Monte-Carlo Simulations of Flash Memory Array Retention

One of the major scalability limitations of flash memories is anomalous SILC, which strongly endangers device reliability and data retention. Therefore, an accurate evaluation of SILC statistics on large arrays is crucial for reliability predictions and new Flash technology development. In the last years, oxide leakage currents were deeply investigated and modeled, neglecting SILC statistics and effects on large Flash arrays. More recently, analytical models relating Flash statistical threshold voltage (VT) distributions to defect statistics and leakage current were proposed. However, these models rely on several simplifying assumptions such as the equivalent cell concept and an uniform defect population. Still, these models do not account for the initial VT distribution and neglect the role played by trap energy and effective field. In this scenario, the purpose of this paper is to present a Monte-Carlo (MC) simulator reproducing flash VT distribution, which overcomes the above model limitations. We will show that this model can be used to 1) investigate effects of defect features and technology parameters on VT distribution, and 2) analyze the impact of temperature and voltage accelerated stresses on final VT distribution.