This PDF file contains the front matter associated with SPIE Proceedings Volume 7339, including the Title Page, Copyright information, Tabe of Contents, Introduction (if any), and the Conference Committee listing.

Mode-locked lasers have applications in signal processing and communications such as analog to digital conversion, arbitrary waveform generation and wavelength division multiplexing. For such applications low noise and phase coherent frequency stabilized optical combs are needed. In this work we report a low noise, Pound-Drever Hall frequency stabilized, semiconductor mode-locked laser at 10.287GHz centered at 1550nm with 1000-Finesse sealed, ultralow insertion loss intracavity etalon. The output optical power of the mode locked laser is ~5mW.

Harmonically mode-locked semiconductor lasers with external ring cavities offer high repetition rate pulse trains while maintaining low optical linewidth via long cavity storage times. Continuous wave (CW) injection locking further reduces linewidth and stabilizes the optical frequencies. The output can be stabilized long-term with the help of a modified Pound-Drever-Hall feedback loop. Optical sidemode suppression of 36 dB has been shown, as well as RF supermode noise suppression of 14 dB for longer than 1 hour. In addition to the injection locking of harmonically mode-locked lasers requiring an external frequency source, recent work shows the viability of the injection locking technique for regeneratively mode-locked lasers, or Coupled Opto-Electronic Oscillators (COEO).

In this work we present a method for improving the uniformity of the optical spectrum or the temporal intensity profile of a quasi-CW, linearly chirped laser source covering the time interval between subsequent pulses. A novel laser cavity design, referred to as the Theta (Θ) cavity, provides linearly chirped pulses directly from the laser oscillator that having non-uniform optical spectrum, that is mapped into the temporal intensity profile of the pulse, due to the frequency-to-time mapping nature of this cavity design. The system developed in this work has been designed to improve the spectral and temporal intensity profile of lasers for photonic signal processing. A fiberized feed-forward system is implemented to reduce variations in the temporal intensity profile, or the optical spectrum due to the time-to-frequency mapping, input to the system. In the feedforward scheme presented, the quasi-CW pulse train generated from the laser is split and part of it is photodetected, while the electrical signal generated alters the transmittance of the second part of the input as it goes through an amplitude modulator, resulting in increase in the uniformity of the signal. The contrast of the optical spectrum of a chirped pulse at input to the system is improved from 51% to 16%, or 3.1 times.

Time is the most precisely measured physical quantity. Such precision is achieved by optically probing hyperfine atomic transitions. These high Q-factor resonances demonstrate frequency instability of ~10^-18 over 1 s observation time. Conversion of such a stable optical clock signal to an electrical clock through photodetection introduces additional phase noise, thereby resulting in a significant degradation in the frequency stability. This excess phase noise is primarily caused by the conversion of optical intensity noise into electrical phase noise by the phase non-linearity of the photodetector, characterized by its power-to-phase conversion factor. It is necessary to minimize this phase nonlinearity in order to develop the next generation of ultra-high precision electronic clocks. Reduction in excess phase noise must be achieved while ensuring a large output RF signal generated by the photodetector. The phase linearity in traditional system designs that employ a photoreceiver, namely a photodiode followed by a microwave amplifier, is limited by the phase non-linearity of the amplifier. Utilizing high-power handling photodiodes eliminates the need of microwave amplifiers. In this work, we present InGaAs p-i-n photodiodes that display a power-to-phase conversion factor <6 rad/W at a peak-to-peak RF output amplitude of 2 V. In comparison, the photodiode coupled to a transimpedance amplifier demonstrates >44 rad/W at a peak-to-peak RF output amplitude of 0.5 V. These results are supported by impulse response measurements at 1550 nm wavelength at 1 GHz repetition rate. These photodiodes are suitable of applications such as optical clock distribution networks, photonic analog-to-digital converters, and phased array radars.

Avalanche Photodiodes (APDs) are widely used in fiber-optic communications as well as imaging and sensing applications where high sensitivities are needed. Traditional InP-based APD receivers typically offer a 10 dB improvement in sensitivity up to 10 Gb/s when compared to standard p-i-n based detector counterparts. As the data rates increase, however, a limited gain-bandwidth product (~100GHz) results in degraded receiver sensitivity. An increasing amount of research is now focusing on alternative multiplication materials for APDs to overcome this limitation, and one of the most promising is silicon. The difficulty in realizing a silicon-based APD device at near infrared wavelengths is that a compatible absorbing material is difficult to find. Research on germanium-on-silicon p-i-n detectors has shown acceptable responsivity at wavelengths as long as 1550 nm, and this work extends the approach to the more complicated APD structure. We are reporting here a germanium-on-silicon Separate Absorption Charge and Multiplication (SACM) APD which operates at 1310 nm, with a responsivity of 0.55A/W at unity gain with long dark current densities. The measured gain bandwidth product of this device is much higher than that of a typical III-V APD. Other device performances, like reliability, sensitivity and thermal stability, will also be discussed in this talk. This basic demonstration of a new silicon photonic device is an important step towards practical APD devices operating at 40 Gb/s, as well as for new applications which require low cost, high volume receivers with high sensitivity such as imaging and sensing.

OptiComp's WDM optoelectronic transceiver module includes a transmitter optical sub-assembly that is comprised of an array of WDM VCSELs directly coupled into an optical multiplexer, the receiver optical sub-assembly which optically demultiplexes the incoming light for photodetection by PIN photodiodes, and an embedded digital diagnostic monitoring interface that enables real-time control of the transceiver and monitoring of the operation status. The transceiver module package offers scalable, high-speed, high-density interconnects for demanding aerospace and space applications. The highly integrated device designs coupled with the modular approach reduces assembly time, increases yield and offers scalability in upgrading individual transceiver components at both the device-level and sub-assembly level without requiring a change in device design, fabrication process, or manufacturing qualifications.

In this paper we overview the statistical method to three-dimensionally recognize very small number of photon-counted objects by using integral imaging (II). A conventional Poisson probability density function is assumed for modeling the distribution of very small number of photons count observation. For three-dimensional (3D) recognition of the small number of photon counted images, the photon limited elemental images set of a 3D object is obtained using the II technique. Then, the virtual geometrical ray propagation algorithm and the parametric maximum likelihood estimator are applied to the photon limited elemental image set in order to reconstruct the irradiance of the original 3D scene pixels. The sampling distributions for the statistical parameters of the reconstructed image are determined. Finally, hypothesis testing for the equality of the statistical parameters between reference and input data sets is performed for statistical classification of populations on the basis of sampling distribution information. Kolmogorov-Smirnov test is conducted and statistical p-value is measured. It is shown in experiments that very small number of photons counted image can be recognizable by the integral imaging and statistical sampling methods.

An integrated 2 x 2 multimode interference switching device was fabricated with InAs/In_0.15Ga_0.85As quantum dots as the active medium. The device, when probed with a 1.31 μm wavelength laser beam, showed similar responses for TE and TM polarization with initial power splitting ratios of 1:29 (TE) and 1:52 (TM) that were continuously adjustable to 49:1 (TE) and 38:1 (TM) when a change in current of 24 mA was applied through one of the electrodes. This is equivalent to achieving channel-to-channel crosstalk values of better than -15 dB for both polarizations. A 50:50 split ratio was reached at a current of 17 mA. We also present the preliminary results from an integrated variable power splitter that is based on a half-length multimode interference structure.

An all fiber tunable Lyot filter is experimentally demonstrated. The filter can be configured for narrow band or wide band tuning. It is implemented as a Sagnac loop configuration comprised of polarization controllers, birefringent fibers, and a polarizer. Results are present for band stop tuning and narrow band tuning for the wavelengths ranging between 1500nm to 1600nm.

Proposed is a Variable Fiber Optical Attenuator (VFOA) using an electronically controlled, variable focus liquid lens. The demonstrated experiment for the VFOA is shown for operation over the communication C-Band (1530nm- 1560nm).

We report on the optofluidic tuning of MMI-based bandpass filters. It is well known that MMI devices exhibit their highest sensitivity when their diameter (D) is modified, since they have a D² wavelength dependence. In order to increase the MMF diameter we use a special fiber, called No-Core fiber, which is basically a MMF with a diameter of 125 μm with air as the cover. Therefore, when this No-Core fiber is immersed in liquids with different refractive indexes, as a result of the Goes-Hänchen shift the effective width (fundamental mode width) of the No-Core fiber is increased, and thus the peak wavelength is tuned. A tunability of almost 40 nm in going from air (n=1.333) to ethylene glycol (n=1.434) was easily obtained, with a minimum change in peak transmission, contrast, and bandwidth. Moreover, since replacing the entire liquid can be difficult, the device was placed vertically and the liquid was covering the No-Core fiber in small steps. This provided similar amount of tuning as before, but a more controllable tuning mechanism.

Liquid crystal spatial light modulators have important applications in nonmechanical beam steering. In this application, a suitable T/R switch that separates the transmitter and receiver of the system is required. Since LC devices require a certain linear polarization state incident upon it in order to steer an incoming laser beam, previous T/R switches through the use of the combination of a circular polarization state and a polarization beam splitter proved to be a lossy solution. Through the use of a two dimensional reflection type liquid crystal spatial light modulator and a large clear aperture Faraday rotator, it is possible to develop a fairly compact and efficient nonmechanical beam steering device, which can be utilized in laser scanning imaging systems. In this paper, we will present the design of such a T/R switch. Experimental results at a 1.06 μm operation wavelength will be demonstrated.

Show that capturing more information from an HDR scene using an ND filter pattern on a 1280 x 1024 pixel 8 bit monochromatic CCD sensor is feasible and compelling. Simulation is performed on appropriate HDR imagery using Matlab. An example HDR scene is selected consisting of a lit hallway with a doorway into a darkened room. The output consists of fused images with a simple digital cut and paste approach. A decision of continuing the masked sensor effort is made based on evaluation of the output scenes. The next steps would be further simulation as well as obtaining evaluation hardware.

A 1.28GSPS 12-bits optoelectronic analog-to-digital converter is presented. The ADC architecture consists of an optical circuit that optically samples an analog input signal, and optoelectronic circuits that demultiplex different phases of the sampled signal (polyphase) to yield low data rate for electronic quantization. Electrical-in to electrical-out data format is maintained through the sampling, demultiplexing and quantization stages of the architecture thereby avoiding the need for electrical-to-optical and optical-to-electrical signal conversions. The ADC architecture encodes and multiplexes four 320MSPS 12-bits time interleaves quantized data into an aggregated 1.28GSPS 12-bits digital signal in real time. All clock signals at frequencies of 1.28GHz, 640MHz, and 320MHz are generated from a single 320MHz femtosecond laser source, thus eliminating the need for external synchronization and control signals. A Spurious Free Dynamic Range (SFDR) of value 80.6dB, and Effective number of bits (ENOB) of value 8.8 bits, were measured for the system.

Optical implementation of RF filtering provides processing and switching advantages for analog optical links, bringing additional versatility to the field of RF photonics. Many RF photonic notch filter designs have demonstrated this increased link functionality, but most techniques are sensitive to polarization and phase noise, and are tunable over limited ranges, set by the tunability limits of conventional link components such as DFB lasers. We demonstrate an RF photonic notch filter with extremely wide tunability, high extinction ratio, and large range of useable bandwidths as part of a high dynamic range RF photonic link. We discuss methods for improved filter operation, present experimental results for X-band operation of the filter, and describe its use in applications such as co-site interference mitigation.

A chirped fiber Bragg grating with a dispersion of 1651ps/nm is used to generate temporally stretched, frequency chirped pulses from a passively mode locked fiber laser that generates pulses of ~1ps (FWHM) duration at a repetition rate of 20MHz with 3.5mW average power (peak power of 175W). The use of a chirped fiber Bragg grating enables the generation of temporally stretched pulses with low peak power so that non-linear effects in the fiber can be avoided. A fiber based interferometeric arrangement is used for interfering a reference signal with the reflected signal from the target to realize a coherent heterodyne detection scheme. In the RF domain, the detected heterodyne beat frequency shifts as the target distance is changed. A round trip target distance of 14km in air is simulated using 9.3km of optical fiber and a resolution of less than a millimeter is observed.

We perform numerical calculations for the diffraction efficiency of single-sided microprism designs with optical beams spread over several microprisms. For small angles (less than 1.5 degrees) the far-field diffraction efficiency exceeds 90% and the far-field angles can be extended to more than 10 degrees with better than 80% diffraction efficiency. For the microprism array the diffraction angles are discrete and the angles between the diffraction peaks can be covered by applying a tilted phase to the input field.

It is well-known that interconnect issues pose a significant bottleneck with regard to improving the performance of high-speed integrated systems such as a cluster of computer processing units. Power, speed (bandwidth), and size all affect the computational performance and capabilities of future systems. High-speed optical processing has been looked to as a means for eliminating this interconnect bottleneck. Presented here are the results of a study for a novel optical (integrated photonic) processor which would allow for a high-speed, secure means for arbitrarily addressing a multiprocessor system. This paper will present analysis, simulation, and optimization results for the architecture as well as considerations for a proof-of-concept level system design. The architecture takes advantage of spatial and wavelength diversity and in this regard may be regarded as a Multiple Input Multiple Output (MIMO) architecture. A given node to be addressed, rather than having a wired metal contact as an output, has as a radiating laser source that has been modulated with the data to be conveyed to another point in the system. Each processor node radiates a different optical wavelength. Each individual wavelength is chosen, for example, to correspond to the wavelengths associated with a WDM ITU grid. All wavelengths are incident on a coherent fiber bundle which acts as an array receiver. Unlike conventional phased arrays, the receive elements are spaced many wavelengths apart giving rise to a large number of grating lobes. It will be shown that by using appropriate photonic/optical signal processing methods any node of the processor cluster can be randomly and rapidly addressed using high-speed phase shifters (electrooptic or others) as control elements. The diversity techniques employed achieve high gain and a narrow beamwidth in the direction of the desired node and high attenuation with regard to the signals from all other nodes. As is often the case of MIMO-bases systems, overall performance exceeds that of diffraction limited array processing. In addition to the interconnect application discussed, the methods described in this paper can also be applied to other applications where rapid electrical (non-mechanical) optical beamsteering is required such as raster scanned laser radar systems and tracking, guidance, and navigation systems.

In this paper we introduce a new preemptive scheduling technique for next generation optical burst-switched networks considering the impact of cascaded wavelength conversions. It has been shown that when optical bursts are transmitted all optically from source to destination, each wavelength conversion performed along the lightpath may cause certain signal-to-noise deterioration. If the distortion of the signal quality becomes significant enough, the receiver would not be able to recover the original data. Accordingly, subject to this practical impediment, we improve a recently proposed fair channel scheduling algorithm to deal with the fairness problem and aim at burst loss reduction simultaneously in optical burst switching. In our scheme, the dynamic priority associated with each burst is based on a constraint threshold and the number of already conducted wavelength conversions among other factors for this burst. When contention occurs, a new arriving superior burst may preempt another scheduled one according to their priorities. Extensive simulation results have shown that the proposed scheme further improves fairness and achieves burst loss reduction as well.

Spatial Domain Multiplexing (SDM) is a novel technique that allows co-propagation of two or more optical communication channels of the same wavelengths over a single strand of optical fiber cable by maintaining spatial separation between the channels. Spatial multiplexer known as the beam combiner module (BCM) supports helical propagation of light to ensure spatial separation between the channels. It is inserted at the input end of the system. Spatial de-multiplexing is achieved by a unit named beam separator module (BSM). This unit is inserted at the receiving end of the system and it routes the optical energy from individual channels to dedicated receivers. Spatially multiplexed channels exhibit negligible crosstalk. The bandwidth of the fiber optic systems employing SDM technique increases by multiple folds. CAD model of a beam combiner module for a two channel system using commercially available simulation tools is presented here. Simulated beam profile of the output is compared to the experimental data.

In this paper, an optical signal infrastructure using a novel simulation framework is presented for self-consistent optoelectronic circuits and systems. This framework uses a formulation based on modified nodal analysis and can be used for transient and small-signal analysis. A flexible representation of optical signals and elements is developed that is appropriate for circuits/systems which incorporate both electrical and optical devices. With the correct choice of optical state variables it is found that optical interference, reflection and coupling can be modeled efficiently. Optical models for multi-mode fibers, optical connectors and cross-couplers are presented as examples of model development within the framework. To illustrate the use of the framework, results from a number of optoelectronic circuits are presented. These examples include optical links involving lasers, multi-mode fibers, optical connectors and photodiodes. Results from these examples highlight the ability of the framework to handle a wide variety of optical effects and to simulate mixed electrical/optical circuits.

We report on a novel tuning mechanism to fabricate an all-fiber tunable laser based on multimode interference (MMI) effects. It is well known that the wavelength response of MMI devices exhibits a linear dependence when the length of the multimode fiber (MMF) section. Therefore, tuning in the MMI filter is achieved using a ferrule (capillary tube of 127 μm diameter) filled with a liquid with a higher refractive index than that of the ferrule, which creates a variable liquid MMF. This liquid MMF is used to increase the effective length of the MMI filter and tuning takes place. Using this simple scheme, a tuning range of 30 nm was easily achieved, with very small insertion losses. The filter was tested within a typical Erbium doped fiber (EDF) ring laser cavity, and a tunable EDF laser covering the full C-band was demonstrated. The advantage of our laser is of course the simplicity of the tunable MMI filter, which results in an inexpensive tunable fiber laser.

Various oxidation states and location sites of chromium (Cr) in doped crystal have been known. Since the high sensitivity of ligand field on chromium ions, these are attributable to a broadband emission. The wide broadband emission can potentially be exploited in many ways. For instance, a Cr-doped broadband amplified spontaneous emission can improve spatial resolution many times in optical coherence tomography (OCT) diagnostic instrumentation. It is difficult to further extend the emission bandwidth of Cr in a single crystal, although a single crystal is generally considered as the best host. Amorphous as a host, on the other hand, may potentially expand the bandwidth of emission and accommodate higher doping concentration. However, a single crystal is beneficial to a specific valance state through charge compensation technique. To investigate chromium's full potential, both Cr-doped nano-crystalline embedded in an amorphous host and Cr-doped glasses are proposed. The ultimate aims are capable of tailoring oxidation states, site symmetry and concentrations of Cr doping. In the first stage, Cr:YAG-doped silica fiber will be studied as a part of our research scope.

Nitride based photocathodes for image intensifiers are of interest because of the wide span of wavelengths covered by the bandgap of the AlGaInN alloy system. The potential bandgap range for this alloy system is from 6.2 eV for AlN to 0.7 eV for InN. Coupled with microchannel plate technology, this alloy system potentially offers low noise and high gain image intensifiers over a wide wavelength range. Results from L-3 EOS work in this area are presented beginning with a brief summary of unpublished early work carried out from 1992 - 1997 on AlGaN image intensifiers. The early work wrestled with the dual issues of sealing image intensifiers along with improving the quality of the AlGaN epitaxy layer. This is followed by our current results on a GaN image intensifier sealed with a photocathode from SVTA. Imagery using 375nm LED illumination is shown. The quantum efficiency at 300nm was estimated to be 16% measured in transmission mode. This QE was achieved with a 0.15μm thick Mg doped GaN active layer.