Photoferroelectric All-van-der-Waals Heterostructure for Multimode Neuromorphic Ferroelectric Transistors

Van der Waals Inverted-Floating-Gate Transistors for Artificial Intelligence Electronics

An inverted floating gate device architecture is introduced, demonstrated with all-van-der-Waals technology, targeting both logic and neuromorphic circuits. Integrating a top polymorphic multilayer graphene floating gate improves the electrostatic coupling to the ReS2 semiconductor channel by facilitating efficient dynamic conductance tuning and enabling dual-mode reconfigurable logic and memory operations. The non-volatile capability is used to implement compact logic gates for in-memory computing. The device is also shown to emulate synaptic plasticity, with an accuracy of 92% demonstrated in simple artificial neural network simulations. Moreover, spiking neuron circuits for neural networks through a five-transistor design makes it a versatile building block for artificial intelligence electronics. These findings demonstrate the potential of hybrid integration of van der Waals materials to address the limitations of traditional semiconductor technologies and become key to developments of next-generation electronics.

Wordline Input Bias Scheme for Neural Network Implementation in 3D-NAND Flash

In this study, we propose a neuromorphic computing system based on a 3D-NAND flash architecture that utilizes analog input voltages applied through wordlines (WLs). The approach leverages the velocity saturation effect in short-channel MOSFETs, which enables a linear increase in drain current with respect to gate voltage in the saturation region. A NAND flash array with a TANOS (TiN/Al2O3/Si3N4/SiO2/poly-Si) gate stack was fabricated, and its electrical and reliability characteristics were evaluated. Output characteristics of short-channel (L = 1 µm) and long-channel (L = 50 µm) devices were compared, confirming the linear behavior of short-channel devices due to velocity saturation. In the proposed system, analog WL voltages serve as inputs, and the summed bitline (BL) currents represent the outputs. Each synaptic weight is implemented using two paired devices, and each WL layer corresponds to a fully connected (FC) layer, enabling efficient vector-matrix multiplication (VMM). MNIST pattern recognition is conducted, demonstrated only a 0.32% accuracy drop for the short-channel device compared to the ideal linear case, and 0.95% degradation under 0.5 V threshold variation, while maintaining robustness. These results highlight the strong potential of 3D-NAND flash memory, which offers high integration density and technological maturity, for neuromorphic computing applications.

2D Reconfigurable Memory for Integrated Optical Sensing and Multifunctional Image Processing

Recently, the growing demand for data-centric applications has significantly accelerated progress to overcome the "memory wall" caused by the separation of image sensing, memory, and computing units. However, despite advancements in novel devices driving the development of the in-sensor computing paradigm, achieving seamless integration of optical sensing, storage, and image processing within a single device remains challenging. This study demonstrates an in-sensor computing architecture using a ferroelectric-defined reconfigurable α-In2Se3 phototransistor. The three polarization states of the device exhibit a linear and distinguishable photoresponse, with a maximum photoresponse current difference of 2.17 × 10-6 A and a retention time exceeding 500 s. The nonvolatile weight and synaptic properties are programmed by external electrical stimulation, enabling 112 distinct conductance states with a nonlinearity of 0.12. Additionally, the device supports efficient optical writing, electrical erasing, optoelectronic logic, and decoding via combined optoelectronic control. In-sensor computation for image edge detection is simulated by embedding a nonvolatile Prewitt convolution kernel into a 3 × 3 device array. The integrated structure and array design highlight the strong potential of 2D ferroelectric semiconductors for in-sensor computing, providing a promising platform for next-generation multifunctional artificial vision systems.

Two-Dimensional Materials for Brain-Inspired Computing Hardware

Recent breakthroughs in brain-inspired computing promise to address a wide range of problems from security to healthcare. However, the current strategy of implementing artificial intelligence algorithms using conventional silicon hardware is leading to unsustainable energy consumption. Neuromorphic hardware based on electronic devices mimicking biological systems is emerging as a low-energy alternative, although further progress requires materials that can mimic biological function while maintaining scalability and speed. As a result of their diverse unique properties, atomically thin two-dimensional (2D) materials are promising building blocks for next-generation electronics including nonvolatile memory, in-memory and neuromorphic computing, and flexible edge-computing systems. Furthermore, 2D materials achieve biorealistic synaptic and neuronal responses that extend beyond conventional logic and memory systems. Here, we provide a comprehensive review of the growth, fabrication, and integration of 2D materials and van der Waals heterojunctions for neuromorphic electronic and optoelectronic devices, circuits, and systems. For each case, the relationship between physical properties and device responses is emphasized followed by a critical comparison of technologies for different applications. We conclude with a forward-looking perspective on the key remaining challenges and opportunities for neuromorphic applications that leverage the fundamental properties of 2D materials and heterojunctions.

Two-Dimensional Ferroelectric Materials: From Prediction to Applications

Ferroelectric materials hold immense potential for diverse applications in sensors, actuators, memory storage, and microelectronics. The discovery of two-dimensional (2D) ferroelectrics, particularly ultrathin compounds with stable crystal structure and room-temperature ferroelectricity, has led to significant advancements in the field. However, challenges such as depolarization effects, low Curie temperature, and high energy barriers for polarization reversal remain in the development of 2D ferroelectrics with high performance. In this review, recent progress in the discovery and design of 2D ferroelectric materials is discussed, focusing on their properties, underlying mechanisms, and applications. Based on the work discussed in this review, we look ahead to theoretical prediction for 2D ferroelectric materials and their potential applications, such as the application in nonlinear optics. The progress in theoretical and experimental research could lead to the discovery and design of next-generation nanoelectronic and optoelectronic devices, facilitating the applications of 2D ferroelectric materials in emerging advanced technologies.

Full van der Waals Ambipolar Ferroelectric Configurable Optical Hetero-Synapses for In-Sensor Computing

The rapid development of visual neuromorphic hardware can be attributed to their ability to capture, store and process optical signals from the environment. The main limitation of existing visual neuromorphic hardware is that the realization of complex functions is premised on the increase of manufacturing cost, hardware volume and energy consumption. In this study, we demonstrated an optical synaptic device based on a three-terminal van der Waals (vdW) heterojunction that can realize the sensing functions of light wavelength and intensity as well as short-term and long-term synaptic plasticity. In the image recognition task, we constructed an optical reservoir neural network (ORNN) and a visible light communication system (VLC) composed of this optical synaptic device. The ORNN has a recognition rate of up to 84.9% for 50 000 color images in 10 categories in the CIFAR-10 color image dataset, and the VLC system can achieve high-speed transmission with an ultra-low power consumption of only 0.4 nW. This work shows that through reasonable design, vdW heterojunction structures have great application potential in low-power multifunctional fusion application tasks such as visual bionics.

Bio-Plausible Multimodal Learning with Emerging Neuromorphic Devices

Multimodal machine learning, as a prospective advancement in artificial intelligence, endeavors to emulate the brain's multimodal learning abilities with the objective to enhance interactions with humans. However, this approach requires simultaneous processing of diverse types of data, leading to increased model complexity, longer training times, and higher energy consumption. Multimodal neuromorphic devices have the capability to preprocess spatio-temporal information from various physical signals into unified electrical signals with high information density, thereby enabling more biologically plausible multimodal learning with low complexity and high energy-efficiency. Here, this work conducts a comparison between the expression of multimodal machine learning and multimodal neuromorphic computing, followed by an overview of the key characteristics associated with multimodal neuromorphic devices. The bio-plausible operational principles and the multimodal learning abilities of emerging devices are examined, which are classified into heterogeneous and homogeneous multimodal neuromorphic devices. Subsequently, this work provides a detailed description of the multimodal learning capabilities demonstrated by neuromorphic circuits and their respective applications. Finally, this work highlights the limitations and challenges of multimodal neuromorphic computing in order to hopefully provide insight into potential future research directions.

Hafnia-Based Ferroelectric Transistor with Poly-Si Gates for Gate-First Three-Dimensional NAND Structures

Ferroelectric transistors based on hafnia-based ferroelectrics have emerged as promising candidates for next-generation memory devices. Additionally, hafnia-based ferroelectric transistors are suggested for three-dimensional (3D) memory devices, such as 3D ferroelectric NAND. This paper investigates the utilization of poly-Si as a gate material for hafnia-based ferroelectric transistors in 3D NAND structures. Conventional gate materials, such as TiN or W, are usually deposited in 3D NAND structures by using the gate-last process, which requires an additional gate replacement process. We demonstrate that poly-Si can be used as a gate material for hafnia-based ferroelectric transistors. We show that the 3D ferroelectric NAND based on the poly-Si gate can be fabricated by a simpler gate-first process without requiring a gate replacement process. Our findings underscore the potential of poly-Si as a gate material for ferroelectric transistors and 3D ferroelectric NAND.

Optical-Electrical Coordinately Modulated Memristor Based on 2D Ferroelectric RP Perovskite for Artificial Vision Applications

Traditional artificial vision systems built using separate sensing, computing, and storage units have problems with high power consumption and latency caused by frequent data transmission between functional units. An effective approach is to transfer some memory and computing tasks to the sensor, enabling the simultaneous perception-storage-processing of light signals. Here, an optical-electrical coordinately modulated memristor is proposed, which controls the conductivity by means of polarization of the 2D ferroelectric Ruddlesden-Popper perovskite film at room temperature. The residual polarization shows no significant decay after 109-cycle polarization reversals, indicating that the device has high durability. By adjusting the pulse parameters, the device can simulate the bio-synaptic long/short-term plasticity, which enables the control of conductivity with a high linearity of ≈0.997. Based on the device, a two-layer feedforward neural network is built to recognize handwritten digits, and the recognition accuracy is as high as 97.150%. Meanwhile, building optical-electrical reserve pool system can improve 14.550% for face recognition accuracy, further demonstrating its potential for the field of neural morphological visual systems, with high density and low energy loss.

Interspecies-chimera machine vision with polarimetry for real-time navigation and anti-glare pattern recognition

Cutting-edge humanoid machine vision merely mimics human systems and lacks polarimetric functionalities that convey the information of navigation and authentic images. Interspecies-chimera vision reserving multiple hosts' capacities will lead to advanced machine vision. However, implementing the visual functions of multiple species (human and non-human) in one optoelectronic device is still elusive. Here, we develop an optically-controlled polarimetry memtransistor based on a van der Waals heterostructure (ReS2/GeSe2). The device provides polarization sensitivity, nonvolatility, and positive/negative photoconductance simultaneously. The polarimetric measurement can identify celestial polarizations for real-time navigation like a honeybee. Meanwhile, cognitive tasks can be completed like a human by sensing, memory, and synaptic functions. Particularly, the anti-glare recognition with polarimetry saves an order of magnitude energy compared to the traditional humanoid counterpart. This technique promotes the concept of interspecies-chimera visual systems that will leverage the advances of autonomous vehicles, medical diagnoses, intelligent robotics, etc.

Applying the Wake-Up-like Effect to Enhance the Capabilities of Two-Dimensional Ferroelectric Field-Effect Transistors

For traditional ferroelectric field-effect transistors (FeFETs), enhancing the polarization domain of bulk ferroelectric materials is essential to improve device performance. However, there has been limited investigation into the enhancement of polarization field in two-dimensional (2D) ferroelectric material such as CuInP2S6 (CIPS). In this study, similar to bulk ferroelectric materials, CIPS exhibited enhanced polarization field upon application of external cyclic voltage. Moreover, unlike traditional ferroelectric materials, the polarization enhancement of CIPS is not due to redistribution of the defect but rather originates from a mechanism: the long-distance migration of Cu ions. We termed this mechanism the "wake-up-like effect". After incorporating the wake-up-like effect into the graphene/CIPS/WSe2 FeFET device, we successfully increased the hysteresis window and enhanced the current on/off ratio by 4 orders of magnitude. Moreover, the FeFET yielded remarkable achievements, such as multilevel nonvolatile memory with 21 distinct conductance levels, a high on/off ratio exceeding 106, a long retention time exceeding 103 s, and neuromorphic computing with 93% accuracy at recognizing handwritten digits. Introducing the wake-up-like effect to 2D CIPS may pave the way for innovative approaches to achieve advanced multilevel nonvolatile memory and neuromorphic computing capabilities for next-generation micro-nanoelectronic devices.

Graphene Magnetoresistance Control by Photoferroelectric Substrate

Ultralow dimensionality of 2D layers magnifies their sensitivity to adjacent charges enabling even postprocessing electric control of multifunctional structures. However, functionalizing 2D layers remains an important challenge for on-demand device-property exploitation. Here we report that an electrical and even fully optical way to control and write modifications to the magnetoresistive response of CVD-deposited graphene is achievable through the electrostatics of the photoferroelectric substrate. For electrical control, the ferroelectric polarization switch modifies graphene magnetoresistance by 67% due to a Fermi level shift with related modification in charge mobility. A similar function is also attained entirely by bandgap light due to the substrate photovoltaic effect. Moreover, an all-optical way to imprint and recover graphene magnetoresistance by light is reported as well as magnetic control of graphene transconductance. These findings extend photoferroelectric control in 2D structures to magnetic dimensions and advance wireless operation for sensors and field-effect transistors.

Single Wavelength Operating Neuromorphic Device Based on a Graphene-Ferroelectric Transistor

As global data generation continues to rise, there is an increasing demand for revolutionary in-memory computing methodologies and efficient machine learning solutions. Despite recent progress in electrical and electro-optical simulations of machine learning devices, the all-optical nonthermal function remains challenging, with single wavelength operation still elusive. Here we report on an optical and monochromatic way of neuromorphic signal processing for brain-inspired functions, eliminating the need for electrical pulses. Multilevel synaptic potentiation-depression cycles are successfully achieved optically by leveraging photovoltaic charge generation and polarization within the photoferroelectric substrate interfaced with the graphene sensor. Furthermore, the demonstrated low-power prototype device is able to reproduce exact signal profile of brain tissues yet with more than 2 orders of magnitude faster response. The reported properties should trigger all-optical and low power artificial neuromorphic development based on photoferroelectric structures.

Reconfigurable Multifunctional van der Waals Ferroelectric Devices and Logic Circuits

Emerging reconfigurable devices are fast gaining popularity in the search for next-generation computing hardware, while ferroelectric engineering of the doping state in semiconductor materials has the potential to offer alternatives to traditional von-Neumann architecture. In this work, we combine these concepts and demonstrate the suitability of reconfigurable ferroelectric field-effect transistors (Re-FeFETs) for designing nonvolatile reconfigurable logic-in-memory circuits with multifunctional capabilities. Modulation of the energy landscape within a homojunction of a 2D tungsten diselenide (WSe2) layer is achieved by independently controlling two split-gate electrodes made of a ferroelectric 2D copper indium thiophosphate (CuInP2S6) layer. Controlling the state encoded in the program gate enables switching between p, n, and ambipolar FeFET operating modes. The transistors exhibit on-off ratios exceeding 106 and hysteresis windows of up to 10 V width. The homojunction can change from Ohmic-like to diode behavior with a large rectification ratio of 104. When programmed in the diode mode, the large built-in p-n junction electric field enables efficient separation of photogenerated carriers, making the device attractive for energy-harvesting applications. The implementation of the Re-FeFET for reconfigurable logic functions shows how a circuit can be reconfigured to emulate either polymorphic ferroelectric NAND/AND logic-in-memory or electronic XNOR logic with a long retention time exceeding 104 s. We also illustrate how a circuit design made of just two Re-FeFETs exhibits high logic expressivity with reconfigurability at runtime to implement several key nonvolatile 2-input logic functions. Moreover, the Re-FeFET circuit demonstrates high compactness, with an up to 80% reduction in transistor count compared to standard CMOS design. The 2D van de Waals Re-FeFET devices therefore exhibit promising potential for both More-than-Moore and beyond-Moore future of electronics, in particular for an energy-efficient implementation of in-memory computing and machine learning hardware, due to their multifunctionality and design compactness.

Voltage-Mode Ferroelectric Synapse for Neuromorphic Computing

Ferroelectric materials with a modulable polarization extent hold promise for exploring voltage-driven neuromorphic hardware, in which direct current flow can be minimized. Utilizing a single active layer of an insulating ferroelectric polymer, we developed a voltage-mode ferroelectric synapse that can continuously and reversibly update its states. The device states are straightforwardly manifested in the form of variable output voltage, enabling large-scale direct cascading of multiple ferroelectric synapses to build a deep physical neural network. Such a neural network based on potential superposition rather than current flow is analogous to the biological counterpart driven by action potentials in the brain. A high accuracy of over 97% for the simulation of handwritten digit recognition is achieved using the voltage-mode neural network. The controlled ferroelectric polarization, revealed by piezoresponse force microscopy, turns out to be responsible for the synaptic weight updates in the ferroelectric synapses. The present work demonstrates an alternative strategy for the design and construction of emerging artificial neural networks.

Engineering Ferroelectricity and Large Piezoelectricity in h-BN

Hexagonal boron nitride (h-BN) is a well-known layered van der Waals (vdW) material that exhibits no spontaneous electric polarization due to its centrosymmetric structure. Extensive density functional theory (DFT) calculations are used to demonstrate that doping through the substitution of B by isovalent Al and Ga breaks the inversion symmetry and induces local dipole moments along the c-axis, which promotes a ferroelectric (FE) alignment over antiferroelectric. For doping concentrations below 25%, a "protruded layered" structure in which the dopant atoms protrude out of the planar h-BN layers is energetically more stable than the flat layered structure of pristine h-BN or a wurtzite structure similar to w-AlN. The computed polarization, between 7.227 and 21.117 μC/cm2, depending on dopant concentration and the switching barrier (16.684 and 45.838 meV/atom) for the FE polarization reversal are comparable to that of other well-known FEs. Interestingly, doping of h-BN also induces a large negative piezoelectric response in otherwise nonpiezoelectric h-BN. For example, we compute d33 of -24.214 pC/N for Ga0.125B0.875N, which is about 5 times larger than that of pure w-AlN (5 pC/N), although the computed e33 (-1.164 C/m2) is about 1.6 times lower than that of pure w-AlN (1.462 C/m2). Because of the layered structure, the rather small elastic constant C33 provides the origin of the large d33. Moreover, doping makes h-BN an electric auxetic piezoelectric. We also show that ferroelectricity in doped h-BN may persist down to its trilayer, which indicates high potential for applications in FE nonvolatile memories.

2D Ferroic Materials for Nonvolatile Memory Applications

The emerging nonvolatile memory technologies based on ferroic materials are promising for producing high-speed, low-power, and high-density memory in the field of integrated circuits. Long-range ferroic orders observed in 2D materials have triggered extensive research interest in 2D magnets, 2D ferroelectrics, 2D multiferroics, and their device applications. Devices based on 2D ferroic materials and heterostructures with an atomically smooth interface and ultrathin thickness have exhibited impressive properties and significant potential for developing advanced nonvolatile memory. In this context, a systematic review of emergent 2D ferroic materials is conducted here, emphasizing their recent research on nonvolatile memory applications, with a view to proposing brighter prospects for 2D magnetic materials, 2D ferroelectric materials, 2D multiferroic materials, and their relevant devices.