Analog Switch Research Articles

Parallel Operation of Self‐Limited Analog Programming for Fast Array‐Level Weight Programming and Update

Memristive neural networks perform vector matrix multiplication efficiently, which is used for the accelerator of neuromorphic computing. To train the memristor cells in a memristive neural network, the analog conductance state of the memristor should be programmed in parallel; otherwise, the resulting long training time can limit the size of the neural network. Herein, a novel parallel programming method using the self‐limited analog switching behavior of the memristor is proposed. A Pt/Ti:NbOx/NbOx/TiN charge trap memristor device for the programming demonstration is utilized, and a convolutional neural network is emulated to train the MNIST dataset, based on the device characteristics. In the simulation, the proposed programming method is able to reduce programming time to as low as 1/130, compared with the sequential programming method. The simulation suggests that the programming time required by the proposed method is not affected by array size, which makes it very promising in a high‐density neural network.

Tuning the analog and digital resistive switching properties of TiO2 by nanocompositing Al-doped ZnO

Resistive switching (RS) is a versatile effect that facilitates the fabrication of non-volatile memory and brain-inspired computing devices by utilizing digital and analog switching. In this work, we demonstrated analog and digital RS properties of titanium dioxide (TiO2) by nanocompositing aluminum-doped zinc oxide (AZO). In particular, we achieved a transition from analog to digital RS by appropriately nanocompositing the AZO (0–25 wt %) in the TiO2 active switching layer. The active switching layers were characterized by different spectroscopic and microscopic techniques, suggested that TiO2 has a tetragonal anatase phase. The morphological study revealed the synthesis of uniform nanocomposite particles. Furthermore, good crystalline nature and phase purities of the nanocomposites were confirmed by HRTEM results. The bare TiO2 possessed rectifying analog RS property whereas, digital RS was observed for 5 to 20 wt % TiO2/AZO-based nanocomposite devices. The digital RS was again converted to analog RS for 25 wt % TiO2/AZO-based device with good hysteresis area and memory window. The time-domain charge and charge-flux characteristics were calculated for all devices, suggesting memristive like properties. The conduction model fitting results suggested that the Schottky and space-charge limited current conduction are responsible for the analog and digital type RS, respectively. Considering the electrical results, we present a possible RS mechanism. The results of the present investigation are useful for the development of non-volatile memory and brain-inspired computing devices.

Principles and methods of non-contact DC contactors calculation for alternative power systems

It is shown that the principle of control and power supply of a non-contact bi-directional DC contactor by the electrical line voltage is implemented by connecting the electronic control circuit by one output between two opposite powered gate-controlled thyristors of the bilateral switch to the first pole of the contactor and the second one – to the output clamp. As a result, the electrical line voltage is fed to the electronic circuit only for a short time (up to ms), that is determined by the charge time of the switching capacitor to the breakdown voltage of the threshold zener diode. The peculiarities of the heating process of the power semiconductor contactor in different operating conditions were determined, that allowed to develop a methodology that with sufficient accuracy determines such important parameters as rated and operational current of the contactor. In addition, it was found that in order to reduce significantly the capacity of the switching capacitor, and consequently the size and cost, it is necessary while turning on of the switching thyristor the discharging current of the switching capacitor was equal to the current of its charging circuit, and their values should not be less than the maximum allowable current of turning off the gate-controlled thyristor. It is allows to define justified the parameters of the elements that provide a secure turning off the bilateral switch. The researches have also shown that the non-contact contactors due to the introduction of the network voltage control as compared with existing ones have peculiarities which rise competitive abilities. In particular, they are more reliable, they do not need to supply an additional power source, they exclude standard drivers, and minimize energy consumption. Thus, an exemplary aspect of using the obtained scientific result is the ability to create competitive reliable non-contact DC contactors for voltage up to 1000 V and currents of 100-600 A for alternative energy. Key words – non-contact bi-directional DC contactor, switching capacitor, switching thyristor, power semiconductor switch, alternative energy.

Initial states and analog switching behaviors of two major tantalum oxide resistive memories

A systematic study was performed to understand the two major operation modes, valence change memory (VCM) and electrochemical memory (ECM), of tantalum oxide resistive memory, which were categorized by the type of conductive filament. The tantalum oxide insulator layer was predictably fabricated to modify the pristine oxygen-vacancy concentration in the insulator layer, where tantalum and copper top electrodes were used to achieve VCM and ECM modes. The distinguishable characteristics of the initial state and analog resistive switching operation revealed the prerequisites for switching capabilities and demonstrated the effects of oxygen vacancies in the two switching modes.

Physically Transient Resistive Memory With Programmable Switching Behaviors in MgO-Mo Based Devices

In this letter, physically transient resistive memory devices based on MgO with programmable switching behaviors by embedding Mo nanolayer were proposed. The devices with a 2-3 nm Mo nanolayer (W/MgO/Mo(2-3 nm)/MgO/W) exhibit bipolar analog switching behavior and synaptic functions that are promising for neuromorphic electronics. When the Mo nanolayer is increased to 7-8 nm, complementary resistive switching (CRS) behavior is demonstrated which can be used to solve sneak current issue in crossbar array. Additionally, triggered electrical failure was achieved after immersing the devices in DI water for 4 min. This transient device with programmable switching behaviors by tuning Mo nanolayer thickness might provide guidelines to advance the security, biocompatible and implantable data storage and neuromorphic computing systems.

Energy Efficient Transmit-Receive Hybrid Spatial Modulation for Large-Scale MIMO Systems

We consider a point to point large-scale multiple-input multiple-output (MIMO) system operating in the millimeter wave (mmWave) band and an outdoor scenario. Novel transmit and receive spatial modulation (SM) schemes are proposed for uplink (UL) and downlink (DL) data transmission phases based on a novel energy efficient hybrid user terminal architecture. The analog circuitry of the proposed hybrid architecture is divided into two stages: phase shifters and analog switches. The phase shifting stage assures high gain and overcomes the severe path-loss caused by outdoor mmWave propagation. The analog switching stage smartly allocates the antennas to be used at the phase shifting stage and combats the spatial correlation. We provide the analysis of the spectral efficiency $({{SE}})$ of the UL and DL systems. Next, we propose a reduced complexity algorithm that jointly optimizes the analog beamformer and combiner design of the UL and DL circuitry to maximize the energy efficiency $({{EE}})$ . Finally, we compare and evaluate the performance of the proposed algorithm in terms of the ${{SE}}$ and ${{EE}}$ assuming both stochastic and realistic channel models.

Experimental and computational analysis of thermal environment in the operation of HfO2 memristors

Neuromorphic computation using nanoscale adaptive oxide devices or memristors is a very promising alternative to the conventional digital computing framework. Oxides of transition metals, such as hafnium (HfOx), have been proven to be excellent candidate materials for these devices, because they show non-volatile memory and analog switching characteristics. This work presents a comprehensive study of the transport phenomena in HfOx based memristors and involves the development of a fully coupled electrothermal and mass transport model that is validated with electrical and thermal metrology experiments. The fundamental transport mechanisms in HfOx devices were analyzed together with the local and temporal variation of voltage, current, and temperature. The effect of thermal conductivity of substrate materials on the filament temperature, voltage ramp rate, and set/reset characteristics was investigated. These analyses provide insight into the switching mechanisms of these oxides and allow for the prediction of the effect of device architecture on switching behavior.

Impact of interfaces on bipolar resistive switching behavior in amorphous Ge–Sb–Te thin films

Electrochemical metallization memories have received much attention as candidates for next generation non-volatile memory applications, due to their fast switching speed, simple structure, high scalability and low energy consumption. Chalcogenide compounds like Ge–Sb–Te-based materials are extensively studied solid electrolytes for such devices and considered within the most promising candidates. In this work, the influence of different electrode materials on bipolar resistive switching characteristics of amorphous Ge–Sb–Te thin films with close composition to 2:2:5, prepared by pulsed laser deposition, is studied. Detailed investigations by current-voltage measurements, secondary ion mass spectrometry and cross-sectional transmission electron microscopy in conjunction with energy-dispersive x-ray spectroscopy are presented. Depending on the utilized electrode material, analog switching (Cu, Ag, Ti), digital switching (Co, Cr, Ta, Al) or no switching (Pt, Au and Cr-CrOx) occurs. This work shows that the switching behavior of Ge–Sb–Te strongly depends on the solubility and reactivity of the electrode material.

Characteristics of analog memristor on thin-film Pt/Co0.2TiO3.2/ITO

In this work, thin films of Pt/Co0.2TiO3.2 (23 nm)/ITO were investigated showing their memristive capacity related to analog switching properties. A temporal analysis of current and voltage in nonlinear I–V graphs shows pinched hysteresis loop with gradual distribution in its electrical conductivity, typical of analog systems. The presence of trap centers such as oxygen vacancies suggests that the specific conduction mechanism is space-charge-limited current (SCLC). In addition, we study its performance under repetitive voltage pulses as well as at different durations. We further conducted data retention tests in the memristor high resistance state (HRS) and low resistance state (LRS).

Synaptic learning and memory functions in SiO2:Ag/TiO2 based memristor devices

In this study, a TiO2 thin layer is deliberately inserted in the Pt–Ag/SiO2:Ag/p++-Si memristor to achieve enhanced resistive switching performance in terms of retention and cyclic endurance characteristics. The memristor device exhibits a bidirectional analog switching behavior corresponding to a wide dynamic range of synaptic weight modulation as potentiation and depression under the consecutive positive or negative DC bias sweeps and pulse trains. Moreover, the learning and memory functions, including paired-pulse facilitation and pulse repetition-induced short-term memory to long-term memory transition, are also demonstrated. The relation of the memory stability to the repeated input pulse stimulus and the device conductance with time have also been revealed by the stretched exponential relaxation model. It is shown that the conductance of our newly fabricated Pt–Ag/SiO2:Ag/TiO2/p++-Si memristor device can be modulated by successive Ag-filament growth and dissolution in the SiO2:Ag/TiO2 bilayer oxide structure, which is essential for potential applications in memory devices and artificial neural systems.

Harmonic Distortion Optimization for Sigma-Delta Modulators Interface Circuit of TMR Sensors.

The tunneling magnetoresistance micro-sensors (TMR) developed by magnetic multilayer material has many advantages, such as high sensitivity, high frequency response, and good reliability. It is widely used in military and civil fields. This work presents a high-performance interface circuit for TMR sensors. Because of the nonlinearity of signal conversion between sensitive structure and interface circuit in feedback loop and forward path, large harmonic distortion occurs in output signal spectrum, which greatly leads to the reduction of SNDR (signal noise distortion rate). In this paper, we analyzed the main source of harmonic distortion in closed-loop detection circuit and establish an accurate harmonic distortion model in TMR micro-sensors system. Some factors are considered, including non-linear gain of operational amplifier unit, effective gain bandwidth, conversion speed, nonlinearity of analog transmission gate, and nonlinearity of polycrystalline capacitance in high-order sigma-delta system. We optimized the CMOS switch and first-stage integrator in the switched-capacitor circuit. The harmonic distortion parameter is optimally designed in the TMR sensors system, aiming at the mismatch of misalignment of front-end system, non-linearity of quantizer, non-linearity of capacitor, and non-linearity of analog switch. The digital output is attained by the interface circuit based on a low-noise front-end interface circuit and a third-order sigma-delta modulator. The digital interface circuit is implemented by 0.35μm CMOS (complementary metal oxide semiconductor) technology. The high-performance digital TMR sensors system is implemented by double chip integration and the active interface circuit area is about 3.2 × 2 mm. The TMR sensors system consumes 20 mW at a single 5 V supply voltage. The TMR sensors system can achieve a linearity of 0.3% at full scale range (±105 nT) and a resolution of 0.25 nT/Hz1/2(@1Hz).

The Design and Verification of an Active SAMSR Ultrasonic Guided Wave Monitoring System with Ultra-Low Crosstalk.

Due to long propagation distance and high sensitivity to a variety of damages, ultrasonic guided wave technologies have been widely applied in the damage detection or health monitoring of pipe networks and large plate-like structures. However, there are two important problems to be solved when applying this technology; namely, the large scanning time required for monitoring large-scaled structures and the serious crosstalk between the actuation and receiving signals, especially when monitoring hot-spot regions. Therefore, this study mainly designed key parts, such as the matrix switcher and attenuation circuit. The single-actuation and multiple-simultaneous-reception (SAMSR) mechanism based on an analog switching matrix and a low noise charge amplifier circuit was designed and integrated with the SPI control bus to shorten the scanning time. Moreover, a two-stage attenuation circuit with an interlocking isolation structure is presented to effectively isolate the receiving signals from the actuation signals to obtain ultra-low crosstalk even under a high voltage actuation source. In this study, the designed matrix switcher and other components were integrated into the developed ultrasonic guided wave monitoring system. Several experiments were conducted on a stiffened composite structure to illustrate the effectivity of the developed SAMSR ultrasonic guided wave system by comparing the signals collected with those from a commercial ultrasonic guided wave system.

A Low-Current and Analog Memristor with Ru as Mobile Species.

The switching parameters and device performance of memristors are predominately determined by their mobile species and matrix materials. Devices with oxygen or oxygen vacancies as the mobile species usually exhibit a great retention but also need a relatively high switching current (e.g., >30 µA), while devices with Ag or Cu as cation mobile species do not require a high switching current but usually show a poor retention. Here, Ru is studied as a new type of mobile species for memristors to achieve low switching current, fast speed, good reliability, scalability, and analog switching property simultaneously. An electrochemical metallization-like memristor with a stack of Pt/Ta2 O5 /Ru is developed. Migration of Ru ions is revealed by energy-dispersive X-ray spectroscopy mapping and in situ transmission electron microscopy within a sub-10 nm active device area before and after switching. The results open up a new avenue to engineer memristors for desired properties.

Differential Work-Function Enabled Bifunctional Switching in Strontium Titanate Flexible Resistive Memories.

Multifunctional electronic memories capable of demonstrating both analog and digital switching on-demand are extremely attractive for miniaturization of electronics without significant drain on energy consumption. Simultaneously translating functionality onto mechanically conformable platforms will further enhance their suitability. Here, we demonstrate the ability to engineer multifunctionality in strontium titanate (STO)-based resistive random-access memories (ReRAM) on a flexible polyimide platform. By utilizing different bottom electrodes of various work functions while the top electrode is fixed, differential work functions are induced in STO, to induce bipolar or complementary switching behaviors whenever required. This work-function difference-induced bifunctional switching on the flexible platform reveals a streamlined route for achieving flexible artificial neural networks, high density integration, and logic operation using a single ReRAM.

Improvement of analogue switching characteristics of MoS2 memristors through plasma treatment

Memristive devices based on 2D materials, such as WSe2 and MoS2, have been demonstrated to exhibit analogue switching characteristics and enable emulation of ionic interactions involved in synaptic activities. These attractive features hold great potential for construction of energy-efficient artificial neural networks. However, the memristors made from pristine 2D materials typically exhibit a small dynamic range and poor linearity of switching characteristics. The neural network simulated in the basis of such switching characteristics has a poor learning accuracy of ~43%. In this work, we find that Ar plasma treatment can greatly improve both the dynamic range and linearity of analogue switching characteristics of few-layer MoS2 memristors. The neural network consisting of such plasma-treated memristors is simulated to be able to result in a significantly improved learning accuracy of 94.3% for the MNIST handwritten digits dataset. Our additional Auger electron analysis in combination with electronic characterizations indicates that Ar plasma enhances the concentration of movable S vacancies in MoS2 channels, which leads to improvements in the analogue switching properties. Especially, the average dynamic range of MoS2 memristors are increased from 1.8 to 14.8 after plasma treatment. This work provides scientific insights for controlling the switching characteristics of 2D memristors and provides technical instruction for construction of practical neural networks based on 2D materials.

A Physics-Based Analytic Model of Analog Switching Resistive Random Access Memory

Many efforts have been reported to improve the linear conductance modulation of analog RRAM for neuromorphic computing, while the related model for analog switching is still lacking. In this letter, we propose a physics-based analytic model to account for analog behavior of HfO2-based RRAM. The switching from abrupt SET process to gradual SET is analyzed quantitatively by considering the electrical and thermal effects of intermediate modulation layer (IML). The proposed model well describes the conductance update in potentiation and depression, which is validated with the experimental data. The results show that for the case of introducing IML to improve linearity, the electrical resistivity of IML has a greater effect than thermal conductivity on the control of oxygen vacancy generation rate, thus playing a dominant role in the nonlinearity optimization of analog RRAM.

Hybrid precoding based on adaptive RF-chain-to-antenna connection for millimeter wave MIMO systems

Hybrid precoding has recently received significant attentions as a promising alternative to fully digital precoding. The existing hybrid precoding architectures can be classified into two categories: fully-connected (FC) and partially-connected (PC) phase shifter (PS) network structures (FC-PS and PC-PS structures). However, the FC-PS structure involves a large number of PSs which leads to a prohibitive power consumption, and the PC-PS structure suffers from a severe spectral efficiency loss in spite of low power consumption. This paper considers a novel adaptive RF-chain-to-antenna connection based on switch network (AC-S) hybrid precoding architecture, which consists of a digital precoder, an analog switch network and an analog PS network. The advantage of the proposed architecture lies in achieving a satisfactory spectral efficiency while maintaining a relatively low power consumption due to the adaptive connection from a radio frequency (RF) chain to all antennas via a switch network. Based on the aforementioned architecture, an effective hybrid precoding algorithm is presented. In the proposed algorithm, the analog PS precoding matrix, the analog switch precoding matrix and the digital precoding matrix are optimized via a block coordinate descent method, and a stable solution can be obtained by several iterations. Simulation results demonstrate that the proposed algorithm with the AC-S structure not only obtains considerably better spectral efficiency than the PC-PS structure but also offers a lower power consumption than the FC-PS structure.

Enhanced Switching Properties in TaOx Memristors Using Diffusion Limiting Layer for Synaptic Learning

To move towards a new generation powerful computing system, brain-inspired neuromorphic computing is expected to transform the architecture of the conventional computer, where memristors are considered to be potential solutions for synapses part. We propose and demonstrate a novel approach to achieve remarkable improvement of analog switching linearity in TaN/Ta/TaO x /Al 2 O 3 /Pt/Si memristors by varying Al 2 O 3 layer thickness. Presence of the Al 2 O 3 layer is confirmed from the Auger Electron Spectroscopy study. Good analog switching ratio of about $100\times $ and superior switching uniformity are observed for the 1 nm Al 2 O 3 based device. Multilevel capability of the memristive devices is also explored for prospective use as a synapse. More than 10 4 and $4\times 10^{4}$ cycles nondegradable dc and ac endurances, respectively, alongwith 10 4 second retention are achieved for the optimized device. Improved linearities of 2.41 and −2.77 for potentiation and depression, respectively are obtained for such 1 nm Al 2 O 3 -based devices. The property of gradual resistance changed by pulse amplitudes confirms that the TaO x memristors can be potentially used as an electronic synapse.

Digital and Analog Switching Characteristics of InGaZnO Memristor Depending on Top Electrode Material for Neuromorphic System

In this study, we demonstrate both of digital and analog memory operations in InGaZnO (IGZO) memristor devices by controlling the electrode materials for neuromorphic application. The switching properties of the devices are determined by the initial energy barrier characteristics between the metal electrodes and the IGZO switching layer. Digital switching characteristics are obtained after the forming process when Schottky junction occurs at both of top and bottom electrodes. On the other hands, analog resistive switching is achieved when Schottky and Ohmic junctions exist at each side because the applied voltage modulates the Schottky barrier height through the Ohmic contact. In addition, the weight-update properties of the devices are verified depending on identical and incremental pulse schemes. The incremental pulse trains improve the linearity and variation of weight modulation, leading to the stable learning characteristics of neuromorphic system in terms of pattern recognition with MNIST hand-written digit images.

Improved Low Power Multichannel Pulse Amplitude Analyzer

Multichannel pulse amplitude analyzer is the core part of ray energy spectrum analysis system. In order to meet the requirements of ultra-low power consumption and portable high accuracy of the instrument, this design is characterized by the application of differential circuit in the screening circuit, which eliminates the analog switch, capacitance and diode, and greatly reduces the power consumption of the instrument. Without the use of diodes, circuits can discriminate pulses of any magnitude, especially in gamma-ray spectrometer systems. The overall structure is 89C52 MCU as the control core, simple design, external RAM storage measurement data, through RS-232 serial port and computer communication. According to the requirements of the measured frequency, switching values of R and C can be applied to a variety of energy spectrum analyzer systems.