Polymer Analog Memristive Synapse with Atomic-Scale Conductive Filament for Flexible Neuromorphic Computing System.

With the advent of artificial intelligence (AI), memristors have received significant interest as a synaptic building block for neuromorphic systems, where each synaptic memristor should operate in an analog fashion, exhibiting multilevel accessible conductance states. Here, we demonstrate that the transition of the operation mode in poly(1,3,5-trivinyl-1,3,5-trimethyl cyclotrisiloxane) (pV3D3)-based flexible memristor from conventional binary to synaptic analog switching can be achieved simply by reducing the size of the formed filament. With the quantized conductance states observed in the flexible pV3D3 memristor, analog potentiation and depression characteristics of the memristive synapse are obtained through the growth of atomically thin Cu filament and lateral dissolution of the filament via dominant electric field effect, respectively. The face classification capability of our memristor is evaluated via simulation using an artificial neural network consisting of pV3D3 memristor synapses. These results will encourage the development of soft neuromorphic intelligent systems.

Metal-etching-free direct delamination and transfer of single-layer graphene with a high degree of freedom.

A method of graphene transfer without metal etching is developed to minimize the contamination of graphene in the transfer process and to endow the transfer process with a greater degree of freedom. The method involves direct delamination of single-layer graphene from a growth substrate, resulting in transferred graphene with nearly zero Dirac voltage due to the absence of residues that would originate from metal etching. Several demonstrations are also presented to show the high degree of freedom and the resulting versatility of this transfer method.

Preventing Vanishing Gradient Problem of Hardware Neuromorphic System by Implementing Imidazole-Based Memristive ReLU Activation Neuron.

With advances in artificial intelligent services, brain-inspired neuromorphic systems with synaptic devices are recently attracting significant interest to circumvent the von Neumann bottleneck. However, the increasing trend of deep neural network parameters causes huge power consumption and large area overhead of a nonlinear neuron electronic circuit, and it incurs a vanishing gradient problem. Here, a memristor-based compact and energy-efficient neuron device is presented to implement a rectifying linear unit (ReLU) activation function. To emulate the volatile and gradual switching of the ReLU function, a copolymer memristor with a hybrid structure is proposed using a copolymer/inorganic bilayer. The functional copolymer film developed by introducing imidazole functional groups enables the formation of nanocluster-type pseudo-conductive filaments by boosting the nucleation of Cu nanoclusters, causing gradual switching. The ReLU neuron device is successfully demonstrated by integrating the memristor with amorphous InGaZnO thin-film transistors, and achieves 0.5 pJ of energy consumption based on sub-10 µA operation current and high-speed switching of 650 ns. Furthermore, device-to-system-level simulation using neuron devices on the MNIST dataset demonstrates that the vanishing gradient problem is effectively resolved by five-layer deep neural networks. The proposed neuron device will enable the implementation of high-density and energy-efficient hardware neuromorphic systems.

Imidazole-based artificial synapses for neuromorphic computing: a cluster-type conductive filament via controllable nanocluster nucleation.

Memristive synapses based on conductive bridging RAMs (CBRAMs) utilize a switching layer having low binding energy with active metals for excellent analog conductance modulation, but the resulting unstable conductive filaments cause fluctuation and drift of the conductance. This tunability-stability dilemma makes it difficult to implement practical neuromorphic computing. A novel method is proposed to enhance the stability and controllability of conductive filaments by introducing imidazole groups that boost the nucleation of Cu nanoclusters in the ultrathin polymer switching layer through the initiated chemical vapor deposition (iCVD) process. It is confirmed that conductive filaments based on nanoclusters with specific gaps are generated in the copolymer medium using this method. Furthermore, by modulating the tunneling gaps, an ultra-wide conductance range of analog tunable conductive filaments is achieved from several hundreds of nS to a few mS with a sub-1 V driving voltage. Through this, both reliable and stable analog switching are achieved with low cycle-to-cycle and device-to-device weight update variations and separable state retention with 32 states. This approach paves the way for the extension of state availability in synaptic devices to overcome the tunability-stability dilemma, which is essential for the synaptic elements in neuromorphic systems.

Spatially isolated neutral excitons via clusters on trilayer MoS2.

In spite of having a large exciton binding energy, two-dimensional (2D) transition metal dichalcogenides (TMDs) are limited as light-emitting materials because the spectral weight of neutral excitons decreases exponentially with increasing the excitation density. That is, neutral excitons easily transfer to trions, and exciton-exciton annihilation (EEA) occurs due to the strengthening of exciton kinetic energy in the layered structure. In here, we come up with an isolated neutral exciton system, maintaining its high spectral weight when the carrier density increased, which is achieved via MoS2 clusters on a MoS2 trilayer directly synthesized by metal-organic chemical vapor deposition (MOCVD). While increasing the excitation density, trions are decomposed by spatial confinement at the saturation level of its full width at half maximum (FWHM), and simultaneously the spectral weight of neutral excitons restarts to increase. Furthermore, we reveal the causality relationship between trions and B excitons, providing a keen insight into organic interactions among radiative recombination processes in 2D TMDs.

Wafer-Scale Uniform Growth of an Atomically Thin MoS2 Film with Controlled Layer Numbers by Metal-Organic Chemical Vapor Deposition.

The growth control of a molybdenum disulfide (MoS2) thin film, including the number of layers, growth rate, and electrical property modulation, remains a challenge. In this study, we synthesized MoS2 thin films using the metal-organic chemical vapor deposition (MOCVD) method with a 2 inch wafer scale and achieved high thickness uniformity according to the positions on the substrate. In addition, we successfully controlled the number of MoS2 layers to range from one to five, with a growth rate of 10 min per layer. The layer-dependent optical and electrical properties were characterized by photoluminescence, Raman spectroscopy, differential reflectance spectroscopy, and field effect transistors. To guide the growth of MoS2, we summarized the relation between the growth aspects and the precursor control in the form of a growth map. Reference to this growth map enabled control of the growth rate, domain density, and domain size according to the application purposes. Finally, we confirmed the electrical performance of MOCVD-grown MoS2 with five layers under a high-κ dielectric environment, which exhibited an on/off current ratio of 10∼6 and a maximum field effect mobility of 8.6 cm2 V-1 s-1.

Conductive-bridging random-access memories for emerging neuromorphic computing.

With the increasing utilisation of artificial intelligence, there is a renewed demand for the development of novel neuromorphic computing owing to the drawbacks of the existing computing paradigm based on the von Neumann architecture. Extensive studies have been performed on memristors as their electrical nature is similar to those of biological synapses and neurons. However, most hardware-based artificial neural networks (ANNs) have been developed with oxide-based memristors owing to their high compatibility with mature complementary metal-oxide-semiconductor (CMOS) processes. Considering the advantages of conductive-bridging random-access memories (CBRAMs), such as their high scalability, high on-off current with a wide dynamic range, and low off-current, over oxide-based memristors, extensive studies on CBRAMs are required. In this review, the basics of operation of CBRAMs are examined in detail, from the formation of metal nanoclusters to filament bridging. Additionally, state-of-the-art experimental demonstrations of CBRAM-based artificial synapses and neurons are presented. Finally, CBRAM-based ANNs are discussed, including deep neural networks and spiking neural networks, along with other emerging computing applications. This review is expected to pave the way toward further development of large-scale CBRAM array systems.

Low-Thermal-Budget Doping of 2D Materials in Ambient Air Exemplified by Synthesis of Boron-Doped Reduced Graphene Oxide.

Graphene oxide (GO) doping and reduction allow for physicochemical property modification to suit practical application needs. Herein, the challenge of simultaneous low-thermal-budget heteroatom doping of GO and its reduction in ambient air is addressed through the synthesis of B-doped reduced GO (B@rGO) by flash irradiation of boric acid loaded onto a GO support with intense pulsed light (IPL). The effects of light power and number of shots on the in-depth sequential doping and reduction mechanisms are investigated by ex situ X-ray photoelectron spectroscopy and direct millisecond-scale temperature measurements (temperature >1600 °C, < 10-millisecond duration, ramping rate of 5.3 × 105 °C s-1). Single-flash IPL allows the large-scale synthesis of substantially doped B@rGO (≈3.60 at% B) to be realized with a thermal budget 106-fold lower than that of conventional thermal methods, and the prepared material with abundant B active sites is employed for highly sensitive and selective room-temperature NO2 sensing. Thus, this work showcases the great potential of optical annealing for millisecond-scale ultrafast reduction and heteroatom doping of GO in ambient air, which allows the tuning of multiple physicochemical GO properties.

Enhanced electrical percolation due to interconnection of three-dimensional pentacene islands in thin films on low surface energy polyimide gate dielectrics.

The role of lateral interconnections between three-dimensional pentacene islands on low surface energy polyimide gate dielectrics was investigated by the measurement of the surface coverage dependence of the charge mobility and the use of conducting-probe atomic force microscopy (CP-AFM). From the correlation between the electrical characteristics and the morphological evolution of the three-dimensionally grown pentacene films-based field-effect transistors, we found that during film growth, the formation of interconnections between the three-dimensional pentacene islands that are isolated at the early stage contributes significantly to the enhancement process of charge mobility. The CP-AFM current mapping images of the pentacene films also indicate that the lateral interconnections play an important role in the formation of good electrical percolation pathways between the three-dimensional pentacene islands.

Flexible Nonvolatile Polymer Memory Array on Plastic Substrate via Initiated Chemical Vapor Deposition.

Resistive random access memory based on polymer thin films has been developed as a promising flexible nonvolatile memory for flexible electronic systems. Memory plays an important role in all modern electronic systems for data storage, processing, and communication; thus, the development of flexible memory is essential for the realization of flexible electronics. However, the existing solution-processed, polymer-based RRAMs have exhibited serious drawbacks in terms of the uniformity, electrical stability, and long-term stability of the polymer thin films. Here, we present poly(1,3,5-trimethyl-1,3,5-trivinyl cyclotrisiloxane) (pV3D3)-based RRAM arrays fabricated via the solvent-free technique called initiated chemical vapor deposition (iCVD) process for flexible memory application. Because of the outstanding chemical stability of pV3D3 films, the pV3D3-RRAM arrays can be fabricated by a conventional photolithography process. The pV3D3-RRAM on flexible substrates showed unipolar resistive switching memory with an on/off ratio of over 10(7), stable retention time for 10(5) s, excellent cycling endurance over 10(5) cycles, and robust immunity to mechanical stress. In addition, pV3D3-RRAMs showed good uniformity in terms of device-to-device distribution. The pV3D3-RRAM will pave the way for development of next-generation flexible nonvolatile memory devices.

Stretchable, multiplexed pH sensors with demonstrations on rabbit and human hearts undergoing ischemia.

Stable pH is an established biomarker of health, relevant to all tissues of the body, including the heart. Clinical monitoring of pH in a practical manner, with high spatiotemporal resolution, is particularly difficult in organs such as the heart due to its soft mechanics, curvilinear geometry, heterogeneous surfaces, and continuous, complex rhythmic motion. The results presented here illustrate that advanced strategies in materials assembly and electrochemical growth can yield interconnected arrays of miniaturized IrOx pH sensors encapsulated in thin, low-modulus elastomers to yield conformal monitoring systems capable of noninvasive measurements on the surface of the beating heart. A thirty channel custom data acquisition system enables spatiotemporal pH mapping with a single potentiostat. In vitro testing reveals super-Nernstian sensitivity with excellent uniformity (69.9 ± 2.2 mV/pH), linear response to temperature (-1.6 mV °C(-1) ), and minimal influence of extracellular ions (<3.5 mV). Device examples include sensor arrays on balloon catheters and on skin-like stretchable membranes. Real-time measurement of pH on the surfaces of explanted rabbit hearts and a donated human heart during protocols of ischemia-reperfusion illustrate some of the capabilities. Envisioned applications range from devices for biological research, to surgical tools and long-term implants.

Elastomer surfaces with directionally dependent adhesion strength and their use in transfer printing with continuous roll-to-roll applications.

In this paper we present mechanics and materials aspects of elastomeric stamps that have angled features of relief on their surfaces, designed to enable control of adhesion strength by peeling direction, in a way that can be exploited in schemes for deterministic assembly by transfer printing. Detailed mechanics models capture the essential physics of interface adhesion in this system. Experiments with cylindrical stamps that have this design demonstrate their potential for use in a continuous, roller mode of operation.

Electrochemical transistors with ionic liquids for enzymatic sensing.

We report an enzymatic sensor based on an organic electrochemical transistor that uses a room temperature ionic liquid as an integral part of its structure and as an immobilization medium for the enzyme and the mediator.