After the dream, the humdrum reality

“THERE IS ONLY one reason for America to subsidise nuclear,” says Ernest Moniz, of MIT, “and that is the climate.” He has a point. By 2020, carbon emissions since the start of the 21st century will have surpassed those of the entire 20th. There is a real risk that emissions on such a scale will bring disaster to humans, or to the natural world, or both. Nuclear power, which produces no direct carbon-dioxide emissions, should be able to make things better.

Robert Socolow, of Princeton University, and his colleagues calculate that if the world were to replace 700GW of coal-fired plant with nuclear reactors over 50 years—which would more or less triple its current nuclear capacity—it could reduce its annual emissions of carbon dioxide by 3.7 billion tonnes. Allowing for the need to replace most of the current fleet over the same period, that would mean deploying nuclear reactors at three times the speed of China's planned record-breaking deployment between now and 2020, and doing it for five decades straight. But even that would make only a minor dent in the problem. In 2010 the amount of carbon dioxide emitted by industry was about 30 billion tonnes, and was growing at 3% a year. At that rate, the savings from such a beefed-up nuclear-power programme would compensate for just four years of emissions growth.

No technology can solve the climate problem on its own. Even in combination, today's remedies—renewables, nuclear and energy efficiency—hardly seem up to the job. To have a reasonable chance of keeping down the rise in temperature to less than 2°C, industrial economies need to reduce emissions by 80% by 2050. The true scale of this challenge is not widely understood. A thorough study of options for such cuts in California, long a leader in energy efficiency, concluded that with today's technology and plausible extrapolations of it, 60% was the best that could be done. If California can't do better than that, says Jane Long, of Lawrence Livermore National Laboratory, who led the study, “neither can anyone else”.

Even getting close to such goals, though, is easier with more technologies than fewer. Even if nuclear can make only a small contribution, it could be worth having. The IEA's 2011 World Energy Outlook calculates that, between now and 2035, an emissions path that keeps the 2°C limit plausible would cost $1.5 trillion more if OECD countries were to stop building nuclear plants and other countries halved their nuclear ambition, largely because much more would have to be spent on renewables.

Germany, long keen on renewables and squeamish about nuclear, provides an example. Its decision after Fukushima to phase out nuclear power entirely will mean that most of the lost capacity will need to be made up with even more renewables, though it will also build new fossil-fuel plants and import electricity from nuclear plants in France. As the new fossil-fuel plants will probably run on gas, emitting less carbon dioxide than do coal plants which are also due for retirement, this may keep the carbon in check. But electricity prices for industrial customers, according to an analysis by UBS, will rise by more than 60% in real terms by 2020. Ottmar Edenhofer of Berlin's Technical University says this is a pretty middle-of-the-road projection.

The market, too, will probably need some re-engineering. Systems with a lot of renewables make life hard for fossil-fuel generators, which have to shut down when it is sunny and windy and take up the slack when it is not. To get the fossil-fuel investment it needs, Germany may well have to pay for the capacity built even if it stands idle, or guarantee rates of return.

But though renewables are expensive, so is building new nuclear plants; the bills Britain will face as it tries to meet carbon goals with new nuclear should keep any Schadenfreude in check. The cheap new supply-side route to lower carbon-dioxide emissions is to replace old coal-fired stations with new gas-fired ones, which emit half as much carbon dioxide per megawatt hour. Plumping for renewables or nuclear will cost a lot more.

Still, renewables are getting cheaper, through technological change and through the benefits of mass production and market competition. In the long run, technologies that get cheaper can be expected to edge out a technology that has only ever got more expensive. In a low-emissions world, the role for nuclear will be limited to whatever level of electricity demand remains when renewables are deployed as far as possible.

That is a large enough role for some Greens to have become nuclear converts. For the most part, though, they are thinking about a nuclear programme more exciting than the slow, expensive and only marginally helpful deployment of PWRs. There are many alternative reactor designs, and each has its champions. An international body called the Generation IV International Forum (GIF), co-ordinated by the NEA, is drawing up plans for prototypes using such ideas, all claiming to offer improvements over the current crop.

Yet the problems these new reactors solve are for the most part those that the industry wishes it had, rather than those it actually faces. The GIF designs, and others, are mostly “fast” reactors that use highly fissile fuel and unmoderated neutrons; they can both burn plutonium and create it in copious amounts. If fissile material were in short supply that might be an advantage. But uranium is not currently in short supply, and it makes up only a small part of nuclear energy's costs. The ability to make new nuclear fuel solves a problem that reactors will run into only if their use becomes massively more widespread. What new reactors need is an advantage that will make them popular in the first place.

Indeed, at present the ability to make plutonium is a disadvantage. Dissuading countries with nuclear programmes, or that want nuclear programmes, from reprocessing their fuel to produce plutonium is one of the core priorities in anti-proliferation work (the other is trying to keep newly nuclear countries from developing their own enrichment systems). If established nuclear powers were to stop reprocessing (as Britain is doing), it might help to persuade others, such as South Korea, that it is better not to start. A new generation of plutonium breeders would completely undermine that effort.

Admittedly, other kinds of breeders are available. Molten-salt reactors, which keep their fuel in liquid form, could be used to turn thorium, of which the world has an abundant supply, into a type of fissile uranium not found in nature, U-233. This would be rather unsuitable for bomb-making and gets round the continuing use of U-235 or plutonium, so thorium molten-salt reactors offer the possibility of breeding fuel in a way that does not facilitate proliferation.

Like some of the other GIF designs, molten-salt reactors also have novel safety features; but although safety is a condition of getting into the game, it is hardly a means of winning it. If generation III reactors, well operated, prove safe, why upgrade? If they are not safe, who would trust generation IV? The way to win will be on price.

At the moment, those who want to bring down the cost of nuclear power are not, for the most part, looking at big generation IV reactors that will not be built for 20 years, if ever. Instead, they are thinking small. Particularly in America, small modular reactors (SMRs) of up to 300MW are all the rage. Some, such as the 100MW mPower reactor offered by Babcock and Wilcox, are scaled-down PWRs. Others are more exotic.

Think small

Such reactors can reach markets which today's big reactors cannot. Many utilities—and smaller countries—have little interest in gigawatt-scale plants: they prefer to build around 100MW of capacity at a time. Small reactors might also open up new applications, perhaps in desalination, district heating or even transport. He Zuoxiu, a Chinese physicist critical of his government's rush to build lots of big PWRs, has suggested that SMRs for ships, both military and merchant, would be a good way to train up a cadre of engineers and designers.

SMRs can also be slotted into underground silos, which cuts down on civil engineering costs. Perhaps most promising of all, they would be built in factories, not on site. That should make them less subject to delay than manufacture in the field. And a factory building ten such reactors a year for years on end might be able to make significant cost reductions through incremental improvements—economies of number as opposed to economies of scale.

Explore our interactive guide to nuclear power around the world

But these advantages do not add up to a conclusive case for a small modular future. Babcock and Wilcox claim overnight costs per kilowatt of capacity for the mPower roughly on a par with those of big PWRs like the AP1000. But in an industry that has long pursued economies of scale, many are unconvinced that smaller reactors can deliver the same costs per kilowatt. Atam Rao, who led the design of an advanced generation III reactor, GE's ESBWR, describes such claims as “complete BS”. Things like control systems are needed for all reactors, big or small. Providing them for each small reactor is bound to push up costs. Will any utility really think it makes sense to field ten SMRs with ten control systems and ten safety systems rather than one big PWR? Only if it has seen it done economically elsewhere.

In the end, that is the biggest problem for proponents of new approaches to nuclear energy. If a radically new technology, as opposed to an incremental one, is to take off, it needs not only to be researched and developed; it needs to be deployed, and industry will not do this until it has seen the technology work. It was the American navy's deployment of nuclear reactors that convinced the world that they could be used as power generators. And it was the experience of deploying them that allowed Admiral Rickover, in the 1950s, to sum up the gap between ideas that might work and those that are in fact working, in a way that still seems spot on 60 years later:

An academic reactor or reactor plant almost always has the following basic characteristics: (1) It is simple. (2) It is small. (3) It is cheap. (4) It is light. (5) It can be built very quickly. (6) It is very flexible in purpose. (7) Very little development will be required. It will use off-the-shelf components. (8) The reactor is in the study phase. It is not being built now. On the other hand a practical reactor can be distinguished by the following characteristics: (1) It is being built now. (2) It is behind schedule. (3) It requires an immense amount of development on apparently trivial items. (4) It is very expensive. (5) It takes a long time to build because of its engineering development problems. (6) It is large. (7) It is heavy. (8) It is complicated.

Innovators need to be able to take risks, to try variations on their ideas and to be able to learn. They flourish in unregulated markets. They frequently depend on venture-capital funds which are dwarfed by the cost of even a single utility-scale power plant. They also need rewards. Yet makers of nuclear reactors cannot take risks that might compromise safety, and they cannot try lots of different things because it would be too expensive. And even if they succeed, all they will be making is commoditised electricity. Power stations are not conducive to radical innovation.

Nuclear reactors, as Philippe Jamet notes, last for centuries; the technology is, by its own standards, still young. Longevity and inertia ensure that even a disaster like Fukushima cannot wipe it from the world. But they also ensure that it cannot grow fast. In energy in general, technologies mature and succeed each other over decades. Nuclear seems likely to lag behind even in this slow field. That does not mean it will not, eventually, play a larger role, but that it will get there slowly. Inside a reactor, things can change in milliseconds. Outside, it takes lifetimes.