MULTIWAVE: Cost-effective MULTI-WAVElength Laser System
[November 2005 – October 2007]
[November 2005 – October 2007]
The MultiWave project has demonstrated a multi-wavelength platform capable of generating source signals with channel spacing in the range of 12.5G, 25G, 50G, 100G or higher on the ITU grid, and covering the S, C, and L band. Error-free operation of the modulated channels in the C-band was obtained with performance equal to or better than commercial DFB lasers. MultiWave has demonstrated a cost-effective platform for upgrading present and future broadband fiber optic links.
Summary of project context and objectives
The deployment of wavelength-division-multiplexed (WDM) systems has allowed for unparalleled network upgrading in network capacity and transmission lengths. As WDM technology advances towards cost-sensitive Metropolitan Area Networks (MAN) and even Access Networks, a major problem identified is the high complexity and cost of WDM transmitters. WDM test and network architectures currently rely on large banks of continuous-wave lasers or DFB lasers, which are often tunable in wavelength. Each single laser source acts as an optical source for a single wavelength (channel), requiring its own drive electronics and current/temperature controlling. Aside from the high initial cost of this approach, upgrading such high capacity WDM network means adding a laser source for each additional channel required leading to unacceptable installation cost.
The MultiWave laser source can replace complex and costly dense-wavelength-division-multiplexed (DWDM) transmitters with a single, cost-effective optical source. The developed laser source is capable of producing a large number of channels from a single system, reducing the complexity of the transmitter architecture and relaxing the demands on power, heat, inventory and space of optical network terminals, thus reducing the cost per channel of future DWDM systems. This single source system can be used as a high-channel-count, multi-wavelength optical source with a potential to cover all optical telecommunication windows.
The approach in MultiWave was based on a passively mode-locked pulse-generating laser with consecutive passive spectral broadening and subsequent passive channel selection through spectral filtering. The MultiWave laser system consists of three fundamental building blocks (see 1 & 2): the initial pulse generating laser source (2), the creation of ultra-broad spectrum (1) and the channel spacing selection stages (3). The most apparent advantage is that all components work mainly on a passive basis, therefore the need for control and monitor electronics is minimized, as well as power consumption, generated heat, space requirements and cost. There is only simple low frequency electronics necessary, which is mostly already available off the shelf. All the components are compact, integratable on a single motherboard and easy to assemble.
MultiWave had two main goals:
- Demonstration and evaluation of each fundamental building block of the MultiWave laser MultiWave consortium has performed the R&D work required for the design and development of the laser source building blocks: a turn-key passively mode-locked Er:Yb:glass lasers operating at 12.5 GHz and 25 GHz, Photonic Crystal Fibers operating successfully across S-, C-, and L-band and having optimized connector solutions, and channel spacing upgrade modules up to 400 GHz using efficient spectral selection.
- Integration of all the required building blocks, and development – demonstration – characterization of the turn-key single Multi-wavelength laser system locking to the ITU-grid. channel spacing: GHz with options for 25 GHz, 50 GHz and 100 GHz. Bandwidth: C-band with options for S- and L-bands.
Main Scientific and technical (S&T) results and foregrounds
The MULTWAVE workplan was divided into 6 workpackages. Workpackages 1 – 4 deal with the development of the individual key components and core know-how on which the proposed Multi- wavelength laser system is based on. Workpackage 5 dealt with the integration of the key components into a single platform and evaluates the performance. Workpackage 6 covered consortium management and assessment of progress and results related tasks.
The main achievements per workpackage are summarized in the following:
Workpackage 1: PGL Fundamentals
The main objectives of WP1 were to understand pulse shaping effects in multi-gigahertz Er:Yb:glass lasers, to demonstrate short pulse laser operation, to investigate wavelength tuning behavior of Er:Yb:glass lasers and to provide optimized laser design for full C-band tuning and to increase the pulse repetition rate of Er:Yb:lasers.
All requirements for the MultiWave system were achieved:
- optical signal to noise ratio OSNR >40 dB (target >30 dB)
- spectral linewidth <100 kHz (target <1 MHz)
- relative intensity noise RIN <0.1%, resp. 0.6% (target < 1%)
- long-term spectral stability <5 pm shift during 18 hours (target <20 pm shift during one hour)
All deliverables and milestones of WP1 were achieved on time. The fundamental physical properties of multi-gigahertz passively mode-locked Er:Yb:glass lasers were successfully analyzed, and new sources with state-of-the-art performance were experimentally realized, leading to multiple scientific publications and optimized laser sources for the MultiWave system. The generated knowledge (e.g. simulation results, new laser design for higher repetition rates, SESAM development) was transferred from CO1-ETHZ to CR3-TBP for commercialization.
Workpackage 2: ERGO development
The main objectives of WP2 were to engineer OEM style packages of 12.5 GHz and 25 GHz ERGO lasers, to demonstrate efficiently working OEM style packages of 12.5 GHz and 25 GHz ERGO lasers, to demonstrate an in-house low profile EDFA in order to increase the laser average output power in excess of 20 dBm, to perform CAD-studies for 3.125 GHz, 6.25 GHz, 50 GHz and 100 GHz ERGO prototypes, to demonstrate wavelocking of the 12.5 GHz ERGO, and to test the ERGO’s lifetime and temperature cycling performance.
All goals were achieved. Lasers at 12.5 GHz and 25 GHz were designed, constructed, and used by the project partners in the full MultiWave system. These products have further been commercialized by TBP. Additionally, TBP was able to demonstrate wavelength stability approximately 10 times better than current typical commercial DFB laser specifications. Initial lifetime measurements of >1500 hours with no measurable power degradation, and temperature cycling from 10 deg C to 40 deg C with less than 5% power fluctuations were achieved.
Turn-key OEM style 12.5 GHz ERGO laser
Modified 10 GHz cavity
The first part of this deliverable, the 12.5 GHz ERGO laser, was completed in May 2006 and shipped to the NTUA group for use in their system experiments. This first unit used a modified 10 GHz design which allowed us to decrease the cavity size while maintaining suitable laser mode sizes in the laser gain material and on the SESAM device, such that we could achieve the repetition rate of 12.5 GHz. Since the laser repetition rate is inversely proportional to the laser cavity round trip time, the cavity for the 10 GHz laser (nominally 15 mm in free space) had to be reduced by 20% to approximately 12 mm. The key challenge to achieve this was the constraints on the positions of the other optical elements in the ERGO housing, and on the available curvatures of the laser mirrors. We were able to achieve this with an output coupler mirror that we positioned as close as possible to the laser gain element. In this case, our key challenge at the end was the mechanical interference of the two components i.e. the output coupler to the laser glass component. For example, during the first attempts to construct the laser and operate it at 12.5 GHz, we experienced mechanical interference between the output coupler mirror and the laser crystal holder, i.e. these two parts ran into each other. Careful mechanical modifications of the laser housing components allowed us to overcome this difficulty.
We discovered an additional problem with the modified 12.5 GHz ERGO, which was identified in QMR #2 – specifically, that outgassing of glue inside the PGL was identified to cause a slow degradation (over hours or days) of the laser output power. We were able to solve this problem by a combination of careful glue placement, plus better controlled curing of the glue in an oven. Additionally, a test stand for outgassing properties of the components was developed, which will allow us to more systematically test individual components for outgassing.
Figure 3 below shows an overview of the internal mechanical configuration of the 12.5GHz OEM ERGO.
Figure 4 shows the internal mechanical view of the ERGO with the 25 GHz laser head and PGL block. Most of the parts, except for the PGL block, are the same as for the 12.5 GHz, and the two PGL heads shown separately to the right compare a 25 GHz PGL block to a 10 GHz/12.5 GHz PGL block.
Building the 25 GHz laser clearly revealed an important potential limitation: while building a 25 GHz laser still works fine with the current mechanical design, a further increase of the repetition rate up to 50 GHz or ultimately 100 GHz will be hard to reach with this approach. This is mainly due to tolerances of the mechanical parts used in the assembly and to the forces applied to the parts when locking them down in their final position. Additionally, there are always mechanical moments (torque) which occur when the parts are locked down with screws, forcing the parts to move out of the desired position. Although this movement is typically very small (few microns to 10’s of microns) it results in the laser moving outside of it optimum performance range, and this becomes simply more critical due to the smaller tolerances for mechanically smaller higher repetition rate cavities. We believe that a new mechanical approach has to be developed for repetition rates substantially higher than 25 GHz.
To achieve a low-profile EDFA, we chose a highly-doped erbium-doped fiber, the MetroGain M12 fiber from Fibercore. Simulations showed that we would achieve >100 mW output power with 5 meters or less of fiber and the available 300 mW of pump power. Figure 5 following shows the initial mechanical packaging of this device using an existing ERGO laser housing and pump diode electronics.
Workpackage 3: PCF Development
The main objectives of WP3 were to realize first PCFs for supercontinuum generation, to develop a low-loss connectorization solution for the first generation PCFs, to determining best-suited design route for second generation of fibers, and to connectorize suitable amount of first generation fiber to be tested experimentally in a supercontinuum setup.
All goals were achieved. The designed fibers were fabricated with emphasis on reducing the fiber attenuation and improve the structural control. Production steps and methods were analyzed and optimized with respect to general contamination and OH contents. Fiber drawing conditions was optimized to improve the structural stability and uniformity leading to better dispersion control. All fabricated pieces, including nonlinear PCFs and contamination test fibers were optically characterized.
A new method was developed for interfacing PCF to standard transmission fiber including both non- PM and PM fiber types, thus allowing for efficient splicing and connectorizing with reduced losses.
The results of the work on PM-PCF will be commercialized by CF.
Workpackage 4: Development of Channel Spacing Upgrade Modules
The main objectives of WP4 were to theoretically investigate the channel spacing upgrade up to 400 GHz through spectral selection, to design x4 and x8 channel spacing upgrade modules by selecting appropriate FP filter parameters for high channel isolation, to fabricate bulk FP etalons for x4 channel spacing upgrade reaching 50 GHz and 100 GHz WDM channel separation, and finally to demonstrate and evaluate the performance of two complete (x4) channel spacing upgrade modules for 50 GHz and 100 GHz from direct input repetition laser harmonics at 12.5 GHz and 25 GHz, respectively. Further objectives included fabrication of bulk FP etalons for WDM channel separation with target spacing of 200 GHz and 400 GHz and to implement the channel spacing upgrade subsystems for x4 and x8 upgrade factors with target channel spacing equal to 100 GHz as well as 200 GHz and 400 GHz respectively. Finally, the ultimate objective of WP4 was the demonstration and full performance evaluation of the 100 GHz, 200 GHz and 400 GHz channel spacing upgrade subsystems.
All project goals and milestones were achieved.
The FSR 50GHz and 100GHz etalons supplied by SLS to NTUA were constructed using the following components:
Mirror substrates (pair):
10.0mm diameter x 3.0mm centre thickness synthetic fused silica
Back surfaces wedged 30 arc minutes (wedges opposed in the finished etalon, to maintain incident beam direction)
Coatings (deposited over the central 5mm diameter of the substrate):
Front surface coating: 96%R at 1550nm / 0deg (soft coating based on zinc sulphide and cryolite)
Back surface coating AR 1530nm-1570nm / 0deg (hard coating based on zirconium dioxide and silicon dioxide)
Zerodur Class 0, 1.50205mm thick (used doubled for FSR 50GHz)
The spacers were optically contacted to the front surfaces of the two mirrors (at the uncoated periphery), so that the mirror coatings faced each other (separated by the air gap).
Open-ended (ring-shaped) aluminum cell for maximum flexibility in subsequent alignment at NTUA
Work-package 5: Demonstration of a Multi-wavelength laser system
The main objectives of WP5 were to perform initial characterization of supercontinuum generation (SCG) in Photonic Crystal Fibers (PCFs) at 12.5 GHz and to perform initial experimental evaluations of the SCG effect in PCFs, using the 12.5 GHz mode-locked ERGO laser, then to optimize and fully characterize supercontinuum generation (SCG) in highly nonlinear PCF at 12.5 GHz as well as to implement and fully characterize SCG at 25 GHz, using the 12.5 GHz and 25 GHz ERGO laser respectively. Moreover, WP5 targeted the implementation of the Multi Wavelength laser systems with channel spacing of 50 GHz, 100 GHz, 200 GHz and 400 GHz, by combining the SCG subsystem with the channel spacing upgrade subsystems that were developed in WP4.
All project goals and milestones were achieved.
In order to assess if the developed multi-wavelength sources are suitable for use as transmitters in a DWDM system, the generated CW channels from the 50 GHz and 100 GHz MWS were modulated at 10 Gb/s and their performance was evaluated through eye and BER diagrams. The optimized 50 GHz MWS achieved improved performance compared to the preliminary experiment that was carried out during the first reporting period. Increasing the channel spacing to 100 GHz yielded equal performance. Comparisons were made with standard DFB lasers and the MultiWave system showed no performance degradation compared to standard DFB lasers. These results are summarized in Figures 1 & 2 of this summary.
In the final stages of the project, ICCS/NTUA undertook the system packaging of the complete MultiWave system. The specific approach involved the packaging of all optical components in small modules that can be stacked together. The pulse-generating laser is packaged into one module and fiber pigtails are used for the interconnection to the second module. This module is stacked on top and contains a high-power (30 dBm) erbium-doped fiber amplifier gain module. The output of the second module is inserted into an identical package that contains the PCF and FP filter. A series of experiments were carried out using the complete and packaged MultiWave system and results showed similar performance with previous experiments from WP4 and WP5.
Moreover, we successfully demonstrated wavelength-locking of a 100 GHz Multi-Wavelength Source based on a 12.5 GHz ERGO laser, with a very high accuracy of ±1.43 pm peak to peak and 0.67 pm rms, measured with an optical spectrum analyzer. These results exceed the wavelength stability requirements of standard off-the-shelf DFB laser systems.
25 GHz Multi-Wavelength Source
Amplitude variation reduction can be achieved by further optimizing the PCF and/or by increasing the launched optical peak power, which will enhance the effect of the nonlinearity. To overcome this limitation, a high-power EDFA was developed, that achieved output power equal to 30 dBm.
50 GHz multi-wavelength operation by x4 times spectral selection and SCG
The specific activity involves the integration of the 12.5GHz mode-locked ERGO laser (pulse generating laser), the super-continuum generation (SCG) module (broad spectrum creation stage) and the x4 channel spacing upgrade module. These building blocks of the MultiWave source were presented previously in WP2, WP3 and WP4 section, respectively. CR3-TBP and CR5-CF has delivered to CR2-ICCS/NTUA the ERGO laser and spools of PCF, respectively, for the purposes of this demonstration.
Multi-wavelength transmitter application of the MultiWave source
To assess whether the generated CW channels are appropriate for use as WDM sources, each channel was modulated in a LiNbO3 modulator at 10 Gb/s and eye diagrams and bit-error-rates (BER) were measured. This work improves the performance of the 50 GHz MWS that was demonstrated in the first reporting period and also extends this work to the 100 GHz spaced MWS.
The 12.5 GHz ERGO laser was used as the seed laser, followed by 125 m of highly nonlinear PCF as well as a x4 and a x8 channel spacing upgrade subsystem respectively (Figure 10). The 12.5 GHz ERGO was selected in order to prove the upgradeability of the MultiWave source and to investigate possible performance variations depending on the channel spacing upgrade factor. Each channel was modulated in a LiNbO3 modulator at 10 Gb/s with a 231-1 PRBS data pattern, and eye diagrams and bit-error-rates (BER) were measured. Figure 11 (e) and (f) show the obtained results for the 100 GHz and 50 GHz grid respectively. Selected channels are presented in a 34 nm range, which practically covers the C-band. The results show similar performance between adjacent and distant channels. The BER also shows that even two adjacent channels perform equally well. A commercially available DFB laser was also employed as the CW source and showed similar BER performance, proving the high quality of each line generated by our multi-wavelength source.
Packaging of the MultiWave system
In the final stages of the project, ICCS/NTUA undertook the system packaging of the complete MultiWave system. The specific approach involved the packaging of all optical components in small modules that can be stacked together. The pulse-generating laser is packaged into one module and fiber pigtails are used for the interconnection to the second module. This module is stacked on top and contains a high-power (30 dBm) erbium-doped fiber amplifier gain module. The output of the second module is inserted into an identical package that contains the PCF and FP filter. ICCS/NTUA implemented a single-stage high-power EDFA with output power 30 dBm in an open board. The PCF spools and the fiber-based Fabry-Perot filter were packaged successfully in small aluminium housing provided by TBP. However, the high-power EDFA could not be packaged in such a housing due to additional heat requirements of the EDFA and more specific the power dissipation of the high-power multi-mode pump used. It is expected that through further development and optimization of the housing, the heat issues can be solved. The new housing can be altered by incorporating an inverted heat sink together with a small fan to provide forced air cooling to the high-power pump. Concerning the MultiWave system, a commercially-available EDFA was used. A series of experiments were carried out using the complete and packaged MultiWave system and results showed similar performance with previous experiments from WP4 and WP5.
The aim of MultiWave was to develop an enabling key component, allowing network operators to meet current and future demands in bandwidth for much lower cost than with today’s state-of-the art technology. The Multi-wavelength laser system, developed within the frame of MultiWave, replaced banks of single wavelength lasers (100 lasers or more) with a single plug-and-play device, therefore reducing the cost per channel of future DWDM systems, the complexity of the DWDM system architecture and relaxing the demands on power budget, inventory and space requirements.
MultiWave provided the telecommunication industry a versatile tool for a broad range of applications. The most evident application with the highest expectable revenue was found in the DWDM network market. Network operators had to upgrade the transmission capacities of their optical fiber networks in the coming years and therefore look out for a cost-effective upgrade solution. Analysis of the market requirements and cost structure indicated that the MultiWave system started to provide competitive economic benefits for a full C-band system with 50 channels at 50 GHz channel spacing, but provided substantial and compelling benefits and cost reductions for systems covering both C & L bands with 50 GHz channel spacing, or for C-band systems with 25 GHz channel spacing.
Due to the modular setup of the within the frame of MultiWave developed component, upgrading, maintenance and redundancy costs for the end-user are very low. This single laser device could potentially be used as a high-channel-count, multi-wavelength optical source with a potential to cover all optical telecommunication windows, as MultiWave was conceived and designed giving highest priority to cost-effectiveness and upgradeability.
Dissemination and Exploitation Activities
Academic Activities and Publications
MultiWave assisted in the education of 8 Ph.D. students specializing in laser and photonics, and resulted in a number of conference and scientific publications:
- C. Zeller, R. Grange, T. Südmeyer, U. Keller, K. J. Weingarten, “77-GHz pulse train at 1.5 µm directly generated by a passively mode-locked high repetition rate Er:Yb:glass laser”, European conference on optical communication (ECOC 2006), Cannes, France, Sept. 2-28, 2006
- Paraskevas Bakopoulos, Efstratios Kehayas, Andreas E. H. Oehler, Thomas Sudmeyer, Kurt
- Weingarten, Kim P. Hansen, Christos Bintjas, Ursula Keller, Hercules Avramopoulos, Talk OWJ2, “Multi-wavelength laser source for dense wavelength division multiplexing networks”, Optical Fiber Communication Conference 2007 (OFC ’07), Anaheim, California, USA, March 25-29, 2007
- Bakopoulos, E. Kehayas, A. E. H. Oehler, T. Sudmeyer, K. J. Weingarten, K. P. Hansen, C. Bintjas, U. Keller, H. Avramopoulos, “Agile and Upgradeable Multi-wavelength Source for Dense Wavelength Division Multiplexing Networks”, Lasers and Electro-Optics Society Annual Meeting, LEOS 2007, 21-25 Oct. 2007.
- C. Zeller, T. Südmeyer, U. Keller, K. J. Weingarten, “77-GHz pulse train at 1.5 μm directly generated by a passively mode-locked high repetition rate Er:Yb:glass laser”, Talk CTuC2, Conference on Lasers and Electro-Optics (CLEO ’07), Baltimore, USA, May 8-10, 2007
- E. H. Oehler, S. C. Zeller, T. Südmeyer, U. Keller, K. J. Weingarten, “Moving towards 100 GHz from a passively modelocked Er:Yb:glass laser at 1.5 μm”, Talk CI6-1-THU, CLEO Europe 2007, Munich, Germany, June 17-22, 2007
- C. Zeller, T. Südmeyer, K. J. Weingarten, U. Keller, “Passively mode-locked 77-GHz Er:Yb:glass laser”, Electron. Lett., vol. 43, issue 1, pp. 32-33, 2007
- Schlatter, S. C. Zeller, R. Paschotta, U. Keller, “Simultaneous measurement of the phase noise on all optical modes of a mode-locked laser”, Appl. Phys. B, vol. 88, pp. 385-391, 2007
- J. H. C. Maas, B. Rudin, A.-R. Bellancourt, D. Iwaniuk, T. Südmeyer, and U. Keller, “High Precision Optical Characterization of Semiconductor Saturable Absorber Mirrors (SESAMs)”, submitted to Conference on Lasers and Electro-Optics CLEO 2008
Additionally, a follow-on program also called MultiWave has been funded at the Swiss Federal Institute of Technology for partner CO1-ETH. This will continue to support further enhancement to key MultiWave components, exploitation of the complete system for research applications and include the activities of 2 Ph.D. students for a further 2 years.
Major conferences and exhibitions attended where MultiWave results were disseminated:
MultiWave results were promoted at more than 15 international trade shows, of which the major ones include: OFC 2007, ECOC 2006, ECOC 2007, Laser 2007, CLEO 2007 and Photonics West 2007
Commercial Products derived from MultiWave:
CR3-TBP offers a new product line of stable high repetition rate lasers (the ERGO-XG) with 10 GHz, 12.5 GHz and 25 GHz. Reliable operation was achieved and verified by temperature cycling and long term operation during the MultiWave program.
- CR4-PL offers a new product line of high quality, low loss laser mirrors with radius of curvatures below 1 mm. The manufacturing yield was strongly improved by new production methods, increased cleanliness in the handing, and optimized coating New inspection methods were implemented for quality control.
- CR5-CF has commercialized the dispersion flattened, highly nonlinear fiber for telecom applications. The recently developed highly nonlinear, polarization-maintaining fiber will be commercialized
- CR6-SLS offers a new product line of high quality etalons with precisely defined free spectral range
MultiWave follow-on program
The excellent collaboration between academic partners and SME partners will continue in the future. Sufficient manpower on the academic side is assured by a Swiss follow-up research project that we also called MultiWave (grant TH-04 07-3, duration February 2008 February 2010).
The project targets three main topics of research
- Further enhancement of passively modelocked solid-state lasers and demonstration of several state-of-the-art laser systems in terms of repetition rate and noise
- The supercontinuum generation at high repetition rate will be investigated more in detail, and sources based on new designs, such as the combination of adiabatic compression stages in standard highly nonlinear fibers and dispersion flattened photonic crystal fibers, will be investigated.
- Methods for advanced noise characterization, which were developed in the MultiWave project, will be applied to the modelocked laser sources and the supercontinuum sources, and noise mechanisms in solid-state lasers and supercontinuum generation will be
The academic partners will continue to benefit from the state-of-the-art MultiWave source that will be made available to them via the SME partners. The SME partners will benefit from the further characterization and development of the system.
The Project’s Public web-site: (http://www.multiwave.eu.com/)
The official website of the MULTIWAVE project (http:// www.multiwave.eu.com /) contains all the updated information regarding the concept, the consortium, the objectives and the publications providing to the viewer an easy way to explore the aforementioned topics through a friendly interface.