This PDF file contains the front matter associated with SPIE Proceedings Volume 12890, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Light-Emitting Diodes (LEDs) exhibit a typical Lambertian emission, usually requiring reshaping by means of secondary optics. We review the potential of various integrated photonic architectures to address the demand for system miniaturization and high efficiency in emerging applications. Photonic structures that can be potentially integrated in LED devices include metalenses, photonic crystals and reflective metasurfaces. We embed periodic nanoantennas in InGaN/GaN multi-quantum well (MQW) LEDs to control their far-field emission directionality and enhance collection efficiency. We propose exploiting mechanisms such as surface lattice resonances, which rely on the near-field coupling between the quantum wells and the nanoantenna array. Multiple experimental and modeling studies demonstrate the benefits and challenges of optimized integrated metasurfaces to enable efficient SSL sources without the need of bulky secondary optics for directional beam control.

Convolution neural networks (CNNs) have garnered significant attention in today's digital world due to their remarkable ability to analyze medical images, recognize speech, translate languages, and make intelligent decisions. However, the convolution operation, which plays a vital role in neural networks, imposes a high computational cost, increasing power consumption and latency, particularly when dealing with large matrices. Integrated photonics presents an exciting opportunity to drastically accelerate CNNs with its virtual O(1) cost and speed-of-light latency, by leveraging the properties of Fourier optics. In this study, we propose the design, fabrication and testing of an integrated photonic system for convolutional neural networks based on the 4F system. Moreover, we introduce an innovative backscattering method to accurately adjust the phase of each micro-disk modulator (MDM), ensuring precise on-chip Fourier transform operations. Furthermore, we demonstrate the system's potential for achieving higher speeds through its scalability with multiple wavelengths and optical buffers.

An approach to measuring chemical concentrations using a micro-ring resonator (MRR) is proposed which is robust to thermo-optic noise and spectral shifts caused by temperature changes. The method uses a modified ResNet50 with varied kernel size and achieved a mean-square error (MSE) of 4.548E-4, and performance is compared to other machine learning methods including VGG20 and XGBoost. The model was trained to read the transmission spectra of a slotted MRR etched into heavily doped silicon and output the concentrations of chemicals in the surrounding analyte. The chemicals tested on were water, ethanol, methanol, and propanol, with concentrations ranging from 0-100%, with a dataset containing . This occurs over the mid-infrared wavelengths and within the temperature range of 290-310 K. Transfer learning was also utilized to retrain the models on several other datasets, consisting of 528 transmissions each. These datasets operated over different temperature ranges (310-320), and the other with a different set of chemicals, (water, ethanol, methanol and butanol). Similar results were achieved, with both networks achieving similar MSE. We then perform the same process on another design with the same chemicals, also operating over the infrared range, demonstrating the robustness of the method. All datasets used in the study were obtained through simulation, although we hope to test on real data.

In this study, we propose on-chip deep photonic networks with custom-designed Mach-Zehnder interferometers (MZI), allowing for the design of devices with arbitrarily defined dispersion profiles. We demonstrate a custom simulation and optimization framework to optimize dispersion and transmission profiles of these MZI-based networks. We experimentally show a proof-of-concept two-port photonic network with a highly nonintuitive, triangular dispersion profile in the C-band, while simultaneously achieving a flat band transmission with less than 0.7 dB insertion loss. We also demonstrate capabilities with multi-port photonic networks to enhance design freedom for customizable dispersion profiles, opening up new possibilities for on-chip dispersion engineering.

We report on the progress of our efforts to apply silicon nitride photonic integrated circuits (PIC) to the miniaturization of optical coherence tomography (OCT) with the goal of facilitating its widespread use in ophthalmology at the point of care. In particular, we highlight the design and optical characterization of photonic building blocks allowing the realization of a silicon nitride PIC-based multi-channel swept-source OCT system in the 1060 nm wavelength region. Apart from waveguide structures, these building blocks include 3D-printed microlenses on the PIC end facets for efficient light coupling to and from the PIC.

This study presents the development and optimization of AlN-based optical waveguides and photonic devices for 1.55 μm applications, utilizing a CMOS-compatible sputtered thin film technology. The fabrication employs a two-step photoresist technique to fabricate AlN core waveguides with low propagation losses and other optical devices, such as a 1×2 multimode interference (MMI) coupler, Mach-Zehnder interferometers and ring resonators. The results of this study highlight the potential for efficient and cost-effective fabrication of AlN-based integrated optical circuits and pave the way for further research and development in the field of integrated optics.

Quantum photonic devices are typically based on either discrete variable (DV) or continuous variable (CV) principles. In either case, their performance must meet strict “fidelity” requirements, especially when scaled using integrated photonic circuitry. DV devices must be robust to imperfections such as timing delays between photons, unwanted spectral correlations present in photons from nonideal single photon sources and variations in circuit components. Furthermore, it becomes increasingly important to have a methodology for characterizing the performance of devices in the presence of such errors. Here, we present a method for simulating quantum photonic integrated circuits (PICs) and use it to model the behavior of a circuit constructed to model the evolution of a quantum state subject to a Bose-Hubbard Hamiltonian, introducing variations to the coupling gaps in the directional couplers inside the circuit. Input states consisting of both indistinguishable and distinguishable photons are modelled. Compact models for the directional couplers and other photonic circuit elements that are parameterized for a specific fabrication process can be incorporated into the circuit using a foundry-associated compact model library, ensuring consistency between circuit design and manufactured component. In CV applications, the degree of squeezing provided by some nonlinear element is often a key metric. Here we describe how the degree of squeezing produced by spontaneous four-wave mixing (SFWM) in a microring resonator can be modelled in the low-power limit, accounting for the effects of self-phase modulation (SPM), cross-phase modulation (XPM), and component losses.

Plasmonic sensing using surface plasmon resonances remains an area with unclear detection limits. The introduction of metallic nanostructures as nanoantennas can improve plasmonic sensing through localized surface plasmon resonances (LSPR), complex lattice effects, and simultaneous propagating plasmons. The combination of plasmonic effects through near or far-field coupling lead to more complex phenomena called plasmonic hybridization. Multivariate analysis methods significantly improve the sensing performance in terms of figures of merit by exploiting multiple features of the resonance curves. Likewise, multivariate design approaches provide optimized plasmonic substrates through electromagnetic simulations when the same curve features are used as in multivariate data analysis to achieve optimized performance metrics. Here, we demonstrate with one-dimensional gold nanogratings that multivariate design concepts lead to optimized performance in plasmonic sensing, which is also confirmed experimentally.

Since the turn of the century, Silicon Photonics (SiPh) has advanced data communication and computing via diverse integrated optical functions and multiplexing strategies. However, conventional design methodologies limit scalability, and inverse designs lead to features sensitive to fabrication process variations. This talk explores the harnessing of machine learning (ML) to predict and rectify these deviations in the design phase. This technique enhances design fidelity and device performance, while facilitating smaller design features, bypassing constraints of traditional methods. Highlighting PreFab, an innovative ML technology, applicable to both conventional and inverse designs, it predicts and corrects fabrication deviations, enabling refined design processes.

In this paper we present the developments carried out at III-V lab in collaboration with CEA-Leti on the hybrid integration of III-V chips with ultra-low-loss silicon nitride circuits for optical detection and RF systems. Different types of hybrid circuits will be presented: 1- a narrow linewidth laser based on the combination of an InP R-SOA and a long Bragg grating on SiN, 2- a Kerr frequency comb based on a high-power InP DFB laser butt-coupled to a high-Q SiN micro-ring, 3- a long cavity mode-locked laser based on an R-SOA and a SiN circuit with a delay line and a Bragg grating.

Optical isolator is indispensable for establishing stable photonic systems by eliminating backward reflections. We have investigated waveguide optical isolators for photonic integrated circuits using magneto-optical (MO) effects. There are two challenges; difficulty in integration method of MO crystal and difficulty in Faraday rotation in planar waveguide systems. In this talk, recent progress of the integrated MO isolators is reviewed.

We report the demonstration of an integrated silicon-rich silicon nitride wavelength converter based on the Bragg scattering intermodal four-wave mixing process. This broadband wavelength converter incorporates on-chip mode conversion, multiplexing and de-multiplexing functionalities. The system allows for broadband signal conversion with a 3dB bandwidth exceeding 70 nm.

While highly coherent light has its advantages in certain applications, it also leads to some unwanted effects, such as the formation of speckles, negatively impacting imaging, lighting, and projection/display quality. This has prompted the development of edge-emitting semiconductor light sources with partial coherence, such as broadband lasers and superluminescent diodes (SLDs), which have found use in display applications, optical coherence tomography, and random bit generation. However, the edge-emission design of these sources complicates the fabrication of two-dimensional arrays, preventing their use in many applications. On the other hand, vertical-cavity surface-emitting lasers (VCSELs) have gained popularity in consumer electronics due to the unique benefits of surface emission. Yet, they face challenges related to high coherence. We demonstrate that by leveraging the distinctive properties of chaotic cavities, we can reduce the coherence of VCSELs without enlarging their emission area. This enables high-speed wireless communication with low-speckle illumination, as confirmed through comprehensive interference studies and high-speed testing. We show that the chaotic-cavity VCSELs can provide better illumination quality while maintaining high data rates (12.6 Gb/s over short distances and 6.5 Gb/s over 3 m). In addition, the use of a carefully designed AC signal for communication can further lower the speckle contrast, which improves the illumination performance. This paves the way for the potential use of chaotic-cavity VCSELs as versatile light sources for illumination, imaging, projection, communication, and various other applications.

Guaranteeing high signal-to-noise ratio and good optical performance over time of Ge-based photodetectors (PD) is challenging. Encompassing both measurement methodology and understanding the physics of failure. In this paper, we highlight recent developments in this field. Methodology-wise, we present a lifetime-assessment for package-level qualification as well as a wafer-level qualification test, including stress and relaxation study of the degradation. Regarding physics-of-failure, we propose a methodology to differentiate dark current degradation originating from the Ge-bulk, the Ge/SiO₂ interface and the Ge/Si interface.

The fully epitaxial integration of IR laser sources into modern photonic circuits built on Si or SOI wafers is severely limited by the thermal- and lattice-constant mismatch between the substrate and the III-V layers that are required to achieve efficient solid-state lasing in the near infrared range. To overcome these limitations, modern commercially-available SiPh technologies employ selective wafer/device bonding techniques that allow to separately grow the III-V emitter and to integrate it into the photonic integrated circuit (PIC) at a later processing stage. Conversely, state-of-the-art devices leverage InAs quantum-dot (QD)-based active regions to highly reduce the sensitivity of the laser diode to the presence of the extended defects, which are generated as a consequence of the heteroepitaxial growth. In term of reliability, high levels of maturity have been demonstrated by both types of sources, either on the field or at laboratory level. Despite this, several degradation mechanisms still affect the long-term operation of such devices, thus limiting their useful lifetime. The aim of this paper is to discuss on the dominant degradation processes related to integrated laser sources for silicon photonics. This goal is achieved by summarizing some of the most recent results that have been obtained on two different, but well representative, classes of solid-state lasers: heterogeneously integrated quantum well (QW)-based emitters, such as vertical-cavity silicon-integrated lasers (VCSILs) emitting at 945 nm and III-V 1.55 m lasers bonded on silicon-on-insulator (SOI) substrates, and InAs QD laser diodes (LDs) epitaxially-grown on silicon, emitting in the 1.31 m window.

We present a non-intuitive but still practically implementable silicon meta-atom with strong electrical tunability. We used topological optimization on a 20×20-variable silicon-on-insulator grid and achieved tunability by using JRD1, a stable electrooptic polymer. The electric potential is applied on two ITO electrodes, and a peak shift of 0.6 nm V⁻¹ is achieved in the optical communication E and S bands. The proposed device can be regarded as a demultiplexer or channel selector on a coarse wavelength-division-multiplexing system, with an extinction ratio between adjacent channels of ∼14 dB.

A solid-state FMCW LiDAR system on a photonic chip can be subdivided into its scanning engine and ranging engine. The scanning engine consists of the laser source and beam steering optics, while the ranging engine consists of a modulated transmitted signal and a coherent receiver. Here, we present the practical considerations of using the wavelength of a swept laser source for both FMCW ranging and beam steering through a dispersive optical phased array (DOPA). Controlling both functions with a single variable (the laser wavelength) is very attractive from a system perspective. The DOPA maps the wavelength injected in its input port to a unique spot in the far-field. Thanks to reciprocity, this mapping also works in reverse and so the reflected beam is sent back to the same input port. Consequently, the DOPA not only provides a simple mechanism for the steering and receiving optics, but also a strong rejection of unwanted signals, since the DOPA acts both as a spatial and a spectral filter. Steering based on sweeping the wavelength can be combined with the frequency ramping for FMCW ranging. The sweep rate and tuning range of the laser determines the far-field coverage, as well as the depth resolution and precision of the ranging engine. The design freedom of the system can be extended by feeding multiple lasers in parallel into the DOPA, without the need for extra wavelength filters. The use of multiple lasers simultaneously (each illuminating different far-field locations), combined with the discretized DOPA architecture, represents a promising approach to system scalability. We analyze the trade-offs and constraints arising in our system, due to the coupling of the parameters of the scanning engine and the ranging engine. In addition, we model the imperfections along the path of the optical signal, starting from its generation by the tunable laser up to its detection on the coherent receiver. The model is then used to analyze the impact of the system parameters on the LiDAR performance metrics.

In this paper characteristic of the microstrip patch antenna for X-band and Ku-band has been validated while deploying photonic –assisted excitation to demonstrate efficient transmission and distribution of the RF modulated optical carrier. To demonstrate the microstrip patch antenna behaviors under photonic-assisted excitation a four element antenna array excited by corporate feed network is design and fabricated for efficient optical beam steering application. In this paper antenna performance characteristics are analyzed in the presence of RF and photonic excitation. It has been found that the performance of the antenna through photonic-assisted feeding is much superior to the RF feeding of the antenna. The proposed antenna has > 100 MHz of bandwidth corresponding to 9.0, 13.0, and 18.0 GHz resonant frequency respectively. The max peak gain of the proposed antenna is 14.6 dBi.

Integrated optical phased arrays (OPAs), fabricated in advanced silicon-photonics platforms, enable manipulation and dynamic control of free-space light in a compact form factor, at low costs, and in a non-mechanical way. In this talk, I will highlight our work on developing OPA-based platforms, devices, and systems that enable chip-based solutions to high-impact problems in areas including augmented-reality displays, LiDAR sensing for autonomous vehicles, optical trapping for biophotonics, 3D printing, and trapped-ion quantum engineering.

A 1550 nm LiDAR receiver is demonstrated that is capable of collecting 3D image data at point rates exceeding 80M points/second enabling a new level of LiDAR data density for sensor fusion and convolutional neural network driven applications. The receiver uses a 512-element InGaAs avalanche photodiode (APD) detector array die that is flip-chip bonded to a Si readout integrated circuit (ROIC). The ROIC comprises 512 differential TIA receiver channels operating in parallel, each with integrated discrimination and timing circuitry. The resulting receiver system has a timing resolution of 250 ps and is capable of ranging up to 600 m. Noise equivalent input energy for this receiver is 90 photons and ranging return sensitivities near this level are possible due to the high LiDAR data density which enables the use of spatial and temporal noise filtering techniques.

Reconfigurable photonics enable the realization of diverse functionalities using a single photonic device. Such devices could play a prominent role in a large variety of applications, including neuromorphic computing, telecommunications, data communication networks, or optical sensing. The silicon photonics platform is the ideal candidate to implement those devices due to its unique capability for handling large scalability and mass-manufacturing at low cost. However, switching and modulation functionalities offered by current silicon platforms are based on the plasma dispersion effect or the thermo-optic effect, which yields devices with large footprints or reduced bandwidth, preventing scalability. Therefore, combining silicon photonics with materials with unique optoelectronic properties is emerging as the most promising path towards developing ultra-compact devices with competitive performance. In this context, new functionalities not yet offered by current silicon platforms may be implemented by the integration of phase change materials (PCMs) and transparent conducting oxides (TCOs) in silicon structures. Hybrid PCM/Si and TCO/Si devices will be presented for implementing weighting operation, reconfigurable activation functions, and optical storage, which are crucial functionalities in artificial neural networks and the emerging field of neuromorphic computing.

In this report, we will compare advanced sub-1nm CMOS technologies for photonic, millimeter wave, and quantum processors. Ultra-low series resistance threshold-less LED or laser, tunnel microwave diode, and photon sensors can be integrated in CMOS as one integral transistor. These devices form a positive feedback look for higher output current and switching speed. CMOS LED or laser can respond to DC signals. For millimeter wave diodes, an additional AC source needs to be connected to the CMOS in order to produce microwaves. Since both laser and microwave diodes are in the same CMOS drain region, generation of microwaves can be amplified by the photons from the laser, especially when avalanche breakdown millimeter wave diodes (IMPATT, BARITT) are fabricated in parallel with the photon emitting diodes. Millimeter wave CMOS output power (dBm) vs. materials (InP, GaAs, Si, SiGe) and input frequencies will be presented. CMOS switching speeds and output currents can be substantially improved using special layout design integrating both laser and microwave devices in the same CMOS transistor. S parameter analysis will be discussed.