Creation of an Ultracold Neutral Plasma

ResearchBlogging.orgThis is the last of the papers I was an author on while I was in grad school, and in some ways, it's the coolest. It's rare that you get to be one of the first people to do an entirely new class of experiment, but that's what this was. It kicked off a new sub-field (or sub-sub-field...), the history and status of which was written up in Physics a little while back.

The ultracold plasma experiment may be the ultimate version of what we jokingly called the "NIST Paradigm" of cold-atoms physics research, which could be summarized as "I wonder what will happen if we stick this other laser in?" It's based on the realization that in the metastable xenon system we were working with, it's possible to ionize the atoms by hitting them with a pulse of green light. A photon at about 514 nm, combined with an 882 nm photon from the trapping laser, gives enough energy to strip one electron off a xenon atom.

At that point, the trapped sample of atoms becomes a plasma-- definitions of "plasma" vary, but it's essentially just a vapor composed of charged particles. The plasma we made was neutral, as there were equal numbers of free electrons and positive ions created (more or less by definition), and it was exceptionally cold, because the atoms we started with were at a temperature of something like 10 microkelvin (10 one-millionths of a degree above absolute zero), and the laser pulse does not significantly heat the atoms.

Of course, we didn't know quite what to expect when we started the experiment. We just kind of blasted the laser in, and what we saw was this:


So, what's going on, here?

The graph shows the number of electrons detected at a charged-particle detector mounted above the trap as a function of time after the green laser pulse. The different traces represent different intensities of the green laser, and increasing intensity should mean more ionized atoms. A short time after the green laser pulse, we applied a small electric field, which increased linearly in time (the diagonal line from left to right), to sweep any electrons still in the trap toward the detector.

What we see is that a big clump of electrons come out almost immediately, and for low laser power (top curve), that's pretty much it. For higher green laser intensity, there's a second lump of electrons that only come out when we start applying a field, and the number of these increases as the laser intensity increases.

What's going on here? Well, the initial pulse is due to a smallish number of electrons that pick up enough energy to fly out of the trapping region immediately, and get picked up by the detector. If the total number of ionized atoms is small, that's the whole story-- electrons are stripped off atoms, and then fly away.

When there are a lot of ionized atoms, though, the departure of the first group of electrons changes things for those left behind. The initial sample is perfectly neutral, with as many positive ions as negative electrons. As electrons leave, though, the remaining sample is left with a positive charge, that increases as more electrons leave-- the positive ions have a couple hundred thousand times the mass of the electrons, and thus move very slowly. The positive charge of the ions remaining in the trap region (they're not trapped, just slow) will pull on later electrons, and hold them back.

At some point, the positive charge left in the trap region is enough to keep the remaining electrons from escaping, and they become trapped. At this point, the ionized sample turns into a trapped "neutral" plasma (it's not perfectly neutral, as there's a net positive charge, but the charge imbalance is pretty small). The plasma sticks around for some time (several microseconds), and the trapped electrons don't come out to be detected until after the electric field is put on to pull them out.

How do we check that this is really what's going on? Well, this model suggests that there should be a critical number of ionized atoms at which we start to see that second lump of trapped electrons, indicating the formation of the plasma. This number should depend on the exact energy of the green photons-- if we go to a slightly higher frequency, we give more energy to the electrons that are stripped off the atoms, and thus need a larger positive charge to hold them back.

If we look at the number of trapped electrons as a function of the number of ionized atoms for several different laser frequencies, this is exactly what we see:


The graph on the left shows the raw number of trapped electrons for different numbers of ionized atoms. Starting at small numbers, we see that there are no trapped electrons at the lowest values, and as we increase the total number of ionized atoms, we hit a point where the number of trapped electrons increases very suddenly. That's the point where the neutral plasma forms. If we look at different laser frequencies, we see that sharp increase occurs at higher numbers for higher laser energies.

Of course, the ultimate test of whether you really understand something is to see how it matches up with theory. It's fairly easy to calculate the critical number at which the plasma should form, and we can re-plot the data with everything scaled to the critical number for the appropriate laser frequency. If we do that, all the data should fall on the same curve, as they do in the graph on the right. The solid line is the prediction of a simple theoretical model, and matches the data beautifully.

As plasmas go, this is pretty tiny-- you can see from the graph that we're only dealing with tens of thousands of atoms, so you might wonder why anyone should care. These experiments are interesting not for the size of the plasma, but for its very low energy. The ion temperature is only a few millionths of a degree above absolute zero, and the electron temperature is higher, but still very low. This means that the kinetic energy of the particles making up the plasma is actually smaller than the energy due to the interactions between neighboring charges.

Such a plasma is referred to as a "strongly coupled" plasma, and it's a very difficult regime to reach with normal high-temperature plasmas. "Strongly coupled" plasmas have been made in ion storage rings, and through experiments involving exploding wires, and they occur in nature in places like the cores of white dwarf stars. Having a relatively easy way to explore this regime is a really attractive, and there's now a community of people (groups at Rice, BYU, Maryland, UConn, and Michigan, and I'm probably forgetting someone) studying these sorts of ultra-cold plasmas, and turning up all sorts of interesting stuff.

And it all starts from "I wonder what would happen if we hit the atoms with this other laser?"

T. C. Killian, S. Kulin, S. D. Bergeson, L. A. Orozco, C. Orzel, S. L. Rolston (1999). Creation of an Ultracold Neutral Plasma Physical Review Letters, 83 (23), 4776-4779 DOI: 10.1103/PhysRevLett.83.4776


More like this

Really interesting - thanks for sharing. Of course it sounds so simple when you describe it in a couple thousand words, but I can just imagine the hours and hours of frustrating labor that people put in to get the data.

It's great to see examples of the "let's just do this and see what happens" scientific paradigm. I'm reminded of Feynman and his story of how figuring out the wobbling plate in the Cornell cafeteria reawakened his love of physics and led to his Nobel prize work. Sometimes the most important work comes from quite unassuming beginnings.

That's also why it bothers me when people say the "purpose" of the LHC is to find the Higgs. It doesn't have any purpose other than to smash some really fast things together and see what happens! And there's absolutely nothing wrong with that.

From my perspective, "I wonder what would happen if we hit the atoms with this other laser?" is a completely random thing to think. But then again, I have no experience in hitting things with lasers.

Is there a level of familiarity with this kind of thing that allowed you to suspect that something interesting would happen when you added another laser? Or was it really just fooling around with a laser?

Is there a level of familiarity with this kind of thing that allowed you to suspect that something interesting would happen when you added another laser? Or was it really just fooling around with a laser?

There was a little more to it than just "I wonder what would happen..." Steve Rolston, the PI on the xenon papers, had thought about this a bit, and the laser wavelength we were using was carefully chosen in advance-- it's not like we just had this laser lying around, and decided to stick it in for no reason and stumbled across the plasma thing. The goal was to make an ultra-cold plasma all along.

What I said isn't too much of a distortion, though, in that we really didn't know what we were going to see at all. We got signals immediately, and then spent several months trying to figure out what the hell they meant. It took quite a while to sort out what the various components of the signal meant, and confirm that we actually were seeing a plasma.

That makes sense.

I'd be curious to hear any lab stories where you (or, y'know, your friend) tried something like this without such glowing success. Especially if they have an interesting reason why.

Wait, TK?

Holy shit.

See! People want to hear about failed experiments! Maybe you can trick somebody else into starting the Journal of What Went Wrong?

By Andrew York (not verified) on 02 Oct 2008 #permalink

This is probably a stupid question, but how do you reach the first ionization potential 10-11 eV, if I recall), with a combination of 514 and 882nm lasers.

This is probably a stupid question, but how do you reach the first ionization potential 10-11 eV, if I recall), with a combination of 514 and 882nm lasers.

The IP for xenon is 12.4 eV, if I recall correctly.
We get there with 514nm + 882nm because the trapped atoms we start with are in a metastable state, 8.32 eV above the ground state. It's got a lifetime of 40+ seconds, and serves as an effective ground state for all the trapping and cooling that we do.