I recently met two people who work at the MIT Plasma Science Fusion Center (PSFC). One works on Alcator C-Mod, the world’s most powerful tokamak. The other works on the Levitated Dipole Experiment (LDX), a new approach to fusion that I think of as the conjugate of a tokamak. I immediately set about cajoling them into giving me a tour, which happened Friday afternoon.
We visited the LDX control room first. We had scheduled our visit to coincide with a time when they were making plasma. After snaking through the industrial corridors of the PSFC, formerly a Nabisco bread factory, we found the LDX control room. The control room is a bare white medium-size conference room, with computers set up on tables around the edges, and another table in the center. There are snacks cluttering the center table, and papers piled on a handful of filing cabinets. In one corner, there are two innocuous black and white television sets that look like 1987, showing nothing. These are wired to the cameras pointed into the vacuum chamber. They are the only live view of what happens during a shot, though the data is also digitized, and everyone is looking at it on their computer screens a few seconds later.
The vacuum chamber can be filled with a number of different gases. While we were there, “el jefe” (his title, written on a piece of paper taped to the back of his standard-issue swivel-chair) called for an Argon shot, and so a burst of argon was allowed into the chamber. The microwave heaters were activated, and the TV screens suddenly lit up. The only source of light in the chamber is the plasma itself, and so the camera sees a diffuse glow, illuminating the chamber walls, the levitating torus, and the spring-loaded catcher positioned beneath it.
In a way, it’s surprising that you can see anything at all. At its highest operating densities, LDX runs at a pressure of about 10^-5 Torr. That’s a better vacuum than low earth orbit; there’s really almost nothing there. To get a visible glow from plasma at that density requires astonishing temperature.
I was surprised to see a lot of flicker in the plasma. My host explained that the flicker is exactly what it looks like: density waves in the plasma, rolling around a loop through the center of the torus. The brilliant thing about LDX is that these fluctuations are themselves stable; they don’t turn into a runaway reaction that throws the plasma against the vessel wall, although they do make life interesting for the levitation feedback circuits.
When they turned off the heating magnetrons, the plasma decayed back to blackness in about one second. My host noted that this was actually due to the high density. At lower densities, the glow persists longer, indicating that containment is better, presumably because the mean free time is longer and scattering events are rare.
After the LDX control room, the Alcator control room was a bit of a shock. It looks like an anarchic version of NASA mission control. The room is a huge, high-ceilinged rectangle, with desks along the edges and two rows of tables down the middle. Every desk surface is claimed by somebody’s stuff. One desk even had two ornate shaded lamps, the better to catch those late-night equations. At capacity, the room must have well over a hundred people, packed in mouse by keyboard.
At the far end of the room is a row of (matching!) modern computers, running identical software, and positioned above them are three enormous projection screens. The leftmost screen is a live video feed from inside the tokamak, and the others present inscrutable diagnostic information in black and green diagrams.
Each “shot” on Alcator is announced by a recorded female voice, referred to by all as “the sexy voice”. “Shot number 31 will commence in 12 minutes”, she said when we arrived. In the meantime, a friendly physics major showed us some footage of previous shots, so we would know what to expect. A normal shot looks much like on LDX: a diffuse glow lighting the inside of the chamber. What’s more interesting, of course, is a failed shot. Tokamaks employ the world’s most sophisticated feedback control mechanisms. Unlike LDX, plasma fluctuations in a tokamak will naturally spiral out of control unless caught and corrected electromagnetically. We watched the recording from a recent failed shot. First the video camera starts to malfunction as charged particles disrupt its electronics, and the image bounces haywire around the screen. Then a shower of sparks, and darkness. In this particular incident, a negatively charged jet escaped the confinement fields, and punched an inch-deep hole in the solid molybdenum inner cladding. This is especially impressive when considering that, as in LDX, the density of the jet was comparable to outer space.
The sexy voice initiated a fifteen-second countdown for the shot, and our hosts warned us to expect a vibration, which, when it came, was unmistakable. Alcator’s primary containment electromagnet draws 15 megawatts during a shot, about as much power as a city. No power line could possibly transfer energy this fast, and so instead, Alcator buffers energy in a giant flywheel weighing many tons, down the street from the control room. For each shot, the power is drawn back out of the flywheel, and the torsional vibrations of the generator shake the foundations.
Once firing was done for the day, we were able to visit the actual apparati themselves. They are both susceptible to similar descriptions: enormous round inscrutable devices several stories tall, surrounded by gratings and staircases for access at different levels. Alcator lives behind a particularly impressive door over a meter thick, next to a tangle of high-power electrical equipment that continues into the distance. Entry to LDX requires snaking through an entrance designed to ensure that radiation cannot have a straight exit path. LDX lives in a warehouse-like room seemingly taller than its width, and topped by a gantry crane. The whole room is carefully RF-shielded to avoid perturbing the measurements.
These guys are awesome. Now I’m definitely rooting for ITER, Polywell, “LDR”, and anyone else working on this stuff.