Phase nucleation and evolution mechanisms in heterogeneous solids

Anirudh Udupa, Purdue University

Abstract

Phase nucleation and evolution is a problem of critical importance in many applications. As the length scales are reduced, it becomes increasingly important to consider interfacial and micro-structural effects that can be safely ignored at larger length scales owing to randomness. The theory of phase nucleation has been addressed usually by the classical nucleation theory, which was originally derived for single component fluid systems, after making an assumption of equilibrium. The criterion has not been rigorously derived for solids, which are far from equilibrium due to dissipation by multiple physical drivers. In this thesis, a thermodynamically sound nucleation criterion is derived for systems with multiple interacting physical phenomena and multiple dissipating mechanisms. This is done, using the tools of continuum mechanics, by determining the change in free energy upon the introduction of a new nucleus into the system. The developed theory is demonstrated to be a generalization of the classical nucleation theory (CNT). The developed theory is then applied to the problem of electromigration driven void nucleation, a serious reliability concern for the microelectronics industry. The void grows and eventually severs the line making the chip nonfunctional. There are two classes of theories at present in the electromigration literature to address the problem of void nucleation, the vacancy supersaturation theory and the entropic dissipation theory, both of which are empirical and based on intuition developed from experimental observations. When the developed theory was applied to the problem of electromigration, it was found to be consistent with the vacancy supersaturation theory, but provided the correct energetic quantity, the chemical potential, which has contribution from both the vacancy concentration as well as the hydrostatic stress. An experiment, consisting of electromigration tests on serpentine lines, was developed to validate the developed nucleation theory. The experimental results are consistent with the developed theory and show that the theory of entropic dissipation is incorrect. A diffuse-interface computational technique was then developed to simulate the problem of electromigration driven void nucleation and growth in arbitrary geometries. Experimentally known results such as Black's law, existence of the Blech length, effect of interface adhesion energy were reproduced. The simulations were also used to infer the numerical value of the nucleation criterion, based on experimental results in the literature. The problem of electromigration is the result of species diffusion due to imparted momentum from the electrons, and the resulting motion of interface is influenced by surface diffusion along the interface, bulk diffusion, and the current density. Similarly, the formation of intermetallic compounds (IMC) and the resulting interface shape in many systems is the result of limiting effects of bulk diffusion, interfacial reaction, surface energy, and surface diffusion. Thus, the dynamics and stability of the interface formed when Cu and Sn react to form the IMC compound Cu6Sn5 is explored next. This system is of significant relevance to modern microelectronic chip assemblies, where solder joints with significant Cu6Sn5 volume fraction are known to be prone to brittle fracture and shorter useful life. Prior experimental observations have shown the interface to possess either a scalloped, flat or needle shaped morphology. The governing mechanism leading to the observed shape of the interface is not clearly known, and is the focus of the present study. In research unrelated to diffusion driven phase evolution, but involving interfaces nevertheless, in the appendix, the problem of interfacial delamination in Through Silicon Vias (TSV) is studied analytically. Three-dimensional (3D) packages utilizing TSVs are seen as enablers of increased performance and "More than Moore" functionality at the present time. However, the use of TSVs introduce a set of reliability concerns, one of which is the thermo-mechanical stress caused by the mismatch in coefficient of thermal expansion (CTE) between the copper via and the surround- ing silicon. The CTE mismatch, causes high stress zones in and around the copper TSVs, which in turn impede the mobility of electrons in the regions surrounding the TSVs. Further, proximal placing of TSVs for improved electrical performance may be restricted by additional stress induced by TSV-TSV interaction. The increased stress of the region surrounding the TSV also increases the risk of interfacial delamination. In order to ensure reliable functioning of 3D chip stacks, design guidelines are necessary on the excluded "keep-out" zone where stress induced by TSVs will impede transistor functionality. Towards this end, we analytically derive, using elasticity theory, the stress field in and around a doubly periodic arrangement of TSVs subjected to a uniform thermal excursion. The model for stress is used to analytically estimate the conditions for interfacial cracks to propagate, as a function of the system geometry and material properties. (Abstract shortened by ProQuest.)

Degree

Ph.D.

Advisors

Subbarayan, Purdue University.

Subject Area

Mechanics|Materials science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS