Understanding the mechanical behavior of composite materials requires extensive knowledge of fracture behavior as a crack approaches an interface between the bulk material and the reinforcement structure. Overall material toughness can be greatly influenced by the propensity of an impinging crack to propagate directly through the substrate or deflect along an interface boundary. As the basis for this thesis; the assertion that an impinging crack may encounter a reinforcement structure at various incident angles is explored. This requires the ability to predict crack penetration/ deflection behavior not only normal to the reinforcement, but at various incident angles. Previous work in the area of interface fracture mechanics has used a stress or energy based approach, with recent advances in the field of a combined cohesive-zone method. Work presented here investigates the interaction between strength and toughness when using the cohesive-zone method on the problem of an impinging crack not normally incident to the interface of a composite material. Computational mechanics methods using Abaqus and user-define cohesive elements will be applied to this angled incident crack problem. A circular model based on the displacement field equations for mode-I fracture loading is introduced and verified against well-established LEFM solutions. This circular model is used to study the effects of incident crack angle on the penetration vs. deflection behavior of an impinging crack at various angles of incidence. Additionally, the effects of angle on the load applied to the model at fracture are explored. Finally, a case study investigating how the interaction between strength and toughness found using the cohesive-zone method helps to explain some of the inconsistencies seen in the interface indentation fracture test procedure. / Graduation date: 2013

No-flow underfill materials that cure during the solder reflow process is a relatively new technology. Although there are several advantages in terms of cost, time and processing ease, there are several reliability challenges associated with no-flow underfills. When micron-sized filler particles are introduced in no-flow underfills to enhance the solder bump reliability, such filler particles could prevent the solder bumps making reliable electrical contacts with the substrate pads during solder reflow, and therefore, the assembly yield would be adversely affected. The use of nano-sized filler particles can potentially improve assembly yield while offering the advantages associated with filled underfill materials. The objective of this thesis is to study the thermo-mechanical reliability of nano-filled epoxy underfills (NFU) through experiments and theoretical modeling. In this work, the thermo-mechanical properties of NFUs with 20-nm filler particles have been measured. An innovative residual stress test method has been developed to measure the interfacial fracture toughness. Using the developed residual stress method and the single-leg bending test, the mode-mixity-dependent fracture toughness for NFU-SiN interface has been determined. In addition to such monotonic interfacial fracture characterization, the interface crack propagation under thermo-mechanical fatigue loading has been experimentally characterized, and a model for fatigue interface crack propagation has been developed. A test vehicle comprising of several flip chips was assembled using the NFU material and the reliability of the flip-chip assemblies was assessed under thermal shock cycles between -40oC and 125oC. The NFU-SiN interfacial delamination propagation and the solder bump reliability were monitored. In parallel, numerical models were developed to study the interfacial delamination propagation in the flip chip assembly using conventional interfacial fracture mechanics as well as cohesive zone modeling. Predictions for interfacial delamination propagation using the two approaches have been compared. Based on the theoretical models and the experimental data, guidelines for design of NFUs against interfacial delamination have been developed.

Reliability of integrated circuit (IC) packages is in great demand for the automotive industry, as they are used in almost every electronic components. IC packages consist of essentially molding compound (MC), lead frame (LF), adhesive and a silicon chip. The elastic mismatch between the components makes the interfaces susceptible to crack initiation, propagation and eventual failure. The main reason of the failure is the thermo-mechanical cycles during the service time. This work presents the robustness estimation and the reliability based robustness improvement of an IC package by minimization of both crack driving force and its standard deviation at the MC and the LF interface with respect to the fatigue fracture toughness. The robustness evaluation and robust design optimization were performed by taking the uncertainty in geometrical parameters into account. Evidently, there are more robust and reliable designs than the current design which have less crack driving force and show less variation. In order to quantify the reliability with respect to the variation of the crack driving force, the fatigue fracture toughness of the interface was characterized under isothermal conditions at 25 ◦C and −40 ◦C with a three point bending test apparatus. The interface characterizations at low temperatures like −40 ◦C is a main concern due to large stress generation during the reliability tests. After then, a test methodology was prepared to validate the fatigue fracture toughness of the interface in the package level. Artificial cracks were introduced at the MC-LF interface in IC packages to predict the crack growth under thermal cycling over a temperature range of −50 ◦C to 150 ◦C. A prediction quality assisted to validate, whether the fatigue fracture toughness, which was obtained mechanically under isothermal conditions, could be used to predict the crack growth in the IC package under thermo-mechanical cycles. Material characterization of the MC and the LF was performed to acquire the fatigue fracture toughness and the crack length by the compliance calibration method as accurate as possible. The mechanical modeling of both materials was accomplished with elasticity plus plasticity at the room temperature. Then the material models were verified by using the behavior of the bi-material structure under three point bending. As the numerical simulations were used to calculate the fracture toughness, this thesis also presents a comparison between the methods in the literature by using finite element simulations. The results were compared with the analytic solution according to their accuracy, ease of implementation and mesh independence. Simultaneously, various crack tip elements were analyzed in contrast considering their capability of fracture toughness calculation. The analyses were included different fracture mechanical concepts from linear elastic to elastic plastic fracture mechanics. The comparison led to a more convenient method and crack tip element preference for the interface characterization.