A precise method to determine the stress current to be applied to electromigration test structures

The density of the stress current flowing through metal stripes for electromigration (EM) median time to failure (MTF) tests should vary as little as possible among all the structures under test. This paper proposes an alternative method to the procedure described in ASTM F 1260, to determine the stress currents for these tests, resulting in smaller stress current density differences among different test structures. This method uses the average resistivity of the metal film to calculate the stress current for each individual test structure. In the standard method (ASTM F 1260), the average film thickness is used to calculate the cross-sectional area, which is used to calculate the stress current. The resistivity of the deposited film (almost always) varies less over a wafer than the film thickness. The alternative method results in an improved precision of the stress current density. A comparison between the two methods is made and two experiments are used to illustrate the difference. For Al-1at.% Si layers deposited in our sputtering system, the precision of the stress current density is improved by a factor of three.