With construction starting on ITER and the first plasma burst expected in 2019, the European JET facility is preparing to restart after its upgrade to pave the way. E&T met the man in charge of the UK’s fusion effort, Professor Steve Cowley.
The possibility of generating electricity from nuclear fusion has intrigued scientists since the late 1950s, but delivering an economical and reliable fusion reactor has proved to be a tough nut to crack.
Nuclear reactors are capable of producing huge amounts of energy, which is achieved either by nuclear fission – the splitting of elements of high atomic numbers, or fusion – the joining of elements with low atomic numbers. Nuclear fission is well developed and supplies around 14 per cent of the world's electricity but the advent of nuclear fusion is still some way off.
The most efficient reaction to use fusion is the DT fusion reaction where the nuclei of two hydrogen isotopes – deuterium and tritium – are forced together to produce a helium nucleus and a neutron – both with very high kinetic energy.
The three parameters required to achieve this fusion – plasma temperature, density and confinement time – need to be achieved simultaneously.
Because the plasma comprises of charged particles, confinement is achieved through powerful magnetic fields that isolate the plasma from the walls of the containment vessel. In a magnetic field the plasma is forced to spiral along the magnetic field lines inside a toroidal shape inside a tokomak, of which until ITER is completed, JET is the largest.
To achieve the required temperatures, in excess of 150 million degrees, several heating methods are used in JET. Currents of five million amperes are induced in the plasma producing several megawatts of heating. Beams of high energy deuterium or tritium atoms are injected into the plasma, transferring their energy to the plasma via collisions with the plasma ions. Finally, as the plasma ions and electrons are confined to rotating around the magnetic field lines in the tokomak, electromagnetic waves of a frequency matched to the ions or electrons are introduced.
But now it looks as though the finishing line is finally in sight. An international consortium has agreed the funding for ITER, a fusion reactor which is a larger version of the tokomak JET, located at the Culham Centre for Fusion Energy (CCFE). JET has been in operation there since 1984, until it was shutdown last year for a refurbishment in preparation for its role as an ITER proving ground. The upgrade, which is now almost complete, includes installing a new beryllium tungsten wall and increasing its power output.
The man charged with managing CCFE is Professor Steve Cowley, chief executive officer of United Kingdom Atomic Energy Authority and head of EURATOM/CCFE Fusion Association. Professor Cowley received his BA from Oxford University and his PhD from Princeton.
Q Can you tell me what you hope will be achieved from the JET upgrade?
A There are two things that we have to sort out before ITER starts. The first is that when we ran JET in 1997 and made those fusion records of 16MW of power, we discovered that the carbon-tiled walls were not suitable. They worked, but they were not suitable because they acted like a sponge and soaked up the tritium. So we had to come up with another solution for the wall. That is when it was conceived that we would have the beryllium tungsten wall on ITER when it does its half a gigawatt of power.
People say that carbon is forgiving because it doesn't melt. So it's nice in the sense that you don't get liquid carbon flowing around the walls, but it is not forgiving in that it allows the tritium to absorb into the walls up to a depth of about a millimetre – you lose a lot of tritium that way which is not good in an operating system. We have to show that a beryllium tungsten wall can work on JET because it's never been used before.
The other thing that we are doing with the upgrade is putting on more power, so when the JET comes back later this year it will be a different machine. No machine has ever operated like this before. We expect it to perform better, but we will have to get used to it first and take up the scenarios on JET that ITER is going to operate in.
We want to get through this phase of doing science experiments as quickly as possible because we want to produce power plants, making fusion a commercial reality. We don't want to waste years fiddling around on ITER getting it right. So this next period of time with JET will be to start up with the new wall, ramp up the power so that we can get record breaking power into it, then put in a tritium fuel and break all the world's fusion records in 2015.
Q Was there no desire to go straight from JET to a demonstration power plant?
A ITER is not very far off being the right size for a commercial reactor. We reason that a real commercial reactor will be less than 20 per cent bigger than ITER, but it will have to have a lot more systems in it. It will have a turbine, all the power engineering of generating electricity, and it will have to have a system for making its own tritium – it's called the blanket system.
ITER will only try out baby versions of that, called test blankets – little squares of wall that will do the testing. It will not breed its own tritium in totality.
The reason is money. When we originally designed ITER in the early 1990s we designed a machine that did everything; could actually be an electricity producing reactor. I think people backed away from that because it was too expensive and you were taking too many risks at once. We just want to do experiments on it. It was felt less risky to make ITER a machine that didn't actually produce any electricity.
Q Sceptics are keen to point to the fact that the longest sustained plasma burst to date is only five seconds?
A JET at full power can only do five seconds. The shots of JET are what we have used to extrapolate the performance of ITER – it was essentially designed around those 97 shots where we made 16MW.
It was not actually the 16MW shot that we designed ITER on, it was the 5MW for the full five seconds – you turn the machine on and it makes fusion for the full five seconds, then you turn the machine off and it stops.
We have no problem sustaining it, you just can't run JET for more than five seconds because the copper magnet gets really hot, all the systems get hot, and they won't last longer than five seconds at full power.
But ITER can because it's a superconducting machine that has its engineering systems designed to maintain a burst longer than that. It's not that the fusion can't go on longer than five seconds, it's just that the machine can't.
Q Is there a particular performance that ITER has been specifically designed for?
A What we call the baseline of ITER is 500MW of fusion power for 400 seconds. But ITER could go on for thousands and thousands of seconds and some of the modes of operation that we have been working on at JET probably mean that, if we adopt them on ITER, we will do thousands of seconds.
Personally I think we should be more ambitious. I think we will probably get more out of ITER than that, but nobody wants to promise that. What we are promising I think is a fairly conservative estimate of how well it can perform. I am hoping that we will get more than 500MW and that the gain will go up to infinity. That means you can turn off your external power sources and just let it cook itself.
Q What have we achieved and learned to date from running JET?
AIt took a little time to get JET going, but by the 1990s we were able to make sustained shots with central temperatures of 170 million degrees.
By the late nineties we had a mode of operating JET that could sustain it for the whole time you could run the machine and get a central temperature to 170 million degrees. You couldn't have done that at the beginning of JET's operation, we didn't know how.
In 1997 we discovered some very interesting new regimes of running JET; one of which is called the hybrid mode. The hybrid mode is when the electric current in the plasma is actually a little less and is hollow – there is more current flow outside than there is in the middle.
Those kind of discharges sustain their own current to some extent: if you're very lucky they will almost entirely sustain their own current. This means that you don't have to put in an electric field to drive the current around the loop, but in fact the plasma does it itself – it's called the bootstrap current.
In the two years before the shutdown we had managed to get hybrid modes on JET that were high performances. If those extend in the next period of time in JET up to 2015, the best way to run ITER will probably be in hybrid modes.
Q What are the big challenges that you can foresee fusion potentially having to face?
A We would like to bring down the turbulence inside the plasma even further. If you ask the physics reason that JET's performance has improved over time, it is because we are suppressing more and more of the turbulence inside.
Inside the plasma it is bubbling; it is 170 million degrees in the middle and 10,000 degrees at the edge. That huge temperature gradient is like applying heat to a saucepan, it bubbles away. And those bubbles are little turbulent eddies inside the plasma that causes the heat to lean out.
The less turbulence you have, the better the fusion reactor. The big success of the last two-and-a-half decades of JET's operation is a reduction in the turbulence. But we can go further. On MAST, for instance, we have regimes where there is almost no turbulence inside the plasma. That is the perfect confinement device.
The second big challenge is that every now and then the plasma will erupt and throw itself against the wall. You are storing several hundred mega joules of energy inside the ITER plasma and you don't want it to throw itself against the walls. So we have to have ways of making sure that doesn't happen in a commercial reactor.
We are working on ways for ITER, but by the end of the ITER project we had better have a very good idea when it's going to go unstable. Because when it does, and it is rare, it can damage the walls.
Q What about the engineering components, will they have to advance as well?
A For fusion itself, for making commercial power, the engineering components have to be incredibly reliable. They have to be built from materials that can survive in this very harsh environment.
Inside a fusion reactor there is a huge flux of neutrons hitting the wall, bombarding the wall, shaking the atoms in the wall. We want to have a material in which every atom in the wall is moved 150 times before you have to replace it.
Q Fusion is still subject to some bad press, normally focusing on the fact that it is always 30 years from delivering electricity. What are your thoughts on this?
AYou read these articles in the press where some scientist says 'how can you have a sun in a bottle?' – I would say we have done that! That is what JET does. The temperature in the middle of JET is a reactor working temperature; it is a sun in a bottle.
That is not the difficulty. The difficulty is a sun in a bottle that is cheap enough that it will make electricity at the sort of cost you want to pay for your electricity. I don't know if we can solve that problem.
I know we can make a sun in a bottle, but I don't know if we can produce electricity at five cents a KWh. That is the challenge and these days, it is as much an engineering challenge as a physics challenge.
Q In the current, post-Fukushima climate, the fact is that fusion is still a nuclear process with all the associated dangers. How do you tackle this?
A Waste from a fusion power station is an interesting problem. We produce helium, it is just ordinary helium. But the neutron that comes out when it strikes steel not only moves the atoms around, it can make some of the nuclei in the steel radioactive, so the steel becomes radioactive.
That is the only waste that we really worry about, that radioactive steel – we have designed a special sort of steel called a low activation steel, so the radioactivity is short lived.
So the analysis of our proposed power plant shows that at the end of the lifetime of a fusion reactor, the waste that you have is a thousand times less active than a fission plant and that it decays in 200 years to have the same radioactivity as coal.
Tritium itself is quite a hazardous substance. It's a radioactive gas with a half life of 12 years. You don't store much tritium at a fusion plant because you make tritium from lithium and you take that tritium and put it in the plant.
The neutron comes out of the plasma and goes into the wall. In the wall is this blanket of lithium, where the energy of the neutron is deposited and goes into your coolant to power your turbine. The neutron hits lithium and makes a tritium, which you have to take and put back in your plasma to make more fusion, so actually you make tritium and consume tritium.
If you've got a clever plant you actually keep very little tritium around, maybe a kilogramme, so if you ever had an accident the amount of tritium that could be released would be relatively small.
In the European Fusion Power plant design that we participated in, you wouldn't need anybody to be evacuated beyond the site boundary in the event of a worst case scenario.
Q What are the future landmarks that you envisage for fusion power?
A We won't know if we are right about ITER until after 2025. But by 2030 we will have to start building the first energy production reactor, something we call Demo.
Demo won't have to be perfect, it's a prototype, and it's not the last word. But it will have to be good enough to produce day in and day out electricity, so it's a demonstration that fusion works.
My goal here is that when eventually JET shuts down we become a design centre for the first commercial reactor, here on British soil. We want to design the first breed of commercial fusion reactors, probably with industrial partners.
So JET is probably going to run until towards the end of this decade, just before ITER starts up.
Beyond the end of JET I'd expect the conceptual design of the demonstration reactor to start here and the true construction and final engineering design right at the end of the 2020s.
Q Is that lengthy timeline working under any financial constraints?
A We are trying to hit something that is well under 10 cents per KWh, and I think we can do that on the timeline that I have just talked about.
If you wanted us just to build a reactor right now, I think we would build something conservative; we wouldn't stress the engineering at all. We would build something very, very big.
Fusion works much better when it's big. It is called the energy confinement time. So ITER is twice the size of JET and gets eight times the power.
If you say 'I want to make sure the reactor works' you would build something 50 per cent bigger than ITER and you would be absolutely sure you had a working reactor.
It would be very expensive and the electricity would be very expensive, but we could actually build it now. Industrialists have said to us that is probably the best strategy, to go ahead and do that, to build something that actually makes electricity and then let the engineering gradually bring the cost and the scale down. Start where you want to be and work backwards, but at the moment we don't have that kind of money.