AMO Physics meet chips.

I recently had the pleasure of attending a small workshop on the topic of doing atomic physics on chip-scale apparatus. The presentations and discussions were very interesting, but unfortunately do not lend themselves to a blog report for a couple of reasons. This kind of get-together discusses ongoing projects, some or all of which are not-ready-for-primetime, i.e. things haven’t been written up and published, so I wouldn’t be at liberty to discuss details, and a lot of the really cool stuff was shown in the pictures, some of which also haven’t been published. So while I can’t go into certain details, I can give an overview, and provide some links to representative work that has been published or is otherwise available to the public.

There’s quite a bit of physics that is described as being “tabletop” physics — in part to distinguish it from the physics that requires large collaborations and trips to an accelerator lab, or access to a big telescope or a satellite, etc. However, “tabletop” in atomic physics usually refers to an optical table, which can be a behemoth — an 8′ x 4′ table can weigh 1000 lbs or more. Add to that the requisite optics, optomechanics, vacuum systems and the support electronics, and you still have quite a lot of equipment, and it’s not going to be very portable.

As I had mentioned before, research layouts tend to be bigger than they strictly need to be, in order to provide some flexibility, and you can transition to more compact layouts if you have just one application in mind. If you take the concept a little further, and start thinking of the possibility that you might want an AMO experiment (e.g. a MOT or BEC, or trapped ions) to be applied to some specific problem, and require portability, you have to shrink not only the laser system, but the trap, and everything mentioned above — the vacuum pumps, power supplies and control electronics. Fortunately, a lot of this is coupled: if the trapping volume is decreased you don’t need as much pumping capacity or laser power, so your power demand is also smaller. If you can survive the decrease in signal, you can still go pretty good measurements, though new challenges tend to crop up working close to surfaces rather than far away. Instead of using general-purpose power supplies and electronics, you use devices specifically built for the application. You mount components together, rather than spaced apart, and use fibers rather than mirrors to transport light.

In the last several years there has been a push for “chip-scale” atomic physics experiments that do these very things. Chip means literally a computer chip-sized lithographed circuit, perhaps several cm on a side, or smaller, that will have the electronics for generating the electric and/or magnetic fields for trapping atoms or ions, and acting as guides for moving the particles around. The lithography usually isn’t anything like the nanoscale work you’d find on an integrated circuit, but there are micro-machined instruments, called Micro-Electro-Mechanical Systems (MEMS), that fall into this category, that have components that are quite small.

Here is a basic example of a chip that might be used (without the fly, which is there for scale. From Ted Haensch’s group), or this one which is presumably of similar size. (I’m pretty sure the text is supposed to say “magnetic moment” rather than “magnetic momentum”)

Once you have the chip, you can start coupling it with other components in various ways. One example is NIST’s chip-scale atomic clock, and another more polished-looking one being developed by Symmetricom. And if you take a atomic clock-like device and make a few tweaks to become sensitive to magnetic fields, you can make a magnetometer.

You can make BECs on chips, and manipulate the atoms to give you interference, which makes it possible to sense rotations and accelerations. Various manipulations can be done to split a BEC and allow the two ensembles to sample different potentials, and cause interference. One thing I saw presented was doing is microgravity experiments using the BEC by dropping the whole apparatus in a tower.

Another set of experiments involve ion traps. An advantage of ions is that you have really strong coupling, so you can manipulate the atoms a little bit more easily than neutral atoms, and so ion work is a little ahead in quantum computing experiments, but there are other applications as well.

Two of the funding sources for this type of work are the Office of Naval Research and DARPA’s Defense Science Office, so one can see the drive here is applications that might be useful to an individual in the military, and a device that can be transported easily, even to the point an individual could carry it around. Small size, small power consumption. And DSO has several ideas in development, because DARPA doesn’t like to be surprised. (and a few animations can be found here within the guided BEC interferometry project; one should note that they are targeted for an audience a few notches below the scientists doing the work) But as with all emerging technology, you don’t know what else might be out there, lurking as a future application, and so the push to miniaturize these aspects of AMO experiments may bear other fruit.