Quantum Switching of Light

ResearchBlogging.orgPhysics World posted a somewhat puzzling story a few days back, headlined Ultra cold atoms help share quantum information:

Scientists in the US have demonstrated a novel "light-switch" in an optical fibre that could become a new tool in the communications industry. The device created by Michal Bajcsy at Harvard University and colleagues could be developed to share both classical and quantum information.

Quantum information systems could bring a revolution to global data-sharing, by encrypting, processing, and transmitting information using the properties of quantum mechanics. However, as strings of "1s" and "0s" are represented by the quantum states of individual subatomic particles, such as the polarization of photons, they are very delicate and information can be easily lost. Prototype quantum devices have been developed but the move towards commercial applications requires more robust systems to compete with established "classical" technologies.

This is puzzling not because of the quantum stuff, but because it seems to have very little to do with the actual research being described, which was written up in Physics a few weeks ago (where, incidentally, you can get a free copy of the original paper). The work in question uses quantum mechanics, to be sure, but the business about quantum information isn't in the paper at all, and appears to be a garbled reference to something about three steps removed from what's actually being done.

The work described in the original paper is plenty cool. The paper comes from a collaboration between the research groups of Vladan Vuletic at MIT and Misha Lukin at Harvard, and it would be hard to find a more intimidatingly smart pair of PI's. They have worked out a way to trap a few thousand laser-cooled rubidium atoms inside a 3-cm piece of hollow-core optical fiber, which is impressive in its own right, but they then went on to use those cold atoms to demonstrate all-optical switching of the light passing through the fiber: they could determine whether a beam of light sent into the fiber was absorbed or transmitted by sending in a second beam of light, at a different frequency.

The basic process involved goes by the unassuming name "Electromagnetically Induced Transparency" ("EIT" for cognoscenti), and is the same phenomenon involved in "slow light" experiments. The process makes use of the fact that, to paraphrase a famous misquotation of Bill Phillips, "there are no two-level atoms, and rubidium is not one of them."

The idea is actually fairly simple, though the implementation details involve some funny quantum stuff. If you put a whole bunch of idealized atoms containing only two energy levels inside an optical fiber, and send light into the fiber whose frequency corresponds to the energy difference between those two levels, the fiber will be opaque-- the atoms inside the fiber will absorb the photons from the light source, and re-emit them in random directions, and the amount of light making it all the way through the fiber will be essentially zero. What EIT does is to use the presence of other states to, well, induce transparency.

If the atoms inside the fiber have three energy levels instead of two, with two of the three levels at very low energy relative to the third, you can get some light through the fiber-- the atoms will only absorb light at very specific frequencies corresponding to the energy differences between the upper and lower states. Atoms in one of the two lower states (state 1) are unaffected by light tuned to the separation between the other two states (states 2 and 3), and won't absorb it, and atoms in state 2 won't absorb light tuned to the separation between states 1 and 3. If you can arrange for all of the atoms to be in state 2, say, then the gas in the fiber becomes transparent, and light will pass through even at a frequency where you would expect it to be absorbed by atoms in state 1.

The real situation turns out to be a little more complicated than that-- for best results, you actually want to apply a small amount of light at the frequency corresponding to the energy difference between states 2 and 3 in order to maximize the transmission of light at the 1-3 frequency, but the basic idea is the same: by applying the right mix of light fields, you arrange for the atoms to be in a "dark state" that won't absorb any light. The authors demonstrate EIT in the cold-atom-filled fiber with very small numbers of photons-- roughly 10,000 photons are enough to almost completely turn off the absorption in the fiber, and let a pulse of light pass through.

You might think that this would suffice to make a switch-- turn the EIT laser on if you want light to pass through, turn it off if you want light absorbed-- but the authors go on to demonstrate switching by adding yet another laser field, connecting state 2 to another energy level, state 4 (rubidium atoms have lots of levels...). By applying a pulse of light at the 2-4 frequency, they can turn off the EIT effect, getting a 50% reduction in transmission with only 700 photons.

It's not entirely clear why you would opt for this, instead of just turning off the EIT laser, but I suspect the answer has to do with the "slow light" effect related to EIT. When the 1-3 light is allowed to pass due to the EIT effect, it does so at a reduced velocity-- in this work, about 3km/s instead of the usual 300,000 km/s. This delay may be useful, and the three-laser switching scheme would allow you to switch light on and off while also controlling the delay. I'm just speculating, though-- readers with more direct knowledge (Hi, Dave) are encouraged to correct me.

I suspect the "slow light" connection is also the origin of the weird stuff about quantum information in the Physics World article, as these effects are often cited as a technology that might be connected to quantum information. I suppose you could also imagine using this sort of thing to entangle pulses of light with each other, since small numbers of photons are enough to control the transmission of pulses (though I would think that the absorption process would kill most of the quantum character). And both the Physics paper and the actual Phys. Rev. Letter mention this as a tool for creating effective interactions between photons, which would open up all sorts of weird quantum phenomena.

Just what the opening of that Physics World piece is talking about, though, I have no idea.

Dawes, A. (2009). Optical switching with cold atoms Physics, 2 DOI: 10.1103/Physics.2.41

Bajcsy, M., Hofferberth, S., Balic, V., Peyronel, T., Hafezi, M., Zibrov, A., Vuletic, V., & Lukin, M. (2009). Efficient All-Optical Switching Using Slow Light within a Hollow Fiber Physical Review Letters, 102 (20) DOI: 10.1103/PhysRevLett.102.203902

More like this

Well, the Physics World thing is just an example of a common technique for reporting on quantum physics experiments when the journalist is not quite sure why the experiment is interesting, i.e. vague oversimplified description of the experiment followed by general unrelated blurb about why quantum things are relevant to technology. This usually results from garbling the result of an interview with the scientists.

I'm not an expert on the experimental side of things, but I imagine that this and/or related technologies could indeed have applications in quantum information. For one thing, we don't necessarily need interactions or entanglement because there are things like quantum cryptography that do not need to make use of them. For example, if a variation on this theme could be used to create a very accurately timed single-photon pulse then that would be very useful for quantum communication applications.

I do recall seeing several papers on using EIT in quantum information on the arXiv a while back, but I did not read them because I did not know too much about EIT or its experimental status at the time. It might be worth taking another look now.

Forgive the silly question from a history major, but how do you cool an atom with a laser? Shouldn't the laser be heating particles?

Because I'm a lazy blogger, I will just refer you to two old posts on laser cooling. The short version is that you can use lasers to exert forces on atoms, and if you arrange things just right, you can use those forces to slow atoms down. Slow equals cold, so this lets you use lasers to make things cold.

The first idea for laser cooling exploits the Doppler effect. If you are moving towards the source of the light (or sound...) you see/hear the frequency raised, and if you are moving away you see/hear the frequency lowered, compared to if you were at rest. (Think of the change in the pitch of a police siren as the car passes by: first it was moving towards you, then away from you).

Atoms only like to absorb light at certain frequencies (These give the spectral lines that can be used to identify which type of atom it is). If we raise or lower the frequency away from one of these preferred values the absorption decreases.

If we have two laser beams shining on the atoms from opposite directions, and tune them to slightly below an absorption frequency, then we will get Doppler cooling. If atoms are moving towards one of the lasers then they will see the frequency raised (closer to the preferred value) and at the same time see the frequency of the "chasing" laser lowered. So, they will preferentially absorb the on-coming photons, which reduces their forward momentum, slows them down, and cools them. (Only along one axis - need to add more lasers to cover full 3D situation).

Turns out other mechanisms are also at work (such as Sisyphus cooling) which greatly increase the effect beyond pure Doppler.