Mirko Hansen

A double barrier memristive device.

We present a quantum mechanical memristive Nb/Al/Al2O3/NbxOy/Au device which consists of an ultra-thin memristive layer (NbxOy) sandwiched between an Al2O3 tunnel barrier and a Schottky-like contact. A highly uniform current distribution for the LRS (low resistance state) and HRS (high resistance state) for areas ranging between 70 μm2 and 2300 μm2 were obtained, which indicates a non-filamentary based resistive switching mechanism. In a detailed experimental and theoretical analysis we show evidence that resistive switching originates from oxygen diffusion and modifications of the local electronic interface states within the NbxOy layer, which influences the interface properties of the Au (Schottky) contact and of the Al2O3 tunneling barrier, respectively. The presented device might offer several benefits like an intrinsic current compliance, improved retention and no need for an electric forming procedure, which is especially attractive for possible applications in highly dense random access memories or neuromorphic mixed signal circuits.

Memristive Hebbian Plasticity Model: Device Requirements for the Emulation of Hebbian Plasticity Based on Memristive Devices

In this work we present a phenomenological model for synaptic plasticity suitable to describe common plasticity measurements of memristive devices. We show evidence that the presented model is basically compatible with advanced biophysical plasticity models, which account for a large body of experimental data on spike-timing-depending plasticity (STDP) as an asymmetric form of Hebbian learning. The basic characteristics of our model are a saturation of the synaptic weight growth and a weight dependent learning rate. Moreover, it accounts for common resistive switching behaviors of memristive devices under voltage pulse application and allows to study essential requirements of individual memristive devices for the emulation of Hebbian plasticity in neuromorphic circuits. In this respect, memristive devices based on mixed ionic/electronic and one exclusively electronic mechanism are explored. The ionic/electronic devices consist of the layer sequence metal/isolator/metal and represent todays most popular devices. The electronic device is a MemFlash-cell which is based on a conventional floating gate transistor in a diode configuration wiring scheme exhibiting a memristive (pinched) I-V characteristic.

The role of ion transport phenomena in memristive double barrier devices

In this work we report on the role of ion transport for the dynamic behavior of a double barrier quantum mechanical Al/Al2O3/NbxOy/Au memristive device based on numerical simulations in conjunction with experimental measurements. The device consists of an ultra-thin NbxOy solid state electrolyte between an Al2O3 tunnel barrier and a semiconductor metal interface at an Au electrode. It is shown that the device provides a number of interesting features such as an intrinsic current compliance, a relatively long retention time, and no need for an initialization step. Therefore, it is particularly attractive for applications in highly dense random access memories or neuromorphic mixed signal circuits. However, the underlying physical mechanisms of the resistive switching are still not completely understood yet. To investigate the interplay between the current transport mechanisms and the inner atomistic device structure a lumped element circuit model is consistently coupled with 3D kinetic Monte Carlo model for the ion transport. The simulation results indicate that the drift of charged point defects within the NbxOy is the key factor for the resistive switching behavior. It is shown in detail that the diffusion of oxygen modifies the local electronic interface states resulting in a change of the interface properties.

Kinetic simulation of filament growth dynamics in memristive electrochemical metallization devices

In this work we report on kinetic Monte-Carlo calculations of resistive switching and the underlying growth dynamics of filaments in an electrochemical metallization device consisting of an Ag/TiO2/Pt sandwich-like thin film system. The developed model is not limited to i) fast time scale dynamics and ii) only one growth and dissolution cycle of metallic filaments. In particular, we present results from the simulation of consecutive cycles. We find that the numerical results are in excellent agreement with experimentally obtained data. Additionally, we observe an unexpected filament growth mode which is in contradiction to the widely acknowledged picture of filament growth, but consistent with recent experimental findings.

Synchronization of two memristive coupled van der Pol oscillators

The objective of this letter is to convey two essential principles of biological computing—synchronization and memory—in an electronic circuit with two van der Pol (vdP) oscillators coupled via a memristive device. The coupling was mediated by connecting the gate terminals of two programmable unijunction transistors through a resistance-capacitance network comprising an Ag-TiOx-Al memristive device. In the high resistance state the memristance was in the order of MΩ, which leads to two independent self-sustained oscillators characterized by the different frequencies f1 and f2 and no phase relation between the oscillations. Depending on the mediated pulse amplitude, the memristive device switched to the low resistance state after a few cycles and a frequency adaptation and phase locking were observed. The experimental results are underlined by theoretically considering a system of two coupled vdP equations. This experiment may pave the way to larger neuromorphic networks in which the coupling parameters (t...

UV capillary force lithography for multiscale structures

Rapid advances in information technology rely on novel patterning techniques. The authors present a simple UV capillary force lithography process, which allows one to imprint a multiscale system, consisting of 250 nm wide nanobridges and a 8–20 μm wide wiring in one lithography step. An additional annealing step for 5 min at 75 °C improved the capillary rise.

Memristive stochastic plasticity enables mimicking of neural synchrony: Memristive circuit emulates an optical illusion.

Memristive devices help address the binding problem: Their memory supports a transient connectivity in oscillator networks. The human brain is able to integrate a myriad of information in an enormous and massively parallel network of neurons that are divided into functionally specialized regions such as the visual cortex, auditory cortex, or dorsolateral prefrontal cortex. Each of these regions participates as a context-dependent, self-organized, and transient subnetwork, which is shifted by changes in attention every 0.5 to 2 s. This leads to one of the most puzzling issues in cognitive neuroscience, well known as the “binding problem.” The concept of neural synchronization tries to explain the problem by encoding information using coherent states, which temporally patterns neural activity. We show that memristive devices, that is, a two-terminal variable resistor that changes its resistance depending on the previous charge flow, allow a new degree of freedom for this concept: a local memory that supports transient connectivity patterns in oscillator networks. On the basis of the probability distribution of the resistance switching process of Ag-doped titanium dioxide memristive devices, a local plasticity model is proposed, which causes an autonomous phase and frequency locking in an oscillator network. To illustrate the performance of the proposed computing paradigm, the temporal binding problem is investigated in a network of memristively coupled self-sustained van der Pol oscillators. We show evidence that the implemented network allows achievement of the transition from asynchronous to multiple synchronous states, which opens a new pathway toward the construction of cognitive electronics.

A memristive spiking neuron with firing rate coding.

Perception, decisions, and sensations are all encoded into trains of action potentials in the brain. The relation between stimulus strength and all-or-nothing spiking of neurons is widely believed to be the basis of this coding. This initiated the development of spiking neuron models; one of todays most powerful conceptual tool for the analysis and emulation of neural dynamics. The success of electronic circuit models and their physical realization within silicon field-effect transistor circuits lead to elegant technical approaches. Recently, the spectrum of electronic devices for neural computing has been extended by memristive devices, mainly used to emulate static synaptic functionality. Their capabilities for emulations of neural activity were recently demonstrated using a memristive neuristor circuit, while a memristive neuron circuit has so far been elusive. Here, a spiking neuron model is experimentally realized in a compact circuit comprising memristive and memcapacitive devices based on the strongly correlated electron material vanadium dioxide (VO2) and on the chemical electromigration cell Ag/TiO2−x/Al. The circuit can emulate dynamical spiking patterns in response to an external stimulus including adaptation, which is at the heart of firing rate coding as first observed by E.D. Adrian in 1926.

An enhanced lumped element electrical model of a double barrier memristive device

The massive parallel approach of neuromorphic circuits leads to effective methods for solving complex problems. It has turned out that resistive switching devices with a continuous resistance range are potential candidates for such applications. These devices are memristive systems - nonlinear resistors with memory. They are fabricated in nanotechnology and hence parameter spread during fabrication may aggravate reproducible analyses. This issue makes simulation models of memristive devices worthwhile. Kinetic Monte-Carlo simulations based on a distributed model of the device can be used to understand the underlying physical and chemical phenomena. However, such simulations are very time-consuming and neither convenient for investigations of whole circuits nor for real-time applications, e.g. emulation purposes. Instead, a concentrated model of the device can be used for both fast simulations and real-time applications, respectively. We introduce an enhanced electrical model of a valence change mechanism (VCM) based double barrier memristive device (DBMD) with a continuous resistance range. This device consists of an ultra-thin memristive layer sandwiched between a tunnel barrier and a Schottky-contact. The introduced model leads to very fast simulations by using usual circuit simulation tools while maintaining physically meaningful parameters. Kinetic Monte-Carlo simulations based on a distributed model and experimental data have been utilized as references to verify the concentrated model.

Double-Barrier Memristive Devices for Unsupervised Learning and Pattern Recognition

The use of interface-based resistive switching devices for neuromorphic computing is investigated. In a combined experimental and numerical study, the important device parameters and their impact on a neuromorphic pattern recognition system are studied. The memristive cells consist of a layer sequence Al/Al2O3/NbxOy/Au and are fabricated on a 4-inch wafer. The key functional ingredients of the devices are a 1.3 nm thick Al2O3 tunnel barrier and a 2.5 mm thick NbxOy memristive layer. Voltage pulse measurements are used to study the electrical conditions for the emulation of synaptic functionality of single cells for later use in a recognition system. The results are evaluated and modeled in the framework of the plasticity model of Ziegler et al. Based on this model, which is matched to experimental data from 84 individual devices, the network performance with regard to yield, reliability, and variability is investigated numerically. As the network model, a computing scheme for pattern recognition and unsupervised learning based on the work of Querlioz et al. (2011), Sheridan et al. (2014), Zahari et al. (2015) is employed. This is a two-layer feedforward network with a crossbar array of memristive devices, leaky integrate-and-fire output neurons including a winner-takes-all strategy, and a stochastic coding scheme for the input pattern. As input pattern, the full data set of digits from the MNIST database is used. The numerical investigation indicates that the experimentally obtained yield, reliability, and variability of the memristive cells are suitable for such a network. Furthermore, evidence is presented that their strong I–V non-linearity might avoid the need for selector devices in crossbar array structures.