The device, dubbed a polarization synchrotron by its inventors, consists of a 2 metre-long gently curving arc of alumina (a dielectric material), with a series of electrodes fitted at regular intervals along its length. Applying a sinusoidal voltage across each electrode and displacing the phase of the voltage very slightly from one electrode to the next generates a sinusoidally-varying polarization pattern that propagates along the device. By carefully adjusting the frequency of the voltage and the phase displacement the researchers say they can make the wave travel at greater than the speed of light (even though no physical quantity of charge travels superluminally).

This principle is based on a model of pulsars -- rapidly spinning neutron stars -- developed by one of the group, Houshang Ardavan of Cambridge University. Ardavan believes that the well-defined pulses of radio waves emitted by these astronomical objects are caused by the pulsar's rotating magnetic field polarizing the surrounding plasma. As the magnetic field sweeps round so too does the region of polarized plasma, and far enough away from the pulsar this region will sweep round at faster than the speed of light.

Singleton's group -- which includes Ardavan's son, Arzhang Ardavan (of Oxford University) -- believes that its polarization synchrotron, like a pulsar, emits radiation in a well-defined beam. They argue that the electromagnetic wavefronts generated by each point within the polarization pattern build up behind that point like sound waves from a supersonic aircraft. Interference between these wavefronts then reinforces the radiation along a spiral trajectory -- the beam -- that travels away from the source.

The researchers claim that the intensity of this beam is proportional to 1/r, where r is the distance from the transmitter, rather than the 1/r2 associated with spherically decaying radiation. They carried out tests on their device at the Turweston Aerodrome in Northamptonshire between May 2003 and February 2004, measuring the intensity of the emitted radiation at a range of distances up to 900 metres and mapping the three-dimensional shape of the emission.

According to Singleton, the polarization synchrotron could transmit radio messages with very little power or over vast distances. A scaled-down version of the device could be used in mobile phones to allow direct communication with satellites, rather than having to rely on relay stations. He says the device could also be used in radar systems, since the beam's unusual shape would make it difficult to trace the beam back to its source.

However, other researchers are sceptical. John Hannay, a theoretical physicist at Bristol University points out that conventional radio sources can generate slowly decaying radiation over limited distances. He has previously said that Singleton and co-workers must test their device over tens of kilometres rather than hundreds of metres.