Fusebox

I've been stuck for several weeks on an idea for a post, as well as doing some family visiting, so I've haven't been here for a while, but during one of those visits, I got another great question from Papou (paraphrased): I've been reading about plans for fusion power plants. How do they work?

Fusion is a type of nuclear power, but unlike the typical fission plants in use today, does not require radioactive elements such as uranium. In a fission reaction, the fuel naturally releases particles which heat the surrounding material, as well as knocking more particles off other parts of the fuel. Fission power relies on the natural release of neutrons from the fuel to spur further reactions. However, fission results in nuclear waste – radioactive materials not strong enough to supply power, but still dangerous to living things.

Fusion works along the opposite path: Forcing two atoms together until they bond, which releases energy. This is made difficult by the structure of an atom: protons and neutrons are bound together by the strong nuclear force in the atom's nucleus, which is surrounded by a cloud of electrons held by the electric Coulomb force. When we try to force two atoms together, initially the Coulomb force between their electrons will resist, until they get close enough for the strong nuclear force to pull the nuclei together.

To put some numbers to this idea, we can look at the potential energies of the two forces involved. The Coulomb potential is given by

where the first term is some constants, the qs are the charges of the two particles, and r is the distance between them. The nuclear potential is a bit more complicated, and requires making some approximations. I used the Reid potential, which uses terms of the form exp(-m*r)/m*r. We can look at how these two potentials behave near the nucleus:

The x-axis is measured in femptometers, or 10^-15 m, about the size of a proton. Notice that the Reid energy has a large dip at around that size. This dip is enough to make up for the rising Coulomb energy, and it's what makes fusion power an attractive energy source. The trouble is that at larger distances, the Coulomb force pushes the atoms apart. We can look at the point where the forces cancel:

You can read a potential energy plot like an elevation map: Objects go downhill, faster for steeper slopes. Anywhere flat, an object will stay still, but if it's the top of a hill, like above, it's an unstable equilibrium and will fall one direction or the other with a slight disturbance. Here, you can see that the crossing point appears between 5 and 6 fm. Getting atoms that close is extremely difficult, and a major obstacle to fusion power.

One way to do it is by heating and compressing the material. This is how stars like our Sun achieve fusion: Gravity squeezes the hydrogen into a tight area, increasing both the temperature, and the chances of atoms hitting each other. The fusion then releases more heat, continuing the process. To scale that down to a power plant though, we need a different way to squeeze the atoms together.

The solution to that is to create a plasma. By heating the hydrogen to a little over 150,000 °C, we can make it ionize, or separate into individual protons and electrons. That makes it susceptible to electric and magnetic fields, which we can use to compress it. This is the method used by tokamak reactors, including the ITER, which is set to be the biggest such reactor beginning operation in 2025.

The problem such reactors need to overcome for power generation is that heating the hydrogen into the plasma state, and containing it with the electromagnetic fields requires a significant input of power. If everything works correctly, the output from the fusion can surpass that, but so far the best we've been able to do is generating 70% of the input power, a net loss. The plan for ITER predicts it will achieve a 10-fold increase in generated power over the input. It's exciting to see the progress being made in this field, but we should keep in mind the tangible results may still be a long way off. Thanks for another great question, Papou!