An international group of researchers has developed a new way of controlling light using nanotechnology. The technique focuses on the boundary between two media, such as air and water, treating the boundary itself as a third medium. This allows the scientists to manipulate the reflected and refracted beams in ways that are not possible with natural materials, creating "designer light".

The scientists, based at Harvard University in the US, claim that their discovery has inspired them to derive a more general expression of Snell's law, which predicts the path determined by a beam of light travelling from one medium to another. This could help in designing new optical components such as planar lenses and polarizers.

At the boundary

Reflection and refraction occur whenever light crosses the boundary between two different media, at an angle. It is this incident angle and the optical properties of the two media that decide the angles of refraction and reflection, according to classical optics. But now, Nanfang Yu and colleagues from the Capasso research group have shown that if the boundary contains structures on the nanoscale, these laws need to be updated.

Standard reflection and refraction treats the boundary between media as a homogenous interface separating the two media. "What motivates us is the question: 'Why not treat the interface as a third 'active' medium?'" says Yu, who is also lead author of a paper on the research published in Science. "We realized that if we artificially structure the interface using nanotechnology, it can introduce an abrupt phase shift and a resultant time delay between the incident light beam and the reflected and refracted beams," he explains.

Yu says that this is the first time anyone has manipulated the boundary between media in the optical regime. "Interestingly, decades ago people working on microwaves and millimetre-waves demonstrated the so-called "reflectarrays" and "transmitarrays" that can shape the reflected and transmitted beams. The connection between that and our results is that both use abrupt phase changes associated with antenna resonances," says Yu. But that research was not at the nanoscale and the structures involved cannot be regarded as an interface or a boundary because the spacing between the array elements was larger than the wavelength.

The light fantastic

The Harvard team uses gold V-shaped plasmonic antennas – or pixels – patterned on silicon wafers as optical resonators. The array is structured on a scale much smaller than the wavelength of the incident light, allowing the engineered boundary between the air and the silicon to impart an abrupt phase shift or "phase discontinuity" to the light passing through. Yu points out that, while previous research concentrated on enhancing the near-field properties of optical antennas, his group uses "a somehow overlooked property of such structures – their phase response". The phase difference between the incident and scattered light varies considerably over one antenna resonance. By operating the antennas at different resonance conditions, a wide range of phase – and therefore time – delays are achieved. Effectively, each antenna captures the incident light, stores it for a given time and then reemits the light into the free space.

The researchers' interface is designed pixel by pixel as a series of optical resonators, such that the structure of the array determines the phase shift. By doing this, they can tailor the interface to reflect or refract in arbitrary directions, allowing a great degree of freedom in "shaping" the light. "For example, light coming in at an angle can be reflected back towards the light source – we call this phenomenon "negative" reflection because ordinarily the reflected beam is directed away from the light source," says Yu. There is also "negative" refraction, where the refracted light bends in the "wrong" direction as compared with the prediction of Snell's law. Yu says that there are two critical angles for total internal reflection, depending on the relative direction of the incident light and that of the gradient of the phase delay along the interface.

In one of the experiments they conducted, the scientists made the light ray hit the interface perpendicularly, from below, where the scattered light propagated at an angle, rather than perpendicular to the surface ( which is how it would naturally propagate), due to the varying structure of the antennae (see image "Perpendicularly incident light ray"). They also produced a vortex beam – a helical, corkscrew-shaped stream of light – from a flat surface (see image "Vortex beam and other strange optical effects").

Integrated optics

The researchers are now working on applications such as planar lenses that could focus an image without the necessity of a compound lens to correct aberrations. "The advantage of the plasmonic interface is that it moulds optical wavefront right after the light passes through it, unlike conventional optical components like bulk lenses, which rely on gradual phase accumulation along the optical path to change the wavefront of propagating light. This makes our design favourable for integrated optics," says Yu. He claims that some of their designs – such as the vortex beam – perform so well that they do not expect major difficulties in producing useful planar optical components for the long-wavelength (mid- and far-infrared) range. For the shorter wavelength range, however, they need to find a better non-metal resonator design.