Get up to speed on tunable laser developments * Progress * Potential * Purveyors

January 10, 2003

32 Min Read
Tunable Lasers Revisited

Nearly two years ago, Light Reading published a groundbreaking report introducing tunable lasers – devices that can switch from one wavelength to another on demand (see Tune In!).

Tunable lasers promised to revolutionize the optical networking industry by keeping a lid on the problem of the escalating cost of buying, stocking, and managing spares for DWDM equipment. And that was just for starters. Other applications for tunable lasers stretched from enabling automated network provisioning to reconfigurable network architectures.

How much of this has come to pass? Sadly, a lot less than expected. The bursting of the optical bubble has delayed carriers' plans for deployment of tunable lasers and put financial pressure on the companies developing the technology.

But tunable laser technology refuses to die. In fact, vendors remain quietly confident that their patience will be rewarded in the not too distant future. "In two years time, every laser will be a tunable laser," contends James Regan, managing director, Europe, for Agility Communications Inc. Regan says his confidence comes from talking with service provider customers that are now making concrete plans for deployment, as opposed to just talking the talk, as they did a few years ago.

If true, then the market is poised for takeoff, making it one of the few areas of optical components technology that looks set to do so – all the more reason to revisit the topic of tunable lasers.

Like the previous report, this article aims to provide a balanced overview of the market opportunity, technologies, and vendors of tunable lasers. The competitive landscape has changed considerably from two years ago, as some vendors have been swallowed up by the optical depression, and new ones have appeared to replace them. Vendors have also refined the business case for using tunable lasers, and this is detailed in the report.

This report can be read sequentially, or alternatively, simply click to the page of interest using the following hyperlinks:

  • Applications

  • Business Case

  • Market Overview

  • Challenges

  • Technology & Vendor Survey:

    • Narrowband Tunable Lasers

    • DFB Laser Arrays

    • 'DBRs on Steroids'

    • External Cavity Lasers

    • VCSELs

Here's some background reading that may also prove useful:

  • Beginner's Guide: Laser Basics

  • Beginner's Guide: Wavelength Division Multiplexing (WDM)

  • Beginner's Guide: Distributed Feedback (DFB) Lasers

  • Beginner's Guide: Vertical Cavity Surface Emitting Lasers (VCSELs)

  • Beginner's Guide: Tunable Lasers

— Pauline Rigby, Senior Editor, Light Reading

As usual, this report was previewed in a Light Reading Webinar sponsored by two leading vendors of tunable lasers: Agility Communications Inc. and Iolon Inc.
The Webinar is archived here.

Some readers may remember our story from two years ago – Researchers Unveil All-Optical Advances – describing how tunable lasers would usher in an era of all-optical networks. Separate teams of researchers in the U.S. and Norway carried out experiments on a new network architecture in which tunable lasers directed traffic to different places by selecting the appropriate wavelength for each destination.

Wavelength routing, as this application is called, certainly fired up imaginations. But it was too revolutionary to be the first step for a new technology and still remains firmly in the research phase today.

However, the growth of DWDM systems gave carriers another, more immediate reason to consider tunable lasers. As channel counts increase, so does the cost for buying, storing, and managing spares for the system, one spare being required per wavelength. A single tunable laser could replace four, eight, or even all the channels in a DWDM system, leading to considerable reductions in both capital and operating expenses.

The next obvious step is to use tunables as both the primary and backup lasers in the system. This places more stringent demands on performance, which needs to be as good as any standard DWDM laser. Benefits come from the fact that the carrier is dealing with a reduced number of part numbers for the overall system, not only for the spares.

The idea of using tunable lasers as spares has been the main driver for their development. The business case for this application is considered on page 3. However, it's not the only potential application, as the following list goes to show:

  • Sparing
    Use tunables to reduce the number of line cards needed to back up all the different wavelengths in a system.

  • One-time provisioning
    Tunable lasers are used as both primary and backup transmitters. The application is called one-time provisioning because, once the wavelength has been set, it is fixed for the lifetime of the equipment. It cannot be changed at will, because most filters and other wavelength muxes are wavelength specific.

  • Dynamic provisioning
    The wavelength of the tunable transmitter can be changed once the system has been deployed; for example, to meet the wavelength requirements of a new customer's equipment that has just been attached to the network. Requires the use of tunable filters or wavelength multiplexers.

  • Reconfigurable optical add/drop multiplexers (ROADMs)
    A simple, more flexible architecture for ROADMs has been proposed, which relies on the use of both tunable lasers and tunable filters. Tunable lasers can add any wavelength to the system, while tunable filters can remove any wavelength from the system.

  • Optical Crossconnects
    Tunable lasers can remove wavelength-blocking issues in OXCs. Right now, most OXCs use optical-electronic-optical (OEO) conversions on both sides of the fabric to avoid this issue. If tunable lasers are used at the input to the OXC, the wavelength can be chosen to make sure it has a clear path all the way to the destination.

  • Dynamic restoration
    When a DWDM channel fails, a tunable laser could automatically restore service. For this to work, the laser must be able to tune and lock onto the failed wavelength in around 10 milliseconds or less – in order to keep the total restoration time under the Sonet requirements of 50 ms.

  • Wavelength routing
    Tunable lasers could be used to eliminate complicated all-optical switches altogether, allowing them to be replaced with simple, fixed crossconnects. This requires a change in the way networks route signals. A mesh of virtual connections would exist, with each wavelength connecting to a unique destination. To route the signal, a tunable laser needs to tune its channel to the appropriate wavelength for that destination.

  • Optical packet switching
    True optical packet switching requires signals to be wavelength routed on a packet-by-packet basis. For this to become a reality, it must be possible to switch the laser in a very short time – on the order of nanoseconds – so that the switching doesn't introduce too much latency into the system.

The following figures come from a study by Agility Communications Inc., based on a simple model of a DWDM metro ring network comprising four nodes and 32 channels at each point on the ring.

In the fixed-wavelength case, we assume that the sites share a central sparing resource, which is stocked with 32 line cards – one per wavelength on the ring. The cost of a fixed-wavelength line card is $8,000. A total of 160 line cards are required for the network, of which 20 percent are spares. The total cost of the spares is $256,000.

If tunable line cards are used as spares, the situation changes. Rather than having a shared sparing center, each site now holds two spare line cards. The total number of line cards is now 136, with eight of them as spares (6 percent). The increased cost of the line cards must also be factored in, to allow for a premium in the cost of a tunable laser compared to a fixed wavelength laser. In this example, we assume that there is a $1,000 premium for the laser, so the line card itself costs $9,000. Despite the premium, the cost of spares is reduced to $72,000, a savings of 72 percent compared to the fixed-wavelength example above.

A similar example, with the same assumptions, but only 16 channels populating the system, yields savings in the region of 44 percent. Clearly, tunable lasers afford big savings even at low channel counts.

In real life, however, there are many other factors to take into account. More sites may be served by the same sparing center. However, it is more likely that spares would be located at each node, rather than at a central site, to reduce the number of truck rolls, in which case the savings would be much larger.

Try this interactive calculator from Agility Communications to work out savings for different network configurations.

This model is based on capital costs only, and ignores the savings in operational expenditure (opex), which result from the reduced cost of stocking and managing the inventory of lasers. This is difficult to quantify, but tunable laser vendors do expect it to constitute a significant portion of the overall savings.

How big is the market?

This report has already considered a range of applications and shown that there is a strong business case for deploying tunable lasers in the most immediate application: sparing. Two questions immediately spring to mind. One, what is the status of the tunable laser market today? And two, how big is the market opportunity going forward?

In answer to the first question, the market today is small. Any forecast that claims otherwise is out of date. In fact, the size of the market is so small that many analysts have dropped out of it altogether (see Tough Times for Tunable Lasers).

Small doesn't equal zero, however. One particular type of tunable laser – the narrowband tunable – has been deployed in volume. For example, Fujitsu Network Communications Inc. (FNC) claims to have shipped more than 5,000 of its four-channel tunable lasers in the period up to June 2002. (See page 7 for a closer look at narrowband tunable lasers.)

One of the reasons narrowband tunable lasers reached deployment earlier than their widely-tunable cousins is because they are based on a technology that's very similar to standard DFBs. That means the technology is more mature. There's also a higher "comfort factor" with service providers and carriers, which is important.

Nevertheless, widely tunable lasers were expected to break into the market a lot sooner than they have. Why didn't they?

Bad timing for one. A few years ago, when carriers were massively building out their networks, widely tunable laser technology wasn't ready. So, carriers have huge stocks of fixed-wavelength devices to use before they consider investing in new technology. Now that tunable laser technology is ready, there are fewer network buildouts. Another factor is that carriers have delayed the adoption of more advanced applications – those that fixed-wavelength lasers can't address.

Vendors believe that excess inventory is the dominant factor affecting the tunable laser market. "Even if tunable line cards had been ready a year ago, I don't think that would have changed the deployment story very much," contends Saied Aramideh, VP for product development at Iolon Inc.

Looking forward

The question of market opportunity for tunable lasers going forward is tricky, especially without a convenient market report to fall back on.

Instead, let's look at some recent design wins for tunable lasers. There were a number of customer announcements by tunable laser companies in 2000 and 2001, but most of these turned out to be premature. Many of these customers were startups, which either changed their plans or went out of business (see Scattered Signals for Tunable Lasers).

The quality of design wins in 2002 looks a lot more promising (see Table 1). They include a couple of Tier 1 players, like Lucent Technologies Inc. (NYSE: LU). In some cases, such as Lightscape Networks Ltd., which is retrofitting an existing product with tunable lasers, there appears to be a reasonable chance of the design win turning into a significant amount of sales.

Table 1: Widely Tunable Laser Design Wins



Read more about who's buying tunable lasers:

  • Iolon Lands Lucent Deal

  • Clouds Lift on Tunable Lasers

  • Mahi Gets a Fresh $75M

  • Iolon and Innovance Deal: Good Omen?

  • Bandwidth9 Scores a Coup

Tunable laser technology is significantly different from that of fixed-wavelength lasers, and as a result it's had to tackle and overcome plenty of challenges.

In a nutshell, the tunable laser manufacturer has to meet the key concerns of a systems vendor or carrier: performance, reliability, and expense.

There is also the additional challenge of manufacturing: It is one thing to make a high-performance, reliable component, and quite another to do so in volume. Manufacturing has a huge influence on both performance and cost. (Note that cost and price are not exactly the same thing!)

Performance

It's difficult to generalize here, since the performance limitations of tunable lasers depend on the technology inside the laser. For example, in distributed Bragg reflector (DBR) lasers, there is a compromise between output power and tuning range. VCSEL-based lasers also tend to deliver extremely low powers, although this is independent of tuning range.

One way around the problem of low power is to integrate Semiconductor Optical Amplifiers (SOAs) with the lasers, but this too has drawbacks of increased manufacturing complexity, and a higher noise, which ultimately leads to more bit errors on a long-haul link.

Different tuning mechanisms have different tuning speeds. Thermal tuning is slowest, taking up to several seconds for the laser wavelength to stabilize. Electronic tuning is the fastest and can be done in milliseconds or less.

For more details, read the individual technology overviews.

Reliability

Even when vendors have got their prototypes up and running, they face many hours of testing before carriers will be willing to install their products in live networks. Fixed-wavelength DWDM lasers have to pass the requirements of Telcordia Technologies Inc. standard GR-468-CORE, and this is the standard that tunable lasers have also set out to meet. The standard tests device lifetime, among other things, which is why it's impossible to speed up testing.

The importance of this qualification cannot be overstated. As Arlon Martin, Agility Communications's VP of marketing says: "Telcordia testing has to happen before deployments can start with carriers."

But so far, only two vendors – Agility and Coretek (now defunct) – claim to have passed Telcordia, although others say they are getting close. Agility claims to have qualified its first product only, a 4mW laser (see Nortel and Agility in Tiff Over Lasers).

However, there is an unresolved issue, in that Telcordia GR-468-CORE was designed for fixed-wavelength devices, and it doesn't take into account certain key tunable laser parameters. It is based on proving the laser can withstand harsh environmental conditions – if you hit it with a hammer will it still work? Additional tests will almost certainly be required to prove the reliability of tunable lasers to the satisfaction of customers, but these haven't yet been standardized, which means each vendor is testing them in a different way.

Expense

To start with, systems vendors appeared prepared to pay a considerable premium for tunable laser technology, but with the downturn in the market, that willingness has gone away. Now systems vendors are talking about a premium of 10 to 20 percent maximum.

However, there is some confusion over exactly what this means. Does it refer to the premium of the laser itself, compared to a fixed-wavelength one, or the premium for the whole line card? Assuming that all other components remain the same, a 100 percent premium at the laser level can translate to a mere 10 percent premium at the line-card level.

Whichever way it works out, the upshot is that there is increased pricing pressure on tunable laser vendors.

Iolon claims to have met this challenge, saying that it has struck a deal with Lucent Technologies to deliver tunables at a "single digit" premium (see Iolon Lands Lucent Deal). But Iolon, like all other tunable laser vendors, declines to give actual figures.

From the tunable laser vendors' point of view, the key to bringing prices down lies mostly with manufacturing – can they design products that are simple and cost-effective to manufacture?

Manufacturing

In general, the designs of tunable lasers are more complicated than those of fixed-wavelength devices, requiring either additional semiconductor process steps, or more parts that must be assembled. Many components vendors believe that the key to manufacturing is automation. Not only can this bring down the cost of manufacturing itself, by eliminating fiddly, time-consuming hand-assembly processes, but it is also liable to produce more consistent results, leading to higher yields for the overall process. Yields affect costs in a big way.

A related challenge for certain types of tunable laser is simply characterizing them. Every laser is slightly different when it rolls off the production line. As a result, each one needs comprehensive testing to work out which combination of currents will produce the correct output. Doing this manually would take days – which is why several startups have focused on producing test gear and software that will speed up this process. These include Fiberspace Inc. (through its acquisition of Tunable Photonics), Intune Technologies, and Tsunami Photonics Ltd. (see Tsunami Joins Tunable Party).

There's no such thing as a typical tunable laser. However, it is possible to group them according to technology type.

This is a worthwhile exercise, because the underlying technology influences the performance, manufacturability, and reliability of tunable lasers. In the early days, it was very much a battle among technologies, each being characterized with particular strengths and weaknesses.

To a certain extent, the technology war is subsiding. Vendors have been concentrating on improving the weaker areas of their technology, and that's starting to pay off.

At the end of the day, customers for these tunable lasers want to be able to distance themselves from the underlying technology completely. They simply need to know that a tunable laser module meets their performance requirements, and how to design it into their systems.

Steps have been made towards removing the distinctions among different technologies, as tunable laser vendors start to develop multisource agreements. However, it's still early days for MSAs (see Tunables Get an MSA, OIF Sets Component Specs, and Components Standards: Key to Survival).

The rest of this report reviews each technology category, and the vendors in it, in turn. Click on the following hyperlinks to dig into the details.

Narrowband Tunable Lasers

  • Distributed Feedback (DFB)

  • Distributed Bragg reflector (DBR)

Widely-Tunable Lasers

  • DFB laser arrays

  • 'DBRs on steroids'

  • External cavity lasers

  • Vertical-cavity surface-emitting lasers

Narrowband tunable lasers have a tuning range that's much less than the full C-band (the wavelength range 1625 to 1665 nm used for DWDM transmission). Coverage of four to eight channels is typical. This stands in contrast to so-called widely tunable lasers, which are expected to cover the complete C-band with a single device.

As a result of their limited tuning range, narrowband tunable lasers are only really suited to the inventory management applications of sparing and one-time provisioning. More advanced applications will require widely tunable lasers.

Technology Description

There are two main types of narrowband tunable lasers: distributed feedback (DFB) and distributed Bragg reflector (DBR).

DFB lasers are based on the same technology as fixed-wavelength devices. Like the lasers found inside CD players and the like, light bounces back and forth between two mirrors, which are formed by cleaving the facets of the semiconductor crystal. The region between the two mirrors is called the gain region, because that's where amplification of light takes place in the material.

Unlike a CD laser, a DFB laser has a grating etched into the gain region. This grating serves the purpose of stabilizing the frequency of the laser, making the wavelength precise enough to use in DWDM systems. However, this inherent stability also eliminates most methods of tuning the laser. The only method that works is heating and cooling the laser, which allows tuning over a range of around 3 nm – far less than the 35 nm range of DWDM systems.

Tunable DFB lasers have been around for some time. They are relatively simple to make, have simple controls, and have performance characteristics similar to those of fixed-wavelength devices, including the ability to emit high powers (up to 20 mW).

A second, more complicated type of laser, the DBR laser, has been developed, which extends the tuning range to about 12 nm. In this type of laser, the gain region and the grating are spatially separated, so it becomes possible to adjust the properties of the two sections independently.

A typical configuration is a three-section device, consisting of a grating, a gain region, and a so-called "phase region." Current can be injected into the phase region to alter its refractive index, without affecting the refractive index of the other two sections, especially the grating. This change in refractive index alters the cavity length and, hence, the wavelength of light coming out of the laser.

DBRs require more complicated control systems. The device is controlled by up to four input currents. Since every laser is slightly different when it rolls off the production line, each wavelength requires a unique combination of currents, and so the calibration process can be time consuming. Once calibrated, the usual method of control is to use a microprocessor with calibration data burned into memory. In addition, a tunable DBR usually requires a wavelength locker, which generates a feedback signal that is used to stabilize the wavelength.

On the upside, electrical tuning is much faster than thermal tuning, on the order of ten milliseconds, compared to ten seconds.

Vendors

Alcatel Optronics (Nasdaq: ALAO; Paris: CGO.PA)

Alcatel announced a thermally-tuned DFB laser in October 2001. The device is CW (continuous-wave), covers eight channels at 50 GHz, and offers output powers up to 20 mW.

Fujitsu Ltd. (KLS: FUJI.KL)

Fujitsu develops thermally tuned DFB lasers. It started with single devices covering four channels and has now moved on to DFB laser arrays, to expand the tuning range.

Intel Corp. (Nasdaq: INTC)

Intel is spreading its bets when it comes to tunable lasers. As well as acquiring two tunable laser firms, it has internally developed an eight-channel tunable transponder that supports OC192 and 10-gigabit Ethernet transmission (see {doclin k20725}).

JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU)

JDS Uniphase introduced a CW tunable transmitter covering 24 channels at 50 GHz in September 2002. The output power is 20 mW. It has also developed thermally tuned DFB devices.

Multiplex Inc.

Multiplex's device is a 16-channel EML (electroabsorption modulated laser) – one that contains a DBR plus an integrated modulator. It offers both 2.5- and 10-Gbit/s transmitters, as well as a tunable transponder suitable for both OC192 and 10-gigabit Ethernet transmission.

Nortel Networks Corp. (NYSE/Toronto: NT)

Nortel's main claim to fame in the tunable laser space is its purchase of Coretek, a vendor of widely tunable lasers. But Nortel also had developed its own range of thermally tuned devices offering four channels at 100 GHz spacing. It's not clear whether this was part of the package acquired by Bookham, which has its own tunable lasers, or if Nortel's products were discontinued.

TriQuint Semiconductor Inc. (Nasdaq: TQNT)

Triquint acquired the optical components business of Agere Systems (NYSE: AGR/A) in a deal that closed last week (see TriQuint to Acquire Agere's Optics).In September 2000, Agere, which was then still part of Lucent, had introduced a 2.5-Gbit/s EML covering 20 channels at 50 GHz spacing. It followed this up with a 10-Gbit/s transmitter in March 2001.

Technology Description

Laser arrays achieve a wider tuning range by integrating a number of DFB lasers on a single chip. Individual devices are designed to cover a sequential part of the overall tuning range, so only one of them need operate at any one time. Regardless of which device is operating, the optical signal must go to a common output port on the chip. Researchers have investigated parallel and serial configurations of lasers, but all commercial devices use the parallel version.

This diagram from Fujitsu Quantum Devices Ltd. shows a typical device. There are eight lasers side-by-side, a combiner, and in this case an amplifier, which makes up for optical losses in the combiner. Other vendors have different ways of merging the outputs from the individual lasers in the array. Santur Corp., for example, uses a MEMS mirror to select among the outputs. It claims that this approach has lower optical losses, which removes the need for additional amplification inside the device.

The biggest advantage of this kind of laser is that its performance can be similar to that of fixed-wavelength devices, with high output powers, good wavelength stability, and simple operation.

On the negative side, tuning is slow because it is based on temperature. There's also a question mark over yields, because of the complexity of the laser chip.

Vendors

Fujitsu Quantum Devices Ltd. – a division of Fujitsu Ltd. (KLS: FUJI.KL)Fujitsu has developed tunable laser arrays with up to 44 channels, which are for internal consumption. Right now, it is shipping 22-channel devices with the Flashwave DWDM system (see Fujitsu Improves Flashwave).

According to its scientists, Fujitsu's laser arrays comprise one laser stripe per channel, which eliminates the need for thermal tuning, but adds to the size and complexity of the overall chip. "A 22-wavelength tunable laser would have at least 22 stripes," says a spokesperson. "It may include several extra stripes to allow for the inevitable semiconductor fabrication yield issues."

Quantum Devices Inc. (QDI) (not the same as Fujitsu Quantum Devices!)

In September 2002, QDI announced development of a temperature-tunable, eight-element DFB laser array chip integrated with a combiner and an amplifier. The device covers 20 nm or 40 channels at 50 GHz spacing.

Santur Corp.

Santur's device contains twelve DFB lasers to cover different parts of the spectrum through heating and cooling. The signals are coupled to the output using a MEMS mirror that, the startup claims, introduces hardly any optical loss (see Tunable Lasers: Back in Fashion?).

Technology Description

Christened "DBRs on steroids," these are DBR lasers with enhancements to widen their tuning range.

These lasers have at least four sections – typically two Bragg gratings, one gain block, and one phase module for fine tuning. This means that they're longer than regular DBRs, so the light has to travel farther before it escapes, by which time it's lost some of its power.

Other sections can be added, such as a modulator section and an amplifier section to boost the output power.

For more technical details, read the individual vendor profiles.

Vendors

Agility Communications Inc.Proper name: Sampled-Grating DBR (SG-DBR).

What this means: Each end of the laser has a section where a number of samples of Bragg gratings are located. In other words, there's a stretch of corrugations at a certain spacing, then a space, then a stretch of corrugations at a different spacing, then a space, and so on. This has the effect of creating a "comb" of different wavelengths. The Bragg grating samples at either end of the laser create different combs of wavelengths, and when light is bounced between them the two become superimposed – the result is a much wider range of wavelengths.

Related stories in Light Reading:

  • Clouds Lift on Tunable Lasers

  • Nortel and Agility in Tiff Over Lasers

  • Agility Turns Out Tunables

  • Agility Gets $83M Third Round

  • Agility Packs Three Into One

  • Agility Unveils Long-Haul Laser

  • Agility Launches First Product

  • Agility Communications, Inc

Bookham Technology plc (Nasdaq: BKHM; London: BHM)

Bookham acquired its tunable laser technology by purchasing Marconi Optical Components in 2002.

Proper name: Digital Supermode DBR (DS-DBR).

What this means: The back end of the laser contains a sampled grating, while the front end contains a novel type of grating that simplifies the "tuning map" of the laser. In other words, there are fewer possible wavelengths that it can reflect, making it easier for the laser to lock onto a useful channel. As a result, only two currents are needed to tune the laser – a coarse tuning current and a phase current for fine tuning. The front grating is also shorter than a sampled grating, which enables the device to produce higher output powers.

Related stories in Light Reading:

  • Bookham Gets a Bargain

  • Marconi Claims Tunable Laser Advance

  • Marconi Components: Up For Sale

NTT Electronics Corp. (NEL)

Proper name: Superstructure-Grating DBR (SSG-DBR).

What this means: This laser works in a similar way to the SG-DBR laser, except for the design of the gratings at each end of the laser. The gratings are designed with a sampled grating and merged with another sampled grating of a different spacing, resulting in a superstructure grating. This provides the advantage that the reflection peaks obtained from these gratings are more uniform, hence reducing the output power variations during tuning.

Honorable Mentions

ADC/Altitun

Swedish startup Altitun was one of first to market with tunable lasers. In 2000, it was bought by ADC Telecommunications Inc. (Nasdaq: ADCT); but two years on, ADC decided to quit the optical components business, either by selling its interests in this area or closing them down. At the time of writing, no buyer has been found for Altitun, and its assets are to be sold by Dovebid on February 11.

Proper name: Grating-assisted Co-directional Coupler with Sampled Reflector DBR (GCSR-DBR).

What this means: Altitun's devices are made in four sections: a gain block; a phase module, used for fine tuning the wavelength; a coupler; and a Bragg grating. The coupler acts like a coarse tuner, transferring power vertically between the two waveguides – one that runs forward to the gain block, and another one, above it, that runs backward into the phase and Bragg grating sections. Note that the grating or phase sections are active in the top waveguide only.

Related stories in Light Reading:

  • ADC Tunes Out

  • ADC Announces Tunable Laser

  • ADC Scores a Coup on Tunable Lasers

Sparkolor

Sparkolor's laser is/was similar in concept to Altitun's, the main difference being that Sparkolor had opted to make different sections of the laser from different materials: indium phosphide for the gain and phase sections, silica for the grating. Sparkolor did announce a product, but it was later snapped up by Intel Corp. (Nasdaq: INTC). It's not clear how Intel plans to use the technology.

Related stories in Light Reading:

  • Intel Buys Sparkolor's Assets

  • Sparkolor Secrets Surface

  • Sparkolor Plays Catch Up

Technology Description

Tunable lasers for test and measurement applications are external cavity lasers (ECLs), containing a gain chip and separate gratings or mirrors to form a cavity. Such lasers are too bulky to be used for DWDM, but vendors have come up with ways of shrinking down these designs by using MEMS (micro-electro-mechanical system). The actual configuration of the cavity varies from vendor to vendor. Some turn a grating, just like test and measurement lasers. Others use a device more like an airbag accelerometer to move a mirror back and forth.

Most ECLs use MEMS – in other words, they have moving parts, which presents a potential reliability problem in the eyes of carriers. But vendors dispute this. "There is no issue in terms of vibrational sensitivity," claims Saeid Aramideh, VP of product development at Iolon. "We have Telcordia testing, and also our customers' testing to back this up."

On the upside, this type of laser is capable of delivering very high output powers, of at least 20 mW and up to 40 mW in some cases.

Vendors

Iolon Inc.

Iolon's laser is continuously tunable over the whole C-band. Tuning is done by turning a mirror to direct light onto different parts of a diffraction grating. The mirror is turned using a tiny actuator made using MEMS technology.

Individual parts of Iolon's laser are simple to manufacture and easy to assemble, the company claims. In fact, Iolon outsources all its manufacturing, which is an important factor in helping keep costs down.

Related stories in Light Reading:

  • Iolon Lands Lucent Deal

  • Iolon and Innovance Deal: Good Omen?

  • Corning Backs Laser Startup

Intel Corp. (Nasdaq: INTC)

Acquired the tunable laser business from New Focus Inc. (Nasdaq: NUFO). No product has been announced to date.

Related stories in Light Reading:

  • Intel Scoops Up New Focus Laser Unit

  • New Focus, New Laser

Picarro Inc. – formerly BlueLeaf Networks

Picarro is still in stealth mode, but given the properties it is claiming for its laser, it is likely to be using an external cavity device.

  • A Discreet Peek at Picarro

  • Blue Leaf: From Gas to Glass

Princeton Optronics Inc.

Having announced that it was making a VCSEL-based laser similar to that of Coretek, Princeton Optronics then went very quiet for the next couple years. Now Andy Quinn, the startup's CEO tells Light Reading that it has abandoned VCSEL design in favor of "something else," but he isn't saying what that something else might be.

In September 2001, the startup demonstrated tuning over the entire C-band at 25 GHz spacings (see Princeton Optronics Demos Tunable).

Related stories in Light Reading:

  • Princeton Optronics Demos Tunable

  • Princeton Optronics Takes On Coretek

Technology Description

Vertical Cavity Surface Emitting Lasers (VCSELs) give out laser light from the surface and have a laser cavity that is vertical – hence the name. This type of laser dominates in datacom transmission, which uses a shorter wavelength of 850 nm. Making VCSELs that operate at 1550 nm is an altogether different and trickier proposition, however, because of materials issues concerning mirror reflectivity and heat generation in the device (for more details see Laser Blazers).

As a result, several vendors have used optically pumped configurations – using a short-wavelength laser to "excite" another, which emits at 1550 nm. This adds to the manufacturing cost and complexity.

Tuning is usually via a MEMS mirror on top of the device, which moves up and down to physically change the length of the optical cavity.

VCSELs are typically very low powered devices, producing outputs in the region of 1 mW – considerably less than the 10 or 20 mW specified for long-haul networks. However, the devices may find a niche in metro networks, which demand lower powers.

Vendors

Bandwidth9 Inc.

Bandwidth9 has developed a monolithic structure for its laser, meaning that the entire device can be produced in a single epitaxial growth process. Furthermore, the device is electrically, as opposed to optically, pumped, so it doesn't require a second laser. Both factors simplify manufacturing, which helps keep costs down.

Tuning is via a cantilever structure on top of the device.

Bandwidth9's most recent product integrates an "amplified modulator" – a modulator that also delivers gain – from CyOptics Inc. (see Bandwidth9 Picks CyOptics Modulator). The company claims that this allows the device to reach 600 km at 2.5 Gbit/s without any dispersion compensation – despite the fact that it's still a low-power device.

Related stories in Light Reading:

  • Bandwidth9 Behind Schedule?

  • Bandwidth9 Cuddles Up With Corning

  • Bandwidth9 Scores a Coup

  • Bandwidth9 Claims Laser Breakthrough

BeamExpress Inc.

In late 2002, Swiss startup Beam Express revealed its plans to make optically pumped tunable VCSELs and fixed-wavelength VCSEL arrays (see BeamExpress Tunes VCSEL). Its technology hinges on a process called wafer-bonding to fix different kinds of material together.

Honorable Mention

Coretek

Coretek, which produced an optically pumped tunable laser, was bought by Nortel Networks Corp. (NYSE/Toronto: NT) for the monster sum of $1.43 billion. Those with knowledge of the product say it was a good one, but that hasn't helped the technology survive. When Nortel decided to quit the optical components business, it ended up selling the majority of its components business to Bookham Technology plc (Nasdaq: BKHM; London: BHM). The sale excluded the former Coretek, whose technology appears to have been buried – possibly on purpose: Multiple sources suggest that part of the deal with Bookham involved keeping Coretek's tunable laser technology (which would compete with Bookham's) out of the market.

Related stories in Light Reading:

  • Bookham Buys Nortel's Components Biz

  • Coretek Is Closed

  • Nortel Gambles $1.43 Billion On Tunable Lasers

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like