Sawsan Abdul-Majid

Electro-optic up-conversion mixer amenable to photonic integration

A novel electro-optical up-conversion mixer architecture comprising four electro-optical phase modulators situated in the arms between an interconnected 1 × 4 distribution tree and a complementary 4 × 2 combination tree is proposed. The distribution and combination trees are based on multi-mode interference couplers (MMI). The novelty lies in the use of the intrinsic phase relations between the MMI ports to realize a broadband and free of drift design requiring no static phase shift elements. A transfer-matrix approach is followed to represent the main building blocks in the proposed design, and hence to describe the operation of the entire optical up-conversion mixer. The concept is demonstrated by computer simulations. A single side-band modulation with carrier suppression is obtained at the output of the proposed architecture, which is in agreement with the analytical development. Scenarios considering both ideal and imperfect power balances and phase relations in the MMIs, as well as imperfect phase relations of the electrical drives to the phase modulators are analyzed.

Radio-over-Fibre access for sustainable Digital Cities

Pervasive broadband access will transform cities to the net social, environmental and economic benefit of the e-City dweller as did the introduction of utility and transport network infrastructures. Yet without action, the quantity of greenhouse gas emissions attributable to the increasing energy consumption of access networks will become a serious threat to the environment. This paper introduces the vision of a ‘sustainable Digital City’ and then considers strategies to overcome economic and technical hurdles faced by engineers responsible for developing the information and communications technology (ICT) network infrastructure of a Digital City. In particular, ICT energy consumption, already an issue from an operating cost perspective, is responsible for 3 % of global energy consumption and is growing unsustainably. A grand challenge is to conceive of networks, systems and devices that together can cap wireless network energy consumption whilst accommodating growth in the number of subscribers and the bandwidth of services. This paper provides some first research directions to tackle this grand challenge. A distributed antenna system with radio frequency (RF) transport over an optical fibre (or optical wireless in benign environments) distribution network is identified as best suited to wireless access in cluttered urban environments expected in a Digital City from an energy consumption perspective. This is a similar architecture to Radio-over-Fibre which, for decades, has been synonymous with RF transport over analogue intensity-modulated direct detection optical links. However, it is suggested herein that digital coherent optical transport of RF holds greater promise than the orthodox approach. The composition of the wireless and optical channels is then linear, which eases the digital signal processing tasks and permits robust wireless protocols to be used end-to-end natively which offers gains in terms of capacity and energy efficiency. The arguments are supported by simulation studies of distributed antenna systems and digital coherent Radio-over-Fibre links.

Limitation Factor Analysis for Silicon-on-Insulator Waveguide Mach–Zehnder Interference-Based Electro-Optic Switch

The Mach-Zehnder interference (MZI) structure has played a significant role in research and development of the optical modulator/switch and silicon-on-insulator (SOI) waveguides have been increasingly developed to implement highly integrated photonic devices. In this paper, for the SOI-waveguide MZI-type electro-optic (EO) switch with free-carrier dispersion (FCD) effect, the extra optical absorption (EOA) loss caused by the FCD effect is analyzed and modeled. An intrinsic limitation factor existing in this device is found to be the tension between the EOA loss and the interaction length, resulting in a negative impact upon the device performance. The numerical calculations show that the millimeter-order interaction length has the lowest optical on-chip (OC) loss of about 0.8 and 1.8 dB at the OFF- and ON-state, respectively, and even a lowest OC imbalance of 1.0 dB between the two switching states. The influence of the coupling ratio of 3 dB waveguide directional coupler used in the MZI structure upon the switch performance is also studied, and a push-pull modulation scheme is proposed as an efficient solution to leveraging this intrinsic limitation caused performance decay with a combination of injection and depletion processes for the FCD effect. As a result, the optical OC loss is reduced to 1.0 dB, its imbalance is compressed to 0.2 dB, and the crosstalk at the OFF-state is also better than -21 dB. The relationship between the switching speed and the interaction length is also analyzed. As a vital condition for the FCD-based EO modulation of the switch, the dependence of free-carrier concentration modulation on the drive voltage and electrode gap is simulated via MEDICI software.

Free space and waveguide Talbot effect: phase relations and planar light circuit applications

Optical fields that are periodic in the transverse plane self-image periodically as they propagate along the optical axis: a phenomenon known as the Talbot effect. A transfer matrix may be defined that relates the amplitude and phase of point sources placed on a particular grid at the input to their respective multiple images at an image plane. The free-space Talbot effect may be mapped to the waveguide Talbot effect. Applying this mapping to the transfer matrix enables the prediction of the phase and amplitude relations between the ports of a Multimode Interference (MMI) coupler– a planar waveguide device. The transfer matrix approach has not previously been applied to the free-space case and its mapping to the waveguide case provides greater clarity and physical insight into the phase relationships than previous treatments. The paper first introduces the underlying physics of the Talbot effect in free space with emphasis on the positions along the optical axis at which images occur; their multiplicity; and their relative phase relations determined by the Gauss Quadratic Sum of number theory. The analysis is then adapted to predict the phase relationships between the ports of an MMI. These phase relationships are critical to planar light circuit (PLC) applications such as 90° optical hybrids for coherent optical receiver front-ends, external optical I-Q modulators for coherent optical transmitters; and optical phased array switches. These applications are illustrated by results obtained from devices that have been fabricated and tested by the PTLab in Si micro-photonic integration platforms.

Numerical Study of Dual Mode Generation Using a Sampled-Grating High-Order Quantum-Dot Based Laterally-Coupled DFB Laser

We propose a fabrication-friendly dual-mode laser source based on a sampled surface-grating, quantum-dot (QD), third-order, and laterally-coupled distributed feedback (LC-DFB) laser composed of alternating grating and Fabry-Perot sections. The dynamic behavior of this device is investigated through numerical modelling, and mode spacing in the millimeter-wave domain (60 GHz) was achieved. We extended a time-domain travelling-wave algorithm, including Streifers terms, to numerically study the dynamic behavior of the modified high-order LC-DFB lasers. We also incorporated an active QD region via a set of rate equations that considers both in homogeneous broadening because of spatial distribution of QD and homogeneous broadening because of the scattering or polarization dephasing rate. It was found that stable dual mode operation in the millimeter-wave range can be achieved with a dual-side-mode-suppression-ratio as high as ~ 50 dB.

Low polarization-sensitive asymmetric multi-quantum well semiconductor amplifier for next-generation optical access networks

A broadband and low-polarization-sensitive multi-quantum well semiconductor optical amplifier with an asymmetric structure is reported for operation in the E-band wavelength range. A gain peak of 20 dB for a bandwidth of more than 50 nm is measured for both TE and TM polarizations. A maximum polarization sensitivity of 3 dB is measured for a broad wavelength range from 1340 to 1440 nm.

Metamaterial Lüneburg waveguide lenses for switch fabric on-a-chip applications

Optical information processing has traditionally been demonstrated using 3D free-space optical systems employing bulk optical components. These systems are bulky and unstable due to the stringent alignment tolerances that must be met. Taking advantage of the alignment accuracy offered by planar light circuits, these issues may be overcome by confining the light in a planar slab waveguide. The limitation on scaling, consequent on the loss of one dimension is offset by the nanoscale component footprints attainable in a silicon integration platform. A key component of this free-space-opticson- a-chip concept is a waveguide lens. Waveguide lenses are of general utility but our specific application is their use to implement the complex crossover interconnections of a switch fabric. The graded refractive index of the lens is engineered by patterning the silicon layer of silicon on insulator slab waveguides into a dense distribution of cylinders; either solid (silicon) or voids (air); using a single etch step. The cylinders have variable diameters and are placed on a regular square or hexagonal grid with sub-wavelength pitch. In the case of voids, the patterned silicon may be suspended in air to form the core of a symmetric slab waveguide. Solid cylinders must be supported by the Si02 layer leading to an asymmetric waveguide of reduced effective index range. Advantageously, the patterning of the metamaterial region within the slab-waveguide requires only a single etch step. Photonic wire feeder waveguides at different positions around the lens may be used to launch light into the lenses or collect light from the lenses. A method is developed to determine the local effective media index of a periodic metamaterial in terms of the parameters of its unit cell. This method is used as a calibration to lay out a metamaterial with graded parameters. The operation of a metamaterial Lüneburg lens telescope is verified by FDTD simulations and shown to be capable of near zero insertion loss and crosstalk. The careful approximation of the graded index of the Lüneburg lens by a metamaterial introduces minimal impairments.

Characterization of an Asymmetric InGaAsP/InP Multi Quantum Well Semiconductor Optical Amplifier

An experimental characterization of broadband semiconductor optical amplifiers (SOAs) at 1360 nm is reported. In addition to their inherent small size, fast dynamics, and feasibility of integration with other optoelectronic components, the relevance of the multi quantum well (MQW) asymmetric SOAs here reported relies on the achievement of a flat and broad 3 dB amplification bandwidth. SOAs are composed of nine In1-xGaxAsyP1-y 0.2% tensile strained MQW layers separated by latticed matched InP barriers. The asymmetry of the active region is based on the difference of the molar concentrations, with Ga (x) ranging from 0.46 to 0.47 and As (y) ranging from 0.89 to 0.94. Devices under test have 7 degrees tilt cleaved facets and feature different geometries: ridge widths from 2 to 4 μm in steps of 0.25 μm, and cavity lengths of 600, 900, 1200, and 1500 μm. Fabry-Pérot (FP) lasers with the same material composition as the SOAs and within the same wafer are used as test structures for parameters extraction, providing a feedback mechanism for further design improvement. The ridge width of the FP lasers varies from 2 to 8 μm, in steps of 2 μm. All the devices have been designed and characterized at the Photonics Technology Laboratory, Centre for Research in Photonics, fabrication was done at Canadian Photonics Fabrication Centre (CPFC), Canada and supported by CMC Microsystems. Devices under test are DC-biased and temperature controlled at 25°C. A single pass gain of 13.5 dB is measured for a 3 dB bandwidth of 60 nm centred at 1360 nm. Light-current plots obtained from the FP lasers show that the threshold current varies with the cavity length, with a minimum of 80 mA for a cavity length of 600 μm and a ridge width of 2 μm. A thermal roll-off occurring at high injection currents is observed, especially with the smallest cavity length. In conclusion, asymmetric MQW SOAs featuring different ridge widths and cavity lengths have been

Effect of the degree of phase-correlation of laser sources on the transmission and optical coherent detection in radio-over-fibre systems

The deployment of high capacity Radio-over-Fiber (RoF) systems rely, among many aspects, on the capability to efficiently generate, transport, and detect millimeter-wave carriers modulated at high data rates. Photonic approaches based on the heterodyne beating of two free-running laser sources have been proposed as an alternative to generate multi-Gbps quadrature phase modulated signals imposed on millimeter wave carriers. Implementing photonic approaches in the down-link avoids the need for electronic generation of high frequency carriers and decreases the requirements at the base band electronics. In addition, implementing complex modulation formats overcomes some of the typical issues found in intensity modulation direct detection approaches such as non linearity, receiver sensitivity and dynamic range. In this work, the performance improvement of a coherent RoF system carrying 10 Gbps QPSK signals is numerically analyzed in terms of both the frequency linewidth and the degree of phase correlation between the lasers utilised at the down-link (for the optical heterodyne beating) and at the up-link (for the optical coherent detection). Relative to phase correlated lasers featuring linewidths of 5 MHz, the peak power of the 60 G Hz carrier generated at the down-link is reduced by 8 dB for un-correlated lasers. In addition, the error vector magnitude of the received signal at the up-link is improved from over 20% (for un-correlated lasers and linewidths of 5 MHz) to around 15% (for correlated lasers) at an optical received power of -30 dBm. The results obtained reinforce the idea of using coherent comb laser sources with phase correlated modes located at the Central Office. It also motivates the eventual deployment of techniques to control the degree of phase correlation between the lasers used as signal and local oscillator at the optical coherent receivers.

An intrinsic limitation to silicon-on-insulator waveguide Mach-Zehnder interference-based electro-optic devices

Mach-Zehnder interference (MZI) construction is broadly exploited to implement optical switches and modulators in the field of integrated optical/photonic technology. Silicon-on-insulator (SOI) waveguides have been increasingly developed to implement highly integrated photonic devices and systems. In this work, for the SOI-waveguide MZI-type electro-optic (EO) modulated devices with free-carrier dispersion (FCD) effect, the FCD-induced extra optical absorption (EOA) loss and its negative impact upon the device performance are studied. An intrinsic limitation to this type of device is found to be the tension between the EOA loss and the interaction length for a half-wave modulation. The numerical calculation and professional software simulation show the EOA loss of <1.0dB determines an interaction length of 4-mm. The performance decay processes of both EO switch and modulator due to the modulation-induced EOA loss are modeled. The numerical calculation shows the optical on-chip (OC) loss is 1.0dB and the isolation between two outputs is 21dB for the EO switch, while for the EO modulator the OC loss is 1.0dB and 2.5dB at the off- and on-state, respectively, and the extinction ratio only approaches 20dB. The negative influence of this intrinsic limitation to the bandwidth of EO modulator is also analyzed.