Integrated optoelectronic circuits that are capable of very high speeds or high functionality have been demonstrated using InP-based heterojunction bipolar transistors (HBTs). Optoelectronic receivers contain photodetectors fabricated from the same epitaxial material structure as the HBTs. High-functionality digital receivers, analog receiver arrays as well as analog-to-digital converters have been realized. Optoelectronic modulation circuits for signal transmission also contain separately grown, surface-coupled multiple- quantum-well (MQW) modulators.

We review the electroabsorption (EA) device as a linear mixer for frequency conversion. We explain the operating principle and compare the operation of the EA device as a modulator/mixer with the operation as a detector/mixer. We present experimental results of EA devices with the quantum- confined Stark effect and the Franz-Keldysh effect. Using a InAsP/GaInP multiple-quantum-well EA device, a conversion loss of 18.4 dB is obtained at 10-mW optical local oscillator power when the device is operated as a detector/mixer. The two-tone measurement on the same device shows a sub-octave, spur-free dynamic range of 120.0 dB-Hz. Phase noise measurement demonstrates that very low phase noise is added in the conversion process. We propose a novel full-duplex operation of the EA device as a detector/mixer for the application in the antenna remoting or fiber-radio systems.

We report our recent development of a fabrication method for monolithic semiconductor waveguided optoelectronic integrated circuits (OEICs) using selective area epitaxy. Selective-area growth (SAG) may can produce an energy band gap variation in the waveguide structure within the same wafer. Therefore, devices such as amplifiers, modulators, and splitters that require different energy band gaps for the same operating wavelength may be fabricated monolithically to avoid the high cost and the signal degradation associated with the hybrid integration. However, epitaxial regrowth is still used in most SAG device fabrication. To further lower the cost and simplify the OEICs fabrication, we proposed a one growth and one waveguide fabrication method. This method requires a single metal-organic-chemical-vapor-deposition (MOCVD) growth of the whole InGaAsP multiple quantum well-waveguide core and InP bottom and top cladding layers on an InP substrate masked by parallel SiO₂ strips. A localized shift of the band gap, that is induced by variations in epilayer thickness and composition, can be controlled by varying the width of the SiO₂ strips and the gap between the strips. We have developed a special waveguide device processing technique for this type of non-uniform SAG materials. The challenge is to obtain the right energy bandgaps, right quantum well widths, right p-i-n electric fields in different waveguide device sections, and maintaining appropriate optical mode profile when the waveguides pass through the layer thickness and composition variations as well as bending curvatures. For the devices discussed here, extensive optical, electrical, and SEM characterizations have been performed to optimize the structure design and processing parameters. A few combinations of integrated waveguide splitters, modulators, and amplifiers have been designed and fabricated. Preliminary device testing has been performed.

A new method to compensate for the column multiplexer (MUS) in photo detector arrays is described and analyzed. It is based on a transimpedance amplifier with improved feedback structure. The focus is on arrays where the photo current is directly used as signal. If this current is transferred to the global transimpedance amplifier by a multiplexer, the resistance of te MUX transistor causes a significant speed degradation. Measurements of a test chip in 0.19micrometers CMOS are presented. It turns out that using the proposed circuit it is possible to almost fully compensate for the MUX in terms of speed and signal amplitude with only little layout area penalty.

This paper presents a 1-Gb/s/ch optical receiver with full rail-to-rail output swing designed in a standard 0.25 micrometers CMOS technology. The receiver consists of two parts, the front part is a transimpedance preamplifier which is used to convert the current signal into the voltage signal for the subsequent process, and the end one is a postamplifier which performs the linear and limiting amplification. The voltage data stream, which is converted by the transimpedance preamplifier, is changed into the adjusted digital data stream by comparator which is followed by the preamplifier. During this process, however, it is very critical to suppress the distortion of the data effectively when the input data is regenerated into the adjusted digital ones. For these requirements, our design consideration in this work is as follows. We detected the middle point of the input signal for determining the reference value. The peak detectors determine the positive and negative peak levels of amplitude of the input signal by a comparator with one N-MOS or P-MOS and a capacitator. Then the output voltage signal is divided by the same resistance and its value is decreased in half.

Conductive cladding layers for electrode poled nonlinear optic (NLO) polymer waveguides have been investigated for use in electro-optic (EO) switching devices. Dropping most of the poling field across the NLO core material optimized poling efficiencies, which in turn allows for lower poling voltages and in-situ poling. In addition, this method has the potential for realizing higher EO coefficients, shorter devices, and lower operating voltages than conventional devices with passive claddings. Poling efficiency enhancement has been achieved for 2-layer structures resulting in the lowest poling voltage to date, 300 V, for a 2 micrometers thick NLO polymer core layer and a 2 micrometers thick Baytron P/PVA)polyethylene dioxythiphene/polystyrene sulfonate in polyvinyl alcohol) conductive cladding layer. This has rendered up to 13 times increase in the effective EO coefficient. New results are presented from extending the concept to 3-layer device structures using the conductive polymer material for both the top and bottom cladding. Our results show that conductive claddings for NLO polymer waveguides will be realizable in actual devices.

Electro-optic (EO) polymer modulators have demonstrated a high frequency response (>50 GHz) and therefore hold promise to become the preferred solution for high-speed data transmission. Typically the modulation is achieved via a Mach-Zehnder waveguide configuration with the active polymer sandwiched between two passive cladding layers. This arrangement has been shown to not be optimal in regard to the efficiency for conversion of RF to optical signal. The cladding layers present an unwanted electrical load when either electric field poling or modulation of the active core polymer is desired. We propose the use of thin layers of Indium Tin Oxide (ITO) deposited on either side of the core to enable direct and highly efficient poling of the central active region while maintaining acceptable optical losses in the waveguide geometry. Presented are the results of poling studies which indicate that out RF sputtered ITO is a viable electrode poling material, while the results of spectrophotometric and ellipsometric results indicate that the optical loss coefficient is within a range that should allow for acceptable use in the proposed geometry. This approach would eliminate the need for conductive claddings that also increase the device poling efficiency.

Photonic crystals have impacted the diverse domain of electromagnetics. With a growing interest, due to their unique ability to control the propagation of light through a periodic dielectric structure, photonic crystals can be integrated with opto-electronic devices to enhance their optical performance. In this paper we investigate two applications of photonic crystals. First, as a multi- channel drop filter, which consists of a waveguide and a frequency selective element (microcavity); and, secondly, as a hetero-structure beam splitter. We also compare the performance of the hetero-structure with that of a uni- structure beam splitter. A Finite-Difference Time-Domain technique was used to stimulate both applications.

There is a clear demand in optoelectronic design for analysis techniques that provide accurate results with minimal computational effort. The spectral Index Method (SIM) is well established as a very accurate yet rapid semi- analytical technique for determining the modal properties of air-clad optical rib waveguides. The present work generalizes the spectral index approach to the analysis of propagation in fully 3D rib waveguide problems. The method is founded upon three practically observed features: fields are polarized, there is little penetration into the air cladding and continuous reflections are very weak with strong reflections occurring only at discrete boundaries. The new propagation algorithm retains the simplicity and computational advantages of the SIM. A full theoretical development of the methods will be presented, along with a discussion of the implementation and application to a typical practical problem illustrative of including taper-based spot-size converters. Comparisons with direct numerical methods show the new technique to be sufficiently accurate for the design of many optoelectronic components. Calculation times for the new method are significantly lower than those for purely numerical methods, making it well suited for use within an iterative design environment.

The geometrical structure of both the ground and excited state of the azo-dyes: Disperse Orange 3 (DO3) and Ch₂- C₄H₄-N=N-C_{44/H(subscript 4}-CH₂ molecules have been investigated applying the Hartree Fock (HF), density functional theory (DFT) methods with the Berny geometry optimization and Ames Laboratory determinant (ALDET)single- double-triple-quadra configuration interaction (SDTQ-CI) method. The investigations proved that the above-mentioned molecules can not rotate around the -N=N- bond. Thus, the alternative mechanism of the isomerization of the DO3 molecule per linear transition state has been suggested and investigated. The obtained one- and many- electron energy diagrams have been drawn. According to these diagrams the above mentioned isomerization way is possible. The mechanism of the isomerization of the azo-dyes molecule per linear transition state is suggested and investigated. The results of light induced internal molecular motions in azo-dyes molecules have been used for the design of light driven logically controlled molecular machines composed form photoactive organic molecules such as carbazole, 1,4- phenylenediamine (PhDA) and 4,5-dinitro- 9(dicyanomethylidene)-fluorene (DN9(CN)₂F), Dithieno[3,2-b:2',3'-d]thiphene and Ferocene (C₁₀H₁₀Fe) molecules joined with -CH=CH- or -N=N- bridges. Ab initio DFT B3PW91 model using 6-311G** and Watanabe (WBTS) basis sets calculations show the stability of Sc₃N and ErSc₂ molecules which exist inside endohedral fullerene C₈₀ derivatives: Sc₃NC₈₀. Analysis of electronic structure of inside clusters allowed proposing that these endohedral fullerenes might be used for electro- optical and magneto-optical switches and for information storage. We performed design of molecular logical devices based on organic electron donor and electron acceptor molecules, fullerene C₆₀ substituted derivative CH₂C₆₀, electron donor-bridge-electron acceptor dyads and triads. Design of new series molecular implementations (MI) of two variable logic functions: AND (NAND), OR (NOR) is based on geometry optimization procedure. Molecular triggers and molecular dynamic memory were designed based on investigations of photo-excitation movements and charge transfer of aza-fullerene supermolecule (NH₂)CH-NC₅₉-NC₅₉-CH(NO₂). Our ab initio DFT B3PW91/LanL3DZ calculation of HOMO-LUMO gap in CdS nanocluster without four phenyle fragments gives value equal to 3.85 eV and the same method calculation of CdS nanocluster with four phenyle fragments gives HOMO-LUMO gap value equal to 3.66 eV. 121

Low cost, reliable packaging that meets essential performance requirements is key to future wide-spread applications of future of planar waveguide integrated optics components. Currently, such integrated optics components are being fabricated using a wide variety of design parameters and materials that dictate the optimum packaging approach to be used in each case. However, there are certain universal packaging issues that must be addressed to some degree for all types of integrated optics. These are mechanical support, environmental protection, thermal control, electrical connection, and optical connection. Most applications, particularly those in the telecommunications arena, require that the entire package stand up to wide temperature excursions, humidity, and mechanical shock and vibration.

One of the main promising applications of micro-machining in All-Optical-Networks (AON) is represented by free-space electro-mechanical Optical Cross Connects (OXCs); free-space components show lower attenuation and lower cross-talk than concurrent technologies based on wave-guides. Although micromachined electro-mechanical OXCs have been recently introduced into the market, further commercialization of devices with increased input/output port-count will certainly require decreasing insertion losses by proper design techniques of both, the electromechanical devices and the system packaging. Among different proposed micromachined architectures, one suitable for small- to medium-size cross-connects id defined by a two-dimensional array of tw-state microreflectors where signal routing occurs in a single plane parallel to the substrate; in this architecture, signals come from and go into Single-Mode Fibers (SMF) collimated by GRIN lenses. This paper discusses different losses mechanisms in micromachined free- space mirror-based OXCs, including both fiber/mirror misalignments on insertion losses. A third section analyzes the effects of imperfect micro-mirrors on insertion losses including finite mirror-size, orientation error, reflectivity and scattering. Finally, conclusions on the required fiber-packaging accuracy and the scalability of micromachined free-space OXCs are presented.

This paper reports on monolithic integration possibilities between a novel micromachined packaging technology and MEMS/MOEMS devices. A high-density lead connection scheme is tailored to fulfill the needs of emerging high density optoelectromechanical/microelectromechanical devices. The packaging structure employs double polysilicon as device and package fabrication and anodic bonding as the sealing technology. Thermal stresses arising due to sealing temperatures have been investigated for various device dimensions. The advantages of the micromachined lead transfer and packaging scheme has been explored and it is indicated that it can incorporate both surface and also bulk micromachined devices as post or intermediate processing steps.

We describe a novel electrically tunable fiber Bragg grating (FBG) filter with a chemically etched cladding and an evaporated metal coating on one side of the fiber. Wavelength tuning ranges up to 2.5 nm with efficiencies greater than 8.5 nm/watt have been demonstrated along with modulation bandwidths up to 14 Hz. We have also demonstrated tuning efficiencies over 200nm/Watt for 30micrometers diameter etched cladding devices tuned under moderate vacuum conditions. Such tunable FBG filters have potential applications in dense wavelength division multiplexing (DWDM), optical signal processing, or in wavelength tunable fiber lasers. In our paper we investigate the tuning efficiency and modulation bandwidth of etched cladding FBGs integrated with Silicon V-grooves that utilize temperature tuning of the filter. By passing an electrical current through a thin metal coating deposited onto an etched cladding FBG, the temperature of the grating can be controlled to tune the spectral characteristics of the FBG. Additionally, to simplify the fabrication process, we evaporate the metal coating onto the etched FBG from only one side. This radially asymmetric metal coating is simpler to fabricate, since it does not require any mechanical fixture to rotate the fiber during metal deposition. The etched cladding FBG is placed in a Si V-groove that serves the multiple functions of holding the FBG during etching and evaporation, and also provides a simple and compact means for scaling up to arrays of tunable FBGs.

To keep pace with the increasing demand for higher throughput, lower cost per unit and tighter specifications, manufacturers of fiber optic devices are now looking towards a new generation of automated alignment tools. The ideal alignment tool has six degrees-of-freedom (DOF), (X,Y,Z, Yaw, Pitch, Roll) repeatability better than 50 nanometers, travel greater than 10 millimeters and is fully automated. In this paper we describe the use of INCHWORM motor technology to produce a new nano-robotic system that demonstrates a major advancement toward the ideal photonics alignment tool. The INCHWORM actuator is uniquely suited to provide nanometer resolution movements over tens of millimeters of range with very high stiffness and stability. The clamp-extend-clamp-retract stepping sequence produces direct linear motion with no backlash. INCHWORM motors are integrated in cross roller bearing stages to achieve 20 nanometer and 0.1 arc second closed-loop resolution. The high stiffness and stability of the solid-state piezoelectric actuators hold position to a single count on a glass scale encoder while generating zero heat. Mounting fixtures hold optical elements so that their geometric centers coincide with the virtual point of rotation. Active alignment processes for selected photonics components, as well as intensity maps of components are presented.

In high volume manufacturing of optoelectronic devices in an industrial setting, automation is the key to success. The devices need to be fundamentally designed with automated manufacturing in mind, otherwise they will not render themselves to automated packaging processes. In addition, it is desirable that the manufacturing machines used to process and assemble the parts to be fully automated turnkey systems that can function with minimal operator interaction. However, other key elements that bring the machine and the product together are the toolings and the fixtures used to hold, manipulate and shape the products. This paper reviews the various issues and design considerations associated with machine tooling as applicable to high volume optoelectronic device production. Recent innovations in this field are discussed. In addition we present and analyze some examples of state-of-the-art machine toolings used in automated photonics assembly lines as well as automated test and measurement device platforms.

We present a novel design of a microoptical fiberoptic 2 x 2 and 1 x 2 switch. The fiber input and output ports are realized with the use of silicon V-grove fiber arrays. The input beams are collimated by a microlens array consisting of two lenslets, separated by the same pitch as the fibers. These two collimated beams which are extremely parallel to each other, are focused into one spot by a simple plan- convex lens. The focal length of the microlenses and this plano-convex lens determine the magnification of the mode field diameter of the fiber. Switching is performed by using a special two-sided high reflectivity mirror placed in the focal plane of the plano-convex lens, with the use of a high speed piezoelectric actuator. For the microlens array- fiber array mounting process a special setup has been built up, allowing for semi-automatic alignment. The assembly technology for all the single components and modules is described, gluing is used as the main fixing techniques. First prototypes show excellent optical parameters (1.5 dB insertion loss, -70 dB crosstalk) and very short switching time of 0.3 ms.

Agilent Laboratories, in a DARPA supported effort, has been developing a series of low-cost VCSEL based, high- performance, surface mount, parallel channel optical communications modules. Separate transmit and receive modules with 12 channels operating at 1.25 and 2.5 GBd per channel have been demonstrated with a datacom application target. Current designs have Vertical Cavity Surface Emitting Lasers (VCSELs) located in tandem with a driver IC on a copper base, which performs both a mechanical and thermal function. A flex circuit substrate laminated on the base provides signal interconnects. As channel speeds increase, the need for co-locating the VCSEL array and its associated output drivers to minimize inductance becomes more important. One solution is to piggyback the VCSELs on the transmitter IC. However, the design is constrained by the need to maintain the VCSEL junction temperature at an acceptable level. Finite element thermal modeling of the tandem mounted and piggybacked VCSELs are discussed, with a focus on comparison of the two schemes at current and projected future power levels, the effect of wirebonds on IC to VCSEL heat transfer, and performance issues related to temperature. Thermal imaging data are used to validate models. The overall module packaging scheme is also discussed.