Breakthrough optics pave way for new technologies

A new approach to designing optical technologies — based on a single device known as a Mach-Zehnder interferometer — could pave the way for a new class of technologies that could give optics the kind of versatility we see in electronics.

Optics in computing, telecommunications links and switches, and virtually any other optical component, could be created simply by configuring a mesh of light-controlling devices, ie, interferometers.

Optical technologies have the potential to greatly reduce the power consumption of computers, speed telecommunications and enhance the sensitivity of chemical and biological sensors. The basic building blocks of traditional optics, however, mirrors and lenses, lack the versatility to readily perform these functions and are difficult to scale to the small sizes needed for many applications. The new approach could overcome these limitations.

“Recently, optical researchers have begun to understand that these interferometers can be thought of as universal ‘building blocks’ that could enable us to construct essentially any optical device we could imagine,” said Dr David AB Miller, Stanford University, California, USA, and author of a letter describing the potential of interferometers published in The Optical Society’s new high-impact journal Optica.

Previously, this approach would have only been feasible if the Mach-Zehnder interferometers were able to achieve perfect performance — a seemingly unattainable goal. The new approach described in this paper, however, presents an alternate pathway.

Rather than engineering a perfect, single component, researchers propose it’s possible to create a mesh, or array, of interferometers that, when properly programmed, could compensate for its less-than-perfect parts and deliver overall perfect performance. “It’s this larger scheme that allows us to use reasonable but imperfect versions of these components,” explained Miller.

What are interferometers?

Interferometers are basically any device that separates and recombines light waves. Like sound waves, light waves can be combined so their signals add together. They also can ‘interfere’ and cancel each other out. This basic ‘on/off’ capability is what would allow interferometers to be harnessed and configured in a variety of ways.

Mach-Zehnder interferometers are specialised versions of these devices that split light from one or two sources into two new beams and then recombine them. They are already used for some specific applications in science and for switching beams in optical communications in optical fibres. Their more general use in consumer and other applications, however, has been obstructed because of the way that the light is initially split as it enters the device. Ideally, the beams would be split in perfect 50/50 symmetry. In reality, however, the split is not nearly so perfect. This means that when the interferometer recombines the signal it cannot be completely cancelled, preventing engineers from completely controlling the optical path.

The ability to combine or cancel the signals along a particular path is critical for technology. Researchers realised, however, that if Mach-Zehnder interferometers could be assembled in large meshes and controlled it would be possible to create a system that achieved the necessary perfect performance. This would allow the meshes to, in principle, perform any so-called ‘linear’ optical operation, much like computers are able to perform any logical application by controlling on/off functions of semiconductors.

Automatic control

The final element that enabled this process was the invention of algorithms — essentially the control software — that allowed the meshes to be ‘self-configuring’, adjusting how they directed the light paths based on the signal received by simple optical sensors embedded in the system.

This self-correcting algorithm allowed the researchers to propose meshes of interferometers with some imperfections and then compensate to make them behave as if they were perfect. The algorithms could then control the ‘phase shifters’ in the interferometers, determining if the signals combined or cancelled, by simply monitoring the optical power in various detectors.

“With this development, we are starting to do some things in optics that we have been doing in electronics for some time,” observed Miller. “By using small amounts of electronics and novel algorithms, we can greatly expand the kinds of optics and applications by making completely custom optical devices that will actually work.”