Effects of Bi Contamination on Sn/Pb Eutectic Solder

Estimated reading time: 7 minutes

By Dr. Jennie S. Hwang and Dr. Zhenfeng Guo

Figure 1. How 62Sn/36Pb/2Bi, or alloy 1, was alloyed.

Within the compositional range explored in this article, the Sn/Pb/Bi ternary solder alloys that result from doping Sn/Pb eutectic with Bi have greater strength and a higher fatigue life than 63Sn/37Pb eutectic solder. Melting temperature decreased with the addition of Bi from 183°C to a range of 175° to 182°C.Aternary phase diagram of Sn/Pb/Bi1 indicates the presence of a ternary eutectic composition of 16Sn/32Pb/52Bi with a melting temperature of 96°C. Additionally, two low-temperature binary eutectic reactions 43Sn/57Bi at 139°C and 43.5Pb/56.5Bi at 125°C extend into the ternary domains. Thus, there is concern for the effects of Sn/Pb coating on Bi-containing lead-free solder joints and the impact of Bi on 63Sn/37Pb joints. This article addresses Bi impact on 63Sn/37Pb solder joints.Because the available ternary phase diagram only manifests the liquidus surface, the Bi threshold content required to induce such low-temperature phases is not known. Moreover, the effects of Bi contamination on the mechanical behavior of 63Sn/37Pb solder joints are not established. Therefore, it is worthwhile to examine how adding up to 5 weight percent Bi to 63Sn/37Pb affects basic mechanical properties and melting temperature.

Experimental

Solder compositions were prepared by melting, in sequence, master solder alloys and pure metals in a Pyrex glass beaker in an electric resistance chamber furnace.* The master solder alloys of 60Sn/40Bi, 42Sn/58Bi and Sn in a bar form were used. Alloying procedures were as follows:

Figure 2. Tensile stress vs. strain curves at a strain rate of 6.56 x 10-4 per second and 300?K for alloy 1 (62Sn/39Pb/2Bi), alloy 2 (61Sn/37Pb/2Bi), alloy 3 (63Sn/35Pb/2Bi), and 63Sn/37Pb.

1. The furnace was heated and controlled to a temperature of 500°C.2. Master solder alloys with the highest melting points were melted initially for 10 minutes.M3. Pure metals and solder alloys with lower melting points were then dissolved sequentially. The resultant mixtures were thoroughly stirred.4. The resulting alloyed mixtures were homogenized for 20 minutes before cast.The procedures for alloying 62Sn/36Pb/2Bi (alloy 1), for example, are detailed in Figure 1.A stainless steel cast mold was configured to make mechanical specimens for testing. The standard tension specimens were tested in accordance with the ASTM Standard E8-96 (Standard Test Methods for Tension Testing of Metallic Materials), ASTM Standard E139-95 (Standard Practice for Conducting Creep, Creep-rapture and Stress-rupture Tests of Metallic Materials), and ASTM Standard E606-92 (Standard Practice for Strain-controlled Fatigue Testing).

Figure 3. Tensile stress vs. strain curves at a strain rate of 6.56 x 10-4 per second and 300°K for alloy 4 (60.5Sn/34.5Pb/5Bi), alloy 5 (58Sn/37Pb/5Bi), alloy 6 (63Sn/32Pb/5Bi), and 63Sn/37Pb.

Standard tensile and low-cycle fatigue tests were run in uniaxial tension at room temperature (300°K) using an electromechanical material testing system.** Load was measured using a 5 kN capacity static load cell, strain was measured using a 1"gauge length extensometer, and test data were acquired by operating software.*** Specifically, tensile tests were performed at a cross-head speed of 1 mm per minute, which gave a strain rate of 6.56 x 10-4 per second. Basic mechanical properties were determined in accordance with ASTM Standard E8-96, including yield strength (sy) at 0.2 percent offset plastic strain, ultimate tensile strength (sTS) at maximum load, and plastic strain (ep) at fracture after a drop in load to 10 percent of the maximum values. Low-cycle fatigue tests were carried out under strain control at a frequency of 0.1 Hz and an R-ratio of 0.8. The fatigue life (Nf) was defined as the number of cycles at which the tensile load on the hysteresis loop of load vs. elongation was dropped to a value of 50 percent of the maximum load, according to ASTM Standard E606-92.

The lead-free alloy melting temperatures were measured through cooling curves produced by a thermocouple technique.

Results

Two and 5 percent Bi dosages were added to 63Sn/37Pb by separately replacing Sn or Pb. Tests also were performed on alloys with Bi replacing an equal amount of both Sn and Pb. This resulted in six solder alloy compositions. The compositions, along with their melting temperatures (Tm), sy, sTS, Young's modulus (E), ep at fracture and Nf at a total strain of 0.2 percent, are summarized in Table 1. All compositions are expressed in weight percent unless otherwise specified. Also included in Table 1 is the 63Sn/37Pb reference solder alloy.

Figure 4. Bi effects on sy, sTS and ep for the Sn/Pb/Bi system at 58 to 63 percent Sn and 32 to 37 percent Pb.

The addition of 2 percent Bi depressed the original melting temperature of 63/37 by 2° to 3°C. There was almost no distinction in the melting temperature change when 2 percent Bi replaced Sn, Pb, or both Sn and Pb in an equal amount. At 5 percent Bi, both the alloy liquidus melting temperatures and the alloy solidus melting temperatures were lowered. The melting temperature for the solder alloys with 5 percent Bi in place of Sn (alloy 5) was about 2° to 3°C lower than that with 5 percent Bi in place of Pb (alloy 6). This indicates that 5 percent Bi lowers the melting temperature of the Pb-rich phase more effectively than the Sn-rich phase. Based on the cooling or heating curves, the formation of low-temperature eutectic phases was not detected.

Figure 2 compares the tensile stress (s) vs. strain (e) curves of the solder alloys containing 2 percent Bi with 63Sn/37Pb. The tensile flow of solder alloys typically consists of an elastic region, a strain hardening region, a stress-recovery region and a cracking region. Strain hardening continued until necking occurred at the maximum load or sTS. Necking was caused by an inhomogeneous plastic deformation somewhere in the gauge length, associated with strain localization. Stress-recovery mechanisms are believed to be dominant in the region after necking and before abrupt fracture for high-temperature plastic deformation. The Bi addition largely increased the alloy strength and plasticity. There were no measurable differences in the tensile behavior among the solder compositions containing 2 percent Bi in place of Sn (alloy 2) or Pb (alloy 3), or both Sn and Pb equally (alloy 1).

Figure 5. Bi effects on Nf for the Sn/Pb/Bi system at 58 to 63 percent Sn and 32 to 37 percent Pb.

When the Bi content in 63/37 increased to 5 percent, strength was maintained, but the alloy plasticity was reduced (Figure 3). The differences in tensile behavior among the solder compositions containing 5 percent Bi in place of Sn (alloy 5) or Pb (alloy 6), or both Sn and Pb equally (alloy 4), were within the data-scattering range.Figure 4 plots sy,sTS and ep at fracture against Bi contents for the Sn/Pb/Bi system at 58 to 63 percent Sn and 32 to 37 percent Pb. Up to 2 percent Bi was the most effective amount to increase both alloy strength and plasticity. Any further increase exhibited little effect on alloy strength but significantly reduced alloy plasticity.

Nf increased with up to 5 percent Bi content (Figure 5). This is in contrast to the plasticity reduction at 5 percent Bi, which results because the fatigue strain range (De = 0.2 percent) for Nf is well below the plastic strain at fracture (ep = approximately 12 percent). Under the relatively low fatigue strain range, high strength is a leading factor contributing to high fatigue resistance.

Discussion

The Pb/Bi binary phase diagram2 indicates a solid solubility of 23.5 percent Bi in Pb, while the Sn/Bi binary phase diagram2 indicates a solid solubility of 21 percent Bi in Sn. When Sn/Pb/Bi constructs a ternary system, the underlying basic physical interactions are not expected to change grossly. It is recognized that low-temperature binary and ternary eutectics will not exist unless Bi content exceeds the primary solid solubility. Because the addition of up to 5 percent Bi to 63/37 is well below the primary solid solubility, the low-temperature eutectic phases are not expected to be formed in the alloys containing a low dosage of Bi. This is consistent with the present heating and cooling thermal measurement. The fact that strength and fatigue life increase with Bi content further supports the absence of the low-temperature phases. Bi atoms up to 5 percent are expected to enter the Sn crystal lattice, Pb crystal lattice or both.

Conclusion

The Sn/Pb/Bi ternary solder alloys containing 2 percent Bi demonstrated greater strength, a higher fatigue life and higher plasticity than 63Sn/37Pb eutectic solder, while melting temperature slightly decreased. With an increase to 5 percent Bi, the strength and fatigue life of Sn/Pb/Bi still remained higher than 63Sn/37Pb, but plasticity decreased significantly.

*Thermolyne 48000.** Instron 4411.*** Instron Merlin series IX.

REFERENCES

1. J.S. Hwang, Solder Paste in Electronics Packaging, Van Nostrand Reinhold, 1989, p. 399.
2. M. Hansen and K. Anderko, Constitution of Binary Alloys, McGraw-Hill Book Co. Inc., 1958.

DR. JENNIE S. HWANG and DR. ZHENFENG GUO may be contacted at H-Technologies Group Inc., 5325 Naiman Parkway, Cleveland, OH 44139; (440) 349-1968; Fax: (440) 349-9961; Web site: www.smtnet.com/h-tech.