Memristor Research Articles

Effect of RF-sputtering temperature of ITO electrodes on the resistive switching behaviour of ITO/RbPbI3/Cu devices

In this work, we deposit Indium doped Tin Oxide (ITO) on a glass substrate using the Radio-Frequency Sputtering technique at different temperatures and studied its structural and optical properties. Various structural, optical, and morphological studies are conducted on the ITO substrates. In addition, we have used these ITO films as conducting electrodes in MIM (metal-insulator-metal) type structures for resistive memory application. A resistive switching memory device based on RbPbI3 with sputtered ITO films as the bottom electrode is fabricated and studied for its electrical performance. It has been found that the device showed the write-once-read-many (WORM) nature at a sputtering temperature of 300 °C with a good OFF/ON ratio. The devices with ITO film sputtered a temperature below 300 °C do not show any switching behavior due to lower conductivity and improper surface morphology.

Spike Optimization to Improve Properties of Ferroelectric Tunnel Junction Synaptic Devices for Neuromorphic Computing System Applications.

The continuous advancement of Artificial Intelligence (AI) technology depends on the efficient processing of unstructured data, encompassing text, speech, and video. Traditional serial computing systems based on the von Neumann architecture, employed in information and communication technology development for decades, are not suitable for the concurrent processing of massive unstructured data tasks with relatively low-level operations. As a result, there arises a pressing need to develop novel parallel computing systems. Recently, there has been a burgeoning interest among developers in emulating the intricate operations of the human brain, which efficiently processes vast datasets with remarkable energy efficiency. This has led to the proposal of neuromorphic computing systems. Of these, Spiking Neural Networks (SNNs), designed to closely resemble the information processing mechanisms of biological neural networks, are subjects of intense research activity. Nevertheless, a comprehensive investigation into the relationship between spike shapes and Spike-Timing-Dependent Plasticity (STDP) to ensure efficient synaptic behavior remains insufficiently explored. In this study, we systematically explore various input spike types to optimize the resistive memory characteristics of Hafnium-based Ferroelectric Tunnel Junction (FTJ) devices. Among the various spike shapes investigated, the square-triangle (RT) spike exhibited good linearity and symmetry, and a wide range of weight values could be realized depending on the offset of the RT spike. These results indicate that the spike shape serves as a crucial indicator in the alteration of synaptic connections, representing the strength of the signals.

Synthesis of LaXO3 (X = Fe, Mn, Cr, Ni) Thin Films Using a Simple Spin Coating Set-Up for Resistive Switching Memory Devices

This study focuses on the preparation of LaXO3(X=Fe,Mn,Cr,Ni) perovskite thin films using a simple set-up spin coating technique and the evaluation of their properties for application in switching memory devices. The properties of as-deposited films were thoroughly characterized using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), UV-Vis spectroscopy, and a vibrating sample magnetometer (VSM). The results obtained revealed that the as-deposited films have a polycrystalline cubic structure. The film surfaces were uniform and densely packed without any voids, cracks, or pinholes. In addition, irregularly shaped grains were observed having an average size of 140, 120, 89, and 70 nm for LaFeO3, LaMnO3, LaCrO3, and LaNiO3 films, respectively. VSM analysis demonstrated that LaFeO3 film exhibited superior magnetic properties compared to the other films. Furthermore, memory devices with Au/LaXO3(X=Fe,Mn,Cr,Ni)/FTO structures were fabricated, and their I-V characteristics were measured. In order to assess their performance, an endurance test was conducted. The findings indicated that Au/LaFeO3/FTO device exhibited higher RHRSRLRS ratio (~7×105), low Set/Reset voltages, lower power consumption (2.7×10−4 W), and stable endurance with no significant degradation was observed in the LRS and HRS after 20 sweep cycles. These favorable parameters can be attributed to the reduced thickness, larger grain size, and excellent magnetic properties of the active-layer LaFeO3. Moreover, the conduction mechanism of the fabricated devices was investigated, revealing that the conduction in the LRS is primarily dominated by Ohmic behavior, while the HRS exhibited different conduction mechanisms.

A 28-nm RRAM Computing-in-Memory Macro Using Weighted Hybrid 2T1R Cell Array and Reference Subtracting Sense Amplifier for AI Edge Inference

Non-volatile computing-in-memory (nvCIM) can potentially meet the ever-increasing demands on improving the energy efficiency (EF) for intelligent edge devices. However, it still suffers from limited input parallelism due to the parasitic effects, signal margin degradation due to device non-idealities, and large hardware cost for analog readout. In this work, we present a two-transistor-one-resistor (2T1R) resistive memory (RRAM) nvCIM macro featuring: 1) a macro structure with decoupled memory and computing data paths; 2) the weighted hybrid 2T1R (WH-2T1R) cell array; 3) the redundant sub-array mapping scheme of the most-significant-bit (RSM-MSB); and 4) reference-subtracting current sense amplifier (RS-CSA). A test-chip is silicon-verified using the 28-nm high-k/metal-gate (HKMG) logic process with foundry-developed RRAM. The test-chip performs linear analog multiply-and-accumulate (MAC) operations over 32 accumulation channels and achieves 30.34–154.04 TOPS/W with 1-bit input (IN), 3-bit weight (W), and 4-bit output (O). Evaluations with the ResNet-18 model show that the MSB-RSM scheme results in 0.96% and 2.83% improvement on CIFAR-10 and CIFAR-100 inference accuracy, respectively.

Environment-friendly resistive memory based on natural casein: Role of electrode and bio-material concentration

It is well-known that moisture plays a crucial role on the stability of bio-material-based resistive memory devices. In this work, we report a bio-material-based resistive memory device that does not depend on moisture to attain a stable performance. Sodium caseinate (NaCas) obtained from bovine milk, is used as the active material with silver (Ag) as the top electrode (TE) and indium tin oxide (ITO) as the bottom electrode (BE) for the implementation of the device. The device exhibits Write-Once-Read-Many times (WORM) type current-voltage characteristics with a sufficiently large data retention of >106 s, a cyclic endurance of >6 × 103, and a switching voltage of ∼1.06 V. The use of the less reactive Ag for the TE facilitates the removal of the capacitive nature of the device characteristics. It is found that increasing the NaCas concentration improves the memory performance. The role of the Ag electrode in removing the capacitive nature and possible switching mechanism of the device is suggested.

Resistive switching properties of CdTe/CdSe core–shell quantum dots incorporated organic cow milk for memory application

Our study focuses on the resistive switching memory characteristics of devices containing active layers of CdTe/CdSe core–shell quantum dots (QDs) dispersed in organic cow milk. We fabricated devices containing CdTe/CdSe particles per volume of milk using a direct-dipping method, with particle concentrations of 2.4 × 10[Formula: see text] (S1), 4.8 × 10[Formula: see text](S2), and 7.2 × 10[Formula: see text](S3). This method was cost-free. Distinct memory characteristics were observed among devices featuring these concentrations. S1- and S2-based devices exhibited memory behavior with ‘S-type’ and ‘O-type’ hysteresis, respectively. The device based on S3 exhibited an initial asymmetric ‘N-type’ behavior with a large ON/OFF ratio ([Formula: see text]104). The memory attribute of the aforementioned device disappeared after the initial three cycles but was subsequently restored by modifying the scan voltage step from 10 mV to 1 mV. The observed results indicate typical symmetric ‘N-type’ behavior of the device, accompanied by threshold switching under positive voltage bias. Additionally, the switching was observed to be as low as 0.04 V. The S1- and S2-based devices were found to exhibit hopping conduction and Schottky emission in the OFF- and ON-state, respectively, while the S3-based device showed conductive bridge resistive switching as the conduction mechanism. The findings indicate that it is possible to produce biodegradable and disposable memory devices using full cream cow milk and CdTe/CdSe core–shell QDs. The device’s switching and memory functions can be manipulated by regulating the quantity of CdTe/CdSe particles present in the milk. Finally, we have demonstrated that the switching behavior of ReRAMs based on milk can be influenced by the voltage steps used during scanning.

Coexistence of Interfacial and Filamentary Resistance Switching in Ti/SiO x /Au Resistive Memory Devices

Coexistence of Interfacial and Filamentary Resistance Switching in Ti/SiO_<i>x</i>/Au Resistive Memory Devices

Effect of Annealing on Resistive Switching Properties of Glancing Angle Deposition‐Assisted WO3 Thin Films

Herein, the impact of postdeposition annealing on resistive switching behavior of radio frequency magnetron sputtered WO3 thin films is reported. Films are deposited under glancing angle deposition (GLAD) configuration of sputtering at varying GLAD angle from 65° to 80°. Structure transition from monoclinic to orthorhombic phase in deposited WO3 films is perceived after ex situ annealing at temperature of 400 °C. Resistive switching properties show shift from bipolar to unipolar switching on postdeposition annealing. WO3 films show unipolar switching behavior after ex situ annealing for all prepared samples. The value of resistance in high resistance state is lowered after ex situ heating treatment and interestingly switching voltage also reduces to 3 V from 7 V after annealing treatment. The ratio of high to low resistance state for annealed WO3 film fabricated at 70° GLAD angle is achieved to be maximum (≈219). A detailed charge transport mechanism shows that ohmic behavior is dominant current conduction mechanism at lower applied voltage, while space charge limited current and Child's law are dominant at higher applied voltages. Obtained results encourage utilization of prepared WO3 thin films toward a wide variety of applications in optoelectronics, microelectronics, and environmental engineering along with advanced electronics such as resistive memory devices.

Activation processes during operation of an Ag/SnSe/Ge2Se3/W ion memristor with a self-forming current-conducting channel

In an Ag/SnSe/Ge2Se3/W ionic type memristor, the activation energy of two main processes responsible for its operation has been determined, namely: the activation energy for the formation of a conductive channel and the activation energy for memristor degradation. By measuring the current-voltage characteristics, the electrical conductivity of the memristor in low- and high-resistance operating modes was assessed. To determine the activation energy, the Arrhenius law and the provisions of the thermodynamics of irreversible processes were used, in particular the second postulate of Onsager, according to which the growth rate of the irreversible part of the entropy of a system tending to equilibrium is proportional to the sum of the products of the flows occurring in the system and the generalized thermodynamic force corresponding to each flow. The equilibrium state of the memristor was taken to be the state in which the memristor lost the ability to function as a resistive memory cell. The flow of Ag+ ions – electromigration was used as a substance flow. For the first process, the activation energy was 0.24 eV, and for the second, 1.16 eV. The different values of activation energy reflect the difference between the agglomeration mechanism of formation of a current-conducting channel, typical of an Ag/SnSe/Ge2Se3/W memristor, and the “standard” mechanism of substance transfer based on a group of point defects, which accompanies the process of memristor degradation.

Nitrogen and aluminum-nitrogen doped graphene for non-volatile resistive memory applications

Doping is a key strategy used to tune the electronic properties of graphene. Calculations have shown that graphene doped with Al has potential applications in the sensing of toxic gases. However, the realization of Al-doped graphene in experiments has proven to be a challenge. In this study, we successfully doped graphene with both N and a combination of Al and N using a microwave process suitable for large-scale production. Several characterization methods such as Raman spectroscopy, x-ray photoelectron spectroscopy (XPS), and energy dispersive x-ray spectroscopy (EDX) were used to confirm the composition, and therefore the successful doping, of the graphene. Doping and co-doping are found to increase the conductivity of graphene. It is also demonstrated that the doped graphene shows well-defined resistive memory properties when included in a sandwich structure consisting of polystyrene/doped graphene/polystyrene. The on-and-off voltage is found to be 2 V, making the Al-N co-doped graphene suitable for applications in resistive memory devices.

Optimized chalcogenide medium for inherently activated resistive switching device

The suitability of thin films of the chalcogenide germanium telluride (GeTex) with various chemical compositions for use as an active medium in a conductive bridge resistive switching memory device with Cu and TiN as the active and counter electrodes, respectively, was examined. Experimental results showed that all of the tested resistive switching devices showed identical current-voltage hysteresis curves for the first and second cycles, the electroforming (EF)-free characteristic, which is favourable for achieving high operational reliability. From a comparative study of different active electrodes, we concluded that the intrinsic electronic traps and diffused Cu ions in the GeTex layer were responsible for the EF-free characteristic. Furthermore, as the Te content increased, the retention characteristics of the programmed resistance state became unstable since the short-range-ordered Te-Te clusters in the GeTex layer caused the Cu conductive filament to dissolve in Cu-Te locally. Experimental results indicated that the chalcogenide materials were good candidates for a highly reliable active layer in a conducting-bridge resistive switching memory device. This report presents various microscopic analysis results to show the superior performance of chalcogenide-based resistive switching memory devices.

Double-Forming Mechanism of TaOx-Based Resistive Memory Device and Its Synaptic Applications.

The bipolar resistive switching properties of Pt/TaOx/InOx/ITO-resistive random-access memory devices under DC and pulse measurement conditions are explored in this work. Transmission electron microscopy and X-ray photoelectron spectroscopy were used to confirm the structure and chemical compositions of the devices. A unique two-step forming process referred to as the double-forming phenomenon and self-compliance characteristics are demonstrated under a DC sweep. A model based on oxygen vacancy migration is proposed to explain its conduction mechanism. Varying reset voltages and compliance currents were applied to evaluate multilevel cell characteristics. Furthermore, pulses were applied to the devices to demonstrate the neuromorphic system's application via testing potentiation, depression, spike-timing-dependent plasticity, and spike-rate-dependent plasticity.

Improved Uniformity of TaOx-Based Resistive Switching Memory Device by Inserting Thin SiO2 Layer for Neuromorphic System.

RRAM devices operating based on the creation of conductive filaments via the migration of oxygen vacancies are widely studied as promising candidates for next-generation memory devices due to their superior memory characteristics. However, the issues of variation in the resistance state and operating voltage remain key issues that must be addressed. In this study, we propose a TaOx/SiO2 bilayer device, where the inserted SiO2 layer localizes the conductive path, improving uniformity during cycle-to-cycle endurance and retention. Transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) confirm the device structure and chemical properties. In addition, various electric pulses are used to investigate the neuromorphic system properties of the device, revealing its good potential for future memory device applications.

Activating M1 muscarinic cholinergic receptors induces destabilization of resistant contextual fear memories in rats

Destabilization of previously consolidated memories places them in a labile state in which they are open to modification. However, strongly encoded fear memories tend to be destabilization-resistant and the conditions required to destabilize such memories remain poorly understood. Our lab has previously shown that exposure to salient novel contextual cues during memory reactivation can destabilize strongly encoded object location memories and that activity at muscarinic cholinergic receptors is critical for this effect. In the current study, we similarly targeted destabilization-resistant fear memories, hypothesizing that exposure to salient novelty at the time of reactivation would induce destabilization of strongly encoded fear memories in a muscarinic receptor-dependent manner. First, we show that contextual fear memories induced by 3 context-shock pairings readily destabilize upon memory reactivation, and that this destabilization is blocked by systemic (ip) administration of the muscarinic receptor antagonist scopolamine (0.3 mg/kg) in male rats. Following that, we confirm that this effect is dorsal hippocampus (dHPC)-dependent by targeting M1 receptors in the CA1 region with pirenzepine. Next, we show that more strongly encoded fear memories (induced with 5 context-shock pairings) resist destabilization. Consistent with our previous work, however, we report that salient novelty (a change in floor texture) presented during the reactivation session promotes destabilization of resistant contextual fear memories in a muscarinic receptor-dependent manner. Finally, the effect of salient novelty on memory destabilization was mimicked by stimulating muscarinic receptors with the selective M1 agonist CDD-0102A (ip, 0.3 mg/kg). These findings reveal further generalizability of our previous results implicating novel cues and M1 muscarinic signaling in promoting destabilization of resistant memories and suggest possible therapeutic options for disorders characterized by persistent, maladaptive fear memories such as PTSD and phobias.

Highly uniform resistive switching characteristics of KGM resistive memory through air plasma technology

In this study, a highly stable resistive switching behavior was obtained by utilizing konjac glucomannan, a biomaterial, as a dielectric layer through a solution-based process, and the ITO bottom electrode was modified by using an air plasma technique. Electrical results showed that the pretreatment device did not observe significant switching characteristics (∼101). The device, which underwent additional plasma treatment, was recognized for its significant improvements in this area, demonstrating a reduced set voltage (Vset = 0.37 V) with a 6% coefficient of variation, more than 200 consecutive cycles, and up to 90% electrical yield. Moreover, the treated device exhibited a dramatically higher memory window (&amp;gt;104) due to the relatively low off-state current. According to the XPS and UPS analysis, the work function increased from an untreated 4.3 to 5.79 eV as the treatment time increased to 180 s. The conductive filaments, which stemmed from oxygen vacancies, were introduced through plasma treatment to enhance the proportion of oxygen vacancies in the ITO films, thereby creating a stable and consistent filament path. Consequently, the systematic and reproducible resistive switching phenomenon was intensified. The reported results confirmed that the reliability and uniformity in bioelectronic devices can be accomplished through a simple and effective plasma technique. This approach paved the way for alternative applications of these devices.

Investigations on RF sputtered TiN thin films and Cu/TiO2/TiN devices for resistive switching memory applications

Investigations on RF sputtered TiN thin films and Cu/TiO2/TiN devices for resistive switching memory applications

Intramolecular Hydrogen Bonding Interactions Induced Enhancement in Resistive Switching Memory Performance for Covalent Organic Framework‐Based Memristors

AbstractCovalent organic frameworks (COFs) are promising active mediums for high‐performance data storage devices stemming from their high crystallinity, ordered porous channels, predetermined topology, and rigid architectures, while relative studies are still limited. Here, two COFs (i.e., COF‐EtD and COF‐EtA)‐based resistive‐switching memory devices are proposed to demonstrate the impact of intramolecular hydrogen bonding interactions, both of which exhibit nonvolatile write‐once‐read‐many (WORM) memory characteristics. Comparatively, COF‐EtD with the hydroxyl motif decoration illustrated a lower SET voltage (i.e., turn‐on voltage) and a higher ON/OFF current ratio as well as exceptional high‐temperature endurance and solvent resistance. Mechanism explorations suggest that the resistance‐switching behavior of COF‐EtD may be controlled by the synergistic effect of space‐charge‐limited current and conductive filaments. This finding is the first example of simply varying the functional group of COFs to regulate the WORM resistive switching behavior, providing some inspiration for the advancement of electronic memories and positively contributing to the realization of high endurance and high‐density electric storage materials.

Multilevel resistive switching memory in lead-free double perovskite La_{2}NiFeO_{6} films

Dense and flat La_{2}NiFeO_{6} (LNFO) films were fabricated on the indium tin oxide-coated glass (ITO/glass) substrate by sol–gel method. The bipolar resistive switching behavior (BRS) could be maintained in 100 cycles and remained after 30 days, indicating that the LNFO-based RS device owned good memory stability. Surprisingly, the multilevel RS characteristics were firstly observed in the Au/LNFO/ITO/glass device. The high resistance states (HRSs) and low resistance state (LRS) with the maximum ratio of sim 500 could be remained stably in 900 s and 130 cycles, demonstrating the fine retention and endurance ability of this LNFO-based RS device. The BRS behavior of Au/LNFO/ITO/glass devices primarily obeyed the SCLC mechanism controlled by oxygen vacancies (OVs) dispersed in the LNFO layer. Under the external electric field, injected electrons were captured or discharged by OVs during trapping or detrapping process in the LNFO layer. Thus, the resistive state switched between HRS and LRS reversibly. Moreover, the modulation of Schottky-like barrier formed at the Au/LNFO interface was contributed to the resistive states switchover. It was related to the change in OVs located at the dissipative region near the Au/LNFO interface. The multilevel RS ability of LNFO-based devices in this work provides an opportunity for researching deeply on the high density RS memory in lead-free double perovskite oxides-based devices.

Effect of HBr additive on the performance of all-inorganic Cs3Bi2Br9 halide perovskite resistive switching memory

Non-toxic, stable Cs3Bi2Br9 halide perovskite films prepared using a one-step solution method have become the active layer of resistive switching memory. Here, we control the nucleation and growth of Cs3Bi2Br9 perovskite by introducing an appropriate amount of HBr to adjust the supersaturation of the perovskite precursor solution, thus improving the film formation quality and optimizing the device performance. The memory prepared for this film has a significant bipolar resistive behavior with a high switching ratio of 105, a cycle time of 300 and a retention time of 104 s. In addition, the current transient response of the device under pulsed voltage is investigated. By different the pulse voltage parameters (pulse amplitude, duration and time interval) to study the variation of the device current value, the term "TJUT" is finally encoded and decoded using different the number of pulses and repetitions.

A Capacitor-Based Synaptic Device with IGZO Access Transistors for Neuromorphic Computing

IGZO TFT has extremely low leakage current characteristics compared to Si-CMOS, so it can be applied not only to displays but also to various research fields. One of the important applications using IGZO TFT is analog-based synaptic device for the efficient implementation for deep learning. The application of IGZO TFT to a synaptic device was more widely studied after a synaptic device based on Si-CMOS and capacitor was devised in 2017(1). Since insufficient retention time was a remaining problem in Si-CMOS and capacitor-based storage synaptic devices, the combination of IGZO TFT with low leakage current and capacitor was a very attractive solution. However, because the viable p-type metal oxide TFT does not exist, linearity and symmetry of weight update, a major requirement of on-chip trainable synaptic devices, has been difficult to achieve. In our research, for the first time, we obtained linear and symmetric weight update characteristics by combining an IGZO TFT and a capacitor.In addition, we verified the leakage characteristics of IGZO TFT and the retention characteristics of synaptic devices, which are the most important characteristics. Through the single TFT measurement result, we were able to obtain a leakage current per channel width 1µm of 10-17A, which is much smaller than the Si transistor. Lower leakage current will lead to better retention characteristics than conventional Si-CMOS-based devices, which will result in higher accuracy in deep neural network training. Based on the published results, the IGZO TFT synapse is projected to be able to achieve the retention time of ~4×109s using only 10fF capcacitor.Another important factor that affects deep neural network training is cell-to-cell and device-to-device variation. As opposed to capacitive synaptic devices, resistive synaptic devices such as Resistive RAM and Phase Change Memory suffer from large device variation. We also verified the device variation characteristics, which are advantages of charge storage devices based on transistors and capacitors. The variation characteristics of the device were evaluated through the ΔG+3σ/ΔG-3σ(%) formula and an excellent result of ~200% was obtained.In our research, we verified the main characteristics for on-chip training of synaptic devices based on IGZO TFT and capacitor: linearity and symmetry of weight update, retention characteristics and variation characteristics. The remarkable characteristics of synaptic devices based on IGZO TFT and capacitor will be a good candidate for on-chip training.1. S. Kim, T. Gokmen, H. M. Lee, and W. E. Haensch, Midwest Symp. Circuits Syst., 2017-Augus, 422–425 (2017).AcknowledgementsThis research was supported by National R&D Program through the National Research Foundation of Korea(NRF) funded by Ministry of Science and ICT(NRF-2020M3F3A2A01081240)