How do you catch atoms? In very tiny traps

Trapping atoms is near and dear to my heart, regardless of whether you do it for sport or food or to knit the tiny pelts into a shawl. So it’s disappointing to find conceptual holes in an article on the subject.

Traps for neutral atoms come in three different flavors. One is called an optical molasses trap, where the Doppler shift is used to turn the atom’s translational motion into light, which prevents atoms from flying away from the trap.

In short, no. Optical molasses uses radiation pressure to slow atoms down — atoms scatter photons and slow down, much like throwing pong-pong balls at a a bowling ball will slow it down. “turn the atom’s translational motion into light” is somewhere between awkward wording and flat-out wrong. The rest is OK up until this part

If I use three pairs of laser beams all facing a cloud of atoms, then no matter what direction the atoms are moving, they always get driven back to the place where the laser beams meet.

For optical molasses, this isn’t true — it’s a dissipative or damping force, which is very strong (hence the use of molasses — near resonance the accelerations can be hundreds of g’s), but there is nothing that pushes the atoms to a particular place. Atoms will random walk— very slowly — until they leak out, like a drunken frat boy in a field, wearing heavy boots. Now, you can get the atoms to tend to collect at a point with the addition of a quadrupole magnetic field and the appropriate polarization of your light, because there is a restoring force that depends on the field strength. This is now a drunken frat boy with heavy boots and the center of the field is a low spot. That’s called a Magneto-Optic Trap, or MOT. (Which, of course is “TOM” backwards, so it’s ridiculously easy to remember). In very rough numbers, a MOT will confine an order of magnitude or so more atoms than a molasses, at least in my experience with a few species of alkali atoms. The atoms in a MOT want to be both slow and at the zero point of the field, whereas in a molasses they just want to be slow. But there’s a cost — the atoms in a MOT tend to be hotter, so if you are going for cold atoms what you can do is trap your atoms in a MOT and the turn the magnetic field off. The cloud expands since there’s no longer a restoring force but it does so slowly because you still have a molasses. I did a video of this a few summers back, though the system is not optimized to give the nice expansion into molasses (the price of getting to play with it)

The big drawback of this trap is that it puts atoms into an excited state. From there they have many options, only one of which involves emitting the absorbed photon—atoms that don’t choose this path escape the trap. Furthermore, the excited state destroys a BEC (remember, all the atoms need to be in the same state to form a BEC), so the trap must be switched off at some point in the preparation process and cannot be used to manipulate the BEC.

Atoms in a molasses are neither dense nor cool enough to form a BEC, so this is a bit of a non-sequitur. There’s no BEC to manipulate. Typically the next step is to load the atoms into a magnetic trap, which the article explains next, and do evaporative cooling — you change the shape of the trap to let the most energetic ones out, which lowers the temperature, similar to evaporation.

Next up is the dipole force trap.

Take an atom sitting just to the right of the center of the laser beam. As photons pass through the electron cloud of this atom they are deflected—those on the left-hand side are deflected to the right, while those on the right are deflected to the left.

Another instance of ummmm, no. The author is likely confusing the explanation of forces on macroscopic spheres (as in optical tweezers) for the atomic response. This indeed explains what happens to a polystyrene sphere, but an atom? The light used is of order a micron in wavelength, which is orders of magnitude larger than the atom itself. So there is no way for a photon to go through one side or the other. What’s going on here is that the intense light has an electric field, and because the light has an intensity gradient, there is a gradient of the electric field. This induces an electric dipole in the atom, which gives rise to a force in the direction of the gradient — the atom is pushed toward the region of highest intensity. There’s no dissipation here, though, so no cooling is going on. Once again, you typically cool the atoms first and then load them into the trap.

Plasmonics, the last section, is not something I’m “up” on, so I can’t really critique anything, but given the rest of the article, there’s a decent chance of a large conceptual gaffe in it.