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

No abstract

Memristors with tunable resistance states are emerging building blocks of artificial neural networks. However, in situ learning on a large-scale multiple-layer memristor network has yet to be demonstrated because of challenges in device property engineering and circuit integration. Here we monolithically integrate hafnium oxide-based memristors with a foundry-made transistor array into a multiple-layer neural network. We experimentally demonstrate in situ learning capability and achieve competitive classification accuracy on a standard machine learning dataset, which further confirms that the training algorithm allows the network to adapt to hardware imperfections. Our simulation using the experimental parameters suggests that a larger network would further increase the classification accuracy. The memristor neural network is a promising hardware platform for artificial intelligence with high speed-energy efficiency.

Using memristor crossbar arrays to accelerate computations is a promising approach to efficiently implement algorithms in deep neural networks. Early demonstrations, however, are limited to simulations or small-scale problems primarily due to materials and device challenges that limit the size of the memristor crossbar arrays that can be reliably programmed to stable and analog values, which is the focus of the current work. High-precision analog tuning and control of memristor cells across a 128 × 64 array is demonstrated, and the resulting vector matrix multiplication (VMM) computing precision is evaluated. Single-layer neural network inference is performed in these arrays, and the performance compared to a digital approach is assessed. Memristor computing system used here reaches a VMM accuracy equivalent of 6 bits, and an 89.9% recognition accuracy is achieved for the 10k MNIST handwritten digit test set. Forecasts show that with integrated (on chip) and scaled memristors, a computational efficiency greater than 100 trillion operations per second per Watt is possible.

A novel Ag/oxide‐based threshold switching device with attractive features including ≈1010 nonlinearity is developed. High‐resolution transmission electron microscopic analysis of the nanoscale crosspoint device suggests that elongation of an Ag nanoparticle under voltage bias followed by spontaneous reformation of a more spherical shape after power off is responsible for the observed threshold switching.