We review a number of optical devices made from microporous silicon. Particular emphasis is placed on the fabrication method of porous silicon laser-mirrors, optical microcavities and one-dimensional photonic crystals with true photonic bandgap.

Micro-ring resonators, by virtue of the geometrical birefringence of the strongly-confined waveguides used to construct them, are polarization dependent, and thus of limited application in fiber networks unless integrated with a laser or polarization controller. This paper addresses the polarization sensitivity issue by proposing a novel design of an MMI-coupled resonator with substantially reduced polarization sensitivity, while maintaining single-mode and low-loss conditions. The polarization-independent, single-mode waveguide is obtained by using the proper combination of ridge width and etching depth. The design is applicable to relatively large resonators, and may provide an intrinsic solution to small-FSR (free spectral range) polarization-independent applications.

A novel tunable optical filter structure based on an Opto-VLSI processor is proposed in this paper. The architecture is capable of dynamically tuning multiple pass-bands through reconfiguration of the size and shape of holographic diffractive gratings generated by the Opto-VLSI processor. Results for an experimental 3-passband tunable filter are presented confirming over 25dB of dynamic range and passband bandwidth of 2 nm.

Ridge waveguides with smooth and vertical sidewalls are essential in photonic circuits. We have investigated waveguide realization with reactive ion etching of InP and InP-based structures using a SiN_x in a Cl₂/H₂/CH₄ chemistry in an ICP plasma. Depending on ICP power and RF power, etching rates can be obtained from 200 nm/min up to > 2μm/min. A maximum etching selectivity of InP vs SiNx of 12 was obtained at 2000 W of ICP power. Deep etched waveguides, fabricated in an InP/InGaAsP double heterostructure, show typical losses of 2 dB/cm. This low value shows the potential of ICP technique in the fabrication of photonic circuits.

Features such as large bandwidth, low drive voltage, compact size, and feasibility for monolithic laser integration make electro-absorption modulators (EAM) attractive candidates for ultra high-speed fiber-optical time division multiplexing (TDM). EAM with traveling-wave (TW) electrodes have successfully been demonstrated as a way to considerably increase the modulation bandwidth without sacrificing modulation efficiency. However, for reasonable modulation efficiency a low characteristic impedance (≈25Ω) has to be accepted. Termination with a matched load is necessary to benefit from the TW configuration. Thus, TWEAM with continuous impedance-matched transmission lines (TML) provide very high bandwidths, but suffer from high electrical return loss when using a 50Ω driver. A solution to this problem is to split up the modulator and insert passive TML segments between the active parts. The passive segments can be designed for a higher characteristic impedance than the active modulator parts with their inherently low impedance. In this way, the impedance seen at the electrical modulator input can be tailored for values that deliver optimum performance in combination with the available driving electronics (usually 50Ω). Only little bandwidth is sacrificed with the segmented design. Recently, we have demonstrated state-of-the-art performance of segmented TWEAM. These devices exhibit low electrical return loss (<-15dB) and a flat small signal modulation response in the characterized frequency range of 0.04 to 50GHz. 50Gb/s operation is demonstrated. An extinction of 10dB with 3Vp-p is achieved at 40Gb/s.

In this article photonic implementations of two oversampled analog-to-digital converter architectures are discussed. The first, and simplest design is that of pulse-code-modulation. Simulations and an experiment are developed employing a Multiple-Quantum-Well p-i-n diode comparator. Agreement between simulation and experiment is demonstrated and the design is subsequently extended to a higher performance first order unipolar sigma-delta architecture. Results of the signal-to-noise ratio as a function of input amplitude are presented for an oversampling ratio of 100.

This paper describes a study conducted into the limit on spectral resolution due to the dynamic range of a T-ray spectrometer. The pulsed nature of terahertz time-domain spectroscopy (THz-TDS) sets a fundamental limit on its spectral resolution. The spectral resolution of THz-TDS can be improved by increasing the duration of the temporal measurement, but is limited by the dynamic range of the system in the time-domain. This paper presents calculations and experimental results relating the temporal dynamic range of a THz-TDS system to its spectral resolution. We discuss three typical pulsed terahertz sources in terms of their dynamic range and hence achievable spectral resolution.

A broadband width semiconductor optical amplifier is reported using an asymmetric quantum well structure. By applying tensile strain to the quantum wells it is possible to both increase the optical bandwidth while reducing the polarization sensitivity. The effects of the ordering of the asymmetric quantum wells, length of the device and carrier density all affect the performance of the device. Using the asymmetric struture, a 90 nm region of polarization sensitivity less than 1.2 dB within the -3dB band width is achieved.

MicroPhotonic broadband RF signal processors utilize the capability of photons to perform true-time delay processing at very low loss that is unattainable by conventional electronic methods. In this paper, we present a novel MicroPhotonic interference mitigation filter architecture that utilises a CMOS Si photoreceiver/VCSEL array in conjunction with a true-time-delay multi-cavity optical substrate to realise an adaptive transversal RF processor with arbitrary response. Results show that the proposed MicroPhotonic structure can synthesize adaptive interference mitigation with a shape factor (ratio of the -40dB bandwidth to the -3dB bandwidth) as low as 2 and passband ripples less than 0.25dB.

We report on the fabrication and characterization of 2D photonic crystals (PhCs) in InP/InGaAsP/InP heterostructures. It is demonstrated that Ar/Cl2 based chemically assisted ion beam etching (CAIBE) is a very promising method to obtain high aspect ratio etching of PhCs in the InP-based materials. With this process, it is possible to obtain PC-holes as deep as 3 microns even for feature (PhC-hole) sizes as small as 200-250 nm. The optical characteristic of the fabricated PhC-based elements/devices such as line-defect waveguides, in-plane resonant cavities and drop-filter based on contra-directional coupling will be reported. The devices were measured using end-fire coupling and the obtained results were simulated using the 2D finite difference time domain (FDTD) method including an effective loss-approximation. The etched PhC-waveguides show low transmission losses, less than 1 dB/100 μm. A quality factor of 400 for a 6 micron long cavity with 6-hole mirrors is obtained. Finally, drop-functionality in a PhC-based filter using contra-directional coupling is demonstrated.

The bandgap of indium nitride has long been accepted to be 1.9 eV. However, recent results have cast doubt over this as modern epitaxy techniques have allowed experimental studies of high quality material. Single crystalline and polycrystalline indium nitride films have been grown on (0001) sapphire and silica glass using plasma assisted molecular beam epitaxy (PAMBE). Optical measurements on the films revealed a luminescence feature in the vicinity of 0.8 eV for all films, both on sapphire and glass. No feature around 1.9 eV could be identified above the background noise. To our knowledge this is the first report of polycrystalline InN exhibiting the 0.8 eV feature. Ion beam analysis of the material could find no measurable oxygen contamination in the bulk of the films. These results along with recent reports of blue shifting of the absorption onset of InN films with increasing oxygen content appear to point towards oxygen contamination as being the source of the previously reported higher bandgap. Like other groups we observed a small anomalous blue shifting of the luminescence with increasing temperature when using a germanium detector. We have verified that this is a real feature by measuring the temperature dependent PL with a lead sulphide detector. Two distinct growth regimes were identified. High In:N flux ratios lead to spotty RHEED accompanied by a morphology of flat plateaus separated by narrow valleys. Low In:N flux ratios lead to rough films consisting of facets largely disjoint from each other. Surprisingly, this regime gave streaky RHEED, suggesting high levels of crystal alignment between facets and high crystal quality within facets.

The combinatorial material chip method has been used to study the emission efficiency of InAs/GaAs quantum dots. The photoluminescence spectroscopy is performed to obtain the rule of emission efficiency on the proton implantation dose. A pronounced enhancement of room temperature emission efficiency has been obtained by the optimized quantum dots process condition. The increment of emission efficiency up to 80 itmes has been observed. This effect may be resulted from both the proton passivation and carrier capture enhancement effects. The maximum photoluminescence peak shift is about 23 meV resulted from the intermixing of quantum dots. A linear dependence behavior has been observed for both the non-radiative recombination time and carrier relaxation time on the ion-implantation dose. The maximum enhancement of the photoluminescence is observed in the proton implantation dose of 1.0 x 10¹⁴ cm^-2 followed by rapid thermal annealing at 700°C. These effects will be useful for the QDs' optoelectronic devices.

Thermal characteristics of heavy-boron-doped Si resistor micro-bridge in infrared (IR) emitters, or so-called IR image simulators, have been studied with the micro-Raman scattering measurement. By illuminating the Si bridge using microscope-objective-focused intensive laser power and taking the advantage of the suspension nature of the Si bridge, the shone spot has been heated up, resulting in changes in Raman spectrum. By taking into account of Fano interference between the inter-valence-band transitions and the optical phonons, the line-shape of the recorded Si Raman spectrum has been analyzed, yielding the Raman peak position and the intensity ratio of the Stokes to anti-Stokes scatterings. The temperature of the measured point has been calculated using these parameters. Line-scanning Raman measurement along some typical directions and Raman mapping over the complete surface of the Si bridge have been performed and the temperature of each measured point has been determined. Because the boundary conditions are different at different places, the same laser illumination leads to different elevated temperature, revealing the heat conduction capabilities at different parts of the Si bridge. The temperature distribution over the Si bridge has been schematically displayed. A finite element simulation analysis has also been carried out and compared with the experimental data. While the thermal characteristics concluded by the simulation are symmetric and uniform, the experimental results, in addition to the agreement to the calculated ones in general, give some case-dependent information, which is more important to reflect the features of actual devices and provides the basis for device design and optimization.

A new bulk micromachined Fabry-Perot modulator fabricated using (110) silicon is presented. The modulator has been developed to reduce the alignment and packaging restrictions imposed on surface micromachined Fabry-Perot modulators. The Fabry-Perot modulator consists of two vertical micromachined cantilever mirrors. The modulator is capable of a modulation depth of 7.8dB for modulation frequencies of up to 100kHz making it suitable for multiplexing low bandwidth sensors. Analytical results of the performance of the modulator compared to current microelectromechanical systems (MEMS) modulators are given and experimental results of the fabrication process indicate the practical realisation of the modulator. The design aspects of the modulator are analysed including the trade-off between bandwidth, beam length and drive voltage.

Wavelength Division Multiplexing has become the leading technology for optical transmission systems which operate at 1550 nm. One of the key components of such systems are tunable and wavelength selective receivers. In this paper we present a fibre-coupled two-chip receiver front end, which is highly wavelength selective and tunable over a wide wavelength range. The device is a bulk-micromachined Fabry-Perot pin-photodiode, which features a high finesse of more than 220 with a sufficient tuning range (> 40 nm) to cover wide wavelength region. The bandwidth (full-width half maximum) of the device is < 0.2 nm (25 GHz). The photocurrent crosstalk from an adjacent channel (100 GHz spaced apart) is below -30 dB. The wavelength tuning is achieved by a change in the resonator length, formed by the two chips. This is realized by current induced thermal heating on top of the membrane mirror suspensions, which deflects the membrane. The optical-electrical conversion takes place in the pin-photodiode. This integration reduces the need for any additional components. Fiber-coupling is achieved with a fiber-coupled lens that tailors the Gaussian beam to match with the Fabry-Perot cavity. The alignment process of the two-chip structure, forming the wavelength selective cavity, has been simplified to the point where a simple place-and-fix strategy can be applied.

Current infra-red detectors are limited to detect broad windows in transmission. By adding Fabry-Perot filtering to these detectors multi- and hyper-spectral detectors can be fabricated. However, filtering will reduce the signal available to the detector. In order to decrease the noise (thereby increasing the signal to noise ratio), the detector can be moved into the resonant cavity of the filter. The design of the mirrors is changed by placing the detector with the resonant cavity. Materials for the design of a resonant cavity enhanced mercury cadmium telluride detector are investigated in this paper.

Light-current, relative intensity noise and spectral characteristics of a group of gain-guided InGaAsP/InP MQW buried heterostructure laser diodes have been investigated under accelerated aging conditions. The operating current, I_op, at constant output optical power increases logarithmically with time in stable devices, which indicates that the life expectancy of the lasers exceeds 2x10⁵ hours or greater than 20 years. The emitted light wavelength at constant output power shifts by 2-2.5Å during 3000 hours of overstressing, mainly due to the increase in I_op. High-frequency RF signal-induced transient chirp for relative stable LDs at constant output power shows very little change with time.

We report the results of an effort to understand the effect of surface electronic structure of indium tin oxide (ITO) on luminance efficiency of organic light-emitting devices (OLED)s. Nitric oxide (NO) plasma was used to modify the ITO. NO plasma induced an increase in the sheet resistance of ITO. The surface electronic structure of ITO was studied using X-ray photoelectron spectroscopy. An approximately 4-nm thick low conductivity layer with a production of N-O type species was formed near the ITO surface region. It is demonstrated that the barrier for hole-injection from an ITO anode to a hole transporting layer can be engineered by NO plasma treatment. The increase in luminance efficiency of the OLEDs reflects an improved current balance in the device.

Fabrication processes such as laser micromachining and laser based photolithography are well established within the automotive, electronics and aeronautic industries, however, except in a handful of cases, laser based micro-processing is infrequently used in the fabrication of photonic devices. In this paper novel laser assisted methods for rapid prototyping of photonic devices such as periodically poled lithium niobate and 2-D photonic crystal structures are reviewed.

We investigate the patterning-accuracy limits of proximity-effect corrected (PEC) electron-beam lithography applied to the fabrication of photonic crystals (PhC's). Energy-intensity distribution simulations reveal that conventional dose-modulation PEC techniques present a lower limit of the best attainable hole-radius variation of approximately 1% for a generic PhC structure, while a PEC method proposed by Watson theoretically should yield perfect correction. Simulation results were verified experimentally and additionally we introduce a new method to determine the beam-broadening parameter α. We analyzed the impact of geometrical key parameters of PhC's on achievable patterning accuracy and showed that proximity effects impose severe limitations on the patterning of structures with large filling factors and/or small lattice constants. Furthermore, we performed a sensitivity analysis on the proximity parameters and showed that overestimation of the backscatter efficiency can actually improve the lithographic accuracy and mimic the Watson-PEC method to a certain degree.

The surface relief phenomenon is one of the interesting properties of shape memory alloys. In this paper, this phenomenon in a NiTi shape memory alloy rod is investigated quantitatively using a temperature controllable atomic force microscope. The surface reflection, which depends heavily on the surface roughness, is studies quantitatively with different incident angles and incident directions. It is found that the incident angle α has a slight effect on the distribution of θ, which is the angle between the real reflective direction and the ideal one. The material is almost uniform in any direction, despite that only one incident angle and four directions have been analyzed. At room temperature (martensite phase), θ mainly falls to around 3°. While at 100°C (austenite phase), a large portion of θ shifts to around 20°. In order to obtain a ratio of change over 0.5, θ is expected to be at around 3°.

For the waveguide having a pillar-missing-line defect in a photonic crystal (PhC) with dielectric pillars in air, a pillar-missing-line defect can support the low-loss transmission when the mode frequency falls inside the photonic bandgap (PBG). In this paper, the waveguide with the single-missing-hole-line defect in the silicon on insulator (SOI) PhC slab, which has square air holes in silicon on a square array, is studied by the FDTD simulation. It seems that even in this structure, a high transmission coefficient is obtained in the bandgap-guided mode. However, a low-loss transmission was observed at outside the PBG. This is the propagation of the index-guided mode based on the refractive index of the core being larger than average index of the cladding medium. Appropriately, choosing the design parameters (the silicon core is 1000-nm thick, lattice constant is 500-nm, and air hole size is 250-nm square), we find that the present PhC has a TE gap ranging from 95-THz to 135-THz. These values approximately agree with the 2D theoretical gap ranging from 92-THz to 133-THz because the silicon layer is thick. The low-loss transmissions are obtained at the frequencies outside the PBG in the longitudinal and transverse directions except for the low-loss transmission frequency in the oblique direction on the original PhC. This is the propagation of the index-guided mode. Since the omni-directional bandgap of PhC with a square lattice is generally smaller than that for a triangular lattice, it is considered that PhC with the square lattice is unsuitable to forming sharp waveguide bends. However, in this paper, we propose the fundamental structure of a ring resonator with sharp U-turns such that the possibility of obtaining a high Q value can also be expected. Assuming a 300-nm-thick silicon layer, the 420-nm lattice constant and 300-nm square air holes, we find good optical confinement near 210-THz. At the resonance of 211-THz, we obtain the Q-value of 100 from the saturated tuning curve in the limit to the Fourier transform. These facts indicate that the resonator with sharp U-turns has Q value greater than 100. Since the leakage to the lower and upper layers is not considered in the present study, the possibility of obtaining a high Q value can also be expected.

This paper reports on the scalability of MEMS optical cross-connect (OXC) switches which use 2D planar waveguide architecture and an integrated waveguide lens to enhance free space propagation. The optical loss, total device area and ease of integration with MEMS micromirrors are considered for three competing layout configurations.

We have fabricated several two-dimensional photonic-crystal (2DPC) slab waveguides by using fine EB lithography and dry etching. The 2DPC waveguides include straight, bend and directional coupler on the GaAs/AlGaAs substrate as an application to the ultra-small and ultra-fast all-optical switching device (PC-SMZ). Site-controlled InAs quantum dots (QDs) responsible for nonlinear phase shift in the PC-SMZ have been investigated by developing a nano-probe assisted in-situ process. Optical linear and nonlinear properties of stacked QDs were characterized. The result exhibited the π/2 phase shift required for the SMZ-type optical switching operation. These results are capable of achieving the PC-SMZ.

We present and characterize hexagonal point defects in a two dimensional photonic crystal based on macroporous silicon. These point defects are prepatterned periodically, forming a superstructure within the photonic crystal after electrochemical etching. Spatially resolved, optical investigations related to morphological properties, like defect concentration and pore radius, are compared to bandstructure calculations. The confined defect states are identified and their interaction is evaluated quantitatively.

Free-space optical interconnects (FSOIs) utilize arrays of vertical-cavity surface emitting lasers (VCSELs), microlenses, and photodetectors to effectively overcome the communication bottleneck caused by the poor performance of electrical interconnects. We derived a comprehensive FSOI link equation which can be used to determine the interconnect performance parameters, such as the receiver carrier-to-noise ratio. The link equation includes both optical and electrical noise components. The optical noise component is caused mainly by laser beam diffraction. We have simplified the modeling of optical noise by using the recently introduced Mode Expansion Method. The optical noise component strongly depends on the modal content of the incident VCSEL beam. The models used in the literature assume that the cross-sectional profile of the emitted laser beam resembles the fundamental Gaussian mode. Our link equation takes into account the modal structure of a multimode VCSEL beam. We have investigated the FSOI performance and we found that for each merit function there exists a single set of design parameters yielding the optimal performance. We have also found that the presence of higher-order modes negatively affects the performance. Our results show that FSOIs based on multimode VCSELs can be utilized in chip-level interconnects despite increased beam diffraction.

In this paper, a numerical study was conducted on spreading of the current in a bottom emitting Vertical Cavity Surface Emitting Laser (VCSEL) with oxidation at the substrate. It was found that the current density profiles of etched VCSELs with small active diameters (< 125 μm) are similar to unetched VCSELs with a 500 μm active diameter. Larger active diameters of 150 μm to 225 μm also have higher density profiles than unetched VCSELs. The simulated current density profiles of large p-contact diameters are dependent on the oxide aperture diameter rather than the contact diameter. For smaller p-doped contact diameters, the density profiles are dependent on the contact diameter rather than the oxide aperture diameter. From current density profiles, higher output powers in the 980 nm wavelength regime are theoretically obtainable at lower threshold currents than previously reported. Maximum output powers of 489 mW, 690 mA and 787 mA at current thresholds of 102 mA, 271 mA and 442 mA were calculated for contact diameters of 50 μm, 100 μm and 150 μm, respectively, with a 50 μm oxide aperture. Depending on the geometric ratios of the simulated devices, required high output power VCSELs can be designed for specific applications.

The morphologies and nonlinear optical properties of vanadylphthalocyanine (VOPc) thin films on polymer and KBr substrates remain unclear. This paper investigates the morphologies and nonlinear optical properties of VOPc thin films prepared on polycarbonate (PC) and KBr substrates by UV/Vis spectrum measured with UV/Vis spectroscopy, X-ray diffraction (XRD) and third-order harmonics (TH), as well as second-order harmonics (SH) detected using the Maker fringe method. The UV/Vis spectra of VOPc thin films prepared on a PC substrate with different evaporating source temperatures have a peak at 840 nm in the Q-band region. This indicates that the morphology of the VOPc thin films prepared is Phase II. Moreover, the UV/Vis spectra of each sample do not change with increasing thickness. Therefore, this also suggests that interaction between a VOPc molecule and the surface of a polycarbonate film is strong. The number of VOPc molecules oriented at 60° to a substrate markedly increases with the increase in evaporating source temperature. This indicates that there is an optimum evaporating source temperature for the orientation of VOPc thin films prepared on a PC substrate. In VOPc thin films prepared on a KBr substrate, the phase morphology of VOPc thin films changes from pseudoepitaxy to epitaxy by increasing the annealing time. Moreover, the incident angle dependence of second harmonic (SH) intensity shows a lower convex curve when irradiated with a p-polarized laser light, and the SH intensity is enhanced with the increase of the annealing time. Moreover, the incident angle dependence of third harmonic (TH) intensity shows upper convex curves and the maximum values of TH intensity are enhanced with the increase of the annealing time. These findings indicate that the orientation of VOPc thin films is improved with the increase in annealing time. It is closely related to molecular diffusion during annealing.

Absorption of electromagnetic waves in electronic systems coupled to intense terahertz waves is calculated. We formulate a theoretical framework suitable for calculating the frequency-dependent electrical current under an intense THz radiation. This first principle method is based on the time-evolution of electron density matrix and it includes electron-photon coupling to all orders. We first obtained the time-dependent electronic states as a function of terahertz field and frequency. The electron-impurity scattering is included to the second order. The absorption of electromagnetic waves of a probing field via various electron-terahertz-photon coupling is then obtained in terms of frequency-dependent dielectric functions.

Tunable vertical cavity surface emitting lasers (VCSELs) are very attractive candidates for employment in various areas of interest, like optical communications or gas sensing. During the last decade these types of components have been demonstrated. In this paper we present a micromechanically realization of an optically pumped 1.55 μm tunable VCSEL and its characteristics. The investigated device comprises two chips. The first chip, a half-cavity VCSEL chip, contains the bottom Bragg-mirror and the active layers. The second chip is a micromechanically manufactured Bragg-mirror membrane chip. After aligning, both chips form together the VCSEL cavity. Wavelength tuning is achieved by thermal actuation of the membrane mirror due to current flow and the subsequent deflection of the mirror membrane resulting in a change of the resonance wavelength. Such a micromachined two-chip VCSEL device is investigated and discussed. In particular, properties like the side mode suppression ratio, relative intensity noise and polarization during actuation and their dependence on the properties of the micromachined mirror-membrane are analyzed. Remarkable results are e.g. the side mode suppression ratio dependence on the pump spot size, the dependence of the relative intensity noise of the VCSEL on the pump laser noise, and stable polarization due to the membrane design.

With the increasing demand on the access network in the local and residential areas, there is a growing need for more scalable and dynamic optical access network architectures. In this paper, variable optical splitter is utilized in the optical access network as branching device. By changing the number of branches at the variable optical splitter or tuning the optical power distribution between these branches, a more flexible architecture can be realized. This will enable more customers to access the network flexibly and dynamically. It also has the added advantage that it can provide link protection for the network.

In this work we present detailed analysis of the emitted radiation spectrum from tris(8-hydroxyquinoline) aluminum (Alq₃) based OLEDs as a function of: the choice of cathode, the thickness of organic layers, and the position of the hole transport layer/Alq₃ interface. The calculations fully take into account dispersion in glass substrate, indium tin oxide anode, and in the organic layers, as well as the dispersion in the metal cathode. Influence of the incoherent transparent substrate (1 mm glass substrate) is also fully accounted for. Four cathode structures have been considered: Mg/Ag, Ca/Ag, LiF/Al, and Ag. For the hole transport layer, N,N'-diphenyl-N,N'-(3-methylphenyl)-1,1'-biphenyl-4,4'-diamine (TPD) was considered. As expected, emitted radiation is strongly dependent on the position of the emissive layer inside the cavity and its distance from the metal cathode. Although our optical model for an OLED does not explicitly include exciton quenching in vicinity of the metal cathode, designs placing emissive layer near the cathode are excluded to avoid unrealistic results. Guidelines for designing devices with optimum emission efficiency are presented. Finally, the optimized devices were fabricated and characterized and experimental and calculated emission spectra were compared.

In this paper we investigate for the first time the effect of the crosstalk introduced due to laser beam imaging in a free-space optical interconnect (FSOI) system. Due to the overfill of the transmitter microlens array by the vertical cavity surface emitting laser (VCSEL) beam, one part of the signal is imaged by the adjacent microlens to another channel, possibly far from the intended one. Even though this causes increase in interchannel and intersymbol interference, to our knowledge this issue has been neglected so far. The numerical simulation has been performed using a combination of exact ray tracing and the beam propagation methods. The results show that some characteristics of stray-light crosstalk are similar to that of diffraction-caused crosstalk, where it is strongly dependent on the fill factor of the microlens, array pitch, and the channel density of the system. Despite the similarities, the stray-light crosstalk does not affect by an increase in the interconnection distance. As simulation models for optical crosstalk are numerically intensive, we propose here a crosstalk behavioral model as a useful tool for optimization and design of FSOIs. We show that this simple model compares favorably with the numerical simulation models.

Photonic crystals (PCs) have been intensively investigated both theoretically and experimentally in recent years due to their possibility to manipulate and control light. A number of different methods have been proposed and demonstrated to fabricate two- or three-dimensional photonic crystal structures. Among them, the holographic lithography method, in which multi-beam interference is employed, offers a number of advantages, including its ability to create large volume of periodic structures through an irradiation process, the uniformity of period, and more degrees of freedom to control the structures. In this study, a multi-beam interference model is presented for predicting the three-dimensional photonic crystal structures. Various parameters, including beam propagation and polarization directions, beam intensities, and phase shifts are considered. Calculations have been carried out to simulate a four-beam configuration which has been popularly used in the fabrication of photonic crystals. It has been demonstrated that the contours of the interference pattern are related to the polarization states and the intensity ratios among the four beams. Therefore, by controlling the beam intensities and polarization directions, different structures can be obtained. The results presented in this study provide a useful guide for choosing various optical parameters and selecting proper photoresists to fabricate three-dimensional photonic crystals.

In a self-mixing type laser range finder the current of the laser is modulated with a triangle wave to produce a range of optical frequencies. However, the electrical signal does not produce a perfect linear sweep in optical frequency due to thermal and other effects in the laser. This leads to errors in the accuracy and resolution of the range finder. In this paper, we describe and implement a method in software to systematically determine the optimal shape of the injected waveform needed to eliminate these thermally induced measurement errors. With this method we do not require the more complicated and expensive optical techniques used by other researchers to recover the optical frequency variations with regard to injection current. The averaging of a reasonable number of samples gave sub-millimeter accuracy when the optimal current shape was used. The uncertainty in the average measurements are improved by roughly six times compared to the conventional triangular modulation. The reshaping also results in the range finding system being less sensitive to changes in ambient temperature.

A different method to determine α (Alpha) parameter for different number of Quantum-wells is presented. The Levenberg-Marquardt algorithm is used to train the Artificial Neural Networks (ANNs) which has a quadratic speed of convergence. Both the computed and the test results are in very good agreement with the experimental results reported elsewhere.

A different method and single model to determine the linewidth enhancement factor [α (Alpha) parameter] for narrow and wide GaAs Quantum-wells (QWs) as a function of modal peak gain and current density is presented. Based on the Artificial Neural Network (ANN) modeling approach, different learning algorithms are trained and tested. Both the training and the test results are in very good agreement with the experimental results reported elsewhere.

Titanium dioxide (TiO₂) cap layers were deposited onto C-doped InGaAs/AlGaAS QW laser structures by electron-beam evaporation in order to investigate their effect on atomic interdiffusion. In comparison to the as-grown sample, a negligible shift of the photoluminescence peak was observed after annealing at 900°C, indicating that the atomic interdiffusion was greatly suppressed by TiO₂ capping layers. For the uncapped sample, the high temperature annealing step significantly improves the threshold current density in laser diode devices but leaves the internal efficiency unchanged. We attribute these effects to the activation of the carbon p-type dopant, as demonstrated by electrochemcial C-V capacitance-voltage and X-ray measurements. SIMS analysis shows that the carbon atomic profile does not significantly change after annealing. In contrast, a similar Zn doped laser structure shows an almost flat Zn profile after annealing at 925°C, due to considerable indiffusion from the highly doped p⁺⁺ GaAs top contact layer in to the rest of the structure.

The GaAlAs/AlAs one-dimensional photonic band gap structure (1D-PBG) was proposed according to the transmission theory. This structure was utilized to tailor the EL spectrum of the commercial GaAlInP red double-heterojunction distributed Bragg reflector LED (DH DBR-LED), whose EL spectrum distributes from 620nm to 670nm (inspired at 120mA). The designed 1D PBG was employed to tailor the spectrum at the area from 620nm to 635nm, 640nm to 655nm, and 660nm to 670nm, left the windows at about 630nm, 640nm, 655nm, and 672nm. This 1D PBG was caped on the DBR-LED surface, and realized by MOCVD method. The EL spectrum of the sample has the illumination peak at 631nm, 640nm, 655nm and 672nm. The result is very consistent with the calculation. The spectrum of the LED was tailored according to the design. The influence of the 1D-PBG to the light emitting angle was studied too.

Implantation of Ti at MeV energies has been investigated as a means of locally doping sapphire with Ti to form Ti:sapphire: a highly valued laser material. Previous investigations of this approach to synthesis of Ti:sapphire have shown that the Ti³⁺ luminescence yield continuously increases with increasing annealing temperature up to 1500°C when a reducing ambient of 96% Ar + 4% H is used in an alumina tube furnace. Here we present preliminary results for anneals performed in the high temperature regime of 1500-1700°C in either an Ar or 96% Ar + 4% H ambient in a graphite furnace. Optical characterisation using photoluminescence (PL) has been combined with structural analysis using Rutherford backscattering spectrometry and ion channeling to study the annealing behavior of the ion implanted layers, the evolution of the Ti profile and the dependence of formation of Ti³⁺ on the implantation conditions, annealing ambient and temperature in this regime. At these annealing temperatures the Ti diffuses substantially. For anneals up to 1500°C, the reducing ambient produces the higher PL yields. At 1700°C, annealing in Ar results in higher PL yields whilst the reducing ambient leads to oxygen loss from the near-surface region and a concurrent reduction in PL yield. Annealing at 1700°C in Ar allows higher concentrations of Ti to be incorporated into sapphire in the 3+ oxidation state.

InP self-assembled structure were grown by MOCVD method on GaInP epitaxial layer mismatched to the GaAs substrate and were measured by employing the Near-field Scanning Optical Microscopy (NSOM) and SEM. The distribution of self-assembled islands was analyzed from the NSOM images and SEM results based on the scaling theories. It is found the distribution periodicity of the islands along [110] direction is improved and 1μm separation is obtained. The regular distribution was found along [1-10] direction. It shows and the mismatched epitaxial layer could improve the distribution periodicity of the islands along [110] and [1-10] direction. The experiment gives a potential way to realize the ordered two-dimensional distribution of the self-assembled structure. A mode, based on the shear force boundary and QD sphere, was established to explain the difference of our results between the topography and NPC. The size of islands could be evaluated by NSOM if the diameter of the probe has been taken account on.

An optoelectronic feedback (OEF) and optical reflected power (ORP, optical feedback) system theory of the single-mode laser diode is developed with the inclusion of Volterra model in this paper. The bandwidth of the whole system is increased compared to the usage of an OEF only. The critical values for both feedbacks are also determined for stable operation conditions.

We have developed highly strained GaInAs/GaAs quantum well (QW) vertical cavity surface emitting lasers (VCSELs) emitting at 1.1 - 1.2 μm wavelength band. Excellent temperature characteritsics, enabling uncooled operations, have been realized. We successfully extended the emission wavelength of highly strained GaInAs QWs up to 1.2 mm without degradations of crystal qualities. We demonstrated uncooled low threshold current operations and high speed operations up to 10 Gb/s. We realized error-free data transmission in a standard single-mode fiber using a GaInAs single-mode VCSEL. For further upgrade of high speed LANs, 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 as large as 190 nm. The multiple-wavelength array and wavelength engineering of VCSELs may open up future ultra-high capacity short reach systems.

The development of terahertz radiation (T-rays) is spurring new applications in spectroscopy and imaging. To maximize the use of T-rays in more applications, a high average terahertz power is needed. Rather than using fast diodes or laser sources, this paper will show that a synchrotron can generate high average power T-rays. This is achieved by creating an electron bunch in the synchrotron ring with high intensity in the terahertz frequency region via Thomson scattering.

Recent progress toward wavelength-scale photonic crystal lasers is summarized. To realize the ultimate laser, one needs to have a wavelength-scale photonic crystal cavity that is lossless. As a candidate for this ultimate laser, the two-dimensional unit-cell photonic crystal laser compatible with current injection is proposed. Experimental demonstration of the low-threshold two-dimensional photonic crystal lasers in the triangular lattice and the square lattice will be discussed. The very high quality factor in excess of 1,000,000 is theoretically predicted from the wavelength-scale resonator supporting the whispering-gallery-like photonic crystal mode.