Nortel Gambles $1.43 Billion On Tunable Lasers
A good deal? The simple answer is that it's too early to tell. Coretek appears to have a lead in developing so-called VCSEL (vertical cavity surface-emitting lasers), but that isn't the only tunable laser technology around. And right now, there's a couple of big question marks over VCSEL lasers. First, will they be reliable, bearing in mind that they have moving parts, albeit microscopic ones? Second, will Coretek be able to manufacture them in large quantities, bearing in mind that even its CEO, Parvis Tayebati, describes the technology as "very difficult to master and develop"?
Coretek is hoping to begin pilot scale manufacturing in the fourth quarter of this year but Tayebati admits it's a challenge. "We are working very hard to get there fast," he says. The uncertainty is reflected in the "up to" inserted before the $1.43 billion in stock that Nortel will pay for the startup. The full amount won't be triggered until manufacturing comes on line.
One competitor reckons that Nortel is banking on Coretek being able to deliver tunable filters - devices at the receiving end of fiber links that sense incoming light rather than pumping it out - because they're a lot easier to make. Nortel is "probably taking a long term punt" on VCSEL lasers being viable, adds Rob Plastow, CTO of Altitun AB http://www.altitun.com, which makes tunable lasers based on alternative DBR (distributed Bragg reflector) technology. Plastow burst out laughing when he heard how much Nortel was paying.
It's worth noting that Nortel itself makes tunable lasers based on DFB (distributed feedback) technology. With this technology, the laser is heated up or cooled down to change the frequency of light that it pumps out, a process that can't deliver a very wide range of frequencies. Continuous heating and cooling also reduces the working life of such lasers.
Greg Mumford, Nortel's president of optical networks, now dismisses DFB lasers as "first generation technology" and says Coretek's VCSEL lasers offer several benefits. In particular, they can be tuned over a very wide range of frequencies (40 nanometers, according to Tayebati), they can switch from one frequency to the other very quickly, and they can lock on to a specific frequency. Theoretically, they can switch from one wavelength to the other in "nanoseconds" but in practice, the time it takes to lock on to a particular frequency is key, says Tayebati. That can take " a millisecond or so" according to Mumford.
Why the big interest in switching speeds? Because tunable lasers might end up replacing cross-connects in carrier networks, and might even go a step further and make it possible to switch light on a packet-by-packet basis in the longer term future (see Researchers Unveil All-Optical Advances). In general, one millisecond is fast enough for protection systems - automatically setting up alternative wavelengths to route traffic around failures. When switching speeds go down to nanoseconds, packet-by-packet routing becomes feasible.
Coretek's achievements were recognized at the recent Optical Fiber Communications (OFC) conference in Baltimore, where a panel of experts gave it a best of show prize for its developments (see What's Hot At The OFC)
But Coretek's technology has a potential drawback, because it's mechanical, albeit on a microscopic scale. Specifically, the frequency of its output is adjusted by raising and lowering a metal oxide layer using electrical currents. Altitun's Plastow says this mechanism can jam, and might also wear out over time, raising a big question mark over the technology's reliability. Tayebati refutes this, saying Coretek has conducted extensive tests that demonstrate that its lasers will work for 800 years.
-- by Peter Heywood, international editor, Light Reading http://www.lightreading.com