High-Performance Neuromorphic Computing and Logic Operation Based on a Self-Assembled Vertically Aligned Nanocomposite SrTiO₃:MgO Film Memristor

Abstract

Neuromorphic computing based on memristors capable of in-memory computing is promising to break the energy and efficiency bottleneck of well-known von Neumann architectures. However, unstable and nonlinear conductance updates compromise the recognition accuracy and block the integration of neural network hardware. To this end, we present a highly stable memristor with self-assembled vertically aligned nanocomposite (VAN) SrTiO3:MgO films that achieve excellent resistive switching with low set/reset voltage variability (4.7%/–5.6%) and highly linear conductivity variation (nonlinearity = 0.34) by spatially limiting the conductive channels at the vertical interfaces. Various synaptic behaviors are simulated by continuously modulating the conductance. Especially, convolutional image processing using diverse crossbar kernels is demonstrated, and the artificial neural network achieves an overwhelming recognition accuracy of up to 97.50% for handwritten digits. Even under the perturbation of Poisson noise (λ = 10), 6% Salt and Pepper noise, and 5% Gaussian noise, the high recognition accuracies are retained at 95.43%, 94.56%, and 95.97%, respectively. Importantly, the logic memory function is proven experimentally based on the nonvolatile properties. This work provides a material system and design idea to achieve high-performance neuromorphic computing and logic operation.

Self-Powered Optoelectronic Synaptic Devices for Neuromorphic Computing with the Lowest Energy Consumption Density

“…We further evaluated the perovskite quality by using X-ray diffraction (XRD) and atomic force microscopy (AFM) measurements, as shown in Figures d and S1 respectively, which exhibited an oriented (100) facet for the perovskite film with a small arithmetic mean roughness of ∼8 nm, − thereby confirming the promising photoelectric response of the devices. The typical dark current–voltage curves in Figure e exhibit obvious hysteresis from a batch of 18 consecutive back-and-forth scans, thus indicating that the device has good memristive characteristics, which is expected to benefit neuromorphic computing. − Figure f demonstrates the optical absorption spectrum of CsFAMA perovskite film ranging from 400 to 800 nm and the external quantum efficiencies (EQEs) of the whole structured device. Arising from the high visible-light absorbance of the triple cation perovskite and efficient carrier collection with a well-designed configuration, the overall EQE of the device exceeds 80% in the most light responding range, which assures the mimicking of the synaptic functions under various light wavelengths with limited energy consumption densities.…”

Enhanced Memristive Performance via a Vertically Heterointerface in Nanocomposite Thin Films for Artificial Synapses

“…Some studies have achieved STDP simulation through pulse engineering . The STDP function could also be realized by employing a straightforward pulse sequence approach (inset on Figure d), which has been widely employed to achieve STDP in devices operating via mechanisms involving OV conductive wires. ,− Figure d presents the STDP property of the nanocomposite film device, which was trained by the inset sequence. The representation of synaptic weight change is defined as normalΔ w = ( G normala − G normalb G b ) × 100 , where G a and G b represent the device conductivity before and after pulse stimulation, respectively.…”

“…However, there is limited work on utilizing the vertical heterointerface in vertically aligned nanocomposites (VANs) as the conductive channels in resistive random access memory for artificial neuromorphic computing, except the very recent work on the SrTiO 3 –MgO-based memristor . While certain prior investigations have examined VANs and their enhanced resistive switching properties, they predominantly focused on ionic conductive materials, with limited exploration into their applicability in artificial neuromorphic computing. , Here, the ZnO–BaTiO 3 (BTO) nanocomposite thin film has been designed as the functional layer and a Au/ZnO–BTO/0.7 wt % Nb-doped SrTiO 3 (NSTO) memristor device has been fabricated, as illustrated in Figure a.…”

PBDB-T/Pentacene-Based Organic Optoelectronic Synaptic Transistor with Adjustable Critical Flicker Fusion Frequency for Dynamic Vision

“…In the era of big data, processing and storage of massive visual information pose heightened demands on real-time image processing and storage devices. − The human visual system possesses the capability for rapid, efficient, and energy-efficient visual information processing in complex environments. , Inspired by this, neuromorphic vision systems integrate photosensors, information processing units, and data storage units to achieve complex image processing. − Within this context, light-gated organic synaptic devices (organic optoelectronic synapses) have been developed to enhance the light-sensing capability to traditional electronic synapses. , This additional capability opens up opportunities for analog image preprocessing and recognition. ,− Moreover, it is imperative to recognize that beyond the realm of static image recognition and processing, the identification and detection of moving objects have become equally significant areas of research. Several studies have proposed novel optoelectronic transistor arrays to implement dynamic visual systems. , The challenge in device recognition of moving objects lies in determining object velocity or the photosensitivity frequency of the device itself.…”