Themed collection Memristors and Neuromorphic Systems

Xiaodong Chen, Cheol Seong Hwang, Yoeri van de Burgt and Francesca Santoro present a themed collection in Materials Horizons and Nanoscale Horizons gathering the latest developments in memristive materials, device fabrication, characterization and circuit design for neuromorphic systems.

Redox- and spin-active macrocyclic vanadyl(IV) complexes and polyoxovanadates(V) have promising structural and physicochemical characteristics for innovative bottom-up hybrid electronics.

This review explores various mechanisms enabling threshold switching in volatile memristors and introduces recent progress in the implementation of neuromorphic computing systems based on these mechanisms.

The proliferation of data has facilitated global accessibility, which demands escalating amounts of power for data storage and processing purposes.

The research found that after doping with rare earth elements, a large number of electrons and holes will be produced on the surface of AlN, which makes the material have the characteristics of spontaneous polarization.

Ferroelectric memory devices such as ferroelectric memristors, ferroelectric tunnel junctions, and field-effect transistors are considered among the most promising candidates for neuromorphic computing devices.

A review and universal benchmark for assessing the artificial neuron and synapse device performance from an energy perspective, with the introduction of neuromorphics from biological to electronics and their prospects.

A self-rectifying ferroelectric tunnel junction that employs a HfO₂/ZrO₂/HfO₂ superlattice (HZH SL) combined with Al₂O₃ and TiO₂ layers is proposed.

Neuromorphic computing based on spin–orbit torque driven domain wall (DW) devices is promising for energy-efficient computation. This study demonstrates energy efficient operations of DW neurons and synapses by novel reading and writing strategies.

A one-selector-one-memristor crossbar array was developed, capable of driving Monte Carlo DropConnect network. This could be achieved through a hardware and algorithm co-design approach, involving mutual improvement of them.

A bionic organic memristor network, inspired by ligand-gated ion channels, is built on the Co–BTA coordination polymer, featuring multiple switching sites with varying activation energies for selective manipulation via in situ redox reactions.

Vacancy-induced filament formation and rupturing across Magnéli phases of VO₂. The resistive switching shows the variations in threshold voltage with discreteness as a function of number of cycles.

A bimodal approach to enhance cognitive accuracies by tuning the magnetoresistance and multistate in a compound spin synapse is developed.

We have fabricated 4-CECAM (with keys encoded into 8-long binary arrays and 8 FTJs) with a content density of 0.75 bits per switch, which highlights 50% increase in content density compared to that of the conventional TCAM (0.5 bits per switch).

Organic neuromorphic platforms have recently received growing interest for the implementation and integration of hybrid systems, acting as a bridge between biological tissue and artificial computing architectures.

Pyridinium-templated one-dimensional (1D) halide perovskites are studied as crossbar memristive materials for artificial neural networks with higher on–off ratios, enhanced endurance, and superior retention characteristics.

We quantified the oxygen tracer diffusion in amorphous hafnium oxide thin films. These tracer diffusion values are consistent with the experimentally measured retention times of hafnium oxide resistive memory devices.

Black phosphorus (BP) field-effect transistors with ultrathin channels exhibit unipolar p-type electrical conduction over a wide range of temperatures and pressures.

This research demonstrates an OTS-based temperature-sensing afferent neuron that features low power consumption and a compact circuit structure.

The recent co-optimization of memristive technologies and programming algorithms enabled neural networks training with in-memory computing systems.

We fabricate a nano-device that laterally confines the switching oxide and filament to 10 nm. Electrical measurements demonstrate lower variability and reduced ionic noise compared to unconfined filaments, which is supported by our 3D simulation.

This work demonstrated a high-performance ferroelectric field-effect transistor (FeFET) via the integration of HZO and 2D MoS₂, showing efficacy in processing diverse real-life tasks within a reservoir computing system.

We propose a self-powered flexible optoelectronic synapse based on PEA₂SnI₄ films for multicolor vision perception and skin sunburn warning.

Fully printed memristors based on silver and water-based 2D material inks were demonstrated on rigid and flexible substrates. The Ag filaments formation depends on the annealing process and can be inhibited by integrating CVD graphene in the device.

Neuromorphic platforms are gaining popularity due to their superior efficiency, low power consumption, and adaptable parallel signal processing capabilities, overcoming the limitations of traditional von Neumann architecture.

In this work, Milano et al. reported on quantum conductance effects in memristive nanowires, unveiling the origin of deviations of conductance levels from integer multiples of the conductance quantum and analyzing conductance fluctuations over time of memristive devices.

The mixture of leaky and self-rectifying memristors in the reservoir increases the accuracy in MNIST-digits recognition task.

A fully stretchable artificial pain perception nociceptor is prepared. The device is stretched in various ways with stable synaptic and nociceptor simulation performance. This work contributes to the development of flexible sensor devices.

A 3D simulation of conductive nanofilaments (CNFs) in multilayer hexagonal-BN memristors is performed.

Respiratory detection and contactless human–computer interaction capacity is achieved by a humidity-sensitive memristor.

Self-formed hierarchical structures of Al and Ag closely mimicking the biological neural network offer wide range synaptic plasticity with ultra-low energy usage. Al islands can be exploited as contact pads to introduce multiple sensory signals.

MLP-Mixer neuromorphic network based on nanocomposite memristive synapses has been developed for efficient and robust classification of images.

We demonstrate the formation of stable monolayers of polyoxometalates on a graphite substrate, which can be electronically multi-level switched and nanostructured without physical contact by the electric field of a scanning tunnelling microscope tip.

A high-dimensional in-sensor reservoir computing system with optoelectronic memristors is demonstrated utilizing optical and electrical masks. Handwritten digit classification and human action recognition are successfully achieved with high accuracy.