This PDF file contains the front matter associated with SPIE Proceedings Volume 11286, including the Title Page, Copyright Information, Table of Contents, Author and Conference Committee lists.

Large enterprise and hyperscale datacenters are presently undergoing a technology transition from 100 Gb/s to 400 Gb/s machine-to-machine connectivity. Soon, 800 Gb/s and 1.6 Tb/s interconnects will be required. The state-of-the-art and future technical directions for optical interconnects in datacenters are reviewed. Particular attention is given to cases where copper and fiber connections are technically and economically competitive or soon will be – in fiber-to-the-server and disaggregated chassis Ethernet switch applications. Further attention is given to the impact of silicon photonics and the co-packaging of optics with Ethernet switching ASICs.

This paper will introduce the evolution of optical interconnect since the adoption of optical fiber as the primary medium for long distance communication back in the 1970’s. The author will share the recent technological development of optical interconnect and also market trends that are driving innovation and adoption in the interconnect industry. Key industries driving optical interconnect include Data Centers & Photonic Integrated Circuits.

Communications satellites are now facing a challenge to meet the growing demand for connectivity and to handle large volume of data at very high speeds but they must be a profitable solution for operators and photonics-related technologies offer great promise. The increasing volume of information being transferred through the satellite puts more constraints on the payload electronics that has to transfer data at rates in excess of 10 Gigabit per second per lane (Gbps/lane). At those rates, electrical signal pre-compensation is done at a higher cost in terms of power consumption on standard copper-based RF electronic and switching to optical fiber medium become a preferable option. GEO and LEO satellites optical interconnect technical requirements are: mechanical ruggedness, low weight, low power consumption and radiation tolerance. If they integrate specific and advanced material and technologies, optical interconnect can meet space mission’s requirement. Constant temperature cycling, operation under vacuum and operation under radiation are just few of the environmental conditions that the optical modules must perform in. Lifetime expectancy for GEO mission has to exceed 15 years. In order to have reliable communication link over 15 years mission, the electro-optics parameters must be selected properly. The effect of the VCSELs bias on the communication performance at the beginning of life (BOL) has to be studied in order to maximize the end of life (EOL) communication link budget. We will show how those parameters have been obtained. Illustrations of hardware and description of the relevant qualification tests and results will be also be shown to explain how the challenges have been overcome.

With the aim to satisfy the scalable growth in both network traffic volume and connected endpoints while decreasing the cost and the energy consumption, transparent optical DC networks (DCNs) based on fast optical switches have been considered, featuring the data rate and format transparency and eliminating the power consuming O/E/O conversions. In this work, we propose and experimentally assess novel optical DCN architectures based on distributed and buffer-less nanoseconds WDM photonics integrated switches. The WDM photonic integrated switches are capable to switching in the wavelength, space, and nanoseconds time domain to provide full flexibility and the required speed to achieve high throughput DCN networks. Disaggregated DCN architectures enabled by the fast WDM PIC switch will be also presented.

The ever-increasing energy consumption of Data Centers (DC), along with the significant waste of resources that is observed in traditional DCs, have forced DC operators to invest in solutions that will considerably improve energy efficiency. In this context, Rack- and board-scale resource disaggregation is under heavy research, as a groundbreaking innovation that could amortize the energy and cost impact caused by the vast diversity in resource demand of emerging DC workloads. However disaggregation, by breaking apart the critical CPU-to-memory path, introduces a challenging set of requirements in the underlying network infrastructure, that has to support low-latency and high-throughput communication for a high number of nodes.

In this paper we present our recent work on optical interconnects towards enabling resource disaggregation both on Rack-level as well as on board-level. To this end, we have demonstrated the Hipoλaos architecture that can efficiently integrate Spanke-based switching with AWGR-based wavelength routing and optical feedforward buffering into highport switch layouts. The proof-of-concept Hipoλaos prototype, based on the 1024-port layout, provide latency performance of 456ns, while system level evaluations reveal sub-μs latency performance for a variety of synthetic traffic profiles. Moving towards high-capacity board-level interconnects, we present the latest achievements realized within the context of H2020-STREAMS project, where single-mode optical PCBs hosting Si-based routing modules and mid-board optics are exploited towards a massive any-to-any, buffer-less, collision-less and extremely low latency routing platform with 25.6Tb/s throughput. Finally, we combine the Hipolaos and STREAMS architectures in a dual-layer switching scheme and evaluate its performance via system-level simulations.

Scalability of silicon photonics packaging is required to support the growing demand for bandwidth in data centers and other emerging datacom and telecom applications. Connecting fiber to chip is still considered one of the main challenges of silicon photonics today due to the need for high alignment accuracy, which in turn requires expensive assembly machines and in most cases active alignment protocols. Furthermore, current fiber assembly technologies are not suitable for multiple emerging applications with large port count such as co-packaged optics. We present here the Photonic-plug technology, a unique self-aligning optical arrangement, which enables large assembly tolerance suitable for passive alignment protocols and for scalable silicon photonics port count packaging. The Photonic-plug technology accomplishes die stacking geometry with efficient wideband surface coupling as well as with grating coupler based silicon photonic chips. The combination of the Photonic-plug's large assembly tolerances and surface coupling geometry enables efficient silicon photonics wafer level testing capabilities prior to dicing. The Photonic-plug technology takes advantage from wafer level fabrication processes for planar and accurate implementing of optical elements. A library service model, called Photonic-bump, is incorporated as part of the Photonic-plug technology through silicon photonic wafer manufacturing process for complete removal of the fibers' mechanical constraints from wafer manufacturing process. The Photonic-plug takes fiber-to-chip packaging away from specialized, low throughput and expensive tools to standard, automated and high volume flip-chip packaging machines. Standardizing optical packaging through Photonic-plug methodology will affect further silicon photonics application to thrive such as photonic FPGA, optical interposers and chip-to-chip optical connectivity.

The development of compact low-power-consuming nano-optoelectronic devices is one of the great challenges to tackle for the convergence of microelectronics and photonics. During this talk, I will review our recent results on InP-on-SOI photonic crystal devices. I will show how these wavelength scale structures allow the achievement of nanolaser diodes and amplifiers which can be densely integrated in a photonic circuit.

High-speed high-power photodiodes are key components in microwave photonic applications including antenna remoting, generation of microwave signals, and analog optical links. With the rapid emergence of Si photonics, it is becoming increasing important to develop approaches to incorporate these photodiodes into a Si platform. For integration of III-V photodiodes on silicon, various integration approaches have been reported, including adhesive wafer bonding, direct molecular wafer bonding, and direct III-V material growth on Si. This paper will report the integration of InGaAs/InGaAsP/InP modified uni-traveling carrier photodiodes using each of these techniques. In addition, high-power Ge on Si photodiodes will be discussed.

We are developing a 1x8 single mode (SM) optical interface to facilitate the adoption of dense wavelength division multiplexing (DWDM) silicon photonic (SiPh) optical interconnects in exascale computing systems. A common method for fiber attachment to SiPh transceivers is ‘pigtailing’- the permanent adhesive bonding of fiber/v-groove arrays to onchip grating couplers (GC). This approach precludes standard high throughput surface mounting and solder reflow assembly of the transceiver onto system printed circuit boards. Our approach replaces the fixed pigtail with a low profile, small form factor, detachable expanded beam optical connector which consists of four essential parts: a GC array, a surface mount glass microlens array chip, an injection molded solder reflowable optical socket, and an injection molded SM light turn ferrule. The optical socket and ferrule are supplied by US Conec Ltd. To design the GC, we developed an optical simulator that considers CMOS foundry constraints in the optimization process. On-wafer measurements of the GC coupling loss to SMF28 fiber at 1310nm is ~1.4dB with a 1dB bandwidth of ~22nm. This ensures a wide low loss spectral window for at least 16 DWDM channels. The geometry of the optical system is arranged so that only a simple spherical lens is required for efficient mode matching in the expanded beam space. The fiber to fiber insertion loss through the light turn ferrule, two microlenses and GCs, and a looped back SOI waveguide ranged from 4.1-6.3dB, with insertion loss repeatability of 0.2dB after multiple mating cycles.

Silicon (Si) photonics is well-positioned to provide high-speed and low-cost optical interconnects. The extraction of data from cryogenically cooled integrated circuits (ICs) has become of great interest for low-power data readout. Utilizing wavelength division multiplexing (WDM), a high capacity optical interconnect can be realized using remoted Si photonic based ring resonator modulators (RRMs). Results include operation up to 20 Gbps and BER < 1E-12 using a 2 Vpp signal, consuming < 100 fJ/bit in the cold environment. Lastly, Si photonic RRM device and interconnect optimizations for operation at temperatures ≤ 77 K will be presented.

Optical interconnects in data centers have traditionally used 850 nm GaAs-based vertical-cavity surface-emitting lasers (VCSELs) in combination with multimode fiber, having a reach up to 100 m in length. Longer links typically use standard single-mode fiber in conjunction with either InP-based edge-emitting lasers or silicon photonic transmitters operating in the 1310 nm or 1550 nm window. Single-mode GaAs-based VCSELs operating at 1064 nm offer another path for achieving longer system reach. Potential advantages of these VCSELs include better power efficiency, modulation speeds reaching 50 Gbps and large-scale fabrication volumes. The longer wavelength is also beneficial due to the lower attenuation and chromatic dispersion of optical fibers at that wavelength. However, one practical issue for single-mode transmission is that the G.657 standard for single-mode fiber requires that the 22-meter cable cutoff wavelength be less than 1260 nm, and these fibers are typically few-moded at 1064 nm. The large differences between the group velocities of the LP01 and LP11 modes can lead to degradation of the system performance due to multi-path interference if the higher order modes are present. To resolve this quandary, we have designed and validated the performance of a new optical fiber which is single-moded at wavelengths less than 1064 nm, but also has G.657- compliant mode field diameter and dispersion characteristics that enable it to be used in the 1310 nm window.

In high performance computer systems and large-scale data centers, data movement becomes a critical problem. To overcome this problem, co-packaged optics are attracted much attention. For high performance large-scale integration (LSI) packages with the co-packaged optics, we have studied an optical and electrical hybrid package substrate. The hybrid package substrate is an organic package substrate with single-mode polymer waveguides. For the hybrid package substrate, we proposed a low-cost MT-ferrule-compatible optical connector and demonstrated its sub-micron passive assembling to connect the polymer waveguides and a single-mode fiber array. In this work, the assembled connector was glued to the substrate with a UV curing adhesive, and then, passive optical coupling between the polymer waveguides and the MT fiber connector was demonstrated. The 12 polymer waveguides including 4 dummy patterns were buttcoupled with the standard MT fiber connector and optical coupling for all polymer waveguides was achieved. The coupling loss penalty compared with active alignment was ~3 dB.

Various fiber-optic connectors have been developed during the 40 years since optical fiber communications systems were first put into practical use. During the first two decades, as the domain of optical fiber communication systems expanded from trunk lines to subscriber lines and customer premises, the main focus changed from performance improvement to miniaturization and cost reduction. After that, new interconnect technologies were continuously developed to meet new requirements such as high-power transmission, high-density on-board wiring, and multicore fiber interconnection. Despite many improvements, the principle and basic structure of the optical connector have hardly changed. Zirconia ferrules and split sleeves are used for single fiber coupling, MT ferrules are used for multifiber connectors, and both types use physical contact technology. This paper describes the key technologies for fiber-optic connectors, a brief history of optical connector improvements and recent technical issues related to fiber-optic interconnect technologies.

Polarization maintaining fibers arrays are key enablers to process high bandwidth data, representing a powerful part within the photonic integrated chip technology. The different channels increase the information density and allow to multiple singles through one fiber bulk at the same time. Due to fiber’s small dimensions (ø125 μm) they can be integrated in existing infrastructure easily and are very flexible at the same time. However, the compact design together with the flexible material properties demands for new precise tools and technologies to reach the necessary precision during packing.

The Fraunhofer-Institute for Production Technology IPT develops, together with their partners Phix and Aixemtec, new handling and assembly tools, as well as processes as one of the leading companies in this field. In the self-developed assembly cell, the fiber handling tool-head operations automatically to pick up, manipulate and tack single fibers to a glass plate or fiber to chip. Each fiber is moved by a portal robot within the assembly cell with micrometer accuracy but also can be rotated with a repetition accuracy less than 0.01°. Advanced illumination units observation techniques allow to package fibers arrays much quicker and more robust than before. Therefore, additional camera systems and material characteristics are used to develop smart alignment routines. As a result, the observation of the orientation of the PM-fiber core as well as the fiber layout during the assembly process leads to high quality products within fast production cycles. Due to the flexible construction of the assembly call also PIC packaging and fiber-to-chip coupling is possible.

The growing deployment of optical fiber in industrial, avionic, military and transportation applications, is leading to the increasing use of optical connectors requiring low insertion loss and mating workability. Connector contaminants are the main cause of power loss and signal degradation and the continuous need of cleaning services is a disadvantage for the fast replacement of copper wiring. Therefore, there is an increased need for reliable and maintenance-free optical connectors with insensitivity to dust, especially in a harsh environment. For this purpose, a novel and unique modular, single-mode lensed ferrule has been developed. Its compact form-factor fits into standard connectors bodies (E2000, MPO …) and enables the insensitivity to dust and mating-cycles reliability typical of bulky harsh environment lensed connectors. It is known that the alignment accuracy between a fiber and a lens must be under 1 um in order to attain acceptable insertion losses, which is particularly difficult to obtain. The use of exceedingly precise components and novel assembly techniques has made it possible, keeping ultra-low mechanical tolerances in the processing of ceramics and opto-mechanical parts. The result is highly repeatability and low insertion loss compared to standard state-of-the-art harsh-environment lensed connectors found in the field today. This opens new perspectives in the context of optical connections, where mating-cycles, insensitivity to dust, reliability and low losses ease the transition from copper to fiber. Furthermore, it enables new solutions for industrial data transmission and laser technology for many different applications in the field of spectroscopy, lidars, rotatory joints and much more.

Juniper’s silicon photonics technology employs an entirely new approach to optics manufacturing that leverages design, wafer manufacturing, packaging and test infrastructure and methodologies from the microelectronics ecosystem, resulting in a fabless optical transceiver manufacturing flow cable of scaling, in both cost and performance, with the needs of the networking industry. Juniper’s fully integrated “Opto-ASIC” transceiver leverages densely packaged electronic and photonic die in a single, low-cost package that supports common module form factors as well as on-board and on-package optics applications. The ability to incorporate all optical components, including lasers, within a single, common silicon die is a key enabler of the approach, fundamentally changing and simplifying how an optical transceiver can be assembled and tested. We will review lessons learned and provide a preview of the road ahead.

The polymer waveguide taper structures are designed for interfacing 3 μm silicon-on-insulator (SOI) devices to standard single mode SM fiber. The structure consists of polymer waveguide with its optical facets compatible with the modes of single-mode (SM) fiber and silicon waveguide. Simulations suggest that the coupling efficiency of -0.63 dB from the SM fiber to 3 μm strip waveguide can be achieved. We show experimentally the transmission loss of 0.6 dB/cm in polymer waveguide and coupling loss of 0.7 dB from the SM fiber to 2.1 μm polymer waveguide.

Recently, with the increasing traffic in data centers (DCs) and the evolution of high-performance computers (HPCs), large-capacity data transmission has been required in DC and HPC networks. To take advantage of optical interconnects with the capability of transmitting high-speed signals under high wiring density, we have proposed active optical modules for inter-rack interconnects in DC and HPC networks. In order to satisfy the higher bandwidth demand, not only the inter-rack interconnects but also on-board interconnects are required to have a high density. In this paper, we propose a high-density optical module utilizing graded-index (GI) polymer waveguide for on-board optical interconnects.

Multimode polymer waveguides have attracted considerable interest for use in high-speed on-board communication links as they provide low loss (<0.04 dB/cm at 850 nm) and high bandwidth (>40 GHz×m) and can be cost-effectively integrated onto standard PCBs. The fabrication of such waveguides on flexible substrates can provide additional advantages: shape flexibility, lightweight and reduced thickness which are particularly important in the aviation and automotive industries. Such flexible and lightweight optical connections will play an important role in next-generation airplanes and driverless cars connecting the multitude of peripheral sensors with the central processing unit at high speed and low latency. However, in such applications, flexible polymer waveguides are required to be bent to meet their stringent space requirements and twisted or stretched when connecting movable parts. Under sharp flexure, the bending or twisting loss dominates the waveguide loss limiting their practical use. In this work therefore, we present a new waveguide design for flexible polymer waveguides with improved bending performance and derive useful layout rules for minimizing twisting losses in such samples. The proposed waveguide structure only requires one additional fabrication step and achieves bending losses below 0.5 dB for a 3 mm bend. In comparison, the conventional waveguide design yields a 2 dB loss under the same bending radius and launch condition. Additionally, useful equations relating the maximum allowed number of twisting turns for low excess loss with sample thickness and width are proposed. Bending and twisting measurements on flexible waveguide samples are presented validating these methods and demonstrating the potential of this technology.

We present a new technology which enables the local resolution production of polymer-based micro-optical components, e.g. in photonic 3D packaging systems. The main advantage of this innovative technology is the capability to dispense liquids of a wide viscosity range (200 – 10.000 mPa*s) on a picoliter scale enabling the use of solvent-free liquid polymers. The newly engineered picoliter dispensing system features the possibility to place liquids on substrates with high positioning accuracy and high flexibility of volume variation and shape. The placement of active and passive micro-optical components for photonic packaging plays an important role in improving optical interfaces, especially in data and telecom applications and optical sensor technology. In this work the capability of this picoliter dispensing system is exemplarily demonstrated on single detached microlenses as well as microlens arrays (MLA) using solvent-free, viscous UV curable hybrid polymers OrmoComp® and OrmoClear®FX. The preliminary results of optical characterisations of the fabricated components verify the advantage of this novel technology over competitive manufacturing methods such as inkjet printing in terms of printability of solvent-free polymers since a comparable optical performance can be obtained while saving solvent evaporation steps.

The rapid growth in data center traffic is driving the need for increased performance and overall bandwidth of networking equipment, including optical interfaces and Ethernet switches, which are based on pluggable transceivers today. But looking just a few years ahead, bandwidth scalability challenges are looming in terms of density, cost, and power; challenges that require tighter integration of optics and networking silicon. This paper reviews the motivation for integration, the enabling technology elements, and discusses how advanced packaging and Silicon Photonics enable higher density, reduced power per bit, and ultimately the continued scalability of network bandwidth and performance.

Integrated photonic systems become more difficult and complex nowadays. That imposes more complex solutions for interconnects, in particularly for dense interconnects framework. Motivated by the potential applications for dense interconnects, we presented a compact solution for dense on-chip interconnection utilizing grating-to-grating coupling technology. In this work, we designed, fabricated and performed measurement of compact chip-based fan-in/fan-out structures fabricated on a silicon chip for multicore fibers (MCFs) applications. We demonstrated a miniaturized chip-based component that can be applicable as an interposer for coupling light from 7-core MCF to individual channels on chip or as a chip-to-chip link between photonic integrated circuits (PICs). A simulated coupling efficiency of this interconnect at chip-to-chip interface has been shown as high as -5.09 dB. Combining compact fully-etched subwavelength grating couplers and flip-chip technology, we present an approach for optical interconnection between photonic chips that can operate over broadband spectral range with reasonable loss penalties and ease of fabrication. Our future efforts will be focused on a demonstration of chip-to-chip optical interconnection with integrated optical and electrical functionalities.

We present an 1x16 silicon optical phased array (OPA) based on an electro-optic phase shifter that enables low power consumption and high speed operation. The phase shifter operates by carrier injection based on the p-i-n structure and is designed to optimize the width between doped regions to 2.5 μm. Phase tuning efficiency of 1.7 mW/π and operating speed of 20 MHz were measured from fabricated devices. The difference between the measured values and the values calculated in the simulation was analyzed through additional simulations to indicate that the error occurred in the actual fabrication. In addition, transversal beam steering was achieved in the range of 45° and the average power of 39.6 mW was consumed for phase tuning for each radiation beam.

One promising solution for the ever increasing transmission capacity demand from fundamental research and data centers is the silicon-photonic integrated WDM transmitter. We designed an easily scalable, high-bandwidth transmitter unit composed of radiation-hardened Mach-Zehnder modulators (MZMs) and Echelle grating (de-)multiplexers (EG-DMUXs).

Our 3 mm MZMs have customized slabs with a reduced etch depth to improve their radiation hardness. Our current MZMs feature a V_π∙L of 4.6 V∙cm and an insertion loss of 4.84 dB. Additionally, an error-free transmission was achieved successfully at a speed of 11.3 Gb/s while driving the modulators with a PRBS-7 signal and an amplitude of less than 2 V_pp.

The Echelle grating (de-)multiplexers were designed and simulated numerically. The presented 1x7 device is compact and low-loss: the on-chip footprint is 680 μm × 380 μm, the channel spacing is 800 GHz, and the measured average insertion loss and crosstalk are 2.5 dB and -22 dB, respectively. With optimized components, higher bandwidth systems with more channels are achievable.

The robustness of optical transceivers is a key consideration for new designs and implementations. Since the majority of transceivers are designed for data center type applications, Telcordia GR-468¹ has been the reference standard that dictates the testing demonstrates both design and process robustness. However, when one considers the industrial, military and aeronautical markets, this qualification is no longer enough to ensure adequate performance.

Instead, the test definitions are covered by pass-down requirements of the original project that are conveyed up the supply chain meaning that the transceiver vendor has little opportunity to propose alternatives. Different applications are covered by different standards with MIL-STD 810² being most widely cited for military/aeronautical (mil/aero) applications and AEC-Q102³ for automotive. Some of the more challenging tests that these standards cover include salt fog, fungal growth, temperature/humidity and shock/vibration.

These tests required additional processing, compared to the commercial world, in terms of protective, conformal coatings that ensure protection against corrosion, fungus and humidity. Such coatings are generally polymer-based compounds that are applied as a liquid and then cured to provide the protection. IPC-CC-830C⁴ defines the type and qualification requirements for different materials. Other protective coatings are available that can offer additional properties such as being hydrophobic or thixotropic.

Although protective coatings are commonly used in PCB processing and systems used for mil/aero applications, they are still considered a niche market for optical transceivers. This is because the optical die and lens surfaces need to be sealed against the coating as these lead to problems such as optical loss or reliability issues. Traditional methods for sealing, including TO-cans or hermetic packages, generally lead to larger or more expensive solutions which limit adoption. Therefore, being able to adapt an existing commercial, high volume part, that allows easy sealing and coating has many advantages.

We report on the design and performance of a compact, low power, micro-transceiver chip suitable for high volume, long life deployment in high temperature (>100°C) environments. The I/O Core micro-transceiver comprises a 5mm x 5mm silicon photonics chip with a multimode interface and incorporates 4 transmit and 4 receiver channels operating at 25 Gb/s in the O-band. We will demonstrate how the I/O Core micro-transceiver may be deployed in a wide variety of packages and applications spaces including data centers, HPCs, 5G, AI and automotive systems.

In this paper, an integrated optical receiver module with demultiplexer chip-to-chip optical interconnects is proposed. The receiver and demultiplexer (demux) were realized on a single chip in a 0.18 μm TSMC CMOS technology. The topology uses a 1:2 for the demux while the receiver (Rx) module consists of transimpedance amplifier (TIA) and a limiting amplifier (LA). The receiver achieved a gain of 90 dBΩ and a 3-dB bandwidth of 4.5 GHz. Clear eye diagrams were observed with a voltage swing of 375 mVpp up to 2.5 Gbps output from the Rx-Demux at 5 Gbps input signal. The integrated chip occupies an area of 594x1089 μm² with power consumptions of 122 mA under a 1.8 V power supply This solution could be applied to CPU/memory interface where bidirectional signaling is required.

Hybrid injection-molded ferrules are presented which consist of a polymer body and an over-molded glass insert. The average coefficient of thermal expansion observed at the front face of the ferrules is 8 ppm/C from room temperature to 100 C. This design could be applicable for direct heterogeneous re-matable connections between fiber ribbons and photonic integrated circuits which exhibit low thermal expansion and operate at elevated temperatures.

This paper presents an overview and application of thin-film-lithium-niobate (TFLN) modulators that have low drive voltages, i.e., V_π. Such modulators are critical components for realizing high-speed operation in a modern telecommunication networks, wireless communications, and RF-Photonic applications. Recent developments in crystalion- slice TFLN have enabled a new class of electro-optic modulators that have a tighter mode confinement, compact footprint, ultra-high bandwidth, and low modulating voltages. However, lithium-niobate suffers from difficult microstructuring in comparison to silicon-based materials, since it can have an etch resistance greater than many metal-based hard masks. To overcome this challenge, a hybrid material system combining the electro-optic properties of TFLN with the ultra-low propagation loss of silicon nitride has been developed. In this work, we demonstrate an integrated hybrid phase modulator, based on a silicon-nitride strip loaded waveguide on a TFLN material platform, which provides tight optical mode confinement, without the need to etch lithium-niobate. As a result, the drive electrodes can be placed closer to the optical waveguide thereby resulting in a strong RF and optical mode overlap. A 2.4 cm long phase modulator and Mach-Zehnder modulator with a demonstrated V_π of 1.5 V and 0.875 V, respectively in DC are presented along with other candidate and demonstrated devices, such as multimode interference coupler, micro-ring, and racetrack resonator.

Unlike PT and SUSY lasers, here we show an extra loss-free non-Hermitian laser engineering approach to realize single mode lasing operation for the first time. By selectively enhancing the fundamental mode’s quality factor, we obtain single mode operation with higher output power per cavity since all cavities in this system contribute to the laser output, in contrast to other non-Hermitian approaches. Furthermore, we show that this approach interestingly allows reducing the number of to-be-designed cavities in super-partner array as compared with, for example, the SUSY approach, thus leading to reduced design complexity upon coupled cavity scale up of laser arrays. In summary, the ability to engineer coupled laser systems where each laser cavity contributes to coherent light amplification opens up a new degree of laser-design freedom leading to increased device performance and simultaneous reduced design and fabrication complexity.

Parasitics such as wirebond inductances and bond pad capacitances that result from hybrid opto-electronic integration pose a challenge towards achieving data rates beyond 50 Gb/s. The effect of bond pad capacitance on the receiver transimpedance limit is shown, which demonstrates the significant advantage of monolithic versus hybrid integration. An analysis of three receiver topologies is presented. These all employ the same Cherry-Hooper voltage amplifier for the core electronics. A comparison across several design metrics of the three Transimpedance amplifier (TIA) variants is then provided. The TIAs are implemented monolithically in the IHP 250-nm SiGe BiCMOS EPIC process (fT = 190 GHz). Measurement results are then presented for 50 Gb/s OOK. PAM4 simulations are also shown.

Integration of photonic devices with CMOS circuits has recently generated much interest for numerous applications ranging from Internet networks to sensor technology. For example, integrated optical interconnects have emerged as a strong contender to replace bandwidth-hungry and energy-guzzling electrical interconnects in data centers and inside microprocessors. However, the lack of group IV semiconductor materials suitable for light emission has thus far prevented the realization of such important photonic-integrated circuits. In this paper, we will present our research works on strain-engineered group IV light sources on silicon that will help enable photonic-integrated circuits.

A wavelength filter is a key component for numerous photonic integrated circuit applications in optical communication. Researchers put forward several methods to design wavelength filters for which the Echelle grating de-multiplexers (EGDMUXs) are popular and have been extensively studied. In comparison with the traditional EG-DMUXs based on the Rowland mounting, EG-DMUXs based on the two stigmatic points (TSP) method were reported rather late. This paper will present the basic design theory and a high-performance device fabricated on a 250 nm silicon-on-insulator (SOI) platform for validation. The simulation and measurement results of this 1×7 EG-DMUX with 800 GHz channel spacing will be presented and compared. Although the fabricated device has the merits of compact on-chip footprint, low insertion loss and low crosstalk, its narrow 1-dB bandwidth may limit its application in practice. We present our solution to widen the transmission spectrum based on the TSP EG-DMUXs and multimode interferometers.

It is anticipated that once silicon switch I/Cs reach 51.2Tbps, there will be a need to migrate from electrical I/Os to optical I/Os. This drives the need to co-package optical engine transceivers with the high-performance silicon switch. In this paper, we explore the challenges, compare the various solutions, and provide guidance from an assembly and optical connector design perspective. Our focus is on single-mode optic solutions privileged for mid to long range optical links. Topics covered include optical interconnect density to the silicon photonics chip, integrated optical interconnects vs pigtail fiber ribbons, socketable vs. μBGA optical engines, laser source location and special requirements for the optical connectors. The ability to embed optics in the first level package is a disruptive capability needed to meet the everincreasing bandwidth demands for data communication. The benefits of such configurations are discussed, as well as the challenges for thermal management and system yields.

This Conference Presentation, "Photonics packaging: from pluggable transceivers to co-packaged optics," was recorded at Photonics West 2020 held in San Francisco, California, United States.

Wavelength-selective switches have been proposed for datacenter use to enhance their scalability and to help in meeting ever-increasing traffic demands. We have previously demonstrated a 4 × 4 ring-based crossbar silicon photonic switch in which each cross-point contained three ring pairs to partition the free spectral range (FSR) into three equal regions to reduce wavelength tuning range per ring pair— thereby reducing both the tuning power consumption and stress on the rings—while maintaining full routing flexibility. However, the question of scalability remains for the crossbar switch in which 96 signal pads—routed to each ring—are required to fully control it. The pad count scales by 2𝑀𝑁² for an 𝑁 × 𝑁 switch with 𝑀 ring pairs. In this paper we present a 4-port silicon photonic ring-assisted Mach-Zehnder interferometer (RAMZI) switch, fabricated in the AIM Photonics process, with multiple-sized rings per switching elements in a Beneš network configuration to reduce the number of electrical pads required to 36 signal pads. The switch is 500μm 3mm in size and is packaged on a custom PCB. In such a switch, the pad count scales by 2𝑀(𝑁 log₂ 𝑁 − 𝑁). Another advantage the RAMZI switch has over the crossbar switch is that the loss through the switch is not path-dependent due to its balanced path configuration. In the crossbar switch, the difference between the shortest and longest paths is 2(𝑁 − 1) switching elements.

One of the main challenges in the assembly of Silicon Photonics components in optoelectronic systems, is the very precise bonding of the dies and their interconnection. Not only is the necessary accuracy much higher than for other conventional optical systems, but from an industrial/business perspective, the high throughput is key. This can be achieved by a combination of techniques, using highest end passive alignment processes, whereby spread of positioning accuracy, below 1.5 µm aim for the best of two worlds: low loss coupling and high throughput. Such accuracy can be reached using highest-end dedicated equipment, whereby the alignment time can be reduced by a factor of 2 to 3 in comparison with active alignment methods.

For miniaturized optical fiber coupled MOEMS systems, fiber coupling on chip level is necessary. Therefore a silicon chip based optical fiber coupling with high position accuracy is introduced. In this paper, we present the fiber chip coupling on two examples: A superconducting single photon detector (SSPD) and a miniaturized fiber Bragg grating sensor. In case of the SSPD position accuracy between SSPD and optical fiber of ± 1 μm is necessary.

In this paper we show the developed alignment system and the proof of the position accuracy on silicon test chips. Further, we show first experiments of a fiber coupled superconducting test structure in a closed cycle cryostat with regard to stability of the chip stack and thermal connectivity to the cryostat.

The fiber coupling of the fiber Bragg grating sensor is used to miniaturize the sensor overall construction. The fiber Bragg sensor consists of two stacked Silicon photodiodes. The top photodiode is fabricated in a cavity within a remaining 50 μm Silicon membrane and therefore detects only the shorter wavelength range. The bottom photodiode detects the transmitted longer wavelengths. The fiber coupling chip is mounted on top of the photodiode stack. This leads to a compact chip stack with included fiber coupling, without the need for large fiber connectors or ferrule holders. Further, we demonstrate the mounting of the fiber Bragg sensor on a flexible PCB and its performance.

This Conference Presentation, "Fundamental limitations for phase-locking of integrated laser arrays," was recorded at Photonics West 2020 held in San Francisco, California, United States.

Oxide-free VCSELs can be reduced in cavity dimension to micron size apertures, enabling ultalow bit energy operation based on directly modulated laser operation. The oxide-free VCSEL can be used at room temperature, high temperature or cryogenic temperatures including 77 K or 4 K operation. Important applications for these optical links include extracting optical data from cryogenic focal plane array cameras, or from superconducting circuits at low temperature. Oxide-free VCSELs enable more than an order of magnitude reduction in cavity volume over oxide VCSELs, and promise high reliability for applications in data centers and cryogenic data links for military and private sector applications. This talk is an invited presentation.

Capacity demand for network connections within and between datacenters is increasing relentlessly, fueling the need for deployment of new and improved optical communications equipment. Confronted with the task of developing innovative solutions to address this challenge, engineers must deal with and consolidate countless design choices that are influenced by a large variety of constraints. To name just a few, an optimum solution may depend on technological requirements such as minimum data rate, maximum latency, electrical and optical bandwidth, link distance, upgradability (to higher speeds and/or other/more wavelengths), as well as the need to comply with standards and how these evolve. In this sense, automated design tools for simulating and comparing alternative solutions are indispensable. We present design examples at the system- and component-levels, illustrating the challenges in modeling, analyzing and optimizing technology choices and equipment parameters of optical interconnects for intra- and inter-datacenter applications. Of critical importance for tuning the performance of transceiver components is the integrated co-design of the corresponding electronic and optical parts. We demonstrate a seamless design flow linking simulations of the electronic circuits at the transmitter/receiver (such as serialization/deserialization, DAC/ADC, driver amplifier/TIA, etc.) with simulations of the optical fiber link, enabling investigation and optimization of the overall system performance. Further, we compare advantages and challenges of multimode infrastructure solutions utilizing, for instance, PAM4 modulation of multimode VCSELs with transmission over wide-bandwidth multimode fibers, and single-mode solutions employing Mach-Zehnder modulators with tunable DFB lasers in WDM operation over SMF-links.

Heterogeneous integration of dissimilar materials are essential in achieving advanced functionalities in integrated photonics and electronics integrated chips. Over the years, transfer printing technique has emerged to be a viable and practical process for heterogeneous integration of functional nanomembranes of different crystalline semiconductor materials. Here we report the development of transfer printing automation tools and processes based on commercial Ficontec optical characterization and assemble tool. Custom-made polydimethylsiloxane (PDMS) stamp printing head was developed along with the automatic pick and printing processes. A wide range of materials were also investigated including thin film silicon and monolayer 2D materials. Improved yield and device performance were achieved based on the characterizations of optical and light-matter interactions.

Optical wireless data transmission is an emerging complementary technology compared to the RF transmission in 5G and 6G communications. The optical phased array (OPA) would be a promising component for such high-speed wireless transmission requiring beam-steering function. We achieved 25Gbps data transmission using silicon-based OPA and commercialized photodetector (PD). OPA was designed as 64 channels having output beam divergence angle less than 1 for both horizontal and vertical direction. The diverging beam is received by high-speed PD with active diameter less than 30μm. The demonstration of 20Gbps data transmission was practiced with OPA’s output beam power as 0dBm and a freespace range of 1.5m. Received signal from photodetector has an eye-diagram with an extinction ratio of 5dB. Depending on our link power budget, higher output beam power and smaller beam divergence angle concludes to longer free-space range and larger bandwidth. Improvements on beam concentration by lens system could also increase range and bandwidth.

We propose a vertical-cavity surface-emitting laser (VCSEL) laterally coupled to multiple transverse cavities (MTCCs) providing modulation bandwidth beyond the relaxation-oscillation-frequency. We show a 3dB modulation bandwidth of 45 GHz, five times greater than it’s conventional VCSEL fabricated on the same epi-wafer.

Metallic optical benches that are useful in photonics packaging and optical interconnects can be manufactured with ultra-high precision stamping processes. They contain microscale features suitable for passive alignment of optical components. A recent innovation is the ability to stamp micro mirrors into metallic optical benches. These mirrors have finishes better than Sa = 5 nm without any secondary processes. Micro mirrors expand the utility of metallic optical benches by enabling light beams to be folded or shaped using aspherical micro mirrors that focus or expand the light beams. This is accomplished without any additional cost of refractive for diffractive lenses, and the mirrors are aligned to other bench features by the stamping process. We present exemplary metallic optical benches with micro mirror arrays for connecting optical fibers to photonic devices such as photonic integrated circuits and discrete photodiodes.