Elevated temperature aging of solder joints based on Sn-Ag-Cu: Effects on joint microstructure and shear strength

Abstract

The shear strength behavior and microstructural effects after aging for 100 h and 1,000 h at 150°C are reported for near-eutectic Sn-Ag-Cu (SAC) solder joints (joining to Cu) made from Sn-3.5Ag (wt.%) and a set of SAC alloys (including Co-and Fe-modified SAC alloys). All joints in the as-soldered and 100-h aged condition experienced shear failure in a ductile manner by either uniform shear of the solder matrix (in the strongest solders) or by a more localized shear of the solder matrix adjacent to the Cu 6 Sn 5 interfacial layer, consistent with other observations. After 1,000 h of aging, a level of embrittlement of the Cu 3 Sn/Cu interface can be detected in some solder joints made with all of the SAC alloys and with Sn-3.5Ag, which can lead to partial debonding during shear testing. However, only ductile failure was observed in all solder joints made from the Co-and Fe-modified SAC alloys after aging for 1,000 h. Thus, the strategy of modifying a strong (high Cu content) SAC solder alloy with a substitutional alloy addition for Cu seems to be effective for producing a solder joint that retains both strength and ductility for extended isothermal aging at high temperatures.

Evaluation of Thin Copper Wire and Lead-Free Solder Joint Strength by Pullout Tests and Wire Surface Observation

“…It was confirmed by SEM observation that the actual bonded area on the bottom surface of the copper wire is about 1, independent of the specimen. The value of the tensile interface strength is a little high in comparison with the other results (σf = 50 to 90 MPa) obtained for similar lead-free solders and copper [9,10], and that of the shear interface strength is lower than the previous result (τf = 40 MPa) [11]. The constraint of plastic deformation at the bottom surface of the copper wire [12,13], the size and shape of the specimens, and the data dispersion (which will be described below) may be the reason for these differences.…”

Microstructural and mechanical analysis on Cu–Sn intermetallic micro-joints under isothermal condition

“…Obviously, the entire IMCs interlayer can be divided into two sub-layers mainly composed of Cu 3 Sn, growing and emerging from the two opposite sides of copper base with small amount of remanent Cu 6 Sn 5 4 in the middle. It is worth noticing that a special layer contains some ultra-fine equiaxial Cu 3 Sn grains accompanied with micro-voids as seen in Fig.…”

“…Lifshitz-Slezov-Wahner (LSW) theory [25] was established to describe the homogenizing process based on the diffusive decomposition in the supersaturated solid solutions, known as the classic Ostwald ripening process through decreasing the interfacial energy between phases in a conservative system. Another Flux-driven ripening (FDR) [26] theory was also developed later to describe the Cu 6 ), which is likely to obey the LSW theory of ripening, but this phenomenon herein is more probably characterized by the FDR theory. Although no liquid tin phase is involved in the full IMCs micro-joints during phase transformation process discussed here, the reaction such as Cu 6 Sn 5 + 9Cu  5Cu 3 Sn is expected to take place throughout the dwelling period.…”

“…The formation and microstructural evolution of the solder joints at a relative large scale (e.g. >50μm) during the reflow and thermal aging process have been well studied [4][5][6]; however, further investigations are still needed to establish on the full IMCs interconnection without existence of any remnant solders, in particular, the discontinuity that are associated with the crystallite microstructure and mechanical integrity.…”

“…And also, the minimum thickness of initial solder layer needed to form a pore-free joint has also been discussed in Bosco's paper [13], where at least 10µm of initial Sn layer was suggested. However the estimation was primarily based on the growth of the Cu 6 Sn 5 grain without considering the growth of Cu 3 Sn, thereby unsuitable for the case where the Sn layer is even thinner than 10µm and the growth of Cu 3 Sn layer can be comparable against the Cu 6 Sn 5 layer.…”