Over the past year or so, we have seen massive strides in the construction of ITER, which we are sure will bring long (up to half an hour is a long time in plasma physics!) periods of net energy gain from fusion. The NIF (America's National Ignition Facility laser fusion experiment) has also made some significant progress. But as we see the building which will surround ITER's beating toroidal heart rise from the ground, I've noticed a somewhat troubling trend elsewhere.
I have seen more and more proposals for what promise to be small, cheap and easy-to-realise fusion devices. With US defence giant Lockheed Martin claiming they could have a working “truck sized” prototype fusion reactor in 5 years, these proposals are no longer just from lone inventors writing from their basements. Could any of them have any credibility? Might they have stumbled on something simple and brilliant, which the mainstream fusion research community have missed?
Firstly, let's consider the conventional wisdom of Deuterium-Tritium fusion: all else aside, this is the easiest reaction to achieve (highest cross section). That's why I think we're only likely to first achieve breakeven from fusion with this reaction; exploring others is going to be useful only after we master it – we should “learn to walk before we run” with regards to fusion.
What about how we actually try to do fusion? Certainly, we need to somehow overcome the repulsion of positive nuclei to make them fuse. There have been some rather spectacular suggestions for processes which can cheat the large Coulomb repulsion between nuclei, many of which rely on unproven science. The burden of proof is certainly on the people making such claims, but I wouldn't hold my breath.
So, we are left with having to accelerate the nuclei to such high speeds that they overcome repulsion. In fact, there is an optimal speed to achieve a fusion reaction. Proposals keep cropping up to do this directly – everything from firing frozen pellets together to colliding beams as in a particle accelerator – but they are fundamentally flawed. The problem with having many particles all travelling and colliding at one speed is that they will tend to redistribute their energy into a random, but predictable thermal profile.
So we are left with having to do thermonuclear fusion – having the nuclei at a temperature where the most “popular” speed is ideal for fusion. This leads us to the typically high temperatures (100 million °C) we see in fusion, making our fusion fuel incredibly difficult to contain – imagine stopping your cooking pot boiling over if it were a million times hotter. Do we need a machine as large as ITER or NIF to contain it? Many experiments have been done over the years which very much suggest so. In any case, if we use the Deuterium-Tritium reaction, we will also need to stop and utilize the neutrons produced, or work even harder to confine the fuel if we use another reaction. We think we probably need a few metres of material to do this, which pretty much rules out anything the size of a truck.
What I've talked about is hardly groundbreaking – every physics student can tell you as much. I think that the phenomenon of budget fusion ideas is something like proposals for perpetual motion machines – perhaps their authors have forgotten some basic physics.