High resolution electromigration measurements for reduction of the test time

Abstract

This paper presents a faster method to assess the electromigration performance of metal tracks. The method is based on high accuracy measurements (in the PPM range) which allow monitoring of resistance variations in NIST metal tracks. It is shown that, due to the high accuracy, differences in degradation can be observed after a relatively short stress time and the failure criterion for time-to-failure (TTF) can be decreased. By decreasing the failure criterion for resistance changes from the typical range of 20 to 30% to a range of 1 to 5%, electromigration test time can be reduced by a factor of four. It was also shown that reducing the failure criterion has no impact on the determined values of thermal activation energy E, and current acceleration factor n. Finally, a new strategy is proposed: by using a 1% failure criterion for determination of E/sub a/ and n (at 3 temperatures and 3 current densities) and a 30% failure criterion for determination of the cumulative failure curve (1 temperature and 1 current density), the total test time duration can be reduced by a factor of four and provides the same information than a conventional electromigration test.