Mid-infrared light-emitting diodes with InGaSb/AlGaAsSb triple-quantum-well active region have been integrated into arrays of either 200×200 μm² or 40×40 μm² square pixels. Two generations of arrays have been designed, fabricated and tested. The first, "sparse" 6×6 array provided valuable information on optimal electrode design and fabrication parameters that was used in the design and processing of the second generation "dense" 11×11 array.

Design analysis of III-Nitride based intersubband quantum well absorption in the mid-IR regime (&lgr; ~ 3-5 &mgr;m) is presented. The use of lattice-matched AlInGaN materials is advantageous because of its extremely fast intersubband relaxation time &tgr;_rel ~ 150-fs. The ability to engineer lattice-matched AlInGaN layer with GaN should allow realization of multiple pairs of AlInGaN / GaN quantum well structures, which would otherwise be challenging due to the cracking issues that might develop in conventional multiple pairs AlGaN / GaN heterostructures. The large conduction band offset in III-Nitride heterostructures is also beneficial for minimizing dark current and thermal noise.

This paper summarizes the main steps of the derivation of a set of simple analytical expressions obtained after a series of consistent approximations, starting from a more complete nonequilibrium many-body Keldysh Green's functions equations describing the coupling of light with intersubband excitations. The approach is valid under inverted medium conditions for which Hamiltonian approaches based on bosonic approximations cannot be applied.

Our group has employed the use of modern graphics processor units (GPUs) for the acceleration of finite-difference based computational electromagnetics (CEM) codes. In particular, we accelerated the well-known Finite-Difference Time-Domain (FDTD) method, which is commonly used for the analysis of electromagnetic phenomena. This algorithm uses difference-based approximations for Maxwell's Equations to simulate the propagation of electromagnetic fields through space and materials. The method is very general and is applicable to a wide array of problems, but runtimes can be very long so acceleration is highly desired. In this paper we present GPU-based accelerated solvers for the FDTD method in both its 2D and 3D embodiments.

In this work, numerical simulations are performed and the performance comparison of four types of dispersion compensating Raman/EDFA hybrid amplifiers configurations in terms of gain, noise figure and nonlinear effect induced penalty is presented. A numerical simulator is presented for the analysis and design optimization of Raman/EDFA hybrid amplifiers for multi-wavelength operation. This simulator combines a steady-state model of a discrete Raman Amplifier (RA) with a spectrally resolved model for the Erbium Doped Fiber Amplifier (EDFA). The numerical simulator allows us to calculate the overall gain, the noise figure (NF), the optical signal-to-noise ratio (OSNR) and the signal and ASE spectrum at the output of a Raman/EDFA hybrid amplifier. We concluded that performance tradeoffs should be considered when an optimum hybrid amplifier configuration for a given fiber optic transmission system is to be decided.

The combination of excellent electro-optical, acousto-optical and non-linear optical properties makes lithium niobate (LiNbO₃) an attractive host material for integrated optical components such as electro-optical modulators, acousto-optically tunable wavelength filters and Bragg gratings. In the last few years Erbium doped LiNbO₃ waveguide optical amplifiers (EDWA's) have attracted increasing interest. The combination of the amplifying properties of erbium with the excellent acousto-optical and electro-optical properties of the waveguide substrate LiNbO₃ allows the development of a whole class of new waveguide devices of higher functionality. The optical gain achievable in Ti:Er:LiNbO₃ waveguides by optical pumping could compensate or even over compensate these scattering, absorption and insertion losses leading to "zero loss" devices with net optical gain. The different types of lasers and amplifiers can be combined with other active and passive devices on the same substrate to form integrated optical circuits (IOC's) for a variety of applications in optical communications, sensing, signal processing and measurement techniques. The analysis of Er-doped diffused channel waveguides is, hence, required for design of amplifying integrated optical circuits in order to optimize the performance of these gain devices. The coupled differential equations, which govern the evolution of, pump power (1484nm), signal power (1485 to 1600nm) and amplified spontaneous emission, involve integrals which depend explicitly on the modal fields at the pump and signal wavelength in the diffused channel waveguide. In general, it is not possible to obtain analytical forms for the modal fields and propagation constant, hence, to obtain them various approximate or numerical methods (BPM, finite difference or finite element) are used. In this paper the modal field profiles obtained by the variational analysis are further approximated to an appropriately chosen Gaussian function, which leads to analytical forms of coupled differential equations with no integrals for the calculation of gain and ASE characteristics of the amplifying waveguide. Thus, computations are simplified and computation time is also reduced.

Proton exchange is becoming widely accepted as a complementary technique to titanium indiffusion for the fabrication of integrated optical waveguides in LiNbO₃. In this paper, we propose, for the first time to our knowledge, a novel approach to analyze the optical distributed waveguides formed by local index variation, combining two processes, titanium indiffusion (Ti), and a localized patterned proton exchange (PE), yielding to the Ti-PE: LiNbO₃ distributed parameter waveguides. We present a straightforward extension of the Wentzel Kramers and Brillouin (WKB) technique covering the computation of a single diffusion's effective index, to multiple successive diffusions having a specific graded-index profile. The efficiency of the method is shown by varying some of its parameters, like the index modulation or the proton exchanged depth, for example. The simulation's result proves that the evolution of the reflectivity's spectrum has been found to be in a good agreement with prediction.

Mutually contradicting previously reported theoretical results on the effects of a dispersive active medium in a ring laser on its sensitivity to rotation are critically analyzed. A measure of the rotation rate in the active ring resonator is the beating frequency &Dgr;&ngr; between the counterpropagating waves whose frequencies are shifted in opposite directions due to the Sagnac effect. &Dgr;&ngr; is shown to be inversely proportional to the group index of the medium filling the cavity. A comparison of the results obtained considering the resting and rotating frames confirms the applicability of Fresnel's original expression for the optical drag effect to the analysis of the Sagnac effect in an active laser gyro. The frequency splitting &Dgr;&ngr; is sensitive to dispersion via the group refractive index.

Laser interferometers detect gravitational waves with a degree of accuracy and efficiency unimaginable even a few years ago. The semiconductor lasers that are the primary light source for these devices are small, lightweight, durable and energy-efficient. On the downside, the devices currently available are still marked by broad oscillation spectra, and heightened sensitivity to fluctuations in injection current and /or ambient temperature. By applying a small sine wave to the injection current, we modulate the oscillation frequency. This frequency-modulated beam is introduced to the Avalanche photo diode through the Rb cell in the saturated absorption optical setup. The resulting signal and a reference signal are detected simultaneously and combined, to produce an error signal, which, when fed back to the injection current, stabilizes the diode's oscillation frequency at 2.12x10^-12 ⩽ &sgr;(2,τ) ⩽ 5.88x10^-11 in the averaging time between 0.4s to 65s. An optical feedback method, which introduces the laser beam reflected by a mirror or a grating to the semiconductor laser itself, is reported to narrow oscillation linewidth and improve frequency stability. We are now combining these two techniques to further improve frequency stability.

A method of detecting gravitational-field variations using laser diodes is described. While the GRACE project is currently using the Doppler microwave system to measure the velocities of satellites flying in tandem, in the future, more advanced laser interferometry will be employed. It is hoped that we will be able to measure infinitesimal changes in their velocities, by using frequency-stabilized lasers rated at better than 10^-13 in the square root of the Allan variance (&sgr;) for 1s < τ < 100s. As laser light sources, these devices will be notable for their compactness, energy efficiency, lightweight and high frequency-stability. This thesis describes the improved frequency stabilization obtained through the use of the magneto-optical effect of the Rb-D₂ absorption line, and the adaptation of the PEAK method, in order to obtain a precise control signal. The method allows us to modulate the reference frequency of the stabilization system (the absorption spectrum of the Rb-D₂ absorption line) by modulating the magnetic field applied to the Rb absorption cell, instead of the oscillation frequency of the laser diode. In so doing we are able to achieve a frequency stabilized laser diode (&sgr; = 9 x 10^-12), while maintaining its linewidth, at an averaging time of 40s. In the next stage, we will test frequency-stabilized laser optical sources that are to be used in detecting and observing gravitational waves.

External cavity diode laser (ECDL) systems are presently experiencing a surge in popularity as laser light-sources, in advanced optical communications- and measurement-systems. Because such systems require that their external reflectors be precisely controlled, to eliminate low frequency fluctuations (LFF) in optical output, we conducted experiments with a two-cavity version, which easily eliminated LFFs, as expected. The technique has the added advantage of a narrower oscillation-linewidth than would be achievable, using a single optical feedback. However, the ECDL's oscillation frequency is susceptible to the influences of the drive-current, as well as changes, both in the refractive index, and the overall length of the external reflector that results from fluctuations in atmospheric temperature. We made every effort to maintain the length of the ECDL cavity, while evaluating oscillation-frequency stability. We used a Super-Invar board as the platform for our compact ECDL system to minimize the influence of thermal expansion, because of its low expansion coefficient. We then compared the effect of atmospheric temperature variations between two experimental conditions, with the Super-invar board and without it, and finally took note of the improvement in performance, using the board.

Great efforts and vast investments have been put into the research and development of widely tunable lasers in the last 25 years. Tunable lasers have become critical components in the implementation of next generation telecommunication networks and systems, to provide dynamic wavelength provision for channel restoration, reconfiguration and protection. Some stringent requirements have been imposed on tunable lasers by telecommunication applications. Consequently, ultra-high optical output power (⩾100 mW), wide tunability (tuning range ~ 40nm), narrow linewidth (< 2MHz), and side-mode suppression ratio (SMSR > 40dB) have become the main objectives for the development of the future telecommunication tunable lasers. Facet output power is the fundamental decisive factor among these targets. Original design ideas and novel approaches to the design of ultra-high power InGaAsP/InP based multisection widely-tunable laser gain section have been reported by the authors previously, mainly including: firstly, a bulk balance layer structure is placed above the InP buffer layer and below the MQWs stack, which enables a large reduction of free-carrier absorption loss by greatly shifting the optical field distribution to the intrinsic and n-doped sides. Secondly, an InP spacer layer is placed below the ridge and above the multiple quantum wells (MQWs) stack, so as to introduce extra freedom in the control of widening the single mode ridge width. This paper will focus on the optimization on the implementation of the above design ideas and approaches, regarding single mode ridge width, optical confinement in the MQWs, optical overlap with the p-doped epilayers, output power, threshold current, and slope efficiency.

We compare the threshold gain and modal discrimination of a range of large aperture VCSELs intended for high-power single-mode operation. The threshold gain is calculated using a gain eigenvalue solver that enforces the threshold condition of the mode (gain = loss). In this way, gain guiding is included automatically. We find that confining the gain to a smaller area than the mode results in excellent threshold gain and modal discrimination, due to the large difference in modal overlap factors between the fundamental and the second order mode.

A stable far-field and single-mode performance is of great interest for many applications in sensing or communications. In this contribution an analysis of the far-field stability versus current and temperature is performed for a long-wavelength vertical-cavity surface-emitting laser (VCSEL) emitting around 1310 nm. Furthermore, the single-mode stability is investigated by means of a technology computer aided design (TCAD) tool. The electro-opto-thermal multi-dimensional simulations are fully-coupled and use microscopic models. The optical modes are obtained by solving the vectorial Helmholtz equation, using a finite element approach. The impact of temperature, free carrier absorption and gain on the refractive index is accounted for. The far-field is calculated using Green's functions. The investigated VCSEL features an InP-based cavity with multiple quantum wells and a tunnel junction as well as wafer-fused AlGaAs/GaAs distributed Bragg reflectors. The comparison of simulated and measured L-I, V-I characteristics and far-field as well as the wavelength-shift show good agreement for different ambient temperatures as well as driving current values. The simulations reveal the impact of temperature, gain and carrier effects on the far-field. The design of optical guiding structures (such as oxides or tunnel junctions) and its impact on the far-field behaviour over ambient temperature and bias current is discussed.

We explore the possibility of coating semiconductor nanowires with metal (Ag) to reduce the size of nanowire lasers operating at photon energies around 0.8-2 eV. Our results show that the material gain of a typical III-V semiconductor in nanowire may be sufficient to compensate Joule losses of such metal as Ag. The most promising mode to achieve lasing is TM₀₁ near its cutoff. To calculate the guiding properties of metal coated nanowires, we developed a finite-difference discretization approach, the details of which we also present. This approach allowed us to treat accurately the large index contrast of the nanowire/metal interface and to include nonperturbatively the imaginary parts of dielectric constants of the semiconductor core and metal coating.

We have performed an analysis of harmonic contents of the optical output power for a diode laser and described the results in details. In the first step the absolute value of power for each harmonic is obtained in terms of various diode laser parameters, and the variations of external parameters such as modulation current, bias current and frequency are discussed. The analysis is done by direct solution of rate equations of an arbitrary diode laser for carrier and photon densities. We conclude that the maximum power occurs at isolated peaks and their loci have been investigated and shown to be predictable by theory. It is known that the optical power has a nonlinear dependence on frequency, and the maximum optical power of each harmonic attained in its resonance frequency. The resonant frequency is shown to be tunable by bias current; thus in the next step we obtain the transfer function for different harmonic contents and have achieved exact expression for each, allowing better optimization to gain improved results. We extend the approach to higher harmonics and numerically calculate the THD (Total Harmonic Distortion) versus related parameters such as frequency, bias current and modulation current. Furthermore we found an effective approach to reduce SHD (Second Harmonic Distortion). The sequence for every arbitrary laser structure is also possible to be developed by the approach presented in this work.

An important feature of coupled laser arrays is that the gain in one cavity is modulated by the radiation from another cavity. This is a generic effect under many forms of coupling, including fringe field interactions in closely packed arrays and reflection feedback from external mirrors. Cross-cavity gain depletion can occur in many types of laser arrays, including edge emitting semiconductor lasers, VCSELs and fiber laser bundles. The interplay between frequency pulling and the characteristic cavity oscillations creates a rich behavior, setting the properties of active photonic lattices apart from the well known passive photonic lattices involving radiation interference due to real index variations. The case of planar VCSEL arrays is chosen as generic example for studying the physics of active photonic lattices. Results of theoretical calculations and numerical simulations are presented, addressing the following issues: (a) Non-linear phase-locked Bloch eigenmodes and boundary layer formation for finite arrays (b) Lattice defects, including sites that fail to lase, and defect tolerance (c) Excitation of stable, slow-light, lattice waves and photonic sound propagation (d) Unstable lattice behavior at high coupling strengths, with self-excited array oscillations and chaotic transitions (e) Phase locking in realistic arrays, with random variations in the cold-cavity parameters (manufacturing tolerances), via self-regulated frequency pulling (f) Existence and properties of randomly phase-locked arrays with "fuzzy" eigenmodes.

The dispersion relation of a cavity surrounded by multi-layered photonic crystals is obtained using a fast, accurate and generalized round trip operator. This will assist in the optical design of photonic crystal patterned lasers. A 2D quasi-bandgap was obtained for the lowest order mode of a 1D multi-layered photonic crystal. Although the method is demonstrated for 1D photonic crystal layers, the method is general and can be extended to two dimensional systems.

We proposed an effective index perturbation method to investigate the intrinsic characteristics of three-dimensional photonic-crystal-slab based microcavity with two-dimensional numerical simulation tools such as two-dimensional finite-difference time-domain (2D-FDTD) and plane-wave expansion (2D-PWE) techniques, for reduced computational requirements and fast design feedback. Less than 2% computational error in predicting cavity spectral locations was obtained for two widely used single defect and line defect air hole photonic crystal cavities, by adjusting the effective index to match the dielectric band edge for donor-like defect mode. The correlation between the modified effective index and the cavity (lasing) mode with the highest quality factor Q offers an efficient tool in the design of defect mode based photonic crystal microcavities.

We have designed and simulated compact photonic crystal (PC) slab waveguide based wave plate. We have numerically investigated the optical path of the TE-like and TM-like waves propagating through a triangular based photonic crystal (PC) slab waveguide. The PC slab waveguide is formed by removing one row of the air holes along &Ggr;K direction. The plane wave expansion and three-dimensional finite-difference time domain (3D-FDTD) methods were employed for the design and simulation of the PC slab waveguide. The thickness of 0.75a, a is the lattice constant, for the PC slab waveguide provides both TE-like and TM-like modal guiding within the normalized frequency band of 0.26-0.268. Spatial Fourier transform (SFT) of the electromagnetic field distribution in the propagation direction was used for the analysis of the dispersion properties of the guided modes of the PC slab waveguide. It was found that the effective refractive indices of the TM-like modes were substantially larger than that of the TE-like modes. The large birefringence of this structure suggests that the PC slab waveguide is useful for the construction of compact wave plates. The birefringence larger than 12.6 % within the modal guiding frequency band was achieved for the PC slab waveguide. Thus, the PC slab waveguide with the length of 6 &mgr;m provides first-order half wave plate within the normalized frequency band of 0.26-0.268.

GaAs-based PIN detectors with mesa sizes 1, 2.5, 5, 7.5 and 10 mm were fabricated and characterized for alpha particle response using a Po-210 alpha source. By decoupling the neutron conversion process of a proximity moderator, we were able to directly probe the alpha response characteristics of the PIN detectors as a function of device area. Dark current levels in the PIN detectors ranged from 6.1 to 9.5 pA at zero bias. The dark current values were higher for larger devices and a linear relationship between mesa size and dark current was observed. The PIN detectors were found to have a strong alpha response of up to 5 nA/mm² with a linear relation between the response current and mesa area. The measured responsivity of the detectors was 0.014 A/W. The average device efficiency was determined to be 31.5%. Using the measured alpha response properties of the GaAs PIN diodes one is able to select the optimal device area for a given moderator and application specific neutron flux.

In this contribution, substrate modes in edge-emitting lasers in the material system Gallium-Nitride are analyzed by means of comprehensive measurements and simulations. The simulations are complex vectorial optical mode calculations using a finite-element method. The simulation domain comprises the ridge waveguide and the full substrate with open boundary conditions on the sides. Therefore, the coupling mechanisms of the waveguides formed by the ridge and the substrate can be analyzed in a realistic setup. The characterization data include the optical loss spectrum obtained from Hakki-Paoli measurements, optical near field, and farfield measurements. The devices used for characterization are ridge waveguide quantum well lasers grown on GaN substrates. A comparison of the measurement data with the simulations explains the characteristics of the substrate modes in a consistent way, and shows very good agreement for the optical loss oscillations, farfield angle, and nearfield pattern. It is shown that material losses, material dispersion and optical diffraction are key ingredients for the analysis of substrate modes.

In attempt to prepare a high performance Al_xGa_1-xN based UV-B LED, a computer simulation has been performed on a typical UV-LED structure to find out the effect of threading dislocations on non-radiative recombination process. UVB LED structures were formed on using GaN and AlN based layers for comparison. Cracks were generated in the device structure formed on GaN underlayer. No cracks were observed on the device structure formed on AlN under layer. Much better structure was formed when the base AlN was grown by high temperature MOVPE.

We report on low-defect-density non-polar a-plane and m-plane GaN films grown by sidewall epitaxial lateral overgrowth (SELO) technique. Dislocations and stacking faults were decreased markedly over the whole area, and surface roughness was decreased with decreasing defect density. The photoluminescence intensity of SELO a-plane and m-plane GaN was about 200 times higher than that of a-plane and m-plane GaN template. We also fabricated and characterized LEDs on a-plane and m-plane GaN using SELO technique. The light power of LEDs increased with decreasing of threading dislocation.

In this work, thermal effects on the optoelectrical characteristics of green InGaN/GaN multiple quantum well (MQW) light-emitting diodes (LEDs) have been investigated in detail for a broad temperature range, from 30^oC to 100^oC. The current-dependent electroluminescence (EL) spectra, current-voltage (I-V) curves and luminescence-current (L-I) curve have been measured to characterize the thermal-related effects on the optoelectrical properties of the InGaN/GaN MQW heterostructures. Experimentally, both the forward voltages decreased with slope of -2.6 mV/K and the emission peak wavelength increased with slope of +4.5 nm/K with increasing temperature, indicating a change in the contact resistance between the metal and GaN layers and the band gap shrinkage effect. With increasing injection current, it has been found the strong current-dependent blueshift of -0.048 nm/mA in EL spectra. It was attributed to not only the stronger band-filling effect but also the enhanced quantum confinement effect, resulted from the piezoelectric polarization and the spontaneous polarization in InGaN/GaN heterostructures. The junction temperature calculated by forward voltage was from 25.6 to 14.5^oC and by emission peak shift was from 22.4 to 35.6^oC.

Blue and green LEDs have been simulated. Changing LED performance characteristics, depending on In concentration and at different temperatures were simulated. It was suggested that a LED having p-n junction area S₀ can be considered as a sum of "SmallLEDs (SLEDs)" electrically connected in parallel, each SLED has its own In-content and its own p-n junction area S(X). Values of ratio S(X)/S₀ are described by Gauss distribution function in the range X = 0.15-0.25 for blue LEDs and X = 0.25-0.35 for green LEDs. Reasonable correspondence of simulation and experimental results (current-voltage characteristic (C-V Ch), Spectral Ch) can be observed.

A time domain model for reflective semiconductor optical amplifiers (RSOAs) is developed based on the carrier rate equation and wave propagation equation. In this model, the gain saturation effect and the dependence of spontaneous carrier lifetime on carrier density are explicitly included, and the evolution of carrier density and the optical power in time and space under current modulation is considered in detail. Using the time domain model, the performance of RSOAs with different active layer lengths is investigated under different inject current densities and input optical powers. Numerical simulations reveal that the carrier spontaneous lifetime is the foremost limiting factor of RSOA modulation speed, but increasing photon density improves RSOA performance. With increased bias currents or optical input powers, the small signal frequency response is improved and the eye closure penalty under large signal on-off key modulation is reduced, but the extinction ratio of the optical output signal is decreased. Under the same bias current density and optical input power, RSOAs with longer active layers exhibit improved frequency response and smaller eye closure penalty.

We present a simple analytical model to approximately analyze the TDFA for silica glass as well as fluoride glass based optical fibers. We have verified the validity of method using the in-house fabricated Tm-doped alumino-germano-silicate glass fiber as well as reported results for the Tm-doped fluoride glass fiber. The spectral variation of gain obtained with the silica glass fiber TDFA after pumping with pump power of 275 mW at 1064 nm showed good agreement between the simulated and the measured values. The pump power dependence of gain in the fluoride glass fibers calculated by our method also showed a good agreement with the experimental results reported. The maximum error in determination of gain was limited to 12%.

We report new experimental and theoretical findings relating the design of dopant profile to the bandwidth and optical saturation power of high speed InGaAs/InP unity gain photodetectors. We demonstrate significant improvements in each of these performance metrics by means of quasi-unipolar operation, and interpret results on the basis of full band ensemble Monte Carlo simulation.

In this paper we have presented a mathematical treatment for tailoring the Transmission Spectrum of the Long Period Gratings by variation in the length of the grating. Our results show that the transmission spectra show a rather varied behavior depending on the length of the grating. We have formulated generalized expressions for finding positions of maxima and minima as well as power retained by the core mode at these positions in the transmission spectrum of the grating. This facilitates the designing of gratings to be used as in line loss elements in applications like gain flattening, gain equalization etc.

The presence of multiple polarized beams can limit the polarimetric discrimination capability of a wire-grid polarizer (WGP). In this study, the effect of linearly polarized background on the polarimetric performance of a WGP has been investigated and compared with that of a perfect polarizer. Simulation results based on rigorous coupled-wave analysis indicate that while a WGP mimics a perfect polarizer in discrimination characteristics, the range of the object polarization angle that can be discriminated against polarized background is fairly limited. The negative impact of Rayleigh anomaly is also discussed. The detectability of the object polarization can be strongly enhanced by employing a multi-cell WGP with multiple polarization orientations.

The analysis of the RIN spectrum of the LCDL devices allows important information on the nonlinear behaviour of Laterally Coupled Diode Lasers (LCDL). Under certain bias conditions, the second resonance frequency characteristic of the LCDL devices is at double the relaxation oscillation frequency and, for the first time, a period doubling is observed. This effect is similar to a single diode laser when is current modulated at the double of the relaxation oscillation frequency. So, in this paper a comparative study between period doubling evolution obtained in LCDL device versus the single laser diode is made.

Nonlinear dynamics of semiconductor lasers has found many interesting applications in microwave photonics technology. In particular, a semiconductor laser under optical injection of proper strength and optical frequency detuning can enter into the dynamical period-one (P1) state through Hopf bifurcation. The resulting optical output carries a broadly tunable high-speed microwave modulation without employing any expensive microwave electronics. It is therefore a desirable source for radio-over-fiber (RoF) applications. The P1 state can also be adjusted to have a nearly single sideband (SSB) optical spectrum. It is an advantageous property for long distance fiber transmission because it minimizes the microwave power penalty that is induced by chromatic dispersion. In this work, we investigate in detail the properties of the P1 state and the effect of fiber dispersion as a function of the injection conditions. Based on a well-established rate equation model, the results show that the generated microwave frequency can be several times higher than the intrinsic relaxation resonance frequency of the laser. With a large injection strength and an injection detuning frequency higher than that required for Hopf bifurcation, the generated microwave power is nearly constant and the optical spectrum is close to SSB. We simulate the effect of fiber chromatic dispersion and the result shows a maximum microwave power penalty of less than 2 dB. The characterization of the P1 state is useful in guiding the design of RoF systems based on optically injected semiconductor lasers.

The time evolution of optically excited carriers in semiconductor quantum wells and quantum dots is analyzed for their interaction with LO-phonons. Both the full two-time Green's function formalism and the one-time approximation provided by the generalized Kadanoff-Baym ansatz are considered, in order to compare their description of relaxation processes. It is shown that the two-time quantum kinetics leads to thermalization in all the examined cases, which is not the case for the one-time approach in the intermediate-coupling regime, even though it provides convergence to a steady state. The thermalization criterion used is the Kubo-Martin-Schwinger condition.

The performance of lasers with self assembled quantum dot active regions is significantly affected by the presence of the two dimensional wetting layer and the other states necessary for carrier injection due to the manner in which carriers are distributed amongst the various states. In this work we describe three approaches to overcome the low value of maximum saturated gain, which has been observed by many groups worldwide, and explain the approaches in terms of the impact on the distribution of carriers within the available states. We present results of direct measurements of the modal gain and measurements that indicate the form of the carrier distribution within the samples to justify our argument. The structures examined include the use of a high growth temperature to smooth the matrix layer, the use of p-type modulation doping and the use of InAlAs capping layers and all have been grown by solid source molecular beam epitaxy. We demonstrate CW operation at 1.3&mgr;m for 1mm long devices with uncoated facets and very low threshold current density (< 40Acm^-2) in longer devices. We also demonstrate that the negative T₀ (reducing threshold current density with increasing temperature) obtained around room temperature in our p-doped devices is due to the temperature dependence of the gain.

We study the optical properties of semiconductor quantum dots by means of a quantum-kinetic theory. The excitation-induced dephasing and the corresponding line-shifts of the interband transitions due to carrier-carrier Coulomb interaction and carrier-phonon interaction are determined and used in conjunction with the usual ingredients of a gain calculation like Coulomb enhancement and State filling to set up a microscopic calculation of the quantum dot gain. We find that for very high carrier densities in QD systems the maximum of the optical gain can decrease with increasing carrier density due to a delicate balancing between state filling and dephasing.

The modulation characteristics of multi-section gain lever quantum dot lasers are investigated in this paper. A 20-dB enhancement in the amplitude modulation efficiency is observed in a two-section quantum dot laser. Based on rate equation analysis a novel modulation response equation is derived to describe the device dynamics. In addition the dependence of the modulation efficiency enhancement and 3-dB bandwidth on the length of the modulation section is discussed. A conservative estimate of the gain lever value of 33 is derived from the measured results.

Spatial mode dynamics in continuous-wave high-power wide-aperture quantum dot lasers is studied within the framework of Maxwell-Bloch equations. The effects of phase-amplitude coupling, junction temperature profile, carrier diffusion - capture - escape phenomena on the lateral modal dynamics is included in the model. It is shown that for low values of alpha factor and at normal operating points, there is an improvement in spatial coherence of the emitted light. It is found that active layer temperature and carrier-diffusion plays a key role in deciding the spatial mode structure in the device.

We develop a model that describes the optical response of a semiconductor quantum dot microcavity pumped above transparency but kept slightly below threshold. The model takes into account the inhomogeneous broadening of the dot emission, the coupling mechanisms between quantum dots and the wetting layer and incorporates gain asymmetry factors in the thermo-emission and capture coefficients. The role of asymmetries with respect to alpha factor and pattern formation is investigated. We then study the conditions for the onset of bistability and modulational instability and characterize the patterns formed.

Quantum-confined Stark effect in zero-dimensional semiconductor quantum-dot (QD) has attracted considerable interest due to the potential applications in electro-optic modulation and quantum computing. Composition interdiffusion occurs easily during the high temperature epitaxial growth or ex situ annealing treatment, therefore understanding the effects of interdiffusion is essential for device design and modeling. However, relatively little attention has been devoted to a systematic study of this effect. In this paper, the effects of isotropic interdiffusion on the optical transition energy of self-assembled InAs/GaAs QD structure under an electric field have been investigated theoretically. Our three-dimensional QD calculation is based on coupled QDs with different shapes arranged periodically in a tetragonal superlattice, taking into account the finite band offset, valence-band mixing, strain, and effective mass anisotropicity. The electron and hole Hamiltonians with the interdiffusion effect are solved in the momentum space domain. Our results show that isotropic three-dimensional In-Ga interdiffusion will makes the Stark shift become more symmetry about F=0 in asymmetric lens-shaped and pyramidal QDs, implying the reduced build-in dipole momentum. The interdiffusion also leads to enhanced Stark shift with more prominent effects to QDs that are under larger electric fields.

Here we present a full 3-D numerical model based on Quantum-Finite-Difference-Time-Domain (Q-FDTD) method, with Perfectly Matched Layer (PML) boundary condition, as a versatile tool to accurately analyze 3-D QD nanostructure with arbitrary shape. Model solid theory has been utilized to determine the 3-D band lineup of the QD heterostructure. The effects of strain distribution, and effective mass distribution on the band structure of the QD nanostructure are also taken into account in the model. The Q-FDTD computation has been applied for analyzing MOCVD-grown InGaAs QDs with GaAs_1-xP_x barriers on GaAs. The Q-FDTD simulation, using the QDs shape measured by TEM and AFM, shows good agreement with the experimental results obtained from the as-grown InGaAs QDs with GaAs_1-xP_x barriers.

The unique properties of semiconductor nanowires pose promising applications in optoelectronics such as photo-detectors and lasers. Owing to the increased surface/volume ratio, nanowire-based p-n junctions exhibit qualitatively different properties from those of bulk cases. These include weaker electrostatic screening and stronger fringe field effects. This work employs a general device simulator, PROPHET, to numerically investigate the unique electrical properties of p-n junctions in single nanowires and nanowire arrays. The implications of such effects in nanowire-based photo-detector design are also examined.

A simple and systematic algorithm based on the perfectly matched layer (PML) method and spectral element method (SEM) is introduced to solve the 3-D Schrodinger equation with tensor effective mass. This algorithm extends the lead regions of a device into artificial PML media, where a modified Schrodinger equation is satisfied. The interface between the physical and PML media has zero reflection coefficients, thus waves attenuating rapidly into the PML region before transmitting to the contact boundary. This algorithm provides a highly effective open boundary condition in solving quantum transport problems. The additional PML region can be designed such that less than -100 dB incoming waves are reflected by this artificial material with the implementation of the spectral element method. Consequently, the solution of the Schrodinger equation and thus the current in the original device region do not deviate from the correct solution. In this algorithm, the transmitted wave function is treated as a total wave instead of being decomposed into waveguide modes, therefore, it significantly simplifies the problem in comparison with conventional open boundary conditions. The implementation of the tensor effective mass provides an excellent tool to study strain effects along any arbitrary orientation. Within this PML implementation, the spectral element method has been applied to achieve an error that exponentially decreases with the increase of the polynomial order and sampling points. This accuracy has been demonstrated by comparing the numerical and analytical results from waveguide examples, and its utility is illustrated by multiple-port devices and nanotube devices.

We discuss recent progress using the radiative emission of single quantum dots as a triggered source of both single photons, and photon pairs displaying polarization entanglement. Excitation of a quantum dot with two electrons and two holes leads to the emission of a pair of photons. We show that, provided the spin splitting of the intermediate exciton state in the decay is erased, the photon pair is emitted in an entangled polarization state. Using quantum dots to generate quantum light has the advantage of allowing a robust and compact source to be realised with contacts for electrical injection. A cavity may be integrated into the semiconductor structure to enhance the photon collection efficiency and control the recombination dynamics. We detail a process to form a sub-micron current aperture within the device, allowing single quantum dots to be addressed electrically.

We have experimentally demonstrated purely optical manipulation of wide-gap semiconductor CuCl quantum dots in superfluid helium. The superfluidity provides an ideal cryogenic frictionless environment for the manipulation. In order to introduce the quantum dots into liquid helium, small particles of CuCl with a broad size-distribution ranging from 10 nm to 10 &mgr;m in radius have been fabricated from a bulk sample by laser ablation in a helium cryostat. We irradiated these particles with laser light covering the excitonic resonance levels of the quantum dots smaller than 50 nm to push them by using resonant radiation force. As a result, we have found that many quantum dots of which sizes range from 10 to 50 nm were transported and sorted over a macroscopic distance, ~1 cm. Importantly, the excitonic resonance condition was crucial for this optical manipulation. The result means that the resonant radiation force for the quantum dots is much stronger than the gravitational force. Feasibility of size-selective manipulation is also discussed.

In this work, the optical characteristics of monolithic passively mode-locked lasers (MLLs) fabricated from 1.24-&mgr;m InAs dots-in-a-Well (DWELL), 1.25-&mgr;m InGaAs single quantum well (SQW), and 1.55-&mgr;m GaInNAsSb SQW structures grown using elemental source molecular beam epitaxy (MBE) are reported. 5 GHz optical pulses with sub-picosecond RMS jitter, high pulse peak power (1W) and narrow pulse width (< 10 ps) were demonstrated in monolithic two-section InAs DWELL passive MLLs. With the 42% indium InGaAs SQW MLL, a record high-temperature performance for a monolithic passively mode-locked semiconductor laser is found. Compared with the typical operating range of the InAs DWELL devices (<60°C), the operation is in excess of 100 °C. The first 1.55-&mgr;m GaInNAsSb SQW MLL operates at a repetition rate of 5.8 GHz and has a 3-dB bandwidth of 170 kHz in the RF spectrum indicating respectable jitter.

In this paper, two-section mode-locked lasers consisting of monolithic quantum dot gain and absorber sections are studied as a function of absorber voltage, injected current to the gain region, and relative section lengths. We map the regions of stable mode-locking as measured by the electrical and optical spectra. A simple algorithm is presented that evaluates the quality of mode locking and allows automated characterization of devices. The relative advantages of increasing the absorber length compared to increasing the absorber reverse bias voltage are analyzed. Initial data indicate that doubling the absorber length from 1.4 to 2.8-mm in a 5 GHz repetition rate device increases the region of stable mode-locking by at least 25%, while increasing the absorber reverse bias can more than double the mode-locking regime. Nonetheless, in these devices, stable mode-locking over greater than a 100 mA bias range is realized with a grounded absorber making single bias control of a passively mode-locked semiconductor laser feasible.

We recognize that the superposed light beams do not interact with each other to re-distribute their energy in space or time in the absence of interacting material dipoles. This platform requires that we re-visit the physical model behind the generation of pulses from the so-called "mode-locked" lasers. In the process, we come across the mathematical models behind formulating (i) the autocorrelation due to pulsed light and (ii) the group velocity of pulse propagation are based on the direct summation (integration) of non-causal infinite Fourier frequencies as if the EM waves can actually modify their energy distribution in the time domain. Accordingly, we show by modeling results and proposed experiments that time-frequency Fourier theorem can give rise to self-contradictory predictions, verifiable by simple laboratory experiments. Based on these results, we propose that we replace the paradigm of "interference of light" by "superposition effects due to light beams" as reported by the material dipoles of detectors and beam splitters.

Advanced types of QD media allow an ultrahigh modal gain, avoid temperature depletion and gain saturation effects, when used in high-speed quantum dot (QD) vertical-cavity surface-emitting lasers (VCSELs). An anti-guiding VCSEL design reduces gain depletion and radiative leakage, caused by parasitic whispering gallery VCSEL modes. Temperature robustness up to 100°C for 0.96 - 1.25 &mgr;m range devices is realized in the continuous wave (cw) regime. An open eye 20 Gb/s operation with bit error rates better than 10^-12 has been achieved in a temperature range 25-85°C without current adjustment. A different approach for ultrahigh-speed operation is based on a combination of the VCSEL section, operating in the CW mode with an additional section of the device, which is electrooptically modulated under a reverse bias. The tuning of a resonance wavelength of the second section, caused by the electrooptic effect, affects the transmission of the system. The second cavity mode, resonant to the VCSEL mode, or the stopband edge of the second Bragg reflector can be used for intensity modulation. The approach enables ultrahigh speed signal modulation. 60GHz electrical and ~35GHz optical (limited by the photodetector response) bandwidths are realized.

The operation and performance of an InGaAs/InP uni-travelling-carrier photodiode (UTC-PD) has been studied using a commercial device simulator. We compare the UTC-PD with conventional PIN photodiodes, focusing particularly on evaluating the intrinsic device performance.

The interaction of semiconductors with terahertz radiation is discussed. The main ingredients of a consistent microscopic description are presented. The theory is evaluated to analyze direct terahertz emission features of semiconductor systems.

The compression and reshaping of optical pulses is a key issue for many of the applications in which ultrashort optical pulses are present since dispersion, nonlinearity and losses degrade their quality. We present a novel numerical procedure for designing pulse compressors based on Nonlinear Optical Loop Mirrors (NOLM). To exemplify the performance of the model, we apply this tool to the design of a NOLM intended to compress and reshape low energy pulses obtained by means of diode laser pulsed sources. This way, the quality of the pulses generated with this techniques can be improved.