 |
08 Nov 2006
An optical memory based on a semiconductor ring laser is under development in a new project by a pan-European consortium of universities and companies.
Photonics scientists have long sought an optical memory that could store large amounts of digital data and pave the way for all-optical computing. A new European research project, headed by Siyuan Yu at the University of Bristol in the UK, believes that the solution could lie in semiconductor ring lasers (SRLs).
The IOLOS project (Integrated optical logic and memory using ultra-fast micro-ring bistable semiconductor lasers) is aiming to exploit SRLs to develop integrated optical logic and memory devices. The project, which started in September 2006, will receive funding totaling €1.35 m over its three-year lifetime.
The SRL has been around for at least 20 years. Yu explains, "In the semiconductor ring laser the cavity shape is a sometimes a circular disc but the types we are working on are either a sort of racetrack shape or rectangular with mirrors at the corners. As the size of the full circular or the racetrack SRL goes down the degree of curvature increases as does the confinement requirement. This can cause a few problems which we hope rectangular cavity can avoid."
The team is working on the principle that light will encircle an SRL in one of the two possible rotational directions (clockwise or counter clockwise), which means that a laser beam will be emitted in only one of two directions. The direction of light emission can therefore represent the "1" or "0" state in a digital data storage system.
The device can "remember" its state, since recirculation is endless as long as power supply to the SRL is not interrupted. To set the direction of the recirculation a beam of light must first be launched in the desired direction, much in the same way as a "1" or "0" is written into a computer memory.
Background to SRL
Yu started researching the potential of the semiconductor ring laser about 15 years ago, while working at Glasgow University. The SRL idea was championed by Peter Laybourn and lead researcher at the time was Thomas Krauss, now Professor at St Andrews University.
"We were exploring the potential of the SRL as an alternative source to the Fabry-Perot laser, only to realize that it's a highly non-linear device," says Yu. "So together with people such as Guido Giuliani (Pavia) and Alessandro Scire (UIB) we worked on possible applications of this nonlinearity."
There are three potential applications of the SRL which the IOLOS project will explore:
• A digital response device that reshapes optical signals that have become distorted, effectively acting as a "gate".
• If the SRL operates under the control of another optical clock pulse it can compensate for timing "jitter" and retime a pulse train for optical transmission applications.
• The optical storage function already mentioned.
Yu and partners are hoping to reduce the size of the device and increase both the switching speed and the number of SRLs that can be installed on a chip.
Performance
Indium phosphide is this particular SRL's material of construction, and the laser light output is centered on the 1.55 µm telecoms wavelength. Pulse reshaping and retiming functions could occur at up to 10 Gbit/s, while for storage and recovery the experimental rate is likely to be 2.5 Gbit/s.
"Ultimately, the real area of interest will probably be at 40 Gbit/s, but we don't yet have to demonstrate that capability," says Yu. "That possibility is currently limited by our research budget but there should be no fundamental scientific obstacle to it."
Many other types of optical storage devices has been investigated -- such as bistable devices based on Fabry-Perot lasers, DFB lasers and non-linear crystals -- but Yu argues that each type has had its disadvantages. Some systems need an extra optical input to set the initial "1"/"0" state, while others have been limited by their sensitivity to environmental factors.
In contrast, SRLs are robust enough to function under a range of operational conditions. Indeed, the proposed IOLOS bistable device is self-holding and will operate over a current variation of 10 mA and at room temperature with some latitude. SRLs can also be miniaturized, which is crucial for mass producing multiples on a single semiconductor chip.
"Our proposed design has the modest aim of storing just one bit, that is, one instance of on or off," says Yu. "However, I would be fairly confident that we could extend this to one byte (8 bits or one keystroke/character)."
"We believe that the indium phosphide technology will gradually mature and become as easy to process and manipulate as silicon-based architectures as the yield continues to improve," he adds.
The project is still in a preliminary stage, but can benefit from many years of research in this area. Other partners in the project include the universities of Glasgow, Bristol, Pavia, the Balaeric islands, Brussels Frei University and industrial partners Intense, Bookham Technology (Caswell) and Siemens SA (Lisbon).
The European Commission has invested €1.25 million in the project, while Intense Photonics and Siemens are together providing a further €100,000. Siemens in particular is also working on commercialization potential of devices and will be supplying specifications from a systems point of view. Bookham is playing the role of general advisor with respect to potential manufacturing.
 |  |  |  |  |  |  |
| © 2024 SPIE Europe |
|
