Toward Nonvolatile Switching in Silicon Photonic Devices

Laser & Photonics Review, Journal Year: 2021, Volume and Issue: 15(6)

Published: May 5, 2021

Abstract Nonvolatile switching is still a missing functionality in current mainstream silicon photonics complementary metal‐oxide‐semiconductor platforms. Fundamentally, nonvolatile stands for the ability to switch between two or more photonic states reversibly without needing additional energy hold each state. Therefore, such feature may push one step further potential of by offering new ways achieving reconfigurability with ultrasmall consumption. Here, detailed review developments that enable waveguide devices provided. Nonvolatility successfully demonstrated either based on device engineering hybrid integration waveguides materials exhibiting unique optical properties. Furthermore, several approaches high evolving toward behavior enhanced performance are also being explored. In most cases, many development steps necessary ensure reliable devices. However, this research field expected progress coming years boosted and emerging applications benefiting from functionality, as paradigms computing advanced reconfigurable circuits programmable systems.

Language: Английский

---

Resonant multilevel optical switching with phase change material GST

Nanophotonics, Journal Year: 2022, Volume and Issue: 11(15), P. 3437 - 3446

Published: June 22, 2022

We demonstrate a multilevel optical memristive switch based on silicon Fabry-Perot resonator. The resonator is constructed by pair of waveguide Bragg gratings at the ends multimode interferometer (MMI) covered with sub-micrometer-size Ge

Language: Английский

---

Resistive Switching Properties in Memristors for Optoelectronic Synaptic Memristors: Deposition Techniques, Key Performance Parameters, and Applications

ACS Applied Electronic Materials, Journal Year: 2023, Volume and Issue: 6(1), P. 73 - 119

Published: Dec. 29, 2023

Due to the fast evolution of information technology, high-speed and scalable memory devices are being investigated for data storage data-driven computation. Resistive switching random access (RRAM) is one very popular types memristors, because its quick program/erase speed high density as a result simple two-terminal structure, low power consumption, cost fabrication. In this review, market promising memristor-based memories smart systems examined. The described classified according their I–V behavior underline physical mechanisms. various filament mechanisms RRAM memristor, including valence change mechanism (VCM), electrochemical metallization (ECM), thermochemical (TCM), especially highlighted in structures. performance resistive with different reported electrode materials layers summarized. This study provides detailed review deposition techniques used surface modification coating. two major areas interest, chemical vapor influence parameter on switching, discussed detail. prospective applications memristors fields such security, neuromorphic computing, non-volatile logic artificial intelligence addressed briefly well future outlook.

Language: Английский

---

Non-volatile tunable optics by design: From chalcogenide phase-change materials to device structures

Materials Today, Journal Year: 2023, Volume and Issue: 68, P. 334 - 355

Published: Aug. 18, 2023

Language: Английский

---

Towards low loss non-volatile phase change materials in mid index waveguides

Neuromorphic Computing and Engineering, Journal Year: 2021, Volume and Issue: 1(1), P. 014004 - 014004

Published: July 16, 2021

Abstract Photonic integrated circuits currently use platform intrinsic thermo-optic and electro-optic effects to implement dynamic functions such as switching, modulation other processing. Currently, there is a drive field programmable photonic circuits, need which only magnified by new neuromorphic quantum computing applications. The most promising non-volatile components employ phase change materials GST GSST, had their origin in electronic memory. However, the optical domain, these compounds introduce significant losses potentially preventing large number of Here, we evaluate two newly introduced low loss materials, Sb 2 S 3 Se , on silicon nitride for future implementation computing. We focus study Mach–Zehnder interferometers that operate at O C bands demonstrate performance system. Our measurements show an insertion below 0.04 dB μm −1 lower than 0.09 cladded devices both amorphous crystalline phases. effective refractive index contrast SiNx was measured be 0.05 1310 nm 0.02 1550 nm, whereas it 0.03 highlighting device.

Language: Английский

---

Comparison of the phase change process in a GST-loaded silicon waveguide and MMI

Optics Express, Journal Year: 2021, Volume and Issue: 29(3), P. 3503 - 3503

Published: Jan. 9, 2021

In the past decades, silicon photonic integrated circuits (PICs) have been considered a promising approach to solve bandwidth bottleneck in optical communications and interconnections. Despite rapid advances, large-scale PICs still face series of technical challenges, such as large footprint, high power consumption, lack memory, resulting from active tuning methods used control waves. These challenges can be partially addressed by combining chalcogenide phase change materials (PCMs) Ge 2 Sb Te ­5 (GST) with photonics, especially applicable reconfigurable circuit applications due nonvolatile nature GST. We systematically investigate process induced electrical pulses GST-loaded waveguide multimode interferometer. Using pulse excitation amorphize GST has clear advantage terms operation speed energy efficiency, while is more suitable for integration because it does not require complex routing. This study helps us better understand push forward further development Si-GST hybrid platform, bringing new potential applications.

Language: Английский

---

Phase-change materials for energy-efficient photonic memory and computing

MRS Bulletin, Journal Year: 2022, Volume and Issue: 47(5), P. 502 - 510

Published: May 1, 2022

Abstract Neuromorphic algorithms achieve remarkable performance milestones in tasks where humans have traditionally excelled. The breadth of data generated by these paradigms is, however, unsustainable conventional computing chips. In-memory hardware aims to mimic biological neural networks and has emerged as a viable path overcoming fundamental limitations the von Neumann architecture. By eliminating latency energy losses associated with transferring between memory central processing unit (CPU), systems promise improve on both speed energy. Photonic implementations using on-chip, nonvolatile memories are particularly promising they aim deliver energy-efficient, high-speed, high-density within photonic multiplexing advantages optics. In this article, we overview recent progress direction that integrates phase-change material (PCM) elements integrated optoelectronics. We compare performances PCM devices optoelectronic programming schemes show consumption can be significantly reduced 60 pJ picosecond (ps) optical pulse plasmonic nanogap approaching 1 GHz. With energy-efficient waveguide memories, concepts in-memory implemented based crossbar arrays. Compared digital electronic accelerators: application-specific circuits (ASICs) graphics units (GPUs), cores 1−3 orders higher compute density efficiency, although much more work toward commercialization is still required. Graphical abstract

Language: Английский

---

Fabrication and integration of photonic devices for phase-change memory and neuromorphic computing

International Journal of Extreme Manufacturing, Journal Year: 2023, Volume and Issue: 6(2), P. 022001 - 022001

Published: Dec. 13, 2023

Abstract In the past decade, there has been tremendous progress in integrating chalcogenide phase-change materials (PCMs) on silicon photonic platform for non-volatile memory to neuromorphic in-memory computing applications. particular, these non von Neumann computational elements and systems benefit from mass manufacturing of integrated circuits (PICs) 8-inch wafers using a 130 nm complementary metal-oxide semiconductor line. Chip based deep-ultraviolet lithography electron-beam enables rapid prototyping PICs, which can be with high-quality PCMs wafer-scale sputtering technique as back-end-of-line process. this article, we present an overview recent advances waveguide PCM cells, functional devices, systems, emphasis fabrication integration processes attain state-of-the-art device performance. After short discuss properties layer well light guiding layer, namely, germanium platforms. Next, cleanroom flow devices thin films nanowires, waveguides plasmonic microheaters electrothermal switching mixed-mode operation. Finally, photonic–electronic is reviewed. These consist arrays associative learning, matrix-vector multiplication, pattern recognition. With large-scale integration, neuromorphicphotonic paradigm holds promise outperform digital electronic accelerators by taking advantages ultra-high bandwidth, high speed, energy-efficient operation running machine learning algorithms.

Language: Английский

---

Low‐Temperature Sputtered Ultralow‐Loss Silicon Nitride for Hybrid Photonic Integration

Laser & Photonics Review, Journal Year: 2023, Volume and Issue: 18(4)

Published: Dec. 31, 2023

Abstract Silicon‐nitride‐on‐insulator (Si 3 N 4 ) photonic circuits have seen tremendous advances in many applications, such as on‐chip frequency combs, Lidar, telecommunications, and spectroscopy. So far, the best film quality has been achieved with low pressure chemical vapor deposition (LPCVD) high‐temperature annealing (1200°C). However, high processing temperatures pose challenges to cointegration of Si pre‐processed silicon electronic devices, lithium niobate on insulator (LNOI), Ge‐on‐Si photodiodes. This limits LPCVD a front‐end‐of‐line process. Here, ultralow‐loss photonics based room‐temperature reactive sputtering is demonstrated. Propagation losses 5.4 dB m −1 after 400°C 3.5 800°C are achieved, enabling ring resonators highest optical factors > 10 million an average factor 7.5 million. To knowledge, these lowest propagation temperature . ultralow loss enables generation microresonator soliton combs threshold powers 1.1 mW. The introduced process offers full complementary metal oxide semiconductor (CMOS) compatibility front‐end electronics photonics. could enable hybrid 3D integration waveguides integrated lasers insulator.

Language: Английский

---

Hybrid photonic integrated circuits for neuromorphic computing [Invited]

Optical Materials Express, Journal Year: 2023, Volume and Issue: 13(12), P. 3553 - 3553

Published: Nov. 13, 2023

The burgeoning of artificial intelligence has brought great convenience to people’s lives as large-scale computational models have emerged. Artificial intelligence-related applications, such autonomous driving, medical diagnosis, and speech recognition, experienced remarkable progress in recent years; however, systems require vast amounts data for accurate inference reliable performance, presenting challenges both speed power consumption. Neuromorphic computing based on photonic integrated circuits (PICs) is currently a subject interest achieve high-speed, energy-efficient, low-latency processing alleviate some these challenges. Herein, we present an overview the current platforms available, materials which potential be with PICs further hybrid devices neuromorphic computing.

Language: Английский

---