Using a gradient echo memory, we experimentally demonstrate cross phase modulation (XPM) between two optical pulses; one stored and one freely propagating through the memory medium. We explain how this idea can be extended to enable substantial nonlinear interaction between two single photons that are both stored in the memory. We present semi-classical and quantum simulations along with a proposed experimental scheme to demonstrate the feasibility of achieving large XPM at single photon level.

Memory-Enhanced Noiseless Cross Phase Modulation

All-Dielectric Metamaterial Fabrication Techniques

https://onlinelibrary.wiley.com/doi/pdf/10.1002/inf2.12014

Low-dimensional nanomaterial/Si heterostructure-based photodetectors

Deep learning is revolutionizing many aspects of our society, addressing a wide variety of decision-making tasks, from image classification to autonomous vehicle control. Matrix multiplication is an essential and computationally intensive step of deep-learning calculations. The computational complexity of deep neural networks requires dedicated hardware accelerators for additional processing throughput and improved energy efficiency in order to enable scaling to larger networks in the upcoming applications. Silicon photonics is a promising platform for hardware acceleration due to recent advances in CMOS-compatible manufacturing capabilities, which enable efficient exploitation of the inherent parallelism of optics. This article provides a detailed description of recent implementations in the relatively new and promising platform of silicon photonics for deep learning. Opportunities for multiwavelength microring silicon photonic architectures codesigned with field-programmable gate array (FPGA) for pre- and postprocessing are presented. The detailed analysis of a silicon photonic integrated circuit shows that a codesigned implementation based on the decomposition of large matrix-vector multiplication into smaller instances and the use of nonnegative weights could significantly simplify the photonic implementation of the matrix multiplier and allow increased scalability. We conclude this article by presenting an overview and a detailed analysis of design parameters. Insights for ways forward are explored.

Silicon Photonics Codesign for Deep Learning

Integrated optical signal processors, in combination with conventional electrical signal processors, are envisioned to open a path to a new generation of signal processing hardware platform that allows for significant improvement in processing bandwidth, latency, and power efficiency. With its well-known features and potential, silicon photonics is considered as a favorable candidate for the device implementation, particularly with high circuit complexity, and hence has been the focus of the study. As an outlook from the previous discussions on such processors, we are considering new building blocks in the silicon photonics platform for further extending the processor capabilities and adding practical features, particularly the miniaturized devices that enable ultra-dense integration of complex circuits into such processor chips. As enlightening examples, we review here our recent contribution together with representative works from other groups of compact designs of silicon photonics devices that enrich functionalities of processor building blocks such as multiplexing, polarization handling, and optical I/Os. The results shown in this review reflect the significance and maturity of the state-of-the-art photonic fabrication technology and contribute to the implementation of high-capacity, general-purpose optical signal processing functionalities on the chip scale.

Miniaturized Silicon Photonics Devices for Integrated Optical Signal Processors

Abstract Multimode silicon photonics is attracting more and more attention because the introduction of higher-order modes makes it possible to increase the channel number for data transmission in mode-division-multiplexed (MDM) systems as well as improve the flexibility of device designs. On the other hand, the design of multimode silicon photonic devices becomes very different compared with the traditional case with the fundamental mode only. Since not only the fundamental mode but also the higher-order modes are involved, one of the most important things for multimode silicon photonics is the realization of effective mode manipulation, which is not difficult, fortunately because the mode dispersion in multimode silicon optical waveguide is very strong. Great progresses have been achieved on multimode silicon photonics in the past years. In this paper, a review of the recent progresses of the representative multimode silicon photonic devices and circuits is given. The first part reviews multimode silicon photonics for MDM systems, including on-chip multichannel mode (de)multiplexers, multimode waveguide bends, multimode waveguide crossings, reconfigurable multimode silicon photonic integrated circuits, multimode chip-fiber couplers, etc. In the second part, we give a discussion about the higher-order mode-assisted silicon photonic devices, including on-chip polarization-handling devices with higher-order modes, add-drop optical filters based on multimode Bragg gratings, and some emerging applications.

https://www.degruyter.com/downloadpdf/j/nanoph.2019.8.issue-2/nanoph-2018-0161/nanoph-2018-0161.xml

Multimode silicon photonics

Silicon photonic filters

A silicon-based on-chip add-drop optical filter is realized for the wavelength-band of 2 μ m by integrating a multimode waveguide grating (MWG) and a pair of two-channel mode (de)multiplexers. The realized add-drop optical filter has low loss and high sidelobe suppression ratio in a broad band, achieved by optimizing the corrugation depth, the apodization strength, as well as the grating length. The measurement results agree very well with the theoretical simulation results. The demonstrated ∼100 μ m-long MWG has a large 0.5 dB bandwidth of 6∼26 nm and a sidelobe suppression ratio of 22∼27 dB. The thermal-tunability of the present optical filter is also demonstrated experimentally, and the measured tuning efficiency is about 0.16 nm/mW.

Silicon Multimode Waveguide Grating Filter at 2 μ m

A silicon-based four-channel coarse wavelength- division multiplexing (CWDM) (de)multiplexer working in the O-band is proposed and realized. The present on-chip four-channel CWDM (de)multiplexer is composed of four cascaded optical filters with flat-top spectral responses, which are realized by combining a multimode waveguide grating (MWG) and a two-channel mode (de)multiplexer. For the realized flat-top CWDM (de)multiplexer with a ~18-nm channel-spacing, the measured 1-dB bandwidth is as large as 15 nm, which is ~75% of the channel spacing. The thermal sensitivity is also reduced from ~85 pm/°C to 46 pm/°C by introducing the compensation provided by an SU-8 upper-cladding with a negative thermo-optic coefficient.

Four-Channel CWDM (de)Multiplexers Using Cascaded Multimode Waveguide Gratings

A low-loss add-drop microring resonator (MRR) with an ultra-large free spectral range (FSR) is demonstrated by introducing an ultra-sharp multimode waveguide bend and bent asymmetrical directional couplers (ADCs). The multimode microring waveguide is introduced to achieve a low bent loss, even with a small radius (e.g., R = 0.8 μm). The bent ADCs are used to suppress the resonance of higher-order modes. For the fabricated device, the transmission at the drop port has a narrow 3 dB-bandwidth of 0.8 nm and a low excess loss of 1.8 dB. A record large FSR of 93 nm is achieved to the best of our knowledge.

Dajian Liu

Papers

Multimode silicon photonics

Silicon photonic filters

Silicon Multimode Waveguide Grating Filter at 2 μ m

Four-Channel CWDM (de)Multiplexers Using Cascaded Multimode Waveguide Gratings

Submicron-resonator-based add-drop optical filter with an ultra-large free spectral range.