Lab Visit Report: Cold Plasmas

Another of the labs I visited while in DC was Steve Rolston’s lab at the University of Maryland. This actually contains the apparatus I worked on as a graduate student, including many of the same quirky pieces of hardware– Steve was the PI (Principal Investigator) for the metastable xenon lab in the Phillips group at NIST, and when he left NIST to take a faculty position at Maryland, he took the apparatus with him.

The xenon lab is now dedicated to work on ultracold plasma physics, which they were just starting when I graduated. The idea is that you can use laser cooling to accumulate a large number of atoms in a magneto-optical trap (MOT) at reasonably high densities– we used to get something like 107 atoms at a density of 1010 per cubic centimeter. That’s something like a billion times less dense than air, but it’s high enough for there to be frequent collisions between the trapped atoms, which is what I studied for my thesis.

It turns out that if you take a green laser (around 500 nm light) and shine it into a xenon MOT, you can make ions. The atoms will absorb one photon from the MOT laser, and one from the green laser, and those two photons can provide enough energy to either strip an electron off an atom altogether, or to excite it to a very loosely bound state (a “Rydberg state”), where it can gently fall apart. If you use a short, intense pulse of green light, you can excite most of the atoms in the MOT, and quickly convert the trap from neutral atoms to ions and free electrons, which is a two-component neutral plasma, at the same very low temperature and relatively high density of the original MOT.

We used to joke about there being two approachs in cold-atom physics: the “French Paradigm” after the groups of Jean Dalibard and Claude Cohen-Tannoudji at ENS in Paris, and the “NIST Paradigm,” after our group. The French groups are famous for their exacting calculations, and in the French paradigm, you sit down and calculate exactly the effect you expect to see, then you go in the lab, and measure it to agree with the theory.

The NIST paradigm, on the other hand, starts with the sentence “Hey, we have this other laser– I wonder what happens if we stick that in there?” The ultra-cold plasma experiments are a shining example of the NIST paradigm– when we started doing this, we just sort of blasted the green laser in there, to see what would happen, and we found all sorts of interesting physics.

The earliest measurements are in the preprint linked above, and subsequent experiments with the xenon apparatus looked at plasma oscillations and measurements of the plasma temperature. It turned out to be a much richer system than we had expected, and we really didn’t know what we were getting into when we started looking at the system back in 1999. It spawned a whole new sub-field, which is summarized in a really nice review article by Tom Killian of Rice, who was a post-doc at NIST back in the day, and has really been one of the leaders in the field.

The basic technique used in the xenon experiments is the detection of charged particles. The MOT is placed between two metal screens (90% transparent, so they don’t really affect the trapping lasers) that can have a large voltage placed on them. Ions or electrons in the plasma are pulled out by the resulting electric field, and accelerated to charged-particle detectors above and below the cloud. Atoms in high-energy Rydberg states can be ionized by the field, and the amount of electric field required to ionize them depends on the energy of the state, providing a way to measure the distribution of states.

Recent work in the Rolston lab has dealt with recombination of the plasma, where an ionized atom will capture a free electron (during a collision with a third particle, either an ion or an electron), returning to a neutral state. The rate at which this happens depends very strongly on the temperature of the plasma– the more energy the electrons have, the harder it is to capture one, so colder plasmas recombine at much higher rates than warm ones.

They measured the recombination rates over a fairly wide range of parameters, and found that it agreed very well with the theoretical prediction. As a result, they were able to turn things around, and use the measured recombination rate (which is easy to measure by looking at the charged particles coming out of the plasma), and use the rcombination rate to measure the temperature (which is tricky to measure otherwise). They see some interesting cooling behavior, with the electron temperature eventually ending up below 1K.

Another interesting effect doesn’t turn up on the arXiv, but it sounded pretty cool. In certain experiments involving applying a magnetic field to the plasma, they saw weird periodic “spikes” of electrons emitted from the plasma. This originally looked like noise, but after a week or two of searching for the noise source, they convinced themselves it was real, and the signature of some sort of instability in the plasma.

This is one of the ways in which the “NIST Paradigm” gets kind of weird, because it turns out that there are dozens if not hundreds of different kinds of plasma instabilities out there. They’re well known to plasma physicists, of course, but not to atomic physicists dabbling in new fields. It turns out, though, that what they were seeing wasn’t one of the common instabilities– in fact, it had some of the local theorists stumped, until somebody stumbled over an old Russian paper about “Hall Thrusters”. These are a type of ion thruster that was popular with the Russian space program, but not so much in the west, until recently. It looks like the instability they see may be related to an instability that’s been seen in these Hall thrusters.

So, this is what “Hey, let’s try the other laser!” leads to– you start with a simple system of cold atoms, and end up learning about obscure Soviet-era spacecraft propulsion techniques. It’s what makes doing physics in the NIST paradigm so much fun…

Comments

  1. #1 perry
    March 11, 2008

    Bill P, Steve, Trey and all of those folks are damned good! I think Luis did his sabbatical the year Steve and y’all got that working. I know Bill has great physical intution as well as an ability to know, or figure out quickly, what questions will have interesting answers. The cool thing is that they have tons of lasers and stuff so they can DO what they dream about! Of course the French group is no gang of slouches either, in terms of intuition and gear.

    I hope to get over to Md. for a few days later this semester, whilst me poor grad students are doing a take home quantum exam :-)

    What is the temp for Doppler/Sisyphus for Xenon, before you blast it? 1K may be cold for a plasma but not so much compared to cold atoms. But you did just blast the things after all…….

  2. #2 Chad Orzel
    March 13, 2008

    I think Luis did his sabbatical the year Steve and y’all got that working.

    Yep. He came down to do that experiment and one other, and also forced us to clean up the rat’s nest of BNC cables running all over the lab. For a brief, shining moment, we had nice, clean, organized patch boards handling all our connections. That lasted about two weeks.

    I know Bill has great physical intution as well as an ability to know, or figure out quickly, what questions will have interesting answers. The cool thing is that they have tons of lasers and stuff so they can DO what they dream about! Of course the French group is no gang of slouches either, in terms of intuition and gear.

    I’m slightly in awe of the precision with which the French manage to do everything. Their theory curves always seem to go exactly through the middle of all their data points. I saw Alain Aspect give a talk once where he showed a data set, and compared it to successive theoretical calculations, each adding a higher-order correction, and honestly, I would’ve stopped two models before he did, and said “Yep, looks pretty good! Next project…”

    What is the temp for Doppler/Sisyphus for Xenon, before you blast it? 1K may be cold for a plasma but not so much compared to cold atoms. But you did just blast the things after all…….

    The Xe MOT was usually at around 100 microK or so. That 1K figure is for the electron temperature, which (as we found out) is considerably higher because the electrons pick up some extra energy in the ionization process. I think the limiting factor is the bandwidth of the laser pulse used to do the excitation.