Planar Integrated Optics
Planar integrated optics plays a key role in many fundamental optoelectronic devices. These devices can be conveniently classified according to the materials on which they are based.
- PDF / 901,666 Bytes
- 58 Pages / 439.37 x 666.142 pts Page_size
- 13 Downloads / 238 Views
9.1 Objectives, New Devices and Challenges Planar integrated optics plays a key role in many fundamental optoelectronic devices. These devices can be conveniently classified according to the materials on which they are based. Thus, doped silica/silica systems are used for the fabrication of passive integrated devices with an extremely tight tolerance. As an example of such devices, we may mention phased arrays or ‘PHASARs’, which are integrated spectrometers intended at multiplexing or demultiplexing signals. Semiconductors, and more specifically III-V semiconductors, are essential to the operation of laser diodes and compact optical amplifiers used for generating and reinforcing signals. Almost all such devices include a waveguide for maximizing the interaction between light and matter. Semiconductors also allow the implementation of a large variety of optoelectronic functions and nonlinear optical functions, provided however that the carrier dynamics can be adequately controlled, in particular at the picosecond time scale, which now corresponds to bit rates much higher than 100 Gbits/s. Dielectric materials with high electro-optical coefficients are more particularly interesting for the realization of modulators or very fast nonlinear functions (without carrier) used for the generation of signals by the frequency sum or difference of two signals. Lithium niobate (LiNbO3) is an example of such a material, which in addition can be polarized. Finally, attempts at introducing organic materials are continually being made, for the realization either of low-cost devices or of optical backplanes in board-to-board interconnects. In addition, some of these organic media present a thermal drift of the refractive index dn/dT < 0, in direct contrast to the standard case, where dn/dT > 0. Such a property is of particular interest for the compensation of the thermal drift. As suggested by the mention of temperature compensation, integrated optics is already a few decades old, enough to find solutions to such problems as the insensitivity to polarization or parasitic reflections. For all the functions mentioned above, the pioneering devices have been indeed developed in the 1960s. Contrary to electronic integration, optical integration has not developed at the fast pace of Moore’s law: no more than two or three functions can currently be cascaded on sixteen channels, and more commonly on two or four channels only. The reasons
288
9 Planar Integrated Optics
for the limits imposed on integration are far from being trivial. A first limitation is the higher cost of fabrication of small-scale device. Another limitation lies in the necessity of a coupling of these devices to single-mode fibres with at least a sub-micrometer resolution, without any excessive reflection: the easy mounting of ‘optical chips’ on ‘optical printed circuit boards’ still remains a remote prospect. A third limitation is the existence of incompatible technologies, and the necessity of resorting to a hybrid integration of devices: for instance, chips realize
Data Loading...
