Article
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Neuromorphic RRAM circuits typically need currents of several mA when many binary memristive devices are activated at the same time. This is due to the low resistance state of these devices, which increases the power consumption and limits the scalability. To overcome this limitation, it is vital to investigate how to minimize the amplitude of the read-out inference pulses sent through the crossbar lines. However, the amplitude of such inference voltage pulses will become limited by the offset voltage of read-out circuits. This paper presents a three-stage calibration circuit to compensate for offset voltage in the wordlines of a memristor-array read-out system. The proposed calibration scheme is based on adjusting the bulk voltage of one of the input differential pair MOSFETs by means of a switchable cascade of resistor ladders. This renders the possibility to obtain calibration voltage steps less than 0.1 mV by cascading a few number of stages, whose results are only limited by mismatch, temperature, electrical noise and other fabrication defects. The system is built using HfO2-based binary memristive synaptic devices on top of a 130-nm CMOS technology. Layout-extracted simulations considering technology corners, PVT variations and electrical noise are shown to validate the presented calibration scheme.

No full-text available

Request Full-text Paper PDF

Request the article directly
from the authors on ResearchGate.

High density spin-transfer torque (STT)-MRAM based on cross-point architecture
  • W Zhao
  • S Chaudhuri
  • C Accoto
  • J.-O Klein
  • D Ravelosona
  • C Chappert
  • P Mazoyer
W. Zhao, S. Chaudhuri, C. Accoto, J.-O. Klein, D. Ravelosona, C. Chappert, P. Mazoyer, High density spin-transfer torque (STT)-MRAM based on cross-point architecture, 2012 4th IEEE International Memory Workshop (IMW), 2012, p. 4.
HfO 2 -based OxRAM devices as synapses for convolutional neural networks
  • D Garbin
  • E Vianello
  • O Bicher
  • Q Rafhay
  • C Gamrat
  • G Ghibaudo
  • B Desalvo
  • L Perniola
D. Garbin, E. Vianello, O. Bicher, Q. Rafhay, C. Gamrat, G. Ghibaudo, B. DeSalvo, L. Perniola, HfO 2 -based OxRAM devices as synapses for convolutional neural networks, IEEE Transactions on Electron Devices 62 (8) (August 2015) 2494-2501.
The stochastic I-Pot: a circuit block for programming bias currents
  • R Serrano-Gotarredona
  • L Camuñas-Mesa
  • T Serrano-Gotarredona
  • J A Leñerobardallo
  • B Linares-Barranco
R. Serrano-Gotarredona, L. Camuñas-Mesa, T. Serrano-Gotarredona, J.A. LeñeroBardallo, B. Linares-Barranco, The stochastic I-Pot: a circuit block for programming bias currents, IEEE Trans. Circuits Syst. Express Briefs 54 (9) (Sept. 2007) 760-764.
  • C Mohan
C. Mohan et al. Microelectronic Engineering 198 (2018) 35-47
... When each wordline in the crossbar is simultaneously driven by inference spike pulses, the full memristor array can become active. This makes a majority of them to drive few mA in their LRS, which in turn makes power dissipation severe and limits the maximum crossbar size and the driving capability of the peripheral circuit [24]. ...
Article
Full-text available
Several analog and digital brain-inspired electronic systems have been recently proposed as dedicated solutions for fast simulations of spiking neural networks. While these architectures are useful for exploring the computational properties of large-scale models of the nervous system, the challenge of building low-power compact physical artifacts that can behave intelligently in the real-world and exhibit cognitive abilities still remains open. In this paper we propose a set of neuromorphic engineering solutions to address this challenge. In particular, we review neuromorphic circuits for emulating neural and synaptic dynamics in real-time and discuss the role of biophysically realistic temporal dynamics in hardware neural processing architectures; we review the challenges of realizing spike-based plasticity mechanisms in real physical systems and present examples of analog electronic circuits that implement them; we describe the computational properties of recurrent neural networks and show how neuromorphic Winner-Take-All circuits can implement working-memory and decision-making mechanisms. We validate the neuromorphic approach proposed with experimental results obtained from our own circuits and systems, and argue how the circuits and networks presented in this work represent a useful set of components for efficiently and elegantly implementing neuromorphic cognition.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
... Layout-extracted simulations, taking into account technology-process corners, PVT variations, noise effects and temperature variation, feature that worst-case offset voltage in the order of 3 mV can be compensated down to 200 μV. This compensated DC offset voltage is further limited by fabrication defects in OxRAM such as nanobattery effect, which is due to nonhomogeneous ion distribution in the electrolyte [26,27]. ...
Article
Full-text available
Redox-based nanoionic resistive memory cells (ReRAMs) are one of the most promising emerging nano-devices for future information technology with applications for memory, logic and neuromorphic computing. Recently, the serendipitous discovery of the link between ReRAMs and memristors and memristive devices has further intensified the research in this field. Here we show on both a theoretical and an experimental level that nanoionic-type memristive elements are inherently controlled by non-equilibrium states resulting in a nanobattery. As a result the memristor theory must be extended to fit the observed non zerocrossing I-V characteristics. The initial electromotive force of the nanobattery depends on the chemistry and the transport properties of the materials system but can also be introduced during ReRAM cell operations. The emf has a strong impact on the dynamic behaviour of nanoscale memories, and thus, its control is one of the key factors for future device development and accurate modelling.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
Article
Full-text available
This paper introduces compact models for memristors. The models are developed based on the fundamental constitutive relationships between charge and flux of memristors. The modeling process, with a few simple steps, is introduced. For memristors with limited resistance ranges, a simple method to find their constitutive relationships is discussed, and examples of compact models are shown for both current-controlled and voltage-controlled memristors. Our models satisfy all of the memristor properties such as frequency dependent hysteresis behaviors and also unique boundary assurance to simulate memristors whether they behave memristively or resistively. Our models are implementable in circuit simulators, including SPICE, Verilog-A, and Spectre.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
Conference Paper
Full-text available
Neuromorphic circuits and systems techniques have great potential for exploiting novel nanotechnology devices, which suffer from great parametric spread and high defect rate. In this paper we explore some potential ways of building neural network systems for sophisticated pattern recognition tasks using memristors. We will focus on spiking signal coding because of its energy and information coding efficiency, and concentrate on Convolutional Neural Networks because of their good scaling behavior, both in terms of number of synapses and temporal processing delay. We propose asynchronous architectures that exploit memristive synapses with specially designed neurons that allow for arbitrary scalability as well as STDP learning. We present some behavioral simulation results for small neural arrays using electrical circuit simulators, and system level spike processing results on human detection using a custom made event based simulator.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
Full-text available
Nanoscale metal/oxide/metal switches have the potential to transform the market for nonvolatile memory and could lead to novel forms of computing. However, progress has been delayed by difficulties in understanding and controlling the coupled electronic and ionic phenomena that dominate the behaviour of nanoscale oxide devices. An analytic theory of the 'memristor' (memory-resistor) was first developed from fundamental symmetry arguments in 1971, and we recently showed that memristor behaviour can naturally explain such coupled electron-ion dynamics. Here we provide experimental evidence to support this general model of memristive electrical switching in oxide systems. We have built micro- and nanoscale TiO2 junction devices with platinum electrodes that exhibit fast bipolar nonvolatile switching. We demonstrate that switching involves changes to the electronic barrier at the Pt/TiO2 interface due to the drift of positively charged oxygen vacancies under an applied electric field. Vacancy drift towards the interface creates conducting channels that shunt, or short-circuit, the electronic barrier to switch ON. The drift of vacancies away from the interface annilihilates such channels, recovering the electronic barrier to switch OFF. Using this model we have built TiO2 crosspoints with engineered oxygen vacancy profiles that predictively control the switching polarity and conductance.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
Full-text available
This review focuses on electrochemical metallization memory cells (ECM), highlighting their advantages as the next generation memories. In a brief introduction, the basic switching mechanism of ECM cells is described and the historical development is sketched. In a second part, the full spectra of materials and material combinations used for memory device prototypes and for dedicated studies are presented. In a third part, the specific thermodynamics and kinetics of nanosized electrochemical cells are described. The overlapping of the space charge layers is found to be most relevant for the cell properties at rest. The major factors determining the functionality of the ECM cells are the electrode reaction and the transport kinetics. Depending on electrode and/or electrolyte material electron transfer, electro-crystallization or slow diffusion under strong electric fields can be rate determining. In the fourth part, the major device characteristics of ECM cells are explained. Emphasis is placed on switching speed, forming and SET/RESET voltage, R(ON) to R(OFF) ratio, endurance and retention, and scaling potentials. In the last part, circuit design aspects of ECM arrays are discussed, including the pros and cons of active and passive arrays. In the case of passive arrays, the fundamental sneak path problem is described and as well as a possible solution by two anti-serial (complementary) interconnected resistive switches per cell. Furthermore, the prospects of ECM with regard to further scalability and the ability for multi-bit data storage are addressed.
... The switching behavior of these ReRAM devices depends on the transition material and the metal electrodes. A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. ...
Article
Full-text available
The resistive switching mechanism of 20- to 57-nm-thick TiO2 thin films grown by atomic-layer deposition was studied by current-voltage measurements and conductive atomic force microscopy. Electric pulse-induced resistance switching was repetitively (> a few hundred times) observed with a resistance ratio > 10(2). Both the low- and high-resistance states showed linear log current versus log voltage graphs with a slope of 1 in the low-voltage region where switching did not occur. The thermal stability of both conduction states was also studied. Atomic force microscopy studies under atmosphere and high-vacuum conditions showed that resistance switching is closely related to the formation and elimination of conducting spots. The conducting spots of the low-resistance state have a few tens times higher conductivity than those of the high-resistance state and their density is also a few tens times higher which results in a similar to 10(3) times larger overall conductivity. An interesting finding was that the area where the conducting spots do not exist shows a few times different resistance between the low- and high-resistance state films. It is believed that this resistance change is due to the difference in point defect density that was generated by the applied bias field. The point defects possibly align to form tiny conducting filaments in the high-resistance state and these tiny conducting filaments gather together to form stronger and more conducting filaments during the transition to the low-resistance state. (c) 2005 American Institute of Physics.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
Article
Full-text available
A mathematical model of the prototype of memristor, manufactured in 2008 in Hewlett-Packard Labs, is described in the paper. It is shown that the hitherto published approaches to the modeling of boundary conditions need not conform with the requirements for the behavior of a practical circuit element. The described SPICE model of the memristor is thus constructed as an open model, enabling additional modifications of non-linear boundary conditions. Its functionality is illustrated on computer simulations.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
... To overcome this problem, the opamps used, either in the neuron circuits or the buffers need to be finely calibrated in order to reduce their input DC offset voltage, which ultimately sets the resting voltage level of the crossbar lines higher than electrical noise. However, conventional calibration schemes are able to compensate for offset ranges in the order of few mV [25]. This paper addresses this limitation and proposes a finer calibration technique. ...
Conference Paper
This paper deals with comparison of two discrete methods for digital trimming of the input offset voltage in operation amplifiers designed in 90nm CMOS technology. Two different topologies based on the binary weighed ladder, one using successive approximation register (SAR) and the other employing a simple counter, were compared. Furthermore, a correction circuit was proposed and used to form the mean offset voltage and increase the probability that its value after trimming process will be near zero. Finally, achieved results and improvements are discussed.
... Layout-extracted simulations, taking into account technology-process corners, PVT variations, noise effects and temperature variation, feature that worst-case offset voltage in the order of 3 mV can be compensated down to 200 μV. This compensated DC offset voltage is further limited by fabrication defects in OxRAM such as nanobattery effect, which is due to nonhomogeneous ion distribution in the electrolyte [26,27]. ...
Article
The impact of the recently discovered nanobattery effect on the switching, the endurance, and the retention of resistive random access memory devices is demonstrated. We show that the relaxation of the electromotive force voltage may lead to a shift of the resistance level for high resistive states, which is included into device modeling. Based on the extended memristive device model, which accounts for the nanobattery effects, endurance and retention problems can be explained.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
The conduction mechanism of metal oxide resistive switching memory is debated in the literature. We measured the I-V characteristics below the switching voltages through TiN∕HfOx∕Pt memory stack and found the conduction cannot be described by the commonly used Poole-Frenkel model, because the fitted dielectric constant and the trap energy are unreasonable as compared to their known values. Therefore, we provide an alternate viewpoint based on a trap-assisted-tunneling model. Agreement of the bias polarity∕temperature∕resistance state-dependent conduction behavior was achieved between this model and experimental data. And insights for the multilevel capability due to the control of tunneling distance were obtained.
... The switching behavior of these ReRAM devices depends on the transition material and the metal electrodes. A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. ...
Article
Resistive switching behavior of Nb2O5 prepared by atomic layer deposition was investigated as a promising candidate for next generation nonvolatile memory technology. The crystalline structure of deposited film at 300 °C was found to be polycrystalline by X-ray diffraction (XRD) and the film was estimated to be oxygen deficient by X-ray photoelectron spectroscopy (XPS). The low resistance ON state and high resistance OFF state can be reversibly altered under low voltage about ±1 V. More than 1000 reproducible switching cycles by DC voltage sweep were observed with a resistance ratio above 10, which was large enough to read out for memory applications. Moreover, the HRS and LRS of the devices are stable for more than 5 × 104 s and does not show any degradation during the test.
... The switching behavior of these ReRAM devices depends on the transition material and the metal electrodes. A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. ...
Article
Resistive-switching random access memory (RRAM) using the TiN/AlOx/Pt stack is fabricated with a 50 nm × 50 nm active area. The bipolar switching characteristic is observed using TiN as an anode and RESET current as low as sub-20 μA was achieved by using a current-limiting transistor as a selection device (1T1R) during the SET process. HRS to LRS ratio of 103 for 103 DC endurance test cycles is demonstrated. Switching time less than 10 ns was observed for both SET/RESET operations. By changing the gate bias of the series transistor and input voltage at the transistor drain terminal, multi-level resistance states can be modulated. Both the unipolar and bipolar resistance switching can coexist in such memory stacks and a qualitative model is proposed for the AlOx -based RRAM resistive switching behavior.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
This review covers resistive random access memories which utilize redox processes and ionic motion on the nanoscale as their storage principle (ReRAM). Generic aspects are described in order to provide the physics and chemistry background for the explanation of the microscopic switching mechanism and of the high nonlinearity in the switching kinetics. The valence change memory (VCM) effect is elaborated in more detail. As common features, ReRAM typically show very short switching times, low switching energies, and long data retention times. In addition, they offer a scalability potential down to feature sizes in the order of 5 nm and below.
... A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. The two resistance states are dynamically selected by a control voltage applied to a series-connected MOSFET transistor, leading to the so-called 1T1R structure, as illustrated in Fig. 1. ...
Article
Novel information processing techniques are being actively explored to overcome fundamental limitations associated with CMOS scaling. A new paradigm of adaptive electronic devices is emerging that may reshape the frontiers of electronics and enable new modalities. Creating systems that can learn and adapt to various inputs has generally been a complex algorithm problem in information science, albeit with wide-ranging and powerful applications from medical diagnosis to control systems. Recent work in oxide electronics suggests that it may be plausible to implement such systems at the device level, thereby drastically increasing computational density and power efficiency and expanding the potential for electronics beyond Boolean computation. Intriguing possibilities of adaptive electronics include fabrication of devices that mimic human brain functionality: the strengthening and weakening of synapses emulated by electrically, magnetically, thermally, or optically tunable properties of materials.In this review, we detail materials and device physics studies on functional metal oxides that may be utilized for adaptive electronics. It has been shown that properties, such as resistivity, polarization, and magnetization, of many oxides can be modified electrically in a non-volatile manner, suggesting that these materials respond to electrical stimulus similarly as a neural synapse. We discuss what device characteristics will likely be relevant for integration into adaptive platforms and then survey a variety of oxides with respect to these properties, such as, but not limited to, TaOx, SrTiO3, and Bi4-xLaxTi3O12. The physical mechanisms in each case are detailed and analyzed within the framework of adaptive electronics. We then review theoretically formulated and current experimentally realized adaptive devices with functional oxides, such as self-programmable logic and neuromorphic circuits. Finally, we speculate on what advances in materials physics and engineering may be needed to realize the full potential of adaptive oxide electronics.
... The crossbar consists of a 4 × 4 1T1R array of OxRAM devices. Initially, each OxRAM device is initially at very high resistance state of 100 GΩ, called Pristine Resistance State (PRS), and hence it has to be electro-formed to make the conductive filament in the oxide layer for the first time [28]. After forming, the OxRAM typically reaches LRS, and hence a RESET (or erase) operation has to be carried out to push it to HRS. ...
Article
In this paper, recent progress of binary metal-oxide resistive switching random access memory (RRAM) is reviewed. The physical mechanism, material properties, and electrical characteristics of a variety of binary metal-oxide RRAM are discussed, with a focus on the use of RRAM for nonvolatile memory application. A review of recent development of large-scale RRAM arrays is given. Issues such as uniformity, endurance, retention, multibit operation, and scaling trends are discussed.
... The switching behavior of these ReRAM devices depends on the transition material and the metal electrodes. A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. ...
Article
The resistance switching characteristics of HfO <sub>2</sub> thin films deposited by reactive sputtering were examined as a function of the annealing temperature. The results showed that the Pt / HfO <sub>2</sub>/ Pt devices exhibited reversible and steady bistable resistance states [high-resistance state (HRS) and low-resistance state (LRS)]. Reproducible resistance switching from one state to another state or vice versa could be achieved by applying the appropriate voltage bias. The memory performances were related to the crystal structures of the HfO <sub>2</sub> films, as confirmed by x-ray diffraction. From current-applied voltage analysis of the devices, LRS in the low electric field regime exhibited Ohmic conduction behavior, while HRS in the high electric field was followed by Poole–Frenkel conduction behavior. The resistance ratios of the two states were maintained in the range of around two orders of magnitude during the endurance test. In addition, it was confirmed that the resistance of the on and off states can be well maintained according to the time elapsed.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
This paper is both a review of some recent developments in the utilization of magnetism for applications to logic and memory and a description of some new innovations in nanomagnetics and spintronics. Nanomagnetics is primarily based on the magnetic interactions, while spintronics is primarily concerned with devices that utilize spin polarized currents. With the end of complementary metal-oxide-semiconductor (CMOS) in sight, nanomagnetics can provide a new paradigm for information process using the principles of magnetic quantum cellular automata (MQCA). This paper will review and describe these principles and then introduce a new nonlithographic method of producing reconfigurable arrays of MQCAs and/or storage bits that can be configured electrically. Furthermore, this paper will provide a brief description of magnetoresistive random access memory (MRAM), the first mainstream spintronic nonvolatile random access memory and project how far its successor spin transfer torque random access memory (STT-RAM) can go to provide a truly universal memory that can in principle replace most, if not all, semiconductor memories in the near future. For completeness, a description of an all-metal logic architecture based on magnetoresistive structures (transpinnor) will be described as well as some approaches to logic using magnetic tunnel junctions (MTJs).
... ReRAM memristor technology combines the features of high-speed performance of present Static Random Access Memory (SRAMs) with non-volatile property of flash memory, which can be realized at low power consumption. ReRAM memristive devices are also known for its robustness and integration capability [16]. ...
Article
Various classes of resistive random access memory (RRAM) classes involving chemical effects related to redox processes in the MIM cell are presented. The electrochemical metallization (ECM) cells systems involve Ag and Cu as electrochemically active metals and phase separated amorphous selenides and sulfides and various oxides, acting as solid electrolytes. The polarity of the bipolar switching cycle is determined by many factors such as the work function and the oxygen affinity of the electrode metals. The SET and RESET switching of an individual filament at the surface of a SrTiO3 single crystal shows that the resistance of the filament can be switched between an ONand OFFstate by a voltage applied to the tip of the LC-AFM. The dissolution of cylindric metal filaments studied by electrothermal simulation is described by a thermally activated process of the filament surface.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
... The best case and worst case corners are classified based on different design parameters like mobility, vth variation, resistance of the actives, body coefficient, oxide thickness and PVT variations. 3 Monte carlo simulations include both processes (wafer-to-wafer variations), ...
Article
A memristor is a two-terminal electronic device whose conductance can be precisely modulated by charge or flux through it. Here we experimentally demonstrate a nanoscale silicon-based memristor device and show that a hybrid system composed of complementary metal-oxide semiconductor neurons and memristor synapses can support important synaptic functions such as spike timing dependent plasticity. Using memristors as synapses in neuromorphic circuits can potentially offer both high connectivity and high density required for efficient computing.
... The switching behavior of these ReRAM devices depends on the transition material and the metal electrodes. A variety of such transition materials -HfO 2 , NiO, Al 2 O 3 , Nb 2 O 5 , SrTiO 3 , Pr 0.7 Ca 0.3 MnO 3 , CuO 2 , Ag 2 S and AgGeSehave been experimented and their switching characteristics have been studied in literature [17][18][19][20][21] device, which is operated in binary mode, so that, the resistance of the embedded TiN-Ti-HfO 2 -TiN structure can be switched between two different resistance states, namely: Low Resistance State (LRS)typically in the order of kΩ-and High Resistance State (HRS)in the order of hundreds of kΩ to MΩ [22]. HfO 2 -based OxRAMs are known for its low-switching energy and high endurance, when compared to other oxide-based ReRAMs [23]. ...
Article
The great variability in the electrical properties of multinary oxide materials, ranging from insulating, through semiconducting to metallic behaviour, has given rise to the idea of modulating the electronic properties on a nanometre scale for high-density electronic memory devices. A particularly promising aspect seems to be the ability of perovskites to provide bistable switching of the conductance between non-metallic and metallic behaviour by the application of an appropriate electric field. Here we demonstrate that the switching behaviour is an intrinsic feature of naturally occurring dislocations in single crystals of a prototypical ternary oxide, SrTiO(3). The phenomenon is shown to originate from local modulations of the oxygen content and to be related to the self-doping capability of the early transition metal oxides. Our results show that extended defects, such as dislocations, can act as bistable nanowires and hold technological promise for terabit memory devices.
... The I-pot's conceptual block diagram is shown in Fig. 4. It has a decade current splitter, a fine current splitter and a current sign selector and tester. The current splitter circuits are basically MOS ladder structures [30]. The decade current splitter has 6 current splitting possibilities, where the reference current, i ref is split by 10 in each stage of a MOS ladder structure. ...
Article
This paper considers the viability of compact low-resolution low-power mini digital-to-analog converters (mini-DACs) for use in large arrays of neural type cells, where programmable weights are required. Transistors are biased in weak inversion in order to yield small currents and low power consumptions, a necessity when building large size arrays. One important drawback of weak inversion operation is poor matching between transistors. The resulting effective precision of a fabricated array of 50 DACs turned out to be 47% (1.1 bits), due to transistor mismatch. However, it is possible to combine them two by two in order to build calibrated DACs, thus compensating for inter-DAC mismatch. It is shown experimentally that the precision can be improved easily by a factor of 10 (4.8% or 4.4 bits), which makes these DACs viable for low-resolution applications such as massive arrays of neural processing circuits. A design methodology is provided, and illustrated through examples, to obtain calibrated mini-DACs of a given target precision. As an example application, we show simulation results of using this technique to calibrate an array of digitally controlled integrate-and-fire neurons.
... Memristors have emerged as promising circuit elements for neuromorphic computing circuits. After Chua coined the word 'Memristor' and later when HP labs showed its physical existence, many memristor models were implemented to characterize, study and explore its potential applications [1][2][3][4][5][6]. Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. ...
Article
Anyone who ever took an electronics laboratory class will be familiar with the fundamental passive circuit elements: the resistor, the capacitor and the inductor. However, in 1971 Leon Chua reasoned from symmetry arguments that there should be a fourth fundamental element, which he called a memristor (short for memory resistor). Although he showed that such an element has many interesting and valuable circuit properties, until now no one has presented either a useful physical model or an example of a memristor. Here we show, using a simple analytical example, that memristance arises naturally in nanoscale systems in which solid-state electronic and ionic transport are coupled under an external bias voltage. These results serve as the foundation for understanding a wide range of hysteretic current-voltage behaviour observed in many nanoscale electronic devices that involve the motion of charged atomic or molecular species, in particular certain titanium dioxide cross-point switches.
... A systematic procedure is followed to tune the specifications of the opamp during its design. I-pots are digitally programmable current sources which, from a reference current, can provide any desired current with high precision, down to pA [29]. I-pot circuits are used as current bias source circuits for the opamp. ...
Article
In this brief, we present the "stochastic I-Pot." It is a circuit element that allows for digitally programming a precise bias current ranging over many decades, from pico-amperes up to hundreds of micro-amperes. I-Pot blocks can be chained within a chip to allow for any arbitrary number of programmable bias currents. The approach only requires to provide the chip with three external pins, the use of an external current measuring instrument, and a computer. This way, once all internal I-Pots have been characterized, they can be programmed through a computer to provide any desired current bias value with very low error. The circuit block turns out to be very practical for experimenting with new circuits (specially when a large number of biases are required), testing wide ranges of biases, introducing means for current mismatch calibration, offsets compensations, etc. using a reduced number of chip pins. We show experimental results of generating bias currents with errors of 0.38% (8 bits) for currents varying from 176 muA to 19.6 pA. Temperature effects are characterized.
... Memristors are a good choice of candidate when used as artificial synapses in CMOS-based neuromorphic computing circuits due to its property of non-volatility, analog behavior and continuously distributed resistive states. Different switching mechanisms such as redox-based, phase-change, magnetic junction and ferroelectric, and different physical models such as conductive filament, Schottky barrier, charge trapping and electrochemical migration of point defect, have been investigated to better understand the switching phenomenon [7][8][9][10][11][12][13][14][15]. ...
Article
This paper surveys circuit innovations in ferroelectric memories at three circuit levels: memory cell, sensing and architecture. A ferroelectric memory cell consists of at least one ferroelectric capacitor, where binary data are stored, and one or two transistors that either allow access to the capacitor or amplify its contents for a read operation. Once a cell is accessed for a read operation, its data are presented in the form of an analog signal to a sense amplifier, where it is compared against a reference voltage to determine its logic level. The circuit techniques used to generate the reference voltage must be robust to semiconductor processing variations across the chip and the device imperfections of ferroelectric capacitors. We review six methods of generating a reference voltage, two being presented for the first time in this paper. These methods are discussed and evaluated in terms of their accuracy, area overhead and sensing complexity. Ferroelectric memories share architectural features such as addressing schemes and input/output circuitry with other types of random-access memories such as dynamic random-access memories. However, they have distinct features with respect to accessing the stored data, sensing, and overall circuit topology. We review nine different architectures for ferroelectric memories and discuss them in terms of speed, density and power consumption