No abstract

x direction are assumed. The real situation is, however, different and a more realistic model should take into account the above complications. We intend to develop a mathematical treatment of such a model in the near future. AcknowledgmentThe author would like to acknowledge helpful discussions with ABSTRACT The process whereby CVD SisN4 films are etched in acidic fluoride solutions has been investigated at 25 ~ and 60~ by use of a series of well-characterized etchants. The rate law for the dissolution process has been identified and is found to be linear in both [HF] and [HF2-], but independent of [F-]. Although the exact magnitude of the etch rates of CVD SisN4 is known to depend on the deposition conditions used, it is thought that the general form of the rate law will remain constant. Comparison between the rate constants for SisN4 and those previously reported for SiO2 suggests that the etch rate of SigN4 can be increased relative to that of SiO2 by use of elevated temperatures and very low pH solutions, and by adjustment of the nitride deposition parameters so as to increase the NHJSi ratio in the reaction mixture. ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 138.251.14.35 Downloaded on 2015-03-08 to IP ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 138.251.14.35 Downloaded on 2015-03-08 to IP Vol. I25, No. 2

Solutions of concentrated (49%) aqueous HF in glycerol have been found to etch CVD Si3N4 films faster than thermally grown SiO2 over a wide temperature range. Since silicon dioxide films are etched much faster than silicon nitride films in aqueous HF media, this result is quite surprising, and the mechanism for the etch rate reversal is not understood at present. HF/normalGlycerol mixtures have been used to pattern a variety of Si3N4/SiO2/normalSi composite structures using either photoresist or metal masks.

not Available.

Corrosion sites on IC devices are often difficult to detect before a significant amount of metallization deterioration occurs. The identification of corrosion sites in their early stages would be a powerful tool in studies of metallized IC device reliability and of corrosion processes themselves. Recently, this laboratory has been investigating metallization corrosion phenomena on IC devices. During these studies we explored the use of pH‐sensitive fluorescent dyes to decorate the IC device surface. After application ofpH‐sensitive fluorescent dye molecules to the surface of the IC test device, an external d‐c bias was applied to the device which activated or deactivated localized fluorescence on the device. We believe this observed behavior relates to corrosion phenomena occurring on the metallization of the IC. Here we report the results of these exploratory experiments conducted on aluminum‐ and gold‐metallized IC test devices.