Heterogeneous Integration and Silicon Photonics: Enabling Optics at Scale
Steve Alleston, OpenLight Photonics
Silicon photonics has emerged as the dominant technology in areas such as datacom and optical computing and is seen as enabling the next generation of AI driven datacenter architectures, quantum computing and machine vision. Despite this promise it is still treated as just one technology choice amongst many with significant barriers to entry and many technical drawbacks. Heterogeneous integration addresses the material limitations of silicon and has been significant in some areas but to date it has been a proprietary technology and the domain of a select number of companies with large R and D budgets and plenty of patience. Now with the development of open PDKs and the emergence of commercially available reliable design tools from new and established vendors the ecosystem is becoming established to allow photonics to be developed and deployed much more like electronics with both economic and technical benefits.
Fibre optical parametric amplifiers for optical communications
Vladimir Gordienko, Aston University
Fibre optical parametric amplifiers (FOPA) can advance modern optical communications in many ways owing to their unique features. The FOPAs’ virtually wavelength agnostic operation can complement and provide the required flexibility for emerging multi-band optical communication systems, the hollow-core fibres revolution and free-space optical communications. The FOPAs’ ability for noiseless amplification can be advantageous for low signal-to-noise ratio applications. The FOPAs’ employment for all-optical wavelength conversion can pave the way for flexible and elastic optical networks. The FOPAs’ instantaneous response time is useful for ultrafast and transient-free applications. However, there are many challenges towards realisation of the FOPAs’ potential. In this talk I will review our latest achievements in the FOPAs research.
Mode division multiplexing technology: from ground to space
Feng Wen, University of Electronic Science and Technology of China
The mode division multiplexing (MDM) technology is to open up the new dimension of fibers, enabling significantly increasing the capacity of optical fiber communication systems from the space domain. To implement the MDM technology in the real transmission systems, the great efforts on the low-complex multiple-input multiple-output (MIMO) equalization have been made to reduce the cost of compensation operation for the high-capacity MDM systems. The hybrid algorithms, e.g., the transfer learning (TF), the genetic algorithm (GA) and the ant-colony optimization (ACO) modified MIMO equalizers have dramatically reduced the training cost, up to 100% compared to the conventional scheme. Moreover, the MDM technology is also considered in the free-space transmission scenario, such as the laser satellite communications. The wide-open channel brings the big challenge on the MDM implementation. The equalization algorithm as the powerful solution has been used to mitigate the impacts from the wireless laser channel, improving the transmission performance in the space satellite networks.
3.52 Tbps Dynamic Demultiplexing of Low-loss Spatial Modes in Strong Turbulence Using Reconfigurable Photonics
Ultan Daly, University of Glasgow
Free-space optical (FSO) systems can provide high-capacity communication channels for bridging point-to-point links rapidly where fibre connections are unfeasible or for rapidly re-establishing connection in the event of cable breaks. To maximise the data-rate of these systems it is critical to exploit several degrees of freedom, including wavelength, polarisation, and space. The spatial modes are acutely affected by atmospheric turbulence, where the challenges in overcoming the effects of the environment have prevented the wide-scale deployment of Spatial Division Multiplexing (SDM) in FSO. We will introduce a reconfigurable Si-photonics device that can continually measure the optical path and create optimised spatial modes that remain orthogonal after propagation, enabling SDM in 1 km channels with high optical turbulence. The device can measure the transmission matrix at kHz frequencies, fast enough to recover real time measurements of the atmospheric channel that can have a coherence time of up-to ~1ms. A singular value decomposition (SVD) is performed on the measured transmission matrix, determining a unique low-loss orthogonal basis for the channel, which is then demultiplexed by the Si-photonics platform. We aggregate 88-DWDM each with a 10 Gbps NRZ signal on 4-SVD channels, a 3.52 Tbps data rate is achieved for a simulated 1 km channel in high turbulence.