Image design by Stephen Eisenmann (University of Illinois at Urbana-Champaign) and Tom Wilson.

In classical physics, whenever a wave encounters a change in density a part of that wave is reflected. In 1887, Lord Rayleigh took this concept further by studying what happens if the wave propagates not through a homogeneous medium but through one with a periodic structure. He showed that rays reflected from the multiple interfaces interfere with one another. For a band of wavelengths of similar value to the periodicity of the stack, the interference is destructive so that this bandgap prevents wave propagation through the structure.

By the 1980s, localization of light in artificial structures was a hot topic. Combining localization with the idea of the Rayleigh bandgap, Sajeev John considered, in 1987, how electromagnetic radiation could be trapped in a periodic three-dimensional dielectric material if randomness is introduced. As an illustration of this, consider altering the periodicity at just one point allows the existence of light at a wavelength within the bandgap; however, this light is trapped in the vicinity of the defect because it is forbidden everywhere else. Applied to chains of imperfections, light can then be guided with little loss. The potential of this approach cannot be understated: just as semiconductors have made possible the miniaturization of electrical devices, so photonic crystals hold the promise of microscale photonic circuitry.

Another landmark was set by Eli Yablonovitch with his paper published earlier in 1987. Following the work of Edward Purcell, scientists had started to think about controlling spontaneous light emission by modifying the photonic environment. This is exactly what a photonic crystal does. A quantum light source surrounded by a photonic bandgap is prevented from decaying because the photon that it needs to emit cannot exist. Conversely, the spontaneous-emission rate can be increased if the emitter is placed inside a defect with which it is in resonance.

The next challenge was fabrication. The first proposed design with a full bandgap comprised dielectric spheres in a configuration similar to atoms in a diamond crystal. However, the eventual structure, which was initially demonstrated in 1991 in the microwave regime, used an approach that was better suited to the material-processing abilities at the time by drilling holes in three different directions.

Of particular relevance to practical applications are two-dimensional photonic-crystal designs, which were first realized in 1996. They represent a compromise between a full bandgap and a simpler fabrication that makes possible the integration of both passive and active optical components on a single photonic chip. With the possibility of a revolution on a par with the development of the semiconductor chip, photonic-crystal research looks set to shine.