Search

Optical interconnects enabled by silicon microring-based transceivers offer great potential for short-reach data communication in future high-performance computing systems. However, microring resonators are prone to process variations that harm both the energy efficiency and the yield of the fabricated transceivers. Especially in the application scenario where a batch of transceivers are fabricated for assembling multiple optical networks, how the transceivers are mixed and matched can directly impact the average energy efficiency and the yield of the networks assembled. In this study, we propose transceiver grouping for assembling communication networks from a pool of fabricated transceivers, aiming to optimize the network energy efficiency and the yield. We evaluated our grouping algorithms by wafer-scale measurement data of microring-based transceivers, as well as synthetic data generated based on an experimentally validated variation model. Our experimental results demonstrate that optimized grouping achieves significant improvement in the network energy efficiency and the yield across a wide range of network configurations, compared to a baseline strategy that randomly groups the transceivers.

In this article, a long-wave infrared InAs/GaAs sub-monolayer quantum dot quantum cascade photodetector (SML QD-QCD) grown on GaAs substrate is demonstrated. Temperature- and excitation-dependent photoluminescence measurements are used to study the optical properties of the quantum dot active region, which reveal energetically hybrid ground states between the InAs quantum dot and InGaAs quantum well due to the possible inter-mixing of In and Ga atoms during growth process. The device covers a spectral region from 6.5 to 9 μm. At 77 K, a peak responsivity of 7.5 mA/W is found at 8.3 μm (0 V) and a zero-bias differential-resistance-area ( R0A ) product of 9008 Ω·cm2 is obtained. The white noise-limited detectivity is 6.5 × 109 cm·Hz1/2/W. These results encourage the SML QD-QCD as a strong competitor for long-wave infrared imaging applications that require normal incidence and low power dissipation.

In this work, we present an approach to modelling III-V lasers on silicon based on a travelling-wave rate equation model with sub-micrometer resolution. By allowing spatially resolved inclusion of individual dislocations along the laser cavity, our simulation results offer new insights into the physical mechanisms behind the characteristics of 980 nm In(Ga)As/GaAs quantum well (QW) and 1.3 mu quantum dot (QD) lasers grown on silicon. We identify two effects with particular importance for practical applications from studying the reduction of the local gain in carrier-depleted regions around dislocation locations and the resulting impact on threshold current increase and slope efficiency at high dislocation densities. First, a large minority carrier diffusion length is a key parameter inhibiting laser operation by enabling carrier migration into dislocations over larger areas, and secondly, increased gain in dislocation-free regions compensating for gain dips around dislocations may contribute to gain compression effects observed in directly modulated silicon-based QD lasers. We believe that this work is an important contribution in creating a better understanding of the processes limiting the capabilities of III-V lasers on silicon in order to explore suitable materials and designs for monolithic light sources for silicon photonics.

In this article, we report the demonstration of directly modulated grating-assisted microcylinder surface-emitting lasers (GAMSELs) with low threshold. Two second-order gratings are added onto the side and top of the microcylinder cavity to select a single mode with a radially polarized emission pattern and efficiently emit the mode vertically, respectively. Laser with cavity radius of 6.05 μm is demonstrated with continuous-wave lasing ranging from 1 °C to 50 °C. This radius is approaching the theoretically smallest radius for the microcylinder laser to work at O band. A low threshold of 3.6 mA at 15 °C is obtained for the GAMSEL laser. Single-mode lasing with side-mode suppression-ratio (SMSR) larger than 40 dB is also shown here. GAMSELs with different cavity radii are demonstrated as laser arrays. 3 dB direct modulation bandwidth of 12 and 14 GHz of GAMSELs with radius of about 8 and 10 μm are obtained, respectively. Besides, effects of the duty cycle of the top and side gratings, scattering losses of the cavity, gain peak of the epitaxial wafer and ion implantation are experimentally investigated.

Optical connectivity, which has been widely deployed in today's datacenters and high-performance computing (HPC) systems, is a disruptive technological revolution to the IT industry in the new Millennium. In our journey to debut an Exascale supercomputer, a completely new computing concept, called memory-driven computing, was innovated recently. This new computing architecture brings challenges and opportunities for novel optical interconnect solutions. Here, we first discuss our strategy to develop appropriate optical link solutions for different data traffic scenarios in memory-driven HPCs. Then, we present detailed review on recent work to demonstrate fully photonics-electronics-integrated single- and multi-wavelength directly modulated laser (DML) transmitters on silicon for the first time. Compact heterogeneous microring lasers and laser arrays were fabricated as photonic engines to work with a customized complementary metal-oxide semiconductor (CMOS) driver circuit. Microring lasers based on conventional quantum well and new quantum dot lasing medium were compared in the experiment. Thermal shunt and MOS capacitor structures were integrated into the lasers for effective thermal management and ultra low-energy tuning. It enables a controllable dense wavelength division multiplexing (DWDM) link architecture in an HPC environment. An equivalent microring laser circuit model was constructed to allow photonics-electronics co-simulation. Equalization functionality in the CMOS driver circuit proved to be critical to achieve up to 14 Gb/s direct modulation with 6 dB extinction ratio. Finally, the on-going and future work is discussed towards more robust, higher speed, and more energy efficient DML transmitters.

We measure the FM noise power spectral density of quantum-dot mode-locked lasers (QD-MLLs) and compare this to their measured linewidths as predictors of performance in a digital coherent system. We explain our observations in terms of the non-Lorentzian line shape of the source wherein linewidth is determined by the low frequency part of its FM noise. Investigation of system performance with simulations based on the measured phase sequences and back-to-back coherent transmission experiments show that QD-MLLs with linewidths of several megahertz can have comparable performance to that of a laser with only a few hundreds of kilohertz of Lorentzian linewidth, due to the non-white part of their FM noise. We show that spectral linewidths of lasers with similar spectral properties can underestimate their performance in coherent systems, regardless of the linewidth measurement technique used. We propose a “ Lorentzian-equivalent linewidth ” measure to characterize lasers with non-white FM noise and to estimate their impact in digital coherent optical systems. This measure is obtained from phase variations at frequencies higher than typical frequencies often used to characterize lasers with white FM noise and comparable to the system baud. The proposed measure is shown to be a better predictor of system performance than the measured linewidth, for lasers with non-white FM noise. The impact of non-white FM noise on the optimization of carrier phase recovery and system performance is also discussed.

This paper reviews optical injection locking (OIL) of semiconductor lasers and its application in optical communications and signal processing. Despite complex OIL dynamics, we attempt to explain the operational principle and main features of the OIL in an intuitive way, aiming at a wide understanding of the OIL phenomenon and its associated techniques in the optic and photonic communities. We review and compare different control techniques that enable robust OIL in practical systems. The applications are reviewed with a focus on new developments in the past decade, under the categories of ‘High Speed Directly Modulated Lasers’ and ‘Optical Carrier Recovery’. Finally, we draw our vision for future research directions.

In this paper, we present measurements and simulations of the small-signal modulation response of monolithic continuous-wave 1.3 μ m InAs/GaAs quantum dot (QD) narrow ridge-waveguide lasers on a silicon substrate. The 2.5 mm-long lasers investigated demonstrate 3 dB modulation bandwidths of 1.6 GHz, D -factors of 0.3 GHz/mA1/2, modulation current efficiencies of 0.4 GHz/mA1/2, and K -factors of 2.4 ns and 3.7 ns. Since the devices under test are not designed for high-speed operation due to their long length and hence long photon lifetime, the modulation response curves are used as a fitting template for numerical simulations with spatiotemporal resolution to gain insight into the underlying laser physics. The obtained parameter set is used to unveil the true potential of the laser material in an optimized device geometry by modeling the small-signal response at different cavity lengths, mirror reflectivities, and for different numbers of QD layers. The simulations predict a maximum 3 dB modulation bandwidth of 5 GHz to 7 GHz for a 0.75 mm-long cavity with 99% and 60% high-reflection coatings and ten QD layers. Modeling the impact of dislocations on the dynamic performance qualitatively reveals that enhanced non-radiative recombination in the wetting layer leaves the modulation bandwidth of QD lasers on silicon almost unaffected, while dislocation-induced optical loss does not pose a problem, as long as sufficient gain is provided by the QD active region.

We report for the first-time electro-optic phase and amplitude modulators in GaAs/AlGaAs epitaxial layers grown on misaligned silicon substrates containing germanium buffer layers. Epilayer has a npin doping profile and is equivalent to a pin diode. The 4-mm long electrode Mach–Zehnder modulators have 7.4- and 3.6-V Vπ under single-arm and push–pull drive conditions corresponding to 1.5 ± 0.1 V-cm modulation efficiency. Data on 7-mm long electrode Fabry–Perot phase modulators indicate 2.3-V Mach–Zehnder modulators are possible. These data also indicate less than 1-dB/cm on-chip propagation loss under 5-V reverse bias.

Due to internal optical absorption loss, which varies with carrier density in the optical confinement region, the operating characteristics can be two valued in semiconductor quantum well lasers. Particularly, there can be two lasing thresholds and two branches in the light–current characteristic. While the internal differential quantum efficiency for the first (conventional) branch of the light–current characteristic is less than 1 and decreases with increasing pump current, that for the second (unconventional) branch is greater than 1 and nonmonotonous.

We propose a tunable dual-wavelength heterogeneous quantum dot laser diode. The tunable dual-wavelength laser consists of a quantum dot semiconductor optical amplifier as the optical gain medium and an external cavity fabricated from silicon photonics technology as the wavelength tunable filter. We successfully demonstrated dual-wavelength lasing oscillation by tuning the difference frequency from approximately 34 to 400 GHz.

GaAs-based InAs/InGaAs/GaAs quantum dot (QD) photonic crystal (PC) surface emitting lasers of 1.3 μm wavelength range are fabricated and room-temperature lasing emissions by optical pumping are demonstrated for the first time. Laser devices are characterized in terms of PC parameters and temperature-dependent measurements are also carried out. The impact of wavelength detuning between PC resonant wavelength and QD gain peak is manifested in the experiments. Moreover, simplified simulation reveals that etched PC depth plays an even more critical role in active region of QD compared to that of quantum well.

We present a mathematical analysis and comparison of the performance of integrated on-chip semiconductor ring laser gyroscope (SRLG) fabricated using GaAs/AlGaAs and InP/InGaAsP technologies. The performance parameters of the gyro are modeled in terms of fundamental material, waveguide, and resonator parameters. In addition to this, influence of phenomena specific to semiconductor lasers such as nonlinear coupling, spatial hole burning, gain grating formation, and carrier induced index change on the gyro performance is also included. The analysis helps in identifying critical parameters, which must be optimized to improve the gyro performance.Best achievable performance of integrated SRLG is calculated, and design modifications are suggested to enhance it for high-performance military applications.

This paper describes the design considerations and performance analysis for an XG-PON Raman amplification scheme. The experimental results and theoretical calculations presented in this paper demonstrate the feasibility of Raman amplification in O-band access systems; thus, enhancing the performance of existing access installations while still preserving the passive nature of these networks. With the availability of cost-effective quantum dot laser diodes as high-power pumps, Raman amplification is proving to be an attractive candidate for access applications. We experimentally demonstrate that it is possible to achieve a total pump power of 759 mW using polarization multiplexed quantum dot laser diodes (QD-LDs), and using a backward pumping scheme we can support a purely passive XG-PON system with 50 km reach and 64 splits. With system level numerical simulations, we illustrate the use of forward or bi-directional pumping scheme to increase the splits further. We also confirm that the Raman pump output in the experiments, despite being below the single mode cut off for the standard SMF, was efficiently coupled into the trunk fibre using off-the-shelf components. We also show using numerical simulations that a flat gain spectrum can be obtained with less than 0.7 dB gain variation across the XG-PON upstream wavelength band (1260–1280 nm) using a multi-wavelength pumping scheme with just two pump wavelengths. From a detailed system analysis, we discuss the impact of broadband filtering (20 nm) at the receiver adopted by the XG-PON standards, and show that lower class (N1) receivers can be adopted for Raman gains higher than 16 dB.

We present the design, fabrication, and performance of our latest generation high-speed oxide-confined 850-nm vertical-cavity surface-emitting lasers. Excellent high-speed properties are obtained by strong confinement of optical fields and carriers. High-speed modulation is facilitated by using the shortest possible cavity length of one half wavelength and placing oxide apertures close to the active region to efficiently confine charge carriers. The resulting strong current confinement boosts internal quantum efficiency, leading to low threshold currents, high wall-plug efficiency, and state-of-the-art high-speed properties at low bias currents. The temperature dependent static and dynamic performance is analyzed by current-power-voltage and small-signal modulation measurements.

We previously proposed a photonics-electronics convergence system to solve bandwidth bottleneck problems among large-scale integrations (LSIs) and demonstrated a high bandwidth density with silicon optical interposers at room temperature. For practical applications, the interposers should be usable under high-temperature conditions or rapid temperature changes so that they can cope with the heat generated by the mounted LSIs. We designed and fabricated athermal silicon optical interposers integrated with temperature-insensitive components on a silicon substrate. An arrayed laser diode (LD) chip was a flip-chip bonded to the substrate. Each LD had multiple quantum dot layers with a 1.3-μm lasing wavelength. The output power was higher than 10 mW per channel up to 100 °C. Silicon optical modulator and germanium photodetector (PD) arrays were monolithically integrated on the substrate. The modulators were structured as symmetric Mach–Zehnder interferometers, which were inherently insensitive to temperature and wavelength. The phase shifters composed of p-i-n diodes were stable against temperature change under a constant bias-current condition. The PD photocurrent was also temperature insensitive, and the photo-to-dark current ratio was higher than 30 dB up to 100 °C. We achieved error-free data links at 20 Gbps and a high bandwidth density of 19 Tbps/cm2 operating up to 125 °C without adjusting the LDs, modulators, or PDs. The interposers are tolerant of the heat generated by the mounted LSIs and usable over the extended industrial temperature range without complex monitoring or feedback controls. The bandwidth density is sufficient for the needs of the late 2010s.

A dense array of InGaAs quantum dots formed by MOCVD on a misoriented GaAs substrate has been used as an active medium of microdisk resonators of various types: a cylindrical disk, an undercut disk, and a suspended disk. Single-mode room temperature lasing in a 9-μm microdisk laser is demonstrated with a dominant line around 1.13 μm. Impact of the resonator design on the lasing threshold and its temperature behavior is discussed.

Theoretical analysis of relative intensity noise (RIN) characteristics of p-doped QD lasers has been presented. By considering dynamics of electrons and holes separately at GaAs barrier, wetting layer (WL), and three discrete QDs levels, 12 rate equations have been linearized in presence of the Langevin noise sources. Calculations indicate that RIN level of QD laser reduces slightly through p-doping. Although providing excess holes to WL state decreases the shot noise, resulted from quantum confined levels and photon noise, the shot noise originated from three and two dimensional carrier densities, respectively, inside the barrier and WL states, increases. It is shown that the RIN level declines in p-doped QD lasers by increasing the injection current, which is in agreement with a recent experimental report. It is also demonstrated that the RIN level of QD lasers decreases by increasing the number of QD layers, and a shot noise plays the main role in this reduction.

We present the intensity noise characteristics of an actively mode-locked quantum-dot hybrid soliton pulse source (QD-HSPS) with a linearly chirped Gaussian-apodized fiber Bragg grating. The QD-HSPS is modeled by using an approximate solution for the propagation equation that governs the electric field and the rate equations, including spontaneous and carrier noise sources. Our results show that QD carrier noise is the most effective noise type, and the spontaneous coupling factor, carrier recombination time in the QD, and laser cavity length are the most important noise parameters. In addition we find that a QD-HSPS produces considerably shorter pulses, with or without noise, than a multi-quantum-well HSPS and has a larger mode-locking frequency range.

This paper reports a hybrid integrated light source fabricated on a Si platform using a spot-size converter (SSC) with a trident Si waveguide. Low-loss coupling for 1.55 μm and 1.3 μm wavelengths was achieved with merely the simple planar form of a Si waveguide with no use of complicated structures such as vertical tapers or an extra dielectric core overlaid on the waveguide. The coupling loss tolerance up to a 1 dB loss increase was larger than the accuracy of our passive alignment technology. The coupling efficiency was quite robust against manufacturing variations in the waveguide width compared with that of a conventional SSC with an inverse taper waveguide. A multi-channel light source with highly uniform output power and a high-temperature light source were fabricated with a 1.55 μm quantum well laser and a 1.3 μm quantum dot laser, respectively. The integration scheme we report can be used to fabricate light sources for high-density, multi-channel Si optical interposers.

The 40-year history of research on semiconductor lasers for high-speed long-distance optical fiber communications, so-called dynamic single-mode (DSM) lasers, is reviewed. DSM lasers include phase-shift distributed feedback (DFB) lasers, uniform DFB lasers, wavelength tunable (WT) distributed Bragg reflector lasers, WT distributed reflector lasers, and external reflector lasers. The vertical cavity surface emitting laser is also a type of DSM laser that is applied for rather short-distance communication. Photonic-integrated circuits with monolithic integration of DSM lasers with other devices, as well as photonic crystal lasers, have advanced significantly to support stable operation of photonic subsystems. The DSM laser is currently essential for most long-haul optical communications and transoceanic submarine cables, as well as medium-distance local area networks. Its application to sensing systems is another interesting area.

In this paper we study high-speed pattern effect-free cross-gain modulation (XGM) in quantum-dot vertical-cavity semiconductor optical amplifier (QD-VCSOA) with and without considering Auger effect. XGM is examined for the different surface densities of QDs and bias currents. We show that appearance of the pattern effect strongly depends on the bias current and surface density of QDs. Pattern effect is improved at higher injection current or low dot density. However, at higher currents since Auger scattering is stronger it fills the higher state of QD and causes the drop of stimulated emission resulting in lower modulation efficiency.

The dynamic properties of uniformly and chirpily stacked InAs-InGaAs-GaAs quantum dot lasers are analyzed in terms of relative intensity noise spectra. For uniformly stacked quantum dot laser with ground-state lasing emissions of 1.3 μm, the K-factor limited bandwidth is 13 GHz. The extracted differential gain and gain compression factor are 1.7 × 10-15 cm2 and 2 × 10-16 cm3, respectively. For chirpily stacked quantum dot laser with excited-state lasing emissions of 1.2 μm, the K-factor limited bandwidth is 14 GHz. Yet the nonproportional dependence between resonance frequency and square root of incremental current yields differential gain of 4.3 × 10-15 cm2 and huge gain compression factor of 1.4 × 10-14 cm3.

A polarization-insensitive quantum dot semiconductor optical amplifiers (QD-SOAs) have been studied for use in future optical communication systems. A part of our work shows that the optical polarization property in QDs depends on both their aspect ratio and strain. To control these two parameters, we propose the use of strain-controlled columnar QDs (SC-CQDs), which exhibit a high aspect ratio and have strain-controlled side barriers for polarization-insensitive operation in the 1.5-μ m wavelength band. QD-SOAs with these optimized SC-CQDs demonstrated polarization-insensitive characteristics. They showed a gain of 8.0 dB with polarization dependence of the gain as low as 0.4 dB, -3-dB saturation output power of 18.5 dBm at a wavelength of 1550 nm, and error-free amplification at a bit rate of 40 Gbit/s.

Semiconductor quantum dot lasers have been extensively studied for applications in future lightwave telecommunications systems. This paper summarizes a tutorial that was presented at the Optical Fiber Communication (OFC) 2010. The motivation for quantum dots in lasers is outlined, and the desirable effects of three dimensional quantum confinement are described. Methods for forming self-assembled quantum dots and the resultant laser characteristics are presented. The formation of patterned quantum dot lasers and the results of this type of quantum dot laser are outlined. Finally, a novel inverted quantum dot structure or nanopore laser containing 3-D quantization formed from an engineered periodicity is introduced.

A self-consistent rate equation model is presented to investigate the influence of carrier relaxation on the modulation response of 1.3 mum InAs-GaAs quantum dot lasers. In this model, the carrier dynamics in GaAs barrier, relaxation pathways, and the phonon- and Auger-assisted relaxation are considered. The dependence of 3 dB bandwidth on the relaxation time and relaxation pathway is discussed. It is shown that carrier relaxation via less energy level has better carrier confinement and higher 3 dB bandwidth. The improvement of bandwidth by tunnelling injection QD structure is investigated from the point of view of relaxation pathway. The different effects of tunnelling into ground state and excited state on the 3 dB bandwidth are analyzed. The enhanced carrier relaxation by p-type modulation doping and its effect on the bandwidth are investigated. It is found that there exists a tradeoff on the improvement of bandwidth by p-doping, which is explained as the competition between the bandwidth limitation of K -factor and relaxation dynamics. Increase in the bandwidth of QD lasers by improving both the carrier relaxation dynamics and K-factor limitation is discussed.

To suppress bipolar population and hence electron–hole recombination outside quantum dots (QDs), tunneling-injection of electrons and holes into QDs from two separate quantum wells was proposed earlier. Close-to-ideal operating characteristics were predicted for such a double tunneling-injection (DTI) laser. In the Stranski–Krastanow growth mode, a two-dimensional wetting layer (WL) is initially grown followed by the formation of QDs. Due to thermal escape of carriers from QDs, there will be bipolar population and hence electron–hole recombination in the WL, even in a DTI structure. In this work, the light–current characteristic (LCC) of a DTI QD laser is studied in the presence of the WL. Since the opposite sides of a DTI structure are only connected by the current paths through QDs and the WL is located in the n-side of the structure, the only source of holes for the WL is provided by QDs. It is shown that, due to the zero-dimensional nature of QDs, the rate of the hole supply to the WL remains limited with increasing injection current. For this reason, as in the other parts of the structure outside QDs (quantum wells and optical confinement layer), the parasitic electron–hole recombination remains restricted in the WL. As a result, even in the presence of the WL, the LCC of a DTI QD laser becomes increasingly linear at high injection currents, which is a further demonstration of the potential of such a laser for high-power operation.

We investigate the characteristics of all-optical clock recovery for return-to-zero differential phase-shift keying (RZ-DPSK) and return-to-zero on-off keying (RZ-OOK) signals at 40 Gb/s using mode-locked semiconductor lasers. Two mode-locked lasers with different configurations, a bulk active layer distributed Bragg reflector laser and a quantum-dot active layer Fabry-Perot laser, are used in the experimental investigation. The results reveal the implications of the input signal power, beat spectrum linewidth of the mode-locked laser, and intensity modulation on the jitter characteristics of the recovered clock signal.

The ultra-fast gain dynamics in quantum-dot semiconductor optical amplifiers (QD-SOAs) have been studied for different types of Auger-assisted relaxation processes. The ultra-fast gain recovery time and gain compression are studied for p-type doped and un-doped QD-SOAs using rate equation model. Our calculations show that the ultra-fast gain dynamics is governed by electron-electron Auger-assisted process for un-doped QD-SOA and by electron-hole Auger-assisted process for p-type doped (NA=1.25times1018 cm-3) QD-SOA. We find that the ultra-fast gain recovery time for un-doped QD-SOA is comparable with that of p-type doped QD-SOA when both electron hole and electron-electron processes present in the active region. We find that the percentage of ultra-fast gain compression in un-doped QD-SOA is limited to ~ 72%. While for p-type doped (NA=1.25times1018 cm-3) QD-SOA, we find that the percentage of ultra-fast gain compression increases as the applied current increases where it can reach >95% at very high applied current.

The influence of p-type and n-type doping on the optical characteristics of a quantum-dot semiconductor optical amplifier (SOA) is studied using a rate equation model that takes into account the effect of the multidiscrete energy levels and the charge neutrality relation. Our calculations show that the amplifier optical gain can be greatly enhanced through p-type doping where the doping concentration should not exceed the certain level. We find that increasing the acceptor concentration increases the unsaturated optical gain but at the same time decreases the saturation density and the effective relaxation lifetime. Also our calculation reveals that the use of p-type doping will be associated with an increase in the transparency current where the increase in the transparency current depends on the incident photon energy. On the other hand, we find that it is possible to increase the saturation density and enhance the linearity of the SOA by using n-type doping.

We present simulation results on the propagation characteristics of active plasmonic waveguides at 1.55 mum wavelength based on semiconductors as the active gain media. Three waveguide structures were investigated: metal rib, metal-semiconductor-metal (MSM), and triangular metal groove. In all three structures, we observed strong plasmon mode confinement with nanoscale spot-sizes and corresponding simulated gain values compatible with existing semiconductor technology. We show the effect of systematic modification of waveguide geometry on the required gain for achieving lossless propagation in all the three plasmonic waveguide structures. We demonstrate that lossless propagation with subwavelength spot sizes well below the diffraction limit of light can be obtained by controlling the geometrical parameters of the proposed waveguides.

We report on interplay of epitaxial growth phenomena and device performance in quantum dot (QD) and quantum wire (QWW) lasers based on self-organized nanostructures. InAs QDs are the most explored model system for basic understanding of "near-ideal" QD devices. Vertically-coupled growth of QDs and activated phase separation allow ultimate QD wavefunction engineering enabling GaAs lasers beyond 1400 nm and polarization-insensitive optical amplification. A feasibility of QD semiconductor optical amplifiers at terabit frequencies using InAs QDs is manifested at 1300 and 1500 nm. 1250-1300 nm QD GaAs edge emitters and VCSELs operate beyond 10 Gb/s with ultimate temperature robustness. Furthermore, temperature-insensitive operation without current or modulation voltage adjustment at >20 Gb/s is demonstrated up to ~90 degC. Light-emitting devices based on InGaN-QDs cover ultraviolet (UV) and visible blue-green spectral ranges. In these applications, InN-rich nanodomains prevent diffusion of nonequilibrium carries towards crystal defects and result in advanced degradation robustness of the devices. All the features characteristic to QDs are unambiguously confirmed for InGaN structures. For the red spectral range InGaAlP lasers are used. Growth on misoriented surfaces, characteristic to these devices, leads to nano-periodi- cally-step-bunched epitaxial surfaces resulting in two principal effects: 1) step-bunch-assisted alloy phase separation, leading to a spontaneous formation of ordered natural super lattices; 2) formation of quantum wire-like structures in the active region of the device. A high degree of polarization is revealed in the luminescence recorded from the top surface of the structures, in agreement with the QWW nature of the gain medium. QD and QWW lasers are remaining at the frontier of the modern optoelectronics penetrating into the mainstream applications in key industries.

We have made InGaAs submonolayer (SML) quantum-dot (QD) and InAs QD photonic-crystal vertical-cavity surface-emitting lasers (PhC-VCSELs) for fiber-optic communications in the 990 and 1300 nm ranges, respectively. The active region of the InGaAs SML QD PhC-VCSEL contains three InGaAs SML QD layers, with each of the SML QD layer formed by alternating depositions of InAs and GaAs. The active region of the InAs QD PhC-VCSEL contains 17 undoped InAs-InGaAs QD layers. Both types of QD PhC-VCSELs exhibit single-mode characteristics throughout the current range, with side-mode suppression ratio (SMSR) larger than 35 dB. A maximum output power of 5.7 mW has been achieved for the InGaAs SML QD PhC-VCSEL. The near-field image study of the QD PhC-VCSELs indicates that the laser beam is well confined by the photonic-crystal structure of the device.

We demonstrate a monolithically integrated laser-waveguide device implemented with InGaAs/GaAs quantum dot heterostructures grown on silicon by molecular beam epitaxy. Focused-ion-beam (FIB) etching is utilized to form high-quality laser mirrors for feedback and grooves for coupling and electrical isolation. Based on a transmission matrix and a generalized beam propagation approach in terms of intensity moments and Gouy phase shifts, a self-consistent model is developed to estimate the reflectivity and coupling coefficient at etched grooves and optimize these parameters for real devices. High-quality FIB-etched facets with a reflectivity of R~0.28 and efficient coupling with coupling coefficients of up to 30% for well-defined grooves have been achieved.

We report on the experimental assessment of an all-optical clock-recovery scheme at 40-Gb/s cascading a polarization-insensitive bulk-based self-pulsating (SP) laser and a high spectral purity quantum-dots-based SP laser. It is demonstrated experimentally that such a clock-recovery scheme is polarization insensitive, efficient in the jitter filtering, and tolerant for an input optical signal-to-noise ratio (OSNR) as low as 15 dB/0.1 nm. It is shown theoretically that the jitter-filtering function of the cascade is the product of the transfer functions of both lasers. The contributions of the phase noise of these two lasers to the final jitter are also identified and quantified. The influence of the degradation of the OSNR to the total timing jitter is also analyzed. The approach proposed in this paper offers the real opportunity to realize an all-optical clock recovery with a performance compatible for system applications

This paper reports on timing-jitter analysis of an all-optical clock-recovery scheme at 40 GHz using self-pulsating (SP) lasers. Based on the analogy with injection locking of oscillators, theoretical investigations on phase-noise properties of the recovered clock lead to the demonstration of a filtering function with slope that is compliant with the International Telecommunications Union (ITU) standards and allow us to underline the dependence of the cutoff frequency of the filtering transfer function on the spectral linewidth of the free running SP laser. From this phase-noise analysis, an analytical expression of the timing jitter of the recovered clock is derived, including the optical signal-to-noise ratio (OSNR) of the injected signal. A set of experiments on all-optical clock recovery at 40 GHz is then presented and demonstrates the crucial role of the spectral linewidth on the timing-jitter-filtering function of the SP laser. In good agreement with theoretical results, the impact of the OSNR degradation of the injected signal on the timing jitter is also demonstrated. Finally, the all-optical clock-recovery operation using a quantum-dot SP laser is shown to be standard compliant in terms of timing jitter, even for highly degraded OSNR

Dynamic and static characteristics of high-speed 1.55- and 1-/spl mu/m wavelength tunneling injection quantum-well lasers and 1-/spl mu/m wavelength self-organized quantum-dot lasers, have been measured as a function of temperature. While differential gain of the quantum-well lasers greatly increased with lowering of temperature (by a factor of 50), gain compression increased along with it, resulting in about the same intrinsic damping limit (K-factor) over a wide range of temperatures and only moderate increases in bandwidth (20-35 GHz). This suggests that increase in differential gain alone is not sufficient to improve modulation characteristics directly. Because of the mechanism of gain compression, lasers which are damping limited may not see a large improvement in modulation bandwidth simply by operating at lower temperature. In contrast, the modulation bandwidth of the quantum-dot lasers increased from 5-6 GHz at room temperature to larger than 20 GHz at 90 K. This behavior is explained by considering electron-hole scattering as the dominant mechanism-for electron capture in quantum-dots. The measured temperature dependence of the K-factor is analyzed with consideration of electron-hole scattering, and the value extracted for the electron intersubband spacing from this analysis, 60 meV, agrees with the theoretically calculated value of 56 meV.

A novel micro-laser configuration formed by integrating an InGaAs/InP pillar with a silicon photonic crystal cavity is proposed and analyzed in detail. Special attention is paid to designing the cavity such that it can accommodate large-size pillars without performance compromise. The Purcell effect is studied and predicted to be significant because of the close interaction between the cavity modes and the gain medium. An overall quality factor as high as 1 x 10(5) and a spontaneous emission factor close to unity are predicted. Possible limiting factors for laser performance, such as surface non-radiative recombination and the thermal dissipation properties are analyzed, and it is found that the proposed laser design is very robust. This comprehensive analysis suggests that the proposed micro-laser is a promising candidate for large-scale integration of micro-lasers on silicon for low power consumption applications, such as intra-chip optical communications.

This paper describes a theoretical study of the small-signal modulation behavior of an injection-locked semiconductor laser. Illustrative examples are given, shelling a comparison between the free-running laser and the same laser with light injection. The results show that a substantial reduction of the chirp-to-power ratio (CPR) can be obtained, depending on both the injection level and the frequency detuning between the master and slave lasers. The behavior of the intensity modulation responses is also investigated, with the modulation conditions chosen in the dynamically stable locking range. It appears that the injection-locked laser may present a larger resonance frequency or modulation bandwidth with respect to those of the same laser under free-running operation.

The application of a high-speed signal via the drive current to a long-cavity semiconductor laser (cavity length >/spl sim/1 cm) differs from the short-cavity case in that the variation of the optical field within one round-trip time is not negligible. We theoretically investigate the modulation response for a semiconductor laser of arbitrary length and relate it to experiments on the multifrequency waveguide grating router laser, a long-cavity laser. We find that the largest modulation bandwidth is achieved by having the modulated amplifier and the desired output as far apart as possible, the modulation bandwidth can be further increased by simple electronic precompensation, and because of fast nonlinearities the broadband intensity modulation of long-cavity lasers exhibits substantially less chirp than short-cavity lasers, the chirp essentially equal to that of the amplifiers. Finally, we demonstrate the direct modulation of three channels simultaneously in a multifrequency laser at 1.24 Gb/s each.

Compressively and tensile strained InGaAsP-InP MQW Fabry-Perot and distributed feedback lasers emitting at 1.3-/spl mu/m wavelength are reported. For both signs of the strain, improved device performance over bulk InGaAsP and lattice-matched InGaAsP-InP MQW lasers was observed. Tensile strained MQW lasers show TM polarized emission, and with one facet high reflectivity (HR) coated the threshold currents are 6.4 and 12 mA at 20 and 60/spl deg/C, respectively. At 100/spl deg/C, over 20-mW output power is obtained from 250-/spl mu/m-cavity length lasers, and HR-coated lasers show minimum thresholds as low as 6.8 mA. Compressively strained InGaAsP-InP MQW lasers show improved differential efficiencies, CW threshold currents as low as 1.3 and 2.5 mA for HR-coated single- and multiple quantum well active layers, respectively, and record CW output powers as high as 380 mW for HR-AR coated devices. For both signs of the strain, strain-compensation applied by oppositely strained barrier and separate confinement layers, results in higher intensity, narrower-linewidth photoluminescence emissions, and reduced threshold currents. Furthermore, the strain compensation is shown to be effective for improving the reliability of strained MQW structures with the quantum wells grown near the critical thickness. Linewidth enhancement factors as low as 2 at the lasing wavelength were measured for both types of strain. Distributed feedback lasers employing either compressively or tensile strained InGaAsP-InP MQW active layers both emit single-mode output powers of over 80 mW and show narrow linewidths of 500 kHz. >

Recent progress in high-power semiconductor lasers for erbium-doped fiber amplifiers is described, focusing on 1.48-μm InGaAsP/InP lasers and 0.98-μm InGaAs/GaAs lasers. The experimental output powers exceed 200 mW (the maximum power was 325 mW) for 1.48-μm lasers, and simulation results indicate that over 400 mW could be obtained by optimizing parameters in strained-layer (SL) multiple-quantum-well (MQW) lasers. Stable operation over a few thousand hours under 100-mW power is demonstrated for liquid-phase-epitaxy-grown lasers, MQW lasers, and SL-MQW lasers grown by all-metal organic vapor-phase epitaxy (MOVPE). For 0.98-μm lasers, improvement in the fiber coupling efficiencies and long-term reliabilities are described. Their power coupled into a single-mode fiber has reached over 100 mW, with coupling efficiencies of approximately 40%. Although reliability seems to be one of the drawbacks compared with 1.48-μm lasers, stable operation for over 10,000 h at 50°C-30 mW has been reported.

Numerical calculations of the frequency modulation responses of monolithically integrated external-cavity semiconductor lasers are performed using G.P. Agrawal's (J. Appl. Phys., vol.56, p.3110-15, 1984) two-section model. Theoretical results show good agreement with the reported experimental results. The comparison of the modulation responses of the solitary semiconductor laser diodes, the conventional external-cavity semiconductor lasers, and the integrated passive external-cavity semiconductor lasers demonstrates that the integrated external-cavity semiconductor laser is the best choice for use in direct frequency modulation. >

We present reliability data on widely tunable diode lasers with sampled grating Bragg mirrors. The lasers have a median life over two hundred years with a wear-out failure rate below 200 failures per billion device-hours.

This special issue features 23 articles which focus on nano-optoelectronics and applications.

For the above titled paper (ibid., vol. 26, no. 21, pp. 3534-3555, Nov. 08), there are some additional comments that are presented here.