How to Make Slow Atoms and Molecules 1

Consider the air around you, which is hopefully at something like "room temperature"-- 290-300 K (60-80 F). That temeprature is a measure of the kinetic energy of the moving atoms and molecules making up the gas. At room temperature, the atoms and molecules in the air around you are moving at something close to the speed of sound-- around 300 m/s (give or take a bit, depending on the mass).

If you're a physicist or chemist looking to study the property of these atoms and molecules, that speed is kind of a nuisance. For one thing, the atoms and molecules tend not to stick around long enough to interact with a given atom or molecule for very long-- in ten milliseconds, they move 2-3 meters, which is about as big as atomic and molecular apparatus gets. If you're studying their properties via spectroscopy (which is the best tool we have for looking at what's going on with atoms and molecules), that speed causes another problem: the Doppler shift. The frequency of the light absorbed or emitted by a moving atom or molecule is shifted from the value for a stationary one, and a gas of atoms will contain atoms and molecules moving at a variety of different speeds and directions. This tends to wash out fine details of the spectra.

In order to study atoms and molecules in detail, then, we need some way to make them slow the hell down. Physicists and physical chemists have developed a wide range of techniques for making atoms and molecules move slowly, many of which are quite ingenious. So let's take a couple of posts to look at some of the tricks you can play to slow atoms and molecules down. In this post, we'll look at cold liquids and mechanical techniques for slowing the motion of atoms and molecules.

Cryogenics: The simplest and most obvious of these techniques is just to surround the atoms or molecules you want to study with cold stuff. The basic procedure for making things cold is the same thing that's used in your refrigerator: you compress a gas by a large amount, which tends to heat it. You cool the compressed gas down as much as you can, and then release the pressure. As the gas expands, it cools, leaving you with colder gas than you started with. You can repeat this as many times as you like.

This process will allow you to produce liquid nitrogen from air, which boils at 77 K (at atmosphereic pressure), and is plenty cold enough for all kinds of cool demos. If you keep working, you can make liquid helium, which boils at 4K, and is cold enough that most metals becomes superconductors. If that's not cold enough, you can pump on the helium to lower the pressure around it, and end up with liquid helium at a fraction of a Kelvin.

Once you've got one of these extremely cold liquids, you just take a container full of the substance you want to study, and dunk it in the appropriate cryogenic liquid. If you keep it there, it will eventually reach equilibirum, and the atoms or molecules will be moving very slowly.

This method is limited by the fact that, by the time you reach liquid helium temperatures, everything other than helium has frozen solid. This complicates matters, because the strong interactions between atoms or molecules in the solid state obscures the individual atomic or molecular properties even more than their room-temperature thermal motion. If you want to make atoms cold using cold stuff, you have to be careful not to make them too cold, or you won't have a gas any more.

Of course, as Emmy and I discussed regarding BEC, you can get around this problem if you work at really low densities. If the vapor of interest is sufficiently dilute that the atoms or molecules only rarely encounter one another, they won't interact enough to form a liquid or solid, and you can keep them in the gas phase.

The best application of this sort of technique is probably the "buffer-gas cooling" method pioneered by John Doyle's group at Harvard. The technique uses a dilution refrigerator to cool a glass and metal cell to less than 1K. The cell contains a dilute vapor of helium, into which a sample of other atoms or molecules is introduced, generally by blasting them off a solid piece with an intense laser. Once in the helium vapor, the atoms or molecules will "thermalize" with the helium, reaching a temeprature of a few hundred millikelvin. This is cold enough to be contained in a magnetic trap, made by applying a large spatially varying magnetic field. The atoms or molecules of interest are held by the trap, while the less magnetic helium is not. The helium can then be removed either by rapidly lowering the temperature of the cell (which causes the helium to condense out on the walls), or by opening a valve and pumping the helium out.

This technique will work to cool and trap pretty much any atom or molecule with the right magnetic properties, and it's been demonstrated using a variety of different atoms and small molecules. The starting temeprature is limited by the temperature of the helium buffer, but once the atoms or molecules are trapped, evaporative cooling techniques can be used to lower the temeprature even further, eventually even reaching the Bose-Einstein Condensate transition temperature with metastable helium (article 37 in the Doyle group wiki above).

Mechanical Methods: Another clever trick you can pull is to slow the atoms by moving the source. If you feed gas into a vacuum system, the pressure difference between the gas source and the vacuum chamber determines the speed of the atoms as they enter the chamber. But that speed is the speed relative to the source. If you make the source move rapidly in the direction opposite the flow of gas, you can get a source of atoms that are moving very slowly with respect to the vacuum chamber, and everything else in the lab.

This tactic obviously isn't sustainable in a linear configuration-- if your source keeps moving backwards, pretty soon you're going to run out of vacuum plumbing. There's no reason it can't work indefinitely in a rotating configuration, though, and a couple of different groups have made this work. The Nobel-winning chemist Dudley Herschbach made this work at Harvard, and continues to do slow-atoms studies in his new(ish) gig at Texas A&M. They feed gas into their system through a nozzle mounted on a rapidly rotating arm. To produce slow atoms, they spin the arm so that the gas exits in the direction opposite the motion. The velocity of the atoms relative to the rest of the apparatus is the difference between the speed of the rotor and the speed of the gas, and they can get this down to tens of meters per second, which would correspond to a temperature of 10K or so.

This work also gives rise to one of my very favorite bits of jargon, namely the "swatting limit," which is the minimum speed the gas needs to have to clear the path of the rotor before it comes around again and swats the atoms or molecules out of the beam path. OK, it's not a household term, but I love the image.

The mechanical rotor method is extremely versatile, and will work for absolutely any atom or molecule that you can feed through the rotor. Of course, getting high-pressure gas through a rapidly rotating arm while maintaining a reasonable vacuum in the chamber presents some formidable technical challenges. A variation of the trick that requires less elaborate plumbing is to keep the source of the gas fixed, and bounce the atoms off a rotating "paddle." This was done by Mark Raizen's group at Texas, who managed to cut the speed of a beam of helium almost in half, from 511 m/s to 265 m/s. This isn't quite as versatile-- it depends on the fact that helium reflects nicely off a silicon surface-- but it's not too shabby.

Of course, these methods are not for the faint of heart-- the Raizen group's apparatus uses a 50 cm arm spinning at 42 Hz (2520 rpm). I saw the apparatus when I toured their lab in Austin, and it's a hulking steel vacuum chamber. There was also a big Kevlar sheet that they set up outside the chamber when the rotor is on, because if it lets go, things could get nasty.

There are a few other tricks people have used to make atoms and molecules slow down, but this is getting a little long, so we'll split them off into a second post.