Analysis of a Bow-Free Prestressed Test Specimen

Abstract

Broadening the temperature range in accelerated testing of electronic products is a typical measure to assure that the product of interest is sufficiently robust. At the same time, a too broad temperature range can lead to the shift in the modes and mechanisms of failure, i.e., result in failures that will not occur in actual operation conditions. Application of mechanical prestressing of the test specimen could be an effective means for narrowing the temperature range during accelerated testing and thereby achieving trustworthy and failure-mode-shift-free accelerated test information. Accordingly, simple engineering predictive models are developed for the evaluation of the magnitude and the distribution of thermal and mechanical stresses in a prestressed bow-free test specimen. A design, in which an electronic or a photonic package is bonded between two identical substrates, is considered. Such a design is often employed in some today's packaging systems, in which the “inner,” functional, component containing active and/or passive devices and interconnects is placed between two identical “outer” components (substrates). The addressed stresses include normal stresses acting in the component cross sections and the interfacial shearing and peeling stresses. Although the specimen as a whole remains bow-free, the peeling stresses might be nevertheless appreciable, since the outer components, if thin enough, deflect to a greater or lesser extent with respect to the inner component. The numerical example has indicated that the maxima of the interfacial thermal shearing and peeling stresses are indeed comparable and that these maxima are on the same order of magnitude as the normal thermal stresses acting in the components' cross sections. It is shown that since the thermal and the prestressing mechanical loads are of different physical nature, the stresses caused by these two load categories are distributed differently over the specimen's length. It is shown also that although it is possible and even advisable to apply mechanical prestressing for a lower temperature range, it is impossible to reproduce the same stress distribution as in the case of thermal loading. The obtained results enable one to shed light on the physics of the state of stress in prestressed bow-free test specimens in electronics and photonics engineering.