A brief review of selected aspects of the materials science of ball bonding

Thermosonic ball bonding is a metallurgical process that until recently was rarely the subject of metallurgical analysis. However, in recent years greater focus has been given to the materials science of thermosonic ball-wedge bonding in an attempt to better control and advance its application as an interconnect technology in advanced packaging. As with most materials processes, establishing a scientific understanding of the process requires knowledge from various sub-disciplines of physical science. This article briefly reviews selected aspects of the materials science of ball bonding, focussing on 1st and 2nd bond formation and intermetallic growth.     

The static contact angle hysteresis obtained by different experiments for the system PTFE/water

Different experimental methods have been used to determine the static contact angle hysteresis of the system polytetrafluoroethylene/water and the results compared. While the Wilhelmy plate method is not influenced by methodical variations, contact angles determined by the sessile drop and the pendant bubble methods vary with the drop or bubble diameter up to a minimal diameter d_K of the contact area with the solid. This condition seems to be a universal one and should always be checked to ensure that the measured values are comparable. Contact angles calculated from the geometrical parameters of a drop or bubble should be used with care. The surface energetic characters for the PTFE/water are δθ = 19.5°, θ_{a, e} = 108.5° and θ_{r, e} = 89°.     

Modification of the vermont test for monitoring fat adulteration of dairy products

The Vermont test was developed for the routine screening of dairy products for vegetable fat adulteration. A slight modification of the test procedure made it more rapid and reduced costs. Gas-liquid chromatographic analysis of saponified fatty acids in the first and second Mojonnier extractions were proportional; thus, only the first extraction was needed to determine the purity of a milk lipid sample. This modification saves about 8 min/series of four samples and conserves 48% of the volume of organic solvents necessary to perform the classical Mojonnier extraction.     

Oxidation of Au4Al in gold ballbonds

Two different 4N (99.99% purity) gold wires were ballbonded on 1 μm thick Al-1 wt.% Si-0.5 wt.% Cu bondpad metallisation and subjected to high temperature storage (HTS) at 175 °C in air. Each wire type showed ball lift failures, Type A after 500 h and Type B after 1500 h, which in both cases was a result of Au₄Al oxidation. With wire Type A the dominant compound underneath the ball was Au₈Al₃. A thin layer of Au₄Al (≈ 1 μm thick) was observed between the Au₈Al₃ and the gold ball. Ball lift failures occurred in the Au₄Al layer, which appeared to disintegrate due to oxidation and the resulting by products of oxidation were deposited on the underlying and unoxidised Au₈Al₃. With wire Type B, a double layer Au₄Al was dominant after long term ageing and Au₈Al₃ was confined to the ball periphery. Consequently, because of the much greater volume of Au₄Al, compound oxidation resulted in the formation of a large amount of a completely new microstructure consisting of gold precipitates embedded in a dark oxide matrix. The Au₈Al₃ compound remained unoxidised. It is speculated that internal stress and contamination may accelerate the oxidation reaction.     

New aspects of aging in epoxy networks. II. Hydrothermal aging

An amine cured epoxy is prepared in two different network states for hydrothermal aging. The “low‐crosslinked” network has a considerable amount of residual reactive groups and a relatively high‐molecular mobility. The low‐crosslinked matrix contains high‐crosslinked regions. In contrast, the “highly crosslinked” epoxy system has little reactive groups and a lower molecular mobility. Here, low‐crosslinked regions are found in a high‐crosslinked matrix. Hydrothermal loading for both networks is performed in demineralized water at temperatures below their glass transition. The water plasticizes both kinds of networks which remain in the glassy state, however. As a consequence, in the low‐crosslinked epoxy, the increased molecular mobility promotes an ongoing curing reaction leading to the consumption of epoxy groups until an almost complete network has formed. As a new aging process, phase separation occurs in the highly crosslinked epoxy. The new phase is more mobile than the matrix because it has its own glass transition at a lower temperature. In addition, thermooxidative degradation is observed for both network states. Certainly, these chemical and structural changes in the epoxy networks should influence the performance of an adhesive joint, a coating, or a fiber‐reinforced composite. © 2003 Wiley Periodicals, Inc. J Appl Polym Sci 91: 369–377, 2004     

The Adhesion of Maleic Anhydride on Native Aluminum Oxide: An Approach by Infrared Spectroscopy and Quantum Mechanical Modeling

The bonding between maleic anhydride, a curing agent for epoxy resins, and natively oxidized aluminum is studied by a combination of infrared spectroscopy and quantum mechanical modeling. On native aluminum oxide, maleic anhydride is hydrolyzed to maleic acid in the early stages of adhesion. The acid prefers a bridged chelate bond of one of its acid groups to two Al atoms of the metal oxide. Furthermore, a monodentate bond between a carbonyl oxygen of an acid group and an Al atom is identified. The anhydride itself shows no tendency to adhere on the native aluminum oxide. Since maleic acid is less reactive in the polymerization than the anhydride, the results imply that the chemical curing reaction could be hampered in the vicinity of the Al substrate.     

Adaptive disturbance rejection in an optical trap

This article presents a method for characterizing the system dynamics of a trapped particle in real-time and designing a controller to minimize disturbances to the particle's position. Specifically, adaptive system identification is used to determine the trap characteristics and the actuator transfer function describing the mirror voltage to trap position path. Using an internal model control scheme combined with a filtered-x least-mean-square algorithm, adaptive control was used to create a controller that minimizes a frequency weighted mean-squared-error. The dynamics associated with multiple particle sizes and materials were experimentally determined under different power levels, each case resulting in different system dynamics and demonstrating positive control results. The adaptive system identification and the controller presented automate the process of system identification and control design, enabling the automation of optical trap controller design.     

Regulation of endothelial proliferation by the renin-angiotensin system in human umbilical vein endothelial cells

This study was performed in order to evaluate the role of angiotensin II in physiological angiogenesis. Human umbilical vein endothelial cells (HUVEC) were stained for angiotensin II type 1 receptor (AGTR1) immunocytochemically and for gene expression of renin-angiotensin system (RAS) components. The regulation of the angiogenesis-associated genes vascular endothelial growth factor (VEGF) and angiopoietins (ANGPT1 and ANGPT2) were studied using quantitative RT-PCR. Furthermore, we examined the effect of angiotensin II on the proliferation of HUVEC using Ki-67 as well as BrdU immunocytochemistry and investigated whether the administration of the AGTR1 blocker candesartan or the VEGF antagonist FLT1-Fc could suppress the observed angiotensin II-dependent proangiogenic effect. AGTR1 was expressed in HUVEC and the administration of angiotensin II significantly increased the gene expression of VEGF and decreased the gene expression of ANGPT1. Since the expression of ANGPT2 was not affected significantly the ratio of ANGPT1/ANGPT2 was decreased. In addition, a significantly increased endothelial cell proliferation was observed after stimulation with angiotensin II, which was suppressed by the simultaneous administration of candesartan or the VEGF antagonist FLT1-Fc. These results indicate the potential capacity of angiotensin II in influencing angiogenesis by the regulation of angiogenesis-associated genes via AGTR1. Since VEGF blockade opposed the effect of angiotensin II on cell proliferation, it is hypothesised that VEGF mediates the angiotensin II-dependent effect in concert with the changes in angiopoietin expression. This is the first report of the RAS on the regulation of angiogenesis-associated genes in physiology.