It is almost miraculous that we can see at all. To reach you from across a room, the light bouncing off a friend's face must run a veritable gauntlet -- in the 10 nanoseconds it takes to race into your own eyes, each photon this light is composed of travels through a space jam-packed with other photons: there are literally millions of millions of other photons racing around the room at the same time. Yet instead of bouncing off of each other, these beams of light travel in straight lines, ignoring each other entirely. Even light travelling to telescopes from the ends of the universe, which has been on its way for billions of years, keeps going in a straight line, unperturbed by all the other photons from the billions of galaxies, quasars, and supernovae it has met along the way.
Unfortunately for those of us who are interested in using light as the medium for a suite of next-generation technologies (such as "quantum computers" which might be far more powerful in certain respects than any conceivable computer based on old-fashioned "classical" physics) this has a downside. The basic building block of any computer is some sort of "logic gate," or a "switch" such as a transistor: some element that lets two different bits of information affect one another. But how to build a computer in which information is stored in light, if light can travel for billions of years without paying any notice at all to the other photons it encounters?
The answer is that the right kinds of materials -- even single atoms, if the light is "tuned" just right to the color those atoms can absorb and emit -- can "trick" photons into having an effect on one another. (At the simplest level, you can imagine that if you shine a strong enough laser on some water, the water will heat up and start to boil, and the steam this releases will modify the path of another light beam moving nearby: in this way, the first beam has "affected" the propagation of the second one, thanks to the help of the water.) Still, these effects are weak, and this is why the study of so-called "nonlinear optics" really only took off with the invention of the laser in 1960. This made it possible to focus billions of photons into the same small region of space at the same time, generating rich effects that are important in a wide range of technologies, ranging from optical telecommunications to biological imaging to green laser-pointers.
To implement the sorts of "quantum technologies" that have justifiably obsessed much of the physics community for the past 20 years, however, information must be stored in individual photons, not beams made up of billions of particles. This has led to a worldwide race to see if there are ways to catalyze "giant nonlinearities," effects in which even a single particle of light can have a measurable effect on a second light beam. In the present work, our group at the University of Toronto and the Canadian Institute for Advanced Research used laser-cooled Rubidium atoms to achieve precisely that. We found that a single photon could "rotate" a "probe beam" by an angle of about one one-thousandth of a degree. While still very small, this was large enough that (after averaging several days' worth of data) it was clearly observable. Future work will aim at generating even larger effects, ones which could be observed on a single shot, and perhaps be used to build a "quantum optical logic gate" for next-generation computers. In any case, this and other related work opens the door to a new chapter in optics, one in which all the richness of nonlinear optics coexists with the fascinating new effects that can be observed when working with individual quantum particles.
The technical article, "Observation of the nonlinear phase shift due to single post-selected photons" by Amir Feizpour, Matin Hallaji, Greg Dmochoswki and Aephraim Steinberg, was published online in Nature Physics on 24 August 2015. An openly accessible early draft is available on the arXiv.|