The mysterious attraction between two neutral, conducting surfaces in a vacuum was first described in 1948 by Henrik Casimir and cannot be explained by classical physics. Instead it is a purely quantum effect involving the zero-point oscillations of the electromagnetic field surrounding the surfaces. These fluctuations exert a "radiation pressure" on the surfaces and the overall force is weaker in the gap between the surfaces than elsewhere, drawing the surfaces together. Tiny though it is, the Casimir effect becomes significant at distances of microns or less and actually causes parts in nano- and micro-electromechanical systems (NEMS and MEMS) to stick together.

Now, Leonhardt and Philbin have calculated that the Casimir force between two conducting plates can turn from being attractive to repulsive if a "perfect" lens is sandwiched between them. A perfect lens can focus an image with a resolution that is not restricted by the wavelength of light. Such a lens could be made from a metamaterial made of artificial structures that are engineered to have negative index of refraction -- which means that the metamaterial bends light in the opposite direction to an ordinary material.

According to the researchers, the negative-index metamaterial is able to modify the zero-point oscillations in the gap between the surfaces, reversing the direction of the Casimir force. Indeed, the researchers believe that this repulsive force is strong enough to levitate an aluminium mirror that is 500 nm thick, causing it to hover above a perfect lens placed over a conducting plate (see figure).

Since the Casimir force acts on the length scale of nanomachines, manipulating it could be important for future applications of nanotechnology. "In the nano-world, the Casimir force is the ultimate cause of friction," said Leonhardt. "Our result means we could now envision frictionless machines or novel micromotors."

While physicists have had some success creating perfect lenses from negative-index metamaterials, the technology is still in its infancy. "The work points towards new applications of left-handed materials that are not strictly optical," says Federico Capasso of Harvard University, who studies the effect of the Casimir force on MEMS. "However, the materials are not easy to make so the concept may take a few years to realise."

Capasso adds that there also other methods to achieve a repulsive Casimir force, without the need to use artificial materials. "For example between two plates made of different materials, separated by a fluid that has a significantly different dielectric function compared to that of the plates themselves."

The researchers will report their work in New Journal of Physics.