We review progress in realizing a semiconductor source of single photons and photon pairs based on the emission of individual self-assembled quantum dots. Integration of the quantum dot into a pillar microcavity produces a strong Purcell enhancement of the radiative recombination rate, resulting in photon collection efficiencies into a lens of ~10%. The residual multi-photon emission is found to derive from the emission of other layers within the structure, such that under resonant laser excitation of the dot a greater than 50-fold reduction in the 2-photon rate can be achieved compared to a laser of the same average intensity. The polarization of the emitted photons can be controlled and selected in appropriately designed cavities. Through careful control of the dot growth conditions, we realize a single photon source at the fiber compatible wavelength of 1300nm. This is achieved by utilizing a second critical InAs coverage to produce a low density of large, long wavelength InAs quantum dots. We demonstrate also an electrically driven planar cavity structure with photon collection efficiencies into a lens of ~5%, corresponding to an order of magnitude enhancement in the photon collection compared to dots in a bulk semiconductor LED. Single photon emission is demonstrated for both the biexciton and exciton state of the quantum dot.

We demonstrate the entangled photon pair generation via biexciton-resonant hyper parametric scattering in CuCl crystal. Quantum tomographic analysis of the polarization correlation measurements showed that the generated photon pairs had a high degree of polarization entanglement, in consequence of J=0 angular momentum of the biexciton state.

Single-photon sources rarely emit two or more photons in the same pulse, compared to a Poisson-distributed source of the same intensity, and have numerous applications in quantum information science. The quality of such a source is evaluated based on three criteria: high efficiency, small multi-photon probability, and quantum indistinguishability. We have demonstrated a single-photon source based on a quantum dot in a micropost microcavity that exhibits a large Purcell factor together with a small multi-photon probability. For a quantum dot on resonance with the cavity, the spontaneous emission rate has been increased by a factor of five, while the probability to emit two or more photons in the same pulse has been reduced to 2% compared to a Poisson-distributed source of the same intensity. The indistinguishability of emitted single photons from one of our devices has been tested through a Hong-Ou-Mandel-type two-photon interference experiment; consecutive photons emitted from such a source have been largely indistinguishable, with a mean wave-packet overlap as large as 0.81. We have also designed and demonstrated two-dimensional photonic crystal GaAs cavities containing InAs quantum dots that exhibit much higher quality factors together with much smaller mode volumes than microposts, and therefore present an ideal platform for construction of single photon sources of even higher quality.

High-efficient single-photon sources are important for fundamental experiments as well as for modern applications in the field of quantum information processing. Therefore, both the overall collection efficiency as well as the photon generation rate are important parameters. In this article, we use cascaded two-photon emission from a single quantum dot in order to double the efficient transmission rate in a quantum key distribution protocol by multiplexing on a single photon level. The energetically non-degenerate photons are separated with a single photon add/drop filter based on a Michelson interferometer. For optimizing the collection efficiency, coupling of quantum emitters to microcavities is advantageous. We also describe preliminary results towards coupling of a single quantum dot grown on a micrometer-sized tip to the whispering gallery modes of a microsphere cavity.

For understanding the fundamental processes in QDs and optimizing the design of QD optical devices, it is essential to obtain accurate optical gain and absorption spectra. An improved segmented-contact method is described that subtracts the unguided spontaneous emission that normally introduces error into the calculated gain and absorption. Using the technique a QD gain spectrum is measured to an accuracy of less than 0.2/cm at nominal gain values below 2/cm. This capability also enables precise measurement of waveguide internal loss, unamplified spontaneous emission spectra and Stark shift data.

The characteristics of p-doped 1.1 μm and 1.3 μm self-assembled In(Ga)As quantum dot lasers grown by molecular beam epitaxy have been studied. With optimum p-doping, we demonstrate quantum dot lasers with zero-temperature dependence of the threshold current (T₀ = ∞) and the output slope efficiency. These characteristics are explained through a self-consistent model that includes temperature-dependent Auger recombination in the quantum dots. With tunnel injection, we measure greatly enhanced -3dB frequency response, 25 GHz and 11 GHz in 1.1 μm and 1.3 μm tunnel injection quantum dot lasers, respectively. These devices also exhibit near zero α-parameters and extremely small chirp (< 0.2 Å), in addition to temperature insensitive operation.

We measure, for the first time, the gain compression coefficient and above-threshold linewidth enhancement factor (alpha parameter) in quantum dot (QD) distributed feedback lasers (DFB) by time-resolved-chirp (TRC) characterization. The alpha parameter is measured to be 2.6 at threshold and increases to 8 when the output power of the QD DFB is increased to 3 mW. The dependence of the above-threshold alpha parameter on the optical power is found to be stronger than the optical gain compression effect alone can predict. The inhomogeneous gain broadening, gain saturation at the ground states and carrier filling in the excited states in QDs are proposed to explain the results.

The dynamic characteristics of an InGaAs/InP lattice-matched V-groove quantum wire laser are theoretically considered. The differential gain of the laser is derived from the k•p theory, taking into account nonparabolicity of both bands, the conduction and the valence band. We investigate the bandwidth of the laser versus its optical confinement, driving current and roll-off time, including nonlinear gain suppression. In spite of solid differential gain and the D-factor comparable with the best-performance QW laser, the maximum bandwidth of the laser is limited to about 20GHz due to insufficient optical confinement.

Semiconductor lasers' oscillation wavelengths respond to temperature- and magnetic field-fluctuations. As reported in the 1960's, these wavelengths shortened, at lower than 80K, under a magnetic field larger than 4T. This phenomenon attracted our attention, during preliminary experiments, because, when we exposed bulk-type semiconductor lasers oscillating at 780nm to relatively weak magnetic fields (less than 1.4T), at room temperature (300K), we observed that the oscillation wavelength shifted to the longer (lower frequency) side. For this work, we focused on the longer oscillation wavelength shift, the lower power side optical output-power shift, and the higher voltage side terminal voltage shift of a number of multi-quantum-well (MQW) laser diodes oscillating at 780nm, under the same experimental conditions as those used in our previous works. In discussions of these shift mechanisms, we consider how wavelength (frequency), optical output-power, and terminal voltage shifts are correlated. Our expanded knowledge base has forced us to employ mechanisms completely different from those used in studies dating as far back as the 1960's. We are now introducing a mechanism in which temperature rise and the longitudinal magneto-resistance effect cause the shifts observed in our experiments.

A comprehensive study of the oblique frozen modes in a finite one-dimensional anisotropic photonic crystal is presented. The most attractive property of these waves is near-zero axial group velocity in conjunction with small reflection from the lattice interface. We use a characteristic matrix analysis to describe both monochromatic wave and pulse propagation in one-dimensional anisotropic photonic crystal of finite length. Using a layered structure consisting of air and an example of a suitable material, SbSI, we demonstrate the slow down factor, intensity enhancement, bandwidth, and transmission coefficient of axially frozen modes.

An iterative algorithm for band structure of lossy 2D photonic crystal is presented in this paper. This iterative algorithm is effective and the real part of the eigenvalue converges fast and accurately when the dielectric function is complex (ε(f)=ε₁ + i•ε₂(f)) and dependent on frequency in lossy cases with ε₂(f)≈0.1ε₁. Effects of the frequency dependent dielectric function on band structure of a particular 2D photonic crystal is discussed.

In photonic crystal (PC) microlasers with point defect cavities, effective carriers are reduced by the leakage to outside of the defect, surface recombination, spatial carrier hole burning, and Auger recombination. To estimate these effects, we calculated carrier and photon behavior by solving two-dimensional rate equations in space and time domains. The result clearly shows these effects and their dependence on cavity structure, pump area, and so on. Compared with that for the microdisk laser, higher threshold values are estimated for PC microlasers. However, a comparably low threshold density and a high efficiency are expected for the quasiperiodic PC microlaser, because the hole burning by the whispering gallery mode of this cavity suppresses the carrier leakage.

The photonic crystal fiber (PCF) and the PCF based structure are playing more and more important role in the optical communication and optical sensors fields. Fabrication technique is the key for realizing the design of PCF. In this paper the fabrication process of PCF is described, which includes stacking, jacketing, collapsing, stretching and drawing on a conventional drawing tower. To maintain the uniform air hole structure, positive micro-pressure has been introduced in the drawing processing. The multi-pole method is used to analyze the PCF structure with one hexagonal array of air cylinder photonics crystal fibers. The theoretical and experimental results show that the PCF fabricated under this way has good performance and coincidence indicator. Several PCF based structures have been studied and developed. It is predicted that the PCF based structures have some funny characteristics, which could find important application in the fiber-optic communication and sensing systems.

The Tilted Cavity (TC) concept has been proposed to combine advantages of edge- and surface-emitting lasers (detectors, amplifiers, switches, etc.). Tilted Cavity Lasers (TCL) enable wavelength-stabilized high-power edge and surface emitters (TCSEL) in low-cost single-epitaxial step design. The concept covers numerous applications including mode-locked TCL for light speed control, dispersion and linewidth engineering, GaN-based light-emitters, electrooptic wavelength tunable devices, and other applications. Presently, wavelength stabilized TC operation is realized between -200°C and 70°C in broad TCL diodes with cleaved facets based on quantum dots (QDs). The spectral width is below 0.6 nm in broad area 100 μm-wide-stipe devices. The far fields are: 4° (lateral) and 42° (vertical). Wavelength-stabilized 1.16 μm and 1.27 μm edge-emitting QD TCL lasers are demonstrated. Quantum well TCL demonstrate high-temperature operation up to 240°C with a low threshold, high temperature stability and improved wavelength stability. The tilted cavity approach can also be applied in wavelength-optimized photodetectors, switches, semiconductor optical amplifiers, including multi-channel devices, in optical fibers, in photodetectors, in light-emitting diodes and in many other applications. Moreover, microelectronic devices based on similar tilted angle resonance phenomena in quantum wells and superlattices can be realized in electron- or hole-wavefunction-engineered structures, thus, merging the fields of nanophotonics and nanoelectronics. The tilted cavity concept can be further complimented by lateral patterning and (or) processing of three-dimensional photonic crystal structures further extending horizons of modern optoelectronics.

The microgear cavity is a microdisk surrounded by a circular Bragg grating. It is well known for its high Q Whispering Gallery Mode (WGM) and its modal selectivity. Microgears are often simulated thanks to two-dimensional (2-D) Finite-Difference Time-Domain (FDTD) computations which are limited by their cartesian grid and experience a high numerical complexity. In this paper, a fast and accurate 2-D method describing the WGM in a microgear dielectric resonators is presented. The model is based on the Floquet Bloch formalism. The field is described analytically within the disk and outside the grating. The field within the grating is calculated with a finite difference formalism in polar coordinates. The resonant wavelength and quality factor can be deduced from the eigenvalue problem. Our method has been compared to the Coupled Mode Theory and to 2-D FDTD computations, it proves to be more accurate and much faster than both methods (few seconds versus few hours for FDTD). Moreover, we have demonstrated a polarization effect of the microgear. Finally, our model can be applied to different structures. Micro-flowers and microgear surrounded by multi-layer Bragg reflector are investigated.

Novel method to realize 1300-1550 nm InGaAsNSb quantum well (QW) structure on GaAs substrate is presented and analyzed. The proposed method is based on nitrogen-based inter-diffusion from InGaAsN nanometers-thick barriers into InGaAsSb quantum well layer. This method should allow realization of InGaAsNSb quantum well layer, in particular by metalorganic chemical vapor deposition (MOCVD), which would otherwise be very challenging due to the conflicting nature of optimum epitaxy parameters / conditions that nitride-based and antimonide-based compounds require.

By introducing triangular holes into oxide confined 850nm monolithic vertical-cavity surface-emitting lasers (VCSELs), single-transverse-mode operation has been obtained for large oxide apertures of 14-15 microns. The two-dimensional triangular holes etched on the device surface were aligned circumferentially along the aperture perimeter, with their tips surrounding the device center. When the holes had a relatively large lateral penetration into the oxide aperture, the holey VCSEL lased with a single spot near field pattern with a high side-mode suppression ratio (SMSR) of 45-50dB, and an output power of 2mW. In this case, it is assumed that the triangular holes are acting as a highly mode selective loss mechanism. On the other hand, when the penetration of the holes was relatively small, an SMSR of 40dB was obtained from a large area "floral" near field pattern, with a record high single-mode output power of 7mW. The lasing spectrum and far field intensity profile of this "floral" type emission indicates that it is a somewhat deformed fundamental mode that is extending over the whole device, and oscillating in-phase. The ability of triangular holes to suppress high order modes in large area oxide confined VCSELs should be effective for systems with wavelengths other than 850nm as well.

We analyzed lateral mode characteristics in 850-nm GaAs-based VCSELs with holey structures by FDTD method. Full three-dimensional analysis, which requires huge computer resources, was carried out using cluster computer. As a result, we estimated the good mode selectivity in circular and triangular holey structures, which cannot be obtained in a simple oxide aperture structure. It is explained by the large ra-diation loss from the inter-hole spacing and scattering loss at the bottom of holes particularly for higher order modes. In addition, the experimental result reported by Furukawa, et al., which showed the record high power single mode operation by the floral fundamental mode in the triangular holey structure, was verified. For any changes of standard structural parameters, only the Gaussian-like fundamental mode was confirmed. The floral mode was observed only when we assumed a nonuniform refractive index around the holey structure. Such nonuniformity can be considered in actual devices because of the non-uniform distribution of carriers and temperature at high current injection level.

We describe and demonstrate a method of decreasing the divergence angle of multi-mode VCSELs, and show how we can obtain a low and stable divergence angle. We first explain the relationship between the lateral wave-vectors of resonant modes and the divergence angle. Then we attempt to optimize the oxide aperture and the electrode structure. Here, we calculate the electro-magnetic field of the VCSELs by the Finite Difference Time Domain (FDTD) method and the far-field pattern by combining the diffraction integral and the FDTD. Finally, we compare the theoretical and experimental results of the divergence angle of the VCSELs.

A novel spatially distributed noise model is used in a device simulator in order to describe relative intensity noise and frequency noise for semiconductor lasers. For charge carrier transport, continuity and Poisson equations are used and self-heating is considered by a thermodynamic equation. Spontaneous and stimulated recombination are calculated in the framework of the semiconductor Bloch equations using the second Born approximation to include many-body effects. The optical field is expanded into modes. The temporal behavior is described by a photon rate and a photon phase equation for each mode. Noise is taken into account by spatially distributed Langevin forces. The correlation functions are described directly in the frequency domain assuming small signal noise sources. All relevant equations are solved in a fully self-consistent fashion. Comparison of static characteristics and dynamic characteristics, such as relative intensity noise, with measurements show excellent agreement for a vertical-cavity surface-emitting laser (VCSEL).

We investigate the dependence of the polarization of an optically pumped vertical-cavity surface-emitting laser (VCSEL) on the degree of electron spin polarization in the active region at room temperature. We show that the output polarization of the laser can be unambiguously controlled by the pump polarization even with low spin polarizations. Less than 30% electron spin polarization in the active region is enough to achieve 100% output polarization of the VCSEL. The dependence of the polarization of the VCSEL emission on the degree of electron spin polarization in the active region is investigated at room temperature. Electrical spin injection via ferromagnetic contacts into LED structures has been shown to be possible with efficiencies close to ten percent at room temperature. We suggest to combine ferromagnetic contacts with VCSELs because the nonlinearity of the laser at threshold can potentially be used to convert small spin injection efficiencies into a large effect onto the degree of polarization of the emitted light.

The wavelength of the designed two-section laser diode can be modulated by direct current modulation. The modulation speed of intensity of each wavelength is 1.5 times faster than the speed of a direct modulated single section laser. The modulation depth of output intensity can be 20dB.

In this work a study of the dependence of the coupling in laterally coupled diode lasers (LCDL) with the relative bias conditions is presented. The study is made by the analysis of the spectrally resolved near and far field optical spectrum combined with the frequency responses of these devices at different bias conditions. By the analysis of these measurements it was observed that three different operation regions appear, and are identified by the spectral phase relation between the fields emitted by each laser stripe.

We fabricated 1-cm-long semiconductor ring lasers monolithically integrated with passive waveguides and photodiodes. Longitudinal mode beating and fine tuning is investigated in a single ring and in a system of two traveling-wave ring lasers with opposite lasing directions.

The chaotic dynamics of a semiconductor laser subject to a delayed polarization-rotated optical feedback is investigated theoretically and experimentally. An extension of the usual one-polarization model is derived to account for two orthogonal polarizations of the optical field. The two-polarization model is motivated by observations of lag synchronization in our experiments using polarization-rotated optical feedback and uni-directional injection. Experimental data confirm the predictions of the two-field model. We also show that the two-polarization model can be reduced to the one-polarization model.

There is an increasing need for single-spatial-mode, edge-emitting semiconductor lasers with reliable cw output power of around 1 W for applications such as pumps for rare-earth-doped fiber amplifiers and free-space communications. The design of respective devices is still a challenging task for experimenters, and software can assist very much in doing analyses of potentially perspective designs. We have developed a 3D numerical code supplied with a user-friendly interface for active diode-laser structures, taking into account light diffraction and carrier diffusion. The Watt-Ampere characteristics are calculated by changing the drive current density in the equation for the carrier-number density. To evaluate a single-mode stability limit, a procedure is developed to calculate 3-5 higher order optical modes on a 'frozen background': gain, carrier-induced index variation, as produced by the operated mode at a fixed drive level. Modal gains of these modes are compared to the calculated threshold gains for each mode. Because of non-uniform gain saturation by the operated mode, modal gains for higher-order modes increase with drive current due to beneficial overlap of their fields with the gain. When one of the higher-order modes approaches its threshold, this puts an upper limit for stable single-mode operation. A graphical interface allows for viewing near- and far-field patterns of any mode in the form of 3D surfaces or contour-plots. Scanning of profiles of mode intensity in an arbitrary cross section in the output plane and in far-field zone is available, too. Results of analyses of a number of published designs are reported.

A comparative simulation study of the optical output characteristics of tapered lasers with different epitaxial structure was performed. The simulation model self-consistently solves the steady state electrical and optical equations for the flared unstable resonator and was previously backed by experiments on one of the simulated structures. Three different epitaxial designs emitting at 975 nm were analyzed: a standard single quantum well symmetrically located in the confinement region (s-SQW), a double quantum well also symmetrically located (s-DQW) and an asymmetrically located double quantum well (a-DQW). The symmetric structures have different confinement factor but a similar ratio between the active layer thickness and the confinement factor, dQW/Γ, while the a-DQW has similar confinement factor than the s-SQW, but double dQW/Γ. A better performance is predicted for the a-DQW design, reaching considerably higher output power with good beam quality. The results are interpreted in terms of a lower density of power in the QW in the case of the a-DQW design, thus delaying to higher output power the onset of the non-linear effects that degrade the beam quality. The role of dQW/Γ as a figure of merit for high brightness tapered lasers is emphasized.

Gain Lever, an effect for enhancing amplitude modulation (AM) efficiency in multisection laser diodes1, has been characterized in InGaAs DQW edge emitting lasers that are integrated with passive waveguides. Specifically designed structures which give a range of split ratios from 1:1 to 9:1 have been fabricated and measured to fully characterize the parameter space for operation in the gain lever mode. Additionally the experimental results are compared to a hybrid 3-D simulation involving effective index method (EIM) reduction to 2-D. Gains greater than 6 dB in the AM efficiency can be achieved within the appropriate operating range, but this gain drops rapidly as the parameter range is exceeded. High speed RF modulation with significant gain is, in principle, possible if proper biasing and modulation conditions are used. This phenomenon can also be useful for high-speed digital information transmission.

The gain bandwidth of a quantum-cascade (QC) laser is important for determining the magnitude of the optical gain, the refractive index change, and the linewidth enhancement factor. We investigate the effects of the scattering mechanisms on the gain linewidth of a type-I QC laser and compare our theoretical results with experimental data. The bandwidth of the gain spectrum of a QC laser is related to the electron relaxation rate, which is determined by scatterings that change the electron momentum or energy in the same subband (intrasubband) or different subbands (intersubband). Polar optical phonon scattering, impurity scattering, and electron-electron scattering are the important mechanisms. In this paper, we investigate the magnitude of the linewidth of the optical gain spectrum due to these scattering mechanisms in type-I mid-IR QC laser structures. In particular, the dependence of the scattering rate on the doping position will be shown in the case of the impurity scattering. We also present calculated optical gain, refractive index change, and linewidth enhancement factor spectra. Our theoretical results agree well with the experimental data.

In this paper, we demonstrate clock recovery from a patterned 160Gb/s optical-time-division-multiplexed (OTDM) return-to-zero (RZ) data stream. A cascaded LiNbO3 Mach-Zehnder modulator is employed as an efficient optical-electrical mixer. A phase-locked-loop (PLL) is used to lock the cross-correlation component between the optical signal and a local oscillating signal. As a result, clock signal at 10GHz is extracted from the 160Gb/s optical TDM signal. The measured root-mean-square (RMS) timing jitter of the 10GHz clock signal is ~ 130fs.

X-crossing vertical coupler filters may be used as optical add/drop multiplexers (OADM’s) in wavelength-division multiplexing (WDM) systems to select and route different channels. The basic requirement is to effectively suppress the side-lobes and enhance wavelength selectivity, which is a complex function of structure geometry and material dispersion. To understand the mechanisms and optimize X-crossed OADM, we have studied the relationship between the performance of the coupling structure and some of their geometrical parameters such as the width of drop channel, the cladding layer thickness and the angle between two vertical waveguides. In our designed structure, the cladding layer is undoped InP and the input/through channel and drop/add channel are both in the guiding layer of InGaAsP material. The field-coupling phenomena of the coupling in the layered structures are simulated by three-dimension beam propagation method (3-D BPM). Four distinct parameters have been identified to characterize the device performance, namely the peak wavelength, 3dB bandwidth of the main peak (drop channel), side-lobe level, and crosstalk, which is defined as the relation between drop efficiency and the throughput attenuation at resonance. The results show that to obtain a decreasing trend of the peak wavelength, increasing the width of drop channel is the only efficient way. Additionally, we observe the 3dB bandwidth shrinks because of the above variation. Increasing the angle contributes more to the side-lobe level suppression, whereas the cladding layer thickness improves the crosstalk, maximizing it up to 27.22dB. The simulation gives optimization ranges for each geometrical parameter, outside which the intensity profile tends to split into two. This work provides some useful optimization results for designing the X-crossing vertical coupler based OADM.

The understanding of the phenomena related to the gain recovery of semiconductor optical amplifiers (SOAs) is necessary for the application of these devices as amplifiers or switching elements in future high speed networks. We have measured the gain recovery as a function of probe wavelength in SOAs of two different lengths but otherwise identical structure. The SOAs have InGaAs MQW active regions with peak gain in the 1550 nm window. Pump-probe measurements of recovery are made using a counter-propagating set-up with a gain-switched DFB laser as the pump and a tunable laser as the probe. Measured results for the recovery time in an SOA of length 2 mm show a strong dependence on probe wavelength with a pronounced minimum around 1580 nm, coincident with the peak of the gain spectrum of the device. Results for an SOA of length 1.2 mm indicate a rather shallow minimum around 1570 nm, which also is close to the peak gain wavelength of this device. Numerical simulation of SOA gain recovery is reported using a model that includes the material gain spectrum, saturation effects and the variation of optical intensities along the length of the device. Comparisons of simulated and measured results give a reasonable level of agreement.

A wideband steady-state model and efficient numerical algorithm for a tensile-strained InP/InGaAsP semiconductor optical amplifier is described. The model is applicable over a wide range of operating regimes. The relationship between spontaneous emission and material gain is clarified. Simulations and comparisons with experiment are given which demonstrate the versatility of the model.

The design of compact spiral erbium-doped waveguide amplifiers (EDWAs) is considered under two constraints. The first one is the minimization of the chip area required for obtaining a predetermined gain, and the second one the maximization of the gain available from a limited area. It is shown that considerable benefit in gain and compactness of spiral EDWAs can be achieved by allowing tight bends with relatively high bend-induced losses. The effect of amplified spontaneous emission along with gain degradation on the noise figure in compact spiral EDWAs is analyzed. It is found that the dependence of the optimized limited area EDWA noise figure on the chip area attains a maximum at some definite area value. It is also found that there is a range of area values, where the optimized limited area EDWA noise figure is subjected to pronounced changes under small variations of the chip area. The study is based on a modified rate-propagation equation model that takes into account bend-induced losses in curved waveguides with varying radius of curvature. Numerical calculations based on actual waveguide parameters show the possibility of fabricating high gain, small size EDWAs.

We consider some principles of the binary encoded digital modulation of light based on exploiting collinear three-wave coupled states. The analytical model of shaping acousto-optical three-wave weakly coupled states is presented. The analysis developed is tested experimentally using the collinear Bragg regime of acousto-optical interaction in uniaxial single crystal. The experiments carried out make it possible to observe the dynamics of reshaping acousto-optic three-wave weakly coupled states under variations in the acoustic pulse width and the frequency mismatch. Applying the technique of localizing the coupled states to the problem of binary modulation of light beam, the conversion of multi-bit electronic signals into binary encoded optical pulse trains is demonstrated.

Large size single crystals of beta-Ga₂O₃ with 1 inch in diameter have been grown by the floating zone technique. The stable growth conditions have been determined by the examination of the crystal structure. Wafers have been cut and fine polished in the (100), (010) and (001) planes. These were highly transparent in the visible and near UV, as well as electrically conductive, indicating the potential use of beta-Ga₂O₃ as a substrate for optoelectic devices operating in the visible/near UV and with vertical current flow. Epitaxial growth of nitride compounds by the metalorganic vapor phase epitaxy (MOVPE) technique is demonstrated on beta-Ga₂O₃ single crystal substrates. High-quality (0001) GaN epi-layers with a narrow bandedge luminescence are obtained using a low temperature conductive buffer layer. InGaN multi-quantum well (MQW) structures were also successfully grown. The first blue light-emitting diode (LED) on beta-Ga₂O₃ with vertical current injection is demonstrated.

We have experimentally observed the time evolution of the photoluminescence spectra of InGaN/GaN quantum wells with widths 3 and 4 nm in response to pulsed excitation at room temperature. We find that for both well widths the time evolution of the energy-integrated photoluminescence increases initially then decays and the spectrum displays a blue shift of the peak energy which then reverses. Through an iterative simulation of the carrier density, piezoelectric field and radiative recombination rate we calculate the behavior of these quantum well systems and find good agreement with the experimental data. The internal field present in the InGaN/GaN system is screened as carrier density increases, which combined with band filling and coulomb interactions result in a blue shift as the system is pumped and as recombination of the carriers occur a red shift is simulated. Although screening of the internal fields occurs our calculations show that at laser threshold there is still a large internal field present, 1.0 MVcm^-1, which is 75 % of the unscreened value.

In this work, a systematic modeling study of polarization-induced internal field effects on the gain spectrum and threshold current density was performed. Two laser diodes of technological relevance: the In_0.2Ga_0.8N/GaN hexagonal nitride laser structures grown along the polar c-axis with internal field values up to 1.8MV/cm, and the In_0.1Ga_0.9As/Al_0.15Ga_0.85As (111)B laser structures with internal field values up to 100kV/cm were studied. The gain model is based on a self-consistent solution of Poisson-Schroedinger equations, and takes into account strain effects, free carrier screening, and the field dependence of gain and spontaneous emission rate. In the nitride case, some of our main findings are: (a) assuming a laser structure with a single In_0.2Ga_0.8N/GaN quantum well (QW) and a modal gain Γg=30cm^-1, the optimal QW width in terms of lowest current threshold is ≈3nm. (b) For a 3nm-wide QW and Γg=30cm^-1, the presence of the internal field increases the threshold current over the zero-field value by at least a factor of three. This factor increases further for higher Γg's. (c) The optimal number of In_0.2Ga_0.8N/GaN QWs in the active region of a nitride laser with cavity losses of 30cm^-1 is four, assuming homogeneous QW pumping. In the arsenide case, however, our modeling shows that in some circumstances the internal field is beneficial and can lead to a substantial reduction of the threshold current, especially for cavities with low optical losses. This reduction was confirmed experimentally, by measuring systematically lower threshold current densities in In_0.1Ga_0.9As/Al_0.15Ga_0.85As (111)B laser diodes, compared to (100)-ones, carrying no internal field.

Homoepitaxial and heteroepitaxial ZnO films were grown by plasma-assisted molecular beam epitaxy (P-MBE). Homoepitaxial ZnO layers were grown on an O-face melt-grown ZnO (0001) substrate. Heteroepitaxial ZnO layers were grown on an epitaxial GaN template predeposited by metalorganic chemical vapor deposition on a c-plane sapphire substrate. There exists a residual strain in the heteroepitaxial ZnO, which is ε = -0.25%. Low-intensity excitation PL spectra of ZnO epilayers excited by a He-Cd laser exhibit only bound-exciton emission with phonon replicas. The quality of ZnO epilayers is better than that of ZnO substrate. However, under high-intensity excitation by a N2 laser, the emission due to exciton-exciton collisions dominates the PL spectrum from heteroepitaxial ZnO layer but is not observed from homoepitaxial ZnO layer.

370nm AlGaInN-based light emitting diodes (LEDs) with SiN in the active layer, fabricated on sapphire substrates by one-time metal organic chemical vapor deposition (MOCVD) have been investigated. Atomic force microscopy and cathodoluminesence results indicated that SiN nano-mask was formed in the active layer, and the degree of compositional indium fluctuation in the active layer of UV-LEDs was enhanced. By using this technique, the output power of the LEDs was improved about 1.3 times than that of the same structure without SiN in the active layer.

We have measured the pulsed light-current characteristics of a series of InGaN/GaN quantum well light-emitting diodes which were annealed post-growth at different temperatures as a function of their operating temperature. The light output at a fixed current density increases with the temperature of measurement, reaches a maximum and then decreases for all the diodes. The measurement temperature at which the maximum light output occurs and the magnitude of the light output depend on the post-growth thermal anneal temperature. The thermal anneal temperature is thought to affect the acceptor concentration in the p-doped cap layer, which also changes the carrier mobility. A simulation, incorporating carrier leakage, is used to reproduce the experimental behavior where the acceptor concentration is changed to represent the effects of the different anneal temperatures.

Group III nitride-based blue-light-emitting diodes having a moth-eye structure at the bottom of the semi-insulating transport 6H-SiC substrate were fabricated. The light extraction efficiency and the corresponding output power were increased 3.8 times compared with those of the LED having the conventional structure. We also study a theoretical simulation using based on the RCWA method with three-dimensional Maxwell's equations. The results of theoretical analysis agree with these findings.

Temperature-dependent electroluminescence (EL) of blue InGaN/GaN multiquantum-well light-emitting diodes (LEDs) has been systematically investigated to illustrate the role of multiquantum barrier (MQB) in carrier capture and recombination. With recent advances in nitride-based light-emitting diodes (LEDs), the importance for the development of high brightness as well as high temperature devices is profound. It is found that, when temperature is slightly decreased to 200 K, the EL intensity of the active region efficiently increases in both devices, as usually seen due to the improved quantum efficiency. However, with further decrease of temperature down to 20 K, unusual reduction of the integrated EL intensity of the active region is commonly observed for both of the two diodes, accompany with the appearance of the high energy band which can be assigned as Mg-related transition at relative low temperature. It also clear noted the the diode without MQB shows a faster reduction tendency of the EL intensity than the one with MQB. Based on a rate equation analysis, we found that not only the radiative recombination zone of the quantum well region less shift to the p-type GaN region for device with MQB, but also the intensity of Mg-related transition in the LED with MQB are more than one order of magnitude less than the one without MQB. These results further verify the effective carrier confinement in active region, less carrier leakage over the barrier, and improvement of the luminescence efficiency by MQBs. All the calculations are agreement with the experimental observations.

We have extended the capabilities of the commercial device simulator ATLAS with extra models specific to the simulation of LED devices. This simulator was used to simulate the characteristics of a GaN/InGaN micro-ring light-emitting diode. These results include spectral response and output coupling efficiency

This paper first gives an overview of state-of-the-art simulation of semiconductor laser devices. The relevant physical models for multi dimensional, electro thermal, and optical simulation as well as an advanced active region model are reviewed. The second part of this work deals with the management of laser simulation projects and the extraction of the relevant data from simulation results. A new tool called PCM Explorer is presented that is suitable for the integration of numerical models in the design and manufacturing process of semiconductor lasers. Both the device performance as well as the process yield can be predicted with the combination of a comprehensive device simulator, some measurement data for calibration purposes, and the statistical process evaluation tool.

In this paper, we present a new procedure for computing the band structure of semiconductor-based photonic crystals using the spatial finite-difference and temporal differential formulation. Unlike the conventional plane wave expansion (PWE) method where the wavelet expansion and Fourier transform are required; this new scheme only forces the finite-difference to the Maxwell's equation in the spatial domain. When applying the Bloch’s boundary condition to the algorithm for solving the band-diagram, it will result in a system of first-order differential equations in a matrix format. A band mode can be obtained directly by solving the eigen-value and the eigen-vector of this matrix. Numerical examples to demonstrate the accuracy and efficiency of the proposed approach are given.

High-Speed silicon modulators, based on carrier dependent absorption effects, have recently been reported in the literature. For improved performance, these modulators rely on a MOS configuration to control carrier accumulation, rather than on carrier injection from the contacts, to induce an index perturbation for controlling the phase of a propagating signal. Accurate simulation of the carrier distribution is required for the analysis of such a device. This entails the self-consistent solution of the coupled electro-thermal transport equations. An appropriate absorption model is also required in order to couple the carrier distribution to the propagating optical field, via a complex index perturbation. Finally, in order to determine performance, the full optical problem must be solved throughout the device domain. The present work integrates the Box Integral Method of solving the active device transport equations with the Vector Beam Propagation Method (BPM) typically used to analyze passive waveguide structures. A modified Drude Model and Kramers-Kronig relations are used to determine the carrier density dependent absorption and refractive index perturbations. This complex index perturbation is determined as a function of the applied voltage, and used by a simulator based on the BPM to determine the optical performance of an example silicon modulator. Both steady-state and frequency responses are considered. This comprises a general methodology for analyzing realistic semiconductor photonic devices in which the optical propagation is affected by the electro-thermal transport within the device.

One topic that has attracted attention is related to the behavior of the optical amplifiers under dynamic conditions, specifically because amplifiers working in a saturated condition produce power transients in all-optical reconfigurable WDM networks, e.g. adding/dropping channels. The goal of this work is to introduce the multiwavelength time-driven simulations technique, capable of simulation and analysis of transient effects in all-optical WDM networks with optical amplifiers, and allow the use of control schemes to avoid or minimize the impacts of transient effects in the system performance.

Using Synopsys TCAD tools, several examples of advanced process and device modeling are presented for full-frame CCD image sensors. The topics covered in these examples include channel potential, charge capacity, charge transport, and charge blooming. The simulations provide in depth analysis of the basic principles of operation of CCDs and cover some aspects of antiblooming protection.

The calculated gain spectrum of a semiconductor laser based on a generic modal gain model reveals a gain peak that is higher than measured. Analyses of the model and the gain mechanism in semiconductor lasers suggest that the accuracy of the model can be improved by re-formulating the carrier occupation probabilities associated with the model. As a result, good agreement between the optical gain spectrum calculated using the revised model and measured results is achieved.

Distributed feedback (DFB) laser chips have recently become available at wavelengths that match the D1 and D2 resonance transitions of alkaline atoms. We investigated the spectral properties, tuning characteristics and modulation behavior of continuous wave, single-mode DFB diodes at 778-780 nm and performed high-resolution spectroscopy of Rubidium vapor. The mode-hop free tuning range of the DFB diodes was as large as 2.4 nm (1186 GHz). The line width of the laser diodes was examined both with a heterodyne beat experiment and with high-resolution Doppler-free two photon spectroscopy, yielding a half width of 2-2.5 MHz. The saturation spectra of the D2-line of ⁸⁵Rb and ⁸⁷Rb were recorded with a resolution close to the natural line width. The emission frequency was actively stabilized to Doppler-free transitions with a relative accuracy of better than 4 parts in 10⁹ using commercially available servo devices only. The output power of 80 mW was sufficient to allow for two photon spectroscopy of the 5S-5D-transition of ⁸⁷Rb. We conclude that the performance of the DFB laser equals that of grating-stabilized external-cavity diode lasers (ECDLs), without the mechanical complexity of the latter systems. The DFB diode is thus well-suited to high-resolution applications in alkaline spectroscopy, including laser cooling and optical manipulation of ultra-cold atoms.

The nonlinear contribution into the diode resistance in addition to the nonlinearity of injecting p-n junction is considered associated with a carrier injection into some nominally passive regions adjacent to the active region of the device (effect of the injection-induced conductivity, IIC). A condition the IIC to give a substantial contribution is availability of resistive passive layer or depleted layer in the diode chip that can be undergone to a conductivity modification by carrier injection and leakeage. There are some consequences that look like unusual features: (i) anomaously large the non-ideality factor of I-V curve; (ii) anomalous electrical response on the variation of the optical feedback in the laser diode device; (iii) anomalous sign of the threshold-related kink of the differential resistance of laser diode.

The performance of quantum well GaN/AlGaN light emitting diode (LED) is reviewed for three different barrier compositions; symmetric barrier composition with low Al content, asymmetric barrier composition with higher Al content on p-type cladding layer and lower Al content on n-type cladding layer, and symmetric barrier composition with higher Al content. The study was conducted using ATLAS/BLAZE & LUMINOUS software developed by Silvaco International Inc. Integrated radiative recombination rate was studied on applied voltages up to 5V. Results showed three phases of LED performance with different applied voltages and these were explained using bandgap theory. I-V characteristic for each design agrees with the total additional voltage drop equation for a quantum well structure. The dominant radiative recombination rate regions in LED at low and high supplied voltages are also presented for the best performance LED design.

Semiconductor white sources used for illumination have attracted much attention because of the theoretical high electro-optical efficiency and potential huge application market. However the practical electro-optical efficiency for the white light-emitting diode (LED) is far from the theoretical predict. In this paper we propose a novel white light superluminescent diode (WSLD), which has two active layers, and double-wavelength reflection filter films are coated on the facets of the chip. With this design we can get superluminescent light emission at two peak-gain wavelengths, that is, blue light and green-yellow light, thus producing white light. It is shown that WSLD has much higher electro-optical efficiency and better performance than ordinary white LEDs. This novel WSLD will be a new way to realize high-flux and high-efficiency semiconductor white light source.

A new type of waveguide optical polarization splitter is proposed and investigated theoretically. The waveguide optical polarization splitter consists of an Y branch waveguide in which a dielectric periodic multiplayer is loaded on a core of one branch as outer cladding layer. The dielectric periodic multilayer has large birefringence. The branch with the dielectric periodic multiplayer is designed so as the effective refractive index becomes higher than that of another branch for the y-polarization and that becomes lower for the x-polarization. Therefore, the x- and y-polarized waves propagate for different branches each other. The optical losses for the x- and y-polarization have been calculated by using a beam propagating method. The theoretical insertion loss of the typically designed waveguide optical polarization splitter is 0.3 dB. It has also been confirmed that the crosstalk is <-16 dB.

We consider the paraxial model for a nonlinear resonator with a saturable absorber beyond the mean-field limit. We introduce a general stability analysis to evidence modulational-instabilities leading to the destabilization of a homogeneous field profile, eventually causing the formation of 3D structures. Further on, for accessible parametric domains, we show in simulations the phenomenon of total radiation confinement leading to the formation of 3D localized bright structures. Such structures are a direct generalization of 2D Cavity Solitons, recently observed in broad-area VCSELs, but they are confined also in the propagation dimension. At difference from freely propagating light bullets, here the self-organization proceeds from the resonator feedback/dissipation, combined with diffraction and nonlinearity. We show that such cavity light bullets can be independently excited and erased by appropriate pulses. They can be addressed to form arrays in the transverse field profile as well as serial trains in the longitudinal direction of the resonator thus combining serial and parallel encoding in the same device. Once created, they endlessly travel the cavity roundtrip.

This paper reviews the time-domain travelling wave (TDTW) model for the simulation of active semiconductor waveguides. We outline the theory, present a discussion of its advantages and limitations and show how it can be applied to the design and simulation of a tunable 4-section sampled-grating DBR laser.

We demonstrate temperature-insensitive eye-opening under 10-Gb/s direct modulation of 1.3-μm p-doped quantum-dot lasers without current adjustments, which show 6.5-dB extinction ratio between 20 and 70°C. The active region consisting of ten quantum-dot layers with p-type doping enabled this highly temperature-stable dynamic performance, much superior to conventional 1.3-μm quantum-well lasers. This result opens a way to uncooled 1.3-μm quantum-dot lasers without current adjustments.

For many years, mid-infrared (2-5μm) semiconductor lasers operating at or near room temperature have been sought for use in LADAR, gas sensing, and spectroscopy. Smaller bandgap materials necessary for this range are more susceptible to non-radiative Auger recombination. Further, as laser structures become more complicated, like quantum cascade intersubband and interband lasers, Shockley-Read-Hall losses increase. The simplest structure is Type-I multiple quantum well (MQW), but few QW III-V heterojunction material systems capable of 2-5μm emission have a Type-I offset. One such system with InAsSb wells and AlInAsSb barriers has been unable to exceed 175K under CW operation partially due to poor carrier confinement associated with small valence band offsets. This paper describes the growth and performance of AlInAsSb/InAsSb lasers using a 0.3 mole fraction of Al in the Group III elements. Increased Al content enhances the valence and conduction band offsets, but the AlInAsSb alloy exhibits a miscibility gap above 0.06 Al mole fraction, so a digital alloy technique was used to grow high quality 0.3-2μm thick quaternary films. As Al mole fraction in the barriers was increased from 0.20 to 0.30 an 80-fold increase in photoluminescence (PL) was observed. The corresponding lasers were grown and tested demonstrating lasing at 3.9μm and 50K. Theoretical studies suggest that adding Ga to the barriers, forming an AlGaInAsSb quinary alloy, results in band structures more favorable towards minimizing Auger effects and realizing Type I offset behavior over a wider range of alloy compositions. PL structures were grown and tested, again using a digital alloy technique for the quinary alloy. Preliminary results show promise.

High hydrostatic pressure can be used for wavelength tuning of semiconductor laser diodes in a wide spectral range. Coupling the laser with external grating leads to wavelength tuning within the gain spectrum (i.e. in a narrower range than with pressure) but allows for a narrow emission line and nearly continuous tuning (mode-hop free if anti-reflecting coating is applied). Here we demonstrate a combination of pressure and external-resonator tuning for the GaInNAs laser emitting at 1343 nm at ambient conditions. Using the specially designed liquid pressure cell working up to 20 kbar we shift the emission down to 1170 nm while the external grating (used in Littrow configuration) allows for fine tuning in the ~10 nm range (at each pressure).

We present a novel hybrid light emitting device design based on a standard InAlGaAs/GaAs high-power laser diode array chip as a pump source and a narrow-gap PbSe-layer as active optical material. Maximum cw output powers of more than 1.1 mW and slope efficiencies of 0.4 mW/A are obtained at 25 °C. The external power efficiency amounts to 3.5×10^-2 %. The emission wavelength is 4.2 μm, with a half width of 770 nm (50 meV). Details about the optimization of the emitter material and device design are discussed as well.