The ratio of input to output energy is often called “Q”
.
To increase Q, we can do two things:
- Use less energy
- Produce more energy
It turns out, these two things are closely connected. As there are more and more fusion reactions (i.e. we produce more energy), the plasma heats itself more and more. This means we don’t need to continue to invest as much energy to heat it up. At some point, it could even keep going without any external heat input!
Over the last 60 years, researchers have worked on many different methods to get this right. Sadly, almost all of these designs are stuck at Q values between 0.0001 and 0.000001. However, there is one type (called ‘tokamak’) that has achieved Q=0.65.
Let’s see how it works!
“Tokamak” - the fusion donut
Remember our three problems:
- Heat the plasma to around 100,000,000°C.
- Density.
- Confinement of the plasma for seconds to minutes.
The tokamak simultaneously satisfies the requirements of confinement and density using strong magnetic fields. These magnetic fields force the negatively charged electrons and the positively charged nuclei to move round and round on a path inside a donut-shape. Because magnets surround the plasma, they create a high pressure, which increases the plasma density.
Breathe deeply. Sit back, make sure you got this so far. Because now we will touch on some of the problems fusion researchers work on today.
Where do deuterium & tritium come from?
Deuterium is easy to find and abundant - it’s in ocean water. However, only a few kilograms of Tritium are produced by nature every year and there is no such thing as a “Tritium factory”. Today’s fusion experiments often get it from nuclear fission power plants, where tritium is produced as radioactive waste. But where will we get it from if we stop using fission reactors?
Luckily, there is a way for fusion reactors to generate their own Tritium. In theory, we could reuse that extra neutron to make more Tritium out of Deuterium! The full reaction uses Deuterium (1p1n) and Lithium-6 (3p6n) as inputs to produce helium:
Using Lithium-6 to produce Tritium actually introduces more energy to the system. Doing this in practise is very hard and one of the active areas of research and uncertainty in fusion energy.
How can we get energy out of a tokamak?!
So far, so good. We have deuterium and tritium, we can heat them up, and cause fusion to happen. But how do we get the energy out?
Remember, neutrons don’t have a charge. Magnetic fields only interact with charged particles. This means that the magnetic field, as strong as it may be, can’t contain the fast neutrons coming out of the fusion reactions.
These fast neutrons are both the most valuable and most annoying aspect of fusion. Valuable because their speed is where we get the energy from, and annoying because they damage the reactor’s walls. How can we solve this?
In the (repeated) graphic above, you see a layer called the blanket between the plasma and the magnets. The fast neutrons are slowed down inside the blanket and their kinetic energy transferred to heat it up. The hot blanket, in turn, is used to heat water, which then turns a steam turbine (just like nuclear fission and coal plants do).
This works in theory, but in practice, it’s hard to build a blanket that is efficient and resilient to damage from fast-moving neutrons.
How can we make tokamaks better?
Aside from the issues around blankets, we still haven’t achieved Q>1. There are two particularly important variables in a fusion reactor that we can control to influence how much energy is released in a fusion reactor:
- R: The radius of the tokamak
- B: The strength of the magnetic field
So, how big do the reactors have to be to get to reasonable Q values? ITER is the biggest-ever international science experiment aiming to reach Q=10. How big is the ITER reactor?
Can you see the person down at the bottom? This thing is HUGE.
Because of its size, ITER has cost tens of billions of dollars, and is taking decades to build. Remember the Q graph from before? Wondering why it stopped? Now you know! The reactors got too big, meaning they take too long to build.
The number on the y-axis here is the so-called ‘fusion triple product’. It’s a rough indicator of how much power a fusion reactor produces and is defined as the product of the three key attributes of any fusion reactor:
ITER is a science experiment, not a commercial reactor. A commercial reactor would likely need to be even bigger. Clearly, increasing the radius (R) isn’t promising. What about the magnetic field strength (B)?
Could stronger magnets make reactors smaller and cheaper?
Inducing a magnetic field requires us to run a current through the electromagnetic field coils on the tokamak. In most materials, the current uses up energy, because some electricity is lost as heat because of resistance. However, some materials - called superconductors - have the ability to let a current pass through them without heat loss as there is no resistance to the current.
Recent work on a type of superconducting magnet called REBCO (Rare Earth Barium Copper Oxide) has allowed the magnetic field strength, ‘B’, to almost double! The limiting factor is now the durability of the steel and concrete holding everything together - at full power, the magnets would rip the reactor apart.
Using REBCO magnets is likely an essential step on the path to affordable fusion reactors with Q > 1.
Importantly, ITER still uses the old, weaker superconductors.
Conclusion
Fusion has always been a ‘technology of the future’, but we really are getting close to Q=1.
Private companies and university labs are now working to integrate REBCO magnets into tokamak fusion reactors. Their progress will be a crucial indicator for fusion’s potential, but reaching Q greater than 1 is not the only problem in fusion research.
There are many open problems in blanket technology, tritium breeding, and reactor protection. This is a massive source of uncertainty. Nevertheless, over the coming decades, we will likely find out whether fusion can be the clean and abundant source of energy we hope it will become.
Next Chapter
