Trapping Ions With Light

There was a flurry of stories last week about an arxiv preprint on optical trapping of an ion. Somewhat surprisingly for an arxiv-only paper, it got a write-up in Physics World. While I generally like Physics World, I have to take issue with their description of why this is interesting:

In the past, the trapping of atomic particles has followed a basic rule: use radio-frequency (RF) electromagnetic fields for ions, and optical lasers for neutral particles, such as atoms. This is because RF fields can only exert electric forces on charges; try to use them on neutral particles and there's little effect. A laser, on the other hand, can draw the dipole moments of neutral particles towards the centre of its beam. But the resultant optical trap is relatively weak, and so ions - which are sensitive to stray electric fields - easily escape.

This is not entirely true. RF fields (if you stretch the definition of RF a little to include microwaves) can and have been used to trap neutral atoms: see "Demonstration of neutral atom trapping with microwaves", which came out not too long after I started grad school (the work was done before I got there, the paper after). It's not a tremendously effective trap, but it does work.

The reason RF fields are used to trap ions is that they're almost but not quite DC fields.

The difference between ions and atoms is that ions are charged. Because they have electric charge, they can be pushed around by electrostatic interactions. Neutral atoms, on the other hand, interact with charged objects only through their induced dipole moments (neutral atoms have no permanent dipole moments, unless there's a really tiny one from the electron EDM), and those forces are many times weaker, and drop off much more rapidly with distance. If you want to use electrostatic forces to manipulate a neutral atom, you need to use an enormous potential difference, or put your electrodes right on top of the atoms.

Ions, though, are quite readily pushed around with relatively small electric fields, on account of their electric charge. Which means that if you want to trap an ion, electrostatic forces are the way to go.

The problem is, there's a formal result about electrostatics called Earnshaw's Theorem that shows it's impossible to trap a charged particle using only static potentials. You can try to do the obvious thing, and surround the ion with electrodes having the same sign of charge, so as to always push it back to the center, but when you do that, it sort of squirts out to the side. No matter what you do, there's always a path that the ion can sneak out along.

What you can do, though, is to switch back and forth between two different trap configurations. You put voltages on your electrode that will trap it effectively in the vertical direction, say, but allow it to slip out horizontally. Just when it starts to head out the side, though, you switch the voltages to a configuration that provides strong left-right trapping, but lets the ion escape up or down. When it starts to move out in a vertical direction, you switch back to the first configuration, and so on.

(The third dimension, in and out of the page, you can close off with static fields. You can't make a stable three-dimensional trap with static fields, but you can do two of the three at any given time.)

If you switch the voltages fast enough, the forces the ion feels will sort of average out into a stable trapping configuration. The forces are basically electrostatic forces, but they change direction quickly enough that the ions can't use Earnshaw's theorem to escape-- before they can make any progress in the untrapped direction, the force switches to stop them from going that way.

How fast is fast enough? Well, you want to make sure that you're shifting the voltages faster than the ions can respond, but not so fast that you get close to any of the resonant frequencies in the ion, at which point you have to start treating the oscillating field like light, rather than like a dc field. Ions move on a time scale of a millisecond or so, so you want to switch faster than kilohertz, and most of the relevant resonances are in the several gigahertz ranges, so tens of megahertz are pretty much the sweet spot for ion trap operation. Hence, RF fields.

So, why hasn't anybody trapped ions with light before? Because there hasn't been any reason to. For the majority of the things people have wanted to do with ions, RF traps are just fine, and provide strong trapping with relatively little effort. There's no reason optical forces wouldn't work, it's just that there's no huge upside to using optical traps rather than RF traps.

My reaction on seeing the news stories about this (this is a rare case where I prefer axivblog's take to Physics World's) wasn't "Wow! This opens up exciting new opportunities!", it was "Huh. They haven't done that before now?" It's a nice result, and will allow studies of things like collisions between ions and neutral atoms in the same trap, but I just don't see this as a dramatic, game-changing result.