Genia Photonics (Canada)

Intermodal crosstalk (XT) fundamentally limits the capacity of orbital angular momentum (OAM)-based mode division multiplexing (MDM) systems. This XT arises both from fiber propagation and from imperfections in the multiplexer (MUX)/demultiplexer (deMUX). Under some simplifying assumptions, we isolate these two contributions across different fiber lengths and wavelengths. We measured XT matrices in a 16- mode OAM-MDM system at three fiber lengths and at the center and edges of the C-band. Assuming that fiber XT grows linearly with distance, we predict net throughput for (i) ideal MUX/deMUX and (ii) our free-space laboratory version. Our results show that we begin to lose modes due to XT at 0.9 km in both scenarios. At longer lengths, and particularly at the C-band edges, experimental demonstrations have pessimistic throughput predictions due to strong wavelength-dependent XT in the free-space MUX/deMUX. These findings allow a more accurate understanding of the ultimate capacity limits of OAMMDM for high-capacity, short-reach data center links.

Space division multiplexing (SDM) is a promising solution to increase the capacity of optical communications within data centers. We focus on mode division multiplexing (MDM) architectures with fibers supporting multiple modes. We show that network hardware is significantly less complex for MDM with orbital angular momentum (OAM) compared to MDM with linearly polarized (LP) modes. The advantage stems from low crosstalk among OAM modes, hence limited digital signal processing (DSP) complexity. Unlike demonstrations of MDM with LP, our demonstrations have no multiple-input multiple-output (MIMO) processing. We transmit 16 OAM channels at 56 Gbaud quadrature phase-shift keying (QPSK) over 400 m (distance sufficient for links within data centers). The net throughput of 1.6 Tbit/s per wavelength is the highest reported OAM-MDM throughput, even among demonstrations with much greater DSP complexity. We can span the entire C-band, with throughput of 1.3 Tbit/s at the band edges. To reduce latency and power consumption, 8-channel transmission can achieve a net throughput of 0.8 Tbit/s per wavelength across the C-band with wide margin. The OAM MDM approach can flexibly trade-off throughput, latency, and complexity.

Optical input/output (I/O) must be carefully designed to scale up bandwidth of intra-data center electrical switches, while reducing energy consumption and cost. Innovative receiver architectures are key to achieving these goals. Both cascaded microring resonators (MRRs) and arrayed waveguide gratings (AWGs) can handle multiple wavelengths. For a fixed number of channels, wavelength de-interleavers can lower loss and crosstalk, and also yield more compact solutions. We present a comprehensive model to analyze six receiver architectures; we consider MRR vs. AWG, 1-stage or 2-stage de-interleaver vs. no de-interleaver. We target a 6.4 Tbps per-fiber bit rate and optimize receiver energy efficiency. We sweep combinations of data rate, channel count, and channel spacing. For each receiver under test simulation, we quantify energy consumption per bit and bandwidth density (compactness). Assuming non-return-to-zero (NRZ) modulation, we find that MRR-based receivers can achieve the 6.4 Tbps per fiber target with energy consumption at (or close to) sub-pJ/bit, while AWG-based receivers can only achieve up to 3.2 Tbps per fiber. The MRR-based receivers are generally more energy efficient and have a higher bandwidth density. Area bandwidth density up to 2.9 Tbps/<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\text{mm}^{2}$</tex-math></inline-formula> can be achieved by AWGs, and shoreline bandwidth density up to 21 Tbps/mm can be realized by MRRs. Finally, we review promising ways to further enhance the per-fiber bit rate and lower the energy consumption per bit.

Multidimensional (MD) geometric shaping is an effective approach for achieving spectral efficiency gains in optical communication systems. MD formats also support the joint transmission across different cores of multi-core fibers (MCFs), enabling higher performance improvements through joint decoding. As a structured geometric shaping method, MD Voronoi constellations (VCs) allow low-complexity encoding and decoding without the need for look-up tables, offering superior performance over quadrature amplitude modulation (QAM) formats in terms of bit error rate and mutual information. Moreover, MD VCs can be combined with multilevel coding (MLC) schemes to achieve higher shaping gains after soft-decision (SD) decoding. In this paper, the performance of 16-dimensional VCs with MLC is experimentally demonstrated over a 50 km four-core weakly-coupled MCF transmission system. Compared to 16QAM with bit-interleaved coded modulation at the outer HD-FEC BER threshold of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$ 4.5\times 10^{-3}$</tex-math></inline-formula>, VCs achieve a 6 dB reduction in the required launch power and extend the operating range by 17 dB.

InGaP integrated on a silicon substrate has emerged as a promising platform for nonlinear and quantum photonics, offering high nonlinear conversion efficiency and scalability with silicon-based fabrication infrastructure. This work presents an experimental demonstration of sum- and difference-frequency generation (DFG) in InGaP waveguides. We generate light at 930 nm, 1550 nm, and 2325 nm, achieving conversion efficiencies of 4.5 ± 0.5 W −1 , 1.4 ± 0.2 W −1 , and 0.43 ± 0.04 W −1 , respectively. These results highlight the potential of InGaP-on-insulator for advanced photonic applications, including broadband infrared light generation and quantum-frequency conversion. We discuss a roadmap for this technology to achieve even broader wavelength coverage, higher efficiencies, and quantum-frequency conversion of single photons.

Optical fronthaul for radio access networks will enable the promise of 5G and beyond. We describe how centralized sources and diminutive microring modulators can meet these goals while also reducing footprint and power consumption. © 2022 The Author(s)

We design a fiber amplifier with heterogeneous doping to maximize the power conversion efficiency. Using a full model, high computation cost is needed to predict performance at great accuracy. We show that a radial basis function surrogate model lowers complexity.

Silicon photonics is nothing new. It has been around for decades, but in recent years, it has gained traction as electronic design challenges increase drastically with their atomic-level limitations. Silicon photonics has made significant advancements during this period, but there are many obstacles without an acceptable level of comfort as seen by the lack of semiconductor community involvement. Apart from a series of technological barriers, such as extreme fabrication sensitivity, inefficient light generation on-chip, etc., there are also certain design challenges. In this chapter, we will discuss the challenges and the opportunities in photonic integrated circuit design software tools, examine existing design flows for photonics design and how these fit different design styles, and review the activities in collaboration and standardization efforts to improve design flows.

Abstract An innovative imaging colorimeter measurement system which corrects tristimulus images by referencing an integrated spot spectroradiometer is introduced. Improvements offered to the display test industry by the new measurement system over current colorimeter designs are verified by measurements with comparable instruments and the analysis of a modern mobile display.

4x25Gb/s space and wavelength division multiplexing receiver chips, based on a multi-guide vertical integration platform in InP, are presented. With the state-of-the-art performance and smallest footptrint size in class, they are well suited for use in 100Gb/s fiber-optics interconnects.