This PDF file contains the front matter associated with SPIE Proceedings Volume 10924, including the Title Page, Copyright information, Table of Contents, Author and Conference Committee lists.

We demonstrate single-mode polymer waveguides and devices fabricated by using commercially available UV-curable epoxies and the mosquito method for on-board high-speed optical interconnects application. Both circular core singlemode waveguides operating at 1550 nm with a transmission loss of 0.79 dB/cm and waveguide based devices such as 3D directional coupler and mode (de)multiplexer have been designed and fabricated. The results imply that the proposed single-mode polymer waveguides and devices have good potential in high-density and high-speed on-board optical interconnects application.

Optical technologies are increasingly considered for use in high-performance electronic systems to overcome the performance bottleneck of electrical interconnects when operating at high frequencies and provide high-speed communication between electronic chips and modules. Polymer waveguides are leading candidates for implementing board-level optical interconnections as they exhibit favourable mechanical, thermal and optical properties for direct integration onto conventional printed circuit boards (PCBs). Numerous system demonstrators have been reported in recent years featuring different types of polymer materials and opto-electronic (OE) PCB designs. However, all demonstrated polymer-based interconnection technologies are currently passive, which limits the length of the on-board links and the number of components that can be connected in optical bus architectures. In this paper therefore, we present work towards the formation of low-cost optical waveguide amplifiers that can be readily integrated onto standard PCBs by combining two promising optical technologies: siloxane-based polymers and ultra-fast laser plasma implantation (ULPI). Siloxane-based waveguides exhibit high-temperature resistance in excess of 300°C and low loss at different wavelength ranges, while ULPI has been demonstrated to produce very high dopant concentrations in glass thin films with values of 1.63×10²¹ cm⁻³ recently reported in Er-doped silica layers. Here we present detailed simulation studies that demonstrate the potential to achieve a internal gain of up to 8 dB/cm from such structures and report on initial experimental work on Er-doped films and waveguides demonstrating photoluminescence and good lifetimes.

Multimode polymer waveguides have attracted strong interest for use in high-speed board-level interconnections as they can be cost-effectively integrated onto standard printed circuit boards (PCBs) and flexible substrates using conventional methods of the electronics industry and provide low-loss (<0.04 dB/cm at 850 nm) and high-bandwidth (>30 GHz×m) interconnection. Various high-capacity passive optical backplanes have been demonstrated using this technology while data transmission up to 40 Gb/s using NRZ modulation has been reported. Despite however the intensive research in this technology, very few studies have been reported on mode mixing effects in such multimode waveguides. Mode mixing is a very important phenomenon in highly-multimoded systems as it greatly affects the mode power distribution and therefore, light propagation inside these waveguides. Important transmission characteristics such as their loss and bandwidth performance are affected as well as the behaviour of passive waveguide components such as bends, crossings and couplers due to the different mode distribution at their input. In this work therefore, we present theoretical and experimental studies on mode mixing in polymer multimode waveguides used in board-level optical interconnects. Measurements are carried out on 24-cm long flexible waveguide samples to assess the strength of mode mixing using two common methods used in multimode fibre: mandrel wrapping and micro-bending, while a simple ray tracing model is developed to correlate mode mixing strength with waveguide sidewall roughness. The combination of experimental and theoretical studies can indicate the strength of the effect over the practical range of lengths (~1 m) which are relevant to the application.

This paper demonstrate a large-size directly inscribed optical waveguide for card-to-card optical interconnects applications. The waveguide was fabricated with commercially available UV-curable epoxies by combining use of a needle-type liquid micro-dispenser and a 3-axis robot stage known as the mosquito method. We designed and fabricated an “E” shaped waveguide device with 3 input/output ports and three connection routes with a length of about 21 cm, 21 cm, and 27 cm, respectively, to fit card-to-card connections of the optical backplane. The optical layout was designed to be scalable for conventional backplane sizes. Moreover, both the optical and geometrical parameters of the waveguides are designed to be compatible with commercial devices and can be terminated with MT connectors. The measurement results show good optical performance in insertion loss, crosstalk, misalignment tolerance, and good uniformity for all the 18 channels. The average insertion loss of the 27 cm-long channels is about 4.5 dB and the misalignment tolerance is larger than 20 μm for 3 dB loss penalty. The measured inter-channel waveguide crosstalk is lower than 42 dB. The waveguides also show excellent high-speed transmission performances with a NRZ signal at a data rate of 32 Gb/s. The experimental results imply that the large-size directly inscribed circular core polymer waveguide device is suitable for card-to-card optical interconnects applications and especially useful and versatile for the development of the porotype systems.

Reduction of laser operating energy is a key issue to use the lasers in datacom and computercom networks because internet traffic is still increasing. To reduce operating energy, it is important to increase the optical confinement factor because the modulation efficiency is proportional to square root of optical confinement factor. Thus, the integration of thin membrane laser on SiO2/Si substrate is essential. Employing buried heterostructure (BH), in which the active region is buried with InP layer, is also important because BH provides efficient carrier confinement and thermal conductance. For datacom application, we have developed membrane Distributed Reflector (DR) laser array on SiO2/Si substrate. To fabricate BH, we have employed epitaxial growth of InP layer on a directly bonded InP on SiO2/Si substrate. We demonstrated an 8-channel DR laser array with integrating an SiN arrayed waveguide grating (AWG) filter. We have also developed photonic crystal (PhC) wavelength-scale cavity laser to obtain ultra-low operating energy for computercom application. The device exhibited a threshold current of 22 A, and a 7.3-fJ/bit energy cost directly modulated with a 10-Gbit/s NRZ signal. These results indicate that the membrane BH lasers on SiO2/Si substrate are highly suitable for use as a transmitter in datacom and computercom applications.

We demonstrate an optimized silicon photonic link architecture using components from the AIM PDK that achieves an ultra-low sub-pJ/bit power consumption with an aggregate bandwidth of 480 Gb/s. At the transmitter, micro-disk modulators are cascaded along a bus waveguide to select and modulate wavelength-division multiplexed (WDM) channels. At the receiver, micro-ring resonator (MRR) filters are thermally tuned to match the corresponding disks to select from the multiplexed channels. This link architecture yields an ultra-small footprint compared to Mach-Zehnder designs, improving the system scalability and bandwidth density. Additionally, using micro-resonators to select and drop the desired wavelengths from a single bus waveguide allows for straightforward integration with a frequency comb source. The energy performance of the design is optimized through sweeping over three key parameters: (i) optical power per channel, (ii) channel count, and (iii) bitrate. These parameters are the dominant sources for the crosstalk and power penalty in the link design. We identify ideal points in the design space which minimize the energy per bit while staying below the desired bit error rate (BER) of 10^-12 and maintaining a realistic aggregate bandwidth. Simulations in the Synopsys OptSim environment using the AIM PDK v2.5a models confirm the functionality of the system with a BER < 10^-12, acceptable for both high performance computing (HPC) and data center (DC) applications. Furthermore, optimizing the link energy consumption in the AIM PDK provides a clear path towards low-cost and high-yield fabrication suitable for application in HPC and DC systems.

High-speed photodetectors with high sensitivity and low noise are needed in Lidar applications. Existing p-i-n photodetectors and avalanche photodetectors (APD) are limited by the gain-bandwidth product. Plasmonic optical antenna technology provides an effective approach to increase the gain-bandwidth product with a low noise level. In this paper, we report a pointed-dipole optical antenna enhanced APD with a high gain-bandwidth product and ultra-low noise and ultra-small form factor. This kind of plasmonic optical antenna enhanced APDs can also be used in high-density optical interconnects.

Photonic Integrated Circuit (PIC) technology is becoming more and more mature and the three main platforms that offer Multi Project Wafer runs (Indium Phosphide (InP), Silicon on Insulator (SOI) and the silicon nitride based TriPleX platform) each have their own unique selling points. New disruptive PIC based modules are enabled by combinations of the different platforms complementing each other in performance. In particular the InP-TriPleX combination are two very complementary technologies. Combining them together yields for instance tunable ultra-narrow linewidth lasers extremely suitable for telecom and sensing applications. Also microwave photonics modules for Optical Beam Forming Networks and 5G communication can, and have been realized with this combination. Important part of this combination is the integration of the different platforms in modules via cost effective assembly techniques. This talk will present the combination of both technologies, the interconnection issues faced in the assembly process and latest measurement results on these hybrid integrated devices.

Laser integration and photonics chip packaging are the two key challenges that require attention to drive down the cost/bit metric for silicon photonics based optical interconnects. We try to address the latter by demonstrating optical interfaces that fit well in an overall scheme of 2.5D/3D electro-optic integration needed for a high performance computing environment. A through-substrate coupling interface provides the benefit of bonding a silicon photonic chip face-up on a package substrate such that the device-side of the chip remains accessible for die-stacking and fiber-array packaging, thereby offering a promising alternative to flip-chip based packaging. In this paper, we demonstrate three through-substrate coupling elements to enable alignment tolerant and energy-efficient integration of silicon photonics with board-level or package-level optical interconnects : (i) a downward directionality O-band grating coupler with a peak -2.3 dB fiber-to-silicon waveguide coupling efficiency; (ii) polymer microlenses hybrid integrated onto the substrate of a silicon photonic chip to produce an expanded collimated beam at λ=1310 nm for a distance of more than 600 μm; (iii) a ball lens placed in a through-package via to result in a 14 μm chip-to-package 1-dB lateral alignment tolerance for coupling into a 20×24 µm squared cross-section board-level polymer waveguide.

The ever-increasing demands in traffic fueled by bandwidth hungry applications are pushing data centers to their limits challenging the capacity and scalability of currently established transceiver and switching technologies in data center interconnection (DCI) networks. Coherent optics emerged as a promising solution for inter-DCIs offering unprecedented capacities closer to data centers and relaxing the power budget restrictions of the link. QAMeleon, an EU funded R and D project, is developing a new generation of faster and greener sliceable bandwidth-variable electro-optical transceivers and WSS switches able to handle up to 128 Gbaud optical signals carrying flexible M-QAM constellations and novel modulation techniques. A summary of the progress on the QAMeleon transponder and Reconfigurable Optical Add/Drop Multiplexer (ROADM) concepts is presented in this paper.

The continuous evolution of data centers (DC’s) has fueled a relentless growth of burst-mode (BM) traffic. DC networks have long relied on electrical switching, but the mismatch between optical transmission and electrical switching has become a bottleneck for DC networking. Optical switching, with related BM technologies, can play a pivotal role in overcoming this issue. For intra-DC networks, switching on a packet-by-packet basis requires optical switches with fast re-configurability in the order of nanoseconds. We discuss the requirements these switches have to satisfy and highlight the tradeoffs in their implementations by reviewing three different types of optical switches we have developed; namely a wavelength-based routing switch (WRS) comprising fast tunable laser diodes (TLD) and an arrayed waveguide grating (AWG), a broadcast-and-select (B and S) switch based on electro-absorption modulators (EAMs), and a route-and-select (R and S) switch combining Mach-Zehnder interferometers (MZIs) and EAMs. On the other hand, apart from warehouse-scale DC networks, the pressing demand for reducing data transmission latency has been promoting the distribution of smaller DCs in the metro segment of optical transport networks. We address this demand by proposing the realization of a dynamic and highly scalable inter-DC network based on BM optical switching technologies, and highlight the role of fast space/wavelength switching to fulfill this goal.

The performance shortcomings of multipurpose compute engines have stirred recent excitement in specialized processors, preempted by GPUs. Simultaneously, computational complexity theory NP 'hard' problems scaling as O(n^k) require new hardware solutions. This presents an opportunity for photonic information processors (PIP) building on photonic integration through recent foundry developments. The value proposition for PIPs exist via optical parallelism, small capacitive charging of OE devices, 10's of ps short propagation delays, a natural convolution via optical interference, and an O(n)-scaling Fourier transform. Based on a recently developed photonic NxN router, here we present two photonic processors; a) the residual arithmetic nanophotonic computer (RANC), and b) a reconfigurable graph processor, the latter being a computing-in-switching (CIS) paradigm. PIPs operate with time-of-flight, once the processor is configured (e.g. setting phase), which is on the order of 10-100 ps given the mm-scale photonic integration footprints. This high bandwidth, however challenges the electronic-optic I/O bottleneck. To address this, we further discuss an optical front-end DAC with <100 ps delay enabled by a 2x2 electro-optic switch.

As a tremendous amount of data is being created exponentially day by day, integrated optical computing starts to attract lots of attention recently due to the bottleneck in the continuation of Moore’s law. With the rapid development of micro/nano-scale optical devices, integrated photonics has shown its potential to satisfy the demand of computation with an ultracompact size, ultrafast speed, and ultralow power consumption. As one of the paradigms in optical computing, the electro-optic logic that combines the merits of photonics and electronics has made considerable progress in various fundamental logic gates. It therefore becomes very critical to develop an automated design method to synthesize these logic devices for large-scale optical computing circuits. In this paper, we propose a new automated logic synthesis algorithm based on And-Inverter Graphs (AIGs) for electro-optic computing. A comprehensive component library of electro-optic logic is summarized with several new proposed logic gates. As an example, a large-scale ripple-carry full adder which serves as the core part of the arithmetic logical unit (ALU) is presented. In the design, all the electrical signals could be applied simultaneously at every clock cycle and then the light could process the signals through every bit at the speed of light without any delay accumulated. High-speed experiment demonstrations are carried out, which show its potential in future high-speed and low-power-consumption optical computing.

We introduce a suite of modular test cards designed and developed on the H2020 COSMICC project to allow test and validation of a silicon photonics transceiver at up to 56 Gb/s. The modular test cards include large and small mezzanine cards to house the silicon photonics transceiver under test and a power distribution and sense card, which allows real time measurement of power consumption in a data center environment. The test modules can be driven stand-alone or incorporated into different data center platforms.

A numerical study on the design and loss mechanisms of a Polymer-to-Silicon adiabatic coupling interface is demonstrated. The interface is composed of Silicon Photonics (SiPh) platform, assisted by Silicon Nitride layer (SiN), flip-chip bonded on Polymer optical board. The silicon taper design is based on the coupling strength maximization as well as the calculation of a Constant Loss Taper (CLT). The maximum calculated coupler efficiency value achieved is 87.6 %, for silicon taper length of 590 μm. The spectral behavior of the proposed coupler is investigated, while a fabrication and assembly tolerance study is conducted.

A novel coupling approach with specific coupling ratios depending on the coupling direction (optical bus to electrooptical device and vice versa) is presented. According to that, an asymmetric optical bus coupler (AOBC) is obtained, where a high amount of light is coupled from a device/module into the optical bus, whereas the ratio in the opposite direction is significantly smaller to keep enough energy in the bus for later coupling. This is realized without any filters by a specific bending of the waveguides. Hence, an up to four times higher coupling ratio in the module-to-bus direction is achieved compared to the opposite direction at the same junction. In contrast, current coupling schemes have a 1:1 coupling, waveguides are interrupted or filters are used. In this paper we investigate the influence of the waveguide manufacturing technology on the coupling performance. The standard photolithographic waveguides with rectangular cross-sections are compared to Aerosol Jet printed waveguides with circular segment cross-sections. Furthermore, we monitored the coupling for several days to get results on the stability. The results show a fluctuation of about ± 20 % for the printed waveguides and about ± 5 % for the photolithographic ones.

Vertical cavity surface-emitting lasers (VCSELs) are the most popular choice as the ultra-fast and energy-efficient transmitters for use in optical interconnection of datacenters and supercomputers. Improvements on the wafer level have made much progress. High speed operations at data rates beyond 40 Gbps were demonstrated through the careful optimization of epi-wafers and device structures, and reached at to NRZ 71 Gbps with electrical equalization employed. But there remain difficulties in increasing the frequency response, which is mostly limited by the relaxation oscillation frequency and paracitics. In this paper, we review the progress of transverse-coupled-cavity VCSELs for increasing the modulation bandwidth and the monolithic lateral modulator integration. The noticeable modulation-bandwidth enhancement can be exhibited owing to “photon-photon resonance effect”. We demonstrated the modulation-bandwidth enhancement of a quasi-single mode VCSEL with a passive optical-feedback-cavity. The 3-dB modulation bandwidth can reach at 30 GHz. Eye opening of large signal modulations at 48 Gbps was obtained. The modeling and future prospect toward the bandwidth enhancement beyond 100GHz will also be presented. In addition, we demonstrated the lateral integration of an ultra-compact electro-absorption modulator with VCSEL. We obtained a sub-volt low driving voltage, and the bandwidth beyond 30 GHz. Our ultra-compact modulator integrated VCSEL can boost the modulation speed far beyond the direct modulation bandwidth for use in next-generation computing and data-center networks. We will also discuss on the multi-level modulation format, wavelength division multiplexing and space division multiplexing toward 400 Gbps or higher per fiber.

The limitations of electronic and optical interconnects inside a semiconductor chip could be overcome by the used of plasmonic-polaritonic interconnects whereby information is carried by surface-plasmon-polariton (SPP) waves guided by planar metal/dielectric interface. Investigating the propagation of a pulse-modulated carrier SPP wave when the metal is either abruptly terminated or bent and the dielectric material is air, we have found that SPP-wave-based interconnects can be effective conduits of information.

Discovery of TCC-VCSEL done by Dr. Dalir (PI) in 2013 led to new functionalities of VCSEL structure. In principal, a TCC-VCSEL has same vertical structure as conventional VCSEL. A VCSEL is consists of two distributed Bragg reflector (DBR) mirrors parallel to the wafer surface with an active region consisting of one or more quantum wells for the laser light generation in between. The planar DBR-mirrors consist of layers with alternating high and low refractive indices. Each layer has a thickness of a quarter of the laser wavelength in the material, yielding intensity reflectivity’s above 99%. High reflectivity mirrors are required in VCSELs to balance the short axial length of the gain region. Here we assume a TCC-VCSEL with a coupling of K between the cavities. We pump one cavity with a gain of g, while the other cavity has loss of ɣ. It is noted that the lasing frequency is a function of loss and coupling between the cavities. Assuming a constant coupling (K), tunabilty of the TCC-VCSEL will be adjusted by the loss. The three-dimensional simulation of the single mode operation in TCC structure is performed by employing film mode matching method of FIMMWAVE Photon Design Corp. With a coupling of K= 1.5THz, a 19.7 nm wavelength will be swept in the PT regime crucial for lab-on-a-chip integrated bio-sensor applications.

Silicon subwavelength grating metamaterial (SGM), formed by periodically interleaving two or more types of materials a pitch far less than the operating wavelength, has been attracting substantial interest in recent years. It provides an effective approach to tailor the optical properties of naturally existing materials and thereby develop integrated photonic devices with unprecedented applications. In this paper, we introduce silicon photonics SGM waveguide based devices for optical interconnect and sensing applications. Silicon provides high index contrast and thus device footprints can be made very small. However, high index contrast is a double-edged sword, which also leads to a few devastating limitations such as limited photon-matter interaction and high dispersion. SGM waveguide can potentially resolve these issues. For instance, silicon-organic hybrid SGM modulator demonstrates < 44 GHz 6 dB bandwidth and estimated energy consumption of 2.55 fJ/bit. SGM waveguide based refractive index sensors show eminent improvement on the sensitivity compared to conventional strip waveguide based resonators. In the meantime, unlike evanescent wave based sensors, the surface sensitivity of SGM sensors does not decrease as analytes accumulate on the waveguide surface. SGM waveguide based passive components demonstrate improved bandwidth due to the decrease of waveguide dispersion. Other potential applications of SGM waveguide will also be discussed.

The persistent drive for increased bandwidth while simultaneously reducing cost in short range optical communications systems is placing substantial pressure on component manufacturers to increase the level of integration and adopt lower cost assembly processes. Moving to well established processes used widely in the electronics industry such as wafer scale test and assembly, and high speed vision-based alignment is a key step to satisfying this demand. Ultrafast lasers offer a unique tool for the integration of photonic components. This paper show how they can be used for manufacturing glass-based interconnect components, which are applied to both fiber-fiber and fiber-transceiver applications. This paper shows how these components can be used as optical interfaces or further combined with conventional lithographic processes to help solve challenges in the packaging of photonic components. In addition to this, the paper will also show how these components can be used as a n integration platform for optical sub-assemblies and photonic-electronic interposers.

Densly packed multi-channel fiber arrays have manifold applications in the field of photonic integrated circuits, telecom and many more. The assembly of the fiberoptical connectors is highly challenging. The requirements for the lateral and angular final assembly precision are below 1° and 1 micron. At Fraunhofer IPT an automated assembly machine for PM-fiber connectors with a 127 micon pitch has been successfully implemented and installed in industry. With the patent-pending fiber manipulator and a computer vision-based stressrod inspection system an efficient manufacturing of fiber arrays is possible. The system consists of a tray-based feeding system which can carry up to 32 individual fibers, stripped and cleaved to be automatically assembled. After the operator-guided loading of the machine the remaining process is carried out by the machine autonomously. A handling system picks up individual fiber and automatically clamps the fiber tip on the acrylic coating for the angular alignment about the fiberaxis. Next, the fiber is brought into the field of view of a telecentric camera system that detects both orientation of the V-groove holder and the fiber angle. Through a piezo-based rotational manipulator the alignment between fiber and V-groove holder is done. After alignment a dispensing system applies a precise amount of UV-curing adhesive for the individual fiber fixation. The process is repeated until the defined amount of fibers is assembled. As a last step a lid is mounted for final fixation of the fibers in the V-grooves. The process can be transferred to direct fiber-chip coupling and related applications.

The rapid development of nanophotonic chip devices combines the advantages of ultra-high miniaturization and largescale integrated circuit production, targeting a drastic decrease in the device costs. Such photonic devices based on single-mode waveguide structures are intensively used in optical communications but recently novel high power applications have emerged where it has been demonstrated that nano-sized waveguides support ultra-high optical power densities, giving rise to the generation of non-linear effects. However, all of the applications require a high efficient and stable fiber interface between the chip and the outside world. In this letter, we propose a stacked glass block technology in combination with a sub-micron alignment approach for the development of robust fiber connectors for lateral coupling. This technology offers a high degree of freedom during the assembly of glass parts, compensating undesired misalignments related to the shrinkage of the adhesive. Moreover, adhesive-based issues are minimized by stacking the glass parts with almost zero gap. The design of the glass fiber connector can be adapted for one or more optical fibers. Using an automated laser cutting system, small glass parts from glass panels are cut with high reproducibility and acceptable cut quality. Two different chips for photonic modules of specific output power level with a laser excitation wavelength at 1550 nm are addressed to show the potential use of fiber connectors based on stacked glass blocks.

Single-mode integrated photonics assembly is challenging due to the tight alignment required for optical connections. To address this, we have developed a parallelized fiber assembly process using self-alignment of fiber arrays in V-grooves defined on the photonic chip. This approach is compatible with standard automated high-throughput pick and place tools, thus improving the scalability and cost-efficiency of photonic packaging. We describe our efforts toward increasing the assembly throughput as well as making the photonic connections compatible with high temperatures from downstream microelectronic assembly processes, such as lead-free solder reflows. The ability to survive these higher temperatures allows the pretesting of optical engines and paves the way for true co-integration of photonics and electronics. We have shown that attaching fibers to chips using multiple adhesives with partitioning of their function provides substantial gains in both throughput and reliability at a relatively small cost of dispense complexity. The fibers are tacked in place with lengthy adhesive cures performed in a batch process outside of the placement tool so as not to impact the placement tool’s throughput. This approach allows for strong long term structural integrity along with optimized optical index matching between the fiber and the waveguide coupler of the photonic dies. Not only did we observe the same peak optical performance that we previously reported, but we have also demonstrated no experimentally significant optical penalty after 5x lead-free solder reflows, in operational temperature between -15°C and 150°C and, after aggressive microelectronic environmental stressing going beyond the parameters traditionally used in optics. The ability to embed single mode optics in the first level package is a disruptive capability contributing to enable the high-density interconnects needed to meet the ever-increasing bandwidth demands for data communication. We discuss the benefits of such configurations, as well as the challenges for thermal management and system yields.

Silicon photonics is gaining increasing adoption for mid- to long-reach communication links in datacenters at 100Gbps and beyond. Two significant challenges remain integrated on-chip WDMs, and fiber and laser packaging. The adoption of wavelength division multiplexing (WDM) allows for multiple data signals to be carried on a single fiber, reducing the cost of fiber provisioning compared to parallel implementations and greatly increasing achievable data rates. Designing WDM structures can be challenging on a silicon platform, however, as silicon waveguides are highly sensitive to minute fabrication variations and temperature changes. To account for and compensate for these changes, it is necessary to have a robust testing methodology to characterize the WDM system, as well as an efficient and complete tuning mechanism to dial in the desired performance without sacrificing too much in power consumption, chip area for off-chip electrical connections, and insertion loss. We will present a monolithic silicon WDM design, the implementation of low power tuning elements, and the test and biasing algorithms to align all filters on the desired CWDM grid. Another perennial challenge for silicon photonics has been coupling light on and off the chip, due to the significant mode size mismatch and exacting alignment requirements between silicon waveguides and optical fibers or III-V chips. We will present on packaging developments implementing fully passive alignment using existing CMOS production tools, improving scalability and cost efficiency and enabling the packaging techniques for silicon photonics chips to keep up with the massive volumes generated by silicon waferscale fabrication.

Silicon based photonic integrated circuits (Si-PIC) are becoming a key contender for data center related transceivers. Available Si-PIC platforms use fast and well-integrated modulators based on the plasma-dispersion effect. However, they intrinsically suffer from high-loss and coupled amplitude and phase modulation. Using the Pockels effect enables highly efficient, low-loss phase shifters without residual amplitude modulation. Hybrid barium titanite (BaTiO₃) Siphotonics technology with a strong Pockels effect on large-scale silicon substrates has emerged as a platform to implement such devices. The platform enables scalable, integrated Si-BaTiO₃ devices with large Pockels coefficients for electro-optic modulation. Based on this platform, various electro-optic devices such as low-power tuning elements, monolithically integrated Si-BaTiO₃ modulators, and ultra-compact plasmonic BaTiO₃ modulators have been demonstrated.

The integration of optical sources in Si photonic transceivers has relied so far on externally coupled III-V laser dies within the assembly. These hybrid approaches are however complex and expensive, as there are additional cost-increasing factors coming from the redundant testing of the pre- and post-coupled laser photonic chips. Further optimization of Photonic Integrated Circuits (PICs) cost and performance can be obtained only with radical technology advancements, such as the “holy grail” of Silicon Photonics; the monolithic integration of III-V sources on Si substrates. MOICANA project funded by EU Horizon 2020 framework targets to develop the technological background for the epitaxy of InP Quantum Dots directly on Si by Selective Area Growth with the best-in-class, in terms of losses and temperature sensitivity, in a CMOS fab, i.e. the SiN waveguide technology. In addition, MOICANA will develop the necessary interface for the seamless light transition between the III-V active and the SiN passive part of the circuitry featuring advanced multiplexing functionality and a combination of efficient and broadband fiber coupling. Through this unique platform, MOICANA aims to demonstrate low cost, inherent cooler-less and energy efficient transmitters, attributes stemming directly from the low loss SiN waveguide technology and the QD nature of the laser’s active region. MOICANA is targeting to exploit the advantages of the monolithic integrated PICs for the demonstration of large volume single-channel and WDM transmitter modules for data center interconnects, 5G mobile fronthaul and coherent communication applications.

Indium Phosphide integrated photonics enables the combination of high-speed lasers and modulators with filters, detectors and multiplexers in one wafer-scale process flow. Low-voltage modulation at rates of 50Gigabit/second and above are feasible with combinations of semi-insulating substrates, optimised multi-quantum wells and high-speed electrical design. Advances in monolithic InP platform technologies have created a mechanism to rapidly introduce such high performance building blocks into sophisticated integration processes, enabling photonic integrated circuits with many tens of active components including distributed feedback lasers, tunable lasers and a range of passsive components. Our recent introduction of 192nm deep UV scanner lithography – believed to be a world first for InP integrated photonics – also enables a step change in the performance for the integrated filters and mode control. In this paper, we present recent innovations in the creation of high performance transceiver technologies for optical interconnects. We showcase circuits using InP integrated photonics to create high-speed, energy-efficient, optically-multiplexed circuits. Monolithic polarization multiplexing and wavelength domain multiplexing are reviewed where all components, inclusive of the lasers, are created in the same wafer. Line-rates of up to 320 Gb/s are demonstrated for optically multiplexed circuits using a variety of open access fabrication platforms. The challenges and approaches for Terabit/second class transceiver chips will be addressed, addressing crosstalk management, component miniaturisation, and intimate electronic integration.

Generation and recombination of free carriers in silicon photonics is fundamental to understand several nonlinear optical phenomena and engineer novel devices. Particularly in strip nano-waveguides, the tightly confined optical field results in highly efficient generation of free-carriers, both through linear and nonlinear absorption. Furthermore, the large surface-to-volume ratio results in a nonlinear recombination behavior dominated by a trap-assisted mechanism. Through time-resolved pump-and-probe experiments, we performed a detailed experimental characterization of linear and nonlinear generation rates, as well as recombination dynamics. We developed analytical expressions to determine the carrier density averaged along the waveguide from the measured free-carrier absorption for different input pump power levels. As a result, we were able to discriminate the contributions from two-photon absorption (TPA) and single-photon absorption (SPA), obtaining absorption coefficients of (1.5±0.1) cm/GW and (1.9±0.1) m⁻¹ , respectively. Our results then reveal that the effective TPA within the waveguide is higher than the value reported for bulk silicon, and that SPA plays an important role in carrier generation up to ≈300 mW. With regards to recombination dynamics, our results show a highly nonlinear decay curve with instantaneous carrier lifetime varying as the recombination evolves (initially faster with lifetime of ∼800 ps and slower at final stages of the decay, reaching ∼300 ns). We interpret our results with a theoretical framework based on trap-assisted recombination statistics applied to strip nano-waveguides, and explore its implication to the dynamics of nonlinear nanophotonic devices in which free carriers play a critical role.

There is a growing demand for integrated photonic devices from a wide range of applications, including high-speed communications, medical devices and sensing. There has been significant research and development to address the demand for fiber-based optical interconnect for integrated photonic devices. However, applications such as medical diagnostics and remote sensing often require free-space coupling using micro optical components. As with fiber-based systems, packaging of micro optical components presents challenges in order to maximise coupling efficiencies and ensure packaging processes can be scaled to high-volume for cost-effective manufacturing. In this talk, I will present an overview of challenges in optical interconnect for integrated photonic devices, including Silicon, InP and SiN technologies. Factors including device layout, optical fiber and micro optical lens design will be reviewed. This includes the development of on-chip mode adaptors, improved device facet configurations for easier fiber packaging and specialty fibers which improve optical coupling efficiencies. The talk will also present information about the European Photonic Packaging Pilot Line, PIXAPP. The Pilot Line addresses the gap between early-stage prototypes and scaling to volume manufacturing. PIXAPP achieves this through the implementation of packaging design rules and working towards standard package designs. The talk will also present details of a novel optical test system which is being developed within the Pilot Line to ensure devices are characterised in advanced of packaging. PIXAPP is supporting a large number of companies develop their first photonics-based products and the talk will be of interest to those from a wide range of application areas.

Grating couplers are the most versatile mechanism to couple light efficiently into photonic interconnects, such as waveguides featuring submicronic cross-sections. Usually grating couplers are used in a tilted illumination configuration in order to obtain unidirectional excitation of the waveguide, requiring challenges in assembly and packaging.

In practical applications, tilted illumination of the gratings is not always possible in particular for fully integrated electrooptical printed circuit board (EO-PCB) with a light source (Vertical-Cavity Surface-Emitting Laser (VCSEL)) and an optical layer implanted on each side of the board. In this case, the incoming light hits the gratings couplers at normal incidence and specific strategies are needed to achieve unidirectional excitation of the guided mode.

In this work, a novel unidirectional Data Center coupling concept based on the use of gap plasmon polariton (GPR) grating couplers sustained by Metal-Insulator-Metal (MIM) resonators is introduced. Unlike traditional challenging subwavelength coupling schemes based on plasmonics and Si-Photonics, we consider non-symmetric GPR featuring highly directional scattering efficiency. The plasmonic gratings have been modelled numerically employing a Fourier Modal Method and the results have been confirmed by FEM simulations.

The only way to make silicon photonics technology completely exploitable at industrial level for volume production is to solve the manufacturing bottlenecks that nowadays limit the applicability of the technology, in particular related to the multiple optical interconnection and the III-V chip integration. This paper will give a survey, based principally on the work made at Inphotec labs, on the problems and the challenges related to the optical interconnections to silicon photonic devices and their impact on process automation. It will give some update to what we are currently doing and developing in our labs utilizing a custom state of the art pigtailing-automated bench, robotics, including their present limitations.

We propose a novel Si photonics module that overcomes the issues of conventional Si photonics modules such as package structure and electrical connection. The module incorporates an optical fiber socket fabricated by blind via socket (BVS) technology, which implements backside optical I/O in a photonic IC (PIC) by forming blind via holes on the backside. High-speed high-density electrical connection to both the PIC and an electrical IC (EIC) is also obtained in the module by fan-out wafer level packaging (FOWLP) technology. These technologies achieve a surface-mountable substrate-less fan-out optical module. It realizes a practicable integrated module of optoelectronic devices excellent in terms of electrical characteristics such as signal integrity (SI) and power integrity (PI), heat characteristics, and miniaturization. This paper presents a BVS module that enables optical coupling between a III-V/Si photodiode (PD) fabricated on a Si substrate and a multi-mode optical fiber by passive alignment of only insertion of the fiber into a blind via hole on the backside. High-speed optical signal transmission is also demonstrated with a fan-out optical module in which a BVS and an EIC are integrated by FOWLP and a vertical-cavity surface-emitting laser (VCSEL) or PD is mounted on the BVS.

A numerical study on the design of a novel dual polarization, low loss, and low fabrication complexity, angled 3D Glassto- SiPh coupling interface is presented. The interface is composed of a 2° angled 3D glass waveguide attached to a thin-SOI platform. The coupler design and the optimization methodology relies on the optimum Constant Loss Taper (CLT) silicon taper calculation. The achieved coupling efficiency is 83.50% and 81.22% for TE and TM polarization state, respectively, at 1.55 μm. The results are obtained using an EigenMode Expansion (EME) propagation solver and are verified through the 3D FDTD simulation method.

Single mode with in-phase operation of the laser arrays is one of the main challenges for engineers to achieve high power in phase far field emission. Recently single mode lasing based on non-Hermitian engineering of the laser is investigated and two scenarios based on parity time (PT) symmetry and super symmetry are presented. Here we present a new method based on Q-enhancing to achieve single mode operation. This design device will satisfy requirement for LIDARs in space applications. The output supports up-to 300 Watts for few integrated coupled cavity lasers.

In a three-level system, actively detuning coupling strength between the respective waveguide modes via electrically gating the graphene double-layer allows ultra-fast switching of light propagating through the outer waveguides. In adiabatic elimination regime, the amplitude of the middle waveguide oscillates much faster in comparison to the outer waveguides leading to a vanishing field build up. As a result, the middle waveguide becomes a dark state and hence a low insertion-loss of 1 dB is obtained. Due to the direct dependency of applied voltage with respect to the absorption in graphene, sub-volt operation with modulation-depth of 5 dB is expected.

Optical antennas/nanoantennas are optical elements that have gained highlight in this decade and have potential application in several branches of photonics/plasmonics, such as, optical sensors, lasers, solid state lighting, photovoltaics, microscopy, etc. Additionally, the dipole nanoantennas have been the antennas elements mostly studied in proposals for those applications and, in special, for free-space communication based on plasmonic circuits.

Here we report some advantages of dielectric resonator nanoantennas (DRNAs) as elements for coupling light to plasmonic-based circuits. Fundamental antenna parameters such as reflection coefficient, gain, efficiency, among others, are evaluated and its advantages are highlighted for nanophotonics applications. A study about an optical link for circuits based on metal-dielectric-metal-dielectric (MDMD) nanostrip waveguides operating at the central wavelength of λ₀ = 1.55 μm is evaluated. We studied the coupling of near- and far-fields of a DRNA matched to an MDMD nanostrip waveguide. The results show the advantage of it as an element for optical coupling from/to free-space as well as to establish a wireless optical link for inter-chip communication.