No abstract

No abstract

No abstract

No abstract

Electromigration performance of a variety of test structures has demonstrated a significant dependence of mean time to failure on topography. In particular, the severity and frequency of steps, both over poly and into and out of contacts, are the primary variables influencing the electromigration results. Materials studies show that the most significant materials parameter that changes at steps is grain size. In fact, it is shown that both lithography dimensions and the presence of topographic variations cause a variation in grain size. The conclusion of the study is that both the topography and lithography of the devices must be comprehended in designing test patterns for the estimation of electromigration performance of actual product devices.During standard production electromigration monitoring it was observed that two processes with similar metallization had distinctly different electromigration characteristics. Process A and Process B, having similar metal deposition process and electromigration test conditions (temperature and current), were found to have a significant difference in observed Mean Time To Failure (MTTF). Interestingly Process A, despite having a narrower line width (3.3,um versus 4.Om on Process B), showed a higher MTTF. Figure 1 shows the distributions of observed MTTF for Process A and B, graphical ly illustrating the difference in electromigration performance. In fact Process B results, relative to Process A, were approximately 10 times worse than could be predicted from cross sectional area alone.