Showing papers in "IEEE Journal of Selected Topics in Quantum Electronics in 2020"

Photonic Multiply-Accumulate Operations for Neural Networks

Abstract: It has long been known that photonic communication can alleviate the data movement bottlenecks that plague conventional microelectronic processors. More recently, there has also been interest in its capabilities to implement low precision linear operations, such as matrix multiplications, fast and efficiently. We characterize the performance of photonic and electronic hardware underlying neural network models using multiply-accumulate operations. First, we investigate the limits of analog electronic crossbar arrays and on-chip photonic linear computing systems. Photonic processors are shown to have advantages in the limit of large processor sizes ( ${>}\text{100}\; \mu$ m), large vector sizes ( $N > 500)$ , and low noise precision ( ${\leq} 4$ bits). We discuss several proposed tunable photonic MAC systems, and provide a concrete comparison between deep learning and photonic hardware using several empirically-validated device and system models. We show significant potential improvements over digital electronics in energy ( ${>}10^2$ ), speed ( ${>}10^3$ ), and compute density ( ${>}10^2$ ).

Reprogrammable Electro-Optic Nonlinear Activation Functions for Optical Neural Networks

Abstract: We introduce an electro-optic hardware platform for nonlinear activation functions in optical neural networks. The optical-to-optical nonlinearity operates by converting a small portion of the input optical signal into an analog electric signal, which is used to intensity -modulate the original optical signal with no reduction in processing speed. Our scheme allows for complete nonlinear on – off contrast in transmission at relatively low optical power thresholds and eliminates the requirement of having additional optical sources between each of the layers of the network Moreover, the activation function is reconfigurable via electrical bias, allowing it to be programmed or trained to synthesize a variety of nonlinear responses. Using numerical simulations, we demonstrate that this activation function significantly improves the expressiveness of optical neural networks, allowing them to perform well on two benchmark machine learning tasks: learning a multi-input exclusive-OR (XOR) logic function and classification of images of handwritten numbers from the MNIST dataset. The addition of the nonlinear activation function improves test accuracy on the MNIST task from 85% to 94%.

Digital Electronics and Analog Photonics for Convolutional Neural Networks (DEAP-CNNs)

Abstract: Convolutional Neural Networks (CNNs) are powerful and highly ubiquitous tools for extracting features from large datasets for applications such as computer vision and natural language processing. However, a convolution is a computationally expensive operation in digital electronics. In contrast, neuromorphic photonic systems, which have experienced a recent surge of interest over the last few years, propose higher bandwidth and energy efficiencies for neural network training and inference. Neuromorphic photonics exploits the advantages of optical electronics, including the ease of analog processing, and busing multiple signals on a single waveguide at the speed of light. Here, we propose a Digital Electronic and Analog Photonic (DEAP) CNN hardware architecture that has potential to be 2.8 to 14 times faster while using almost 25% less energy than current state-of-the-art graphical processing units (GPUs).

Analysis of Diffractive Optical Neural Networks and Their Integration With Electronic Neural Networks

Abstract: Optical machine learning offers advantages in terms of power efficiency, scalability, and computation speed. Recently, an optical machine learning method based on diffractive deep neural networks (D2NNs) has been introduced to execute a function as the input light diffracts through passive layers, designed by deep learning using a computer. Here, we introduce improvements to D2NNs by changing the training loss function and reducing the impact of vanishing gradients in the error back-propagation step. Using five phase-only diffractive layers, we numerically achieved a classification accuracy of 97.18% and 89.13% for optical recognition of handwritten digits and fashion products, respectively; using both phase and amplitude modulation (complex-valued) at each layer, our inference performance improved to 97.81% and 89.32%, respectively. Furthermore, we report the integration of D2NNs with electronic neural networks to create hybrid classifiers that significantly reduce the number of input pixels into an electronic network using an ultra-compact front-end D2NN with a layer-to-layer distance of a few wavelengths, also reducing the complexity of the successive electronic network. Using a five-layer phase-only D2NN jointly optimized with a single fully connected electronic layer, we achieved a classification accuracy of 98.71% and 90.04% for the recognition of handwritten digits and fashion products, respectively. Moreover, the input to the electronic network was compressed by >7.8 times down to 10 × 10 pixels. Beyond creating low-power and high-frame rate machine learning platforms, D2NN-based hybrid neural networks will find applications in smart optical imager and sensor design.

III-V-on-Silicon Integration: From Hybrid Devices to Heterogeneous Photonic Integrated Circuits

Abstract: Heterogeneous integration of III-V materials onto silicon photonics has experienced enormous progress in the last few years, setting the groundwork for the implementation of complex on-chip optical systems that go beyond single device performance. Recent advances on the field are expected to impact the next generation of optical communications to attain low power, high efficiency and portable solutions. To accomplish this aim, intense research on hybrid lasers, modulators and photodetectors is being done to implement optical modules and photonic integrated networks with specifications that match the market demands. Similarly, important advances on packaging and thermal management of hybrid photonic integrated circuits (PICs) are currently in progress. In this paper, we report our latest results on hybrid III-V on Si transmitters, receivers and packaged optical modules for high-speed optical communications. In addition, a review of recent advances in this field will be provided for benchmarking purposes.

Deep Neural Network Through an InP SOA-Based Photonic Integrated Cross-Connect

Abstract: Photonic neuromorphic computing is raising a growing interest as it promises to provide massive parallelism and low power consumption. In this paper, we demonstrate for the first time a feed-forward neural network via an 8 × 8 Indium Phosphide cross-connect chip, where up to 8 on-chip weighted addition circuits are co-integrated, based on semiconductor optical amplifier technology. We perform the weight calibration per neuron, resulting in a normalized root mean square error smaller than 0.08 and a best case dynamic range of 27 dB. The 4 input to 1 output weighted addition operation is executed on-chip and is part of a neuron, whose non-linear function is implemented via software. A three feedback loop optimization procedure is demonstrated to enable an output neuron accuracy improvement of up to 55%. The exploitation of this technology as neural network is evaluated by implementing a trained 3-layer photonic deep neural network to solve the Iris flower classification problem. Prediction accuracy of 85.8% is achieved, with respect to the 95% accuracy obtained via a computer. A comprehensive analysis of the error evolution in our system reveals that the electrical/optical conversions dominate the error contribution, which suggests that an all optical approach is preferable for future neuromorphic computing hardware design.

MEMS for Photonic Integrated Circuits

Abstract: The field of microelectromechanical systems (MEMS) for photonic integrated circuits (PICs) is reviewed. This field leverages mechanics at the nanometer to micrometer scale to improve existing components and introduce novel functionalities in PICs. This review covers the MEMS actuation principles and the mechanical tuning mechanisms for integrated photonics. The state of the art of MEMS tunable components in PICs is quantitatively reviewed and critically assessed with respect to suitability for large-scale integration in existing PIC technology platforms. MEMS provide a powerful approach to overcome current limitations in PIC technologies and to enable a new design dimension with a wide range of applications.

Toward Neuromorphic Photonic Networks of Ultrafast Spiking Laser Neurons

Abstract: We report on ultrafast artificial laser neurons and on their potentials for future neuromorphic (brainlike) photonic information processing systems. We introduce our recent and ongoing activities demonstrating controllable excitation of spiking signals in optical neurons based upon vertical-cavity surface emitting lasers (VCSEL-Neurons). These spiking regimes are analogous to those exhibited by biological neurons, but at sub-nanosecond speeds (>7 orders of magnitude faster). We also describe diverse approaches, based on optical or electronic excitation techniques, for the activation/inhibition of sub-ns spiking signals in VCSEL-Neurons. We report our work demonstrating the communication of spiking patterns between VCSEL-Neurons toward future implementations of optical neuromorphic networks. Furthermore, new findings show that VCSEL-Neurons can perform multiple neuro-inspired spike processing tasks. We experimentally demonstrate photonic spiking memory modules using single and mutually coupled VCSEL-Neurons. Additionally, the ultrafast emulation of neuronal circuits in the retina using VCSEL-Neuron systems is demonstrated experimentally for the first time to our knowledge. Our results are obtained with off-the-shelf VCSELs operating at the telecom wavelengths of 1310 and 1550 nm. This makes our approach fully compatible with current optical network and data center technologies, hence offering great potentials for future ultrafast neuromorphic laser-neuron networks for new paradigms in brain-inspired computing and artificial intelligence.

Recent Progress in Low-Loss Hollow-Core Anti-Resonant Fibers and Their Applications

Abstract: In the research field of hollow-core optical fiber (HCF), one type of fiber geometry with a leaky mode nature has unexpectedly taken center stage over the last couple of years: the so-called hollow-core anti-resonant fiber (ARF). The guidance mechanism of this ARF has been elucidated explicitly, the optical performance of the fiber has improved significantly, and the range of potential fiber application areas has expanded steadily. This paper will review our continuous efforts to understand, design, and fabricate this hollow-core ARF with the aim of lower loss and wider bandwidth. We also explore the possibility of using an advanced form of ARF in communications applications. In the long journey of looking for optical fibers that provide better performance than conventional solid-core glass fibers, exploitation of the hidden potential of artificial photonic micro-structures will continue to advance.

Reservoir Computing Using Multiple Lasers With Feedback on a Photonic Integrated Circuit

Abstract: We propose a scheme for reservoir computing using multiple semiconductor lasers with optical feedback arranged in parallel on a photonic integrated circuit, and we investigate the performance of reservoir computing numerically. The virtual nodes are obtained from the temporal waveforms of the outputs of the parallel reservoir lasers. We test the chaotic time-series prediction task, memory capacity, and nonlinear channel equalization task to investigate the performance of reservoir computing. We found that our scheme using multiple lasers outperforms that using a single laser with multiple delay times. Large memory capacity can also be obtained for the multiple lasers. Finally, we investigate the effect of parameter mismatch of the multiple lasers on reservoir computing performance.

Optical Reservoir Computing Using Multiple Light Scattering for Chaotic Systems Prediction

Abstract: Reservoir Computing is a relatively recent computational framework based on a large Recurrent Neural Network with fixed weights. Many physical implementations of Reservoir Computing have been proposed to improve speed and energy efficiency. In this study, we report new advances in Optical Reservoir Computing using multiple light scattering to accelerate the recursive computation of the reservoir states. Two different spatial light modulation technologies, namely, phase or binary amplitude modulations, are compared. Phase modulation is a promising direction already employed in other photonic implementations of Reservoir Computing. Additionally, we report a Digital-Micromirror-based Reservoir Computing at up to 640 Hz, more than double the previously reported frequency using a remotely controlled optical device developed by LightOn, and present new binarization strategies to improve the performance of binarized Reservoir Computing.

Femtojoule per MAC Neuromorphic Photonics: An Energy and Technology Roadmap

Abstract: Photonic artificial neural networks have garnered enormous attention due to their potential to perform multiply-accumulate (MAC) operations at much higher clock rates and consuming significantly lower power and chip real-estate compared to digital electronic alternatives. Herein, we present a comprehensive power consumption analysis of photonic neurons, taking into account global design parameters and concluding to analytical expressions for the neuron's energy- and footprint efficiencies. We identify the optimal design-space and analyze the performance plateaus and their dependence on a range of physical parameters, highlighting the existence of an optimal data-rate for maximizing the energy efficiency. Following a survey of the best-in-class integrated photonic devices, including on-chip lasers, photodetectors, modulators and weighting elements, the mathematically calculated energy and footprint efficiencies are mapped into real photonic neuron deployment scenarios. We reveal that silicon photonics can compete with the best-performing currently available digital electronic neural network engines, reaching TMAC/s/mm2 footprint- and sub-pJ/MAC energy efficiencies. Simultaneously, neuromorphic plasmonics, plasmo-photonics and sub-wavelength photonics hold the credentials for 1 to 3 orders of magnitude improvements even when the laser requirements and a reasonable waveguide pitch are accounted for, promising performance at a few fJ/MAC and up to a few TMAC/s/mm2.

Parallel Programming of an Arbitrary Feedforward Photonic Network

Abstract: Reconfigurable photonic mesh networks of tunable beamsplitter nodes can linearly transform $N$ -dimensional vectors representing input modal amplitudes of light for applications such as energy-efficient machine learning hardware, quantum information processing, and mode demultiplexing. Such photonic meshes are typically programmed and/or calibrated by tuning or characterizing each beam splitter one-by-one, which can be time-consuming and can limit scaling to larger meshes. Here we introduce a graph-topological approach that defines the general class of feedforward networks and identifies columns of non-interacting nodes that can be adjusted simultaneously. By virtue of this approach, we can calculate the necessary input vectors to program entire columns of nodes in parallel by simultaneously nullifying the power in one output of each node via optoelectronic feedback onto adjustable phase shifters or couplers. This parallel nullification approach is robust to fabrication errors, requiring no prior knowledge or calibration of node parameters and reducing programming time by a factor of order $N$ to being proportional to the optical depth (number of node columns). As a demonstration, we simulate our programming protocol on a feedforward optical neural network model trained to classify handwritten digit images from the MNIST dataset with up to 98% validation accuracy.

Ring-Resonator Based Widely-Tunable Narrow-Linewidth Si/InP Integrated Lasers

Abstract: This paper presents recent results on widely-tunable narrow-linewidth semiconductor lasers using a ring-resonator based mirror as the extended cavity. Two generations of lasers on the heterogeneous Si/InP photonic platform are presented. The first-generation lasers, with a total footprint smaller than 0.81 mm2, showed an intrinsic linewidth of ∼2 kHz over a 40 nm wavelength tuning range across C+L bands. The second-generation lasers using ultra-low loss silicon waveguides and a novel cavity design achieved an intrinsic linewidth below 220 Hz. The lasers also possess an ultrawide wavelength tuning range of 110 nm across three optical communication bands (S+C+L). These are records among all fully integrated semiconductor lasers reported in the literature.

Graphene-Based High-Efficiency Broadband Tunable Linear-to-Circular Polarization Converter for Terahertz Waves

Abstract: Broadband tunable terahertz (THz) polarization converter is a vivid necessity due to its great significance in functional utilization. Graphene is a promising material for the evolution of tunable terahertz devices, thanks to its tunable surface conductivity and extraordinary electrical properties. In this paper, we present a high-efficiency graphene-based ultra-thin metasurface manifesting linear-to-circular (LTC) polarization conversion for THz wave. A broadband LTC conversion with ellipticity over 0.95 can be achieved for the entire operating bandwidth of 14 THz to 40 THz for any graphene chemical potential from 0.4 eV to 0.6 eV. It is also shown that tuning the graphene chemical potential may provide much improved ellipticity and efficiency for different frequencies. The proposed LTC converter also provides over 90% efficiency for the complete operating bandwidth for graphene chemical potential from 0 eV to 0.7 eV. Simulation results also reveal that a perfect LTC converter (ellipticity = 1) with more than 90% efficiency can be realized from 16 THz to 37 THz by tuning the graphene chemical potential between 0 eV to 0.7 eV. Altogether, the proposed LTC polarization converter provides the advantage of dynamic regulation, simpler structure, higher efficiency, enhanced ellipticity and greater relative bandwidth.

Integrated 256 Cell Photonic Phase-Change Memory With 512-Bit Capacity

Abstract: All-optical nonvolatile memories enable storage of telecommunication data without detours through electronic circuitry. Phase-change materials provide the means to embed such memories within integrated optical circuits and thus allow combining waveguide devices for information processing with local data storage. Using this concept, we realize an all-photonic memory circuit capable of storing 512 bits of data in an array of nanoscale phase-change devices. We employ multilevel addressing and wavelength multiplexing of microring resonators to write and read a 16 × 16 greyscale image with 2-bit resolution entirely in the optical domain. Our approach holds promise for implementing scalable architectures for on-chip optical data storage with long data retention and ultrafast access times.

Polarization Multiplexing Reservoir Computing Based on a VCSEL With Polarized Optical Feedback

Abstract: We propose to realize the parallel tasks processing by reservoir computing (RC) using two polarization-resolved modes in a single vertical-cavity surface-emitting laser (VCSEL). Two feedback cases, parallelly-polarized optical feedback (PPOF) and orthogonally-polarized optical feedback, are considered in the VCSEL. A time series prediction task and a waveform recognition task are employed to examine the parallel tasks processing performance in two polarization-resolved modes of the VCSEL-based RC system. Through numerical simulations, the dependences of parallel tasks processing performance of the VCSEL-based RC system on the feedback strength, injection strength are analyzed carefully and compared for the two feedback cases. Besides, the effects of injection current, frequency detuning, spontaneous emission noise, and the number of virtual nodes are also examined. It is found that, for both feedback cases, good performance of parallel tasks processing in two polarization modes of the VCSEL-based RC system could be obtained in broad parameter regions, and in these regions a better performance is achieved for the PPOF case. This proposed polarization multiplexing RC based on VCSEL permits parallel tasks processing in two polarization modes in a single VCSEL, and hence is interesting and valuable for low power consumption neuromorphic photonics systems.

Photonics Inverse Design: Pairing Deep Neural Networks With Evolutionary Algorithms

Abstract: Deep Neural Networks (DNN) have shown early promise for inverse design with their ability to arrive at working designs much faster than conventional optimization techniques. Current approaches, however, require complicated workflows involving training more than one DNN to address the problem of non-uniqueness in the inversion and the emphasis on speed has overshadowed the far more important consideration of solution optimality. We propose and demonstrate a simplified workflow that pairs forward-model DNN with evolutionary algorithms which are widely used for inverse gg design. Our evolutionary search in forward-model space is global and exploits the massive parallelism of modern GPUs for a speedy inversion. We propose a hybrid approach where the DNN is used only for preselection and initialization that is more effective at optimization than a standalone DNN and performs nearly as well as a vanilla evolutionary search with a significantly reduced function evaluation budget. We finally show the utility of an iterative procedure for building the training dataset which further boosts the effectiveness of this approach.

Optical Frequency Comb and Nyquist Pulse Generation With Integrated Silicon Modulators

Abstract: We demonstrate nine-line optical frequency comb (OFC) and sinc-shaped Nyquist pulse generation based on two cascaded silicon Mach–Zehnder modulators. The OFC has a flatness of 1.83 dB, a frequency spacing of 5 GHz, and a corresponding Nyquist pulsewidth of ∼22 ps. Moreover, multi-wavelength operation and repetition rate tunability have also been demonstrated, which indicates the potential of our work to generate an on-chip high-speed pulse train demanded in optical communications and microwave photonics applications.

Tunable Beam Steering at Terahertz Frequencies Using Reconfigurable Metasurfaces Coupled With Liquid Crystals

Abstract: Metasurfaces with a spatially varying phase profile enable the design of planar and compact devices for manipulating the radiation pattern of electromagnetic fields. Aiming to achieve tunable beam steering at terahertz frequencies, we numerically investigate metasurfaces consisting of one dimensional arrays of metal-insulator-metal (MIM) cavities infiltrated with liquid crystals (LCs). The spatial phase profile is defined by a periodic voltage pattern applied on properly selected supercells of the MIM-cavity array. By means of the electro-optic effect, the voltage controls the orientation of LC molecules and, thus, the resulting effective LC refractive index. Using this approach, the spatial phase profiles can be dynamically switched among a flat, binary, and gradient profile, where the corresponding metasurfaces function as mirrors, beam splitters or blazed gratings, respectively. Tunable beam steering is achieved by changing the diffraction angle of the first diffraction order, through the reconfiguration of the metasurface period via the proper adjustment of the applied voltage pattern.

A Winograd-Based Integrated Photonics Accelerator for Convolutional Neural Networks

Abstract: Neural Networks (NNs) have become the mainstream technology in the artificial intelligence (AI) renaissance over the past decade. Among different types of neural networks, convolutional neural networks (CNNs) have been widely adopted as they have achieved leading results in many fields such as computer vision and speech recognition. This success in part is due to the widespread availability of capable underlying hardware platforms. In parallel, hardware specialization can expose us to novel architectural solutions, which can outperform general purpose computers for the tasks at hand. Although different applications demand for different performance measures, they all share speed and energy efficiency as high priorities. Meanwhile, photonics processing has seen a resurgence due to its inherited high speed and low power nature. Here, we investigate the potential of using photonics in CNNs by proposing a CNN accelerator design based on Winograd filtering algorithm. Our evaluation results show that while a photonic accelerator can compete with current state-of-the-art electronic platforms in terms of both speed and power, it has the potential to improve the energy efficiency by up to three orders of magnitude.

Photonic Computing With Single and Coupled Spiking Micropillar Lasers

Abstract: We review experimental and theoretical results on the computing properties of single spiking micropillar lasers and present numerical studies of propagation and computing in chains of evanescently coupled micropillar lasers Single micropillar lasers are shown to behave as ultrafast optical neurons with sub-nanosecond spike times They also possess absolute and relative refractory times, spike latency, and show temporal summation With delayed optical feedback, they emulate an autapse These basic neural properties can be used for simple photonic computing We show by numerical simulations of a chain of coupled spiking micropillar lasers that basic logical operations can be implemented in photonic circuits, as well as temporal pattern recognition based on the collision properties of pulses in this chain

Long-Distance Continuous-Variable Quantum Key Distribution With Quantum Scissors

Abstract: We investigate the use of quantum scissors, as candidates for non-deterministic amplifiers, in continuous-variable quantum key distribution. Such devices rely on single-photon sources for their operation and as such, they do not necessarily preserve the Guassianity of the channel. Using exact analytical modeling for the system components, we bound the secret key generation rate for a protocol that uses quantum scissors. We find that, for certain non-zero values of excess noise, such a protocol can reach longer distances than the counterpart with no amplification. This sheds light into the prospect of using quantum scissors as an ingredient in continuous-variable quantum repeaters.

Simultaneous Measurement of Temperature and Refractive Index Using an Exposed Core Microstructured Optical Fiber

Abstract: We have demonstrated a novel scheme for simultaneous measurement of temperature and refractive index by using an exposed core microstructured optical fiber (ECF). The ECF allows for high sensitivity to refractive index due to the small exposed-core, while being supported by a standard fiber diameter cladding making it robust compared to optical microfibers. The sensor combines a fiber Bragg grating (FBG) inscribed into the core of the ECF and a multimode Mach–Zehnder interferometer (MZI). Both the FBG and MZI are sensitive to refractive index (RI) and temperature through a combination of direct access to the evanescent field via the exposed-core, the thermo-optic effect, and thermal expansion. The FBG and MZI respond differently to changes in temperature and RI, thus allowing for the simultaneous measurement of these parameters. In our experiment, RI sensitivities of 5.85 nm/RIU and 794 nm/RIU, and temperature sensitivities of 8.72 pm/°C and −57.9 pm/°C, were obtained for the FBG and MZI respectively. We demonstrate that a transfer matrix approach can be used to simultaneously measure both parameters, solving the problem of temperature sensitivity of RI sensors due to the high thermo-optic coefficient of aqueous samples.

Temporal Information Processing With an Integrated Laser Neuron

Abstract: Spiking neural networks enable efficient information processing in real-time. Excitable lasers can exhibit ultrafast spiking dynamics, and when preceded by a photodetector in an O/E/O link, can process optical spikes at different wavelengths and thus be interconnected in large neural networks. Here, we experimentally demonstrate and numerically simulate the spiking dynamics of a laser neuron fabricated in a photonic integrated circuit. Our spiking laser neuron is shown to perform coincidence detection with nanosecond time resolution, and we observe refractory periods in the order of 0.1 ns. We propose a method to implement XOR classification using our laser neurons, and simulations of the resultant dynamics indicate robust tolerance to timing jitter.

Programmable Photonics: An Opportunity for an Accessible Large-Volume PIC Ecosystem

Abstract: We look at the opportunities presented by the new concepts of generic programmable photonic integrated circuits (PIC) to deploy photonics on a larger scale. Programmable PICs consist of waveguide meshes of tunable couplers and phase shifters that can be reconfigured in software to define diverse functions and arbitrary connectivity between the input and output ports. Off-the-shelf programmable PICs can dramatically shorten the development time and deployment costs of new photonic products, as they bypass the design-fabrication cycle of a custom PIC. These chips, which actually consist of an entire technology stack of photonics, electronics packaging and software, can potentially be manufactured cheaper and in larger volumes than application-specific PICs. We look into the technology requirements of these generic programmable PICs and discuss the economy of scale. Finally, we make a qualitative analysis of the possible application spaces where generic programmable PICs can play an enabling role, especially to companies who do not have an in-depth background in PIC technology.

Silicon Nitride Photonics for the Near-Infrared

Abstract: In recent years, silicon nitride (SiN) has drawn attention for the realisation of integrated photonic devices due to its fabrication flexibility and advantageous intrinsic properties that can be tailored to fulfill the requirements of different linear and non-linear photonic applications. This paper focuses on our progress in the demonstration of enhanced functionalities in the near infrared wavelength regime with our low temperature ( ${^\circ}$ C) SiN platform. It discusses (de)multiplexing devices, nonlinear all optical conversion, photonic crystal structures, the integration with novel phase change materials, and introduces applications in the 2 $\mu$ m wavelength range.

Topological Edge Modes in Non-Hermitian Photonic Aharonov-Bohm Cages

Abstract: We investigate the topological edge modes in a rhombic waveguide array with imaginary coupling, which is realized by incorporating auxiliary waveguide between the coupled waveguides. By suitably tuning both real and imaginary couplings to generate an effective π flux, a non-Hermitian analog of Aharonov-Bohm (AB) cage is formed as the band structures become flat and coalesce into third-order exceptional points (EPs). We show the array can support gapped topological edge modes, which can be explained by mapping the system Hamiltonian into the square root of an anti-parity-time (PT) -symmetric waveguide array. In contrast to the linear power increase of bulk modes, the propagation of edge modes can be conserved or exponentially amplified depending on which termination is initially excited. Our study provides a promising way to realizing robust light propagation.

Control and Calibration Recipes for Photonic Integrated Circuits

Abstract: This article presents the key ingredients and the best practices for implementing simple, effective and robust control and calibration procedures for arbitrary photonic integrated circuit (PIC) architectures. Three main features are presented and discussed: a technique to cancel out the effects of mutual crosstalk among thermal tuners, the exploitation of labelling to identify different optical signals, the use of input modulated signal to automatically reshape the frequency response of the device. Examples of application are then illustrated to show the validity and generality of the approach, namely a cross-bar interconnect matrix router, a variable bandwidth filter and third order coupled microring filter. Further, the automatic and dynamic generation of the lookup table of add/drop hitless filters operating on a dense wavelength division multiplexing grid is demonstrated. The lookup table achieved with the proposed approach can dynamically update itself to new conditions of the chip or new requirements of operation, such as variations in channel modulation format or perturbation induced by neighboring devices due to a change in their working point.

Programmable Integrated Silicon Photonics Waveguide Meshes: Optimized Designs and Control Algorithms

Abstract: Programmable Integrated Photonics is a recent area of research that aims to integrate a very-large scale of reconfigurable photonic components to enable flexible and versatile photonic integrated circuits. In this paper, we review the state of the art of general-purpose waveguide mesh arrangements with a special focus on those that allow the synthesis of optical feedback loops. Moreover, we propose for the first time, a new design approach to generate waveguide mesh patterns with equally-oriented components. This innovation is of special relevance to improve performance and to mitigate one of the main scalability limitations, the integration density. The paper finalizes with an introduction to control algorithms for waveguide mesh arrangements based on derivative methods and non-derivative methods. These control methods provide a proof for the self-reconfiguration of large-scale waveguide mesh arrangements. In particular, we apply the computational optimization algorithms to program a hexagonal waveguide mesh to emulate a 1 × 8 beamforming network and an optical filter based on an unbalanced MZI design. All in all, the paper comprises recipes to achieve truly practical software-defined photonic integrated circuits.