Isothermal versus standard wafer electromigration test for the characterization of metal systems

The isothermal electromigration test has recently been introduced as a potential alternative to the SWEAT technique. In the isothermal test, the stress current is adjusted in a feedback loop so that a constant metal line temperature is maintained. This differs from the SWEAT test, where the mean time to failure (MTF) as predicted by Black's equation is the control parameter. Once at the control temperature, the isothermal test can utilize three alternative stress modes: constant power, constant current or constant resistance (temperature). The isothermal test has several important advantages, including: (1) simplicity of implementation; (2) line temperature is directly controlled, so it is easier to avoid the bulk diffusion region; (3) algorithm does not require empirical adjustment of acceleration factors; (4) alternative stress modes are available to optimize performance. The purpose of this paper is to compare the isothermal and SWEAT tests experimentally in terms of MTF and standard deviation, and to evaluate the three different isothermal test modes of constant power, constant current or constant resistance.