Microelectronic Metallurgical Bonding under Extreme Temperatures

Abstract

The objective of this research project is to further refine and explore existing and possibly new solid-liquid interdiffusion (SLID) systems. The focus of the project was placed on SLID systems which can withstand high temperatures and SLID systems which can be bonded at low temperatures. With this focus considered, the intermetallic systems selected and studied were Ag-(In-Bi) for low temperature bonding and Ni-Sn for high temperature stability. The system Ag-(In-Bi) provided a unique possibility due to limited prior research [1]. With the potential to provide bonding at temperatures as low as 72 ℃, the system addresses the need to join piezo-electric materials (PEM) at temperatures lower than the Curie temperature, eliminating the need for costly magnetic re-poling and improving acoustic coupling [2]. The contributions to the knowledge of metal bonding from the research involving the Ag- (In-Bi) SLID bonding are the following: 1. SLID bonding of Ag-(In-Bi) i) The first micrographs and description of the Ag-(In-Bi) SLID system. ii) Demonstrated bonding at 150 ℃ with an In-Bi foil of 78.5at% In, where the bonds can be theoretically stable to ⁓480–570°C. iii) Discovery of Bi precipitation below the original bonding surfaces. iv) Statistical evidence showing gravity does not affect the precipitation distribution. The Bi precipitations required further exploration to determine their cause, and this led to the publication for solid-state bonding of Ag-(In-Bi). This had the additional advantage of demonstrating even lower bonding temperatures. The scientific contributions are the following: 2. Solid-state bonding of Ag-(In-Bi) i) The first micrographs and description of solid-state bonding with Ag-(In-Bi) ii) Demonstrated low temperature bonding at 65 ℃ with an In-Bi foil of 78.5-at% In and 95-at% In, where the bonds can be theoretically stable to ⁓166 °C. iii) Showed dissolution is required for the unique Bi precipitation which occurred in the Ag-(In-Bi) SLID system. The interest in high temperature stability is due to new developments in sapphire-based systems which can perform at temperatures above 1000 ℃. The Ni-Sn intermetallic system has been demonstrated by many works to provide significant temperature stability. The conference paper “Ni–Sn SLID bonds for assembly at extremely high temperatures“ attempted to create a homogenized bond of only Ni3Sn2 to provide the high temperature stability. The scientific contributions are the following: 3. SLID bonding of Ni–Sn i) Demonstrated bonding at 285 °C, where the bonds can be theoretically stable to ⁓798 °C. Further annealing of these bonds achieved various intermetallics of even higher temperature stability. ii) Based on finite element modelling, Pt was proposed as a candidate to better match the CTE values between Ni and Si compared to the W layer which was used in experimentation. Ultimately the main goal was not achieved in the Ni-Sn bonding due to small bonding times and temperatures. However significant voiding was seen to occur at the middle of the bond lines where impingement due to inter-metallic growth occurred. The cause and reduction of voiding is always a concern for SLID bonding and the explanation of these particular voids prompted the third main publication “Solid-liquid interdiffusion stresses leading to voiding“, which examined the pressure development within the encapsulated liquid Sn pockets. Within this paper the following was achieved: 4. SLID stresses leading to voiding iii) Performed a theoretical study on volumetric voiding in the Ni-Sn system showing the theoretical pressure development within constrained liquid pockets. This showed sufficient pressure to initiate cavitation. iv) Performed a molar volume balance of the Ni-Sn SLID system to determine the volumetric changes. A difference between this value and the thermodynamically calculated value was found. v) If unwettable inclusions are absent cavitation and growth expansion should occur, otherwise the proposed voiding model is some combination of expansion of bifilms and growth due to vacancies.