Programmable, electroforming-free TiOx/TaOx heterojunction-based non-volatile memory devices

Recent advances in flexible memristors for advanced computing and sensing

Conventional computing systems based on von Neumann architecture face challenges such as high power consumption and limited data processing capability. Improving device performance via scaling guided by Moore's Law becomes increasingly difficult. Emerging memristors can provide a promising solution for achieving high-performance computing systems with low power consumption. In particular, the development of flexible memristors is an important topic for wearable electronics, which can lead to intelligent systems in daily life with high computing capacity and efficiency. Here, recent advances in flexible memristors are reviewed, from operating mechanisms and typical materials to representative applications. Potential directions and challenges for future study in this area are also discussed.

Linear conductance update improvement of CMOS-compatible second-order memristors for fast and energy-efficient training of a neural network using a memristor crossbar array

Memristors are two-terminal memory devices that can change the conductance state and store analog values. Thanks to their simple structure, suitability for high-density integration, and non-volatile characteristics, memristors have been intensively studied as synapses in artificial neural network systems. Memristive synapses in neural networks have theoretically better energy efficiency compared with conventional von Neumann computing processors. However, memristor crossbar array-based neural networks usually suffer from low accuracy because of the non-ideal factors of memristors such as non-linearity and asymmetry, which prevent weights from being programmed to their targeted values. In this article, the improvement in linearity and symmetry of pulse update of a fully CMOS-compatible HfO2-based memristor is discussed, by using a second-order memristor effect with a heating pulse and a voltage divider composed of a series resistor and two diodes. We also demonstrate that the improved device characteristics enable energy-efficient and fast training of a memristor crossbar array-based neural network with high accuracy through a realistic model-based simulation. By improving the memristor device's linearity and symmetry, our results open up the possibility of a trainable memristor crossbar array-based neural network system that possesses great energy efficiency, high area efficiency, and high accuracy at the same time.

Performance Regulation of a ZnO/WOx-Based Memristor and Its Application in an Emotion Circuit

The development of a memristor is very important for artificial intelligence and new electronic circuits. In this work, Ag(Al)/ZnO/WO_x/FTO memristors are fabricated by magnetron sputtering, and the device performance is further improved through annealing and oxygen supply during sputtering. The experimental data show that the FTO/WO_x/ZnO-O₂/Ag memristor has the largest high resistance state (HRS)/low resistance state (LRS) resistance ratio and the best durability. Through data fitting and analysis, the switching mechanism of memristors with different top electrodes is investigated. Furthermore, the physical model of the best performance memristor was established by Simulink, and an emotion-monitoring circuit was constructed on this basis. The circuit can be used to monitor and record the mood changes, and the feedback of the emotion monitoring can be fed back to the user to help them adjust the mood.

Electrochemical Resistive Switching in Nanoporous Hybrid Films by One-Step Molecular Layer Deposition

An organic-inorganic hybrid resistive random-access memory based on a nanoporous zinc-based hydroquinone (Zn-HQ) thin film has been constructed with a Pt/Zn-HQ/Ag sandwich structure. The porous Zn-HQ functional layer was directly fabricated by a one-step molecular layer deposition. These Pt/Zn-HQ/Ag devices show a typical electroforming-free bipolar resistive switching characteristic with lower operation voltages and higher on/off ratio above 10². Our nanoporous hybrid devices can also realize multilevel storage capability and exhibit excellent endurance/retention properties. The connection and disconnection of Ag conductive filaments in nanoporous Zn-HQ thin film follow the electrochemical metallization mechanism. Our computational simulations confirm that the existence of nanopores in Zn-HQ thin films facilitates the Ag filament formation, contributing to the high performance of our hybrid devices.

Induced Complementary Resistive Switching in Forming-Free TiOx/TiO2/TiOx Memristors

The undesirable sneak current path is one of the key challenges in high-density memory integration for the emerging cross-bar memristor arrays. This work demonstrates a new heterojunction design of oxide multilayer stacking with different oxygen vacancy contents to manipulate the oxidation state. We show that the bipolar resistive switching (BRS) behavior of the Pt/TiO_x/Pt cross-bar structure can be changed to complementary resistive switching (CRS) by introducing a thin TiO₂ layer in the middle of the TiO_x layer to obtain a Pt/TiO_x/TiO₂/TiO_x/Pt device architecture with a double-junction active matrix. In contrast to the BRS in a single-layer TiO_x matrix, the device with a double-junction matrix remains in a high-resistance state in the voltage range below the SET voltage, which makes it an efficient structure to overcome the sneak path constraints of undesired half-selected cells that lead to incorrect output reading. This architecture is capable of eliminating these half-selected cells between the nearby cross-bar cells in a smaller programming voltage range. A simplified model for the switching mechanism can be used to account for the observed high-quality switching performance with excellent endurance and current retention properties.

Incorporating a redox active entity to attain electrical bistability in a polymer semiconductor

Owing to the advantages of 3-D printable stack, scalability and low cost solution state production, polymer-based resistive memory devices have been identified as the promising alternative for conventional oxide technology. Resistive memory devices based on the redox switch mechanism is particularly found to yield high precision with respect to the operational voltages. Reversible non-volatile resistive state switching was realized with high device yield (>80%), with a redox-active chemical entity conjugated to the polymeric semiconductor, and the control experiments with the model compound confirmed the imperative role of the redox-active anthraquinone center in the polymeric backbone. Highly uniform nanodomains and the trap free layers excluded the possibilities of other known switching mechanisms. Optical studies and the molecular modelling data assert the presence of strong charge transfer characteristics upon optical excitation due to the insertion of the anthraquinone unit, which was detrimental in exhibiting bistable conductive states in electrical bias as well.