SAC 305 evolved into the default Pb-free choice, but many groups have failed to run the alloy through the complete battery of tests.

Inexpensive tests regularly referred to and performed by electronics manufacturers might include the following. A pull test, for one, may be performed to ascertain at what force level a solder joint system will fail by pulling the component out of the joint, destroying the component without destroying the joint, or pulling the component, joint and pad off the board. It is a test to failure. The best outcome would be for the component, joint and pad to pull off the board as the entire joint assembly would then be stronger than the underlying board. The worst outcome would be for the component to pull out of the solder joint. No pass/fail criteria exist for this test, so one is left with the theory that the greater the force exerted to cause failure, the stronger the solder joint. Likewise, a drop test may be performed to determine whether joint failure could occur when a board is dropped from a predetermined height.

Thermal cycling with subsequent tests for shock and vibration are routinely performed on Sn63. Thermal cycling provides accelerated aging information useful in predicting joint reliability. Each solder manufacturer shows test results on their alloys. One thousand thermal cycles (accelerated aging) is an acceptable number of cycles, which loosely represents about 3.5 years of PCB service. The boards are then tested for shock and vibration, again to provide joint reliability data.

Testing of the intermetallic layer is beneficial, since an inadequate intermetallic thickness may indicate insufficient bond formation between the component lead and solder, and the solder and pad. An either very thin or excessively thick intermetallic layer connotes greater solder attachment failure potential. A desired inter-metallic thickness is 1 to 5 µm. Over time, the intermetallic layer may grow. Thermal cycling should show accelerated aging growth of the intermetallic thickness. Too great an intermetallic thickness is detrimental to joint reliability because the joint may become more brittle.

Shear testing may be performed to see at what force a joint component can be “sheared”, thus indicating joint strength. It is thought that the harder the solder alloy, the greater the shear force necessary to shear the component. Since lead is malleable and SnPb solder is softer than Pb-free solder, it stands to reason Pb-free solders would have higher shear strength than SnPb solders.Other studies have confirmed this.¹ However, an examination of IPC standards (IPC-A-610D, J-STD-001, etc.) failed to turn up any performance standards (or pass/fail criteria) for pull, drop or shear testing. The most that can be said about a shear test is that the greater the force required to shear the component, in theory at least, the better and stronger the solder joint.

Much discussion regarding solder voiding and joint reliability has recently occurred. Solder voiding tests might be performed. The sole IPC reference found regarding solder voiding is in IPC-A-610D and regards BGAs, for which a defect is classified as “more than 25% voiding in the ball x-ray image area.”² The IPC Solder Value Products Council recently addressed this issue, stating, “Nine separate methods of statistical analysis comparing cycles to failure looking at both voids greater than 25% of the interconnection area and total voids have been done. Absolutely no correlation between voids and failures under thermal cycling has been demonstrated … there is no evidence that the type of solder joint voiding observed in the SAC alloy solder joints has any significant impact on solder joint reliability.”³ Yet voiding remains an oft-discussed issue.

A red dye test may indicate the potential for joint failure, and may be performed to visibly show micro cracking in solder joint surfaces before and after thermal cycling. It is believed that micro cracks are the point source for additional cracking, which may lead to joint failure.

Temperature/humidity tests can be performed to determine whether an alloy is susceptible to tin whisker formation. Tin whisker formation and dendrytic growth may cause short circuits at an undetermined time after the PCB has been placed in use. iNEMI recently published recommendations to mitigate whisker formation potential. The organization stated that “unalloyed tin electroplating has a long history of whisker formation and growth that has resulted in reliability problems for various types of electronic equipment. … It is generally accepted that the driving force for whisker formation is compressive strength on the tin films.”⁴ Various solutions are promulgated by iNEMI to reduce the possibility of failure.

The authors contracted with an independent laboratory to conduct an exhaustive test comparison between SAC 305 and Cobalt995 (a cobalt enhanced binary alloy consisting of Sn99.5Cu0.5Co). As the industry standard, SAC 305 was the control. Thermal cycling (including subsequent shock and vibration), solder voiding quantification, and pull and shear testing were performed for both alloys. The intermetallic thickness and alloy diffusion into the copper was examined both before and after thermal cycling/aging for both alloys. In addition, the cobalt alloy was subjected to a temperature/humidity test.

Boards hot air leveled with the Sn99.5Cu0.5Co alloy were manufactured. Half the boards used SAC 305 no-clean solder paste for the SMT portion and the same alloy for through-hole. The remaining half used Sn99.5Cu0.5Co alloy in no-clean solder paste for SMT and bar solder for wave soldering. Both pastes were manufactured with the same flux chemistry to reduce potential variables. These boards were then sent to a recognized independent laboratory for testing.

Test Parameters

Thermal cycling. The IPC-9701A protocol was used. Boards of SAC 305 and the enhanced cobalt alloy were subjected to a thermal range of -10° to +110°C. The number of thermal cycles was 1,000. The temperature rate of change was 10° to 20°C/min. The soak time at each temperature extreme was 5 min.

Shock test. The IPC-9701A protocol was used. The shock amplitude and duration was 1500G, 0.5mSec. The number of shocks was five. The direction of shocks was normal to the surface of the printed wiring board (z-axis).

Vibration test. IPC-9701A protocol. The frequency range was 20 to 20,000Hz. The vibration amplitude and duration was 15Grms for one hour and 20Grms for one hour. The direction of vibration was all axis simultaneously +3 rotational displacements (6° of freedom).

Shear test. A crosshead speed of 0.1 mm/min. with various target components using an Instron 3343 tester. A section of board was placed on a shear test fixture with an immovable edge to hold the target component in place. The Instron crosshead was lowered at a rate of 0.1 mm/min. until the device sheared from the board. Three resistors and five capacitors were sheared from each assembly (SAC 305 and enhanced cobalt).

Intermetallic examination and x-ray mapping of the intermetallic. Measurements and photographs were taken of the intermetallic thickness before and after thermal cycling of the cobalt alloy.

Pull test. A metal threaded stud was fixed to the top of the IC package and placed in a fixture to hold down the board edges while tensile force was exerted. The force was directly normal to the surface of the board to provide equally distributed force. An Instron 3343 tensile tester pulled the stud at a rate of 1 mm/min. until fracture. The maximum force was recorded for both the cobalt-enhanced and SAC 305 alloys.

Temperature/humidity parameters. The Jedec JESD22-A104-B protocol was used.

An assembled board soldered with the enhanced cobalt alloy was subjected to temperature and humidity for 500 hr. at temperatures of 85° +/-2°C. Relative humidity was 85% +/-5%. The total test duration was 500 hr. No bias was applied to the board. No monitoring took place until the completion of 500 hr.

Test Results

Thermal cycling. Both the SAC 305 and cobalt alloys exhibited no deterioration in solder attachment integrity after 1,000 thermal cycles (Figure 1). Minor voiding was observed in the solder attachments and TSOP devices for the SAC 305 as well as the cobalt alloy. Voiding was estimated at 2 to 12%, well within the range of IPC-A-610D (25% max.) and much below the IPC SPVC’s recommendation.


Thermal shock. For the SAC 305, minor voiding was observed on tested J-leads and no voiding of TSOP devices (Figure 2). For the cobalt alloy, the TSOP devices exhibited minor voiding, while no voiding was observed on the J-leads. The cobalt-enhanced alloy exhibits a much finer and more uniform grain structure than that of the SAC 305. As indicated on the left in Figure 2, the dark areas are silver-enriched pockets of the SAC alloy.


Vibration. For the SAC 305 alloy, minor voiding and minor cracks were observed on both the J-leads and TSOP devices (Figure 3). The cobalt-enhanced alloy performed the same as SAC 305. The cobalt-enhanced alloy’s grain structure is much finer and tighter than that of SAC 305, providing a significantly brighter solder joint.


Shear. Table 1 shows the cobalt-enhanced alloy outperformed SAC 305 on resistors while SAC 305 outperformed the cobalt-enhanced alloy on capacitors. Averaging all results, the cobalt-enhanced alloy slightly outperformed SAC 305. The smaller deviation of the cobalt-enhanced alloy suggests a greater ability to maintain consistency in all soldering processes.


Intermetallic examination. The cobalt-enhanced alloy showed an average intermetallic layer of 2.18 µm for the unstressed sample, while the average thickness after 1,000 thermal cycles was 2.32 µm (Figure 4). This suggests little or no growth of the intermetallic layer, indicating long term solder joint reliability. Both unstressed and stressed samples were well within the 1 to 5 µm desired thickness. Per Figure 5, the SAC 305 alloy showed an average intermetallic layer of 2.00 µm for the unstressed sample, while the average thickness after 1,000 thermal cycles was 3.43 µm, still within the desired thickness range.


X-ray intermetallic mapping. Both alloys were subjected to x-ray mapping to show copper diffusion at the intermetallic layer level. Copper diffusion was limited strictly to the intermetallic layer for the cobalt enhanced alloy, while the SAC 305 exhibited more copper diffusion into the bulk solder (Figure 6). The greater the copper diffusion beyond the intermetallic layer, the more brittle the solder joint can become.


Pull test. The SAC 305 alloy slightly outperformed the cobalt-enhanced alloy on the three devices pulled with 352.5N for the SAC alloy and 316.5N for the cobalt-enhanced alloy (Table 2). However, the SAC alloy exhibited a larger deviation in the results (41.6N for the SAC alloy and 11.2N for the cobalt-enhanced alloy), which suggests the cobalt-enhanced alloy shows a greater ability to maintain consistency in the soldering process. Since no figures exist for pass/fail criteria, and the test is to failure, it is not known if the forces are indicative of real-world use.


Temperature/humidity. After the parameters discussed above were completed, the cobalt-enhanced solder attachments were examined for defects precipitated by the test. No structural defects on the board were observed at 30x magnification. No tin whiskers were observed (Figure 7).


Conclusions

Although much industry testing has yet to be accomplished and evaluated, independent testing has shown an enhanced SnCu Pb-free alloy can provide benefits equal to, and in some cases better than, those of SAC alloys. In addition, the cobalt-enhanced alloy provides a much brighter and shinier solder joint with a tighter grain structure when compared to SAC 305 joints; in fact, the solder joints are virtually indistinguishable from Sn63 joints. When used in through-hole applications, a binary alloy (such as cobalt-enhanced) is much easier to maintain operating specifications when compared to a tertiary alloy (such as SAC 305); thus, less solder pot maintenance becomes necessary.

References

1. Thomas Siewert, Stephen Liu, David R. Smith, Juan Carlos Madeni, “Properties of Lead-Free Solders Release 4.0,” p. 21, Feb. 11, 2002.
2. IPC, IPC-A-610D, "Acceptability of Electronic Assemblies,” p. 8-83, February 2005.
3. IPC Solder Products Value Council, “The Effect of Voiding in Solder Interconnections formed from Lead Free Solder Pastes with Alloys of Tin, Silver and Copper,” p. 11.
4. iNEMI, “iNEMI Recommendations on Lead-Free Finishes for Components Used in High-Reliability Products,” version 4, pp. 3-4, Dec. 1, 2006.

Howard Stevens is vice president sales and marketing at Metallic Resources Inc. (metallicresources.com); hstevens@metallicresources.com.