This paper describes the design and implementation of wide-band Raman amplifiers for fiber-optic telecommunications systems. All-Raman amplifiers permit 100nm wide systems over spans of over 1500km due to the low noise figure and reduced nonlinear system penalties. First, the enabling technologies for realizing 50-100nm transmission systems are reviewed. Then, the focus is on the key building block for wideband systems, namely the all-Raman optical amplifier. The advantages and challenges of all-Raman wideband amplifiers (WBA) are first reviewed. Then, Section 3 describes the physical principles and engineering design rules for construction of all-Raman WBAs that satisfy gain and noise figure performance requirements of typical long-haul and ultra-long-haul fiber-optic transmission systems. Exemplary wideband system experiments are described in Section 4, and then the paper is summarize and concluded in Section 5.

Bismuth oxide-based erbium-doped fiber amplifiers (Bi-EDFAs) are attractive because of their broad gain bandwidth and high gain per unit length. In this paper, several applications of broadband bismuth oxide-based erbium-doped fiber lasers (Bi-EDFLs) are reviewed including a wideband tunable single frequency Bi-EDFL, a femtosecond passively mode-locked Bi-EDFL, and applications to broadband spectrum generation.

An emission, which covers the spectral range from the O to L-band, was observed from a Ni²⁺-doped zinc-alumino-silicate glass at room temperature. The emission lifetime was about 440 msec at 5 K and decreased with an increase temperature to 240 msec at 300 K. The emission band was identified as the downward d-d ³T_2g(³F)→³A_2g(³F) transition of octahedral Ni²⁺ ions. It was confirmed from X-ray diffraction analyses and transmission electron microscope observations that ZnAl₂O₄ spinel nano-crystals containing octahedral Ni²⁺ ions acted as emission centers. It is expected that the Ni²⁺-doped glass has a potential for a broadband amplification medium.

Enhanced mode selection is facilitated by incorporating compact two dimensional photonic crystal gratings into Fabry-Perot lasers. The gratings are etched directly through the active layer and interact strongly with the guided mode. A high mode-selectivity 2D grating which is significantly wider than the ridge waveguide structure is investigated. The mode selectivity for the structure is -37 dB. The enhanced selectivity afforded by the broad 2D grating is attributable to a novel and robust photonic bandgap mechanism: transverse mode beating.

In this paper, we present self-collimation in three-dimensional (3D) photonic crystals (PhCs) that consist of a simple cubic structure. By exploiting the dispersive characteristics of the photonic crystals, we demonstrate the ability to achieve structureless (defect-free) confinement of light. We also verify that polarization dependence is a key issue in 3D self-collimation. The results hold promise for the high-density PhCs devices due to the lack of structural interaction. Finally, a novel method for the fabrication of three-dimensional (3D) simple cubic photonic crystal structures using conventional planar silicon micromachining technology is presented. It overcomes the disadvantages of the methods hitherto reported in the literature for the fabrication of 3D photonic crystal devices, which include high complexity of multi-step processes, tight alignment tolerances, long turnaround times, and incompatibility with an integrated photonics platform. The method utilizes a single planar etch mask coupled with time multiplexed sidewall passivating deep anisotropic reactive ion etching along with isotropic etch process to create three-dimensional photonic crystal devices. Initial experimental results are presented.

Modal solutions for photonic crystal fibers with circular air-holes in a hexagonal array are presented, by using a rigorous full-vectorial finite element-based approach. The effective indices, mode field profiles, spot-sizes, modal hybridness, modal birefringence and group velocity dispersion values are presented. The effect of external pressure on photonic crystal fibers is analyzed also by using the powerful finite element method. Modal solutions are obtained for both symmetric and asymmetric air-holes and the effect of pressure on the modal properties and mode degeneration are evaluated, presented and discussed.

In conventional optical fibers the light guiding and other optical properties are mainly controlled by the index profile as given by the composition of the doping materials. In this case technological and material properties limit the design flexibility of the optical properties. For example, it is not possible to shift the region of anomalous dispersion to wavelengths shorter than the zero dispersion value of silica. In contrast to these conventional fibers, for microstructured or photonic crystal fibers (PCF) the optical properties are determined by the size, distribution and geometry of the air holes running the entire length of the fiber. Variations of the geometrical parameters offer a wide flexibility in designing different optical properties. It is possible, for example, to modify numerical aperture and mode field area over wide ranges or to shift the zero dispersion wavelength down to the visible spectral region. In the case of hollow core fibers (photonic bandgap fibers) the interaction of the guided light with the glass material is reduced and allows the transmission of higher intensities with reduced nonlinear effects.

One-, two- or three-dimensional periodic structures with an omni-directional photonic bandgap are well known as "photonic crystals". For applications in new or improved functional components used for communications we currently investigated in more detail their spatial and chromatic dispersion. Especially the strong angular difference between phase and group velocity in the vicinity of an optical bandgap may be used in practical devices. These strongly frequency dependent beam-steering effects were demonstrated by one author decades before: they may now be interesting for spatial optical frequency discrimination and thus also for dispersion compensation or pulse shaping. In contrast to photonic crystal fibers 2D-planar "photonic crystal waveguides" offer the ability to use both the spatial and chromatic dispersion characteristics. Thus by changing the geometry of propagation adaptive dispersion management may be possible. Based on our former theoretical and experimental studies, we show actual simulation results for state-of-the art planar photonic circuits technology on the dispersive properties of one- and two-dimensional photonic crystal waveguides. Possible ways for applications of more complex 2D-photonic crystal waveguide devices in optical communications for use as adaptive dispersion compensators will be presented and discussed.

Simple fabrication techniques of polymeric optical waveguides and waveguide elements toward low insertion loss are demonstrated. Serially-grafted waveguides with active layer such as electro-optic polymer and passive layer such as transparent polymer was fabricated using softlithography as well as photo-bleaching. Single-mode waveguide was fabricated using light-induced self-written (LISW) optical waveguide.. For large core optical waveguide, replication technology is adopted for the formation of a platform with a fiber guide.

Polymer waveguides have attracted a great deal of attention for their potential applications as optical components in optical communications, optical interconnections and optical sensors because they are easy to manufacture at a low temperature, and they have a low processing cost. Hot embossing is powerful and effective tools to produce a large volume of waveguides and structure high-precision micro/nano patterns of thin polymer films using a stamp for optical applications. In this work, fabrication techniques of hot embossed polymeric optical waveguides for parallel optical interconnection module, multi-channel variable optical attenuator and optical printed circuit boards are demonstrated. The single- and multi-mode waveguides are produced by core filling and UV curing processes. New approaches to fabricating single-mode polymeric waveguides with the high thermal stability in thermosetting polymers and two-dimensional multi-mode polymeric waveguides for high-density parallel optical interconnections as well as a simultaneous fabrication of single-mode polymeric waveguides with micro pedestals for passive fiber alignment are also reported.

A 16-arrayed polymeric optical modulator is fabricated using an electro-optic (EO) polymer with a large EO coefficient and good thermal stability. The 16-arrayed modulator has lumped type electrodes with a response time of less than one nanosecond. The 16-arrayed modulator has good uniform modulation characteristics between the individual modulators. The deviation of half-wave voltages is 0.2 V and that of insertion losses about 1 dB. Crosstalks range from -28 to -36dB and extinction ratios are more than 21 dB.

We have developed several novel polymer materials for optical waveguides. In the first part of this paper, we described two examples of our developed materials, one is low loss optical waveguide materials for optical communications at near infrared wavelength region, the other is low loss and high heat resistant materials for optical interconnections in the electronic devices at around 850 nm wavelength region, which can resist to the lead-free solder reflowing process. In the second part, we demonstrated two kinds of novel low cost polymer coarse wavelength division multiplexing (CWDM) devices, one is the polymer arrayed waveguide grating (AWG) device and the other is the polymer CWDM device using the polymer diffraction grating. The structures and performances of these devices were described in detail.

All-Optical signal processing and routing avoid the optical-to-electrical and electrical-to-optical conversions and thus eliminates the bandwidth constrains set by conventional electronics technologies. It will form key building blocks for the next generation ultra-high speed and dynamic multi-casting optical networks. Polymer waveguide based photonics integrated circuits are promising in achieving high-performance optical signal processing circuits due to their large Electro-optic/thermo-optic coefficients and structure flexibility. In this paper, design examples such as dynamic channel equalizer, optical CDMA encoder and matched filter are analyzed. Simulation results are presented.

Novel special polymer optical fibres and devices have been developed by several research groups around the world. Special polymer optical fibres and devices have great potential in a range of photonic applications. These polymer optical fibres, including both single-mode and multimode, could be made laser-active, photosensitive, photorefractive or optically nonlinear, either by incorporating functional materials into the fibre cores or by fabricating specially structured fibre cross sections. In this presentation we will report on the progress in developing special polymer optical fibres and also discuss a few prospective device applications of these fibres.

Two-dimensional (2-D) multi-channel 8x8 optical interconnect and processor system were designed and developed using complementary metal-oxide-semiconductor (CMOS) driven 850-nm vertical-cavity surface-emitting laser (VCSEL) arrays and the photodetector (PD) arrays with corresponding wavelengths. We performed operation and bit-error-rate (BER) analysis on this free-space integrated 8x8 VCSEL optical interconnects driven by silicon-on-sapphire (SOS) circuits. Pseudo-random bit stream (PRBS) data sequence was used in operation of the interconnects. Eye diagrams were measured from individual channels and analyzed using a digital oscilloscope at data rates from 155 Mb/s to 1.5 Gb/s. Using a statistical model of Gaussian distribution for the random noise in the transmission, we developed a method to compute the BER instantaneously with the digital eye-diagrams. Direct measurements on this interconnects were also taken on a standard BER tester for verification. We found that the results of two methods were in the same order and within 50% accuracy. The integrated interconnects were investigated in an optoelectronic processing architecture of digital halftoning image processor. Error diffusion networks implemented by the inherently parallel nature of photonics promise to provide high quality digital halftoned images.

A new paradigm for an all-to-all optical interconnect is presented. It could be part of an interconnection fabric between parallel processing elements and the first level of the cache in a computer system. Parallel processing has traditionally aspired to improve performance of such systems. An optical interconnect raises a new possibility: obtain both improved performance and significant cost reduction with respect to standard serial computer system models.

Laser communication has an enormous potential to provide a secure, jam-resistant, low detection probability and high-bandwidth means of communmication to support multimedia, imagery, video, mapping and other command and control functions in battlefield environments. In this paper, we discuss the development of a complete numerical model for free-space laser - PamCom model. We examine results obtained in the preliminary study that validate the feasibility of a selected approach. We review the estimates for atmospheric transmission for various lasercom links at 1.55 μm, analysze spatial-temporal statistics of the scintillation mititigation techniques. In addition, we review results for the link performance analysis including the SNR and BER calculations and examine predictions for the irradiance probability density function (PDF) from various models. Finally, we discuss the composition and top level architecture of the Pam Com model, which will be usable for a wide variety of scenarios, involving terminals on aricrafts, satellites, and the ground, and will be designed for use in DoD and commercial lasercom system design, test, and evaluation.

The expected increase in space and terrestrial services that include two-way fixed, SATCOM, CATV and mobile wireless services require expanding the system capacity. This expansion has created an opportunity for the utilization of the demonstrated photonic transport systems in wireless networks. System demonstrations and architectural developments have been proposed for distribution of communication services over fiber. Termed Fiber Radio and Hybrid Fiber Wireless, these systems offer the potential to improve services and reduce base station costs through increased bandwidth and ease of installation. We have developed and demonstrated DWDM broadband photonic transport systems able to meet the requirements for IS-95 Personal Communications Services operating at 1.9 GHz and Broadband Wireless Internet operating over the band of 2.5 to 2.7 GHz. Each DWDM channel operates from 1 to 3 GHz transporting services up to 80 Km. Solutions are being sought for low cost transmitters to meet DWDM SATCOM system requirements include extending the transmission distance to over 100 Km with a bandwidth that exceeds multiple octaves. These new requirements put high performance demands on the optical components. We have developed high performance transmitters based on electro-absorption modulated lasers (EML) that can meet SATCOM requirements. We have shown that the EML is capable of providing the required CNR of 32 dB for satellite transmission in the band of 950 to 2150 MHz over a 100 Km distance. In addition, we are investigating a new modulation technique, Microwave Photonic Vector Modulation (MPVM), which has the potential for wideband transmission in DWDM systems.

Vertical cavity surface emitting lasers (VCSELs) have been extensively developed and are now key devices in local area networks based on multi-mode optical fibers. Long wavelength VCSELs are currently attracting much interest for use in single-mode fiber metropolitan area and wide area networks. Also, parallel data links including board-to-board interconnections with low threshold VCSEL arrays are also under development. Low threshold single-mode VCSEL arrays will enable us to realize parallel optical interconnects with low power consumption. We have developed highly strained GaInAs/GaAs QW VCSELs emitting at 1.1-1.2 mm band. Excellent temperature characteristics have been realized. We present long wavelength GaInAs VCSELs on GaAs substrates, enabling uncooled operation for high speed data transmission in single-mode fibers. Also, we will discuss a possibility of isolator-free operations of single-mode VCSELs. In addition, we demonstrated a single-mode multiple-wavelength VCSEL array on a patterned GaAs substrate for the wavelength engineering of VCSELs. The maximum lasing span of arrays is over 190 nm. Densely integrated multi-wavelength arrays are presented. Tunable micromachined VCSELs are also attracting much interest for WDM networking, because micromachined tunable VCSELs enable wide continuous tuning. We proposed and demonstrated a micromachined tunable vertical cavity with a strain control layer, which gives us novel functions including temperature insensitive operation, thermal wavelength tuning, and so on. We also propose and demonstrate injection-locked VCSELs for all-optical signal processing. Some results on optical inverters, optical bistable devices and optical regenerators will be reported. Toward other applications including optical storages and sensing, we will describe metal nano-aperture VCSELs for near-field optics.

We have developed highly reliable widely tunable module, whose performances were comparable with fixed-wavelength DFB laser module. To realize wide tunability, 12 l/4-shifted DFB laser array, S-bend waveguides, MMI coupler and an SOA were integrated on a chip. We could achieve 37nm tunability by controlling each chip temperature in the range of 5 to 45°C. High output fiber coupled power of 30mW and very uniform L-I curves out of 12 DFB lasers were achieved even at 50°C. Good quality of lasing spectrum was obtained. Side mode suppression ratio (SMSR) > 45dB. The well-suppressed reflection at chip front facet contributed to the lower noise characteristics, such as RIN < -140dB/Hz and linewidth < 4MHz. The shift of locked frequency was less than 0.4GHz as the case temperature varied from -5 to 75°C. Very small frequency shift was realized by controlling the temperature of locker part, independently. By optimizing TEC design, we could achieve low TEC power consumption less than 4W under Tcase=75°C and the end of life condition of SOA current. The new function by incorporating SOA was VOA. By changing the operating SOA current, we could vary output power from 1mW to 20mW, maintaining SMSR > 40dB, RIN < -135dB/Hz, linewidth < 4MHz. We also performed optical blocking > 40dB, when SOA current was turned off. We examined modules reliability under high temperature storage of 85°C. The change of output power was < ± 10%, and the shift of locked frequency was < ± 5pm after 2000 hours.

We derive the Schawlow-Townes linewidth of mode-locked lasers, and discuss how the timing jitter affects the linewidth in passively/actively mode-locked laser diodes (MLLD's). The linewidth of longitudinal modes of MLLD's is much broader than the Schawlow-Townes linewidth due to the linewidth enhancement factor of the gain medium, and its characteristics depend on whether the laser is mode-locked passively or actively. For the case of active mode-locking, the detuning of the modulation frequency from the natural cavity frequency broadens the linewidth. In the experiment, we measure the linewidth of 10-GHz external cavity MLLD's and investigate the correlation with the noise characteristics of timing jitter and amplitude noise of the pulse train. The experimental results agree well with theoretical predictions.

We have proved that short cavity length and large coupling coefficient structure can increase the relaxation oscillation frequency of directly modulated 1.3mm DFB-LDs for 10Gb/s operation. The grating with large coupling coefficient can be made of high refractive index InGaAsP material (=narrow band gap energy material), which has to be fabricated in the n-type InP cladding layer to prevent large hole and electron accumulation in the grating layer. By using this grating, 1.3mm AlGaInAs quarter lambda phase shifted DFB-LDs successfully revealed excellent eye diagrams with over 10% margin for OC-192 mask at the temperature range of 25-95 deg.C. Furthermore, we newly propose multi-phase shifted gating to 10Gb/s quarter lambda phase-shifted DFB-LDs to suppress the deterioration of eye diagrams due to non-uniform longitudinal distribution of light intensity and carrier density.

All-optical wavelength conversion is one of the key technologies in future wavelength division multiplexed (WDM) transparent photonic network. An electro-absorption modulator (EAM) is one of promising candidates for optical nonlinear element to realize optical wavelength conversion due that it has fast absorption recovery in comparison to slow gain recovery of a semiconductor optical amplifier while it has moderate conversion efficiency in comparison to silica fibers or electro-optic dielectric materials. In this paper, cross-absorption modulation (XAM) and cross-phase modulation (XPM) characteristics as well as absorption recovery time of EAM having multiple quantum well (MQW) absorption layer was quantitatively analyzed. Absorption recovery time less than 10 ps were achieved for the MQW EAM due to fast sweep-out of photogenerated carriers by the external electric field. The phase shift induced by 1.5 pJ pulse injection was about 0.2. All-optical 2R type wavelength converter using the MQW EAM in conjunction with delayed-interferometer was tested at bit-rate of 100 Gbit/s. Error-free conversion was successfully confirmed. The operable wavelength bandwidth was investigated at 40 Gbit/s. The bandwidth of 30 nm covering almost full C-band was confirmed. An electro-optical 3R type wavelength converter utilizing rf-driven EAM was also proposed and demonstrated. The intentionally added timing jitter onto the input signal was reduced with simultaneous contrast ratio improvement after the 3R wavelength conversion due to periodical gate closing by the rf-clock drive. These results show that EAM is suitable for optical wavelength conversion at bit-rate range of 40-100 Gbit/s with a reasonable conversion efficiency.

Considerable progress has been in the last decade in the fields of photonic networks and ultra-fast optics. The past few years has seen the widespread use of wavelength division multiplexing (WDM) to provide enormous point-to-point capacity in the backbone and metro area networks. Remarkable progress in electronics, in terms of both costs and performance speeds, has to some extent alleviated the 'electronic bottleneck'. Developments in fiber-optics such as novel fiber types and Raman amplification have opened up additional wavelength regions of operation resulting in great expansion of usable fiber bandwidth. There exist unique opportunities for ultrafast technologies - a subject of much interest in the last decade and reaching a point of maturity - to complement these advances and spark the next generation networks. In our talk, we will mention two networking environments very different from WDM - (1) optical time division multiplexing (OTDM) and (2) optical code division multiple access (OCDMA). We will look at the potential of both these scenarios for different applications, focusing in particular on the latter as an approach that provides maximum flexibility to utilize the immense bandwidth of the optical fiber. We will also describe various ultra-fast technologies that have been developed e.g. .high repetition rate pulsed lasers, ultra-fast optical switches, time delay elements etc. which have a direct relevance to both these types of networks.

Today, commercial fiber optical transmission systems are operated at a data rate of 10 or 40 Gbit/s per wavelength channel. The optical time division multiplexing (OTDM) technique allows for the investigation of the next generation TDM data rate of 160 Gbit/s while electrical signal processing at 160 Gbit/s is not yet available. In an OTDM transmission system, the key functions of the transmission system are realized using ultrafast all-optical or electro-optical network elements. In this paper, the recent progress in ultrafast technologies for 160 Gbit/s OTDM transmission systems is reviewed. The focus will be on techniques to realize optical pulse sources and ultrafast optical gates which operate on a time scale of a few picoseconds to a few hundred femtoseconds. They are the key elements of a 160 Gbit/s OTDM transmission system and are applied for various functions in the system. The paper focuses in particular on the application of ultrafast optical technologies for measurement purposes. In an optical communication system the quality of the data signal can be inferred from the optical eye diagram of the data signal. Presently, the measurement of optical eye diagrams at a data rate of 160 Gbit/s is inhibited by the limited bandwidth of photodetectors and electrical sampling oscilloscopes. By realizing an all-optical sampling system using the ultrafast technologies discussed before, the bandwidth of such an optical sampling system exceeds 400 GHz and allows for eye diagram measurements up to 320 Gbit/s.

Measuring the temporal electric field of light allows the optimization of optical sources for specific applications and allows a better understanding of the interaction of light with matter. In optical telecommunications, the need for temporal characterization is due to the impact of processes such as chromatic dispersion in optical fibers, nonlinear effects and amplification on the electric field of a data-encoded source. The characterization of these impairments is valuable for system design and operation. Two recent developments in the domain of high-speed characterization of optical signals in the telecommunication environment are reviewed. The spectrogram technique allows the measurement of the electric field of a periodic source, and can therefore be used for the characterization of optical sources, pulse carving and pulse compression setup, as well as the characterization of propagation in a medium, leading to determination of chromatic dispersion or nonlinear index. Linear optical sampling measures samples of the electric field of a source (i.e. amplitude and phase), and is therefore a generalization of a sampling oscilloscope. It therefore allows the measurement of eye diagrams and constellation diagrams of data-encoded optical sources, and is particularly useful to characterize phase-shift keyed sources and partially coherent sources.

All optical XOR functionality has been demonstrated experimentally using an integrated SOA-based Mach-Zehnder interferometer (SOA-MZI) at 20 and 40 Gb/s. The performance of the XOR results has been analyzed by solving the rate equation of the SOA numerically. The high-speed operation is limited by the carrier lifetime in the SOA. In order to solve the limitations imposed by carrier lifetime, a differential scheme for XOR operation has been experimentally investigated. This scheme is potentially capable of XOR operation to > 100 Gb/s.

To cope with the explosive growth of IP traffic, we must increase both the link capacity between nodes and the node throughput. These requirements have stimulated research on photonic networks that use optical technologies. Optical packet switching (OPS) is an attractive solution because it maximizes the use of the network bandwidth. The key functions in achieving such networks include synchronization, label processing, compression/decompression, regeneration, and buffering for high-speed asynchronous optical packets. However, it is impractical to implement such functions by using all-optical approaches. We have proposed a new optoelectronic system composed of a packet-by-packet optical clock-pulse generator (OCG), an all-optical serial-to-parallel converter (SPC), a photonic parallel-to-serial converter (PSC), and CMOS circuitry. The OCG provides a single optical pulse synchronized with the incoming packet, and the SPC carries out a parallel conversion of the incoming packet. The parallel converted data are processed in the smart CMOS circuit, and reconstructed into an optical packet by the photonic PSC. Our system makes it possible to carry out various functions for high-speed asynchronous optical packets. This paper reviews our recent work on high-speed optical packet processing technologies such as buffering, packet compression/decompression, and label swapping, which are key technologies for constructing future OPS networks.

Third-order optical nonlinearity in optical fibers has attractive applications to all-optical signal processing that will be employed in future photonic networks. After reviewing the third-order nonlinear optical property of optical fibers, we describe some examples of all-optical signal processing devices, focussing on self-phase modulation (SPM), cross-phase modulation (XPM), and four-wave mixing (FWM).

In recent years, optical CDMA systems have been proposed for multiple accesses to utilize the vast bandwidth available in optical fiber. Optical CDMA systems are believed to provide asynchronous access for each user in the system, which is especially suitable for usage in LAN. In this paper, we demonstrate a novel optical CDMA scheme in a fiber-based testbed. Using the liquid crystal spatial light modulator (SLM), we are able to construct a reconfigurable optical CDMA system suitable for fiber-optic networks. We address the code for each user in the spectrum domain by using a standard 4-f pulse shaping apparatus. Because of the low coherency of the light source we used in the system, we are able to modulate it in time domain without changing its frequency distribution significantly. We can reconfigure the network connection while keep the information bits un-influenced. Another merit of using analog liquid crystal device is that the transmissions of the different frequency components are analog controllable, we can get a uniform intensity distribution in frequency domain when the spectrum of the light source is not flat. Using the liquid crystal as a programmable optical modulator, the high polarization sensitivity of the components used in the system enables low crosstalk between different codes assigned to different users.

We report on several recently demonstrated microphotonic devices based in ferroelectric single crystals. The device concept is to micro-pattern ferroelectric domains in the shape of lenses and prisms whose refractive index is electric field tunable through the electro-optic effect. Two different devices have been fabricated and tested. The performance of an integrated lens and scanner device is demonstrated with electric field tunable focusing power and a scanning range of 14.6°. A 2-stage cascaded horn-shaped scanner is demonstrated with a scanning range of 25.4° at 5 kHz. Ferroelectric materials offer a unique platform for integrating several optical functions, like focusing, scanning, and frequency conversion, all in the same device.

In this paper, we demonstrate a technique for suppressing the intensity noise of the erbium doped fiber lasers. We show that by introducing negative feedback into the system through the use of a self-biased intracavity loss modulator, the relaxation oscillation noise of the fiber laser could be suppressed by as much as 20 dB. To explain the observed noise reduction phenomenon, a simple theory based on teh linearized rate equation has been developed and also been supported by experimental measurements. We also derive a formula for estimating the maximum achievalbe intesity noise suppression ratio.

Semiconductor optical amplifiers (SOAs) may be used for frequency conversion in cross gain modulation up to a few Gbps. Cross gain has the advantage that the output depends only on power in the incoming signal. At higher bit rates---10Gbps is frequently used in telecommunications--- the optical pulse shape degrades as though an R-C filter had been applied and bit error rates become unacceptable. The intensity variation in an SOA in cross gain is accompanied by a phase change which has the information for the lost part of the pulse. A discriminant filter following the SOA may be used to recover the lost bit signal intensity. The edge of the Bragg discriminant filter acts as a phase to intensity converter. We show by simulation that the bit error rate can be restored at 10 Gbps by the use of the filter. We further show how to optimize the filter frequency and bandwidth in such a way as to avoid unstable regions where precise control circuits would be required.

An arrayed-waveguide grating (AWG) is an attractive device for use in a WDM network because it offers multi-channel operation, design flexibility and suitability for mass production. AWGs fabricated using silica-based planar lightwave circuit (PLC) technologies are excellent in terms of loss and long-term stability, and so are widely used in commercial DWDM systems. This paper reviews recent progress on the AWG as regards improving its performance, especially with a view to eliminating polarization sensitivity. It also describes large-scale and compact AWGs formed using high-index contrast waveguides and fiber-connection methods using spot size converters. The second half of the paper reviews recent studies on functional devices incorporating thermo-optic (TO) switches and variable optical attenuators (VOAs).

Thin film dense wavelength division multiplexing (DWDM) filter is one of the most widely used filtration technologies that made possible the technical advancement of modern optical communication systems. A brief introduction to the thin film DWDM filter technology will be presented. The recent progress in thin film DWDM technology will be reviewed. Applications demonstrating the DWDM filter as an enabling technology for modern optical communication systems will be reviewed in detail.

A microring resonator, a circular waveguide adjacent to a straight waveguide, performs selective band-stop filtering for optical telecommunications. Small size and versatility of functionality allow many microrings to be incorporated into a single integrated optic circuit. We desire small size, less than 20 micron radius; low insertion loss in band-stop outside resonance; perfect cancellation to zero in band-stop at resonance; and low refractive index difference between core and cladding. These requirements are conflicting: conventional silicon waveguides cannot achieve low enough loss at the desired bend radius. Current approaches use different materials with higher refractive index differences. Analysis and simulation provide evaluation of the trade-offs. Further, we propose, for the first time, conventional waveguides using Bragg reflectors surrounding the microring to reduce loss for improved performance. Simulation shows feasibility of this approach.

Using the simple inverse Flourier transformation of the target channels, we can get the index modulation structure of the sampled period for the sampled fiber Bragg gratins (FBGs). In this method, the enable channels are identical wavelength operation while the unable channels are almost suppressed completely, and the enable and unable channels can be established at will based on the applications. However, the efficient of the sampled FBGs of which the index modulation is obtained by the simple inverse Fourier transformation is very slow. To get a high efficient, a particle swarm optimization algorithm is applied to design the phase of the target spectrum. Combing the two methods together, a high efficient multi-level phase sampled FBG that each channel is suppressed or not at will, can be obtained. We also use the particle swarms optimizations (PSO) algorithm to optimize the pure phase sampling profiles of the sampled FGB. Results showed that a much high efficient of the FBG can be obtained compared with that obtained from the simple inverse Fourier transformation technology. If appropriate grating period chirp and sampled period chirp is applied to such a grating, a novel FBGs based OADM or interleaver devices with dispersion or dispersion slope compensation can be designed.

The induced cladding-mode performance in a tilted superstructure fiber grating (SFG) is first experimentally demonstrated. As a slanted SFG is heated to shift wide-band loss dips for obtaining the multi-narrow-band reflection moved at various positions of the loss dip, the strength and central wavelength of the induced cladding-mode depend on the loss depth and the positive/negative sides of the dip, respectively. The characteristics may provide the third physical-parameter measurements in a simultaneous multi-parameters fiber sensor.

With the drastic expansion of internet usage, the demand of 10Gbps transmission optoelectronic devices for local-area-network (LAN) and storage-area-network (SAN) are increasing. The key issues of these applications are to improve cost, manufacturability and reliability of optoelectronic devices in high speed transmission. The authors have demonstrated highly manufacturable and reliable optical front end and trasneiver module for 10Gbps Ethernet applications in this paper. TO-Can package is a way to reduce cost of 10Gbps optical assembly. However, the signal integrity of high speed transmitter optical sub-assembly (TOSA) and receiver optical sub-assembly (ROSA) are limited due to the mismatched characteristic impedance, parasitic inductance and spread capacitance of conventional TO-Can package. In this paper, high performance and high sensitivity of 10Gbps TOSA and ROSA with TO-Can package are discussed and demonstrated to overcome the critical issues mentioned, respectively. In order to improve the signal integrity and manufacturability of 10Gbps OSA in 10Gigabit Ethernet small form factor transceiver module assembly. The authors also integrate high speed flex board and OSA package to extend the signal path, and to minimize the effect of crosstalk in module. The high speed transceiver modules with TO-Can package embedded compliant with XFP multi-source agreement (MSA) are also demonstrated in this paper. The performance of temperature-stabilized transceiver module over awide case temperature range is tested. The optical eye diagram of 10Gbps transmitter developed in this study shows an excellent quality passing the 10Gbps Ethernet mask test between 0 degrees Centigrade and to 70 degrees Centigrade. Finally, the reliability tests regulated by Telcordia GR-468-CORE and MIL-STD-883 are also performed and certified to pave the ways of highly manufacturable and reliable XFP transceiver modules for 10Gbps Ethernet applications.

We examine the effects of multiple incident pulses and determine that for closely spaced pulses adjacent pulse interactions lead to pulse coalescence resulting in increased peak energy and greater absorption. Furthermore, trailing pulses reach a reduced transmission level.

This paper describes a 12-channels parallel optical transmitter module with a MPO-Connector designed for a very short reach OC-192 and SNAP 12 specifications. It is important to design the micro array lens for better couple efficiency of array optical transmitter module. The authors design the high accuracy micro array lens for injection modeling to reduce the price and suit for further mass production. In this 12-channel parallel optical transmitter module, it is very difficult to posit the chip on correct position according to guide pin or guide pin hole. Therefore, the authors develop the method of two steps flip chip bonding to release the difficulty of chip alignment on ceramic substrate without two guide pin holes. The performance of the module is demonstrated to fulfill the requirements of SNAP 12[1]. The extinction ratio of the 12-channels array transmitter module is tested above 6dB, respectively. The optical shift by heat is an important factor affecting the performance of the array module. Thermal analysis of 12-channel parallel optical transmitter module is used to improve the effect of optical shift by heat in this paper. And the temperature among the case of transmitter module is greatly reduced from 52.7 degree to 31.9 degree. In this paper, a 12-channel array transmitter module package and thermal simulation are discussed and tested. This is a low cost package design and is suitable for mass production.

The finite difference beam propagation method (FD-BPM) is an effective model for simulating a wide range of optical waveguide structures. We present results of simulations combining this method with finite element modeling of thermal effects using Comsol FEMLAB software. These simulations were developed and are used to examine propagation of optical signals in polymer waveguides in which inhomogeneous temperature profiles are induced using MEMS microheaters, for example, for use in switching applications. Thermal modeling combined with values of the thermo-optic nonlinearity yields three-dimensional refractive index profiles in the active regions of a variety of waveguide structures. The change in the refractive index profile of cladding induces mode deformations and transmission losses due to leakage at the core/cladding interface, in addition to phase shifts in the propagating beams. These effects are used to design thermo-optic switches in both multimode and single mode waveguides, and study the direct effect on propagation due to changes in the applied heater power. In addition, we demonstrate the utility of applying the method to assessing losses due to thermally induced inhomogeneities in planar lightwave circuits such as optical interconnects.

Theoretical and experimental studies on buried ion-exchanged waveguides using field-assisted annealing (FAA) in erbium-doped phosphate glasses were reported. The procedure for making buried waveguides using a two-step process was demonstrated. An analytical model for the FAA process was developed, and both the refractive index profiles and the effective buried depths of these waveguides are predicted. This model is based upon the available data derived from the ion-exchanged planar waveguide measurements. The mode intensity profiles as well as the buried depths were measured with the use of standard end-fire coupling setup and a calibrated image capture card. The buried depths are found to be in good agreement with the simulated results.

While most OADM systems deployed today are static, carriers offering new wavelength services, are looking for ways to migrate towards reconfigurable OADM and dynamically reconfigurable OADM. ROADMs allow carriers to automate service provisioning, wavelength management and failsafe mechanisms, thereby eliminating slow, costly and error-prone manual network configuration. And although next-generation systems offer this flexibility as an integral part of their architecture, carriers must protect their existing investment, extending the life of their existing networks. This article examines the use of add-on optical switching systems, as a cost- and time-effective means for upgrading legacy networks. Two examples are given. In the first, a static OADM is upgraded into a reconfigurable system. The second example shows how existing OADM systems can be upgraded to support dynamic reconfigurability. The article examines also particular advantages of Planar Lightwave Circuits (PLC) for these applications.

An integrated optical device comprising two arrayed waveguide gratings (AWGs) and an electrooptic modulator array is described. When used in conjunction with a broadband light source, the integrated optic device can provide multiple high data rate signals from a single optical light source. The first AWG spectrally slices the light from the broadband light source. Each optical signal is then modulated using an electrooptically-clad silica waveguide Mach-Zehnder interferometer. The second AWG multiplexes the modulated optical signals onto a single output fiber. The paper describes the design of the modulator-multiplexer circuit. Experimental results of the modulator portion of the circuit show the modulation of light at 1550 nm in a electrooptically-clad silica waveguide at data rates of 1 GB/s.

Advance optoelectronic packaging is dependent on the fundamental understanding of the constituent materials behavior during manufacturing and thermal cycling. Such understanding will allow increasing yield and performance of the optoelectronic packages while decreasing their cost. The combination of issues such as dimensional stability, thermal management, electrical connections, and optical access are not yet well understood. Dimensional stability of the generic optoelectronic package was studied by the in-plane moire interferometry. Using digital image analysis enhanced data processing one can get the in-plane displacement resolution of 20 nm. The 1200 l/mm grating was transferred on the test vehicle surface before it was subjected to the thermal cycling. The effect of the temperature cycling on the test vehicle integrity was analyzed by studying the relative deformation of the components. It was found that ten cycles in the range of 0° C - 100° C resulted in permanent strain of 400 μstrains at the corner of the brass post and the quartz plate. Thus, the obtained results suggest that during thermal cycling there was an irreversible deformation or slip at the corner of the brass post on the boundary with the quartz plate. The presented here methodology may be used for analysis of the various adhesives under cyclic thermal loading. The obtained data will be very useful for the design of the optoelectronic package, since it will allow for prediction of the possible misalignment and may be used for the reliability studies of the optoelectronic packages.

A novel AlGaAs/GaAs asymmetric Y-junction optical switch, with 1.7um-thick core layer for single-mode operation at 1.55um, has been designed and fabricated. The switch is based on total internal reflection (TIR) from a carrier-induced index barrier. At the TIR interface inside the Y-junction, a small tapered gap is introduced to obtain a sharp carrier gradient profile for improved switching efficiency. A double masking technique, suitable for monolithic circuit integration, is used to achieve a nearly perfect gap vertex. The switch's strip-loaded waveguide design has a five-layer-heterostructure with a W-shaped index profile, and can provide high coupling efficiency to a single-mode fiber. We have obtained a low optical loss of 1.0dB/cm on 5um-wide waveguides. The asymmetric Y-junction design is carried out by simulating two-dimensional carrier-induced index change based on the calculated carrier concentration and self-heating temperature profile using finite element method (FEM). Preliminary experimental results indicate that optical switches with small size, low loss, fast switching speed and high extinction ratio can be obtained. Switching times of a few nanoseconds have been obtained.

Gain and output power as a function of pump power has been calculated for short Er/Yb doped single mode fibers for various fiber parameters. The calculation shows (i) small signal gain increases with increasing Er concentration, (ii) long fiber lengths provide both higher gain and higher output, (iii) calculated gain is larger for higher Er emission cross section, and, (iv) > 25 dB gain is feasible for 5 cm long fiber for Er and Yb concentrations of 3 X 10²⁶ m^-3 and 1.2 X 10²⁷ m^-3 . Gain of 14 dB has been observed in a 30 mm long fiber at 400 mW. The observed gain is found to be linearly proportional to the length of the fiber with a slope of 0.44 dB/mm at 400 mW of pump power.

In this paper, we report a simple laser direct writing method using pulsed photoabsorption-induced disordering (PPAID) to create high spatial bandgap selectivity and high processed material quality multiple-wavelength chip in InGaAs-InGaAsP heterostructures. Using this intermixing technology, we have achieved a spatial bandgap selectively of better than 2.5mm from a two-section InGaAs-InGaAsP chip with differential wavelength shift of over 120 nm. A theoretical model has been developed to estimate the limit spatial resolution of the PPAID process. Theoretical analysis has also been performed to investigate the relationship between the irradiation conditions and the bandgap shift. Devices such as bandgap tuned lasers, integrated extended cavity lasers, and multiple-wavelength laser chip have also been fabricated and characterized to verify the integration capability of this postgrowth bandgap engineering technique. The quality of the devices fabricated from the intermixed materials was found to be high, indicating that PPAID is a promising process for the fabrication of photonic integrated circuits.