Skip to content
Skip to main content

About this free course

Download this course

Share this free course

Digital communications
Digital communications

Start this free course now. Just create an account and sign in. Enrol and complete the course for a free statement of participation or digital badge if available.

3.4 Optical switches

Optical space switching has been possible for a long time, but has been slow to find widespread application.

Solid-state optical switching (i.e. switching with no moving parts) can use devices based upon electro-optic materials such as lithium niobate (LiNbO3). An electro-optic material is one whose refractive index changes significantly when an electric field is applied across it.

Figure 23 shows a 2 × 2 switch element which exploits the electro-optic effect. This is a coupler, similar to those previously described, but in the coupling region the light travels in an electro-optic material such as lithium niobate instead of glass. When an electric field is applied across one or both of the light paths, the amount of light crossing between the two paths changes.

Figure 23 A 2 × 2 optical switch based upon a directional coupler

With suitable design, it can be arranged to switch between the ‘cross’ state (light entering I1 emerges from O2, and light entering I2 emerges from O1, as in Figure 24(a) and the ‘bar’ state (light entering I1 emerges from O1 and light entering I2 emerges from O2, as in Figure 24(b) under the control of the applied voltage, V.

Figure 24
The states of a 2×2 optical switch

Electro-optic switches such as these can be operated at high speed – switching between the two states can be as fast as 10 ps – but they are expensive, they generally introduce high losses in the optical signal (up to 8 dB) and can introduce distortion in the transmitted signal due to polarization effects.

A more recent development is the micro-electro-mechanical system switch. Micro-electro-mechanical systems (MEMS) are miniature electrically operated mechanical devices which can be constructed using the same materials and similar processing techniques as for large scale integrated electronic components. For optical switching, miniature movable mirrors can be made, each with dimensions of less than a millimetre. The movement of the mirrors can be controlled by an electrical signal, and incoming light beams from optical fibres can be directed to one of several different output fibres to perform the switching function.

With this technology, compact switching arrays with large numbers of crosspoints can be constructed as illustrated in Figure 25.

Figure 25
Figure 25 A space switch using MEMS technology

Two arrays of mirrors are used, one associated with the input fibres and one with the output fibres. Each of the mirrors on the input array can deflect the input beam to any of the mirrors on the output array. To switch from input fibre i to output fibre j, the mirror i on the input array deflects the beam from input fibre i to mirror j on the output array. Mirror j then deflects this beam to output fibre j.

Figure 25 shows a line of input and output fibres, but this is only to simplify the illustration. In practice the fibres (and mirrors) are in rectangular arrays, allowing more fibres to be packed into a small structure. They can be constructed with large arrays of crosspoints – typically 32 × 32 or 64 × 64 arrays, but up to 1000 × 1000.

Compared to switches based upon electro-optic materials, MEMS switches are slow (switching times of the order of 10 ms) but introduce less attenuation (of the order of 5 dB).

Because of the importance of optical switching, there has been a lot of research effort put into investigating different switch technologies, and the two presented here represent two extremes: fast, small arrays in the case of lithium niobate electro-optic switches and slow, large arrays in the case of MEMS switches. Lithium niobate electro-optic switches have been available for many years but have not yet been used commercially to any great extent. MEMS switches on the other hand have only been developed recently (2002) and are already starting to be exploited commercially.