Photonics roadmap calls for III-V design rules
"Today, the optical components industry stands at the threshold of a major expansion that will restructure its business processes and sustain its profitability for the next three decades."
That's one of the conclusions that has been reached by a four-year effort, co-ordinated by the Massachusetts Institute of Technology (MIT), to produce a roadmap for optical communications similar to that which already exists for the silicon industry.
Back at the height of the telecoms boom in 2000, MIT's Microphotonics Center Industry Consortium commissioned the Communications Technology Roadmap (CTR) to evaluate the vast array of potential technologies confronting the telecommunications industry.
Its study has concluded that the future of optical components technology will be determined by electronic-photonic convergence and short-reach (less than 1 km) interconnections.
For compound semiconductor chip manufacturers, the implications would be enormous.
Despite the massive restructuring over the past few years, the MIT report says that with a large number of suppliers still vying for a share of the photonic component market in telecommunications, one of two options must be followed by the sector:
The first is to move to an outsourced manufacturing model, since the amount of revenue available to individual companies will not allow for profitable support of an internal fabrication facility.
The second option is for yet further large-scale consolidation of the current supply base.
III-V integration focus
Integration in III-V materials has been one of the key focuses of MIT's effort, which included a working group comprising representatives from TriQuint Semiconductor, JDS Uniphase, Vitesse and NTT among others.
In one of many recommendations, it concluded that an industry effort to develop manufacturing processes and standard design rules for III-V technology is now required.
The III-V working group also reported that the integration of photonics and electronics is limited currently by the lack of fabs that are capable of producing large-scale optoelectronic chips on InP.
While much of the existing InP capacity, for example, is focused on 2- and 3-inch processing, these larger chips will need to be made on 4-inch wafers, said the III-V working group.
Since the move to 4-inch manufacturing is currently an unviable option for cash-strapped photonics companies, one solution could be to design products that can be made using standard processes originally developed for InP microelectronics, it added.
The Irish company Eblana Photonics is already working with US-based InP foundry Vitesse on just such a model, using the latter's HBT electronics process to fabricate lasers (see related story).
Critically, added the working group, the real issue that is hampering III-V integration is the lack of a volume market that could drive economies of scale in III-V manufacturing.
One option to solve this problem could be to find a high-volume application in a different market sector, such as photovoltaics or data storage, whose processes could cross over into communications. "Taking advantage of developments in high-volume markets may be the only path for economical III-V integration in telecom," said the working group.
