Hi,

I would like to add one point here . Photonics Integrated circuit is generally immune to the functionality losses related to electromagnetic pulses. Moreover, it can also be integrated with electronic circuit to increase functionality.

Thanks.

Obviously the data rate that you can transmit between the optical tracks and optical chip... Since silicon based integrated circuits are almost reaching its saturation... the growth of transistors and the size of chips are not in right proportion anymore.

However, photonic integrated circuits still need solid optical interconnect research... which is rapidly growing at the moment!

Thanks very much Moschim. Im not into the optical interconnects, but if we could transfer the knowledge track design with stubs acting as cap and inductor into the optical world, then we got a world with photons rather than electrons...

But, we are not there yet.... And yes, polarization is one of the toughest problem, but I believe in such short tracks, it would not be a big problem? Or we should introduce the theory of polarization maintaining fiber into this tracks.... its purely speculation... :)

The transmission data rate is very high in photonic integrated technology...

recently using erbium doped waveguide amplifiers and fiber amplifiers increasing the capacity of the system to fulfill the required demands of data communication.

Hello All,

I would like to contribute with my two cents. Sorry if I was a bit long.

The selling point of PICs is certainly the potential bandwidth and the absence of Joule effect as pointed out by previous answers. Next generation of processors, the so called exascale processing, will require a so dense network of copper interconnections that many serious industries (IBM, INTEL) are exploring the idea to replace them with optical interconnections.I would say that in this case integrated photonics demonstrates a clear potential advantage.

I would also add that many researchers are moving on the track of Quantum-Photonics in integrated devices (for example for quantum-cryptography), that promises to steer photonics in a direction quite orthogonal to electronic technologies. (In essence the advantage of integrated photonics could be simply in fields still not covered by electronics).

However a there are a number of technological constrains (subject indeed of investigation and fierce competition) that makes PICs still unpractical solutions respect to electronic implementations.

First of all, integrated electronics is fully nonlinear: you can switch, amplify, re-address, process, memorize ...etc. Taking advantage of the photonics "speed" requires to implement all those processes at optical level (fast photons control fast photons). Nonlinear optics in integrated photonics is evolving, but still the implementation of many of these functions is very challenging or at least not cost-effective.

There is also the question of the integration. Photonics is fast but electronics is dense.. very dense. When I was MSc student in optoelectronics I used to hear that electronics was approaching its density limit 130nm MOS gates... now we are at 22nm and 10nm gates are planned for 2016. With the exception of very top-notch technologies based for example on plasmonics (still at the very infancy at optical wavelength), the level of the integration in optics is pretty much determined by the wavelength (microns, or fractions of microns). In essence if you want to have a very high number of devices densely (separated let say by 100nm) packed together on a substrates you need to implement a number of tricks that normally have detrimental effects on the general performances.

Another interesting point is inherent in the optical technology. III-V semiconductors are a very well established phonic platform. One of the key reasons is the possibility to embed GAIN and optical sources (we do need light coming from somewhere), and to enable quite strong nonlinear interactions. But III-V are not exceptionally good in terms of losses, basically scattering (the index is quite high) and multiphoton absorption. In addition we are in a world dominated by Silicon (and its standard CMOS processes) which has similar limitations and makes the implementation of sources very difficult (there are some nice research on embedding GAIN in silicon). New technologies like in doped glasses and silicon nitride on silicon are emerging, they are pretty low-losses and nonlinear. Still the integration of sources is a challenge.

Marco

May I change the question: what is the advantage of photonic communication compared with electronic communication? Coaxial cable were used about 100 years ago with bandwidth of about 960 channels mostly for voice during the seventies. I was project manager of coaxial cable projects of more than 8,000 km connecting hundreds of towns during 1869 to 1979. Optical fiber networks have reached 300 wavelengths in research works and about 120 wavelengths commercially and a speed of about 400 Gb/s in research work and 100 Gb/s commercially and with signal to noise ratio of 10 to power of 12 which is incomparable with wireless or electronic communication. The only problem with photons is that there is no photonic memory and for this reason switching in nodes has to be done in electronic domain which has high power consumption , slow and more costly. But now research work is concentrating on how get round this problem, one of the solution is to use burst packet switching.

Another interesting feature is the possibility of using quantum effects. Photons can travel long distances with little interaction which makes them excellent tools for holding quantum information. This has applications in security where quantum cryptography can enable two parties to exchange encryption keys in the knowledge that their transmissions have not been intercepted. Storage of quantum information held on photons is another matter, as it is not easy to stop and store a photon. This storage would be done by transferring to an electronic medium. Photonic integrated circuits make possible photon interactions which in turn allow quantum processing to be performed. These are the quantum equivalent of electronic logic gates, and could be used as part of a quantum computer.