Matthew Marinella

A non-volatile organic electrochemical device as a low-voltage artificial synapse for neuromorphic computing

The brain is capable of massively parallel information processing while consuming only ∼1-100 fJ per synaptic event. Inspired by the efficiency of the brain, CMOS-based neural architectures and memristors are being developed for pattern recognition and machine learning. However, the volatility, design complexity and high supply voltages for CMOS architectures, and the stochastic and energy-costly switching of memristors complicate the path to achieve the interconnectivity, information density, and energy efficiency of the brain using either approach. Here we describe an electrochemical neuromorphic organic device (ENODe) operating with a fundamentally different mechanism from existing memristors. ENODe switches at low voltage and energy (<10 pJ for 103 μm2 devices), displays >500 distinct, non-volatile conductance states within a ∼1 V range, and achieves high classification accuracy when implemented in neural network simulations. Plastic ENODes are also fabricated on flexible substrates enabling the integration of neuromorphic functionality in stretchable electronic systems. Mechanical flexibility makes ENODes compatible with three-dimensional architectures, opening a path towards extreme interconnectivity comparable to the human brain.

A physical model of switching dynamics in tantalum oxide memristive devices

We present resistive switching model for TaOx memristors, which demonstrates that the radius of a tantalum rich conducting filament is the state variable controlling resistance. The model tracks the flux of individual oxygen ions and permits the derivation and solving of dynamical and static state equations. Model predictions for ON/OFF switching were tested experimentally with TaOx devices and shown to be in close quantitative agreement, including the experimentally observed transition from linear to non-linear conduction between RON and ROFF. This work presents a quantitative model of state variable dynamics in TaOx memristors, with direct comparison to high-speed resistive switching data.

Initial Assessment of the Effects of Radiation on the Electrical Characteristics of

Radiation-induced effects on the electrical characteristics of TaOx memristive (or redox) memory are experimentally assessed. 10 keV x-ray irradiation is observed to cause switching of the memristors from high to low resistance states, as well as functional failure due to cumulative dose. Gamma rays and 4.5 MeV energy protons are not observed to cause significant change in resistance state or device function at levels up to 2.5 Mrad(Si) and 5 Mrad(Si) protons, respectively. 105 MeV and 480 MeV protons cause switching of the memristors from high to low resistance states in some cases, but do not exhibit a consistent degradation. 800 keV silicon ions are observed to cause resistance degradation, with an inverse dependence of resistance on oxygen vacancy density. Variation between different devices appears to be a key factor in determining the electrical response resulting from irradiation. The proposed degradation mechanism likely involves the creation of oxygen vacancies, but a better fundamental understanding of switching is needed before a definitive understanding of radiation degradation can be achieved.

{\rm TaO}_{\rm x}

The steady-state solution of filamentary memristive switching may be derived directly from the heat equation, modelling vertical and radial heat flow. This solution is shown to provide a continuous and accurate description of the evolution of the filament radius, composition, heat flow, and temperature during switching, and is shown to apply to a large range of switching materials and experimental time-scales.

Memristive Memories

Charge trapping and slow (from 10 s to >; 1000 s) detrapping in AlGaN/GaN high electron mobility transistors (HEMTs) designed for high breakdown voltages ( >; 1500 V) is studied through a combination of electrical, thermal, and optical methods to identify the impact of Al molefraction and passivation on trapping. Trapping due to 5-10 V drain bias stress in the on-state (Vgs = 0) is found to have significantly slower recovery, compared with trapping in the off-state (Vgs <; Vth, Vds = 0). Two different trapping components, i.e., TG1 (Ea = 0.6 eV) and TG2 (with negligible temperature dependence), in AlGaN dominate under gate bias stress in the off-state. Al0.15Ga0.85N shows much more vulnerability to trapping under gate stress in the absence of passivation than does AlGaN with a higher Al mole fraction. Under large drain bias, trapping is dominated by a much deeper trap TD. Detrapping under monochromatic light shows TD to have Ea 1.65 eV. Carbon doping in the buffer is shown to introduce threshold voltage shifts, unlike any of the other traps.

Isothermal Switching and Detailed Filament Evolution in Memristive Systems

Standard deposition processes for depositing ReRAM oxides utilize mass flow of reactive gas to control stoichiometry and have difficulty depositing a precisely defined sub-stoichiometry within a “forbidden region” where film properties are discontinuous with mass flow. We show that by maintaining partial pressure within this discontinuous “forbidden region,” instead of by maintaining mass flow, we can optimize tantalum oxide device properties and reduce or eliminate the electroforming step. We also show that defining the partial pressure set point as a fraction of the “forbidden region” instead of as an absolute value can be used to improve wafer-to-wafer consistency with minimal recalibration efforts.

Slow Detrapping Transients due to Gate and Drain Bias Stress in High Breakdown Voltage AlGaN/GaN HEMTs

In this paper, we present a review of the state of the art in memristor technologies. Along with ionic conducting devices [i.e., conductive bridging random access memory (CBRAM)], we include phase change, and organic/organo-metallic technologies, and we review the most recent advances in oxide-based memristor technologies. We present progress on 3-D integration techniques, and we discuss the behavior of more mature memristive technologies in extreme environments.

Optimizing TaOx memristor performance and consistency within the reactive sputtering “forbidden region”

A major class of resistive memory devices is based on transition metal oxides, where mobile oxygen vacancies allow these devices to exhibit multiple resistance states. Ta2O5 based devices in particular have recently demonstrated impressive endurance and forming-free results. Deposition of substoichiometric Ta2Ox (x < 5) films is a critical process in order to produce the required oxygen vacancies in these devices. This paper describes a physical vapor deposition (PVD) reactive sputtering process to deposit substoichiometric Ta2Ox films. The desired film stoichiometry is achieved by feedback control of the oxygen partial pressure in the PVD chamber. A calibration procedure based on Rutherford backscattering spectroscopy is described for locating the optimum oxygen partial pressure.

Reconfigurable Memristive Device Technologies

We apply first-principles density-functional theory (DFT) calculations, ab-initio molecular dynamics, and the Kubo-Greenwood formula to predict electrical conductivity in Ta2Ox (0 ≤ x ≤ 5) as a function of composition, phase, and temperature, where additional focus is given to various oxidation states of the O monovacancy (VOn; n = 0,1+,2+). In the crystalline phase, our DFT calculations suggest that VO0 prefers equatorial O sites, while VO1+ and VO2+ are energetically preferred in the O cap sites of TaO7 polyhedra. Our calculations of DC conductivity at 300 K agree well with experimental measurements taken on Ta2Ox thin films (0.18 ≤ x ≤ 4.72) and bulk Ta2O5 powder-sintered pellets, although simulation accuracy can be improved for the most insulating, stoichiometric compositions. Our conductivity calculations and further interrogation of the O-deficient Ta2O5 electronic structure provide further theoretical basis to substantiate VO0 as a donor dopant in Ta2O5. Furthermore, this dopant-like behavior is sp...

Reactive sputtering of substoichiometric Ta2Ox for resistive memory applications

The exponential increase in data over the last decade presents a significant challenge to analytics efforts that seek to process and interpret such data for various applications. Neural-inspired computing approaches are being developed in order to leverage the computational properties of the analog, low-power data processing observed in biological systems. Analog resistive memory crossbars can perform a parallel read or a vector-matrix multiplication as well as a parallel write or a rank-1 update with high computational efficiency. For an N × N crossbar, these two kernels can be O(N) more energy efficient than a conventional digital memory-based architecture. If the read operation is noise limited, the energy to read a column can be independent of the crossbar size (O(1)). These two kernels form the basis of many neuromorphic algorithms such as image, text, and speech recognition. For instance, these kernels can be applied to a neural sparse coding algorithm to give an O(N) reduction in energy for the entire algorithm when run with finite precision. Sparse coding is a rich problem with a host of applications including computer vision, object tracking, and more generally unsupervised learning.