Random and localized resistive switching observation in Pt/NiO/Pt

Abstract

1 Introduction Resistive switching of transition metal oxides has attracted significant attention as promising materials for the future high-density, nonvolatile memories [1]. This is because metal-TMO-metal devices showed high endurance, excellent retention properties, fabrication compatibility with Si, and high-density scalability [2]. However, it has a significant drawback in random fluctuation on each memory switching parameter [3,4], i.e., R on , R off , V SET , and V RESET . Here, R on and R off represent the resistance of low and high resistive states, and V SET and V RESET are the required voltages for the switching from R off to R on and vice versa. The fluctuation causes a memory window failure and thus, it is of paramount importance for the application of the devices. Many researchers [5,6] have made strong suggestion that the resistive switching is the result of the existence of a large number of conducting filaments spanning the insulating matrix, which arise during the electroforming process [7]. The switching occurs in the "formed" devices via the formation and rupture of the filamentary paths, which are considered to be localized and randomly distributed over the devices [8]. Recently the filaments were observed electrically in NiO [9] as well as TiO 2 [10]. In addition, for both NiO and TiO 2 , impedance spectroscopic studies strongly suggested the filamentary

Analysis of the Switching Variability in <inline-formula> <tex-math notation="TeX">$\hbox{Ni/HfO}_{2}$</tex-math></inline-formula>-Based RRAM Devices

“…These different distributions could be attributed to a different size, shape, or number of CFs. This result demonstrates the random nature of the conductive path formation through percolation processes [13], [15], [18]. Fig.…”

Physical electro-thermal model of resistive switching in bi-layered resistance-change memory

“…A uniform doping concentration of n D =1 × 10 22  cm −3 was defined within the CF and TaO x layer as the initial state (i.e., the forming state). The CF size was set to a diameter of 10 nm, in agreement with direct evaluations by using conductive atomic force microscopy23.…”

Random Circuit Breaker Network Model for Unipolar Resistance Switching

“…Note that for all the current maps in this study, we applied a bias of 0.1 V to the C-AFM tip. [22] To explain unipolar RS, we propose a new percolation model, the RCB network model. As shown in Figure 2c, the C-AFM current map of the electrically formed surface, that is, the LRS, reveals several highly conductive spots, typically with diameters of about 3-10 nm.…”