineering fix.” , former chief engineer at AECL, doesn’t even think it needs to be fixed, since a small positive value is fairly easy to manage using neutron-absorbing control rods. “It really doesn’t matter if it’s positive or negative as long as it’s small,” he says.
Smith, Meneley and other independent engineers say there is no sound technical reason why the Maples shouldn’t be operational right now. The stalemate, they argue, is not due to AECL’s failure to deliver the reactor that it promised, but rather the nuclear regulator’s demand that AECL should be held to that promise.
Linda Keen was president of the CNSC from 2001 until she was sacked by the prime minister in January 2008 (New Scientist, 26 January 2008, p 6). Though not a physicist, she is no stranger to regulation, having worked in the role in the food and explosives industries. “I treated AECL like any other licensee,” she says. Her stance of acting as a guarantor of public safety rather than a facilitator of the nuclear industry was breaking 50 years of tradition, she says. “In the past, regulators and the industry had a different relationship. They were buddies.”
In the old days, the regulator would often deal with problems by trying to help iron things out, recalls , a professor emeritus of nuclear engineering at McMaster University in Hamilton, Ontario. “They knew what situations could cause grief, and would work with the company to make sure there was real safety – not just regulatory compliance.” But with the introduction of the Nuclear Safety and Control Act in 2000, the the regulator’s job changed. Keen is clear that the job of the CNSC is not to help out but to protect Canadian citizens. She points out that with the Maples, it was not the positive PCR itself that was the problem. “The CNSC didn’t say that you can’t have a positive power coefficient. We said: ‘You promised. Your safety analysis was based on this. Tell us what your safety analysis is’.”
This goes to the heart of the issue, says nuclear engineer at the École Polytechnique in Montreal. “It is not the fact that the coefficient is positive or negative that is a problem. The problem is that you cannot calculate it,” he told the parliamentary committee. “When you find yourself in a situation where you cannot predict as simple a measure as the power coefficient, then can you be sure that the nuclear safety analyses, which are based on calculations, are correct?”
Small is difficult
Meneley, however, contests claims that the PCR is simple to calculate, especially in a small reactor like the Maple design. In a big reactor, leaving a few neutrons unaccounted for may not make a difference, but in a tiny reactor a very small difference in temperature or in the position of any of the many small components can throw the numbers off. “It’s almost incalculable,” he says. “It was naive to think that such a sensitive coefficient could be calculated within such a high degree of accuracy.”
Meanwhile in May 2009, the Chalk River reactor started leaking heavy water and was again shut down. Doctors and patients around the world now hope it can be fixed before March this year, when the world’s other main source of isotopes, a reactor at Petten in the Netherlands, goes offline for six months of essential maintenance. Between them, these two reactors produce around 70 per cent of the world’s supply of technetium-99m. Any extended shutdown could lead to a shortage, preventing diagnostic tests and delaying vital treatment for millions of patients.
“The world’s other main source of isotopes is due to go offline soon. Any extended shutdown could lead to a global shortage”
With this shortage of medical isotopes looming worldwide, other countries have begun to act. In November, the US House of Representatives voted to spend $163 million over five years to look into developing a domestic supply of medical isotopes, including the possibility of refitting an existing reactor at the University of Missouri. Later the same month, the Netherlands gave the go-ahead for a replacement for the aged Petten reactor to be built at the same site. Called Pallas, it is scheduled to be fully operational by 2016.
November also saw the of a report by an expert panel charged with investigating ways to secure Canada’s supply of medical isotopes, technetium-99m in particular. Its recommendation? To construct a new multipurpose reactor costing C$1.2 billion, built from scratch. The fate of the Maples appears to be sealed.
It pays to be negative
A fission reactor works by maintaining a controlled nuclear chain reaction. First, a neutron is absorbed by the nucleus of an atom of uranium fuel. This causes the uranium nucleus to split – undergo fission – which liberates heat and more neutrons. Some of these neutrons will be absorbed by other uranium nuclei, which will split in turn, resulting in yet more neutrons.
If this chain reaction is allowed to let rip, the result is a nuclear explosion. Reactors are designed to keep it at the desired, steady level, and this means knowing how many neutrons are being absorbed not only by the fuel, but by the coolant, the steel parts of the plant and innumerable other components.
The power coefficient of reactivity (PCR) is a measure of whether a particular reactor’s nuclear neutron population will tend to increase or decrease after an increase in the reactor’s power. If the number of neutrons increases, then the PCR is positive; if it decreases, PCR is negative.
Any sudden excess of neutrons will accelerate the fission reaction and increase the reactor’s power output. This is where having a negative PCR is handy, because it automatically limits the number of neutrons flying through the reactor, so it damps itself down after a momentary power increase.
The Maples were designed to have a small negative PCR: this was billed as one of its safety features and the reactors were licensed on that basis. When they powered up, though, operators measured a positive PCR. It was the maker’s continuing failure to explain why this happened that led to them being refused a licence.
