Point two laser beams so that they cross each other, and each goes through as if the other one did not exist. Light rays cannot interact with other light rays—or can they? With the help of a single atom, physicists have devised a system in which one light beam can turn another on or off. Such a light switch could serve as the basic component of futuristic optical quantum computers and may help open the way to a quantum version of the Internet, which would offer unbreakable data security.

The device makes use of a phenomenon called electromagnetically induced transparency, in which a laser beam can render opaque clouds of atoms temporarily transparent to a narrow wavelength of light. The cloud can then act as a switch for a second beam, either letting it through or blocking it. The result is similar to what happens with transistors in electronic circuits, where a voltage applied at one electrode controls whether current can flow between two other electrodes.

Applications such as quantum computing demand the control of beams down to single photons, the elementary particles of light. For that purpose, single atoms are better than clouds of them, says physicist Martin Mücke of the Max Planck Institute for Quantum Optics in Garching, Germany. He and his collaborators trapped a rubidium atom and aimed two different laser beams at it: one for probing, or transmitting, and the other one for switching. Ordinarily the atom acts as a barrier to photons from the probe beam because it would first absorb them—going from its “ground” state to an “excited” state—and then shoot them back, that is, reflect them. This condition would constitute the “off” state of the device.

But turning on the switching beam changed the atom’s possible states, so that it now had two different ground states. The probe beam then had two different ways of exciting the electron, each starting from a different ground state, but in the math­ematics describing the atom’s quantum-mechanical nature, the two possibilities cancel out, so that no excitation was possible. Thus, the probe beam photons, rather than being absorbed, could get through, marking the “on” state.

Making single photons interact can be useful because a photon can carry the units of quantum information, called qubits. They can exist in two states simultaneously and thereby represent both the 0 and 1 of binary code at the same time. Thanks to this feature, quantum computers could perform certain operations in parallel. In principle, they could quickly perform calculations that a typical computer could not do, at least not before the sun swells up and bakes the earth five billion years from now.

Max Planck’s Gerhard Rempe, the senior researcher on the team, points out that a single-atom device could do more than mere switching. For example, it could store photons and release them at will without damaging their delicate quantum states—an application known as quantum random-access memory, which could be crucial for data routers of a quantum Internet. In such a network, privacy is guaranteed by the law of quantum physics [see “Privacy and the Quantum Internet,” by Seth Lloyd; Scientific American, October 2009].

The new device still needs improvement: in the off position, the atom still lets through 80 percent of photons from the second beam. But the researchers say that straightforward improvements, such as keeping the atom colder, could bring that number down to 10 percent, if not to 0. (A more substantial limitation is that handling single atoms requires a fairly sophisticated physics laboratory.) The team published its results in the June 10 Nature. (Scientific American is part of Nature Publishing Group.)

Right now the device’s low efficiency limits its usefulness, comments Paul G. Kwiat, a quantum optics expert at the University of Illinois at Urbana-Champaign. But if the team can improve efficiency, he notes, it “could open a new, potentially efficient approach to quantum computing.”