Number of traps and trap depth position on statistical distribution of random telegraph noise in scaled NAND flash memory

The dependence of random telegraph noise (RTN) amplitude distribution on the number of traps and trap depth position is investigated using three-dimensional Monte Carlo device simulation including random dopant fluctuation (RDF) in a 30 nm NAND multi level flash memory. The ΔVth tail distribution becomes broad at fixed double traps, indicating that the number of traps greatly affects the worst RTN characteristics. It is also found that for both fixed single and fixed double traps, the ΔVth distribution in the lowest cell threshold voltage (Vth) state shows the broadest distribution among all cell Vth states. This is because the drain current flows at the channel surface in the lowest cell Vth state, while at a high cell Vth, it flows at the deeper position owing to the fringing coupling between the control gate (CG) and the channel. In this work, the ΔVth distribution with the number of traps following the Poisson distribution is also considered to cope with the variations in trap number. As a result, it is found that the number of traps is an important factor for understanding RTN characteristics. In addition, considering trap position in the tunnel oxide thickness direction is also an important factor.