Metallic thin films—the structural backbone of modern integrated circuits (ICs)—exhibit deformation mechanisms fundamentally distinct from those of bulk metals. Their extreme thinness, strong adhesion to substrates, and passivation constraints tig...
Metallic thin films—the structural backbone of modern integrated circuits (ICs)—exhibit deformation mechanisms fundamentally distinct from those of bulk metals. Their extreme thinness, strong adhesion to substrates, and passivation constraints tightly regulate dislocation kinetics, thereby dictating plastic flow and mechanical failure. In addition, fabrication-induced stresses—such as thermal, interfacial, and growth stresses—generate large internal stress fields, leading to cracking, delamination, and other mechanical failures. Under these conditions, understanding the individual motion of dislocations at small scales becomes essential for capturing the true governing mechanisms of thin-film deformation and failure. As IC architectures advance toward highly heterogeneous systems and increasingly complex packaging technologies, a dislocation-level understanding of thin-film mechanics is critical to ensuring mechanical reliability. This dissertation investigates the micromechanical reliability of metallic thin films through a thin-film-specific dislocation-dynamics (DD) framework.
To begin with, the fundamental tensile behavior of thin Cu films was investigated under various surface boundary conditions. Free surfaces promoted dislocation escape and surface nucleation, resulting in nucleation-induced softening, whereas surface passivation suppressed dislocation motion, generating dislocation pile-ups, back stress, and pronounced strain hardening. The film with a free top surface and a passivated bottom—representative of actual IC structures—exhibited a mixed response governed by competing softening and confinement effects, with confinement ultimately dominating and resulting in overall hardening.
Building upon this model, the strain-rate dependence of fatigue cracking was explored. Experiments and DD simulations showed that crack initiation accelerates at low frequencies due to active dislocation glide, while propagation dominates at high frequencies owing to strain hardening under limited relaxation. A coarse-grained continuum model incorporating dislocation nucleation and depletion kinetics reproduced these trends, confirming the consistency between discrete and continuum scale.
Finally, inhomogeneous deformation modes—represented by bending—revealed the role of geometrically necessary dislocations (GNDs) in strain gradient plasticity (SGP). GND accumulation near the neutral axis produced gradient-dependent hardening proportional to the strain gradient, consistent with Eshelby-type predictions. Non-proportional loading tests showed continuous plastic flow without an elastic gap, validating the incremental SGP formulation.
Overall, this dissertation establishes a mechanistic framework that connects dislocation-level behavior to macroscopic reliability in metallic thin films. It demonstrates how surface constraints, strain rate, and strain gradients govern deformation of thin films. These findings offer fundamental insight into the governing mechanisms of small-scale plasticity and guide the design of mechanically resilient interconnects for next-generation electronic systems.