Share This article

A team of researchers from NASA’s Jet Propulsion Laboratory and CalTech have developed a range of ultra-thin, optical components capable of arbitrary manipulation of light. The devices, dubbed metasurfaces, are able to locally modify the properties of a light-field in ways difficult to achieve with standard optics.

So what does it mean to manipulate light? Light, an electromagnetic field that propagates through space, can be completely described at one wavelength by polarization, phase, and amplitude. If we know what the light-field looks like now, we can accurately predict what it will look like in the future by knowing only these properties.

Any optical component that exists can be thought of as essentially modifying one or more of the above properties. For example, in free-space optical systems, polarization is modified using wave retarders, polarizers, and polarization beam-splitters; phase is shaped using lenses, curved mirrors, or spatial phase modulators; and amplitude is controlled via neutral-density absorptive or reflective filters. Therefore, by combining many components, we’re able to build systems that can manipulate the light-field to varying degrees.

However, as you may have guessed, in order for us to have full control, we normally need many components, each of which is usually bulky and expensive. Think of the optics in a telescope or DSLR for example.

What is a metasurface?

Metasurfaces are planar (~2D) structures that locally modify the polarization, phase, and amplitude of light in reflection or transmission, where each sub-pixel is smaller than the wavelength of light. When we say 2D, the vertical dimension is normally <100nm, or ~1,000x smaller than a human hair. Therefore, these flat, highly functional optical components can be manufactured in exactly the same way as state-of-the-art electronics, such as microchips, which use high resolution lithographic techniques.

, A device that separates x- and y-polarized light and focuses them to two
different points. The two different points can be chosen at will

As described by ExtremeTech previously, the holy grail in re-configurable optics is to achieve ultra-thin, tunable, complete phase control with a device.

What have they done?

The team of researchers have developed a new kind of metasurface composed of a single-layer array of amorphous silicon (silicon nanopillars), patterned into differently sized elliptical posts — all sitting upon was is essentially a glass surface. Seen under a scanning electron microscope, the metasurface appears as a cut forest where only the stumps remain.

Each silicon stump, or pillar, has an elliptical cross section, and hence has a different effective refractive index associated with the two different modes that can be excited across the structure. By carefully varying the diameters of each pillar and rotating them around their axes, the scientists were able to simultaneously manipulate the phase and polarization of passing light.

Using the elliptical nanopillar as a sub-wavelength pixel, the researchers produced a range of optical devices, from polarization beam splitters and lenses to phase holograms, all operating at a near-infrared wavelength of 915nm. (Visible light is 400-700nm.)

A device that separates x- and y-polarized light and focuses them to two
different points. The two different points can be chosen at will

Should I care?

To be sure, this is an incremental advance. There is a plethora of research worldwide on metasurfaces, and pretty much every single week there is a new device, which will supposedly “revolutionize” the field of optics and be applicable to every field one might think of (which is a nifty selling tactic of getting your work published). Yet, in reality, the research is a small incremental adjustment based on a nanostructured surface, which manipulates a light field to some degree. For example, the work uses infrared light, because doing this at shorter, visible light runs into all sorts of fabrication problems. Also, nanorods and other geometries have been used previously to do pretty much the same thing.

The work in and of itself is not bad, and is a nice window into the world of metasurfaces. However, until someone cracks full, re-configurable phase control at visible wavelengths, the impact of such work will be short-lived.