Metal chalcogenides for neuromorphic computing: emerging materials and mechanisms

Evaluation of Nanostructured NiS2 Thin Films from a Single-Source Precursor for Flexible Memristive Devices

Herein, we report the first demonstration of a single-step, in situ growth of NiS2 nanostructures from a single-source precursor onto a flexible substrate as a versatile platform for an effective nonvolatile memristor. The low temperature, solution-processed deposition of NiS2 thin films exhibits a wide band gap range, spherical-flower-like morphology with high surface area and porosity, and negligible surface roughness. Moreover, the fabricated Au/NiS2/ITO/PET memristor device reveals reproducible bipolar resistive switching (RS) at low operational voltages under both flat and bending conditions. The flexible device shows stable RS behavior for multiple cycles with a good memory window (∼102) and data retention of up to 104 s. The switching of a device between a high-resistance state and a low-resistance state is attributed to the filamentary conduction based on sulfur ion migration and sulfur vacancies and plays a key role in the outstanding memristive performance of the device. Consequently, this work provides a simple, scalable, solution-processed route to fabricate a flexible device with potential applications in next-generation neuromorphic computing and wearable electronics.

Negative Differential Interlayer Resistance in WSe2 Multilayers via Conducting Channel Migration with Vertical Double-Side Contacts

The inherent interlayer resistance in two-dimensional (2D) van der Waals (vdW) multilayers is expected to significantly influence the carrier density profile along the thickness, provoking spatial modification and separation of the conducting channel inside the multilayers, in conjunction with the thickness-dependent carrier mobility. However, the effect of the interlayer resistance on the variation in the carrier density profile and its direction along the thickness under different electrostatic bias conditions has been elusive. Here, we reveal the presence of a negative differential interlayer resistance (NDIR) in WSe2 multilayers by considering various contact electrode configurations: (i) bottom contact, (ii) top contact, and (iii) vertical double-side contact (VDC). The contact-structure-dependent shape modification of the transconductance clearly manifests the redistribution of carrier density and indicates the direction of the conducting channel migration along the thickness. Furthermore, the distinct characteristic of the electrically tunable NDIR in 2D WSe2 multilayers is revealed by the observed discrepancy between the top- and bottom-channel resistances determined by four-probe measurements with VDC. Our results provide an optimized device layout and further insights into the distinct carrier transport mechanism in 2D vdW multilayers.

Thermal Management in Neuromorphic Materials, Devices, and Networks

Machine learning has experienced unprecedented growth in recent years, often referred to as an "artificial intelligence revolution." Biological systems inspire the fundamental approach for this new computing paradigm: using neural networks to classify large amounts of data into sorting categories. Current machine-learning schemes implement simulated neurons and synapses on standard computers based on a von Neumann architecture. This approach is inefficient in energy consumption, and thermal management, motivating the search for hardware-based systems that imitate the brain. Here, the present state of thermal management of neuromorphic computing technology and the challenges and opportunities of the energy-efficient implementation of neuromorphic devices are considered. The main features of brain-inspired computing and quantum materials for implementing neuromorphic devices are briefly described, the brain criticality and resistive switching-based neuromorphic devices are discussed, the energy and electrical considerations for spiking-based computation are presented, the fundamental features of the brain's thermal regulation are addressed, the physical mechanisms for thermal management and thermoelectric control of materials and neuromorphic devices are analyzed, and challenges and new avenues for implementing energy-efficient computing are described.

Amorphous Boron Nitride Memristive Device for High-Density Memory and Neuromorphic Computing Applications

Although two-dimensional (2D) nanomaterials are promising candidates for use in memory and synaptic devices owing to their unique physical, chemical, and electrical properties, the process compatibility, synthetic reliability, and cost-effectiveness of 2D materials must be enhanced. In this context, amorphous boron nitride (a-BN) has emerged as a potential material for future 2D nanoelectronics. Therefore, we explored the use of a-BN for multilevel resistive switching (MRS) and synaptic learning applications by fabricating a complementary metal-oxide-semiconductor (CMOS)-compatible Ag/a-BN/Pt memory device. The redox-active Ag and boron vacancies enhance the mixed electrochemical metallization and valence change conduction mechanism. The synthesized a-BN switching layer was characterized using several analyses. The fabricated memory devices exhibited bipolar resistive switching with low set and reset voltages (+0.8 and -2 V, respectively) and a small operating voltage distribution. In addition, the switching voltages of the device were modeled using a time-series analysis, for which the Holt's exponential smoothing technique provided good modeling and prediction results. According to the analytical calculations, the fabricated Ag/a-BN/Pt device was found to be memristive, and its MRS ability was investigated by varying the compliance current. The multilevel states demonstrated a uniform resistance distribution with a high endurance of up to 10⁴ direct current (DC) cycles and memory retention characteristics of over 10⁶ s. Conductive atomic force microscopy was performed to clarify the resistive switching mechanism of the device, and the likely mixed electrochemical metallization and valence change mechanisms involved therein were discussed based on experimental results. The Ag/a-BN/Pt memristive devices mimicked potentiation/depression and spike-timing-dependent plasticity-based Hebbian-learning rules with a high pattern accuracy (90.8%) when implemented in neural network simulations.

ZnO nanowire optoelectronic synapse for neuromorphic computing

Artificial synapses that integrate functions of sensing, memory and computing are highly desired for developing brain-inspired neuromorphic hardware. In this work, an optoelectronic synapse based on the ZnO nanowire (NW) transistor is achieved, which can be used to emulate both the short-term and long-term synaptic plasticity. Synaptic potentiation is present when the device is stimulated by light pulses, arising from the light-induced O₂desorption and the persistent photoconductivity behavior of the ZnO NW. On the other hand, synaptic depression occurs when the device is stimulated by electrical pulses in dark, which is realized by introducing a charge trapping layer in the gate dielectric to trap carriers. Simulation of a neural network utilizing the ZnO NW synapses is carried out, demonstrating a high recognition accuracy over 90% after only 20 training epochs for recognizing the Modified National Institute of Standards and Technology digits. The present nanoscale optoelectronic synapse has great potential in the development of neuromorphic visual systems.