Continuous Electrical Tuning of the Chemical Composition of TaOx-Based Memristors

Hewlett-Packard Laboratories, 1501 Page Mill Road, Palo Alto, California 94304, USA.
ACS Nano (Impact Factor: 12.03). 02/2012; 6(3):2312-8. DOI: 10.1021/nn2044577
Source: PubMed

ABSTRACT TaO(x)-based memristors have recently demonstrated both subnanosecond resistance switching speeds and very high write/erase switching endurance. Here we show that the physical state variable that enables these properties is the oxygen concentration in a conduction channel, based on the measurement of the thermal coefficient of resistance of different TaO(x) memristor states and a set of reference Ta-O films of known composition. The continuous electrical tunability of the oxygen concentration in the channel, with a resolution of a few percent, was demonstrated by controlling the write currents with a one transistor-one memristor (1T1M) circuit. This study demonstrates that solid-state chemical kinetics is important for the determination of the electrical characteristics of this relatively new class of device.

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: To facilitate the development of memristive devices, it is essential to resolve the problem of non-uniformity in switching, which is caused by the random nature of the filamentary switching mechanism in many resistance switching memories based on transition metal oxide. In addition, device parameters such as low- and high-state resistance should be regulated as desired. These issues can be overcome if memristive devices have switching limits for both the low- and high-resistance states and if their resistance values are highly controllable. In this study, a method termed self-limited switching for uniformly regulating the values of both the low- and high-resistance states is suggested, and the circuit configuration required for the self-limited switching is established in a Ta2O5/TaOx memristive structure. A method of improving the uniformity of multi-level resistance states in this memristive system is also proposed.
    Advanced Functional Materials 01/2015; DOI:10.1002/adfm.201403621 · 10.44 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We illustrate a heuristic two-state-variable memristor model of charged O vacancy-drift resistive switches that include the effects of internal Joule heating on both the electronic transport and the drift velocity (i.e., switching speed) of vacancies in the switching material. The dynamical state variables correspond to the cross-sectional area of a conducting channel in the device and the gap between the end of the channel and one of the electrodes. The model was calibrated against low voltage pulse-sweep and state-test data collected from a TaOx memristor so that the contributions of the channel gap, area and temperature to switching can be analyzed. The model agrees well with experimental results for long switching times and low-to-intermediate voltage operation.
    Applied Physics A 04/2015; 119(1). DOI:10.1007/s00339-015-9033-3 · 1.69 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Due to its fitting the resistive switching behavior of memristor well, the percolation network model has recently attracted attention in the memristive mechanism field. However, the current 2D percolation network model lacks the pristine states analysis. In this paper, the original model is simplified to study the effects of pristine state on the forming process of conductive percolation channel with the increase of applied voltage. Intuitively, such a percolation network model not only demonstrates the dynamic process of local conducting channels formed in the physical meaning, which verifies that the resistance distribution of the memristor switching is not ideally bistable but can be fitted by Gauss curve; also it contributes to deciphering the unknown conductive mechanisms of memristor with the various types of percolation channel.
    Acta Physica Sinica -Chinese Edition- 05/2013; 62(9):096401. DOI:10.7498/aps.62.096401 · 0.85 Impact Factor