Search

Set and reset characteristics are studied for unipolar and bipolar metal-oxide resistive-switching memory devices. We show a universal dependence of set-state resistance and reset current on the compliance current used during set, with negligible impact of metal-oxide composition and switching condition. An analytical Joule-heating model for universal reset is presented, predicting a weak dependence of reset temperature and voltage on diffusion and migration parameters in both unipolar- and bipolar-switching modes. Data for the reset voltage are shown for a wide range of unipolar and bipolar metal oxides, in support of our calculations. [ABSTRACT FROM AUTHOR]

Resistive-switching memory (RRAM) devices are attracting a considerable interest in view of their back-end integration, fast programming, and high scalability. Prediction the programming voltages and currents as a function of the operating conditions is an essential task for developing compact and numerical models able to handle a large number (106-109) of cells within an array. Based on recent experimental findings on the set and reset processes, we have developed physics-based analytical models for the set and reset operations in NiO-based RRAMs. Simulation results obtained by the analytical models were compared with experimental data for variable pulse conditions and were found consistent with data. The set transition is described by a threshold switching process at the broken conductive filament (CF), while the reset transition is viewed as a thermally driven dissolution and/or oxidation of the CF. Set and reset models are finally used for reliability predictions of failure times under constant-voltage stress (read disturb) and elevated-temperature bake (data retention). [ABSTRACT FROM AUTHOR]

The physical understanding of the programming and reliability mechanisms in resistive-switching memory devices requires a detailed characterization of the electrical and thermal conduction properties in the low-resistance state of the memory cell. The aim of this paper is the characterization of the conductive filament (CF), which controls the localized current flow in the low resistive state of the cell. Based on a new technique for evaluating the CF temperature during operation, we perform a statistical characterization of the critical filament temperature for the reset operation, i.e., the transition to the high-resistance state by the thermal dissolution of the CF. The thermal resistance of the CF and the activation energy for the dissolution mechanism are then evaluated, allowing for a physics-based numerical modeling of the reset operation based on CF thermal breakup. [ABSTRACT FROM AUTHOR]

This paper addresses the numerical modeling of re- set programming in NiO-based resistive-switching memory. In our model, we simulate electrical conduction and heating in the conductive filament (CF), which controls the resistance of the low resistive (or set) state, accounting for CF thermal-activated dissolution. Employing CF electrical and thermal parameters, which were previously characterized on our NiO-based samples, our calculations are shown to match experimental reset and retention characteristics. Simulations show that reset transition is self-accelerated as a consequence of a positive feedback between the thermal dissolution of the CF and local Joule heating in the CF bottleneck, which can account for the abrupt resistance transition. in experimental data. Finally, the model is used to investigate the reduction of the reset current, which is needed for device application. [ABSTRACT FROM AUTHOR]

In this letter, we present the characterization and modeling of the reset statistics of Pt/NiO/W resistive random access memories. The experimental observations show that the Weibull slopes of both Vreset and Ireset cumulative distributions increase linearly with \1/Ron. The value of Vreset63\% is roughly independent of Ron while Ireset63\% increases with \1/Ron. Fully analytical cell-based models based on the thermal dissolution of conductive filament are proposed for the reset switching statistical distributions, which can account for the experimental results with a remarkable agreement. [ABSTRACT FROM PUBLISHER]

Resistive-switching random access memory (RRAM) devices are attracting increasing interest as a potential candidate for high-density nonvolatile memory devices. One of the main issues toward RRAM feasibility is the reduction of the reset current Ireset necessary to restore the high-resistance state in the device. Ireset can be reduced by controlling the size of the conductive filament responsible for the low-resistance state; however, available data only focus on direct-current reset analysis. This letter addresses Ireset reduction under pulsed operation. Unstable reset behaviors, including set–reset and set instability, are shown to occur during relatively fast pulses and starting from set states with relatively large resistance values. These instability effects limit Ireset reduction, posing a potential issue of minimum reset current achievable in RRAM devices. [ABSTRACT FROM AUTHOR]

NiO-based resistive-switching memory (RRAM) is a promising new technology for high-density nonvolatile storage. The main obstacles to practical application are the large and hardly scalable reset (programming) current and the reliability at high temperature. This letter studies temperature-accelerated data retention in RRAM cells from both experimental and theoretical standpoints, addressing the size/nature of the conductive filament and clarifying the tradeoff between data retention and reset current. [ABSTRACT FROM AUTHOR]

The resistive random access memory (RRAM), based on the reversible switching between different resistance states, is a promising candidate for next-generation nonvolatile memories. One of the most important challenges to foster the practical application of RRAM is the control of the statistical variation of switching parameters to gain low variability and high reliability. In this work, starting from the well-known percolation model of dielectric breakdown (BD), we establish a framework of analysis and modeling of the resistive switching statistics in RRAM devices, which are based on the formation and disconnection of a conducting filament (CF). One key aspect of our proposal is the relation between the CF resistance and the switching statistics. Hence, establishing the correlation between SET and RESET switching variables and the initial resistance of the device in the OFF and ON states, respectively, is a fundamental issue. Our modeling approach to the switching statistics is fully analytical and contains two main elements: (i) a geometrical cell-based description of the CF and (ii) a deterministic model for the switching dynamics. Both ingredients might be slightly different for the SET and RESET processes, for the type of switching (bipolar or unipolar), and for the kind of considered resistive structure (oxide-based, conductive bridge, etc.). However, the basic structure of our approach is thought to be useful for all the cases and should provide a framework for the physics-based understanding of the switching mechanisms and the associated statistics, for the trustful estimation of RRAM performance, and for the successful forecast of reliability. As a first application example, we start by considering the case of the RESET statistics of NiO-based RRAM structures. In particular, we statistically analyze the RESET transitions of a statistically significant number of switching cycles of Pt/NiO/W devices. In the RESET transition, the ON-state resistance (RON) is a key parameter to describe the initial state of the CF. Hence, we subdivide the statistical samples (obtained mainly in a single device) in several RON ranges so as to study the switching statistics as a function of RON. In this regard, we have found that the experimental data can be nicely fit to a Weibull model in all the resistance ranges. Moreover, the distributions significantly change with RON. This change might be even more significant than the device-to-device related variations and, hence, mostly determine the overall statistics of switching parameters. In particular, we have found that the Weibull slopes of both VRESET, IRESET, and PRESET cumulative distributions increase linearly with 1/RON, i.e., they increase with the area of the CF. On the other hand, while the scale factor of the VRESET distribution (V63%) is roughly independent of RON, the scale factor of the distribution of IRESET and PRESET (I63% and P63%) linearly increases with 1/RON. Upon a direct analogy with the cell-based analytical percolation model of oxide BD, two simple geometrical cell-based models, the single path model with variable width and the multiple parallel path model, are proposed to address the RESET statistics. In the limit where these two geometrical models coincide, we incorporate a deterministic model for the RESET switching dynamics based on the self-accelerated thermal dissolution of the CF. With these two ingredients, the complete physics-based model for the RESET statistics is constructed. This analytical model is shown to nicely account for the experimental results with remarkable agreement. [ABSTRACT FROM AUTHOR]