Low timing jitter (< 5 ps), high peak power (1 W at the laser facet), narrow pulses (< 10 ps) at a 5 GHz pulse repetition rate is demonstrated by monolithic passively mode-locked quantum dot lasers. Their low cost, compact size and DC-biased operation make them ideal for high speed optical interconnects and optical clocking applications.

Short pulsed laser sources have recently been under investigation for applications in high speed optical communications systems. Studies concerning short pulse laser sources based on Gain Switching techniques and compression devices have been made. Nonlinear Optical Loop Mirrors (NOLM) provide both compression and reshaping of the optical pulses. As achieved pulse widths become shorter, timing jitter becomes crucial. The present work turns the attention to the influence of timing jitter in the performance of the NOLM pulse compression device. The results obtained will be the basis for the design of a compact high quality ultrashort optical pulse source at 40 GBs with monolithic semiconductor compressor. This device is under investigation within the MONOPLA project.

The integration of photonic clocking in microprocessors is anticipated to occur during the 2008-2012 high-volume manufacturing (HVM) cycle. Though photonic clocking can be achieved through electronic modulation or actively mode-locking a laser, a more cost-effective and better solution would be to use internal cavity passively mode-locked semiconductor lasers. Not only do these lasers offer low-cost, simplicity, and ease of integration, but prototypes that are amenable to HVM are currently available. We present such a laser that is scalable by design to clock rates of 9 to hundreds of GHz and wavelengths in the 800 to 1100+ nm range. These lasers utilize internal saturable absorber(s) to passively mode-lock a semiconductor laser with relatively high peak powers. Experimental results from these lasers show an RF spectrum signal peak that is at least 40 dB above the noise floor with a -10 dB width of <1 MHz. The RMS jitter as determined by an oscilloscope with a precision timebase module was found to be ~1 ps which is among the best for this type of laser. Autocorrelation was used to confirm mode-locking and pulse width. In addition to experimental data, a theory and discussion on how the different characteristics of these lasers can be tailored for various commercial applications such as microprocessor clocking will be presented.

Recent progress in data transmission systems requires high output power single transverse mode laser diodes for Raman and Erbium-Doped Fiber Amplifier pump sources. Several approaches have shown high output power emission up to 1 Watt for 14xx-nm InGaAsP/InP pump lasers. In these works, the design rules were mainly related to the active volume increase, through large or flared waveguides or through the use of MMI. Such designs lead to electrical and thermal resistance decrease. MMI waveguide have already demonstrated better behavior in comparison with straight active waveguide, but neither experimental comparison nor modelisations between different active MMI structures have been reported. In this work, we report on 2D-BPM simulations and experimental results between different active MMI 1x1 structures for 14xx-nm pump laser. The MMI 1x1 is a large multimode waveguide connected in both sides center by one single transverse mode waveguide. Several parameters such as p order, wavelength emission, MMI width and MMI length have been studied in order to carry out simple design rules for active MMI waveguide based lasers. Moreover, to achieve high level output power, long lasers are necessary. In this case, we also compare long laser, typically 1.2mm, formed by one long MMI waveguide or formed by several short MMI waveguide. Simulations show a better behavior for short MMI length (L_MMI<250μm) and small MMI width (WMMI<10μm). As expected, experimental results are in good agreement with simulations, and the output power figure is improved using the long laser formed by several short MMI waveguide.

We report on the design and analysis of a surface-normal coupled-quantum-well (CQW) electro-absorption modulator (EAM) using a simulation software APSYS. The modulator is designed as a p-i-n diode operating at 1.55 um, in which the CQW structure is composed of InGaAs-InAlAs. In this paper, detailed analysis of the internal device physics is investigated, which includes the transition energy and the square of the normalized overlap integrals between confined energy levels for electrons and holes. This analysis reveals the mechanisms in the device that are responsible for the redshift of the absorption spectra for achieving an operating wavelength at 1.55 um, which, in turn, provides researchers with insight in the high performance modulators design. The analysis also explains the dependence of the absorption spectra on wavelength and bias voltage. The simulation model self-consistently solves for wavefunctions and the corresponding quantum states energies and then used for absorption spectra calculations within the 2D/3D drift-diffusion framework.

The dynamics of a semiconductor laser subject to polarization rotated (incoherent) optical feedback have been numerical and experimental investigated. The results show that the induced dynamics can be grouped into four regimes (stable, chaotic, pulsed and two-state) and they also show that the TM mode of the laser never lases, at least in the ring cavity configuration studied. In addition, unlike the earlier short cavity work, the boundaries between the regimes in the long cavity case were found to be independent of the external cavity delay time. By focussing on the effects induced by incoherent optical feedback from long external cavities we will show that four regimes of operation can be accessed by varying the external reflectivity. Calculated Lyapunov exponents and a bifurcation diagram clearly delineate these regimes and experimental verification of the transition from stable to chaotic dynamics will be described. The dynamical behaviour exhibited in each regime will be explained in terms of the memory effect introduced by the time delay associated with the external cavity and the natural relaxation oscillations of the solitary laser. It will also be shown that within the pulsed regime the pulse repetition rate is determined by the external reflectivity and injection current and is independent of the external cavity length.

The response of a diode laser resulting from an incoherent delayed optical feedback is considered from numerical and experimental perspectives. We concentrate on a class of solutions that appear as regular square waveforms. A two-field model is used and the bifurcation diagram of these square-wave regimes is studied. Conditions under which they typically appear are determined. The roles of various parameters are examined, particularly with regard to the gains and losses of the two polarization modes. Numerical results are in close agreement with experiments.

External cavity diode laser (ECDL) systems are presently experiencing a surge in popularity as laser light-sources, in advanced optical communications- and measurement-applications. 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 of ECDL system, which was expected to eliminate LFF easily. This technique brings the added advantages of a narrower linewidth than would be achievable via a single optical feedback. However, the ECDL's oscillation frequency is susceptible to the influences of the driving current, changes in the refractive index, and the expansion/contraction of the 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-inver board as the platform for our ECDL system, in order to minimize the influence of thermal expansion. Moreover, our ECDL system combines an Rb cell within an external cavity to improve stability; by restricting the LD frequency to both the external cavity mode and to the Rb saturated absorption spectrum. We used the square root of the Allan variance to evaluate oscillation frequency stability, observing, in the process, that it improved stability about one order of magnitude.

We report on the stabilization of a semiconductor laser's frequency, using Rb absorption lines. In order to improve overall frequency stability within our system, we adjusted the setup used in Rb- saturated absorption spectroscopy, and optimized modulation parameters such as modulation - frequency and -width, to more accurately detect the error signal. When we stabilized laser frequency using a Doppler-free absorption line of Rb atoms, a time-constant of 0.01sec, and a modulation frequency of 7.77kHz, relative optical frequency stability of 2.12×10^-12≤σ(2,τ)≤5.88×10^-11 was achieved, in averaging time for 0.04s≤τ≤65s.

The nonlinear mode interactions (NMIs) in semiconductor lasers produce perturbations of optical parameters in the vicinity of strong-mode frequencies. These perturbations include changes in the complex refractive index, which could either increase or decrease, depending on the frequency detuning. Variation of the refractive index over a small frequency range may produce a rather strong variation of the group index. There are regimes of slow light (large group index), fast light (group index smaller than the phase index), and a range of negative group index. Critically anomalous dispersion conditions occur when the group index changes its sign. Thus, the NMIs in diode lasers and in semiconductor optical amplifiers can be used to obtain slow or fast light. A specific application is considered for gyroscopic laser devices that are sensitive to dispersion. It is shown that the fast light phenomenon can be used to increase the scale factor in a gyro without increasing its size.

Understanding the nature of single-photon propagation and any differences to the classical multi-photon regime is essential to better use the escalating number of single-photon technologies. Novel techniques and results are reported that demonstrate a consistent ability to measure the classical group velocity in an optical fiber to an accuracy in the time-offlight around ±0.2 ns in 30,000 ns. Identical techniques are then applied to single photons to determine their velocity and assess if there is any evidence that Schrödinger frequencies (N+1/2)ω alter the ω-k dispersion diagram. The evidence suggests that photons in a circular fiber show no additional dispersion but travel at the classical group velocity.

A dynamical laser model is coupled to a fully microscopic calculation of scattering rates, allowing effcient calculations without phenomenological parameters. The approach is used to analyze nonequilibrium effects in the switch-on of an optically pumped laser structure. Lasing leads to kinetic hole burning in both electron and hole distribution. The gain spectrum, however, does not show spectrally narrow hole burning but a reduction over a wide range of frequencies compared to the equilibrium gain because of the large homogeneous broadening in the high density lasing system.

We present and analyze a new nitride-based gain media by utilizing type-II InGaN-GaNAs quantum well (QW) on GaN. This novel III-N based type-II QW allows extension of the emission wavelength from blue regime (450-nm) to yellow regime (550-nm) with relatively-low In-content in the QW, while maintaining a large electron-hole wavefunction overlap. High electron-hole wavefunction overlap (≥ 65%-70%) can be obtained by careful energy band engineering to take advantage of the polarization-induced electric field. Our analysis shows this new type-II QW gain media offers wide emission wavelength coverage, from pure blue (~450-nm) to yellow-green (~530-nm). Design and optimization for pure blue (~450-nm), green (~515-nm) and yellow-green (~530-nm) emission structures are also presented. This method may allow realization of green laser diode on GaN, as well as paving the way to low cost, truly monolithic solid-state white light source.

The application of magnetic fields to semiconductor lasers gives rise to significant alterations in their oscillation wavelengths. Reports dating back to the 1960's describe shifts to shorter oscillation wavelengths, at extremely low temperatures (80K) and strong magnetic fields (~4T). These shifts were explained by the effect of the Landau Level. In our preliminary experiments, we exposed bulk-type semiconductor lasers oscillating at 780nm to a relatively weak magnetic field (1.4T), at room temperature (300K), observing, in the process, that the oscillation wavelength shifted to the longer (low-frequency) side. This work looks at the longer side shift, the optical output-power shift on the low-power side, and the terminal-voltage shift seen on the higher voltage side of some multi-quantum-well (MQW) laser diodes oscillating at 780nm, under the same experimental conditions we employed in previous works. In discussions of shift mechanisms, we consider how wavelength (frequency), optical output-power, and terminal voltage-shifts are correlated. Our expanded knowledge base has forced a complete rethink of the mechanisms we relied upon in the 1960's. Current evidence suggests that the temperature rise and the longitudinal magneto-resistance effect were "co-conspirators" in the shifts observed in our experiments, when we applied magnetic fields to laser diodes parallel to the injection current. In an upcoming report, we will describe the results of experiments conducted near the threshold current.

We have proposed all-optical signal processing using an optically injection-locked two-mode vertical-cavity surface-emitting laser (VCSEL). The operating principle is based on an all-optical inverter using transverse mode switching of an injection-locked VCSEL. The nonlinear transfer function in input-output characteristics enables us to reshape degraded input signal. In addition, the polarization state of the output signal can be tightly controlled by two-mode optical injection scheme. Thus, the polarization state of an inverted signal can be fixed for randomly polarized input light. In addition, 3R regeneration can be achieved with electrically driven clock signal. The proposed scheme gives us novel functions such as all-optical 3R regeneration with optical re-polarization function and NRZ-RZ format conversion. We carried out the modeling based on a multi-transverse-mode rate equation analysis which includes two orthogonal polarization modes, exhibiting a potential of high-speed 4R regeneration and format conversion at 10 Gbps.

We present the static and dynamic simulation of a long-wavelength vertical-cavity surface-emitting laser (VCSEL) operating at around 1310 nm. The device consists of AlGaAs/GaAs distributed Bragg reflectors (DBRs) which are wafer-fused to both sides of the InP-based cavity with InAlGaAs quantum wells. A tunnel junction is used for current injection into the active region. The structure is simulated with a modified version of the commercial device simulator Synopsys Sentaurus Device. The fully-coupled two-dimensional electro-opto-thermal simulations use a microscopic physics-based model. Carrier transport is described by the continuity and Poisson equations and self-heating effects are accounted for by a thermodynamic equation. To obtain the opticalmodes, the wave equation is solved using a finite element approach. The optical gain model includes many-body effects. The equations are solved self-consistently. Calibrations of static (L-I, V-I curves) and dynamic characteristics (RIN) show good agreement with measurements at different temperatures. On this basis, the simulations reveal the critical factors that determine the modulation-current efficiency factor (MCEF) of the device.

We report on lasers and light emitting diodes based on the longitudinal photonic bandgap crystal (PBC) concept. The PBC design allows achieving a robust and controllable extension of the fundamental mode over a thick multi-layer waveguide region to obtain a very large vertical optical mode spot size and a very narrow vertical beam divergence. An efficient suppression of high order modes can be realized either by the optical confinement factor selection of the fundamental mode, which is localized at the "optical defect" region and has a higher overlap with the gain region. All the other modes spread across the thicker PBC waveguide. In another approach leakage loss selection can be used to suppress excited modes in case of absorbing substrate or the substrate with a higher-refractive index. In this paper we concentrate on growth and performance of high power single mode visible (650 nm) GaInP/AlGaInP PBC lasers, giving a comprehensive example. The devices show narrow far field pattern (full width at half maximum of vertical beam divergence of about 7°), which is stable up to the highest output powers. Differential efficiency up to 85% is demonstrated. Total continuous wave single mode output power as high as 120 mW is achieved in 4 micrometer-wide stripes. Infrared (980 nm) InGaAs/AlGaAs PBC lasers with a beam divergence down to 4.2 degrees and a high temperature stability of the threshold current are also demonstrated.

Miniaturized optical resonators with spatial dimensions of the order of the wavelength of the trapped light offer prospects for a variety of new applications like quantum processing or construction of meta-materials. Light propagation in these structures is modelled by Maxwell's equations. For a deeper numerical analysis one may compute the scattered field when the structure is illuminated or one may compute the resonances of the structure. We therefore address in this paper the electromagnetic scattering problem as well as the computation of resonances in an open system. For the simulation effcient and reliable numerical methods are required which cope with the infinite domain. We use transparent boundary conditions based on the Perfectly Matched Layer Method (PML) combined with a novel adaptive strategy to determine optimal discretization parameters like the thickness of the sponge layer or the mesh width. Further a novel iterative solver for time-harmonic Maxwell's equations is presented.

Several concepts of coherent control are extended to manipulate light propagating along metal nano-particle arrays. A phase-polarization control strategy is proposed and applied to control the electromagnetic energy transport via nano-array constructs with multiple branching intersections, leading to an optical switch or inverter far below the diffraction limit. An optimal control approach, based on the genetic algorithm optimization procedure, is next generalized to suggest a systematic design tool for plasmonic nano-devices, where both material properties of nano-arrays and incident field parameters are optimized in order to make devices with desired functionality. The proposed schemes are also used to better understand the physics underlying the phenomenon of electromagnetic energy transport via metal nano-constructs. Several applications of the phase-polarization and optimal control strategies are considered.

Laser cooling in semiconductor structures due to anti-Stokes luminescence is reviewed. Theoretical background considering luminescence trapping and red-shifting, the effect of free carrier and back-ground absorption, Pauli band-blocking, and the temperature-dependence of various recombination mechanisms are discussed. Recent experimental results demonstrating record external quantum efficiencies (EQE) in GaAs/GaInP heterostructures are described, and conditions favorable for the first observation of laser cooling in semiconductors are discussed.

Various methods leading to potential improvement of laser cooling in semiconductors are considered. It is shown that using combination of bandgap engineering in electronic domain and photonic bandgap engineering one can attain a factor of few improvement in efficiency provided the nonradiative efficiency does not get affected.

On the road toward experimental realization of laser-induced cooling of semiconductors, theoretical investigations are necessary for a detailed understanding of the microscopic phenomena underlying the cooling process, and for a prediction of optimal parameter regimes where efficient cooling could be expected. A recent realistic theory for cooling of bulk GaAs by Sheik-Bahae and Epstein has focused on the high-temperature regime, where the cooling process involves absorption of and luminescence from an electron-hole plasma. Using a microscopic many-particle theory, we extend the Sheik-Bahae Epstein approach to the low-temperature regime, where excitonic effects, i.e. effects of bound electron hole pairs in quasi-thermal equilibrium with correlated unbound pairs (i.e. the plasma) become important. We use a diagrammatic approach that is non-perturbative in the Coulomb interaction and contains effects of phase-space filling, single-particle renormalization in a partially ionized plasma, and screening. We ensure that our theory contains the relevant limiting cases for (partial) ionization in the low-density regime (Saha equation and Beth-Uhlenbeck formula) as well as the high density regime (Mott transition). Based on our microscopic theory for absorption and luminescence in the quasi-thermal equilibrium regime, we present a detailed study of cooling criteria at low temperature, focusing mainly on the temperature regime between 5K and 100K. In particular, we discuss the transition from the high temperature regime dominated by absorption in the e-h continuum to the low-temperature regime dominated by resonant exciton absorption.

We demonstrate a non-contact, spectroscopic technique to measure the temperature change of semiconductors with very high precision. A temperature resolution of less than 100 μK has been obtained with bulk GaAs. This scheme finds application in experiments to study laser cooling of solids. We measure a record external quantum efficiency of 99% for a GaAs device.

We have calculated the recombination rates of electrons and holes in a quantum dot ensemble. The calculation treats the various recombination mechanisms as localized events at the dot site and uses global population statistics, re-interpreted to provide integer occupancies of the ground and excited states of the dots. The results show that the localization means that unambiguous functional forms describing the recombination in terms of charge carrier density can no longer be determined. The evolution of the recombination rates with increasing injection is complex and the different radiative and non-radiative routes have a similar dependence. Experimentally this means that studies of the light output versus current can no longer be used to assess the dominant recombination process.

The effects of phonon bottleneck effect on the performance of QD laser devices are studied using the rate equation model. Due to phonon bottleneck effect, simultaneous laser oscillations of ground state and excited state can occur. By comparing experimental data and theoretical prediction, the estimated value of intra-dot relaxation time is more than 100ps.

The performances of all-optical logic gates AND, OR, XOR, NOT based on quantum dot semiconductor optical amplifier (QD SOA) devices have been simulated. The saturation power, optical gain and optical phase response of a QD SOA has been analyzed numerically using a rate equation model of quantum dots embedded in a wetting layer. The calculated response is used to model the performance of the logic gates. Impacts of injection current and the input signal power on system quality factor have been studied. For the parameters used in this paper, all-optical logic gates using QD-SOA is capable of operating at speeds of ~ 250Gb/s.

Optical coherence time and nonlinear optical response of excitons localized to quantum dots in a GaAs/AlGaAs single quantum well are investigated by the three-pulse photon echo technique. At low temperature, the optical coherence time is estimated to be 1.8 ns from the decay curve of the photon echo signal. The decay profile for short time delay shows the strong dependence on the input pulse area of the third pulse. Rabi oscillations are also observed. The oscillations exhibit nonlinear behavior, which is affected by the pre-excited polarization. Such behavior can be qualitatively explained by the local field effect due to the exciton-exciton interaction between the excitons.

Photoluminescence from a single semiconductor quantum dot after perfect resonant excitation has been temporally resolved with picosecond time resolution. Making use of non-confocal geometry for the micro photoluminescence setup, we can greatly remove the elastic scattering component, and successfully capture the resonantly-excited spontaneous emission from a single quantum dot. The emission intensity as a function of coherent photoinjection shows peculiar oscillatory dependence, i.e., Rabi oscillations of ground-state excitons. The transition strength of the groundstate exciton is precisely determined through the analysis. The value indicates a good agreement with that expected by the radiative decay rate of the quantum dots.

Properties of atom-like emitters in cavities are successfully described by cavity quantum electrodynamics (cQED). We report on cavity quantum electrodynamics (cQED) experiments in a single quantum dot semiconductor system. CQED, which is a very active research field in optics and solid state physics, can be divided into a weak and a strong coupling regime. In case of weak coupling, the spontaneous emission rate of an atom-like emitter, e.g. a single quantum dot exciton, can be enhanced or reduced compared to the value in vacuum in an irreversible emission process. In contrast, a reversible energy exchange between the emitter and the cavity mode takes place when the conditions for strong coupling are fulfilled. We investigate weak as well as strong coupling in a system based on a low density In_0.3Ga _0.7As quantum dot layer placed as the active layer in a high quality planar AlAs/GaAs distributed Bragg reflector cavity grown by molecular beam epitaxy. Using electron beam lithography and deep plasma etching, micropillars with high Q-factors (up to 43.000 for 4 μm diameter) were realized from the planar cavity structure. Due to the high oscillator strength of the In_0.3Ga _0.7As quantum dots together with a small mode volume in high finesse micropillar cavities it is possible to observe strong coupling characterized by a vacuum Rabi splitting of 140 μeV. The fabrication of high-Q micropillar cavities as well as conditions necessary to realize strong coupling in the present system are discussed in detail.

We investigate the dephasing of excitons in InAs self-assembled quantum dots by using a transient four-wave-mixing technique. A used sample is specially designed to compensate the strain. We observe long-lived coherence of excitons at 5 K which corresponds to the dephasing time longer than a nanosecond, where the photon energy of the excitation pulse is 0.874~eV. We find that a pure dephasing due to exciton-phonon interactions dominates in the exciton dephasing rather than the population decay and the exciton-exciton interaction in the weak excitation region, by analyzing the population lifetime and the polarization-dependent dephasing time.

We theoretically study pump and probe spectra of a quantum dot (QD) embedded in a microcavity in the cavity QED treatment. Energy levels of a QD are modeled by a four-level system consisting of the ground state, right and left-polarized excitons, and a biexciton. As a cavity-QED effect, four spectral peaks appear when the pump field frequency is tuned to the lower or upper level constituting the vacuum-field Rabi splitting. In addition, the spectral intensity dramatically enhances due to the cavity effect. The enhancement rate is independent of the coupling energy between exciton and cavity photons when the well-defined vacuum-field Rabi splitting is formed.

Various samples from the GaInNAs dilute nitride material system are modeled microscopically and good agreement with experiment is shown for the optical gain, linewidth enhancement factor, photomodulated reflectance and photoluminescence. Even though the differential gain is reduced by the inclusion of nitrogen, the linewidth enhancement factor is shown to stay almost unchanged. Radiative decay times are calculated and show a strong change in their density dependence above threshold.

We present results of the analysis of effective masses and optical gain on InGaAsN based quantum-well structures. In our work we have considered self-consistent effects which are determined by a simultaneous solution of the Luttinger-Kohn equations for electron and hole wavefunctions and the Poisson equation for heterostructure potential modification. We describe systematic numerical results for a large range of material and structural parameters. Our results show that significant variation in the value of effective masses and optical gain is possible by adjusting the relevant parameters. We also indicate the effects due to self-consistency.

It is demonstrated that fully microscopic many-body models are required for a correct description of the dominant carrier loss processes in semiconductor lasers, spontaneous emission and Auger recombination, and that they are able to quantitatively predict these losses. The density dependence of the losses assumed in semi-empirical approaches, J=AN+BN²+CN³, is shown to break down already near transparency. For the spontaneous emission it is shown to decrease from quadratic to linear (BN), Auger rates are shown to increase only quadratically (CN²) or even less.

We propose and demonstrate theoretically a method of achieving single lateral mode field patterns in CW broad area laser diodes at high power. The method relies on laterally profiling the thermal conductivity of the bonding solder layer and thereby thermally controlling the lateral index. Various configurations are simulated using a split step FFT beam propagation method including thermal, carrier (alpha-parameter) and optical (Kerr effect) index variations. Thermal effects are described using a 2-D finite element electro-thermal model including current spreading.

A novel 'W' optical waveguide using AlGaAs-GaAs semiconductor material systems is presented and analyzed for achieving a continuously-phase-matched difference frequency generation (DFG) nonlinear process. Our approach utilizes two near-IR lasers to obtain an efficient generation of coherent mid-IR radiation. By maintaining a continuous phase matching condition in our 'W'-waveguide, our studies show that nonlinear DFG process obtained from 920-nm and 1140-nm lasers with input powers at 0.5 Watts (with interaction length of 5-mm) results in coherent continuouswave (CW) power of ~ 0.13 mW at wavelength of 4.77-μm. The normalized conversion efficiency (η_nor) of the DFG process is ~ 40 W^-1m^-2, as a result of low spatial mode overlap. An optimized 'W' waveguide consisting of large index contrast core region structure, with an improved spatial mode overlap of the near and mid-IR modes, allows a significant improvement in the normalized conversion efficiency by up to approximately two orders of magnitude (η_nor > ~ 104 W^-1m^-2).

In this paper, we demonstrate the numerical simulations of the organic light-emitting device (OLED) based on a rigorous and efficient numerical method. The input parameters of such a program include the layer thickness and complex refractive index of each layer, the locations and density of the oscillating dipoles, and the emission photoluminescence spectrum. In evaluating the device performances, the output spectrum, the intensity distribution, and the viewing-angle characteristics of an OLED are concerned. Since the numerical difficulty arising from the large thickness of the glass layer in the OLED is carefully overcome, the program can simulate the optical performances with different glass substrate thickness ranged from less than one to hundreds of μm. When the glass substrate becomes thinner, multi-peak spectrum of an OLED is observed due to the strong interference effect between the two sides of the glass substrate. When the thickness of the glass substrate is reduced to less than 1 μm, the device is identical to a top-emission OLED with a two-microcavity structure. The simulation results are consistent with the Fabry-Perot cavity equation, which can be used as a guideline for optimizing the optical characteristics of an OLED from the normal direction. We have also demonstrated the procedures to maximize the total flux from an OLED which is more important than the luminance from the normal direction for the lighting application. Since our development of numerical algorithms is based on the general electromagnetic theory, the proposed model is, in principle, applicable to an OLED consisting of any number of layers.

We study photonic crystal (PhC)-assisted light extraction from gallium nitride (GaN) light-emitting diodes (LEDs). We focus on the issue of omnidirectional extraction, and we introduce a complex crystal lattice, namely the Archimedean tiling, which enables efficient and omnidirectional light extraction. PhC LEDs with regular triangular lattices and Archimedean lattices are studied experimentally by angle-resolved luminescence, and the theoretical properties of the Archimedean tiling are confirmed experimentally.

Light extraction efficiency of nitride-based blue-light-emitting diodes on SiC substrate is theoretically investigated using rigorously coupled wave analysis method. Rigidly periodic moth eye structure is found to show superior light extraction efficiency compared with conventional rectangular parallelepiped or pseudo-moth eye structures. High-throughput technology by laser interference lithography for the fabrication of rigidly periodic moth eye structure is established using a 244 nm laser as the light source.

A super-luminescent light emitting diode (SLED) operating around 1300nm is simulated. This edge-emitting device is grown on an InP substrate and comprises multiple quantum wells. For the simulation the focus is put on the amplified spontaneous emission (ASE) spectra for different input currents as well as on the output power vs. current curves (light vs. current---LI). Simulated ASE spectra agree very well with measurements over a large wavelength range (more than 80nm). Regarding the LI-characteristics good agreement between simulations and measurements is obtained for input currents from 10mA up to 150mA. The simulated electrical characteristics of the device are obtained by solving drift-diffusion equations, the optical problem is solved by decomposing the vectorial (3D) Helmholtz equation into a transverse and into a longitudinal part.

Optoelectronic integrated circuits incorporating twin ring diode lasers with InAs/InGaAs/GaAs quantum-dot active region have been fabricated and characterized. Directional control and unidirectional operation of ring diode lasers are demonstrated by forward biasing an S-section waveguide incorporated within the ring cavity. Mode-beating spectra from individual ring diode lasers are observed at three bands near 7.5 GHz, 16.6 GHz, and 25.1 GHz, corresponding to single, double, and triple longitudinal mode spacing in the ring cavity. In addition, mode beating spectra between optically independent integrated twin ring diode lasers are also demonstrated, with minimal linewidth of ~4 MHz.

A multimode model based on the rate equation approximation is presented in order to simulate the mode dynamics and noise characteristics of semiconductor microring lasers. The effect of straight waveguide facet residual reflectivity is studied in detail. A complete characterization is performed including calculated optical mode spectra, time traces and relative intensity noise (RIN) spectra. Different operation regimes are demonstrated, stable multimode, stable single mode, and hopping multimode dependent on the injection current and the straight waveguide reflectivity. An expansion of the stable single mode operation regime is observed up to higher injection currents with a respective suppression of the mode-hopping regime, for a power reflectivity ratio of -30dB for both facets. Moreover, the use of different reflectivities for each facet at appropriate values promotes unidirectional operation.

Microring resonators are excellent candidates for very large scale photonic integration due to their compactness, and fabrication simplicity. Moreover a wide range of all-optical signal processing functions can be realized due to the resonance effect. Possible applications include filtering, add/drop of optical beams and power switching, as well as more complex procedures including multiplexing, wavelength conversion, and logic operations. All-active ring components based in InGaAsP/InP are possible candidates for laser sources, lossless filters, wavelength converters, etc. Our work is based on measurement, characterization and proposal of possible exploitation of such devices in a variety of applications. We investigate the spectral characteristics of multi-quantum well InGaAsP(λ=1.55μm)/InP microring structures of various ring diameters and different configurations including racetracks with one or two bus waveguides and MMI couplers. The latter configuration has recently exhibited the possibility to obtain tunable active filters as well as tunable laser sources based on all-active ring-bus-coupler structures. In the case of tunable lasers single mode operation has been achieved by obtaining sufficiently high side mode suppression ratio. The tuning capability is attributed to a coupled cavities effect, resembling the case of multi-section DBR lasers. However, in contrast to the latter, the fabrication of microring resonators is considered an easier task, due to a single step growth procedure, although further investigation must be carried out in order to achieve wide range tunability. Detailed mappings of achievable wavelengths are produced for a wide range of injection current values.

The 1.3-μm semiconductor material systems are numerically studied with a LASTIP simulation program. The optimum active layer materials of AlGaInAs/InP and InGaNAs/GaAs system are suggested. For the AlGaInAs/InP system, we optimize the structure by varying the number of quantum wells, the linear GRINSCH, and the compensated tensile strain in barriers. The optimized active structure possesses four quantum wells, linear GRINSCH, and a compensated tensile strain in the barrier of 0.325% at an emission wavelength of 1.3 μm. The characteristic temperature can be improved to 99.4K, 51.0K, and 68.6K as it is operating among 288K~318K, 318K~348K, and 288K~348K respectively. Furthermore, the optimized structure can also enhance the stimulated recombination rate and reduce the Auger recombination rate because of the compensated tensile strain in barriers. The simulation results show that the active layer with a certain amount of compensated tensile strain in barriers is beneficial for improving the laser performance. On the other hand, the performance of the InGaNAs/GaAs lasers with quantum wells of different compressive strains is investigated. The wavelength of InGaNAs/GaAs system is about 1.3 μm if the Ga composition in quantum wells is 0.54. The results of numerical simulation suggest that the stimulated recombination rate is larger and the Auger recombination rate is smaller when the Ga composition in quantum well is 0.50.

The wavelength of a semiconductor laser diode (LD) is modulated using the self-mixing effect to create the undulation wave. In order to determine the direction of a moving target, the number of undulations is added to the undulations of detected Doppler frequency shift. These modulated undulations are generated through the difference between the rising and falling counts of a triangular modulated current waveform of LD. A theoretical model is proposed to relate the wavelength modulation and current modulation of LD to predict the number of undulations so that the direction of a moving target can be determined.

The study of noise spectrum, in particular of the RIN, in semiconductor diodes lasers is a powerful tool to determine the proper frequencies and the operation regimes of the device. Following this a study of the RIN is of major importance to characterize the laterally coupled diode lasers. In this work a study of the noise spectrum dependence with the lateral separation between waveguides as well as its dependence with relative bias current applied is made. These results combined and compared with spectrally resolved near and far field allows the definition of which are the operating regimes of the device.

The present work discusses methods of stabilizing the frequencies of commercially-available laser diodes. Laser diodes are generally compact and long-lived. The frequency stability, which makes them ideal for onboard laser interferometer light-sources, in applications such as the satellite-to-satellite tracking systems used to verify fluctuations in earth's gravity field, which, in turn, indicate other critical changes in the environment, is the key characteristic of this work. We used the devices typically operating at 780nm, and their frequencies can be stabilized using either of two systems; one, employing the Doppler-free absorption line of Rb atoms and another, sing the Faraday effect of the Rb absorption line. In both cases, the use of the proper modulation frequency and amplitude improved frequency stability, overall, attaining 2.05×10^-12 and 2.73×10^-11, respectively, in the square root of the Allan variance, by measuring the beat-note between two independently-stabilized laser diodes.

A theoretical study of photoluminescence refrigeration in semiconductors has been carried out using a model that takes into account photon recycling and includes the rate equations for both carriers and photons. General expressions for cooling efficiency, cooling power density, and the cooling condition are derived. The investigation of the photoluminescence refrigeration in an intrinsic GaAs slab shows that net cooling is accessible when quantum efficiency and luminescence extraction are high, and that photon recycling contributes strongly to photoluminescence refrigeration when the luminescence extraction efficiency is small.

We modify the Soref single-mode criterion such that it can be applied also to waveguides with small cross sections. This is verified for highly sensitive silicon on insulator (SOI) rib waveguides with core thicknesses of 60 nm for TE- and 220 nm for TM-polarization, respectively, at a wavelength of 1550 nm by a fully vectorial finite element method (FEM) eigenmode analysis. The semi-analytical criterion enables an efficient design of rib waveguides for evanescent wave sensing devices.

An exact analytical solution is derived for the variance of the timing jitter of a dispersion-managed soliton in the presence of synchronous amplitude modulators and filters. We show that, for a given position of these control elements, a total suppression of the timing jitter is possible, which will permit unlimited error free transmission distances of DM solitons. However, if only the modulator or the filter is used, the asymptotic behavior of the timing jitter shows a linear dependence with distance, which is in contrast with the cubic dependence in the uncontrolled case.