03 Feb 2011
Intel-backed OpSIS officially kicked off on February 1, and is now looking for partners and to recruit staff interested in silicon photonics.
The University of Washington (UW) has officially launched a new foundry service program that promises to lower the cost of producing silicon photonic chips in dramatic fashion.
The Optoelectronics Systems Integration in Silicon (OpSIS) project is aiming to emulate the Metal Oxide Semiconductor Implementation Service (MOSIS) effort at the University of Southern California in the 1970s that is credited with ushering in the current era of electronic integrated circuit (IC) manufacturing in silicon.
OpSIS, which is part-funded by Intel, already a major proponent of silicon photonics technology through its own research, will offer so-called “shuttle run” production, where researchers can share the cost of processing silicon wafers between a large number of projects. That is the approach already being used at the ePIXfab silicon photonics manufacturing platform. Run by IMEC and CEA-Leti, this service will perform six shuttle runs in 2011.
Matt O’Donnell, dean of the UW College of Engineering, kicked off the OpSIS opening ceremony, saying: “It’s now clear that silicon photonics is becoming an integral part of the electronics world, and so it’s critical to have [this] type of capability.”
In close collaboration with BAE Systems, which will perform the chip fabrication, OpSIS will aim to offer three shuttle runs per year, each of which could accommodate 30-40 users.
Based within UW’s new Institute for Photonic Integration, the OpSIS service will be offered to any researcher in the world. A handful of users are already participating in so-called “risk runs” to test out the processing protocols in development.
One of those early users is John Bowers from the University of California, Santa Barbara. In collaboration with Intel, Bowers has already developed one of the most advanced silicon photonics devices, in the form of hybrid silicon-InP laser chips.
He says of the foundry offering: “By focusing on many different groups in one process line, that allows you to advance a library of components and processes faster than any one group could do on its own. It enables a faster evolution of photonic devices.”
Demonstrating Intel’s interest in silicon photonics, the company’s CTO Justin Rattner spoke at the OpSIS opening ceremony, saying: “OpSIS will enhance the education of US engineering students, giving them the opportunity to learn the new optical design paradigm.” Rattner added that the impact of silicon photonics would not just be felt in computing and communications – seen as the two most obvious applications – but also in life sciences, where the technology could lead to new low-cost optical sensors.
Intel’s own plans for developing silicon photonics products remain at an early stage, but Mario Paniccia, who leads the research effort at the chip giant, said last year that initial applications were seen in high-speed interconnects for data centers and server farms – with a potential introduction in around 2015.
According to UW’s Michael Hochberg, director of the Institute for Photonic Integration, the design and manufacture of photonic-electronic silicon chips like those developed through OpSIS will constitute a multibillion-dollar industry within a decade. He added that OpSIS' shuttle runs would aid that development by reducing production costs by more than 100 times for an individual research team taking part.
Hochberg was one of the speakers at a panel session on silicon photonics held at the recent Photonics West conference in San Francisco. That panel featured Ashok Krishnamoorthy, a Distinguished Engineer at the computing system giant Oracle who also sits on the advisory board of silicon photonics pioneer Luxtera – which also has close links with UW.
Krishnamoorthy outlined Oracle’s key position regarding the use of silicon photonics – saying that the company was essentially technology agnostic, and that the key issue to focus on was energy efficiency. “The industry rejects solutions that are not efficient,” he said, suggesting that the high power requirements of current photonic interconnects makes a commercial deployment unlikely within the next five years.
However, at the same panel IBM Zürich’s Bert-Jan Offrein hinted that the performance of silicon photonics would see the technology find use in supercomputing applications a little sooner, stating that the “biggest win” for photonics versus electronics was its relative lack of complexity.
• OpSIS is actively encouraging interested parties to participate in forthcoming shuttle runs. To register interest, contact OpSIS directly. The center is also recruiting. Interested parties should contact Michael Hochberg directly.