In this paper, we present the wavelength engineering of surface emitting lasers for use in high speed short reach systems, which may include the wavelength expansion, the wavelength integration and the wavelength stabilization based on fully monolithic VCSEL technologies. We have developed highly strained GaInAs/GaAs QW VCSELs emitting at 1.1-1.2 μm band and GaInNAs/GaAs VCSELs at 1.3 μm wavelength. Excellent temperature characteristics have been realized. We extended the emission wavelength of highly strained GaInAs QWs up to 1.2 μm and demonstrated low threshold operations of 1.3 μm GaInNAs lasers grown by MOCVD. We carried out the growth of highly strained GaInAs/GaAs quantum wells on a patterned substrate for realizing multiple wavelength VCSEL arrays in a wide wavelength span. We demonstrated a single-mode multiple-wavelength VCSEL array on a patterned GaAs substrate covering a new wavelength window of 1.1- 1.2 μm. By optimizing a pattern shape, we achieved multiple-wavelength operation with widely and precisely controlled lasing wavelengths. The maximum lasing span is over 190 nm. We proposed and demonstrated a micromachined tunable vertical cavity with a stress control layer, which gives us novel functions including temperature insensitive operation, thermal wavelength tuning, and so on. Either temperature insensitive operation or wide wavelength tuning induced by temperature change can be realized. The temperature insensitive VCSEL based on this technology may be helpful for decreasing the channel spacing in coarse WDM systems. We demonstrated long wavelength GaInAs and GaInNAs VCSELs on GaAs substrates, enabling uncooled operation for high speed data transmission in single-mode fibers. The multiple-wavelength array and wavelength engineering of VCSELs may open up ultra-high capacity short reach systems.

High efficiency continuous wave operation of 1.53 μm vertical-cavity surface-emitting lasers (VCSELs) grown by metal organic chemical vapor deposition (MOCVD) has been demonstrated. Devices show a high differential quantum efficiency of 58% and a single-mode power of 1.38 mW with a side mode suppression ratio (SMSR) of 45 dB. Lasing wavelength is 1537 nm. Lasing operation up to 90° C has been achieved with 0.16 mW power at 85° C. A symmetric and narrow far field angle of less than 12° was observed. An error free transmission at 2.488 Gbps with a -26 dBm minimum average receiver power was demonstrated.

Surface-micromachined 1.55µm vertical-resonator-based devices, capable of wide, continuous, monotonic and kink-free tuning are designed, technologically implemented and characterized. Tuning is achieved by mechanically actuating one or several membranes in a vertical resonator including two ultra-highly reflective DBR mirrors. The tuning is controlled by a single parameter (actuation voltage). The two different layers composing the mirrors reveal a very strong refractive index contrast. Filters including InP/air-gap DBR's (3.5 periods) using GaInAs sacrificial layers reveal a continuous tuning of up to 9% of the absolute wavelength. Varying a reverse voltage (U=0 .. -3.2V) between the membranes, a tuning range up to 142nm was obtained by electrostatic actuation. The correlation of the wavelength and the applied voltage is accurately reproducible without any hysteresis. Theoretical model calculations are performed for symmetric and asymmetric device structures, varying layer thickness and compositions. Models of highly sophisticated color tuning can be found in nature, e.g. in tunable spectral light filtering by trogon and butterfly wings. Bionics transfers the principles of success of nature into natural science, engineering disciplines and applications (here filters and VCSELs for optical communication on the basis of WDM). Light interferes constructively and destructively with nano- and microstructures of appropriate shape, dimensions and materials, both in the artificial DBR structures fabricated in our labs as well as in the natural ones.

8×8 parallel-channeled optical interconnect systems operating at 1 Gbits/s per channel were designed and developed using complimentary metal-oxide-semiconductor (CMOS) circuits driven 850-nm vertical-cavity surface-emitting laser (VCSEL) arrays and the corresponding photodetector arrays. Low operating threshold and voltage were adapted and facilitated in the design and fabrication of VCSELs and photodetectors in order to achieve the low-power consumption for the entire system. The driver and receiver circuits were fabricated on transparent sapphire substrates using 0.5-μm ultra-thin silicon-on-sapphire (SOS) technology and subsequently flip-chip bonded with corresponding VCSEL and photodetector arrays. The VCSEL transmitter and photoreceiver arrays were biased at 3.3 V and optically coupled in a free-space configuration using compound lens systems. Data communications at bandwidth up to 1.0 Gb/s for each single channel were characterized. Bit-error-rate (BER) was measured to be better than 10^-9 from the eye diagrams. Such interconnect systems were also demonstrated for optical data processing using diffractive optical elements.

Vertical Cavity Surface Emitting Lasers (VCSELs) are now essentially the only source used in short distance high bit rate data communications over multimode optical fiber. First commercially realized in 1996 by Honeywell, the primary application has been single channel links operating Ethernet or Fibre Channel protocols in the LAN and SAN environments. Today, the total bandwidth throughput is being raised to more than 10Gbps per channel, with the potential of several channel operation to yield more than 100Gbps. 850nm VCSELs are beginning to emerge in relatively new application arenas and wavelengths. This paper describes the market readiness for VCSELS in a wide variety of optical networking applications.

An new optical correlator containing a tapped delay line with thousands of taps is described. This enables ultra-high resolution correlation. We apply this to monitoring quality-of-signal by correlating the received, degraded bits with and un-degraded signal. The strength of the correlation signal, which is all optical, is proportional to the quality. Dispersion and attenuation can be evaluated in less than 100 ps at 40Gb/s, and jitter and noise in less than 100 ns. This is a significant improvement over minutes or even hours for bit-error-rate measurements. Simulations show good correspondence to eye-diagram measurements, the conventional (but slow) way to measure signal quality. If a network node can know the quality of all its links in real-time, it can re-route signals around poor links, and provide restoration and protection as well. The key to all this is an optical correlator with a very large number of taps in its internal tapped delay line. Our device uses a White cell and a fixed micro-mirror array. In a White cell, light bounces back and forth between three spherical mirrors. Multiple beams circulate in the same cell without interfering and are each refocused to a unique pattern of spots. We make the spots land on the micro-mirror array to switch between cells of slightly different lengths. Our current design provides 6550 possible delays for thousands of light beams, using only ten mirrors, a lens, and the micro-mirror array. We have developed two routing and protection protocols to exploit having this real-time information available to the network.

In this paper we present a novel technique allowing monitoring the chromatic dispersion or the polarization-mode dispersion characteristics of the information within the optical channel. The technique is based upon mapping of the parameters that are to be monitored into measurements of the required multicasting or the variable attenuation of an optical switch and the minimal readout at the detector.

Using of dispersion-managed fiber optics transmission lines is one of the peculiarities of optical networks with WDM. These fiber optics transmission lines consist of jointed lengths with different parameters. They are piecewise regular transmission lines. Novel method of simulation for such fiber optics transmission lines is described. Proposed method is based on fiber optics transmission lines representation as consistence of Laguerre resonators. By applying presented method we can evaluate a pulse damage during propagation along transmission line. By using tools based on method, it is possible to find an optimal design of dispersion-managed fiber optics transmission lines.

A novel hybrid electrical optical Clos switch network for multiprocessor cluster system was presented. For multiprocessor cluster system of 128 hosts, the novel optical Clos network includes 16 basic modules, a passive optical fiber backplane with (8×15)×16 which has a total of 1920 optical data channels and a signaling control system. The basic module is composed of the input line cards of 8 hosts, a single chip of 16×16 crossbar switch, parallel transmitting VCSEL modules for fan-out of (16-1)optical fiber channels and (16-1)×1 optical combiners. The passive optical fiber backplane of very large capacity and high density, based on linear VCSEL arrays and fiber ribbon technology, is to be used to interconnect between hosts of different sub-clusters. The routing of the optical Clos switch network is decided by a signaling control system. Compared with high performance electronic system, this technology offers a relatively easy and simple means of communicating large amount of information between hosts, and lower delay time.

Performances of a WDM multistage interconnection architecture have been evaluated, by using measurements carried out on a 8×8 MEMS 2-D photonic switch made by OMM Inc.. Maximum and minimum insertion losses resulted 3.96 dB and 1.42 dB respectively at the wavelength of 1540.5nm. Furthermore polarization dependent loss and crosstalk were below 0.7 dB and -79 dB respectively. A rist time of 12 ms and a fall time of 4 ms were measured. The measurements were used to study a 40 Gbit/s WDM network which operates by using two fibers each supporting 8 multiplexed channels for a total of 16 channels whose carriers are spaced by 50 GHz and modulated at 2.5 Gbit/s. The network includes an optical cross connect (OXC), designed to switch part or all of the eight channels from one fiber to the other. The heart of the OXC is a 16×16 strictly non-blocking switching matrix based on a three stage Clos architecture. Insertion losses lower than 7 dB were evaluated over the entire network and the eye diagram spreads by just about 4%.

Shortly after it was proposed, optical micro-electromechanical systems (MEMS) was recognized as the most promising all-optical switching technology for core lightwave-communications networks. Focusing on achieving low loss, high port-count, and bit-rate transparency, this technology provides the best match for the applications of provisioning and restoration in core optical networks, without following the mislead path in the history of optical-switch development that all-optical switches need to achieve fast nanosecond switching time. Although the delay in capacity demand of long-haul optical networks has dramatically slowed down the development and commercialization of this technology, many impressive technological advances have been demonstrated as a result of intense research and development in a brief period of a few years. In this paper, we review various technologies for MEMS optical switches, and discuss their roles in optical communication networks.

Optical CDMA technology has shown promise in optical communications, particularly in local-area optical fiber networks. We present a novel O-CDMA scheme with programmable and reconfigurable bipolar code capability using liquid crystal (LC) Spatial Light Modulators (SLMs). The key to our system performance depends on constructing a decoder that implements a true bipolar correlation using only unipolar signals and intensity detection. This has been accomplished using two unipolar correlations that can be performed optically, followed by a subtraction. In our coding system, the power spectrum of a broadband light source is encoded and decoded by programming the SLMs. The high polarization selectivity of these components coupled with the polarization rotation ability of liquid crystal elements makes switching possible with high extinction ratio and low crosstalk. Experimental results including the correlation measurements are presented. Good contrast between the autocorrelation and cross correlation values shows that a binary information symbol can be recovered by an appropriate threshold operation.

A novel replication technology for fabricating polymer optical waveguides has been developed and named as SPICA (Stacked Polymer optical IC/Advanced). SPICA is superior to current semiconductor fabrication technology because it satisfies the needs of low-cost and high-volume manufacturing. Furthermore, SPICA can be processed and packaged much like as IC devices by using planar fabrication and batch processing. Using SPICA, a single-mode optical waveguide has been fabricated the optical characteristics of which include insertion loss of less than 0.2 dB/cm. Other functional devices, such as a coupler, tap-coupler, optical switch, VOA (Variable Optical Attenuator) and optical transceiver, were also successfully fabricated.

We developed the bidirectional optical transceiver module that combined the two LED light sources of different wavelength and three-dimensional (3-D) optical waveguides. By using a light-induced self-written (LISW) technique, we fabricated and tested 3-D optical waveguide circuits for a plastic optical fiber (POF) WDM full-duplex communication module. Because of the large diameter of the POF, an optical waveguide has the advantages, as compared to conventional lens based modules, of a small size and optical low-loss features. The LISW waveguide enables optical components to connect automatically even if the circuit in the module is complex. In this paper, we demonstrate newly developed bidirectional WDM optical module containing 3-D optical circuits, i.e. a branching waveguide and a reflection waveguide, and their optical properties. The module using commercially available green and red LED was constructed and the two-way communication on IEEE1394-S100 (125Mbps) protocol was verified through 10m-length of POF.

We have tried to develop the technology for selective bleaching of core patterns of waveguides on polysilane films by UV light irradiation without wet development and RIE process. Furthermore, we have examined heat treatment of films in order to make transmission loss low instead of introducing fluorine groups in the polymer structure which conventional polymeric material are applied. As a waveguide device, evaluation of low loss straight-line waveguides and multilayered laminating type waveguides at 850nm for MM, and splitter of 8 branches (1×8) at 1550nm for SM was carried out. This new technology is expected to establish the low-cost manufacturing process of optical waveguides which are important components of passive optical devices for PON (passive optical network) system and optical interconnects.

Replication processes of a fluorinated polyimide film were demonstrated using a polyimide and a quartz glass molds. A silicon-oxide was introduced on the top of the mold, which could significantly improve the separation between the mold and the fluorinated polyimide film without degrading the optical property. A single- and a multimode waveguide film were fabricated using a cladding film with grooves formed by our replication process. The multimode waveguide exhibited low optical propagation loss of 0.36dB/cm at the wavelength of 0.85um. Using the fluorinated polyimide waveguide film, a low cost optical board with vertical-cavity surface-emitting laser (VCSEL) and photodiode (PD) was developed for intra-board level optical interconnection. The edges of the multimode waveguide film were butt-coupled to a commercialized 0.85um multimode VCSEL and PD by the passive alignment technique developed. The waveguide film coupled with the VCSEL and the PD was mounted on a FR-4 printed wiring board (PWB), bending the waveguide film. The total loss of two couplings between the waveguide and optical components (VCSEL and PD) was obtained to be less than 3 dB. We confirmed that the high-bit-rate data, >3Gbps, was transmitted using the optical board on the PWB.

In this work simple fabrication techniques of polymeric optical waveguides and wavegiode elements toward low insertion loss are proposed and demonstrated. For single-mode propagation, light-induced self-written (LISW) optical waveguide is fabricated. For large core optical waveguide, hot-emboss technology is adopted for the formation of platform with fiber guide. Moreover, simultaneous formation of channel waveguide with 45° mirror based on the mold technology is also reported.

We developed an optical fiber collimator, which means one of the key optical systems for optical components and devices with ultra compactness and cost effectiveness. The beam waist diameter of 250 μm through the optical components and devices might be recommended, taking consideration of space between two pair of optical fiber collimators and compactness without any excess insertion losses. We designed and fabricated an aspherical collimating lens with coupling loss of less than 0.2 dB with a single-mode optical fiber. Several innovative optical components and devices based on this common platform have been proposed.

The operational bandwidth of optical fiber amplifiers has been expanded by developing rare-earth-doped fiber amplifiers and fiber Raman amplifiers (FRA), aiming at realizing ultra-broadband optical amplifiers in this decade. Especially, non-silica-based rare-earth-doped fiber amplifiers have played an important role in opening up new amplification bands. This paper reviews a development status of optical fiber amplifiers.

Distributed amplification in fiber-optic transmission systems is quantitatively studied. Optimization techniques to find amplifier parameters that maximize system performance are discussed. A particular emphasis is given to noise properties of Raman amplifiers and a tradeoff between the noise performance and nonlinear impairments. It is shown that Raman amplification in dispersion-managed fibers closely approaches the theoretical limit of an ideal distributed amplifier and further increases the system reach. Recent advances in the Raman-assisted ultra-long-haul transmission are briefly reviewed.

We review recent developments of fiber-based optical parametric amplifiers (FOPA). While a theoretical framework based on highly efficient four-wave mixing is provided, emphasis is on applications enabled by the parametric gain including optical sampling, wavelength conversion, and pulse generation. As these amplifiers offer high gain and low noise at arbitrary wavelengths with proper fiber design and pump wavelength allocation, they are also candidate enablers to increase overall WDM system capacities similar to the better-known Raman amplifiers. A comparison with Raman amplifiers is also made and the future outlook of parametric amplifiers is discussed. As this is yet a very new (enabled by the availability of highly nonlinear fiber and inexpensive high power pump lasers) and not fully explored type of amplifier, there is reason to believe that substantial progress may be made in the future, perhaps involving ”holey fibers” to further enhance the nonlinearity and thus the gain.

In this paper, the development of phosphate glass fiber amplifier with a high gain per unit length is reviewed. The performance of compact multimode pumped erbium-doped phosphate fiber amplifiers is presented. A fiber amplifier with a small signal net gain of 41dB at 1535nm and 21dB over the full C-band was demonstrated using a newly developed 8-cm long erbium-doped phosphate fiber excited with a 1-W, 975-nm multimode laser diode. A theoretical model was developed for the multi-mode pumped amplifier based on modified rate equations and effective beam propagation method. Close agreement between experimental and modeling results is observed.

We present model calculations of gain at 1550 nm for Er-Yb double clad amplifiers as a function of pump power, signal input power, fiber length and temperature. The calculated gain is ~20 dB for 6W pump for 20 mW input signal at 25 C. More than 10 W of amplified power can be obtained using 8m long fiber with ~25 W of pump power. The gain decreases with increasing input signal. We show that a high power Er-Yb co-doped double-clad fiber amplifier also exhibits high gain for Yb transition near 1060 nm. This is not unexpected since the input pump causes population inversion in Yb also. The high gain can lead to high peak power pulses at 1060 nm in the presence of small reflections.

We must expand the operating wavelength range of the optical fiber amplifier if we are to achieve a large scale DWDM and CWDM optical communication system with high performance levels. In this report, we introduce the S-band amplification technique with a Tm³⁺-doped fluoride fiber amplifier and an Er³⁺-doped fiber amplifier, and a fiber Raman amplification technique with a wider application range realized by using tellurite fiber. Furthermore, we describe the use of our proposed wide optical fiber amplifiers in an 8-channel CWDM communication system.

Future networks will require very high throughput, carrying dominantly data-centric traffic. The role of Photonic Networks employing all-optical systems will become increasingly important in providing scalable bandwidth, agile reconfigurability, and low-power consumptions in the future. In particular, the self-similar nature of data traffic indicates that packet switching and burst switching will be beneficial in the Next Generation Photonic Networks. While the natural conclusion is to pursue Photonic Packet Switching and Photonic Burst Switching systems, there are significant challenges in realizing such a system due to practical limitations in optical component technologies. Lack of a viable all-optical memory technology will continue to drive us towards exploring rapid reconfigurability in the wavelength domain. We will introduce and discuss the advanced optical component technologies behind the Photonic Packet Routing system designed and demonstrated at UC Davis. The system is capable of packet switching and burst switching, as well as circuit switching with 600 psec switching speed and scalability to 42 petabit/sec aggregated switching capacity. By utilizing a combination of rapidly tunable wavelength conversion and a uniform-loss cyclic frequency (ULCF) arrayed waveguide grating router (AWGR), the system is capable of rapidly switching the packets in wavelength, time, and space domains. The label swapping module inside the Photonic Packet Routing system containing a Mach-Zehnder wavelength converter and a narrow-band fiber Bragg-grating achieves all-optical label swapping with optical 2R (potentially 3R) regeneration while maintaining optical transparency for the data payload. By utilizing the advanced optical component technologies, the Photonic Packet Routing system successfully demonstrated error-free, cascaded, multi-hop photonic packet switching and routing with optical-label swapping. This paper will review the advanced optical component technologies and their role in the Next Generation Photonic Networks.

In this paper, the numerically efficient finite element based full vectorial modal and propagation approaches are used to design and optimize various photonic components for the manipulation of polarization effects in opto-electronic systems. Designs of directional coupler-based polarization splitters both using a metal cladding and without a metal cladding are presented. It is also shown that a compact polarization splitters both using a metal cladding and without a metal cladding are presented. It is also shown that a compact polarization splitter can be designed by using single section or two-section multimode interference (MMI) guides. In this paper, a novel concept of a single polarization optical waveguide with layered core is introduced and its expected performance is reported, along with its experimental validation. In this paper, the origin of polarization cross-talk and design approaches to minimize its detrimental effects is also presented. Finally, designs of single and multiple sectioned passive polarization rotators using waveguides with slanted side-walls are also presented.

We present the results for a 50GHz drive amplifier for use with a Mach-Zehnder modulator. The MMIC device is packaged using a flexible substrate technology to obtain compact size and broadband performance. The packaged device exhibits well-matched transmission lines on the input and output, and large gain and bandwidth. The MMIC performance is directly related to performance of the drain bias circuit.

The evolution of the dynamic optical components marks an important change in the way optical networks will be configured in future. This new class of optical components can be electronically controlled and configured in software thereby enabling service providers to automate service provisioning with out expensive truck rolls. Numerous component and subsystem manufacturers are pursuing the challenge of developing these innovative products with various technologies like MEMS and Liquid Crystal. These new products in due time will enable network evolution to a dynamic architecture, allowing carriers to optimize bandwidth allocation and implement additional revenue generating services.

All optical regenerations or wavelength conversions using SOA-based polarization discriminated switch injected by a transparent assist light are reviewed. First, the reduction of a gain recovery time in SOA by injection of a transparent assist light wass discussed. A simple measurement technique of cross gain modulation (XGM) and cross phase modulation (XPM) in SOA was shown to confirm that the injection of transparent cw assist light reduced a gain recovery time without significant reduction in the amount of XGM and XPM. All optical regeneration operation 40Gbit/s as well as bit-rate tunable operation from 10Gbit/s to 80Gbit/s were presented. Simultaneous demultiplexing from 80Gbit/s to 2 channels of 40Gbit/s signals with little loss was also demonstrated. Finally, tolerance to amplitude noise and timing jitter was discussed. Those results indicate that the SOA-based polarization discriminated switch is a promising candidate for all-optical regenerator from the practical point of view.

An number of parameters, such as gain, modulation response, linewidth enhancement factor and relative intensity noise in modulation-doped InGaAsP quantum well (QW) laser emitting at 1.55 μm have been theoretically investigated. The results indicate that the relaxation oscillation frequency for p-type modulation doped QW laser is enhanced by a factor of more than 2 compared to that for undoped MQW lasers. The linewidth enhancement factor of p-type modulation doped QW laser is reduced to ½ of that of undoped MQW laser and the relative intensity noise (RIN) is reduced by a factor of > 10 dB compared to that for undoped MQW lasers.

The effect of velocity matching, impedance matching, conductor loss and dielectric loss on the optical bandwidth is reported for Lithium Niobate and GaAs electro-optic modulators by using rigorous numerical techniques. It is shown that by etching Lithium Niobate the switching voltage and bandwidth can be improved and similarly by using a higher aluminium content in the buffer layer the device length for a given optical loss can be shortened. It is also observed that the dielectric loss and impedance matching play a key role in velocity matched high-speed modulators with low conductor loss. Finally it is indicated that by using tantalium pentoxide coating, velocity matching can be achieved for GaAs modulators.

For several years tunable lasers and other WDM sources has been as one of the hottest components topics in photonics, but like other areas the downturn has hit hard. In this paper we will describe the latest developments in the area of optical sources for DWDM systems, including tunable lasers and multiwavelength sources, and we will look at the companies (still) competing in this area. We will start by introducing the basic technologies used for DWDM sources. After that we will give an overview of the source options, and discuss tuning methods and wavelength control issues. The source options under discussion will include monolithic tunable lasers, hybrid structures and external cavity lasers, wavelength selectable laser arrays, tunable VCSELs, and non-semiconductor alternatives. Numerous examples will be shown, and the characteristics and performance of the various devices will be discussed. The key performance parameters, such as tuning range, power and switching speed will be related to the expected areas of application. These areas include sparing, fixed wavelength transmitter replacement, and use in wavelength switched networks.

Wavelength division multiplexing (WDM) has become a key technology to cope with the exploding demand for the communication traffic. A variety of optical technologies, especially device-related technologies, has been developed to put the WDM into a new phase, i.e., next generation WDM. This paper reviews the recent advances in the WDM technologies, among which includes wavelength selective light sources, multi-wavelength integrated VCSELs, and wavelength tunable devices.

Modelocked semiconductor diode lasers are used as compact sources for multiwavelength generation. The generation of multiple wavelengths can be grouped into 2 general categories: 1) multiple continuous wave, phase locked optical frequencies, or 2) multiple, synchronized modelocked wavelength channels. Applications of these sources are demonstrated in areas of optical sampling, access networks, and arbitrary waveform generation.

A tunable waveguide grating router (WGR) design is reported, where a subpicosecond phase shift is obtained by means of the electro-optically induced refractive index change in the arms of an arrayed-waveguide grating (AWG) made of highly nonlinear poled polymer CLD-75/APC. The polymer consists of a guest-host system, formed by a ring-locked phenyltetraene bridged cromophore dispersed in an amorphous polycarbonate, with coefficient r₃₃=55pm/V and propagation losses of 1.7dB/cm. We propose a multilayer structure on Si substrate, where segments of each waveguide of the AWG are sandwiched between a ground gold electrode and electrodes whose length varies over the AWG. Numerical simulations of a device with electrode length difference of 250μm show a tuning range of 11nm centered at 1550nm by varying the applied voltage from -90V to +90V. From the optimized AWG, a WGR operating with 16 channels spaced by 100GHz has been designed. The WGR is made of single-mode rib waveguides and buffers whose thicknesses are respectively 1.8μm and 1.7μm. A broader tunability range is obtained using the push-pull technique, which induces a refractive index change of opposite sign in two halves of the AWG. A crosstalk of -40dB with tuning range of 22nm over the C-band was figured out.

High resolution DWDM devices based on the principles of gratings (planar, Bragg, AWG, etc.) and Fabry-Perots (etalon, Lummer-Gehrke plate, etc.) suffer from inherent limitations due to (i) temporal pulse stretching of data, and (ii) broadening of time integrated spectral (demuxed) fringes. While the relation, dν_Fdt >1, can account for these limitations, our analysis imply that dnF does not represent real, physical frequencies. We explain the broader implications of this interpretation in designing DWDM devices based on gratings and Fabry-Perots and illustrate how to use prisms, photonic crystals and non-linear devices for very high data rate per channel.

We report on various types of ultrafast all-optical signal processing with hybrid-integrated Symmetric-Mach-Zehnder (HI-SMZ) all-optical switches driven by 40-Gb/s or 40-GHz optical pulses. To operate HI-SMZ switches with such high-repetition excitation, we use longer semiconductor optical amplifiers (SOAs) as nonlinear phase shifters than we previously used. We demonstrate that extending the SOA length is useful for increasing carrier injection and thus enhancing the nonlinear phase shift in SOAs. We show 3R regeneration and wavelength conversion at 42 Gb/s using HI-SMZ switches with longer SOAs. We also show error-free optical demultiplexing of 168- or 336-Gb/s signal pulses with HI-SMZ switches driven by 42-GHz control pulses.

Future photonic networks will perform routing and switching in the optical layer based upon ultrafast photonic processing. An ultra-wideband hierarchical hybrid optical time division multiplexing/wavelength division multiplexing (OTDM/WDM) network is proposed for the future core photonic network. As its enabling technologies, continuous C- and L- wavelength-band signal generation, OTDM-WDM multiplexing format conversions, and OTDM wavelength-band conversions are demonstrated.

Photonic crystal fibre, or holey fibre, offers a new paradigm in optical fibre where the effective properties of the holey material can be engineered to differ widely from the bulk properties of the matrix material. This engineering freedom has led to development of fibres with unusual and useful properties for applications throughout physical and biological sciences.

Photonic crystal fibers (PCFs) confine light by the regularly aligned cladding air-holes. PCFs can produce large birefringence of >10^-3 by asymmetrical structures. Since the polarization mode in the slow axis is hard to couple to the fast axis mode due to the large birefringence, such PCFs can be used as polarization maintaining fibers. To induce such anisotropy, some hole patterns can be considered. In this paper, the structures of polarization maintaining PCFs, two defects type and enlarged type are described. In 100 m fiber, the extinction ratio better than -20 dB and -30 dB were obtained for former and latter type PCF, respectively. We have also examined the cross-talk degradation by the bending distortion. In the bending diameter of 10 mm, the cross-talk degradation was no larger than the measurement accuracy of 0.2 dB in such polarization maintaining photonic crystal fibers, but the 20 dB degradation was observed in the PANDA fiber, which was the conventional polarization maintaining fiber.

This paper surveys recent work in several photodetector areas including high-speed, low-noise avalanche photodiodes, high-power photodiodes, solar-blind ultra-violet PIN photodiodes, and quantum dot infrared photodetectors (QDIPs).

In this talk, we will review our research efforts on resonant cavity enhanced (RCE) high-speed high-efficiency photodiodes (PDs) operating in the 1st and 3rd optical communication windows. Using a microwave compatible planar fabrication process, we have designed and fabricated GaAs and InGaAs based RCE photodiodes. For RCE GaAs Schottky type photodiodes, we have achieved peak quantum efficiencies of 50% and 75% with semi-transparent (Au) and transparent (indium-tin-oxide) Schottky layers respectively. Along with 3-dB bandwidths of 50 and 60 GHz, these devices exhibit bandwidth-efficiency (BWE) products of 25 GHz and 45 GHz respectively. By using a postprocess recess etch, we tuned the resonance wavelength of an RCE InGaAs PD from 1605 to 1558 nm while keeping the peak efficiencies above 60%. The maximum quantum efficiency was 66% at 1572 nm which was in good agreement with our theoretical calculations. The photodiode had a linear response up to 6 mW optical power, where we obtained 5 mA photocurrent at 3 V reverse bias. The photodetector had a temporal response of 16 psec at 7 V bias. After system response deconvolution, the 3-dB bandwidth of the device was 31 GHz, which corresponds to a bandwidth-efficiency product of 20 GHz.

It has been shown recently that the noise and speed characteristics of avalanche photodiodes (APDs) can greatly benefit from the presence of a reasonable amount of initial energy stored in the injected carriers that initiate the avalanche process. The benefits range from reduced excess noise factor, increased abruptness in the breakdown probability, as well as an increase in the bandwidth. The key mechanism for the improved performance is the significant reduction of the first dead space in the ionization process. The dead space is the minimum distance a carrier must travel before acquiring sufficient kinetic energy enabling it to impact ionize. The reduction of the first dead space has the effect of localizing the first impact ionization and forcing it to occur quickly near the edge of the multiplication region. This, in turn, will have the effect of nearly inducing two avalanche processes that run in parallel and whose combination will yield the total gain. In this talk, a theoretical model for the avalanche multiplication is utilized to examine the fundamental limits of the gain-bandwidth product in light of the initial-energy effect in practical APD structures.

High bandwidth short distance communications standards are being developed based on parallel optical interconnect fiber arrays to meet the needs of increasing data rates of inter-chip communication in modern computer architecture. To ensure that this standard becomes an attractive option for computer systems, low cost components must be implemented on both the transmitting and receiving end of the fibers. To meet this low cost requirement silicon based receiver circuits are the most viable option, however, manufacturing high speed, high efficiency silicon photodetectors presents a technical challenge. Resonant cavity enhanced (RCE) Si photodetectors have been shown to provide the required bandwidth-efficiency product and we have recently developed a method to reproduce them through commercially available fabrication techniques. In this work, commercially reproducible silicon wafers with a 90% reflectance buried distributed Bragg reflector (DBR) are used to create Si-RCE photodetector arrays for optical interconnects. The Si-RCE photodetectors have 40% quantum efficiency at 860 nm, a FWHM of 25 ps, and a 3dB bandwidth in excess of 10 GHz. We also demonstrate Si-RCE 12×1 photodetector arrays that have been fabricated and packaged with silicon based amplifiers to demonstrate the feasibility of a low cost monolithic silicon photoreceiver array.

Flip-chip-mounting of integrated circuits has been shown as an effective way to connect integrated circuits with substrates providing highest bandwidths. This technique also has been used to connect chips directly without causing significant parasitics. In this paper, hybrid integrated photoreceivers with large bandwidths for fiber-optic data transmission will be discussed. Several photoreceivers with bandwidths of more than 60 GHz have been fabricated by combining photodiodes and amplifiers, which have been optimised in separated wafer runs. Key features of this type of photoreceivers are a waveguide photodiode with a high responsivity (0.8 A/W) and a traveling wave amplifier with low input impedance as well as high amplification to a 50 Ω output termination up to 65 GHz. Further improvement of bandwidth is expected by using metamorphic HEMTs instead of pseudomorphic ones in the amplifier.

Photonic integrated photodetectors are built by combining guided-wave, optoelectronic and microwave devices. They offer high conversion efficiency, extremely high speed operation and high power linearity up to 20 mW of optical power. A careful design of the applied spot size converter allows an optimized fiber-chip coupling and polarization insensitive performance of the photodiodes. Recent results of electro-optic sampling measurements reveal 3 dB- and 6 dB-bandwidth of 64 GHz and 100 GHz, respectively, as well as an excellent phase linearity up to 160 GHz. Further integration of two photodiodes leads to e.g. 40 GHz balanced detectors enabling DPSK transmission for long-haul systems. Using optical pre-amplification an excellent sensitivity of -32 dBm was achieved at 40 Gbit/s.

Multi Form Pluggable (MFP) 8-channel optical link modules for OC-3 (STM-1) and OC-12 (STM-4) intermediate reach (IR-a) applications have been developed. The MFP module can increase packaging density by 60% and reduce power consumption by 60% for virtually the same cost compared to the use of conventional single-channel 8 Small Form Pluggable (SFP) modules.

High power photodetectors have made remarkable progress in performance over the last few years with prospective application in high-bit-rate digital receivers with reduced complexity, CATV, analog-to-digital converter, high power RF distribution and optoelectronic generation of high power microwaves and millimeter-waves in RF photonic systems. An in depth understanding of the failure mechanism and novel techniques in thermal management have dramatically improved the power handling capability of the high power photodiodes. In order to increase the output RF power of the photodiodes, several groups have developed surface-illuminated, waveguide, traveling-wave, uni-travelling carrier, parallel fed traveling wave with MMI couples and distributed photodetector arrays. An impressive improvement in link gain, noise figure, and spurious free dynamic range (SFDR) of externally modulated links has been achieved by several groups. The realization of high power balanced photodetectors has helped suppress the relative intensity noise (RIN) of the laser source and the amplified spontaneous emission noise (ASE) from erbium-doped fiber amplifiers (EDFA) reaching the level of quantum noise floor. This paper will highlight various approaches that different research groups have undertaken and will discuss our recent advances in power handling capability and reliability analysis of velocity-matched distributed photodetectors and balanced photodetectors for RF photonic links.

Charge balance in the depletion region of a photodiode can be used to reduce the space charge effect. A partially depleted absorber (PDA) InGaAs/InP photodetecter with space charge balance and transit time balance has been demonstrated. A 1dB large signal compression current of 24mA was achieved at 48GHz for an 8μm × 8μm photodiode. A 100μm-diameter photodiode achieved a maximum saturation current of 199mA at 1GHz.

The uni-traveling-carrier photodiode (UTC-PD) is a novel photodiode that utilizes only electrons as the active carriers. This unique feature is the key to achieving excellent high-speed and high-output characteristics simultaneously. A record 3-dB bandwidth of 310 GHz and a millimeter-wave output power of over 20 mW at 100 GHz have already been achieved. The superior capability of the UTC-PD for generating very-large high-bit-rate electrical signals as well as a very-high output power in millimeter/sub-millimeter ranges can innovate various systems, such as broadband optical communications systems, wireless communications systems, and high-frequency measurement systems. Achievements include photoreceivers of up to 80 Gbit/s, DEMUX operations using an integrated optical gate of up to 320 Gbit/s, and a 10-Gbit/s millimeter-wave wireless link at 120 GHz. Also achieved has been high-power millimeter generation of 17 mW at 120 GHz with a waveguide-output UTC-PD module, considered for use in the photonic-local system of radio telescopes.

The increased absorption volume of traveling wave distributed photodetectors can be used for high power generation without bandwidth reduction. In these traveling wave photodetectors, in order to not be limited by round-trip bandwidth limit, half of the generated photocurrent, which is traveling towards the input end, has to be absorbed in an input termination. We propose cancellation of the backward propagating current by using a multi-section transmission line to eliminate this loss. The impedances of the individual transmission line sections are chosen such that the backward current (traveling towards input end) generated by each of the diodes is canceled by the reflected fraction of the forward current (traveling towards output end) generated by the preceding diodes. With backward wave cancellation, RF response increases by up to 6dB while maintaining high-bandwidth. We present here the experimental results of a traveling-wave-backward-wave-cancelled photodetector with 38GHz bandwidth and up to -1dBm of linear RF output at 40GHz.

The authors report the most recent progress in Type II InAs/GaSb superlattice materials and photovoltaic detectors developed for focal plane array applications with a cutoff wavelength of ~8 μm. No turn-on of tunneling current was observed even at a reverse bias of -3 V for a 3 μm thick p-i-n photodiodes. The thermally-limited zero bias detectivity under 300 K 2 π FOV was 2~3×10¹¹ cm•Hz^1/2/W at liquid nitrogen temperature, with a current responsivity of 2~3 A/W and a mean quantum efficiency of ~50%. Initial passivation using SiO₂ has shown to decrease the dark current by ~30% at a reverse bias of -1 V. The same detector structure was used for focal plane arrays with silicon readout integrated circuit. Concept proof of imaging was demonstrated with a format of 256×256 at liquid nitrogen temperature.

This paper presents a physical model for dark count rate and single-photon quantum efficiency of single-photon avalanche photodiodes. The model makes direct connections between the performance of single photon avalanche detectors and the physical parameters of the devices, which are useful for choosing commercial APDs to function in single-photon mode, designing APDs specifically for single-photon detection, and setting up suitable device operation conditions for optimal performance. Good agreement between the calculations and the experimental data from commercial InGaAs/InP APDs validates the model.

Silica-based planar lightwave circuits (PLC) are key components of functional devices designed for use in optical fiber communication systems because, compared with bulk optics devices, they offer compactness, excellent stability and reliability in addition to high functionality. This talk reviews the current status of PLC development. The fabrication process, basic characteristics, packaging and reliability are presented, many device applications are described, and experimental results related to optical switches, interleave filters, chromatic/polarization dispersion compensators, gain spectrum equalizers and wavelength multiplexers are reported. Finally, the history of PLC development is summarized, and future target devices are described along with the kind of development needed for such devices designed for use in the next generation of optical communication systems.

Optical Interleaving filters are attractive components in WDM systems especially narrower than 100GHz channel spacing system. Several types of interleaving filters have been proposed and realized. An interleaving filter is required to have the Box-like characteristics such as the periodic response, flat-passband, low insertion loss and low crosstalk. In addition, low chromatic dispersion (CD) is indispensable for DWDM systems. We focus on planar lightwave circuit interleaving filters. In this paper, we present interleaving filters with a tandem configuration of Fourier transform-based MZIs. The circuit using PLC technique is fabricated with high index contrast waveguides of Δ1.5%. We also have demonstrated monolithically integrated 1×4 (50-200GHz) interleaving filters and arrayed waveguide gratings(AWGs) with 200GHz-channel spacing suitable for interleaving WDM system.

Arrayed Waveguide Gratings (AWG) have gained significant popularity in Wavelength Division Multiplexing (WDM) in recent years. One of the most important characteristics of an AWG is to have a flat spectral response in order to maximize the performance of the device and reduce crosstalk. In this paper, we present a novel method of achieving a flat spectral response AWG in silicon-on-insulator (SOI). This is achieved using free carrier doping, which ultimately introduces absorption which alter the intensity field distribution at the output of the array waveguides. By applying Fourier optics to the free space region of the AWG, this field profile is the inverse Fourier transform of the output field at the AWG, hence producing a flat spectral response. The rib of the arrayed waveguides is designed to operate in singlemode condition and exhibit minimum polarization dependence. The AWG is designed to operate at a centre wavelength of 1.55μm at a grating order of 51 with path length differences of 22.61μm. Particular emphasis is paid to the theoretical analysis and development of a flat spectral response AWG, including a number of different implantation doses.

Silica-based planar lightwave circuit (PLC) hybrid integration is a promising candidate for realizing highly functional photonic components and subsystems. This paper reviews the key features of PLC hybrid integration technology, including the PLC platform, semiconductor optical devices suitable for hybrid integration and the assembly technique. Recently developed optical modules that use this technology are also described. They include a bi-directional optical transreceiver module and a multiwavelength laser module, both of which employ a PLC platform with a super high refractive index difference.

We present the design and fabrication of micromachined vertical cavity filters with stress control, which give us novel functions including temperature insensitive operation, thermal wavelength tuning, wavelength trimming and 2-D multi-wavelength integration. The temperature dependence could be freely controlled either for temperature operations or for wide wavelength tuning. In addition, we propose the use of a hollow waveguide in which light can be confined in air for realizing temperature-insensitive waveguide devices. We present various unique features in hollow waveguides and the combination with microelectro-mechanical system (MEMS) gives us novel widely tunable waveguide devices.

Recently, Michelson-Gires-Tournois interferometer (MGTI)-based optical interleavers have gained much attention due to its excellent performance, simple configuration, and great potentials. Optical interleaver extends the lifetime of existing demultiplexer in optical fiber communication systems and networks, and acts as cost-effective driver toward a narrower-channel-spacing (100/50/25/12.5GHz), next generation, ultra-dense multichannel DWDM optical network. The objective of this paper is twofold: (1) provide a technical overview of the numerous existing implementations of optical interleaver, and (2) focus on one type of family of interleavers based on MGTI, which the author originally proposed and developed. Here, we discuss the different proposed variations of this novel interleaver and compare numerical results of their performances in terms of their comparative flattop shapes, optical configurations, and chromatic dispersion performances.

For many years, fiber-optics communication has become an essential part of the development of our modern society. For example, its significance comes from the increasing demands on real-time image transmission, multimedia communication, distance learning, video-conferencing, video telephone, and cable TV, etc. This paper is to develop an automatic transmittance measurement system for a DWDM (dense wavelength division multiplexing) filter. In this system, a grating-based monochromators is devised to generate a collection of monochromatic light with various wavelengths, instead of using an expensive tunable laser. From this approach, the cost of the proposed system will be much lower than that of those having the same functions, by one order. In addition, we simulate the spectral filtering to investigate the resolving power of the system. It appears that our simulations give quite satisfactory results.

This presentation will discuss advancements in thin-film filter technology for telecommunications applications in the following key topic areas: Narrow Band Channel Spacing: Several years ago, it was thought by industry analysts that thin-film filter technology would not be practical for channel spacings below 100 GHz due to loss and center wavelength stability requirements. Thin-film solutions for 50 GHz and 25 GHz channel spacings have been demonstrated with state of the art low-loss thin-film fabrication and new packaging strategies. Chromatic dispersion and dispersion power penalty for narrow band filters will also be discussed. Wide band filters with fewer dropped channels and near theoretical performance: Optical Add-Drop Multiplexing (OADM) is a key application for wide-band thin-film filters. Application requirements call for a minimization of the number of dropped channels with very steep filter sidewalls. The reflection notch function must have very deep isolation to minimize coherent cross-talk effects. This class of filter requires advancements in total film thickness, accurate layer thickness control, and center wavelength temperature control. High angle of incidence filters: Filter designs that use fixed angle of incidence in the 8 degree to 15 degree range are enjoying resurgence in popularity especially for CWDM applications. Wavelength tunable optical add/drop filters using angle-tuning methods are also emerging as an important application. Special designs have been created to optimize performance in these applications.

Digital Planar Holography (DPH) has arrived due to progress in microlithography, planar waveguide fabrication, and theoretical physics. A computer-generated hologram can be written by microlithography means on the surface of a planar waveguide. DPH combines flexibility of digital holograms, superposition property of volume (thick) holograms, and convenience of microlithographic mass production. DPH is a powerful passive light processor, and could be used to connect multiple optical devices in planar lightwave circuits (PLCs), and if combined with active elements on the same chip, may perform not only analog operations but also logical ones. A DPH implementation of a multiplexer/demultiplexer with discrete dispersion is proposed and demonstrated, avoiding communication signal distortion inherent in multiplexers/demultiplexers with continuous dispersion. The concept of discrete dispersion leads to a device with a flat top transfer function without a loss penalty. The dispersion is created with custom-designed bandgaps for specific directions. A DPH hologram resembles a poly-crystal with long-range correlations, and it exhibits the properties of a quasi-crystal. Unlike photonic crystals, light in quasi-crystal may propagate in almost any direction. Single mode planar waveguides are specially designed to suppress parasitic reflections that appear due to mixture of TE-modes, TM-modes, and cladding modes. Demultiplexers with 2-32 channels were demonstrated on planar waveguides with binary single-layer lithography.

Fiber Bragg gratings (FBGs) have been widely used in optical communication systems for wavelength tuning and dispersion compensation. FBGs are made by laterally exposing the core of a single-mode fiber to a periodic pattern of intense ultraviolet light. This proposes that, by exposing the fiber core to other temporary and alternative exposure sources in real time may create a gratings effect which is so called “real time virtual gratings”. The virtual gratings can be classified as a real time adaptive fiber Bragg gratings which may be used in dispersion compensation applications.

The application of the acoustic-induced vibration on a fiber Bragg grating has been proposed as the function of controlling reflectivity levels and switching reflection wavelengths. Moreover, a switchable multi-wavelength optical filter is expected to develop for various applications in optics. Thus an acousto-optic interaction in a superstructure fiber grating (SFG) can provide a multi-wavelength reflective filter with the function of switching the operation wavelength. In this paper, we experimentally demonstrated that the channels of a blazed SFG could be increased or switched as the acoustic waves were launched into the fiber. When the acoustic wave is applied in the fiber and travels along the fiber axis, the cladding modes of a blazed SFG can couple back to the core mode by acousto-optic interaction in the fiber. The grating reflectivity and the number of the induced wavelength channels can be controlled by acoustic flexural amplitude. Thus, this device acts as a switchable multi-wavelength comb filter for the applications in a WDM system, or in fiber lasers or in fiber sensors.

In this paper we evaluate the influence of pumping schemes on the performance of fiber Raman amplifier bidirectionally pumped at multiple wavelengths. The numerical model we used includes bi-directional interactions between all signal and pumping wavelengths induced by stimulated Raman scattering, spontaneous Raman scattering and Rayleigh backscattering. We develop a effective algorithm to numerically solve the equations and show that by changing the ratio of forward to backward pumping power for each wavelength we can get very good tradeoff between gain flatness, optical signal to noise ratio and nonlinear effects.

In this paper the effect of instability and chaos in optical fiber networks based on the Internet is described. Nonlinear optical fiber effect especially Back scattering in networks has emerged as the essential means for the construction of active optical devices used for all-optic in-line switching, channel selection, amplification, oscillation in optical communications and a host of other applications. The inherent optical feedback by the back-scattered Stokes wave in optical networks also leads to instabilities in the form of optical chaos. This paradigm of optical chaos in fiber Internet serves as a test for fundamental study of chaos and its suppression and exploitation in practical application in optical communication. This paper attempts to present a survey and some of our research findings on the nature of Stokes chaotic effect on Internet based optical communication.

This paper delves into the fundamentals of hardware-software partitioning between an intelligent, generalized signal processor and reconfigurable hardware. A novel architecture is proposed for a RADAR control system and a design specification is illustrated with simulation results.

In an earlier approach, the 2-D acoustical field profiles on the substrate region are often calculated with BPM. In this paper, we present a new approach based on the finite element - artificial transmitting boundary method and calculate the 2-D acoustical field on the substrate region.

A rigorous numerical approach, based on the full vectorial finite element method, is used to design various types of monolithically integrated spot-size converters for efficient coupling to an industry-standard optical fiber. Spot-size converters with a tapered core, using uniform directional couplers and MMI structures are discussed and the predicted performance of the system reported.

Optical signal processing in any living being is more complex than the one obtained in artificial systems. Cortex architecture, although only partly known, gives some useful ideas to be employed in communications. To analyze some of these structures is the objective of this paper. One of the main possibilities reported is handling signals in a parallel way. As it is shown, according to the signal characteristics each signal impinging onto a single input may be routed to a different output. At the same time, identical signals, coming to different inputs, may be routed to the same output without internal conflicts. This is due to the change of some of their characteristics in the way out when going through the intermediate levels. The simulation of this architecture is based on simple logic cells. The basis for the proposed architecture is the five layers of the mammalian retina and the first levels of the visual cortex.

Optical components based on geometrical and refractive index variations could be studied by coupled mode theory. For that we introduce a new coupling coefficient which takes into account not only the geometrical variations, but also the core refractive index variations along the propagation axis for the coupled mode theory. The results show that the coupling coefficient between modes can be separated in a sum of two coefficients. The first one is the classic coefficient which takes into account the radius variation along axis. The second one is reported in order to study the index variation. We examine with more details the effect of the core dopants diffusion due to the heating during the fabrication process. The concatenation of a biconical tapers and long period grating is presented.

Dispersion-managed transmission lines for WDM technology applications include optical fibers with different refractive index profiles. USF, DSF, NZ-DSF and DCF are used as the transmission medium. It is also possible to use RDFs. By optimizing refractive index profile of DCM fibers, it is possible to design transmission lines with low non-linearity, low loss and low PMD. Advanced method of refractive index profile optimization for DCM fibers is proposed in this paper.

This work presents the analysis of ring resonator based tuneable optical filters and their proposed application in wavelength division multiplexing transmission systems. Using coupled mode theory and its transfer matrix formalism it has been shown that single-ring resonator add-drop filters with coupling coefficients k can provide similar or better performance than double ring resonator add-drop filters with coupling coefficients k/2, while benefiting from simple and more robust designs, simplified tuning control, and easier manufacturing. Further analysis of the their dynamic range showed that thermally tuneable single-ring add-drop filters can easily cover four optical channels at 0.8nm spacing. Their proposed application is a sparse reconfigurable optical add-drop multiplexer with 0.25 add-drop factor and 32 channels at 0.8nm channel spacing that can be utilized in linear, ring, and mesh all-optical WDM networks/transmission systems.

In this paper we present a novel approach allowing cancellation of the polarization mode dispersion (PMD) using a technique involving periodic polarization modulation. A digital post processing following the optical polarization modulation allows significant attenuation of the PMD component in the bit stream.

For the past few years, we have been researching a novel type of photodetector featuring distributed optical amplification. We call these devices traveling-wave amplifier-photodetectors, or TAP detectors. The distributed combination of gain and absorption seeks a larger efficiency while keeping a low optical power, thus avoiding saturation. In this paper, we present experimental results both of GaAs- and InP-based TAP detectors, showing in the former case an external quantum efficiency larger than 200%, and larger than 100% in the latter. The performance limitation is shown to be related to the competition between the optical input signal and the amplified spontaneous emission (ASE) generated in the amplifier.