Two independent groups of physicists claim to have demonstrated invisibility cloaks that operate for light at optical wavelengths.

Until now researchers had only been able to create invisibility cloaks for the microwave part of the spectrum. But last week Michal Lipson and colleagues at Cornell University uploaded a preprint on arXiv in which they describe the first demonstration of a cloak that can disguise objects from light in the near infrared to the far red. The following day Xiang Zhang and colleagues at the University of California at Berkeley uploaded a preprint in which they describe a cloak for just the near infrared.

Although Lipson’s group has submitted its preprint to Nature Photonics and is awaiting peer review, the work of Zhang’s group will soon be published in Nature Materials.

Rapid progress

The first invisibility cloak was demonstrated by a group at Duke University in the US in 2006. It was based on a metameterial — a manmade material with exotic electromagnetic properties — that could cause light to bend around an object in 2D like water around a stone.

But the cloak only operated over a narrow range of microwave wavelengths, and it was only at the beginning of this year that the same group managed to overcome this drawback with a broadband version.

Based on a design by John Pendry at Imperial College, London, the so–called carpet cloak used a different principle whereby a special conducting sheet appears to flatten the bulge of an object hidden underneath.It worked by reflecting light in such a way that it cancels the distortion produced by the curved surface, thus giving the illusion of a plane mirror.

The Cornell and Berkeley groups both claim to have improved on this type of cloak by fabricating metamaterials with nano-scale — as opposed to millimetre or centimetre scale — features, so that it operates over shorter wavelengths.

In the Cornell group’s device the metamaterial is an array of 50 nm–diameter silicon posts on a silicon–dioxide substrate, and the mirror is a textured pattern known as a distributed Bragg reflector (DBR). In the Berkeley group's device, however, the metamaterial is an array of 110 nm–diameter holes in silicon dioxide, and the mirror is made from gold.

Devices are ‘very similar’

The cloaking properties of both devices substantially overlap. The California group claims their device cloaks between wavelengths of 1400 nm and 1800 nm, which is all in the near infrared. Meanwhile, the Cornell group claims their device cloaks at even shorter optical wavelengths, from 1975 nm, which is in the near infrared, to 1025 nm, which borders on the red.

The two groups could not comment to physicsworld.com about their own work because of Nature’s embargo policy. However, Lipson of Cornell said both groups’ devices are “very similar”, except that the DBR employed by her group should in principle reflect 100% of incident light.

Tomas Tyc, a theoretical physicist at Masaryk University in the Czech Republic, thinks the devices make a step towards true invisibility. “There is no doubt that the experiments are important and present very good achievements, but still are far from real full cloaking where a three–dimensional object looks like a zero-dimensional point, so it is invisible from any direction of view,” he said.

Another disadvantage with the carpet-cloak scheme is that the surface itself — that is, the “carpet” — remains visible. Early this year Tyc, together with Ulf Leonhardt of St Andrews University in the UK, put forward a cloaking scheme exploiting “curved space” geometries that is broadband and leaves no trace of a surface, but this has not yet been implemented.