All-optical switching is a process by which light, usually in the form of digital communication signals, is routed from one transmission channel to another, or modulated, without intermediate conversion to another format.

In previous technologies, the routing usually involves the following steps: detection of an optical signal, an electronic routing decision, electronic triggering of a laser in the appropriate output channel, and generation of a new optical pulse. In the all-optical switching format, the signal remains in the optical domain and its properties are altered so that it is routed to a specific output.

All-optical switching now has two meanings. In the first, the switching operation is controlled electrically. For example, this may be controlled via the electro-optic effect. The induced phase changes are used interferometrically in a 2x2 switch. Switching between more than two channels is implemented by ganging multiple 2x2 switches.

The second meaning for all-optical switching is when even the controlling operation is done optically. That is, an optical beam, rather than an electrical signal, changes the optical properties of the medium and leads to interferometric control of the output. This approach typically utilizes an intensity controlled change in the refractive index.

However, both types require an externally induced change in the refractive index so that the device geometries and their applications are very similar. Electro-optic modulators, principlly using ferroelectric materials are already in system use, whereas the development of the completely optical concept is still futuristic and depends strongly on continuing materials development.

Alternative all-optical switching technologies

The nonlinear optics option has certain features that make it attractive, and perhaps the only viable solution as data rates climb into the terabit range and higher. All-optical switching with light beams relies on the response of the third-order nonlinear properties of materials, which can be as short as femtoseconds.

How does it work?

Any optical signal can be characterized by its spatial location, its arrival time at a specific location, its phase, and its polarization. An optically induced change in any one of these properties can lead to identification and subsequent rerouting of the modified pulse.

An example of an all-optical switching device is the Mach-Zehnder interferometer, usually in a guided wave format.

An Mach-Zehnder interferometer consists of six distinct elements, two of which constitute the input and output channels. The Y branch splits the input signal into equal signals in the two intermediate channels.

In the absence of illumination by an intense optical beam, the two signals propagate in their separate and identical channels and acquire equal phase shifts in each one. Therefore, at the second Y junction, the two signals recombine in phase and the input signal emerges unaltered in the output channel.

If one arm is illuminated with a control beam of intensity I, and that channel has some characteristic nonlinearity, there is an index change induced in the illuminated region. Therefore, there is an additional phase shift in the signal beam introduced into that arm over an illuminated distance and the signal sin the two arms are no longer in phase at the output Y junction. Hence the resulting partial destructive interference reduces the transmitted signal. When they are in a complete destructive phase change, complete destructive interference occurs and all light are scattered and absorbed and no light pass through the second Y function. Thus incur a all-optical modulation.