Improving the accuracy of defect mobilities and observing interface effects of resistive switching memories using the current transient phenomenon

Abstract

Resistive switching memories are a serious contender for high density computing memory in the future. The operation of valence change memories specifically relies on a migration of defects within the oxide layer. This is assumed to be the case in our silicon dioxide devices[1]. It is therefore crucial we understand the properties of the migrating defects, such as their mobilities, so that we can simulate devices more effectively and with better precision. We are investigating a phenomenon observed within oxide capacitors known as the current transient[2] as one potential mechanism to probe defect mobilities and their effect on interface barriers. The current transient phenomenon arises when a constant voltage is applied to a leaky oxide capacitor causing a distinctive transient to be observed. The device current initially increases rapidly up to a maximum, then levels out, and finally decays slowley. This results in a characteristic peak observed in the current-time plot. Although we have previously used it to carry out computations[3], it has also been suggested that oxygen vacancy mobilities can be obtained from such transients[4] by analysing the dynamics using a space charge limited current (SCLC) model[2,5]. However, our investigations into this effect suggest that the SCLC theory may not be an accurate representation of the changes occurring within our own devices. To investigate the validity of the SCLC model we have developed techniques to repeatably induce current transients within devices as well as novel analytical techniques that reveal more information on the dynamics of the transient. In doing so, we have been able to develop an alternative model that may help explain the current transient as a two-part process which is the result of charge injection and a field drive migration of defects occurring in parallel. While this model differs to the SCLC theory, it still predicts defect mobility to be a measurable observable and may provide a more accurate representation of this quantity. Additionally, it may allow us to determine interface barrier heights at the source of electron injection - something which cannot be determined from the SCLC analysis. Our new approach to analysing current transients may prove to be an important probe into the dynamics of defects within resistive switching memories. 1 A. Mehonic, M.S. Munde, W.H. Ng, M. Buckwell, L. Montesi, M. Bosman, A.L. Shluger, and A.J. Kenyon, Microelectronic Engineering 178, 98 (2017). 2 A. Many and G. Rakavy, Physical Review 126, 1980 (1962). 3 D.J. Mannion, A. Mehonic, W.H. Ng, and A.J. Kenyon, Frontiers in Neuroscience 13, 1386 (2020). 4 S. Zafar, R.E. Jones, B. Jiang, B. White, P. Chu, D. Taylor, and S. Gillespie, Applied Physics Letters 73, 175 (1998). 5 M.A. Lampert and P. Mark, (1970).