Domain wall magnetic tunnel junction-based artificial synapses and neurons for all-spin neuromorphic hardware.

We report a breakthrough in the hardware implementation of energy-efficient all-spin synapse and neuron devices for highly scalable integrated neuromorphic circuits. Our work demonstrates the successful execution of all-spin synapse and activation function generator using domain wall-magnetic tunnel junctions. By harnessing the synergistic effects of spin-orbit torque and interfacial Dzyaloshinskii-Moriya interaction in selectively etched spin-orbit coupling layers, we achieve a programmable multi-state synaptic device with high reliability. Our first-principles calculations confirm that the reduced atomic distance between 5d and 3d atoms enhances Dzyaloshinskii-Moriya interaction, leading to stable domain wall pinning. Our experimental results, supported by visualizing energy landscapes and theoretical simulations, validate the proposed mechanism. Furthermore, we demonstrate a spin-neuron with a sigmoidal activation function, enabling high operation frequency up to 20 MHz and low energy consumption of 508 fJ/operation. A neuron circuit design with a compact sigmoidal cell area and low power consumption is also presented, along with corroborated experimental implementation. Our findings highlight the great potential of domain wall-magnetic tunnel junctions in the development of all-spin neuromorphic computing hardware, offering exciting possibilities for energy-efficient and scalable neural network architectures.

Asymmetric magnetization switching and programmable complete Boolean logic enabled by long-range intralayer Dzyaloshinskii-Moriya interaction.

After decades of efforts, some fundamental physics for electrical switching of magnetization is still missing. Here, we report the discovery of the long-range intralayer Dzyaloshinskii-Moriya interaction (DMI) effect, which is the chiral coupling of orthogonal magnetic domains within the same magnetic layer via the mediation of an adjacent heavy metal layer. The effective magnetic field of the long-range intralayer DMI on the perpendicular magnetization is out-of-plane and varies with the interfacial DMI constant, the applied in-plane magnetic fields, and the magnetic anisotropy distribution. Striking consequences of the effect include asymmetric current/field switching of perpendicular magnetization, hysteresis loop shift of perpendicular magnetization in the absence of in-plane direct current, and sharp in-plane magnetic field switching of perpendicular magnetization. Utilizing the intralayer DMI, we demonstrate programable, complete Boolean logic operations within a single spin-orbit torque device. These results will stimulate investigation of the long-range intralayer DMI effect in a variety of spintronic devices.

Anti-friction gold-based stretchable electronics enabled by interfacial diffusion-induced cohesion.

Stretchable electronics that prevalently adopt chemically inert metals as sensing layers and interconnect wires have enabled high-fidelity signal acquisition for on-skin applications. However, the weak interfacial interaction between inert metals and elastomers limit the tolerance of the device to external friction interferences. Here, we report an interfacial diffusion-induced cohesion strategy that utilizes hydrophilic polyurethane to wet gold (Au) grains and render them wrapped by strong hydrogen bonding, resulting in a high interfacial binding strength of 1017.6 N/m. By further constructing a nanoscale rough configuration of the polyurethane (RPU), the binding strength of Au-RPU device increases to 1243.4 N/m, which is 100 and 4 times higher than that of conventional polydimethylsiloxane and styrene-ethylene-butylene-styrene-based devices, respectively. The stretchable Au-RPU device can remain good electrical conductivity after 1022 frictions at 130 kPa pressure, and reliably record high-fidelity electrophysiological signals. Furthermore, an anti-friction pressure sensor array is constructed based on Au-RPU interconnect wires, demonstrating a superior mechanical durability for concentrated large pressure acquisition. This chemical modification-free approach of interfacial strengthening for chemically inert metal-based stretchable electronics is promising for three-dimensional integration and on-chip interconnection.

Multi-Level Pyramidal Microstructure-Based Pressure Sensors with High Sensitivity and Wide Linear Range for Healthcare Monitoring.

Flexible pressure sensors have garnered significant attention in the field of wearable healthcare due to their scalability and shape variability. However, a crucial challenge in their practical application for various healthcare scenarios is striking a balance between the sensitivity and sensing range. This limitation arises from the reduced compressibility of the microstructures on the surface of pressure-sensitive materials under high pressure, resulting in progressive saturation of the sensor's response and leading to a restricted and nonlinear pressure sensing range. In this study, we present a novel approach utilizing multi-level pyramidal microstructures in flexible pressure sensors to achieve both high sensitivity (8775 kPa-1) and linear response (R2 = 0.997) over a wide pressure range (up to 1000 kPa). The effectiveness of the proposed design stems from the compensatory behavior of the lower pyramidal microstructures, which counteracts the declining sensitivity associated with the gradual hardening of the higher pyramidal microstructures. Furthermore, the sensor demonstrates a fast response time of 11.6 ms and a fast relaxation time of 3.8 ms and can reliably detect pressures as low as 30.2 Pa. Our findings highlight the applicability of this flexible pressure sensor in diverse human body health detection tasks, ranging from weak pulses to finger flexion and plantar pressure distribution. Notably, the proposed sensor design eliminates the need for replacing flexible pressure sensors with varying ranges, thereby enhancing their practical utility.

MEMS Fluxgate Sensor Based on Liquid Casting.

Compared with electroplating, liquid casting enables the rapid formation of a three-dimensional solenoid coil with a narrower line width and greater thickness, which proves advantageous in enhancing the comprehensive performance of the micro-electromechanical system (MEMS) fluxgate sensor. For this reason, a MEMS fluxgate sensor based on liquid casting with a closed-loop Fe-based amorphous alloy core is proposed. Based on the process parameters of liquid casting, the structure of the MEMS fluxgate sensor was designed. Utilizing MagNet to build the simulation model, the optimal excitation conditions and sensitivity were obtained. According to the simulation model, a highly sensitive MEMS fluxgate sensor based on liquid casting was fabricated. The resulting sensor exhibits a sensitivity of 2847 V/T, a noise of 306 pT/√Hz@1 Hz, a bandwidth of DC-10.5 kHz, and a power consumption of 43.9 mW, which shows high sensitivity and low power consumption compared with other MEMS fluxgates in similar size.

Schottky Barrier Control of Self-Polarization for a Colossal Ferroelectric Resistive Switching.

Controlling the domain evolution is critical both for optimizing ferroelectric properties and for designing functional electronic devices. Here we report an approach of using the Schottky barrier formed at the metal/ferroelectric interface to tailor the self-polarization states of a model ferroelectric thin film heterostructure system SrRuO3/(Bi,Sm)FeO3. Upon complementary investigations of the piezoresponse force microscopy, electric transport measurements, X-ray photoelectron/absorption spectra, and theoretical studies, we demonstrate that Sm doping changes the concentration and spatial distribution of oxygen vacancies with the tunable host Fermi level which modulates the SrRuO3/(Bi,Sm)FeO3 Schottky barrier and the depolarization field, leading to the evolution of the system from a single domain of downward polarization to polydomain states. Accompanied by such modulation on self-polarization, we further tailor the symmetry of the resistive switching behaviors and achieve a colossal on/off ratio of ∼1.1 × 106 in the corresponding SrRuO3/BiFeO3/Pt ferroelectric diodes (FDs). In addition, the present FD also exhibits a fast operation speed of ∼30 ns with a potential for sub-nanosecond and an ultralow writing current density of ∼132 A/cm2. Our studies provide a way for engineering self-polarization and reveal its strong link to the device performance, facilitating FDs as a competitive memristor candidate used for neuromorphic computing.

Magnetically tunable diffractive optical elements based on ion-irradiated ultrathin ferromagnetic stacks.

We report a novel type of magnetically tunable diffractive optical element (DOE) based on ultrathin ferromagnetic (FM) Pt/Co stacks. The Pt/Co stacks are irradiated by Ar+ ions at selected areas so that the perpendicular anisotropy is spatially modulated and the DOEs can be tuned by an external magnetic field through the magnetooptical effect. Based on this concept, a diffraction grating and a Fresnel zone plate (FZP) were developed, and complementary experimental results corroborate that a magnetic field can simultaneously manipulate both the zeroth and the first diffraction orders of these DOEs. Importantly, this effect can be utilized to enhance or hide the image formed by the FZP. Our studies pave the way toward developing compact and high-precision DOEs with fast and robust tunability, facilitating various applications spanning a wide spectrum range.

Spintronic leaky-integrate-fire spiking neurons with self-reset and winner-takes-all for neuromorphic computing.

Neuromorphic computing using nonvolatile memories is expected to tackle the memory wall and energy efficiency bottleneck in the von Neumann system and to mitigate the stagnation of Moore's law. However, an ideal artificial neuron possessing bio-inspired behaviors as exemplified by the requisite leaky-integrate-fire and self-reset (LIFT) functionalities within a single device is still lacking. Here, we report a new type of spiking neuron with LIFT characteristics by manipulating the magnetic domain wall motion in a synthetic antiferromagnetic (SAF) heterostructure. We validate the mechanism of Joule heating modulated competition between the Ruderman-Kittel-Kasuya-Yosida interaction and the built-in field in the SAF device, enabling it with a firing rate up to 17 MHz and energy consumption of 486 fJ/spike. A spiking neuron circuit is implemented with a latency of 170 ps and power consumption of 90.99 µW. Moreover, the winner-takes-all is executed with a current ratio >104 between activated and inhibited neurons. We further establish a two-layer spiking neural network based on the developed spintronic LIFT neurons. The architecture achieves 88.5% accuracy on the handwritten digit database benchmark. Our studies corroborate the circuit compatibility of the spintronic neurons and their great potential in the field of intelligent devices and neuromorphic computing.

Bioinspired Dual-Mode Stretchable Strain Sensor Based on Magnetic Nanocomposites for Strain/Magnetic Discrimination.

Recently, flexible stretchable sensors have been gaining attention for their excellent adaptability for electronic skin applications. However, the preparation of stretchable strain sensors that achieve dual-mode sensing while still retaining ultra-low detection limit of strain, high sensitivity, and low cost is a pressing task. Herein, a high-performance dual-mode stretchable strain sensor (DMSSS) based on biomimetic scorpion foot slit microstructures and multi-walled carbon nanotubes (MWCNTs)/graphene (GR)/silicone rubber (SR)/Fe3 O4 nanocomposites is proposed, which can accurately sense strain and magnetic stimuli. The DMSSS exhibits a large strain detection range (≈160%), sensitivity up to 100.56 (130-160%), an ultra-low detection limit of strain (0.16% strain), and superior durability (9000 cycles of stretch/release). The sensor can accurately recognize sign language movement, as well as realize object proximity information perception and whole process information monitoring. Furthermore, human joint movements and micro-expressions can be monitored in real-time. Therefore, the DMSSS of this work opens up promising prospects for applications in sign language pose recognition, non-contact sensing, human-computer interaction, and electronic skin.

An SOI-Structured Piezoresistive Differential Pressure Sensor with High Performance.

This paper presents a piezoresistive differential pressure sensor based on a silicon-on-insulator (SOI) structure for low pressure detection from 0 to 30 kPa. In the design phase, the stress distribution on the sensing membrane surface is simulated, and the doping concentration and geometry of the piezoresistor are evaluated. By optimizing the process, the realization of the pressure sensing diaphragm with a controllable thickness is achieved, and good ohmic contact is ensured. To obtain higher sensitivity and high temperature stability, an SOI structure with a 1.5 µm ultra-thin monocrystalline silicon layer is used in device manufacturing. The device diaphragm size is 700 µm × 700 µm × 2.1 µm. The experimental results show that the fabricated piezoresistive pressure sensor has a high sensitivity of 2.255 mV/V/kPa and a sensing resolution of less than 100 Pa at room temperature. The sensor has a temperature coefficient of sensitivity (TCS) of -0.221 %FS/°C and a temperature coefficient of offset (TCO) of -0.209 %FS/°C at operating temperatures ranging from 20 °C to 160 °C. The reported piezoresistive microelectromechanical systems (MEMS) pressure sensors are fabricated on 8-inch wafers using standard CMOS-compatible processes, which provides a volume solution for embedded integrated precision detection applications of air pressure, offering better insights for high-temperature and miniaturized low-pressure sensor research.

Tailored Plasmonic Ru/OV-MoO2 on TiO2 Catalysts via Solid-Phase Interface Engineering: Toward Highly Efficient Photoassisted Li-O2 Batteries with Enhanced Cycling Reliability.

The photoassisted electrochemical reactions are considered an effective method to reduce the overpotential of Li-O2 batteries. However, achieving long-term cell cycling stability remains a challenge. Here, we report a solid-phase interfacial reaction (SPIR) strategy that introduces both oxygen vacancies (OV) and metal centers (Ru) into the MoO2 to synthesize the surface plasmon (i.e., Ru/OV-MoO2). Then, Ru/OV-MoO2 can be uniformly loaded on the TiO2 nanowires by the hydrothermal method. The plasma effect of Ru/OV-MoO2 demonstrates the effective reduction of the photoexcited electron and hole recombination to improve visible light-harvesting ability. The lifetime of electrons and holes can be extended by Ru nanoparticles, which is beneficial for promoting the formation and decomposition of Li2O2. In addition, the generated OV further enhanced the migration of electrons and Li+, thus improving the ORR performance. The Ru/OV-MT/CC cathode corroborates excellent stability and catalytic performance in the photoassisted Li-O2 battery, with an overpotential value of 0.47 V, achieving the highest energy efficiency of 93.94%, retaining at 89.13% after 800 h. This work offers a platform for preparing a stable, bifunctional catalyst with the high total activity of a photoassisted Li-O2 battery.

Visible-near-Infrared Light-Driven Photocatalytic Characteristics of Er3+/Yb3+-Codoped BiOBr Upconverting Microparticles for Tetracycline Degradation.

To settle the unsatisfying efficiency and insufficient light harvesting ability of photocatalysts, we report on the development of Er3+/Yb3+-codoped BiOBr (BiOBr:Er3+/xYb3+) microparticles that were synthesized by a rational high-temperature solid-state reaction method. The prepared microcrystals exhibit high visible upconversion (UC) emissions with maximum intensities at x = 0.01 when excited by a 980 nm laser. Remarkably, the corresponding UC emission process is attributed to a two-photon absorption route. Furthermore, the photocatalytic activities of as-synthesized compounds were further evaluated through analyzing the visible-near-infrared light-triggered tetracycline degradation. Compared with BiOBr:Er3+ microparticles, BiOBr:Er3+/xYb3+ microparticles present superior photocatalytic properties and the optimal status is achieved when x = 0.05, in which h+, ·O2-, and ·OH active species contribute to the photocatalytic mechanism. Additionally, the designed microparticles exhibit better photocatalytic abilities than previously reported photocatalysts (i.e., TiO2, SnO2) upon full-spectrum light irradiation. These results reveal that Yb3+ codoping is able to not only enhance the UC emission properties of BiOBr:Er3+ microparticles but also reinforce their photocatalytic activities. Our findings may put forward a facile strategy to regulate the photodegradation capacity of photcatalysts.

Synergy of Spin-Orbit Torque and Built-In Field in Magnetic Tunnel Junctions with Tilted Magnetic Anisotropy: Toward Tunable and Reliable Spintronic Neurons.

Owing to programmable nonlinear dynamics, magnetic domain wall (DW)-based devices can be configured to function as spintronic neurons, promising to execute sophisticated tasks as a human brain. Developing energy-efficient, CMOS compatible, reliable, and tunable spintronic neurons to emulate brain-inspired processes has been a key research goal for decades. Here, a new type of DW device is reported with biological neuron characteristics driven by the synergistic interaction between spin-orbit torque and built-in field (Hbuilt-in ) in magnetic tunnel junctions, enabling time- and energy-efficient leaky-integrate-and-fire and self-reset neuromorphic implementations. A tilted magnetic anisotropic free layer is proposed and further executed to mitigate the DW retrograde motion by suppressing the Walker breakdown. Complementary experiments and micromagnetic co-simulation results show that the integrating/leaking time of the developed spintronic neuron can be tuned to 12/15 ns with an integrating power consumption of 65 µW, which is 36× and 1.84× time and energy efficient than the state-of-the-art alternatives, respectively. Moreover, the spatial distribution of Hbuilt-in can be modulated by adjusting the width and compensation of the reference layer, facilitating tunable activation function generator exploration. Such architecture demonstrates great potential in both fundamental research and new trajectories of technology advancement for spintronic neuron hardware applications.

Biologically Emulated Flexible Sensors With High Sensitivity and Low Hysteresis: Toward Electronic Skin to a Sense of Touch.

Recently, flexible pressure sensors (FPSs) have attracted intensive attention owing to their ability to mimic and function as electronic skin. Some sensors are exploited with a biological structure dielectric layer for high sensitivity and detection. However, traditional sensors with bionic structures usually suffer from a limited range for high-pressure scenes due to their high sensitivity and high hysteresis in the medium pressure range. Here, a reconfigurable flea bionic structure FPS based on 3D printing technology, which can meet the needs of different scenes via tailoring of the dedicated structural parameters, is proposed. FPS exhibits high sensitivity (1.005 kPa-1 in 0-1 kPa), wide detection range (200 kPa), high repeatability (6000 cycles in 10 kPa), low hysteresis (1.3%), fast response time (40 ms), and very low detection limit (0.5 Pa). Aiming at practical application implementation, FPS has been correspondingly placed on a finger, elbow, arm, neck, cheek, and manipulators to detect the actions of various body parts, suggestive of excellent applicability. It is also integrated to make a flexible 3 × 3 sensor array for detecting spatial pressure distribution. The results indicate that FPS exhibits a significant application potential in advanced biological wearable technologies, such as human motion monitoring.

Reflow Soldering Capability Improvement by Utilizing TaN Interfacial Layer in 1Mbit RRAM Chip.

We investigated the thermal stability of a 1Mbit OxRRAM array embedded in 28 nm COMS technology. A back-end-of-line (BEOL) solution with a TaN-Ta interfacial layer was proposed to eliminate the failure rate after reflow soldering assembly at 260 °C. By utilizing a TaN-Ta interfacial layer (IL), the oxygen defects in conductive filament were redistributed, and electromigration lifetimes of Cu-based damascene interconnects were improved, which contributed to optimization. This work provides a potential solution for the practical application of embedded RRAM beyond the 28 nm technology node.

Long-Term Accuracy Enhancement of Binary Neural Networks Based on Optimized Three-Dimensional Memristor Array.

In embedded neuromorphic Internet of Things (IoT) systems, it is critical to improve the efficiency of neural network (NN) edge devices in inferring a pretrained NN. Meanwhile, in the paradigm of edge computing, device integration, data retention characteristics and power consumption are particularly important. In this paper, the self-selected device (SSD), which is the base cell for building the densest three-dimensional (3D) architecture, is used to store non-volatile weights in binary neural networks (BNN) for embedded NN applications. Considering that the prevailing issues in written data retention on the device can affect the energy efficiency of the system's operation, the data loss mechanism of the self-selected cell is elucidated. On this basis, we introduce an optimized method to retain oxygen ions and prevent their diffusion toward the switching layer by introducing a titanium interfacial layer. By using this optimization, the recombination probability of Vo and oxygen ions is reduced, effectively improving the retention characteristics of the device. The optimization effect is verified using a simulation after mapping the BNN weights to the 3D VRRAM array constructed by the SSD before and after optimization. The simulation results showed that the long-term recognition accuracy (greater than 105 s) of the pre-trained BNN was improved by 24% and that the energy consumption of the system during training can be reduced 25,000-fold while ensuring the same accuracy. This work provides high storage density and a non-volatile solution to meet the low power consumption and miniaturization requirements of embedded neuromorphic applications.

All-Electrical Control of Compact SOT-MRAM: Toward Highly Efficient and Reliable Non-Volatile In-Memory Computing.

Two-dimensional van der Waals (2D vdW) ferromagnets possess outstanding scalability, controllable ferromagnetism, and out-of-plane anisotropy, enabling the compact spintronics-based non-volatile in-memory computing (nv-IMC) that promises to tackle the memory wall bottleneck issue. Here, by employing the intriguing room-temperature ferromagnetic characteristics of emerging 2D Fe3GeTe2 with the dissimilar electronic structure of the two spin-conducting channels, we report on a new type of non-volatile spin-orbit torque (SOT) magnetic tunnel junction (MTJ) device based on Fe3GeTe2/MgO/Fe3GeTe2 heterostructure, which demonstrates the uni-polar and high-speed field-free magnetization switching by adjusting the ratio of field-like torque to damping-like torque coefficient in the free layer. Compared to the conventional 2T1M structure, the developed 3-transistor-2-MTJ (3T2M) cell is implemented with the complementary data storage feature and the enhanced sensing margin of 201.4% (from 271.7 mV to 547.2 mV) and 276% (from 188.2 mV to 520 mV) for reading "1" and "0", respectively. Moreover, superior to the traditional CoFeB-based MTJ memory cell counterpart, the 3T2M crossbar array architecture can be executed for AND/NAND, OR/NOR Boolean logic operation with a fast latency of 24 ps and ultra-low power consumption of 2.47 fJ/bit. Such device to architecture design with elaborated micro-magnetic and circuit-level simulation results shows great potential for realizing high-performance 2D material-based compact SOT magnetic random-access memory, facilitating new applications of highly reliable and energy-efficient nv-IMC.

A Novel RFID Authentication Protocol Based on Reconfigurable RRAM PUF.

Radio frequency identification technology (RFID) has empowered a wide variety of automation industries. Aiming at the current light-weight RFID encryption scheme with limited information protection methods, combined with the physical unclonable function (PUF) composed of resistive random access memory (RRAM), a new type of high-efficiency reconfigurable strong PUF circuit structure is proposed in this paper. Experimental results show that the proposed PUF shows an almost ideal value (50%) of inter-chip hamming distance (HD) (µ/σ = 0.5001/0.0340) among 1000 PUF keys, and intra-chip HD results are very close to the ideal value (0). The bit error rate (BER) is as low as 3.8×10-6 across one million challenges. Based on the RRAM PUF, we propose and implement a light weight RFID authentication protocol. By virtue of RRAM's model ability, the protocol replaces the One-way Hash Function with a response chain mutual encryption algorithm. The results of test and analysis show that the protocol can effectively resist multiple threats such as physical attacks, replay attacks, tracking attacks and asynchronous attacks, and has good stability. At the same time, based on RRAM's unique resistance variability, PUF also has the advantage of being reconfigurable, providing good security for RFID tags.

Reducing Dzyaloshinskii-Moriya interaction and field-free spin-orbit torque switching in synthetic antiferromagnets.

Perpendicularly magnetized synthetic antiferromagnets (SAF), possessing low net magnetization and high thermal stability as well as easy reading and writing characteristics, have been intensively explored to replace the ferromagnetic free layers of magnetic tunnel junctions as the kernel of spintronic devices. So far, utilizing spin-orbit torque (SOT) to realize deterministic switching of perpendicular SAF have been reported while a large external magnetic field is typically needed to break the symmetry, making it impractical for applications. Here, combining theoretic analysis and experimental results, we report that the effective modulation of Dzyaloshinskii-Moriya interaction by the interfacial crystallinity between ferromagnets and adjacent heavy metals plays an important role in domain wall configurations. By adjusting the domain wall configuration between Bloch type and Néel type, we successfully demonstrate the field-free SOT-induced magnetization switching in [Co/Pd]/Ru/[Co/Pd] SAF devices constructed with a simple wedged structure. Our work provides a practical route for utilization of perpendicularly SAF in SOT devices and paves the way for magnetic memory devices with high density, low stray field, and low power consumption.

Hybrid memristor-CMOS neurons for in-situ learning in fully hardware memristive spiking neural networks.

Spiking neural network, inspired by the human brain, consisting of spiking neurons and plastic synapses, is a promising solution for highly efficient data processing in neuromorphic computing. Recently, memristor-based neurons and synapses are becoming intriguing candidates to build spiking neural networks in hardware, owing to the close resemblance between their device dynamics and the biological counterparts. However, the functionalities of memristor-based neurons are currently very limited, and a hardware demonstration of fully memristor-based spiking neural networks supporting in-situ learning is very challenging. Here, a hybrid spiking neuron combining a memristor with simple digital circuits is designed and implemented in hardware to enhance neuron functions. The hybrid neuron with memristive dynamics not only realizes the basic leaky integrate-and-fire neuron function but also enables the in-situ tuning of the connected synaptic weights. Finally, a fully hardware spiking neural network with the hybrid neurons and memristive synapses is experimentally demonstrated for the first time, and in-situ Hebbian learning is achieved with this network. This work opens up a way towards the implementation of spiking neurons, supporting in-situ learning for future neuromorphic computing systems.