Ligand Exchange Reaction Enables Digital-To-Analog Resistive Switching and Artificial Synapse within Metal Nanoparticles

The transition between digital and analog resistive switching in a single memristive device is beneficial for the reduction in power consumption and circuit complexity, the development of in-memory neuromorphic computing, and the discovery of new switching mechanisms. However, achieving such transition is a challenge due to the complex switching mechanisms and device designs. Here, it is shown that the digital-to-analog resistive switching can be realized by the ligand exchange reaction of metal nanoparticles. The field-injected copper cations migrate within carboxyl-functionalized gold nanoparticle (AuNP) layer that are subsequently reduced into metallic filaments, enabling an abrupt resistive switching. Importantly, when the carboxyl groups on the gold nanoparticle are replaced by amino-carboxyl ligands, the copper cations coordinate with the new ligands and create the conductance bridges to reduce the electron tunneling/hopping energy barriers, leading to continuous modulation in conductivity. This analog resistive switching allows to implement several important synaptic functions such as potentiation/depression, paired-pulse facilitation, learning behaviors including forgetting curves and spaced learning effect. In the end, due to the non-volatile characteristics, the gold nanoparticle synapse is used to build single layer perceptron for pattern classification with 100% accuracy.     

Electronic synapses mimicked in bilayer organic-inorganic heterojunction based memristor

Electronic synapses implementing in-memory computing system could overcome the developing limitation on the energy efficiency of traditional von Neumann architecture. Compared with the high sensitivity of biological synapses, lower responsivity of the memristive synapses was found via the electrical stimulations. Here, poly{2,2-(2,5-bis(2-octyldodecyl)-3,6-dioxo-2,3,5,6- tetrahydropyrrolo[3,4-c]pyrrole-1,4-diyl)-dithieno[3,2-b]thiophene-5,5-diyl-alt-thiophen-2,5-diyl} (PDPPBTT)/zinc oxide (ZnO) based heterojunction is found to exhibit stable memristive switching behavior, which originates from the confined formation/rupture of filament in the two-layer interface region as the ions migrate with different transport rates in two layers. The implementing synaptic functions in the sensitive memristive device can realize the short-term plasticity and long-term plasticity when stimulated by the applied electrical signals with different stimulating rate. Similar to the biological synapse, the memory loss, memory transition, and the critical role of stimulation rate on the transition process, can be achieved in the as-prepared memristor device. The systematic demonstrations on the synaptic emulation may facilitate building bio-inspired device-level neuromorphic systems.     

Filament diffusion model for simulating reset and retention processes in RRAM

Resistive switching memory (RRAM) devices are attracting an increasing interest as a possible future technology for ultra-scaled, high-density nonvolatile/dynamic memory. Although the RRAM concept is promising from the integration and scaling viewpoints, the switching mechanism and its controllability are still under debate. This paper addresses the modeling of reset and retention processes in unipolar resistive-switching memory devices. Reset transition and data loss are described in terms of the dissolution of a conductive filament, which is modeled by thermally-activated diffusion of defects/dopants. Carrier transport, Joule heating and diffusion of oxygen ions/vacancies during the electrical pulse and/or the annealing are modeled within a 3D numerical solver. The model can account for the observed dependence of reset voltage on the width of the applied triangular pulse and on the initial resistance for NiO-based RRAM devices. Retention simulations as a function of annealing temperature also agree with available data. The model provides a first example of device simulation tool for the design and the exploration of scaling and performance of RRAM cells.     

Enhancing Memory Window Efficiency of Ferroelectric Transistor for Neuromorphic Computing via Two-Dimensional Materials Integration

In-memory computing, particularly neuromorphic computing, has emerged as a promising solution to overcome the energy and time-consuming challenges associated with the von Neumann architecture. The ferroelectric field-effect transistor (FeFET) technology, with its fast and energy-efficient switching and nonvolatile memory, is a potential candidate for enabling both computing and memory within a single transistor. In this study,  the capabilities of an integrated ferroelectric HfO₂ and 2D MoS₂ channel FeFET in achieving high-performance 4-bit per cell memory with low variation and power consumption synapses, while retaining the ability to implement diverse learning rules, are demonstrated. Notably, this device accurately recognizes MNIST handwritten digits with over 94% accuracy using online training mode. These results highlight the potential of FeFET-based in-memory computing for future neuromorphic computing applications.     

面向自旋存内计算架构的图算法优化设计

图计算广泛应用于社交网络分析、推荐系统等诸多关键领域,然而,传统的大规模图计算系统面临冯诺依曼架构下访存带来的性能瓶颈。新型存内计算架构成为加速大规模图计算非常有前景的方案,尤其是非易失自旋磁存储器(MRAM)具备超高耐擦写性和超快写入等优点,可使图计算的存内实现更为高效。实现这种潜力的关键挑战之一是如何优化存内计算架构下的图算法设计。该文的前期工作表明,三角形计数算法和图连通分量计算算法可以通过按位运算实现,从而高效地部署在自旋存内处理核中加速。该文探索了更多图算法的优化实现,例如单源最短路径、K-core、链路预测,并提出了面向新型存内计算架构的图算法优化设计模型。该研究对于突破冯诺依曼架构下大规模图计算的内存访问瓶颈具有关键意义。     

Resistive Switching Properties of Sol–Gel-Derived V-Doped SrTiO3 Thin Films

V-doped and undoped SrTiO₃ (V:STO and STO) thin films on Pt/Ti/SiO₂/Si substrates were synthesized using a sol–gel method to form metal–insulator–metal (MIM) structures. Coexistence of the bipolar and unipolar resistive switching (BRS and URS) modes in Pt/STO/Pt and Pt/V:STO/Pt structures was observed as a irreversible transition from BRS to URS on adjustment of the compliance current (I _comp). Both states were stable and reproducible over 60 cycles, and the maximum operating voltage of the Pt/STO/Pt was reduced from 10 V to 2 V by doping with V. Linear fitting of current–voltage curves suggests that space-charge-limited leakage was the limiting leakage mechanism for these two devices. Based on these results, a switching mechanism based on filament theory is proposed to explain both resistive switching modes.     

Impact of Al in Cu alloy interconnects on electro and stress migration reliabilities

The reliability of Cu interconnects was successfully improved by applying a CuAl alloy seed. However, the effect of additive Al on the reliability is not fully understood. In order to reveal the reliability improvement mechanism, Cu films using CuAl alloy seed were investigated in detail. As stress induced voiding (SIV) as well as electromigration is caused by migration of vacancies and/or Cu atoms, the measured activation energy value of electromigration using CuAl indicates that the fast diffusion paths are Cu grain boundaries. The analysis using high lateral resolution scanning type secondary ion mass spectrometry (nano-SIMS) clarifies that additive Al in ECP-Cu film is mainly localized at grain boundaries. Furthermore, positron annihilation was used to probe vacancy-type defects in Cu films. The CuAl films before recrystallization contain larger and higher density vacancy-type defects. Whereas, the recrystallized CuAl films after annealing above 250 °C contain smaller and lower density defects. Furthermore, CuAl films with annealing above 350 °C contain less Al inside the grains. These results represent that Al atoms in Cu films with annealing above 350 °C are exhausted from inside grains to the grain boundaries, and the spewed Al atoms existing at Cu grain boundary effectively prevents the diffusion of Cu and/or vacancies.     

(Organic) Switching phenomenon in lateral structures: Tuning by gate voltage

We have fabricated devices based on copper phthalocyanine (CuPc) that exhibited electrical bistability and switching phenomenon. Apart from sandwiched structures, coplanar (lateral) electrodes or field-effect transistor structures have also been characterized. The observation of switching and memory phenomenon in lateral structures with Au/Au electrode combination primarily rules out (1) migration of metal ions into the organic materials, (2) metal filament formation through a redox-driven process, and (3) electroreduction of the molecules as possible mechanisms of electrical bistability in CuPc. The existence two conducting state and bistability have been explained in terms of trapping and detrapping of carriers in the molecular layer. Furthermore, the lateral structures show possibilities to tune switching processes through the bias of the base (gate) electrode.     

Low Power MoS2/Nb2O5 Memtransistor Device with Highly Reliable Heterosynaptic Plasticity

Artificial synapses based on 2D MoS₂ memtransistors have recently attracted considerable attention as a promising device architecture for complex neuromorphic systems. However, previous memtransistor devices occasionally cause uncontrollable analog switching and unreliable synaptic plasticity due to random variations in the field-induced defect migration. Herein, a highly reliable 2D MoS₂/Nb₂O₅ heterostructure memtransistor device is demonstrated, in which the Nb₂O₅ interlayer thickness is a critical material parameter to induce and tune analog switching characteristics of the 2D MoS₂. Ultraviolet photoelectron spectroscopy and photoluminescence analyses reveal that the Schottky barrier height at the 2D channel–electrode junction of the MoS₂/Nb₂O₅ heterostructure films is increased, leading to more effective contact barrier modulation and allowing more reliable resistive switching. The 2D/oxide memtransistors attain dual-terminal (drain and gate) stimulated heterosynaptic plasticity and highly precise multi-states. In addition, the memtransistor devices show an extremely low power consumption of ≈6 pJ and reliable potentiation/depression endurance characteristics over 2000 pulses. A high pattern recognition accuracy of ≈94.2% is finally achieved from the synaptic plasticity modulated by the drain pulse configuration using an image pattern recognition simulation. Thus, the novel 2D/oxide memtransistor makes a potential neuromorphic circuitry more flexible and energy-efficient, promoting the development of more advanced neuromorphic systems.     

Artificial Synapses Based on Multiterminal Memtransistors for Neuromorphic Application

Neuromorphic computing, which emulates the biological neural systems could overcome the high‐power consumption issue of conventional von‐Neumann computing. State‐of‐the‐art artificial synapses made of two‐terminal memristors, however, show variability in filament formation and limited capacity due to their inherent single presynaptic input design. Here, a memtransistor‐based arti?cial synapse is realized by integrating a memristor and selector transistor into a multiterminal device using monolayer polycrys‐talline‐MoS₂ grown by a scalable chemical vapor deposition (CVD) process. Notably, the memtransistor offers both drain‐ and gate‐tunable nonvolatile memory functions, which efficiently emulates the long‐term potentiation/depression, spike‐amplitude, and spike‐timing‐dependent plasticity of biological synapses. Moreover, the gate tunability function that is not achievable in two‐terminal memristors, enables significant bipolar resistive states switching up to four orders‐of‐magnitude and high cycling endurance. First‐principles calculations reveal a new resistive switching mechanism driven by the diffusion of double sulfur vacancy perpendicular to the MoS₂ grain boundary, leading to a conducting switching path without the need for a filament forming process. The seamless integration of multiterminal memtransistors may offer another degree‐of‐freedom to tune the synaptic plasticity by a third gate terminal for enabling complex neuromorphic learning.     

Control of Switching Modes and Conductance Quantization in Oxygen Engineered HfOx based Memristive Devices

Hafnium oxide (HfO_x)‐based memristive devices have tremendous potential as nonvolatile resistive random access memory (RRAM) and in neuromorphic electronics. Despite its seemingly simple two‐terminal structure, a myriad of RRAM devices reported in the rapidly growing literature exhibit rather complex resistive switching behaviors. Using Pt/HfO_x/TiN‐based metal–insulator–metal structures as model systems, it is shown that a well‐controlled oxygen stoichiometry governs the filament formation and the occurrence of multiple switching modes. The oxygen vacancy concentration is found to be the key factor in manipulating the balance between electric field and Joule heating during formation, rupture (reset), and reformation (set) of the conductive filaments in the dielectric. In addition, the engineering of oxygen vacancies stabilizes atomic size filament constrictions exhibiting integer and half‐integer conductance quantization at room temperature during set and reset. Identifying the materials conditions of different switching modes and conductance quantization contributes to a unified switching model correlating structural and functional properties of RRAM materials. The possibility to engineer the oxygen stoichiometry in HfO_x will allow creating quantum point contacts with multiple conductance quanta as a first step toward multilevel memristive quantum devices.     

A Polymer‐Electrolyte‐Based Atomic Switch

Studies on a resistive switching memory based on a silver‐ion‐conductive solid polymer electrolyte (SPE) are reported. Simple Ag/SPE/Pt structures containing polyethylene oxide–silver perchlorate complexes exhibit bipolar resistive switching under bias voltage sweeping. The switching behavior depends strongly on the silver perchlorate concentration. From the results of thermal, transport, and electrochemical measurements, it is concluded that the observed switching originates from formation and dissolution of a silver metal filament inside the SPE film caused by electrochemical reactions. This is the first report of an electrochemical “atomic switch” realized using an organic material. The devices also show ON/OFF resistance ratios greater than 10⁵, programming speeds higher than 1 μs, and retention times longer than 1 week. These results suggest that SPE‐based electrochemical devices might be suitable for flexible switch and memory applications.     

Neurotransmitter-Mediated Plasticity in 2D Perovskite Memristor for Reinforcement Learning

The neuromorphic computing architecture is a promising artificial intelligence for implementing hierarchical processing, in-memory computing, event-driven operation and functional specialization in computing systems. However, current investigations mainly focus on unisensory processing without objective experience which is contrary to the flexible sensory learning capability in the human brain that can sense and process information according to the ever-changing environment. For example, a dominant paradigm for reconfigurable bio-learning features is the emotional experience. The neurotransmitter dopamine is released during arousal, influencing the vital brain functions involved in cognition, reward learning, movement and motivation. Here, the on-demand configuration of a biorealistic synaptic connection based on a 2D CaTa₂O₇ (CTO) device is demonstrated that can be adaptively reconfigured for a reinforcement learning purpose by the light-active resistive switching, which originated from the photon-regulated metaplasticity. The low energy consumption of 12.4 fJ endows the reinforcement learning system with high power efficiency and reliability. Finally, in-sensor computing with a CTO synapse is implemented with a filtering function to process digital data in a neuromorphic engineering manner. This work demonstrates the feasibility of 2D perovskite neuromorphic device with enhanced biological plausibility in the approaching post-Moore era.     

Visualizing Thermally Activated Memristive Switching in Percolating Networks of Solution-Processed 2D Semiconductors

Memristive systems present a low-power alternative to silicon-based electronics for neuromorphic and in-memory computation. 2D materials have been increasingly explored for memristive applications due to their novel biomimetic functions, ultrathin geometry for ultimate scaling limits, and potential for fabricating large-area, flexible, and printed neuromorphic devices. While the switching mechanism in memristors based on single 2D nanosheets is similar to conventional oxide memristors, the switching mechanism in nanosheet composite films is complicated by the interplay of multiple physical processes and the inaccessibility of the active area in a two-terminal vertical geometry. Here, the authors report thermally activated memristors fabricated from percolating networks of diverse solution-processed 2D semiconductors including MoS₂, ReS₂, WS₂, and InSe. The mechanisms underlying threshold switching and negative differential resistance are elucidated by designing large-area lateral memristors that allow the direct observation of filament and dendrite formation using in situ spatially resolved optical, chemical, and thermal analyses. The high switching ratios (up to 10³) that are achieved at low fields (≈4 kV cm⁻¹) are explained by thermally assisted electrical discharge that preferentially occurs at the sharp edges of 2D nanosheets. Overall, this work establishes percolating networks of solution-processed 2D semiconductors as a platform for neuromorphic architectures.     

Emerging MXene-Based Memristors for In-Memory,Neuromorphic Computing,and Logic Operation

Confronted by the difficulties of the von Neumann bottleneck and memory wall, traditional computing systems are gradually inadequate for satisfying the demands of future data-intensive computing applications. Recently, memristors have emerged as promising candidates for advanced in-memory and neuromorphic computing, which pave one way for breaking through the dilemma of current computing architecture. Till now, varieties of functional materials have been developed for constructing high-performance memristors. Herein, the review focuses on the emerging 2D MXene materials-based memristors. First, the mainstream synthetic strategies and characterization methods of MXenes are introduced. Second, the different types of MXene-based memristive materials and their widely adopted switching mechanisms are overviewed. Third, the recent progress of MXene-based memristors for data storage, artificial synapses, neuromorphic computing, and logic circuits is comprehensively summarized. Finally, the challenges, development trends, and perspectives are discussed, aiming to provide guidelines for the preparation of novel MXene-based memristors and more engaging information technology applications.     

基于噪声调制的薄膜电阻加热温度控制系统

基于脉宽调制（PWM）技术的控制系统广泛应用于工业和日常生活中,随机脉宽调制（RP-WM）技术也已提出.利用耦合到系统的环境噪声和系统本身的噪声经加工处理后调制控制信号,由噪声信号的自相关函数和功率谱密度函数给出噪声调制脉冲的平均开关频率和平均加热功率的计算方法.利用模拟器件设计了一个简单的基于噪声调制的薄膜电阻加热温度控制系统.实验表明：控制系统能够稳定地运行,在一定范围内能抑制噪声,实现了高精度温度控制.     

超声楔键合Al/Ni系统高温存储过程界面扩散行为(英文)

应用电子扫描电镜(SEM)及X射线能谱(EDX)分析了直径为25μm的Al+1％Si引线与 Au／Ni／Cu焊盘键合后以及老化过程中界面间的冶金行为,结果表明:接头形成于键合点塑性流动 最大的周边区域;超声引线键合的连接本质是压力使引线发生塑性流动导致Al元素与Ni元素之 间的扩散,而超声一方面使金属引线软化增强了塑性流动的程度,另一方面超声使引线内部产生大 量的缺陷,成为扩散的通道,大大加速了扩散的进行;短路扩散是键合点形成的主要机制。在170℃ 时,键合点空气中高温存储后X射线能谱线扫描分析结果表明:老化10 d时,有明显的Ni向Al引 线内扩散现象;老化30 d时,引线内部出现孔洞以及裂纹,界面出现云状组织,成份分析为15．55％ Ni和78．82％Al;老化40 d时,引线内部出现大量的孔洞并存在方块岛状的Al-Ni组织,其尺寸和 形状显示与Kirkendall孔洞不同。     

Direct Observation of Conducting Nanofilaments in Graphene‐Oxide‐Resistive Switching Memory

Determining the presence of conducting filaments in resistive random access memory with nanoscale thin films is vital to unraveling resistive switching mechanisms. Bistable resistive switching within graphene‐oxide (GO)‐based resistive memory devices, recently developed by many research groups, has been generally explained by the formation and rupture of conducting filaments induced by the diffusion of metal or oxygen ions. Using a low‐voltage spherical aberration‐corrected transmission electron microscopy (TEM), we directly observe metallic nanofilaments formed at the amorphous top interface layer with the application of external voltages in an Al/GO/Al memory system. Atomic‐resolution TEM images acquired at an acceleration voltage of 80 kV clearly show that the conducting nanofilaments are composed of nanosized aluminum crystalline within the amorphous top interface layer after applying a negative bias (ON state). Simultaneously, we observe the change in the crystallinity of GO films by the back‐diffusion of oxygen ions. The oxygen‐deficient regions are clearly confirmed by energy‐filtered TEM oxygen elemental mapping. This work could provide strong evidence to confirm the resistive switching mechanism previously suggested by our group.     

Ionization and diffusion of metal atoms under electric field at metal/insulator interfaces; First-principles study

The ionization and diffusion of metal atoms at metal/SiO₂ interfaces under an electric field are studied by first-principles calculation. It is shown that the ionization of metal atoms occurs when the hybridization of metal-atom electronic states with metal-induced gap states (MIGS) is broken. Moreover, we show that an electric field markedly decreases the penetration barrier of metal atoms from a metal electrode into SiO₂ and enhances the diffusion of metal atoms.     

Resistance switching of the nonstoichiometric zirconium oxide for nonvolatile memory applications

The resistance switching behavior and switching mechanism of nonstoichiometric zirconium oxide thin films were investigated for nonvolatile memory application. The Pt/ZrO/sub x//p/sup +/-Si sandwich structure fabricated by reactive sputtering shows two stable resistance states. By applying proper bias, resistance switching from one to another state can be obtained. The composition in ZrO/sub x/ thin films were confirmed from X-ray photoelectron spectroscope (XPS) analysis, which showed three layers such as top stoichiometric ZrO/sub 2/ layer with high resistance, transition region with medium resistance, and conducting ZrO/sub x/ bulk layer. The resistance switching can be explained by electron trapping and detrapping of excess Zr/sup +/ ions in transition layer which control the distribution of electric field inside the oxide, and, hence the current flow.