An Electronic Synapse Device Based on Metal Oxide Resistive Switching Memory for Neuromorphic Computation

Yu, Shimeng; Wu, Yang; Jeyasingh, Rakesh; Kuzum, Duygu; Wong, Hang

doi:10.1109/ted.2011.2147791

Cited by 720 publications

(491 citation statements)

References 33 publications

Supporting

Mentioning

479

Contrasting

Unclassified

Order By: Relevance

“…Various types of memristors based on ionic drift (drift-type memristor) [4][5][6][7][8] have recently been utilized for this purpose in neuromorphic architectures [9][10][11][12][13][14][15] . Although qualitative synaptic functionality has been demonstrated, the fast switching and non-volatility of drift memristors optimized for memory applications do not faithfully replicate the nature of plasticity.…”

mentioning

confidence: 99%

Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing

et al. 2016

View full text Add to dashboard Cite

The accumulation and extrusion of Ca 2+ in the pre-and postsynaptic compartments play a critical role in initiating plastic changes in biological synapses. To emulate this fundamental process in electronic devices, we developed diffusive Ag-in-oxide memristors with a temporal response during and after stimulation similar to that of the synaptic Ca 2+ dynamics. In situ high-resolution transmission electron microscopy and nanoparticle dynamics simulations both demonstrate that Ag atoms disperse under electrical bias and regroup spontaneously under zero bias because of interfacial energy 2 minimization, closely resembling synaptic influx and extrusion of Ca 2+ , respectively.The diffusive memristor and its dynamics enable a direct emulation of both short-and long-term plasticity of biological synapses and represent a major advancement in hardware implementation of neuromorphic functionalities.CMOS circuits have been employed to mimic synaptic Ca 2+ dynamics, but three-terminal devices bear limited resemblance to bio-counterparts at the mechanism level and require significant numbers and complex circuits to simulate synaptic behavior [1][2][3] . A substantial reduction in footprint, complexity and energy consumption can be achieved by building a two-terminal circuit element, such as a memristor directly incorporating Ca 2+ -like dynamics.Various types of memristors based on ionic drift (drift-type memristor) 4-8 have recently been utilized for this purpose in neuromorphic architectures [9][10][11][12][13][14][15] . Although qualitative synaptic functionality has been demonstrated, the fast switching and non-volatility of drift memristors optimized for memory applications do not faithfully replicate the nature of plasticity. Similar issues also exist in MOS-based memristor emulators [16][17][18] , although they are capable of simulating a variety of synaptic functions including spike-timing-dependent plasticity (STDP). Recently, Lu's group adopted second-order drift memristors to approximate the Ca 2+ dynamics of chemical synapses by utilizing thermal dissipation 19 or mobility decay 20 , which successfully demonstrated STDP with non-overlapping spikes and other synaptic functions, representing a significant step towards bio-realistic synaptic devices. This approach features repeatability and simplicity, but the significant differences of the dynamical response from actual synapses limit the fidelity and variety of desired synaptic functions. A device with similar physical behavior as the biological Ca 2+ dynamics would enable improved emulation of synaptic function and broad applications to neuromorphic computing. Here we report such an emulator, which is a memristor based on metal atom 3 diffusion and spontaneous nanoparticle formation, as determined by in situ high-resolution transmission electron microscopy (HRTEM) and nanoparticle dynamics simulations. The dynamical properties of the diffusive memristors were confirmed to be functionally equivalent to Ca 2+ in bio-synapses, and their operating characteri...

show abstract

mentioning

confidence: 99%

Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing

et al. 2016

View full text Add to dashboard Cite

show abstract

“…In the beginning, synaptic functions were emulated by complementary metal oxide semiconductor (CMOS) neuromorphic circuits, but such CMOS circuits consumed substantially more energy than a biological synapse, and it is hard to scale up the circuits to a size comparable with the brain 5 . Recently, resistive switching memory, memristors or atomic switch has been investigated in biologically inspired neuromorphic circuits [6][7][8][9][10][11] . Important synaptic learning rules such as spike-timing-dependent plasticity (STDP) and short-term memory to long-term memory transition have been demonstrated.…”

mentioning

confidence: 99%

Artificial synapse network on inorganic proton conductor for neuromorphic systems

et al. 2014

View full text Add to dashboard Cite

The basic units in our brain are neurons, and each neuron has more than 1,000 synapse connections. Synapse is the basic structure for information transfer in an ever-changing manner, and short-term plasticity allows synapses to perform critical computational functions in neural circuits. Therefore, the major challenge for the hardware implementation of neuromorphic computation is to develop artificial synapse network. Here in-plane lateral-coupled oxide-based artificial synapse network coupled by proton neurotransmitters are selfassembled on glass substrates at room-temperature. A strong lateral modulation is observed due to the proton-related electrical-double-layer effect. Short-term plasticity behaviours, including paired-pulse facilitation, dynamic filtering and spatiotemporally correlated signal processing are mimicked. Such laterally coupled oxide-based protonic/electronic hybrid artificial synapse network proposed here is interesting for building future neuromorphic systems.

show abstract

“…Then the voltage is swept back (-3 V 0 V), and the HRS current usually shows super-linear I-V relation. Other performance parameters of our samples such as switching speed (~10 ns), endurance (~10 6 cycles), retention (>2 hours @ 100 °C) and multilevel resistance states capability by controlling reset voltages were reported in previous publications [28][29]. Fig.…”

Section: Hrsmentioning

confidence: 78%

Characterization of low-frequency noise in the resistive switching of transition metal oxide HfO2

Jeyasingh

et al. 2012

Phys. Rev. B

Self Cite

View full text Add to dashboard Cite

-Low frequency noise measurements were performed on HfO 2 based bipolar resistive switching memory devices. A 1/f α DC noise power spectral density was observed with α~1 for low resistance state and α~2 for high resistance state. We developed an electron tunneling model to elucidate the conduction process which showed that the 1/f α behavior was due to the distribution of relaxation times of electron tunneling between the electrodes and the traps in the conducting filaments. The transition of the slope index α from 1 to 2 at a certain cutoff frequency indicates that there is a tunneling gap formed between electrodes and the residual of the conductive filaments in the high resistance state.Keywords -low frequency noise, resistive switching (RRAM), transition metal oxide, electron tunneling Currently, transition metal oxide based resistive switching memory is extensively studied as one of the most competitive candidates for future non-volatile memory applications due to its simple structure, fast switching speed, great scalability, and compatibility with silicon complementary metal-oxidesemiconductor (CMOS) technology [9][10][11]. The mechanism of resistive switching phenomenon in oxides is usually attributed to the formation/rupture of conductive filaments (CFs) which may consist of oxygen vacancies or metal precipitates [12]. The set process from high resistance state (HRS) to low resistance state (LRS) is interpreted as a dielectric soft breakdown associated with the migration of oxygen ions toward the anode, leaving behind the oxygen vacancies in the bulk oxide to form CFs connecting both electrodes [13]. To reset from LRS to HRS, there are two modes: in the unipolar reset (the reset occurs at the same polarity as the set), Joule-heating-assisted diffusion of oxygen ions from anode and surrounding oxides rupture the CFs by recombination with oxygen vacancies or re-oxidization of the metal precipitates [14]. In the bipolar reset (the reset occurs at the opposite polarity as the set), electric-field-assisted drift of oxygen ions from the oxygen reservoir rupture the CFs [15]. In the bipolar switching mode, oxidizable electrode materials such as Ti, TiN, TaN are usually used to serve as the oxygen reservoir providing the oxygen ions during the reset process [16]. The device used in this paper exhibits the bipolar switching mode. Although such a phenomenological physical switching picture above has been proposed, the detailed characterization and quantitative theoretical analysis of the conduction mechanism is still lacking in the literature.Low frequency noise (LFN) measurement is a technique that can electrically characterize the trapassisted conduction process in dielectrics [17][18][19] NiO [22] (in this paper f refers to the frequency). In this work, we perform LFN measurement on HfO 2 based resistive switching memory to investigate its conduction and switching mechanism. HfO 2 is chosen because HfO 2 based devices exhibit desirable properties such as ultra-fast switching speed (

show abstract

An Electronic Synapse Device Based on Metal Oxide Resistive Switching Memory for Neuromorphic Computation

Cited by 720 publications

References 33 publications

Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing

Memristors with diffusive dynamics as synaptic emulators for neuromorphic computing

Artificial synapse network on inorganic proton conductor for neuromorphic systems

Characterization of low-frequency noise in the resistive switching of transition metal oxide HfO2

Contact Info

Product

Resources

About