This PDF file contains the front matter associated with SPIE Proceedings Volume 11692, including the Title Page, Copyright Information, and Table of Contents.

We report on the design and performance of a silicon photonic micro-transceiver required to operate in high ambient temperature environments above 105°C. The four channel “I/O core” micro-transceiver incorporates a 1310 nm Quantum Dot laser system and operates at a data-rate of 32 Gbps per lane. The 5mm x 5mm micro-transceiver chip benefits from a multimode coupling interface for low-cost assembly and robust connectivity at high temperatures and an optical redundancy circuit, increasing reliability by over an order of magnitude. I/O core is a photonic building block used to construct more complex application-customized modules such as 512 Gb/s modules for HPC, 5G and AI systems.

We have witnessed a tremendous evolution in the switch industry over the past decade, which has led to a 40x increase in terms of switch Input/Output (I/O) bandwidth (BW). However, as we start reaching the physical limits of Ball/Land Grid Arrays (BGA/LGA), continued BW scaling is getting more and more challenging. A promising solution to overcome BW density and thermal cooling limits is the integration of optics onto the 1st-level package, a.k.a., copackaged optics (CO). The increased escape BW offered by CO can enable high-radix switch implementations of >150 ports, which can be combined with high data rates of ≥400 Gb/s per port. From the network design perspective, there are two key benefits of using CO: (a) the ability to build large-scale fat-tree topologies of >11,000 end points with only two switch layers, and (b) the ability to provide 4x higher bisection BW, reducing at the same time the number of required switch ASICs by an order of magnitude. CO can enable both reduced energy consumption and packet delays since fewer hops are required, i.e., packets traverse fewer SerDes lanes and visit fewer buffers, which reduces network contention and improves the tolerance to network congestion. Simulation results for synthetic traffic patterns with hotspots suggest that CO can enable linear BW scaling and can significantly reduce the mean packet delay and its standard deviation, with improvements reaching up to 71% and 79% for high-load conditions, respectively.

Current high-speed optical interconnects are composed of cost-effective direct-detection (DD) systems. However, the continuous and exponential growth of data traffic is pushing new technologies to be adopted for optical interconnections. One promising technological candidate is the coherent optics. This technology facilitates the use of multi-level modulation format and narrow channel spacing. Despite the supreme performance of coherent optics technology, this technology is still regarded as too expensive to be used for cost-sensitive and very-short-reach applications. Thus, it is anticipated that DD systems would still dominate the market in the near future. Optical single sideband (SSB) modulation is an effective way to double the spectral efficiency of conventional double sideband-based DD systems. The narrow spectral width of the signal can be used to increase the number of wavelength-division-multiplexed (WDM) channels or to relax the requirements of WDM filters. This signal can be readily generated by using an in-phase/quadrature modulator (IQM) driven by the Hilbert transform pair. However, the IQM has a high insertion loss, requires a couple of automatic bias controllers for its stable operation, and occupies a large footprint. In this paper, we present our recent research activities on the dual modulation of directly modulated laser and electro-absorption modulator for the generation of optical SSB signals. This scheme can be implemented into a monolithically integrated semiconductor device in a highly cost-effective manner, just like electro-absorption modulated lasers. We present the modulation conditions of this dual modulation scheme for high-speed optical SSB signals and experimental results.

Silicon photonic switches are recognized as a key element in the applications of telecommunication networks, data center and high-throughput computing due to the low losses, low power consumption, large bandwidth and high integrated density. In this paper our recent works on silicon photonic switches for reconfigurable photonic integrated devices and circuits used in wavelength-division-multiplexing (WDM), mode-division-multiplexing (MDM), as well as hybrid WDM-MDM systems. First, high-performance Mach-Zehnder switches with an ultra-broad bandwidth, polarization-insensitivity, and low phase errors are reviewed. Second, wavelength-selective photonic switches based on MRRs are discussed. Finally, the progresses of multi-channel reconfigurable optical add-drop multiplexers are reviewed.

We review our recent activities on developing surface-normal high-speed modulators with electro-optical polymer embedded inside plasmonic and silicon metasurface structures. By judiciously designing the grating structures, we obtain sharp optical resonances, which can be used to achieve efficient modulation of the reflected and transmitted light.

Hybrid integration of semiconductor optical amplifiers with frequency-selective feedback circuits, implemented using low-loss Si3N4 waveguides, enables robust chip-sized lasers with outstanding properties. Deploying ring resonators as a tunable feedback filter provides single-mode operation over a wide wavelength range. Moreover, these rings resonantly enhance the cavity length, which results in ultra-narrow intrinsic linewidth, as low as 40 Hz. Here, we present an overview on state-of-the-art developments regarding these lasers. We compare linewidth and tuning results for different feedback circuit configurations. Finally, we report on the first demonstration of a hybrid-integrated semiconductor laser that operates in the visible wavelength range.

3D sensing is being widely adopted in consumer, industrial and automotive markets. As an illumination source, VCSELs provide a combination of high efficiency, miniaturized packaging, fast pulse rise times and minimal spectral shift with temperature. This paper will describe advances in VCSEL technology that address the requirements of 3D sensing and LiDAR applications. 3D Sensing based on structured light requires power efficient VCSELs with a narrow beam divergence, compatible with the optics that produce a spot pattern in the far field. Time of Flight or LiDAR also requires high power efficiency, as well as fast rise times for good resolution in the 3rd dimension. For consumer applications, compactness of the illumination module is important, while all versions of 3D sensing benefit from the VCSEL’s narrow spectrum and low spectral shift with temperature. In this paper we will describe recent advances in VCSEL technology that enable improvements in 3D sensing systems. This includes efficiency improvements (greater than 60% power conversion efficiency), multi-junction VCSEL designs (up to 5 junctions in a device), and flip-chip back-side emitting VCSELs that enable miniaturization of illumination modules and large-scale addressability. In addition, we will describe module level integration of illumination sources, particularly for Time of Flight (TOF) and LiDAR applications, that incorporate VCSEL, driver, monitor diode, eye safety measures and optics.

In this paper, we perform a comparison of three modulation formats: NRZ, duo binary and DMT in combination with the state of the art 850 and 910nm VCSELs for their application in short reach high speed optical links. The system for NRZ and DB utilizes feasible for deployment equalization including a 9-tap finite impulse response filter and raised cosine filtering in the transmitter and a 7-tap UI-spaced feed forward equalizer in the receiver. The 100 Gbit/s net link with DB modulation can be realized without applying receiver equalization. For DMT the highest gross data rate of 224 Gbit/s/lambda is achieved.

Over the past decade, in the near-infrared spectral region, optical phased array (OPA) technology has advanced from basic concepts to demonstrations in efficient, high-resolution, wide-angle systems. Extension to the mid-infrared spectral region has only recently begun. In this work, at λ = 4.6 μm, on an InGaAs/InP platform, we demonstrate the operation of a 32-channel OPA beam steering device. Through thermo-optic tuning we steer the beam laterally within a ±11.5° field-of-view.

Commercial telecommunications and internet data centers have made integrated photonic transceivers commodity products. We discuss the evolution of optical transceiver technology from direct detection to culmination in digital and analog coherent optic products for both fiber optic and free-space optical communication. Could this be a path for the convergent evolution of optical logic? We discuss the history of optical and electronic logic devices. We review recent work on coherent all-optical logic. We discuss new approaches using all-dielectric meta-surface structures in silicon photonics.Commercial telecommunications and internet data centers have made integrated photonic transceivers commodity products. We discuss the evolution of optical transceiver technology from direct detection to culmination in digital and analog coherent optic products for both fiber optic and free-space optical communication. Could this be a path for the convergent evolution of optical logic? We discuss the history of optical and electronic logic devices. We review recent work on coherent all-optical logic. We discuss new approaches using all-dielectric meta-surface structures in silicon photonics.

Silicon photonics optical phased arrays have been studied intensively in recent years. As the divergence angle of the light beam is inversely proportional to the size of optical antennas, developing long grating emitters is crucial to the implementation of large aperture optical phased arrays. However, because of the high refractive index contrast between silicon and cladding materials on silicon-on-insulator (SOI) platform and fabrication limitation, the grating strength of a conventional grating is so strong that the light can only propagate a short distance within the grating. Because of the capability to engineer the macro optical properties of materials, subwavelength structures have become important building blocks in integrated photonics. In this paper, we propose subwavelength silicon segments as a promising approach to form long grating emitters. Subwavelength segments are placed a distanced away from a conventional waveguide to assure that they only interact with the evanescent wave of the guided mode. The grating strength can be tailored to any values of interest by optimizing the dimensions and positions of subwavelength segments. As a proof-ofconcept, a millimeter-long, through-etched grating and an apodized grating are designed and fabricated, which shows a divergence angle of 0.081 ° and 0.079°.

The direct optical wiring (DOW) technology which is based on the 3D writing of optically transparent polymer wires using a meniscus-guided method is employed to develop multi-mode optical interconnects for VCSEL-MMF applications. This DOW method, similar in concept to conventional electrical wire bonding, does not require any chemical reactions. DOW bonding is used to optically connect high-speed VCSELs and standard OM3 multi-mode fibers (MMFs). The resulting simplified lens-free transmitter modules show a high coupling efficiency (65%) and 12 Gbit/sec error-free transmission per-channel, which is satisfactory for optical HDMI 2.1 applications. In the presentation, the 4K-60Hz performance transmitted through the HDMI2.1 will be demonstrated and extended single-mode results introduced. We believe that DOW bonding can be a powerful platform technology for optical interconnect and photonic integration for innovative industrial and academic photonic applications.

Co-packaged optics for next-generation data center switches require novel photonic packaging and optical interconnect solutions to increase bandwidth and decrease manufacturing costs. An optoelectronic glass substrate with integrated ion-exchanged (IOX) single-mode waveguides for photonic integrated circuit (PIC) packaging and fiber cable connectivity is demonstrated in an effort to reduce the overall packaging complexity. The single-mode glass waveguides were fabricated and evaluated to be thermally stable at 110ºC for more than 5 years. Laser singulated optical end-facets and laser-formed passive alignment features yield an average connector loss of 0.68 dB when end-coupled to standard MTP- 16™ connectors.

Polymer optical interconnects have been under development for decades and have found use in multimode data center applications, but to date have seen limited use as single-mode optical interconnects, where the preference has been to use optical fiber. However, with the emergence of high volume, high port count silicon photonics, the cost and packaging limitations of single-mode fiber are increasingly evident. We will discuss several innovations we have developed to facilitate optical interconnection between silicon photonics chips, either directly or via an underlying optical printed circuit board, including misalignment tolerant inverse tapers and dry-film direct write refractive index contrast polymers.

We present our latest results on the design and fabrication of mode-field conversion tapers for low-loss optical interconnects. These structures are fabricated by means of two-photon polymerization-based 3D nanoprinting. We experimentally demonstrate that our 3D nanoprinted downtapers outperform conventional lensed fibers for low-loss edge coupling of single-mode fibers with SOI, Si₃N₄ and InP-based photonic integrated circuits. They are also more robust as they allow butt coupling rather than free-space coupling. Non-linear taper profiles allow shortening the length of the downtapers while keeping their performance. We also demonstrate 3D nanoprinted uptapers that allow for relaxation of the lateral misalignment tolerances.

We present the current challenges for high frequency interconnects, especially for calibrated measures of the frequency response of components operating above 100 GHz. This is the challenge addressed by the TERAmeasure Future and Emerging Technologies project, aiming to combine photonics and electronics to develop new paradigm in the millimetre and Terahertz frequency ranges, overcoming the current obstacles to better measurements, eliminating the frequency banded nature of rectangular waveguides and providing metrology-grade results across the full frequency range.

Photonics integration continues to be a main driver for innovation in multiple aspects, including wafer-scale integration, new materials, sub-micron alignment of components and protection from harsh environment. We show cost-effective fabrication technologies of micro-optical components by UV wafer-scale replication into chemically stable polymers. Furthermore, for simplified fiber coupling and packaging, a novel 90° optical interconnect is presented, integrated with self-alignment structures. Replicated, space compliant microlenses on packaged CMOS imagers show improved light sensitivity by a factor 1.8. A laser based, low stress bonding process is explored to generate wafer-scale hermetic enclosures for harsh environment applications ranging from space to implants.

New technological developments such as IoT, Artificial Intelligence and Blockchain are leading our “Data-driven society”, where data generated by physical devices are shared across multiple platforms to improve everyday services. This unique technological evolution also corresponds to new cybersecurity and data-protection risks and challenges as well as computational limitations, which could ultimately impact users’ experience, safety and privacy. The development of products and infrastructure offering long-term security guarantees and stronger computational capabilities is a global priority. Near-term, quantum technologies provide a radically new toolset to realize stronger encryption systems as well as improved randomized algorithms. One such technology is the development of a reliable high-speed and scalable quantum random number generator. In this article, a System in Package (SiP) integration and packaging process is analyzed to bring this component into the low cost and high-volume arena. The proposed SiP solution, combined with existing surface mount technology, offers numerous benefits, such as scalability, smaller physical size, less parasitic effects and lower cost. We will discuss component performance, with insights on interconnect lengths, the shielding effect and the impact of the encapsulants.

III-V materials with quantum wells or quantum dot active regions have proven to be relatively efficient devices for amplifying light. However, integration and scaling of many other functions are moving towards the development of ever more complex photonic integrated circuits (PICs). Assembling these devices into hybrid/heterogeneous PICs poses a challenge in terms of bandwidth and footprint. In this work, we propose a Particle Swarm Optimized methodology to generate non-intuitive structures that couple light vertically from a III-V platform to a silicon-on-insulator chip. By designing heuristically optimized III-V and silicon tapers, we can overcome the limitations of typical linearly-varying spot-size converters in terms of footprint, without sacrificing bandwidth. Furthermore, the optimization parameters are adjusted to fit the usual design rule constraints that are ready for mass production, namely UV-lithography limits.

All envisaged practical implementations of cryogenic processors, including quantum computers and classical processors based on single flux quantum (SFQ) signals, require massive data transfer from and to classical high performance computers (HPCs). Cryogenic computing has recently become a very hot topic, including superconducting quantum computers (QCs), and classical processors based on single flux quantum (SFQ) signals. All envisaged practical implementations of cryogenic processors require massive data transfer from and to classical HPCs. The project aCryComm aims to develop building blocks for cryogenic photonics interconnects and eventually enable this challenging data transfer. The long-term goal is the development of an open-access platform to integrate classical optical interfaces based on low-loss silicon photonics, plasmonics, and nano light sources together with superconducting photonic and electronic devices, including SFQ-based co-processors for HPCs and for QCs.

Commercial silicon photonic (SiP) biosensor architectures rely on expensive swept-tunable lasers that limit their use for widespread, point-of-care applications. An alternative is the use of fixed wavelength lasers integrated directly on a silicon photonic platform. This study investigates the design considerations of such architectures.

The National Aeronautics and Space Administration (NASA) is continuously working in furthering its space and aero communications capabilities required for the successful accomplishment of its aerospace missions. With the ever present demand for higher communication data rates and larger bandwidth required by future space exploration missions, optimization of the communications systems supporting such missions is necessary to ensure that critical scientific data and high definition video and imagery of human and robotic exploration is properly transmitted back to Earth. In the aeronautics side, the envisioned increase in aircraft volumes under the Urban Air Mobility (UAM) and Advanced Air Mobility (AAM) ecosystems could benefit from communications capabilities impervious to interference and free from spectrum limitations constraints. This work discusses examples of GRC’s ongoing technology development and integration efforts relevant to the aforementioned scenarios. In particular, the ability to increase the versatility, affordability, and reliability of ground-based optical receivers for space-to-ground communications will be presented. Our current activities on the development of highly secure airborne laser communications links augmented with Quantum Key Distribution (QKD) will also be discussed, as well as scenarios in which optical communications could be beneficial to UAM/AAM. The status of efforts in quantum communications, high rate optical networks, and on the development of current efforts to demonstrate integrated Radio/optical communications (iROC) will also be addressed.

With an ongoing trend in computing hardware toward increased heterogeneity, domain-specific coprocessors are emerging as alternatives to centralized paradigms. Here I will introduce and discuss key concepts on PIC-based photonic paradigms for machine intelligence. These include parallelization strategies, real-time computing (~ps), non-iterative one-shot execution O(1) run-time complexities, synergistic convolutions, photonic nonvolatile (multistate) memory. These device, circuit, and system design paradigms allow realizing PIC-based special-purpose processors in this ongoing trend towards increased heterogeneity. As an example, I will share a photonic tensor core processor and a massively-parallel convolutional neural network processor.

Photonic switches promise to have a large impact on future HPC and datacenter networks. However, even if we assume ideal zero-energy photonic switches, we cannot achieve significant system-wide benefits if we do not change the rest of the system architecture. This motivates co-design between emerging photonic switches and the rest of the system in order to adapt the system network to best make use of unique features of photonic switches, as well as tailor photonic switches to better support system-wide trends such as resource disaggregation. In this paper, we discuss the architectural impact of several properties of photonic switches. For each, we provide an overview of what system-level capabilities they enable, how they can be adapted to support ongoing trends, and what other synergistic advancements would produce a better system-wide improvement. In this way, we illustrate the potential benefit of closer collaboration between the photonic and architecture communities.

In this paper, we propose the use of two Passive Optical Network (PON)-based network architectures to connect free-space Optical Wireless Communication (OWC) Access Points (APs) within a room with multiple users. We optimize through a Mixed Integer Linear Programming (MILP) model the assignment of mobile OWC users to more than one AP to improve the resilience of the fronthaul network, i.e the OWC system and the wired network linked to APs, and study the impact of users distribution and channel characteristics.

We demonstrate a novel probabilistic Brillouin frequency shift (BFS) estimation framework for both Brillouin gain and phase spectrums of vector Brillouin optical time-domain analysis (BOTDA). The BFS profile is retrieved along the fiber distance by processing the measured gain and phase spectrums using a probabilistic deep neural network (PDNN). The PDNN enables the prediction of the BFS along with its confidence intervals. We compare the predictions obtained from the proposed PDNN with the conventional curve fitting and evaluate the BFS uncertainty and data processing time for both methods. The Brillouin phase spectrum generally provides a better measurement accuracy with reduced measurement time in comparison to the Brillouin gain spectrum-based measurement, for an equal signal-to-noise ratio and linewidth. The proposed method is demonstrated using a 25 km sensing fiber with 1 m spatial resolution. The PDNN based signal processing of the vector BOTDA system provides a pathway to enhance the BOTDA system performance.

The continued acceleration of switching capacity and link transmission bandwidth is driving the need for new connectors in next generation optical networks. With 25.6Tb switch ASICs available in 2020 [1], only two years after the introduction of 12.8Tb switching, the industry is now looking to radical new architectures to achieve 51.2Tb switching, including the advent of optics integrated or co-packaged with ASIC technology. The standard 250μm pitch used in multi-fiber ferrules with 125μm cladding diameter optical fiber is physically too large to support the quantity of optical lanes that will be coupled inside the coming switching platforms. This paper describes a next generation MT-style ferrule designed for fibers with 80μm cladding diameter on a pitch of 165μm. By decreasing the pitch from 250μm to 165μm, up to 24 fibers can be placed in a single row between the pins of the MT-16 alignment structure. This tighter pitch enables higher fiber densities coupled directly to or in the proximity optical Tx/Rx photonic tiles. Endface geometry models combined with connector mating normal forces are based on the traditional 250μm pitch of MT-style ferrules. Fiber tip radii, fiber tip coplanarity, ferrule surface endface angles relative to alignment pin bores were measured empirically and documented on the new design. Varied topologies were combined with different mating forces demonstrating effective physical contact for the new ferrule. Mated pairs were monitored for attenuation changes during exposure to industry standard uncontrolled environment temperature cycling. Subsequent specifications for future end face geometry and mating spring force requirements are proposed.

We are investigating to utilize multi-core fiber for increasing the concentration of optical fiber and the density of optical connector. Until now, from the viewpoint of optical mounting, the bending loss and the crosstalk of the multi-core fiber in a short distance range have not been investigated so much. In this report, we propose low-loss and easily bendable multi-core fiber with a novel double core taper structure to realize highly efficient optical coupling to a silicon photonics integrated circuit, and compact fiber mounting with small bending radius. By BPM simulation, we confirmed that it is possible to suppress the bending loss of less than 0.5 dB and the inter-channel crosstalk to -20 dB or less when the radius of curvature is 3 cm. Based on this result, we fabricated an optical fiber with the double core taper structure in order to verify the characteristics. As a result, the coupling loss was realized within 1.5 dB with low radiation loss. It is therefore considered that the double core taper structure can be sufficiently applied to the multi-core fiber.