Analytical Modeling of RTN with Short Dwell Time for Strict Memory Bit Error Estimation

We derive an analytical model of the average current distribution of random telegraph noise (RTN). This model comprehensively and correctly reproduces experimental current distributions. Moreover, the RTN parameters extracted from experimental data using the proposed model are consistent with those derived by a conventional method. In addition, the average dwell time parameters can be estimated, even when they are shorter than the measurement integration time. Such estimation is difficult using previous methods. Furthermore, the proposed analytical model allows to strictly estimate the RTN-induced memory bit error rates. Hence, this method can be particularly useful for predicting extremely rare bit errors caused by RTN with a very short dwell time.