Highly Controllable and Silicon-Compatible Ferroelectric Photovoltaic Synapses for Neuromorphic Computing

Cheng, Shuangshuang; Fan, Zhen; Rao, Jingjing; Hong, Liang; Huang, Qicheng; Tao, Ruiqiang; Hou, Zhipeng; Qin, Minghui; Zeng, Min; Lu, Xubing; Zhou, Guofu; Yuan, Guoliang; Gao, Xingsen; Liu, Junming

doi:10.1016/j.isci.2020.101874

Cited by 37 publications

(29 citation statements)

References 82 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[30,31] This mechanism is complementary to a recently demonstrated photovoltaic read-out of the ferroelectric state, which likewise circumvents the need for Ohmic contacts. [32] Here, we report the observation of tunable electronic properties of ferroelectric switching at microwave frequencies and on the level of a single ferroelectric nanodomain. Utilizing a scanning probe as a top electrode, we locally switched the domain structure of a thin PbZr 0.2 Ti 0.8 O 3 (PZT) film.…”

Section: Introductionmentioning

confidence: 96%

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film

Burns

Tselev

Ievlev

et al. 2021

Adv Elect Materials

View full text Add to dashboard Cite

information is stored in the polarization orientation. A recent resurgence of interest in electronic properties of ferroelectrics, such as polarization controlled tunneling and conductivity of domain walls, potentially enables a new generation of applications utilizing resistive probes of polarization state, such as synaptic junctions, [1][2][3][4][5] domain-wall resistors, and domain-wall logic. [6][7][8][9] At the same time, ferroelectric control over resistive switching may enable neuromorphic devices without structural phase change and filamentary breakdown, potentially leading to much higher energy efficiency.One of the key properties enabling the coupling of ferroelectric and resistive switching is the electronic conductivity of domain walls, [10] which was experimentally demonstrated in numerous materials, [11][12][13][14] beginning with the work of Seidel et al. [15] The intrinsic coupling of domain walls to applied electric field, their nanoscale dimensions, and the flexibility afforded by deterministic control of ferroelectric, ferroelastic, and ferromagnetic structures containing conducting domain walls [16][17][18][19][20] provide great promise for new memory and electronics-engineering concepts, such as racetrack memories [21] or magnetoelectric spin-orbit devices. [22] A notable challenge on the path of ferroresistive switching is that metallic conductivity and ferroelectricity are fundamentally Ferroelectric materials exhibit spontaneous polarization that can be switched by electric field. Beyond traditional applications as nonvolatile capacitive elements, the interplay between polarization and electronic transport in ferroelectric thin films has enabled a path to neuromorphic device applications involving resistive switching. A fundamental challenge, however, is that finite electronic conductivity may introduce considerable power dissipation and perhaps destabilize ferroelectricity itself. Here, tunable microwave frequency electronic response of domain walls injected into ferroelectric lead zirconate titanate (PbZr 0.2 Ti 0.8 O 3 ) on the level of a single nanodomain is revealed. Tunable microwave response is detected through first-order reversal curve spectroscopy combined with scanning microwave impedance microscopy measurements taken near 3 GHz. Contributions of film interfaces to the measured AC conduction through subtractive milling, where the film exhibited improved conduction properties after removal of surface layers, are investigated. Using statistical analysis and finite element modeling, we inferred that the mechanism of tunable microwave conductance is the variable area of the domain wall in the switching volume. These observations open the possibilities for ferroelectric memristors or volatile resistive switches, localized to several tens of nanometers and operating according to well-defined dynamics under an applied field.

show abstract

Section: Introductionmentioning

confidence: 96%

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film

Burns

Tselev

Ievlev

et al. 2021

Adv Elect Materials

View full text Add to dashboard Cite

show abstract

“…[66,75,76] The process referred to as synaptic plasticity can be categorized into short-term and long-term potentiation (STP and LTP) which corresponds to the STM and LTM of the human brain. [77,78] STP is a temporary potentiation of neuronal activities that lasts only for a few minutes or less, whereas LTP is a permanent potentiation that lasts from hours to years. [71,79] Since communication in neurons usually occurs by transmitting data in the form of electrical and electrochemical signals in the range of few tens of milli-volts, [80] a bias voltage as small as 50 mV is applied in our vdW exfoliated ZnO synaptic device.…”

Section: Resultsmentioning

confidence: 99%

“…Energy consumption is one of the critical parameters www.advopticalmat.de that determine the potential of an artificial synapse network. [78] The optical energy (E opt ) for a single pulse event can be calculated using the formula E opt = P × t × S, [105] where P is the illumination power applied for an optical spike with time duration t and S is the active area of the device. Under the excitation wavelength of 365 nm, with a pulse width of 70 ms, the illumination intensity of 0.5 mW cm −2 , and an active area of 250 µm 2 , the energy consumption of our device is obtained as 87.5 pJ.…”

Section: Resultsmentioning

confidence: 99%

A Visible‐Blind Photodetector and Artificial Optoelectronic Synapse Using Liquid‐Metal Exfoliated ZnO Nanosheets

Krishnamurthi

Ahmed

Mohiuddin

et al. 2021

Advanced Optical Materials

View full text Add to dashboard Cite

that offers novel electrical, electronic, and optical properties that are distinct from their bulk counterparts along with mechanical flexibility and high compatibility with state-of-the-art silicon-based platform. [2][3][4] Among prominently studied materials in the last few decades is zinc oxide (ZnO), a highly versatile tunable material. [5] The optoelectronic properties of ZnO in various morphologies have been investigated widely. [6] Due to its strong absorption in the UV region, ZnO is an attractive candidate for visible-blind photodetectors. [7] With a wide bandgap of 3.39 eV, exciton binding energy as large as 60 meV at room temperature, and the ability to undergo a strong quantum confinement effect, atomically thin ZnO promises an excellent platform for optoelectronic applications. [8] Most importantly, oxygen adsorbed onto the surface of ZnO provides low electron densities that can enable low dark current which is ideal for low energy applications. [9] Though thin nanosheets (<20 nm) ZnO has been used as an active layer for various applications, the lack of a reliable and controllable synthesis technique to obtain large area few atoms thin ZnO has prevented the miniaturization of ZnO based optoelectronic devices slowly making the material less competitive compared to other emerging systems relying on atomically thin functional layers. [10][11][12] Also, ZnO and other planar metal-semiconductormetal (MSM) based UV photodetectors developed to date have Atomically thin 2D materials are highly sought for high-performance electronic and optoelectronic devices. Despite being a widely recognized functional material for a plethora of applications, ultra-thin nanosheets of zinc oxide (ZnO) at a millimeter-scale for developing high-performance electronic/optoelectronic devices have not been reported. This has prevented the exploration of electronic and optical properties of ZnO when it is only a few atoms thick. Here, a liquid metal exfoliation technique is used that takes advantage of the van der Waals forces between the interfacial oxide and the chosen substrate to obtain ZnO nanosheets with lateral dimensions in the millimeter scale and thickness down to 5 nm. Their suitability for applications is shown by demonstrating a visible-blind photodetector with high figures of merit as compared to other ZnO morphologies. At extremely low operating bias of 50 mV and low optical intensity of 0.5 mW cm −2 , the ZnO photodetector demonstrates an external quantum efficiency (EQE), responsivity (R), and detectivity (D*) of 4.3 × 10 3 %, 12.64 A W −1 , and 5.81 × 10 15 Jones at a wavelength of 365 nm. The trap-mediated photoresponse in the ZnO nanosheets is further utilized to demonstrate optoelectronic synapses. Versatile synaptic functions of the nervous systems are optically emulated with the ultra-thin ZnO nanosheets.

show abstract

“…Notably, the polarization switching can induce not only the magnitude change but also the sign reversal of photoresponse 19,21 , enabling a single FE-PS to represent both positive and negative weights and hence reducing the hardware overhead for network construction. Moreover, the nonvolativity, high controllability, and ultrafast switching kinetics (<1 ns) of polarization as demonstrated in various ferroelectric memory and neuromorphic devices [29][30][31][32][33][34] , along with the intimate coupling between polarization and photoresponse 35 , endow the FE-PS with good reliability and high write speed. Also noteworthy are the high photosensitivity and ultrashort photoresponse time (<1 ns) of FE-PS 24,25 , allowing a high-speed readout.…”

Section: Introductionmentioning

confidence: 99%

Ferroelectric photosensor network: an advanced hardware solution to real-time machine vision

Cui¹,

Fan²,

Li³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Nowadays the development of machine vision is oriented toward real-time applications such as autonomous driving. This demands a hardware solution with low latency, high energy efficiency, and good reliability. Here, we demonstrate a robust and self-powered in-sensor computing paradigm with a ferroelectric photosensor network (FE-PS-NET). The FE-PS-NET, constituted by ferroelectric photosensors (FE-PSs) with tunable photoresponsivities, is capable of simultaneously capturing and processing images. In each FE-PS, self-powered photovoltaic responses, modulated by remanent polarization of an epitaxial ferroelectric Pb(Zr0.2Ti0.8)O3 layer, show not only multiple nonvolatile levels but also a sign reversibility, enabling the representation of a signed weight in a single device and hence reducing the hardware overhead for network construction. With multiple FE-PSs wired together, the FE-PS-NET acts on its own as an artificial neural network. It is demonstrated that an in situ multiply-accumulate operation between an input image and a stored photoresponsivity matrix is available in our FE-PS-NET hardware. The FE-PS-NET hardware is faultlessly competent for real-time image processing functionalities, including binary pattern classification with an accuracy of 100% and edge detection with an F-Measure of 95.2%. This study highlights the great potential of ferroelectric photovoltaics as the hardware basis of real-time machine vision.

show abstract

Highly Controllable and Silicon-Compatible Ferroelectric Photovoltaic Synapses for Neuromorphic Computing

Cited by 37 publications

References 82 publications

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film

A Visible‐Blind Photodetector and Artificial Optoelectronic Synapse Using Liquid‐Metal Exfoliated ZnO Nanosheets

Ferroelectric photosensor network: an advanced hardware solution to real-time machine vision

Contact Info

Product

Resources

About

Highly Controllable and Silicon-Compatible Ferroelectric Photovoltaic Synapses for Neuromorphic Computing

Cited by 37 publications

References 82 publications

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr0.2Ti0.8O3 Thin Film

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr0.2Ti0.8O3 Thin Film

A Visible‐Blind Photodetector and Artificial Optoelectronic Synapse Using Liquid‐Metal Exfoliated ZnO Nanosheets

Ferroelectric photosensor network: an advanced hardware solution to real-time machine vision

Contact Info

Product

Resources

About

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film

Tunable Microwave Conductance of Nanodomains in Ferroelectric PbZr_0.2Ti_0.8O₃ Thin Film