By Jon Cartwright

A single photon can now be split into three, thanks to the work of an international team of physicists. The achievement could open up new avenues in the field of quantum information.

The ability to split photons may not sound as extraordinary as other achievements in quantum physics, but for decades it has proved crucial to the success of many experiments. Often researchers need to know that photons are emitted at precisely the same time and are in phase with each other, and this is almost impossible if the photons come from separate sources.

In the past, devices have been able to split a photon only into two. In the typical method used to achieve this, known as parametric down-conversion, a laser beam is shone into a special "non-linear" crystal--crystals that exhibit unusual optical effects under intense laser light. Occasionally, a single photon from the beam converts into two photons, each with a portion of the original's energy and momentum.

Researchers have known that, in theory, it would be possible to split one of these new photons again in a "cascaded down-conversion," making a total of three photons. But there has been a catch: the probability of one photon splitting is normally just one in a billion, making the probability of it happening twice in succession one in a billion billion. Experimentally, this has been too small to contemplate.

Technology has improved over time, however, and now Thomas Jennewein of the University of Waterloo in Ontario, Canada, and his colleagues have made three-photon generation a reality. The results were published July 29 in Nature.

The team used periodically poled potassium titanyl phosphate (PPKTP) and periodically poled lithium niobate (PPLN)--non-linear crystals that are more efficient than those generally used in previous experiments--and inserted a waveguide into one. The waveguide helps to confine the optics along the crystal's length to increase efficiency further, so that the probability of a photon splitting into three is one thousand times greater--a more experimentally friendly one in a million billion.

Hero experiment

In tests, the group recorded the average rate of three-photon detection as about 4.7 per hour--well above the background rate of 0.5 photon triplets per hour.

Yaron Silberberg, an optics researcher at the Weizmann Institute of Science in Rehovot, Israel, praises the group's work as a "hero experiment--very impressive in achievement, although straightforward in concept." However, he calls the rate of triplet production a "crazily low" figure. "Any experiments with these would take many days," he says. "Demonstrating triple entanglement would be one."

Indeed, says Jennewein, demonstrating entanglement--a phenomenon in which quantum entities become inextricably linked to one another over any distance--is next on the group's agenda. Doing this will demand that the researchers measure the three photons' energy and momentum, and verify that these sum to the energy and momentum of the original photon, to prove there is none left over in the photon source as a "memory" of the triplet production.

If the researchers do manage to prove three-photon entanglement, which Jennewein expects to take a year or more, it could allow for more complex experiments in quantum information, such as quantum communication between three parties.

There is also the added advantage that two of the three photons generated by this method come at wavelengths of close to 1,500 nanometers. This is the standard in telecoms, and means that it should be easy to integrate future quantum systems with existing telecoms systems.

"As always, only time will tell where this particular design will eventually lead," says quantum physicist Aephraim Steinberg of the University of Toronto, Canada. "But it already demonstrates the possibility of something that seemed out of reach not that long ago, and shows that the classical theory behind down-conversion seems to remain valid all the way down to the quantum level, 10 orders of magnitude lower in intensity than where it was initially applied."