Characteristics of Resistive Memory Read Fluctuations in Endurance Cycling

We report on new fluctuation dynamics of the high resistance state of Hafnia-based RRAM devices after RESET. We observe that large amplitude fluctuations occur more frequently immediately after programming and their frequency of occurrence decays in the tens of microseconds. The fluctuation amplitude, on the other hand, does not decay noticeably over the entire millisecond read time. While post-programming instability and post-programming resistance dispersion have both been reported in the literature, the relaxation in the frequency of occurrence without a commensurate amplitude decay is new. Since picosecond pulses were used for our RESET operation, post-programming thermalization occurs on the nanosecond time scale. This clearly eliminates a thermally driven mechanism as the cause of the observed fluctuation behavior. Furthermore, reducing the READ voltage by tenfold does not have any effect and also eliminates read disturb as possible cause.

I. Introduction

Scalability, high endurance, and fast switching speeds have solidified Hafnia-based resistive random access memory (RRAM) as a serious candidate for future nonvolatile memory [1]–[3]. Demonstrations of precise resistance state control [4], [5] further extended the application space to include multilevel cells and neuromorphic computing. However, for higher resistance filaments (more desirable for low power operation), observations of large resistance fluctuations threaten RRAM viability as the resistance states are less distinguishable [6], [7]. Mitigation efforts have focused on understanding the origin and impact of these fluctuations [8]–[11]. Unfortunately, most studies focus on the long-term fluctuations that impact retention, not on the much faster fluctuations that impact fast read/write operation. Recently, the resistance distribution was tracked from to 1s after programming and was shown to degrade drastically over that period [12]. This degradation was attributed to two factors – fluctuation amplitude decay over time and stochastic (mean value) relaxation.