The Making of “Creation of an Ultracold Neutral Plasma”

As mentioned in the previous post, the cold plasma experiment was the last of the metastable xenon papers that I’m an author on. My role in these experiments was pretty limited, as I was wrapping things up and writing my thesis when the experiments were going on.

The main authors on this were Tom Killian, now running his own cold plasma lab at Rice and Scott Bergeson, now running his own cold plasma lab at BYU. Scott was a pulsed-laser expert with a remarkably cavalier attitude toward things like anti-reflection coatings on vacuum windows, and Tom came from Dan Kleppner’s hydrogen BEC project, with the very methodical approach you would expect from them. Scott got the laser tuned up and kept it running, and Tom’s the one who really nailed down the model of what was going on. Simone Kulin was another post-doc on the project, and Luis Orozco was on sabbatical from Stony Brook that year, working in our lab.

My role in this project was pretty much to be a reference for the quirks of the xenon apparatus. I was writing up my thesis at the time, and had pretty much shifted my schedule back by 3-4 hours– I would come in just before lunch, spend the afternoon reading over the previous night’s drafts, answering occasional questions, and procrastinating like a mad thing by talking to anyone foolish enough to make eye contact. After everyone else went home, I would buckle down and write and revise the thesis, often until four in the morning, when I’d go home and collapse for a few hours before starting over.

At this point, I had been working on the xenon project for almost six years, so I knew all the ins and outs of the apparatus. The others knew how to work it, but there were occasional weird issues that would pop up to hamper the operation of the trap, and they usually turned out to be things I had seen before. I also took part in the paper torture, and the discussions of the model of plasma formation, which made a nice change of pace from thinking about the same collisional physics problems over and over.

As a result of my detached state, I don’t have that many stories of the making of this experiment. A couple of amusing anecdotes do come to mind, though:

One is a sort of joke that you have to be a physicist to really appreciate. When we started doing these experiments, we used the same ion detector that we had been using for the collision experiments, which was way off on one edge of the chamber, and not particularly optimized for efficient detection. We got good signals out of it, but since we needed to put field plates in anyway, we rigged up a new system that put the detector immediately above the trap, in the center of the chamber.

At which point, all the signals started to look like crap, because we were absolutely killing the detector. At one point as we were trying to figure out how to proceed, I called the manufacturer, and told them what was going on. “What’s the count rate?” they asked. “About 10⁸ per second,” I replied.

There was silence. “For how long?”

“About eight hours a day, give or take. That’s our base rate.”

Another long pause. “And it still works?”

We ended up switching out of pulse-counting mode, and turning the detector voltage way down, lowering the efficiency, and getting us back to a regime where the detector behaved nicely. At around this time, I was at a meeting talking to a bunch of people from other groups, and said something like “Yeah, we’ve been having a miserable time with the experiment, because the signal rate is so high, it’s killing our detectors.”

I think I was lucky to get out of there without being physically beaten. Telling a bunch of physicists “The signal we’re looking for is way too big to measure” is not a way to make yourself popular.

The other memorable thing was an equipment failure, in the pulsed laser. The laser we used was a Nd:YAG laser, which produces pulses of light at about 1064 nm. This was focussed into a crystal to double it to 532 nm, and then into another crystal to generate the third harmonic in the UV, which then pumped a dye laser to make tunable light at 514 nm. Each of these steps is a low-efficiency process, so we needed to start with a pretty high pulse energy– nothing all that huge by pulsed-YAG standards, but more concentrated energy than us CW-laser types were used to.

This was driven home for us the day that a gasket in the cooling system failed. The gasket failure let water leak out inside the laser, and it formed a little waterfall over the face of the YAG rod. This acted like a little lens, and brought the 1064 nm light to a focus on the edge of one of the laser mirrors, which physically blasted the mirror out of its mount. When we shut the laser down, and looked inside, the mirror was twisted halfway around, and barely held from falling into the puddle of water in the bottom of the case.

It’s a good demonstration of how much punch those lasers pack into a hundred nanosecond pulse…