Toward non-volatile photonic memory: concept, material and design

174

127

and constant data movement are needed while working. In the human brain, huge and complex neural networks composed of gigantic amounts of neurons massively interconnected by an even larger number of synapses are in charge of computing and memory. Unlike a digital computer, information storage and processing happen at the same time in the brain. [3] Artificial intelligence (AI) that could rival biological neural networks is being continuously pursued by human being. There are two approaches to implement AI: one is the software simulation of neural networks with the aid of digital computers, another is developing hardware-integrated circuits of neural networks. In fact, AI based on the software simulation is evolving very fast thanks to the advances in the development of algorithm in recent years, and it even outperforms the human brain in specific complex tasks, such as playing the game of Go. [4] However, software-based AI suffers from critical issues of enormous power consumption and massive data throughput. Hardware-based AI systems are more efficient in terms of speed and power consumption; however, complementary metal oxide semiconductor (CMOS) transistors are inherently different from the basic building blocks of biological neural networks, neurons, and synapses, in terms of operation dynamic and behavior. A circuit consisting of at least ten transistors are required to realize the function of one biological synapse, which is impractical for the implementation of large networks, not to mention the human brain (roughly 10 11 neurons interconnected by ≈10 15 synapses). [5,6] Fortunately, the emergence of memristive devices with innate dynamics resembling biological synapses and neurons provides a feasible and simple way to build hardware-based bio-inspired computing systems. [7,8] For instance, only one compact memristive device with a metalinsulator-metal (MIM) sandwich structure is sufficient to reproduce some functions of a biological synapse. Moreover, the vector-matrix multiplication, which is believed to be the most data-intensive work in neural networks, can be naturally performed in a cross-bar array of memristive devices on the physical level based on Ohm and Kirchhoff laws. [9,10] These features endow the memristive devices ideally suited to realize highly efficient bioinspired computing systems in hardware.To realize highly efficient neuromorphic computing that is comparable to biological counterparts, bioinspired computing systems, consisting of biorealistic artificial synapses and neurons, are developed with memristive devices with native dynamics resembling biological synapses and neurons. Tremendous materials and devices have been successfully used to emulate diverse functions of synapses, as well as neurons, in the last decade. Herein, approaches to realize certain synaptic or neuronal functions are introduced with state-of-art experimental demonstrations. First, the dynamics and working principles of biological synapses and neurons are briefly presented to provide guidance for developing biore...

Section: Wwwadvelectronicmatdementioning

confidence: 99%

Memristive Synapses and Neurons for Bioinspired Computing

Yang

Huang

Guo

2019

174

127

“…One aspect is that synapses in the brain are living and growing, that is, not only synaptic weight but also synaptic plasticity itself should be adaptable and developmental. [15] Recent progresses for the target have been made by introducing the electrical/optical third terminal into the memristive devices, [16][17][18][19][20] most of which are based on inorganic semiconductors. The other aspect is that local synaptic plasticity is often accompanied by remote regulation such as homeostatic phenomena in the brain.…”

Section: Selective Uv-gating Organic Memtransistors With Modulable Lementioning

confidence: 99%

Selective UV‐Gating Organic Memtransistors with Modulable Levels of Synaptic Plasticity

Zhong

Gao

et al. 2019

Targeting the emulation of diverse synapses in the brain, organic thin‐film memtransistors gated remotely by UV light are achieved and analytically modeled. They show continuously and reversibly variable device states, and the degree of the variability can also be controlled by selective UV gating. The modulable memristive characteristics cover a wide range of four orders in reading current, all with excellent state retention upon reading or power‐off. The UV‐gating effect stems from the UV‐enhanced cumulative charge trapping/detrapping in the polymer electret layer, expressed as a linear relationship between the characteristic trapping frequency and UV power density. The organic memtransistors are utilized to successfully mimic various synapses at newborn, young, and mature stages that have different levels of synaptic plasticity, as well as the synaptic functions of paired‐pulse facilitation and paired‐pulse depression. The realization of versatile synaptic emulation in a single device may pave the way toward organic‐based neuromorphic circuits, where UV gating could be used for remote regulation of circuit learning rate.

“…[1][2][3][4][5][6] These memories based on the electric programming and erasing have been extensively studied while further development of their practical application is restricted by the high programming/ erasing voltages, which motivates the introduce of the light programming as an exactly novel independent programming method. [1][2][3][4][5][6] These memories based on the electric programming and erasing have been extensively studied while further development of their practical application is restricted by the high programming/ erasing voltages, which motivates the introduce of the light programming as an exactly novel independent programming method.…”

mentioning

confidence: 99%

“…[1][2][3][4][5][6] These memories based on the electric programming and erasing have been extensively studied while further development of their practical application is restricted by the high programming/ erasing voltages, which motivates the introduce of the light programming as an exactly novel independent programming method. [1] However, organic phototransistor memory (OPTM) currently showed unsatisfying performance such as small memory window and short retention time compared with memories with electric programming and erasing. [1] However, organic phototransistor memory (OPTM) currently showed unsatisfying performance such as small memory window and short retention time compared with memories with electric programming and erasing.…”

mentioning

confidence: 99%

“…[7][8][9][10][11] The light programming offers a noncontact programming method and is orthogonal to the electrical read-out process, which provides efficient way to achieve higher accuracy and capacity (multilevel storage) with the high discrepancies between different storage levels. [1,12,13] The basic working mechanism of photonic memories is that the photo-induced charge carriers are trapped into or detrapped from the charge trapping layer across the interface between the active layer and trapping layer during the programming/erasing process, [9] resulting in a shift of threshold voltage during the reading process. Moreover, in order to generate sufficient photo-induced charge carriers and subsequent trapped minority charges for large memory window, sufficient photon energy, light intensity (3.5−188 mW cm −2 ), and illumination time (1 s to continuously) was required, which also seriously hindered further development and application of photonic memory.…”

mentioning

confidence: 99%

See 1 more Smart Citation

High‐Performance All‐Inorganic Perovskite‐Quantum‐Dot‐Based Flexible Organic Phototransistor Memory with Architecture Design

Yang

Liu

et al. 2019

floating gate organic field effect transistors (OFET)-based memory has been considered to have the most application prospects. [1][2][3][4][5][6] These memories based on the electric programming and erasing have been extensively studied while further development of their practical application is restricted by the high programming/ erasing voltages, which motivates the introduce of the light programming as an exactly novel independent programming method. [7][8][9][10][11] The light programming offers a noncontact programming method and is orthogonal to the electrical read-out process, which provides efficient way to achieve higher accuracy and capacity (multilevel storage) with the high discrepancies between different storage levels. [1] However, organic phototransistor memory (OPTM) currently showed unsatisfying performance such as small memory window and short retention time compared with memories with electric programming and erasing. Moreover, in order to generate sufficient photo-induced charge carriers and subsequent trapped minority charges for large memory window, sufficient photon energy, light intensity (3.5−188 mW cm −2 ), and illumination time (1 s to continuously) was required, which also seriously hindered further development and application of photonic memory. [1,12,13] The basic working mechanism of photonic memories is that the photo-induced charge carriers are trapped into or detrapped from the charge trapping layer across the interface between the active layer and trapping layer during the programming/erasing process, [9] resulting in a shift of threshold voltage during the reading process. Therefore, the performance of OPTM significantly depends upon the non-volatile bistability of the charge-trapping layer and the charge transfer process during the charge-trapping and reading process. [14] However, majority researches of light programming memory currently focused on the design of the charge-trapping layer to increase number of discrete charge-trapping sites such as metallic or semiconducting nanoparticles to improve the memory performance. [15][16][17][18] Little attention has been given to two key factors, "trapped carriers effect" and "interface effect" which were related to the charge transfer process and notably impacted the A novel organic phototransistor memory (OPTM) with architecture design is fabricated with all-inorganic perovskite quantum dots (QDs) as charge trapping layer. The novel architecture enables ultrashort channel length and vertical the charge transfer to effectively suppress both trapped carriers effect and interface effect, two key factors which impact photonic memory performance. Additionally, perovskite QDs are introduced to function as both the charge trapping layer and a UV-light-responsive material that emits visible light that can be subsequently absorbed by the active layer, leading to efficient utilization of UV light to generate trapped charges. OPTM exhibits excellent memory properties under significantly improved light conditions, which is superior to previou...