Meanwhile, back at the lab....

Th Edex1
Th Edex1



Under the ACTS umbrella, European companies and universities are pushing the boundaries of fiber-optic technology.

The countries of Europe have spent the better part of the last 50 years seeking ways to work together toward common goals. While debates continue to simmer over the merits of such organizations as the European Economic Community or of individual programs such as the establishment of a common currency, the overall desire to cooperate remains intact.

This spirit of joint endeavor extends to the realm of photonics research, most recently within the Advanced Communications Technologies and Services (ACTS) program. More than 30 research efforts have begun within the ACTS Photonic Domain. With a combined emphasis on basic research and practical experimentation and exploitation, the work now underway within the ACTS program promises to significantly influence the shape of future optical networks. The two programs examined in this article provide a snapshot of how research in Europe will not only benefit communications on the Continent, but around the world as well.

The Table shows the current roster of research projects within the Photonic Domain. Several of these programs have run their course or pollinated other programs. For example, the Pan-European Lightwave Core and Access Network (PELICAN) program builds on work already completed during the Optical Pan-European Network (OPEN), Keys to Optical Packet Switching (KEOPS), Photonic Local Access Network (PLANET), and Management of Photonic Systems and Networks (MEPHISTO) efforts.Th 1099 Page101 Table1

Generally, project partners come from both industry and academia and from several different countries. In keeping with the emphasis on demonstrations and practical applications, carriers have joined several efforts.

The newest programs within the Photonic Domain are part of what is known as the Third Call. These projects include Demonstrating the Evolution of a Metropolitan Optical Network (DEMON), Regeneration of Pulse Shape Amplitude and Timing for Optical Signals and Validation on Testbeds (REPEAT), Large Optical Integrated Switches (LOIS), PELICAN, Application and Control of Tunable Lasers in High-Density Wavelength Division Multiplexing Networks (ACTUAL), Advanced Photonic Experimental Crossconnect (APEX), Horizontal Action on Dissemination of Results in Photonic Domain (horizon-2), Evolutionary Optical Approach for Intersatellite Space Communication Systems (OSC), Photonic Routing of Interactive Services for Mobile Applications (prisma), and Switchless Optical Network for Advanced Transport Architecture (sonata).

As Lightwave has previously given extensive coverage of many of the optical-networking and multiplexing and transport programs, the focus here will be on two programs in the subsystems and components arena: the Customer Access Photonics-An Integral Technology for Low-Cost Devices (CAPITAL) and ACTUAL.

Silica-on-silicon has become one of the most popular methods of creating photonic chips. Basically, these chips consist of multiple layers of silica-based glass deposited on silicon substrates. Photolithography is used to etch waveguides into the silica.

There are a variety of methods for constructing these chips, including flame hydrolysis deposition (see related article on page 30). However, the CAPITAL program investigated the use of a new technique in which a sol-gel is spun-coated onto the silicon substrate. The sol-gel includes liquid metallorganic "precursors" of the target glass composition (e.g., tetra-ethyl silane for SiO2) suspended in solution. Technicians can vary the ratio of these precursors to create the composition and index required. Reaction of these precursors in turn forms suspended particles of the glass. The suspension is spun to a thin layer on the substrate; as the solvent evaporates, a gelatin-like substance is created. Heating the gel reduces residual organics and hardens the material. Repetition of this process produces layers of the desired thickness and composition.

In practice, chip designers can create a "bilayer" that has a buffer of lower index and a core layer of greater index. This core layer acts as the guiding layer and may be doped with other materials as desired. The use of photolithography and reactive ion etching creates channels that can be reflowed to improve the shape and surface quality. Addition of a top-cladding glass completes the process.

Alternatively, light propagating through the core can be guided by shallow ridges on the upper cladding layer. This design avoids etching and reflow of the active layer and is referred to as "strip-loading."

Partners in the program included Imperial College in the United Kingdom; Alenia UN Azienda Finmecanica, Consiglio Nazionale delle Ricerche IROE, and Universita Degli Studi di Padova of Italy; GeeO, Thomson-CSF LCR, and Institut National Polytechnique de Grenoble of France; and INESC of Portugal. The teammates focused on developing an efficient fabrication process for the development and manufacture of photonic chips based on sol-gel technology and looked at new host glasses and at doping the sol-gel with nanocrystals containing rare earth. The hope, according to Imperial College's Dr. Eric Yeatman, was to cap the program by creating photonic chips for erbium-doped amplifiers.

For the most part, the program was a success when it wrapped up early this year. GeeO installed fabrication equipment that has demonstrated its efficiency in manufacturing photonic chips. The equipment shows an ability to produce chips with uniform thickness (with variations less than 2%) and refractive index (with variations of less than 0.002), while providing orders-of-magnitude better yield rate than previous manual processes. The program reports deposition speed of better than 2 min per coating (approximately 6 microns per hour). GeeO is so encouraged by the equipment that it has started a company it calls Teem Photonics to exploit the technology.

Meanwhile, work progressed on developing glasses and dopants for amplifier applications. Because silicate glasses have not proven ideal for the short path lengths required in erbium-doped amplifiers, the program participants sought to develop other oxides and sulfides with low phonon energies. In particular, the team focused on nonsilicate oxide waveguides made from germanate glasses. The glasses have shown what program literature calls "promising characteristics," including high fluorescence lifetimes. Compositions included the crystallite acting as an enhanced matrix in which the active ion is a dopant, or a self-activated crystallite that has the rare earth as an intrinsic part. Researchers examined both praseodymium and erbium as dopants, with the latter showing high-fluorescence lifetimes at high concentrations when fabricated.

Despite the promise of these glasses and dopants, however, Dr. Yeatman reports that the team was not able to produce components that could create the 10-dB gain over short path lengths that erbium-doped integrated amplifiers would require. The CAPITAL team investigated strip-loaded channel waveguides based on silica-titania guides. The use of codopants achieved what the team considers "high quenching concentration" and "good spectroscopic performance." The use of aluminum codoping and processes to reduce titania segregation and OH contamination produced host material that demonstrated low loss and a fluorescence lifetime of 8 msec.

Yet, while relative gain of 1 dB/cm was obtained, the requisite net gain remained unreachable. Program participants discovered that erbium ions impede the equilibrium of the molecular structure during the gel and glass formation stages, causing nucleation of nanocrystallites and resegregation on a nanometer scale. As a result, 40% to 60% of the ions cannot be kept in inversion at reasonable pump levels when high doping levels are used, which quenches gain.

According to Dr. Yeatman, another few years of development will be necessary before the technology investigated by the CAPITAL program will be ready for commercialization. While GeeO and Teem look for ways to exploit the sol-gel fabrication equipment initially for passive components, the scientific and academic project partners will likely continue their research as well.

Meanwhile, Dr. Yeatman acknowledges that other researchers have taken the sol-gel process in other directions. In particular, Lumenon Innovative Lightwave Technology Inc. (Dorval, QC, Canada) has announced imminent commercialization of sol-gel components in partnership with Molex Inc. The principle difference between Lumenon's research and that of the CAPITAL project resides in the sol-gel, according to Dr. Yeatman. While the CAPITAL researchers used strictly inorganic material, the Lumenon process involves a combination of glass and polymers. The CAPITAL program focused on inorganic material because of its mechanical stability and rugged nature as well as the researchers' belief that rare earth dopants could more easily be added with this material, says Dr. Yeatman.

Participants in the two efforts were very much aware of each other, he says. In fact, Dr. Yeatman states, several members of the CAPITAL team now work for Lumenon.

Tunable lasers represent one of the hottest areas in component research. With the number of channels carried by dense wavelength-division multiplexing systems (DWDM) growing each year, service providers face the unenviable position of having to keep a large number of lasers in their inventories for maintenance and equipment repair. At the same time, the cost of DWDM equipment could be minimized if channel counts could increase without the addition of more laser sources.

Participants in the ACTUAL program, which follows up on work conducted during the BLISS project, are looking into the use of widely tunable lasers for DWDM networks. As part of this work, the program will develop control, management, and switching methods for tunable laser modules as well as explore network architectures that could optimally exploit such modules. These modules would contain not only the laser sources themselves, but also drive electronics and microprocessor control.

Two control methods are under investigation. The first consists of an EEPROM that contains a lookup table with such laser characteristics as power and wavelength. The researchers will emphasize control algorithms that require little characterization and very fast programming, along with automation of characterization and parameter extraction. The second method, feedback control, will use direct monitoring of wavelength, power, and other laser parameters during operation. The use of stable optical filters for wavelength monitoring will be an object of focus. A hybrid approach that combines the use of a lookup table with feedback control also will be evaluated.

A network demonstrator will put the modules and their control methods to the test. The researchers will pay particular attention to the power, singlemode stability, frequency stability, and tuning accuracy of the modules. Meanwhile, the demonstrator will also highlight the modules' ability to perform fast wavelength switching and direct intensity modulation, the latter to avoid the cost of an external modulator common to some of the tunable lasers that have been demonstrated at recent conferences.

The ACTUAL team comprises IMEC and the University of Gent in Belgium, Altitun AB of Sweden, Gayton Photonics Ltd. and GEC-Marconi Materials Technology of the United Kingdom, Japan's NTT, Telenor R&D of Norway, and the University College of Dublin, Ireland. According to Gayton's Jens Buus, technical administrator of the ACTUAL effort, the 22-month program is making "very good progress." The network demonstrator is under construction and software and lasers are being tested, he reports.

The ACTUAL program was expected to be a part of the ACTS stand at last month's European Conference on Optical Communications in Nice, France. Meanwhile, program participants have kept the research community updated on their progress via presentations at other conferences. For example, Bjorn Broberg, Pierre-Jean Rigole, Stefan Nilsson, Lars Andersson, and Markus Renlund of Altitun reported at this year's Conference on Optical Fiber Communications (OFC) on a grating coupler sampled reflector (GCSR) laser. The laser combines two common methods of tuning a distributed Bragg reflector (DBR) laser: codirectional couplers and the Vernier effect. In the GCSR laser, a modulated Bragg reflector provides a comb of peaks. A codirectional coupler selects any of these peaks; each peak can then be tuned in the same way as a DBR laser. The presenters claimed the GCSR has the function of approximately 10 DBR lasers with their tuning ranges interleaved to provide wavelength coverage over the tuning range; complete wavelength coverage extends beyond 60 nm, with a total tuning range exceeding 100 nm. A microprocessor provides control.

The wavelength switching properties of the GCSR laser appear to be dependent on the carrier dynamics in the tuning section, with a time constant equal to the spontaneous carrier lifetime on the order of a few nanoseconds. Switching time has been measured from 2 nsec for the neighboring wavelength to 27 nsec for a wavelength 38 nm distant. This performance is similar to DBR lasers.

The two program descriptions presented here represent a mere glimpse of the research activity now underway as part of the ACTS program. As previously stated, the Third Call launched a series of new efforts that will advance optical technology in each of the four areas highlighted in the table.

The key for the fiber-optic community at large will be the transition of ACTS technology into the field. This goal appears obtainable. Besides the efforts of GeeO, Alcatel of France has been particularly active in using ACTS field trials to demonstrate technology-particularly optical crossconnects-that appear certain to become part of the company's product portfolio in the near future. As the Third Call programs wend their way toward conclusion, it's hoped that another call will be issued to researchers across Europe to build on the advancements achieved by projects like CAPITAL and ACTUAL.

More in DWDM & ROADM