This paper presents an extensive literature review and assessment of corrosion failure mechanisms encountered during accelerated tests of microelectronic devices. The failure mechanism of primary emphasis is electrolytic metal migration. The metallurgies of interest are silver, gold, copper, and aluminum. Electrochemical investigations of dendritic growth are also reviewed. Mechanistic results from the electrochemical investigations are discussed in light of the empirical results of accelerated tests.

381ChemInform Abstract (corrosion mechanisms encountered during accelerated tests of microelectronic devices; 152 refs.).

Electrolytic metal migration is a mechanism which causes failure in microelectronic devices. This failure mechanism results in dendritic growths which bridge the electrode spacing between adjacent conductors on an insulating substrate and produce electrical failure. Accelerated life tests have been developed to study and characterize metal migration. The accelerated life test for this investigation consists of a doped aqueous/polyethylene glycol solution placed across a biased copper electrode pair. A chemical cleaning procedure is used to provide a reproducible copper electrode surface. The short-time (an indication of the rate of copper dendritic growth) is dependent on the electrode configuration (plane-toplane, point-to-plane, or point-to-point), the electrode spacing (0.05-0.25 mm), the applied potential (2-12V), the ammonium perchlorate electrolyte concentration (0.45-23.0 mm), the water concentration [48-85 mole percent (m/o)], and the temperature (21~176 The short-time is independent of either fiberglass, alumina, or polyimide substrates. The total charge (the integrated current to failure) corresponding to the short-time data is dependent on the electrode configuration, the electrode spacing, the applied potential, and the electrolyte concentration. The total charge is independent of the water concentration and the temperature. ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 130.179.16.201 Downloaded on 2015-06-14 to IP ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 130.179.16.201 Downloaded on 2015-06-14 to IP

Many accelerated electrolytic metal migration and corrosion investigations have been performed in accelerated test chambers with unknown gas concentration and flow characteristics. This paper outlines the development of a wellcharacterized accelerated test chamber. The length (mean flow path) to hydraulic diameter ratio of the closed-circuit rectangular flow channel is 20. Velocity and concentration measurements showed minimal horizontal or vertical gradients in the test section. The range of standard deviations (m/s) for the velocity profiles in the test section was 0.0037-0.16 at mean velocities of 0.11-2.9 m/s, respectively. Similarly, the range of standard deviations [ppm(vol.) SO2] for the SO~ concentration profiles was 1.0-1.2 for a standard gas mixture of 102 _+ I.i ppm SO2 in the air. The effect of gas velocity upon the rate of metal migration was demonstrated in the described test chamber using a new accelerated test in an SO2 environment. This test, the PEG-400 drop test, is a moderate accelerated test with metal migration rates that are less than the water-drop test but greater than a temperature humidity bias test with or without pollutants. The gas velocity was found to have a significant effect upon the rate of copper migration.The effects of gas velocity, humidity, and pollutant concentrations upon the metal migration and corrosion rates can be investigated in this accelerated test chamber.