The international thermonuclear experimental reactor (ITER) is a highly ambitious project that is intended to remove the last obstacles between civilization and our very own pet suns. During the last year, I have become, and continue to be, peripherally (very peripherally) involved in ITER. It is a huge project and, as yet, my involvement has been limited to attending scientific discussions and writing proposals associated with a side project. Nevertheless, I have gained some feeling for the challenges facing ITER, and I thought it might be interesting to share those with Ars readers.

The consequences of confinement 

First let’s look at what appears to be solved: magnetic confinement of the plasma. None of the graduate students, post-docs, or senior scientists discuss the confinement system except in passing. The impression from these meetings is that this was largely a solved problem on paper, and ITER would really test how well these solutions worked. To put it more succinctly, experimentalists do not have the necessary tokomak to test the physics further than they already have, and theory says it should work. The theorists may be wrong and, if they are, ITER will provide the perfect testbed to get it right. 

The design of the physical geometry and the parameters of the plasma-wall interactions were dictated by the requirements of the confinement system. In as much as the walls were considered at all, it was to the extent of whether they would survive long enough to complete the experiment. This has led to a huge scientific challenge—how to make the walls resistant to immense plasma flows.

Magnetic confinement has leaks wherever the field terminates. It is impossible to design a field without these leaks, so the structure of any magnetic confinement system must take this into account. The solution used with ITER is to have specific walls where the field terminates, which confines the leaks to a well-known area. These walls will be subject to a huge flows of very hot plasma that will lead to both surface etching and ions implanting themselves deep into the wall material.

That leads to two further problems: how do we get rid of the dust? And how do we deal with the contaminated wall panels? ITER will eventually use tritium (a radioactive form of hydrogen) as the fuel for the reaction. Some of the tritium will end up embedded in the walls and in the dust. The radioactivity levels are expected to be such that the cleaning will have to be done remotely, and the walls will be quite radioactive when the material needs to be changed.

This is a huge technical challenge that can probably be dealt with. Unfortunately, even though tritium has a relatively short half-life, it rather destroys the idea that fusion is a clean technology. It will certainly be cleaner than coal-fired and fission-powered reactors, but it is certainly not going to be as green as fusion power is generally viewed. This, of course, can change because it is not essential to use tritium—tritium is the easiest to get ignited, so ITER will use that—so future reactors could be much greener than ITER. But, as I will discuss a bit later, it is still a big problem.

Clearly, even if ITER is a complete success in terms of fusion, the plasma-wall interaction problem may yet remain unsolved. It is still not clear what wall material will work best in ITER—indeed, there may not be a best choice for ITER (just a least-bad choice).

To help sort this out, a testbed for the plasma-wall interaction has been built. Even this presented a serious challenge, because the amount of plasma hitting the wall in ITER is expected to be huge. A pilot facility has been operating for some time, but at temperatures and plasma flows significantly lower than those expected from ITER. In 2007, scientists finally succeeded in generating ITER-like conditions in a plasma source, so the construction of the main test facility is now underway. Nevertheless, it is becoming pretty clear that future fusion reactors are going to have to place a higher priority on plasma-wall interaction than on the magnetic confinement configuration.

Paying the price 

One of the problems faced by all big science projects like ITER is planning. Scientific experiments by their very nature have a lot of unknowns, making it difficult to forecast costs; even making design decisions can be problematic. For instance, absolutely every observation port planned for ITER is already reserved for certain instruments. This is a necessity but, particularly with the plasma-wall interaction, there is a lot to be discovered—and a lot being discovered—meaning that it is very difficult to determine the best instrumentation choices.

To make matters worse, it costs money every time an instrument is changed. One of the past year’s big stories described how ITER project leaders were going back to contributing governments to present a final bill that would be much more than expected. This is, in large part, due to the fact that no one knows what wall material to use or how to monitor those walls. To cope with this problem, ITER has been redesigned so that it will be possible to change the materials used in the walls.

The thing is, none of these sorts of problems are different from those faced by any other big science project. So, why has ITER languished? When we look at big science projects—the LHC springs to mind—they are sold on their ability to succeed. The LHC and neutrino observatories will discover new physics; it is simply impossible for them not to because they are looking at things we have not looked at before. Large banks of sequencing machines and huge databases will lead to biological breakthroughs, because it is simply impossible to gather that much new information and not learn something new.

But ITER doesn’t fit in this mold. It is unlikely to discover any fundamentally new physics. It certainly won’t generate power. It will generate some radioactive waste. It might even lead us to the conclusion that fusion power is not economic. In a sense, it is designed to fail—a power station that uses power—but to fail in such a way that we learn enough to succeed. Unfortunately, governments see the price tag, the “if” statements that go with every science experiment, the lack of certainty… and go weak at the knees. Imagine dropping a few billion on a project, only to have a completely null result—something that is simply not possible for most large science projects because they are instruments.

Scale is really the difference between fusion power and all other alternative energy sources. We cannot build a cheap, small-scale fusion reactor and draw conclusions about its feasibility. Instead, a very large reactor needs to be constructed simply to see if it works. Risk-averse governments see that a price of failure that is very much less if one invests in solar energy, wind power, or tidal power, meaning that ITER gets the short shrift.

My personal opinion is mixed. I really don’t know if ITER would succeed. Nevertheless, energy demand has grown consistently at around three to five percent per year in almost every developed nation on Earth for as long as we have records. Even removing environmental considerations, it is unlikely that this demand simply can be met by current mature technologies. Given this, it is probably worth finding out if fusion research is worth it.

In fact, by my reckoning, we should fund ITER to the level requested by the scientists, in order to discover if it is ever going to be worth building fusion reactors. If it is, that is money well spent. If it isn’t, you have a good reason to scale back fusion research for the foreseeable future and invest money elsewhere. This is probably a good deal cheaper than what we’ve done so far: partially fund smaller scale experiments for an indefinite period of time.