One of the Mysteries of the Universe involves very heavy atoms like Uranium 235. If you sort-of-nudge it by adding a slow neutron to it, the thing splits apart. So how did it manage to form, without splitting as a result of the energy of the particles that combined to make it? Just a little bit too
much energy of those particles...and poof! --no heavy atom of U-235.
Nevertheless, the fact that Nature managed to do it means that we should be able to do such things, too. And we actually do (carefully!); that's how we have managed to extend the Periodic table up to the current total number of something like 114 elements, with most of the ones heavier than uranium being quite unstable. So....
Consider a particle accelerator designed to accelerate deuterium nuclei. Each such nucleus is a particle that consists of one proton and one neutron, and the electric charge on the proton makes it very easy to accelerate in a highly controlled way. That means we should be able to give each deuteron in the beam a very precise amount of energy.
Now imagine we have four of these accelerators. We arrange them such that each is aimed from the corner of a tetrahedron (a regular shape that only has four corners; see link), toward the center, within the volume of that shape.
We now fire all four accelerators at exactly the same time, one deuteron each. We make sure the four deuterons all arrive at the empty center point at the same instant. If we have to adjust the acceleration speeds slightly to ensure it, then fine. We DO have the timing-tech (but perhaps not yet the aiming tech) to be able to do this. (Not to mention, the art of particle-acceleration has improved greatly over the years; these accelerators need not be super-big and super-powerful; remember that one type of fusion reactor can sit on a tabletop --see link.)
The goal of the collision, of the four deuterons, is to construct a nucleus of Beryllium 8, which as you might suppose, will consist of four protons and four neutrons. What you might not know is that Be-8 is extremely unstable, and almost instantly breaks apart into two "alpha particles", each of which is an ordinary nucleus of the atom Helium 4. EXCELLENT!
It is excellent because the reaction is perfectly "clean"; there are no neutrons produced. That is, if we used only two particle accelerators, to directly construct one He-4 nucleus, it mostly won't work, because the fusion process yields a lot of energy that MUST be carried away somehow. (I note in passing that when we use four accelerators to make Be-8, each accelerator needs to be somewhat --but not hugely-- bigger and more powerful than the two accelerators that might try to make He-4, because there is more total electrostatic repulsion to overcome.)
The normal mechanism for carrying away D+D fusion energy is for the result of the reaction to NOT be He-4; instead we usually get Helium 3 and a neutron, or Hydrogen 3 and a proton (50/50 probability). Each of the two resultant particles (say He-3 & n), from the fusion, carries away some of the energy. And, we also don't get anywhere near the total energy that COULD have been released, if He-4 had been the direct result instead. That's because there is no easy way to "balance" the huge amount of energy that just that single nucleus has become energized with. One time out of a million it will emit an extremely energetic gamma ray, and successfully stabilize, but most of the time the He-4 nucleus just can't handle all that energy, so it breaks into something like He-3 & n.
However, we may just about always get ALL that energy of fusion, when we construct Be-8 from four deuterons, and let it break down into two He-4 nuclei! This will be because the two nuclei can between them carry away from the reaction all the energy of the reaction in a balanced way.
There are a couple of big caveats with this Idea, of course.
First, particle-accelerator fusion is traditionally unworkable because it cannot be "scaled up" to accurately handle large quantities of fusing nuclei. Ideally we would like to "stream" them quickly one-after-another toward the fusion site (because each particle can still be accurately aimed), but no particle accelerator design existing can actually do that. One at at time, or clumps at once, OK. But not a nice steadily-accelerating stream of particles.
On the other hand, if the particle accelerators we actually need, to fuse deuterons into Be-8, are fairly small and relatively inexpensive, then what if we had a lot of them? For example, a dodecahedron (see link) has twenty corners, which can be selected as "five groups of four". That means we could activate four of them to start accelerating four deuterons, then activate four more, followed by four more, and so on.
Remember that in each case the total amount of energy that accelerates four atomic nuclei is relatively small, compared to the energy that can be released by the fusion reaction. We merely need extreme precision of aiming, to make sure that the 4H2->Be8->2He4 reaction always, always happens.
Now, IF the total acceleration time was small enough, the first four accelerators would be done and ready for shooting the sixth group of four deuterons, just after the fifth group of accelerators had been activated. If not, well, then, a shape known as a "buckyball" has 60 corners allowing for 15 groups of 4(!). The buckyball shape is associated (see links) with a type of building known as a "geodesic dome", thus giving us the name for this Idea (the reactor needs to be inside a building, right?).
The other big caveat involves the two He-4 nuclei that are produced. They will have a lot of energy and quickly leave the collision site where the Be-8 nucleus had been temporarily created, but what directions are they travelling??? It is likely that they will, at least some of the time, interfere with some of the "next" deuterons that are approaching the collision site. This automatically means that despite extremely accurate aiming of the accelerators, the desired reaction will sometimes not happen.
It MIGHT be possible to influence the direction the two alpha particles take, by unbalancing the four original deuterons in a special way. They still need to all arrive at the same place at the same time, of course, but if one or two of them arrived a little faster than the others (and had been just slightly delayed in getting started), then the total kinetic energy of the four particles will not balance out; there will be some leftover motion that might be "aimed" to avoid the next four deuterons.
It remains to be seen whether or not that will make this Idea workable. But for the HalfBakery, it's perfect.