Once atomic fusion occurs, the heat produced will help to keep the core hot. But unlike a fission reaction that takes place in nuclear power stations and atomic bombs, the fusion reaction is not self perpetuating. It requires a constant input of material or else it quickly fizzles out, making the reaction far safer. And unlike what you might have seen in a recent Batman movie, the chamber cannot be transformed into a nuclear bomb. The neutrons will then be absorbed by the surrounding walls of the tokamak, transferring their energy to the walls as heat, and this in turn will be dissipated through cooling towers.
Because one of the hydrogen isotopes used, tritium, is radioactive (with a half-life of 12 years), the entire site must conform to France's strict nuclear safety laws. And to complicate matters further, the site is also moderately seismically active, meaning that the buildings are being supported on rubber pads to protect them from earthquakes.
These issues, plus the logistics of dealing with multiple nations with their own fluctuating domestic budget constraints, mean that the site won't be ready for the first experiments until 2020. Even then, they will just be testing the reactor and its equipment. The first proper fusion tests, reacting deuterium (a hydrogen isotope abundant in sea water) and tritium (which will be made from lithium), won't take place until 2028.
Those will be the key tests, though. If all goes to plan, the physicists hope to prove that they can produce ten times as much energy as the experiment requires. The plan is to use 50 megawatts (in heating the plasma and cooling the reactor), and get 500 MW out. Larger tokamaks should, theoretically, be able to deliver an even greater input to output power ratio, in the range of gigawatts.
And that is the big gamble. So far, the world's best and biggest tokamak, the JET experiment in the UK, hasn't even managed to break even, energy-wise. Its best ever result, in 1997, achieved a 16 MW output with a 25 MW input. Scale is an extremely important factor for tokamaks, though. Iter will be twice the size of JET, as well as featuring a number of design improvements.
If Iter is successful in its proof of principle mission, the first demonstration fusion plants will be built, capable of actually using and storing the energy generated for electricity production. These plants are slated to begin operation in about 2040 - around 30 years away, in fact...
Despite the seductive promise of finally getting a supply of electricity that's "too cheap to meter", the long wait to readiness and the fact that the technology remains unproven, means that many politicians are hesitant or even hostile to the expensive project. Additionally, because fusion energy won't be ready for decades, even if it works, other low-carbon energy sources must still be pursued in the short-term at least.
But if we do manage to replicate the Sun on Earth, the consequences would be spectacular. An era of genuinely cheap energy – both environmentally and financially, would have far reaching implications for everything from poverty reduction to conflict easement.
It’s exciting to think that the next generation could in some way be fusion powered – perhaps even within the lifetimes of the workman digging below me. But I can’t help but remember the 30-year rule.
Update (14/08): The original text contained factual inaccuracies regarding the fusion reaction within the reactor. This has now been rectified.