Yi Li

Ultrafast Synaptic Events in a Chalcogenide Memristor

Compact and power-efficient plastic electronic synapses are of fundamental importance to overcoming the bottlenecks of developing a neuromorphic chip. Memristor is a strong contender among the various electronic synapses in existence today. However, the speeds of synaptic events are relatively slow in most attempts at emulating synapses due to the material-related mechanism. Here we revealed the intrinsic memristance of stoichiometric crystalline Ge2Sb2Te5 that originates from the charge trapping and releasing by the defects. The device resistance states, representing synaptic weights, were precisely modulated by 30 ns potentiating/depressing electrical pulses. We demonstrated four spike-timing-dependent plasticity (STDP) forms by applying programmed pre- and postsynaptic spiking pulse pairs in different time windows ranging from 50 ms down to 500 ns, the latter of which is 105 times faster than the speed of STDP in human brain. This study provides new opportunities for building ultrafast neuromorphic computing systems and surpassing Von Neumann architecture.

Activity-Dependent Synaptic Plasticity of a Chalcogenide Electronic Synapse for Neuromorphic Systems

Nanoscale inorganic electronic synapses or synaptic devices, which are capable of emulating the functions of biological synapses of brain neuronal systems, are regarded as the basic building blocks for beyond-Von Neumann computing architecture, combining information storage and processing. Here, we demonstrate a Ag/AgInSbTe/Ag structure for chalcogenide memristor-based electronic synapses. The memristive characteristics with reproducible gradual resistance tuning are utilised to mimic the activity-dependent synaptic plasticity that serves as the basis of memory and learning. Bidirectional long-term Hebbian plasticity modulation is implemented by the coactivity of pre- and postsynaptic spikes, and the sign and degree are affected by assorted factors including the temporal difference, spike rate and voltage. Moreover, synaptic saturation is observed to be an adjustment of Hebbian rules to stabilise the growth of synaptic weights. Our results may contribute to the development of highly functional plastic electronic synapses and the further construction of next-generation parallel neuromorphic computing architecture.

16 Boolean logics in three steps with two anti-serially connected memristors

Memristor based logic gates that can execute memory and logic operations are regarded as building blocks for non Von Neumann computation architecture. In this letter, Ta/GeTe/Ag memristors were fabricated and showed reproducible binary switches between high-resistance and low-resistance states. Utilizing a structure with two anti-serially connected memristors, we propose a logic operation methodology, based on which arbitrary Boolean logic can be realized in three steps, and the logic result can be nonvolatilely stored. A functionally complete logic operation: NAND is further verified by HSPICE simulation and experiments. The implementation of logic-in-memory unit may stimulate the development of future massive parallel computing.

Functionally Complete Boolean Logic in 1T1R Resistive Random Access Memory

Nonvolatile stateful logic through RRAM is a promising route to build in-memory computing architecture. In this letter, a logic methodology based on 1T1R structure has been proposed to implement functionally complete Boolean logics. Arbitrary logic functions could be realized in two steps: initialization and writing. An additional read step is required to read out the logic result, which is in situ stored in the nonvolatile resistive state of the memory. Cascade problem in building larger logic circuits is also discussed. Our 1T1R logic device and operation method could be beneficial for massive integration and practical application of RRAM-based logic.

Polarization compression of large dynamic range laser returned signals from water surface and underwater

This paper describes a polarization detection method to compress a large dynamic range of laser returned signals from water surface and underwater in an airborne laser bathymetry system. The experimental results show that the method can compress the dynamic range of returned signals by more than one order of magnitude. The laser reflection on the water surface and underwater backscatter is considerably reduced without significantly affecting the laser returned signals from the underwater target.

Memcomputing: fusion of memory and computing

In this world of ubiquitous computing, the contemporary electronic digital computer has become an essential machine that is available anytime and everywhere. Computing systems come in various forms, including desktops, laptops, tablets, mobile phones, smart watches, and other daily life devices. These systems all originate from the earliest huge, heavy EDVAC and UNIVAC machines of the 1940s, and all share a universal architecture conceived by von Neumann, who divided the computing system into five primary groups: a central arithmetic part (CA), a central control part (CC), memory (M) and outside recording medium (R), input, and output. The CA and CC parts evolved into the central processing unit (CPU), whereas the M and R correspond to the high-speed main memory that stores data and instructions (SRAM and DRAM) and external mass storage, respectively.

Reconfigurable logic in nanosecond Cu/GeTe/TiN filamentary memristors for energy-efficient in-memory computing

Owing to the capability of integrating the information storage and computing in the same physical location, in-memory computing with memristors has become a research hotspot as a promising route for non von Neumann architecture. However, it is still a challenge to develop high performance devices as well as optimized logic methodologies to realize energy-efficient computing. Herein, filamentary Cu/GeTe/TiN memristor is reported to show satisfactory properties with nanosecond switching speed (<60 ns), low voltage operation (<2 V), high endurance (>104 cycles) and good retention (>104 s @85 °C). It is revealed that the charge carrier conduction mechanisms in high resistance and low resistance states are Schottky emission and hopping transport between the adjacent Cu clusters, respectively, based on the analysis of current-voltage behaviors and resistance-temperature characteristics. An intuitive picture is given to describe the dynamic processes of resistive switching. Moreover, based on the basic material implication (IMP) logic circuit, we proposed a reconfigurable logic method and experimentally implemented IMP, NOT, OR, and COPY logic functions. Design of a one-bit full adder with reduction in computational sequences and its validation in simulation further demonstrate the potential practical application. The results provide important progress towards understanding of resistive switching mechanism and realization of energy-efficient in-memory computing architecture.

Model of dielectric breakdown in hafnia-based ferroelectric capacitors

Ultra-thin ferroelectric hafnia-based thin films are very promising candidates for nanoscale ferroelectric random access memories. However, dielectric breakdown is a main failure mechanism during repeated polarization switching. Generalizing Lou et al.s local phase decomposition model, originally for ferroelectric fatigue, we propose a dielectric breakdown model for ferroelectric hafnia. While charging injection during the polarization reversal is regarded as a key step, eventual phase separation of the Hf cluster accounts for the dielectric breakdown. Using this model, we explain why TaN/HfO2/TaN ferroelectric capacitors are more prone to dielectric breakdown than TiN/HfO2/TiN, and conclude that the lower Schottky barrier for the TaN/Pca21-HfO2 interface stabilizes neutral oxygen vacancies within the dielectric. On the other hand, when TiN electrodes are employed, oxygen vacancies tend to be positively charged. They can further pin the domain walls, resulting in ferroelectric fatigue. The relationship between the conductive filament formation, dielectric breakdown, wake up, and fatigue in ferroelectric HfO2 is discussed within the framework of our model.Ultra-thin ferroelectric hafnia-based thin films are very promising candidates for nanoscale ferroelectric random access memories. However, dielectric breakdown is a main failure mechanism during repeated polarization switching. Generalizing Lou et al.s local phase decomposition model, originally for ferroelectric fatigue, we propose a dielectric breakdown model for ferroelectric hafnia. While charging injection during the polarization reversal is regarded as a key step, eventual phase separation of the Hf cluster accounts for the dielectric breakdown. Using this model, we explain why TaN/HfO2/TaN ferroelectric capacitors are more prone to dielectric breakdown than TiN/HfO2/TiN, and conclude that the lower Schottky barrier for the TaN/Pca21-HfO2 interface stabilizes neutral oxygen vacancies within the dielectric. On the other hand, when TiN electrodes are employed, oxygen vacancies tend to be positively charged. They can further pin the domain walls, resulting in ferroelectric fatigue. The relationship b...

Theoretical investigation of the Ag filament morphology in conductive bridge random access memories

Conductive bridge random access memories (CBRAMs) usually involve active Ag or Cu metals, where the formation of metal filaments accounts for the low resistance state. For the application of neuromorphic computation, it is highly desirable to develop artificial neurons and synapses, which utilize the complicated volatile or nonvolatile resistive switching phenomena, respectively. This can be achieved by controlling the morphology and stability of the filaments, which requires a deep understanding of the filament formation and disruption mechanisms. Using ab initio calculations, we explored the physical mechanism behind various Ag filament morphologies and growth modes, using GeSe, ZrO2, SiO2, and a-Si as the examples. The roles of Ag and Ag+ stability inside the dielectric, the migration barrier of Ag+, and the Ag+ solvation effect have been investigated in detail. A comprehensive model has been proposed, which in particular could explain the diverse Ag filament morphology experimentally observed in sputtered SiO2 and PECVD SiO2. Our theoretical approach can serve as a pre-screening method in designing new solid-state electrolyte materials of CBRAM, aiming at new functionalities in neuromorphic computation or in-memory logic computing.Conductive bridge random access memories (CBRAMs) usually involve active Ag or Cu metals, where the formation of metal filaments accounts for the low resistance state. For the application of neuromorphic computation, it is highly desirable to develop artificial neurons and synapses, which utilize the complicated volatile or nonvolatile resistive switching phenomena, respectively. This can be achieved by controlling the morphology and stability of the filaments, which requires a deep understanding of the filament formation and disruption mechanisms. Using ab initio calculations, we explored the physical mechanism behind various Ag filament morphologies and growth modes, using GeSe, ZrO2, SiO2, and a-Si as the examples. The roles of Ag and Ag+ stability inside the dielectric, the migration barrier of Ag+, and the Ag+ solvation effect have been investigated in detail. A comprehensive model has been proposed, which in particular could explain the diverse Ag filament morphology experimentally observed in sputt...

Correlation analysis between the current fluctuation characteristics and the conductive filament morphology of HfO2-based memristor

Memristors are attracting considerable interest for their prospective applications in nonvolatile memory, neuromorphic computing, and in-memory computing. However, the nature of resistance switching is still under debate, and current fluctuation in memristors is one of the critical concerns for stable performance. In this work, random telegraph noise (RTN) as the indication of current instabilities in distinct resistance states of the Pt/Ti/HfO2/W memristor is thoroughly investigated. Standard two-level digital-like RTN, multilevel current instabilities with non-correlation/correlation defects, and irreversible current transitions are observed and analyzed. The dependence of RTN on the resistance and read bias reveals that the current fluctuation depends strongly on the morphology and evolution of the conductive filament composed of oxygen vacancies. Our results link the current fluctuation behaviors to the evolution of the conductive filament and will guide continuous optimization of memristive devices.