A Study of Copper Dissolution in Pb-Free Solder Fountain Systems

ABSTRACT Implementation of a Pb-free assembly process implies that rework and selective solder attachment of through-hole components with Pb-free alloys may also be required. The processes are physically similar to wave soldering, but with longer contact times. Exposure to radiant and convective heat over the solder pot is often the only preheat that the PWA receives, so longer contact times are required compared to wave soldering. The Pb-free alloy in common use for wave soldering, Sn96.5Ag3.0Cu05, (SAC 305), was expected to dissolve copper from the board pads, traces, and component leads at a higher rate than traditional SnPb alloys. The rate at which Cu is dissolved could be dependent upon a number of factors, including solder alloy, solder and board temperature, solder flow rate, copper alloy, copper grain structure, and the physical flow characteristics of the solder over the board (laminar, turbulent). Previously published information indicated the dissolution rates for high-Sn alloys were greater than for SnPb (1,2). This predicted behavior was observed in both soldering and HASL (Hot Air Solder Leveling) operations (3, 4). As a result of the higher expected dissolution rate, the allowable immersion time might not be long enough to allow rework when using high-Sn alloys. Current immersion time requirements range from less than 20 seconds for initial soldering of simple components on low-mass boards to approx. 60 seconds for rework of some connectors on highmass boards. An investigation was conducted in two phases to study the effects of contact time on dissolution of copper. In the first phase, copper wires of various diameters were used as an “analog” of a circuit board trace. Candidate alloys were compared as to the time required to dissolve the wire. The test was conducted using a production mini-wave, to ensure that the flow dynamics of the solder were what a PWA would experience in the actual process. In the second phase, the two alloys with the lowest dissolution rates in Phase 1 were then compared using actual plated traces to verify that the wire dissolution data could be extrapolated to actual PWAs. Sn63Pb37 (SN63) alloy was also tested as a comparison baseline. The findings from this investigation indicate that even small alloying additions to high-Sn solders can drastically alter the dissolution rate. Both the doped SnCu and 4-part alloy showed the lowest dissolution rate of the alloys tested. The dissolution rates were nearly four times lower than for SAC305. Both of these alloys are commercially available from several sources, and both are less expensive than SAC 305. The doped SnCu, however, has more limited reliability data available compared to the 4-part alloy. The 4-part alloy is also a member of the SAC family, benefiting form the lower melting temperature (217°C vs. 227°C) and cross-compatibility with other SAC-based soldering processes (intermixing of SAC alloys has not been shown to be a problem). Even with the reduced dissolution rates, the immersion times required would still resulted in significant reduction in copper on exposed traces and pads. In order to minimize the erosion of copper from the exposed traces and lands, it will be necessary to reduce exposure time through proper process design. Board design can also be optimized to minimize the impact of copper dissolution by incorporating widened traces in exposed areas near lands (pad tear-dropping).