Revealing Strong Flexoelectricity and Optoelectronic Coupling in 2D Ferroelectric CuInP2S6 Via Large Strain Gradient.

The interplay between flexoelectric and optoelectronic characteristics provides a paradigm for studying emerging phenomena in various 2D materials. However, an effective way to induce a large and tunable strain gradient in 2D devices remains to be exploited. Herein, we propose a strategy to induce large flexoelectric effect in 2D ferroelectric CuInP2S6 by constructing a 1D-2D mixed-dimensional heterostructure. The strong flexoelectric effect is induced by enormous strain gradient up to 4.2 × 106 m-1 resulting from the underlying ZnO nanowires, which is further confirmed by the asymmetric coercive field and the red-shift in the absorption edge. The induced flexoelectric polarization efficiently boosts the self-powered photodetection performance. In addition, the improved photoresponse has a good correlation with the induced strain gradient, showing a consistent size-dependent flexoelectric effect. The mechanism of flexoelectric and optoelectronic coupling is proposed based on the Landau-Ginzburg-Devonshire double-well model, supported by density functional theory (DFT) calculations. This work provides a brand-new method to induce a strong flexoelectric effect in 2D materials, which is not restricted to crystal symmetry and thus offers unprecedented opportunities for state-of-the-art 2D devices.

Adaptive Neural Activation and Neuromorphic Processing via Drain-Injection Threshold-Switching Float Gate Transistor Memory.

Hetero-modulated neural activation is vital for adaptive information processing and learning that occurs in brain. To implement brain-inspired adaptive processing, previously various neurotransistors oriented for synaptic functions are extensively explored, however, the emulation of nonlinear neural activation and hetero-modulated behaviors are not possible due to the lack of threshold switching behavior in a conventional transistor structure. Here, a 2D van der Waals float gate transistor (FGT) that exhibits steep threshold switching behavior, and the emulation of hetero-modulated neuron functions (integrate-and-fire, sigmoid type activation) for adaptive sensory processing, are reported. Unlike conventional FGTs, the threshold switching behavior stems from impact ionization in channel and the coupled charge injection to float gate. When a threshold is met, a sub-30 mV dec-1 increase of transistor conductance by more than four orders is triggered with a typical switch time of approximately milliseconds. Essentially, by feeding light sensing signal as the modulation input, it is demonstrated that two typical tasks that rely on adaptive neural activation, including collision avoidance and adaptive visual perception, can be realized. These results may shed light on the emulation of rich hetero-modulating behaviors in biological neurons and the realization of biomimetic neuromorphic processing at low hardware cost.

Simultaneously ultrafast and robust two-dimensional flash memory devices based on phase-engineered edge contacts.

As the prevailing non-volatile memory (NVM), flash memory offers mass data storage at high integration density and low cost. However, due to the 'speed-retention-endurance' dilemma, their typical speed is limited to ~microseconds to milliseconds for program and erase operations, restricting their application in scenarios with high-speed data throughput. Here, by adopting metallic 1T-LixMoS2 as edge contact, we show that ultrafast (10-100 ns) and robust (endurance>106 cycles, retention>10 years) memory operation can be simultaneously achieved in a two-dimensional van der Waals heterostructure flash memory with 2H-MoS2 as semiconductor channel. We attribute the superior performance to the gate tunable Schottky barrier at the edge contact, which can facilitate hot carrier injection to the semiconductor channel and subsequent tunneling when compared to a conventional top contact with high density of defects at the metal interface. Our results suggest that contact engineering can become a strategy to further improve the performance of 2D flash memory devices and meet the increasing demands of high speed and reliable data storage.

Scalable integration of hybrid high-κ dielectric materials on two-dimensional semiconductors.

Two-dimensional (2D) semiconductors are promising channel materials for next-generation field-effect transistors (FETs). However, it remains challenging to integrate ultrathin and uniform high-κ dielectrics on 2D semiconductors to fabricate FETs with large gate capacitance. We report a versatile two-step approach to integrating high-quality dielectric film with sub-1 nm equivalent oxide thickness (EOT) on 2D semiconductors. Inorganic molecular crystal Sb2O3 is homogeneously deposited on 2D semiconductors as a buffer layer, which forms a high-quality oxide-to-semiconductor interface and offers a highly hydrophilic surface, enabling the integration of high-κ dielectrics via atomic layer deposition. Using this approach, we can fabricate monolayer molybdenum disulfide-based FETs with the thinnest EOT (0.67 nm). The transistors exhibit an on/off ratio of over 106 using an ultra-low operating voltage of 0.4 V, achieving unprecedently high gating efficiency. Our results may pave the way for the application of 2D materials in low-power ultrascaling electronics.

Approaching strain limit of two-dimensional MoS2 via chalcogenide substitution.

Strain engineering is a promising method for tuning the electronic properties of two-dimensional (2D) materials, which are capable of sustaining enormous strain thanks to their atomic thinness. However, applying a large and homogeneous strain on these 2D materials, including the typical semiconductor MoS2, remains cumbersome. Here we report a facile strategy for the fabrication of highly strained MoS2 via chalcogenide substitution reaction (CSR) of MoTe2 with lattice inheritance. The MoS2 resulting from the sulfurized MoTe2 sustains ultra large in-plane strain (approaching its strength limit ~10%) with great homogeneity. Furthermore, the strain can be deterministically and continuously tuned to ~1.5% by simply varying the processing temperature. Thanks to the fine control of our CSR process, we demonstrate a heterostructure of strained MoS2/MoTe2 with abrupt interface. Finally, we verify that such a large strain potentially allows the modulation of MoS2 bandgap over an ultra-broad range (~1 eV). Our controllable CSR strategy paves the way for the fabrication of highly strained 2D materials for applications in devices.

A non-linear two-dimensional float gate transistor as a lateral inhibitory synapse for retinal early visual processing.

Synaptic transistors that accommodate concurrent signal transmission and learning in a neural network are attracting enormous interest for neuromorphic sensory processing. To remove redundant sensory information while keeping important features, artificial synaptic transistors with non-linear conductance are desired to apply filter processing to sensory inputs. Here, we report the realization of non-linear synapses using a two-dimensional van der Waals (vdW) heterostructure (MoS2/h-BN/graphene) based float gate memory device, in which the semiconductor channel is tailored via a surface acceptor (ZnPc) for subthreshold operation. In addition to usual synaptic plasticity, the memory device exhibits highly non-linear conductance (rectification ratio >106), allowing bidirectional yet only negative/inhibitory current to pass through. We demonstrate that in a lateral coupling network, such a float gate memory device resembles the key lateral inhibition function of horizontal cells for the formation of an ON-center/OFF-surround receptive field. When combined with synaptic plasticity, the lateral inhibition weights are further tunable to enable adjustable edge enhancement for early visual processing. Our results here hopefully open a new scheme toward early sensory perception via lateral inhibitory synaptic transistors.

High-Performance Planar-Type Photodetector Based on Hot-Pressed CsPbBr3 Wafer.

Considering the disadvantages of the common methods for CsPbBr3 single crystal growth including the high cost of the melt method and the low shape controllability of the solution method, a facile hot-pressed (HP) approach has been introduced to prepare CsPbBr3 wafers. The effects of HP temperature on the phase purity of HP-CsPbBr3 wafers and the performance of the corresponding photodetectors have been investigated. The HP temperature for preparing phase-pure, shape-regular, and dense CsPbBr3 wafers has been optimized to be 150 °C, and the HP-CsPbBr3 wafer based planar-type photodetectors exhibit an ultrasensitive weak light photoresponse. Under the illumination of a 530 nm LED with a light power density of 1.1 µW cm-2, the responsivity, external quantum efficiency, and detectivity of the devices reach 19.79 A W-1, 4634%, and 2.14 × 1013 Jones, respectively, and a fast response speed with a rise time of 40.5 µs and a fall time of 10.0 µs has been achieved.

A Reconfigurable Two-WSe2 -Transistor Synaptic Cell for Reinforcement Learning.

Reward-modulated spike-timing-dependent plasticity (R-STDP) is a brain-inspired reinforcement learning (RL) rule, exhibiting potential for decision-making tasks and artificial general intelligence. However, the hardware implementation of the reward-modulation process in R-STDP usually requires complicated Si complementary metal-oxide-semiconductor (CMOS) circuit design that causes high power consumption and large footprint. Here, a design with two synaptic transistors (2T) connected in a parallel structure is experimentally demonstrated. The 2T unit based on WSe2 ferroelectric transistors exhibits reconfigurable polarity behavior, where one channel can be tuned as n-type and the other as p-type due to nonvolatile ferroelectric polarization. In this way, opposite synaptic weight update behaviors with multilevel (>6 bit) conductance states, ultralow nonlinearity (0.56/-1.23), and large Gmax /Gmin ratio of 30 are realized. By applying positive/negative reward to (anti-)STDP component of 2T cell, R-STDP learning rules are realized for training the spiking neural network and demonstrated to solve the classical cart-pole problem, exhibiting a way for realizing low-power (32 pJ per forward process) and highly area-efficient (100 µm2 ) hardware chip for reinforcement learning.

Ultrafast and stable phase transition realized in MoTe2-based memristive devices.

Phase engineering of two-dimensional transition metal dichalcogenides has received increasing attention in recent years due to its atomically thin nature and polymorphism. Here, we first realize an electric-field-induced controllable phase transition between semiconducting 2H and metallic 1T' phases in MoTe2 memristive devices. The device performs stable bipolar resistive switching with a cycling endurance of over 105, an excellent retention characteristic of over 105 s at an elevated temperature of 85 °C and an ultrafast switching of ∼5 ns for SET and ∼10 ns for RESET. More importantly, the device works in different atmospheres including air, vacuum and oxygen, and even works with no degradation after being placed in air for one year, indicating excellent surrounding and time stability. In situ Raman analysis reveals that the stable resistive switching originates from a controllable phase transition between 2H and 1T' phases. Density functional theory calculations reveal that the Te vacancy facilitates the phase transition in MoTe2 through decreasing the barrier between 2H and 1T' phases, and serving as nucleation sites due to the elimination of repulsive forces. This electric-field-induced controllable phase transition in MoTe2 devices offers new opportunities for developing reliable and ultrafast phase transition devices based on atomically thin membranes.

Complementary Memtransistor-Based Multilayer Neural Networks for Online Supervised Learning Through (Anti-)Spike-Timing-Dependent Plasticity.

We propose a complete hardware-based architecture of multilayer neural networks (MNNs), including electronic synapses, neurons, and periphery circuitry to implement supervised learning (SL) algorithm of extended remote supervised method (ReSuMe). In this system, complementary (a pair of n- and p-type) memtransistors (C-MTs) are used as an electrical synapse. By applying the learning rule of spike-timing-dependent plasticity (STDP) to the memtransistor connecting presynaptic neuron to the output one whereas the contrary anti-STDP rule to the other memtransistor connecting presynaptic neuron to the teacher one, extended ReSuMe with multiple layers is realized without the usage of those complicated supervising modules in previous approaches. In this way, both the C-MT-based chip area and power consumption of the learning circuit for weight updating operation are drastically decreased comparing with the conventional single memtransistor (S-MT)-based designs. Two typical benchmarks, the linearly nonseparable benchmark XOR problem and Mixed National Institute of Standards and Technology database (MNIST) recognition have been successfully tackled using the proposed MNN system while impact of the nonideal factors of realistic devices has been evaluated.

Threshold switching memristor-based stochastic neurons for probabilistic computing.

Biological neurons exhibit dynamic excitation behavior in the form of stochastic firing, rather than stiffly giving out spikes upon reaching a fixed threshold voltage, which empowers the brain to perform probabilistic inference in the face of uncertainty. However, owing to the complexity of the stochastic firing process in biological neurons, the challenge of fabricating and applying stochastic neurons with bio-realistic dynamics to probabilistic scenarios remains to be fully addressed. In this work, a novel CuS/GeSe conductive-bridge threshold switching memristor is fabricated and singled out to realize electronic stochastic neurons, which is ascribed to the similarity between the stochastic switching behavior observed in the device and that of biological ion channels. The corresponding electric circuit of a stochastic neuron is then constructed and the probabilistic firing capacity of the neuron is utilized to implement Bayesian inference in a spiking neural network (SNN). The application prospects are demonstrated on the example of a tumor diagnosis task, where common fatal diagnostic errors of a conventional artificial neural network are successfully circumvented. Moreover, in comparison to deterministic neuron-based SNNs, the stochastic neurons enable SNNs to deliver an estimate of the uncertainty in their predictions, and the fidelity of the judgement is drastically improved by 81.2%.

2D Cu9 S5 /PtS2 /WSe2 Double Heterojunction Bipolar Transistor with High Current Gain.

Bipolar junction transistor (BJT) as one important circuit element is now widely used in high-speed computation and communication for its capability of high-power signal amplification. 2D materials and their heterostructures are promising in building high-amplification and high-frequency BJTs because they can be naturally thin and highly designable in tailoring components properties. However, currently the low emitter injection efficiency results in only moderate current gain achieved in the pioneer researches, severely restraining its future development. Herein, it is shown that an elaborately designed double heterojunction bipolar transistor (DHBT) can greatly promote the injection efficiency, improving the current gain by order of magnitude. In this DHBT high-doping-density wide-bandgap 2D Cu9 S5 is used as emitter and narrow-bandgap PtS2 as base. This heterostructure efficiently suppresses the reverse electron flux from base and increase the injection efficiency. Consequently, the DHBT achieves an excellent current gain (ß ≈ 910). This work systematically explores the electrical behavior of 2D materials based DHBT, and provides deep insight of the architecture design for building high gain DHBT, which may promote the applications of 2Dheterojunctions in the fields of integrated circuits.

Design and tailoring of two-dimensional Schottky, PN and tunnelling junctions for electronics and optoelectronics.

Owing to their superior carrier mobility, strong light-matter interactions, and flexibility at the atomically thin thickness, two-dimensional (2D) materials are attracting wide interest for application in electronic and optoelectronic devices, including rectifying diodes, transistors, memory, photodetectors, and light-emitting diodes. At the heart of these devices, Schottky, PN, and tunneling junctions are playing an essential role in defining device function. Intriguingly, the ultrathin thickness and unique van der Waals (vdW) interlayer coupling in 2D materials has rendered enormous opportunities for the design and tailoring of various 2D junctions, e.g. using Lego-like hetero-stacking, surface decoration, and field-effect modulation methods. Such flexibility has led to marvelous breakthroughs during the exploration of 2D electronics and optoelectronic devices. To advance further, it is imperative to provide an overview of existing strategies for the engineering of various 2D junctions for their integration in the future. Thus, in this review, we provide a comprehensive survey of previous efforts toward 2D Schottky, PN, and tunneling junctions, and the functional devices built from them. Though these junctions exhibit similar configurations, distinct strategies have been developed for their optimal figures of merit based on their working principles and functional purposes.

Enhancing LiAlOX synaptic performance by reducing the Schottky barrier height for deep neural network applications.

Although good performance has been reported in shallow neural networks, the application of memristor synapses towards realistic deep neural networks has met more stringent requirements on the synapse properties, particularly the high precision and linearity of the synaptic analog weight tuning. In this study, a LiAlOX memristor synapse was fabricated and optimized to address these demands. By delicately tuning the initial conductance states, 120-level continuously adjustable conductance states were obtained and the nonlinearity factor was substantially reduced from 8.96 to 0.83. The significant enhancements were attributed to the reduced Schottky barrier height (SBH) between the filament tip and the electrode, which was estimated from the measured I-V curves. Furthermore, a deep neural network for realistic action recognition task was constructed, and the recognition accuracy was found to be increased from 15.1% to 91.4% on the Weizmann video dataset by adopting the above-described device optimization method.

The mechanism of the modulation of electronic anisotropy in two-dimensional ReS2.

Although the anisotropy and strategies for the modulation of the anisotropy in ReS2 have been widely reported, a comprehensive study on the inherent electronic anisotropy of ReS2 is still absent to date; therefore, the mechanism of anisotropy evolution is ambiguous as well. In this study, we have conducted a systematic investigation on the evolution of electronic anisotropy in bilayer ReS2, under the modulation of charge doping levels and temperature. It is found that the adjustability of electronic anisotropy is largely attributed to the angle-dependent scattering from defects or vacancies at a low doping level. At a high doping level, in contrast, the inherent electronic anisotropy can be recovered by filling the traps to attenuate the influence of scattering. This work renders insights into the exploration of electronic anisotropy in 2D materials.

PbSe Quantum Dots Sensitized High-Mobility Bi2O2Se Nanosheets for High-Performance and Broadband Photodetection Beyond 2 µm.

As an emerging two-dimensional semiconductor, Bi2O2Se has recently attracted broad interests in optoelectronic devices for its superior mobility and ambient stability, whereas the diminished photoresponse near its inherent indirect bandgap (0.8 eV or λ = 1550 nm) severely restricted its application in the broad infrared spectra. Here, we report the Bi2O2Se nanosheets based hybrid photodetector for short wavelength infrared detection up to 2 µm via PbSe colloidal quantum dots (CQDs) sensitization. The type II interfacial band offset between PbSe and Bi2O2Se not only enhanced the device responsivity compared to bare Bi2O2Se but also sped up the response time to ∼4 ms, which was ∼300 times faster than PbSe CQDs. It was further demonstrated that the photocurrent in such a zero-dimensional-two-dimensional hybrid photodetector could be efficiently tailored from a photoconductive to photogate dominated response under external field effects, thereby rendering a sensitive infrared response >103 A/W at 2 µm. The excellent performance up to 2 µm highlights the potential of field-effect modulated Bi2O2Se-based hybrid photodetectors in pursuing highly sensitive and broadband photodetection.

Reconfigurable two-dimensional optoelectronic devices enabled by local ferroelectric polarization.

Ferroelectric engineered pn doping in two-dimensional (2D) semiconductors hold essential promise in realizing customized functional devices in a reconfigurable manner. Here, we report the successful pn doping in molybdenum disulfide (MoS2) optoelectronic device by local patterned ferroelectric polarization, and its configuration into lateral diode and npn bipolar phototransistors for photodetection from such a versatile playground. The lateral pn diode formed in this way manifests efficient self-powered detection by separating ~12% photo-generated electrons and holes. When polarized as bipolar phototransistor, the device is customized with a gain ~1000 by its transistor action, reaching the responsivity ~12 A W-1 and detectivity over 1013 Jones while keeping a fast response speed within 20 µs. A promising pathway toward high performance optoelectronics is thus opened up based on local ferroelectric polarization coupled 2D semiconductors.

Nanochannel-Based Transport in an Interfacial Memristor Can Emulate the Analog Weight Modulation of Synapses.

By exploiting novel transport phenomena such as ion selectivity at the nanoscale, it has been shown that nanochannel systems can exhibit electrically controllable conductance, suggesting their potential use in neuromorphic devices. However, several critical features of biological synapses, particularly their conductance modulation, which is both memorable and gradual, have rarely been reported in these types of systems due to the fast flow property of typical inorganic electrolytes. In this work, we demonstrate that electrically manipulating the nanochannel conductance can result in nonvolatile conductance tuning capable of mimicking the analog behavior of synapses by introducing a room-temperature ionic liquid (IL) and a KCl solution into the two ends of a nanochannel system. The gradual conductance-tuning mechanism is identified through fluorescence measurements as the voltage-induced movement of the interface between the immiscible IL and KCl solution, while the successful memorization of the conductance tuning is ascribed to the large viscosity of the IL. We applied a nanochannel-based synapse to a handwritten digit-recognition task, reaching an accuracy of 94%. These promising results provide important guidance for the future design of nanochannel-based neuromorphic devices and the manipulation of nanochannel transport for computing.

Doping engineering and functionalization of two-dimensional metal chalcogenides.

Two-dimensional (2D) layered metal chalcogenides (MXs) have significant potential for use in flexible transistors, optoelectronics, sensing and memory devices beyond the state-of-the-art technology. To pursue ultimate performance, precisely controlled doping engineering of 2D MXs is desired for tailoring their physical and chemical properties in functional devices. In this review, we highlight the recent progress in the doping engineering of 2D MXs, covering that enabled by substitution, exterior charge transfer, intercalation and the electrostatic doping mechanism. A variety of novel doping engineering examples leading to Janus structures, defect curing effects, zero-valent intercalation and deliberately devised floating gate modulation will be discussed together with their intriguing application prospects. The choice of doping strategies and sources for functionalizing MXs will be provided to facilitate ongoing research in this field toward multifunctional applications.

Van der Waals Coupled Organic Molecules with Monolayer MoS2 for Fast Response Photodetectors with Gate-Tunable Responsivity.

As a direct-band-gap transition metal dichalcogenide (TMD), atomic thin MoS2 has attracted extensive attention in photodetection, whereas the hitherto unsolved persistent photoconductance (PPC) from the ungoverned charge trapping in devices has severely hindered their employment. Herein, we demonstrate the realization of ultrafast photoresponse dynamics in monolayer MoS2 by exploiting a charge transfer interface based on surface-assembled zinc phthalocyanine (ZnPc) molecules. The formed MoS2/ZnPc van der Waals interface is found to favorably suppress the PPC phenomenon in MoS2 by instantly separating photogenerated holes toward the ZnPc molecules, away from the traps in MoS2 and the dielectric interface. The derived MoS2 detector then exhibits significantly improved photoresponse speed by more than 3 orders (from over 20 s to less than 8 ms for the decay) and a high responsivity of 430 A/W after Al2O3 passivation. It is also demonstrated that the device could be further tailored to be 2-10-fold more sensitive without severely sacrificing the ultrafast response dynamics using gate modulation. The strategy presented here based on surface-assembled organic molecules may thus pave the way for realizing high-performance TMD-based photodetection with ultrafast speed and high sensitivity.