We evaluate the sensitivity of neuromorphic inference accelerators based on silicon-oxide-nitride-oxide-silicon (SONOS) charge trap memory arrays to total ionizing dose (TID) effects. Data retention statistics were collected for 16 Mbit of 40-nm SONOS digital memory exposed to ionizing radiation from a Co-60 source, showing good retention of the bits up to the maximum dose of 500 krad(Si). Using this data, we formulate a rate-equation-based model for the TID response of trapped charge carriers in the ONO stack and predict the effect of TID on intermediate device states between 'program' and 'erase.' This model is then used to simulate arrays of low-power, analog SONOS devices that store 8-bit neural network weights and support in situ matrix-vector multiplication. We evaluate the accuracy of the irradiated SONOS-based inference accelerator on two image recognition tasks - CIFAR-10 and the challenging ImageNet data set - using state-of-the-art convolutional neural networks, such as ResNet-50. We find that across the data sets and neural networks evaluated, the accelerator tolerates a maximum TID between 10 and 100 krad(Si), with deeper networks being more susceptible to accuracy losses due to TID.
In-memory computing based on non-volatile resistive memory can significantly improve the energy efficiency of artificial neural networks. However, accurate in situ training has been challenging due to the nonlinear and stochastic switching of the resistive memory elements. One promising analog memory is the electrochemical random-access memory (ECRAM), also known as the redox transistor. Its low write currents and linear switching properties across hundreds of analog states enable accurate and massively parallel updates of a full crossbar array, which yield rapid and energy-efficient training. While simulations predict that ECRAM based neural networks achieve high training accuracy at significantly higher energy efficiency than digital implementations, these predictions have not been experimentally achieved. In this work, we train a 3 × 3 array of ECRAM devices that learns to discriminate several elementary logic gates (AND, OR, NAND). We record the evolution of the network’s synaptic weights during parallel in situ (on-line) training, with outer product updates. Due to linear and reproducible device switching characteristics, our crossbar simulations not only accurately simulate the epochs to convergence, but also quantitatively capture the evolution of weights in individual devices. The implementation of the first in situ parallel training together with strong agreement with simulation results provides a significant advance toward developing ECRAM into larger crossbar arrays for artificial neural network accelerators, which could enable orders of magnitude improvements in energy efficiency of deep neural networks.
The III-nitride semiconductors have many attractive properties for field-emission vacuum electronics, including high thermal and chemical stability, low electron affinity, and high breakdown fields. Here, we report top-down fabricated gallium nitride (GaN)-based nanoscale vacuum electron diodes operable in air, with record ultralow turn-on voltages down to ∼0.24 V and stable high field-emission currents, tested up to several microamps for single-emitter devices. We leverage a scalable, top-down GaN nanofabrication method leading to damage-free and smooth surfaces. Gap-dependent and pressure-dependent studies provide new insights into the design of future, integrated nanogap vacuum electron devices. The results show promise for a new class of high-performance and robust, on-chip, III-nitride-based vacuum nanoelectronics operable in air or reduced vacuum.
Digital computing is nearing its physical limits as computing needs and energy consumption rapidly increase. Analogue-memory-based neuromorphic computing can be orders of magnitude more energy efficient at data-intensive tasks like deep neural networks, but has been limited by the inaccurate and unpredictable switching of analogue resistive memory. Filamentary resistive random access memory (RRAM) suffers from stochastic switching due to the random kinetic motion of discrete defects in the nanometer-sized filament. In this work, this stochasticity is overcome by incorporating a solid electrolyte interlayer, in this case, yttria-stabilized zirconia (YSZ), toward eliminating filaments. Filament-free, bulk-RRAM cells instead store analogue states using the bulk point defect concentration, yielding predictable switching because the statistical ensemble behavior of oxygen vacancy defects is deterministic even when individual defects are stochastic. Both experiments and modeling show bulk-RRAM devices using TiO2-X switching layers and YSZ electrolytes yield deterministic and linear analogue switching for efficient inference and training. Bulk-RRAM solves many outstanding issues with memristor unpredictability that have inhibited commercialization, and can, therefore, enable unprecedented new applications for energy-efficient neuromorphic computing. Beyond RRAM, this work shows how harnessing bulk point defects in ionic materials can be used to engineer deterministic nanoelectronic materials and devices.
It is well established that the miniaturization of batteries has not kept pace with the miniaturization of electronics. Three-dimensional (3D) batteries, which were developed with the intent of improving microbattery performance, have had limited success because of fabrication challenges and material constraints. Solid-state, 3D batteries have been particularly susceptible to these shortcomings. In this paper, we demonstrate that the incorporation of a high-conductivity, solid electrolyte is the key to achieving a nonplanar solid-state battery with high areal capacity and high power density. The model 2.5D platform used in this study is a modification of the more typical 3D configuration in that it is comprised of a cathode array of pillars (3D) and a planar (two-dimensional, 2D) anode. This 2.5D geometry exploits the use of a high-conductivity, ionogel electrolyte (10-3 S cm-1), which interpenetrates the 3D electrode array. The 2.5D battery offers high areal energy densities from the post array, while the high-conductivity, solid electrolyte enables high power densities (3.7 mWh cm-2 at 2.8 mW cm-2). The reported solid-state 2.5D device exceeds the energy and power densities of any 3D solid-state system and the derived multiphysics model provides guidance for achieving significantly higher energy and power densities.
The ability to train deep neural networks on large data sets have made significant impacts onto artificial intelligence, but consume significant amounts of energy due to the need to move information from memory to logic units. In-memory "neuromorphic" computing presents an alternative framework that processes information directly on memory elements. In-memory computing has been limited by the poor performance of the analogue information storage element, often phase-change memory or memristors. To solve this problem, we developed two types of "redox transistors" using TiO2 (anatase) which stores analogue information states through the electrochemical concentration of dopants in the crystal. The first type of redox transistor uses lithium as the electrochemical dopant ion, and its key advantage is low operating voltage. The second uses oxygen vacancies as the dopant, which is CMOS compatible and can retain state even when scaled to nanosized dimensions. Both devices offer significant advantages in terms of predictable analogue switching over conventional filamentary-based devices, and provide a significant advance in developing materials and devices for neuromorphic computing.
Materials exhibiting metal-to-insulator transitions (MITs) could enable low power neuromorphic computing, but progress is hindered by insufficient mechanistic understanding. In this issue of Matter, Banerjee and colleagues describe with intricate detail a new MIT mechanism in β′-CuxV2O5, with potential applications to neuromorphic computing. Materials exhibiting metal-to-insulator transitions (MITs) could enable low power neuromorphic computing, but progress is hindered by insufficient mechanistic understanding. In this issue of Matter, Banerjee and colleagues describe with intricate detail a new MIT mechanism in β′-CuxV2O5, with potential applications to neuromorphic computing.
Terahertz (THz) technology has shown promise for several applications, but limitations in sources and detectors have prevented broader adoption. Existing THz detectors are rigid, planar, and fabricated using complex technology, making it difficult to integrate into systems. Here we demonstrate THz detectors fabricated by inkjet printing on submicrometer thick, ultraflexible substrates. By developing p- and n-type carbon nanotube inks, we achieve optically thick p–n junction and p-type devices, enabling antenna-free pixels for THz imaging. By further designing and folding the printed devices, we realize origami-inspired architectures with improved performance over single devices, achieving a noise-equivalent power of 12 nW/Hz1/2 at room temperature with no voltage bias. Our approach opens avenues for nonplanar, foldable, deployable, insertable, and retractable THz detectors for applications in nondestructive inspection.
Inspired by the in-memory computing architectures of biological systems, neuromorphic computing using crossbar arrays of artificial synapses based on non-volatile memory (NVM) devices with variable conductances has emerged as a new paradigm to enable massively parallel and ultra-low power computing hardware for data centric applications. Although inference has been demonstrated successfully using crossbars based on a variety of NMV technologies, efficient learning and scaling to large arrays (>106 elements) remains a challenge due to the synaptic elements' non-ideal electrical characteristics which degrades ANN accuracy. A further challenge is that in the conductive state memristors draw large currents >μA resulting in significant voltage drops in the interconnect wires and increased probability of failure in scaled arrays. We suggest the organic polymer redox transistor (RT) is an alternate approach that could solve many of these challenges, enabling both inference and parallel outer product updates, as recently demonstrated by Fuller et al. An RT consists of redox-active channel and gate electrodes in contact with a liquid or solid electrolyte. lon insertion through the electrolyte controls the channel electronic conductivity, while electron transfer through an external circuit maintains overall charge neutrality. Unlike a rechargeable battery, in the RT the voltage built-up across the electrolyte is kept to a minimum (typically <100 mV) by using the same material for the gate and channel. Elimination of the voltage offset simplifies integration of the RT into programmable arrays by enabling the use of various selectors. RTs based on inorganic and organic materials have been recently demonstrated with conductance tuning occurring at potentials of just a few mV and hundreds to thousands of linearly and symmetrically programmable conductance states, enabling near ideal accuracy in neural network simulations. Introduced in the 1980's, redox transistors with metallic gate electrodes and organic channel materials, also known as organic electrochemical transistors (OECTs), have been explored for a variety of applications such as chem- and bio-sensing, neural interfaces, and low cost printed circuits. A typical channel material for OECTs is the conducting polymer poly(3,4-ethylenedioxythiophene) doped with poly(styrene sulfonate) (PEDOT:PSS). PEDOT is a p-type semiconducting polymer with mobile positively charged polarons that hop chain-to-chain.