A joint research project between Fujitsu and the University of Tokyo may have discovered a way to provide complete data security between two networks.

The two have been working on a viable quantum cryptography system that would allow two parties to share encryption keys via telecommunication networks with full confidence that they have not been compromised en route.

The team has succeeded in generating and detecting a single photon at wavelengths useful for telecommunications, said Yasuhiko Arakawa, director of the Nanoelectronics Collaborative Research Center at the University of Tokyo and leader of the research project.

The reliable generation and detection of single photons is vital if quantum cryptography systems are to leave the laboratory and enter practical use, and the team has managed this through the development of a new photon generator.

If two parties want to exchange encrypted data they need to share the electronic key that will be used to encode it. The data is encoded with a corresponding private key, so using the genuine public key is vital. Should a fake key be substituted for the real one, the data could be read by a third party. Since sharing keys across networks can expose them to tampering, many users exchange keys offline via physical media, such as a floppy disk or CD-ROM.

Quantum cryptography systems, however, will make those keys tamper-proof as each data bit of the key is encoded onto individual photons of light. A photon cannot be split so it can only end up in one place: with the intended receiver or with an eavesdropper. Should a key be completely received the recipient can be sure it hasn't been compromised and should it be incorrectly received there's a chance that it has been intercepted and so a new key can be issued.

The only difficulty with this theory is that it has to be possible to reliably generate a single photon. If two or more photons are generated, the key's security is gone. Until now most experiments involving quantum cryptography have used lasers as their photon source but they haven't proven to be completely reliable generators.

"By reducing the output power of the laser we can create one photon sometimes, however it is impossible to control accurately the number of photons," Arakawa explained. Reducing the laser power also means the overall transmission speed is slowed.

So Arakawa's team has developed a new generator based on materials developed by Fujitsu and Japan's National Institute for Materials Science. The material is embedded with quantum dots, which are like tiny holes into which individual electrons can enter and a photon be produced.

"They are almost comparable to the wavelength of the electron so electron motion is almost zero and the electron cannot move," Arakawa said. "The energy state is fixed. So if we can control the energy of the electron, we can control the number of photons that are emitted."

The wavelength of the photons that are emitted can be controlled by adjusting the size and shape of the quantum dots. Doing so very accurately is difficult, so additional filtering is employed to ensure that only those with a wavelength suitable for transmission down commercial optical fiber networks are let through, said Tatsuya Usuki, a researcher at Fujitsu Laboratories, who also worked on the technology.

Because the accurate generation of single photons is possible and there is no need to throttle back the power, the transmission speed can be increased from a few hundred bits per second to around 400 times that speed, Arakawa said. He estimated a commercial system might be possible to transmit data at up to 100Kbit/s.

The group has also made progress on the detection end of the system. Light coming out of the fiber is split into two and sent to two detectors. By measuring the time at which photons arrive researchers can determine whether one or two photons were generated. In the case photons arrive at the same time at each detector, it means two were generated which was not the case with the new system, Arakawa said.

At present the team has succeeded in generating photons at both 1.3 micron and 1.55 micron wavelengths and verified single photon transmission at the former wavelength. Verification of the latter is one of its next goals. The project hopes to develop a practical single photon generator by 2007 and Arakawa predicts commercial systems based on the technology could be available in five years.