Temperature-dependent standard deviation of log(failure time) distributions

C.K. Chan
DOI: https://doi.org/10.1109/24.87118
IF: 5.883
1991-06-01
IEEE Transactions on Reliability
Abstract:The distributions of failure times at accelerated life-test temperatures are often analyzed assuming a constant- sigma model, where sigma is the standard deviation of log(failure times). Recent experimental and theoretical studies of laser degradation, time-dependent dielectric breakdown, and electromigration suggest that sigma is temperature-dependent. The author derives a general relationship for temperature-dependent sigma from the Arrhenius relation. The relationship is used to analyze published electromigration data. The analysis suggests that the correlation coefficient between the activation energy and log(pre-exponential factor in the Arrhenius relation) close to -1. An example illustrates the use of the relationship to validate the electromigration design rule of a hypothetical microcircuit technology. The author argues that assuming a constant sigma could result in an invalid design rule.<>
engineering, electrical & electronic,computer science, software engineering, hardware & architecture
What problem does this paper attempt to address?