Telecommunication hubs for ultra-fast dataflows

The universities of Eindhoven, Twente and Amsterdam have joined forces to develop a prototype optical chip that can form the basis for ultra-fast telecommunication nodes. The chip is essential to safeguard future communication networks against congestion, which would have disastrous consequences. Without fast communication hubs the communication networks worldwide would become paralysed. The consortium is aiming for a miniaturised optical circuit with a processing speed of at least 1 Terahertz. Optical transmission systems with a speed of 1 Terabit per second and higher are essential to network operators.

The current state of the technology offers 250 GHz as a maximum speed. However, in an experimental environment, Japanese researchers have achieved speeds of 1.28 Terabit per second. This set-up did, however, take up many square meters. The consortium aims to take up the challenge of investigating a technology that will allow for miniaturisation in an integrated circuit as well as achieve a processing speed of at least 1 Terahertz. The research will focus on a number of central points.

New materials, new characteristics

Ultra-fast integrated circuits now appear to be a possibility with new materials. Using nanotechnology, dot patterns, line patterns and three-dimensional optical filters can be realized on an atomic scale. Recent results show that such structures can possess new properties. Quantum dot, quantum wire and photonic bandgap technology in particular are applicable in this context. Quantum dots are minuscule islands in materials with dimensions in the region of one-billionth of a metre. If the islands are thread-shaped, they are referred to as quantum wires. The islands consist of atoms that are larger than those in the surrounding areas, which means that the islands bulge out of the surface almost like pimples, as it were. These quantum dots and quantum wires are now being produced in semiconductor materials. These are materials used to make fixed matter lasers, as used in compact discs. These dots enable the production of extremely small and fast lasers and optical switches.

Reducing the light angle

Photonic bandgap technology can make optical integrated chips even smaller. The existing chips have optical channels with large bends, as the light is inclined to ‘fly out of the corner’. Photonic bandgap technology can provide materials with characteristics that they do not posses naturally. For instance, using a matrix of extremely small distortions in the material, artificial mirrors can be produced in semiconductors used to make lasers and optical chips. Optical channels in chips can therefore, as it were, be made with the aid of very high quality mirrors, which means sharp bends can be realised without much loss. The advantage of structures with sharp bends and twists is that they can be made much more compactly than structures with gentle bends. It is therefore still possible to save space on a chip by making the light angle sharper.

Ultra-brief light pulses

Once it is possible to generate ultra-brief light pulses in a chip, using new technologies, a lot more data can be transmitted via a network. The starting point is a network capacity of 1 Terabit per second, or higher.

Faster switches

Switches are often produced based on resonators. In a resonator, the light must first be bounced ‘backwards and forwards’ before the desired effect is achieved.
By reducing the resonators in size, the effects are achieved more quickly. Micro-resonators, therefore, make it possible to produce ultra-fast switches.

Optical electronics

New technologies can generate ultra-fast processes in optical semiconductor amplifiers. These ultra-fast optical amplifiers can, in their turn, be the basis for the realisation of logical circuits; in electronic format these form the basis of current electronics. Clearly, we find ourselves on the eve of ‘optical electronics’. These optical electronics will therefore, with regard to the speeds to be realised for data processing, leave conventional electronics far behind. It is interesting to note that a number of terms that are common in normal electronics are also found in optical electronics, but then preceded by the word ‘optical’. For instance, optical flip flops, optical buffers and optical gates have already been realised.

Participating parties

Vrije Universiteit Amsterdam (Quantum Electronics Theory Group)
Technical University Eindhoven (Semiconductor physics group, Integrated Opto-Electronic Building Blocks Group and Telecommunication System Group)
University of Twente (MESA+ Institute)

Duration

Autumn 2002 - autumn 2006

Project manager

Daan Lenstra (Vrije Universiteit Amsterdam)
Department of Physics and Astronomy
De Boelelaan 1081
1081 HV Amsterdam
The Netherlands
E-mail: lenstra@nat.vu.nl
Tel.: +31 (0)20 - 444 78 55