The SnCuNi alloy (SCN) has been used in wave soldering applications because of its applicability in achieving acceptable soldering results and its lower rates of reaction with copper (used in PCBs and components) and iron (used as a based material in wave soldering equipment). The lack of precious metals like silver makes it less expensive, and the cosmetic appearance of the final joints is similar to that of SnPb joints; but these benefits are offset by its higher melting point temperature of 227°C in reflow applications. The higher melting point may necessitate reflow profiles with higher peak temperatures and/or longer time above liquidus (TAL) than that of SAC 305 to obtain complete and homogenous mixing of the paste deposits with the component lead/bump. Concerns such as possible damage of heat-sensitive components and joint reliability arise.

A study was carried out to develop reflow processes for SCN solder paste using SAC 305 and SCN-bumped BGA-CSP components. Assembly characterization was performed using cross-sectional analysis, vibration testing and thermal cycling. The objective was to characterize the performance of pure SCN joints and compare them with pure SAC 305 solder joints and mixed SCN paste/SAC 305 sphere solder joints. This was accomplished by designing reflow soldering profiles that reached the same peak temperatures and TAL (above 217°C) optimized for typical SAC 305 assemblies.

The test vehicle was a 0.062" thick, four-layer FR-4 PWB with Cu-OSP surface finish and non-solder mask-defined pads. Each board was populated with 16 256 I/O BGA-CSP components. The design of experiments included different peak temperatures (238°C and 248°C) and TAL (50 and 75 sec.). The corresponded TAL above 227°C was 30 and 50 sec., respectively. Levels for each factor were based on the current SAC 305 process window and recommendations from the SCN solder paste supplier. All boards were reflowed in air, and a small batch of pure SCN joints were reflowed in nitrogen (<100 ppm).

An x-ray automatic program was used to inspect voiding. All solder joints from 12 boards with 16 components each were tested. The percentage of the single largest and overall voiding per solder ball was recorded. Results showed pure SCN shows fewer and smaller voids than the other two metal systems. However, the overall sizes of voids were insignificant. The overall sizes were on average less than 3.5% for the SAC and mixed systems and less than 1.2% for SCN systems. The size of these voids passed IPC-A-610D, which sets acceptance criteria for Class 1, 2 and 3 at a maximum 25% of the ball x-ray image area.

Good solder joint formation and collapse was observed on SCN and mixed assemblies when reflowed at 238°C and TAL of 75 sec. Microstructure analysis showed the main difference between these two systems was the presence of Ag3Sn intermetallic in the mixed joints. Another difference was the thickness of the intermetallic between the PCB and joint. A 60% thicker intermetallic was observed in the mixed system. This might be an indication that the nickel content inhibits the growth of CuSn intermetallic.¹

Vibration testing was performed on 12 boards reflowed in air. The goal was to excite the first resonance (bending movement) at relatively low amplitude to induce high cycle fatigue failures. The boards were mounted with standoffs at the four corners to an electrodynamic shaker. Failure data were divided into four groups depending on component location on the board because they experienced different stress levels. Figure 1 shows a schematic of the board with its groups. Two failure modes were observed: pad cratering (groups 1 and 2) and solder fatigue (groups 3 and 4).

Results showed SAC, SCN, and mixed assemblies performed similarly in each group. The lowest cycles to fail were observed in groups 1 and 2, followed by groups 3 and 4. Further testing is planned to compare alloys in drop testing where the strain rate and stress levels are much higher.

Thermal cycling was performed with temperature ranges from 0° to 100°C, with a dwell time of 10 minutes and ramp rate of 10°C/s. The test was stopped at 1,686 cycles, when more than 50% failures were observed for each board. Table 1 shows the characteristic life (N63) and early failures (N01) for each case. It can be observed that SAC systems had on average better characteristic life followed by SCN (150 fewer cycles) and mixed (191 fewer cycles) systems. The data favor 238°C peak temperature, and there was not a significant difference between TAL.

Early failures, which correspond to 1% of the failure data, showed a different trend. In this case, SAC systems had higher numbers of cycles-to-failed, followed by mixed (90 fewer cycles) and SCN (204 fewer cycles) systems. An improvement in early failures was observed when nitrogen was used in SCN joints, resulting in similar behavior to SAC systems.

SCN shows promise as a replacement for SAC alloys for some reflow applications. SCN shows comparable performance in mechanical and thermal testing to SAC systems. Thermal cycling results suggest that the appropriate process window for the SCN system should have a peak temperature of 238°C and TAL of 50 sec. Thus, a typical SAC profile can be used to assemble pure SCN. At this temperature, heat-sensitive material suitable for Pb-free applications can be used without any problem. The use of a single alloy in wave and reflow processes will benefit the end-user by reducing complexity and cost.

Mixed assemblies, which are mainly SAC alloy (ratio sphere/paste = 3.18), were affected by the content of silver and nickel, which results in a decrease in the characteristic life when compared to pure SAC, but were comparable to SCN assemblies. In general, all three systems performed similarly within an appropriate process window, but more experiments are needed to support this conclusion.

Reference

1. F. Song, J. Lo, J. Lam, T. Jiang and S.W.R. Lee, “A Comprehensive Parallel Study on the Board Level Reliability of SAC, SACX, and SCN Solders,” Electronic Components and Technology Conference, May 2008.

Ursula Marquez de Tino is a process and research engineer at Vitronics Soltec, based in the Unovis SMT Lab (vitronics-soltec.com); umarquez@vsww.com.