The optical chip: a world changing technology

There were no seats left in book store Van Piere on the evening of Tuesday April 24th. Some 80 people, of whom about one third were students, attended the Studium Generale Science Cafe about the optical chip. IPI representatives Ton Backx and Meint Smit informed an obviously interested audience about the vast amount of possible applications of this revolutionary technology, and the current state of affairs in research.

‘Who is working with chips on a daily basis?’, Ton Backx starts by asking the audience. Only a couple of people raise their hands. ‘I think that is not true,’ he replies. ‘Everyone of you is working with chips, all of the time. Your smartphone, your car, even your washing machine: they are all loaded with chips based on micro-electronics. As far as we are concerned, in the future, most of these chips will at least partially be based on photonics.’

Tens of thousands of Netflix movies, per second

Photonics uses light instead of electrons to carry information, he explains. That has some major advantages, since it is possible to transport more data at lower energy costs. ‘We are hungry for data,’ Backx says. ‘And our current technology is reaching its limits. Where with micro-electronics, you can typically download one Netflix movie per second, photonics will enable you to download 57.500 of them in the same time. Datacenters currently work at a typical speed of 100 Gbits per second. In two years, that speed should already be increased tenfold to 1 terabit per second. And where it now costs 1 picoJoule (10-12 Joules) per bit of information being processed, we should be using on tenth of that per bit in five years. The only way to achieve both goals, is by implementing photonics as an additional technolgy to the current micro-electronics.’

Eindhoven as a photonics hotspot

The ambitions of IPI and the Eindhoven region are high, Backx states. ‘Worldwide, there will be three hotspots for photonics, comparable to Silicon Valley for micro-electronics. And we will be one of them.’

‘That is quite an ambition. What makes you think this region can live up to that promise?’, is the immediate question that follows from the audience. ‘Because we are working on active components like lasers and amplifiers, and we have developed a generic technology. No one can get around us,’  Backx swiftly replies. Other questions demonstrate a same amount of broad interest, and variety of backgrounds of the audience: They vary from technical details to ethical debates and obstacles Backx expects along the way toward actual market introduction. To round up this first part of the evening, presenter Gijs van de Sande asks in what way this technology will have impacted our lives in 2050. Backx replies with an anecdote: ‘In the late seventies, IBM presented their first personal computer. At that occasion, the president of the company stated that he expected to sell at maximum 100 of them, “because the world doesn’t need more than 100 computers.” Who could have predicted then our current digitized society, or products like the smartphone? Photonics will have the same revolutionary impact, and it will enable applications beyond any of our current imaginations.’

A trip through thirty-five years of research

After a short break, Meint Smit takes over the floor. As a pioneer in the field of integrated photonics, he is ideally suited to guide the attendees through thirty-five years of research advancements in this field. ‘We started in Eindhoven by making optical waveguides, that transport signals from one place to the other. With these waveguides in our hands, we took the next step, and made so-called optical waveguide multiplexers: devices that split a signal into its different wavelengths, and the other way around, couple different wavelengths into one signal. This multiplexer is the basis for the success of glass fiber technology: since you can couple multiple wavelengths into one fiber, it is possible to achieve much higher bandwidths. That technology we developed in 1988 is now widely used: you can find it at the beginning and the end of each fiber.’

Sharing production costs

Smit gives examples of different technologies derived from those two basic components. Running from phase shifters that enable switches with switch rates of over 10 billion times per second, to light amplifiers and lasers with all kinds of specific features. ‘At TU/e, we were amongst the first to use standardized methods to make components in a reliable and reproducible way. In 2007 we pioneered the field again, and started organizing multiple wafer runs, which enable multiple designs to be printed onto one wafer. Processing one 10 centimeter wafer that is able to host hundreds of chips costs about 200.000 euros. Since in a developmental phase, you typically need less than ten chips of a specific design, that is too costly. But when researchers and companies can share these costs, the developments in optical chip making are sped up significantly.’

Today, over 350 different photonic integrated circuits have been produced this way, he says with pride. ‘These range from lasers, Terahertz and radiofrequency circuits to optical data handling systems and optical switches. Last year, we even managed to produce a record-breaking data handling transmitter that is able to transmit 360 gigabits per second.’
Over the last thirty years, Smit was part of a tremendous development in photonics, he concludes. ‘And now we are finally on the verge of breaking through to the market.’