PRODUCTION TESTING OF Ni-MODIFIED SnCu SOLDER PASTE

Size: px
Start display at page:

Download "PRODUCTION TESTING OF Ni-MODIFIED SnCu SOLDER PASTE"

Transcription

1 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 PRODUCTION TESTING OF Ni-MODIFIED SnCu SOLDER PASTE Karl Seelig, Timothy O Neill, Kevin Pigeon & Mehran Maaleckian AIM Solder Cranston, RI, USA kseelig@aimsolder.com Andy Monson & Walter Machado Hayward Industries, HRI (formerly Goldline Controls) North Kingstown, RI, USA Chrys Shea Shea Engineering Services Burlington, NJ, USA ABSTRACT Low-silver and no-silver lead-free PCB solder alloys represent substantial cost savings over SAC alloys that contain 3% silver. Many wave solder operations use silverfree alloys for through hole soldering, and SMT operations are also beginning to adopt the alloy for surface mount soldering. This paper reviews two case studies that test Ni-modified SnCu solder paste on mixed technology circuit assemblies which currently use SAC305 in production. The major differentiator between the two cases is the reflow profile: In one case, the SAC profile was considered nearly perfect for the SnCuNi alloy, peaking at 245 C. In the other case, thermally sensitive components constrained the profiles to peaking near 235 C, only 8 degrees above the silver-free alloy s 227 C melting temperature. This important case is used to explore the lower limit of reflow capability and compare it against the baseline provided by the typical thermal process. The performance of the solder pastes and joint formation in both cases are evaluated using numerous assembly metrics and reliability tests, including visual and automatic optical inspection, X-ray inspection, in-circuit and functional testing, optical microscopy, thermal aging, thermal cycling and shear testing. The results are presented and discussed. Key words: Lead-free solder paste, silver-free solder paste, Sn100C, Ni-modified SnCu, It is not as commonly used as an SMT alloy, however, due in part to its slightly higher melting temperature in comparison to SAC305. The non-eutectic SAC305 alloy has a solidus temperature of 217 C, a 3-4 degree pasty range, and liquidus temperature between 220 and 221 C. Eutectic SnCuNi is completely liquid at 227 C, only 6-7 degrees higher than SAC 305. Recommended peak temperatures in reflow processes are usually degrees higher than the solder alloy s melting temperature. Peaking reflow temperatures within 10 degrees of a solder s liquidus point can prompt concerns about solder joint quality and integrity. Wetting speeds and intermetallic formation rates are slower at lower temperatures, introducing the potential to create cold, poorly wetted, or otherwise inferior and perhaps unreliable solder joints. The capability of the SnCuNi solder alloy and flux to form quality interconnections on the low end of the reflow process window is tested by running it in a production process and comparing it with SAC305 in the same process, and with both alloys reflowed in a hotter process. PCB ASSEMBLY Test Vehicles The test vehicles used in these experiments are shown in figures 1a through 2. PCBs 1a and 2 are both industrial controller boards, measuring approximately 6x6 and 8x8 respectively. PCB 1b is a display board that accompanies PCB 1a, and measures approximately 3x3.5. INTRODUCTION Nickel-modified SnCu, also known as SN100C, is a commonly used lead-free electronic solder alloy. Its predominant application is in wave soldering, because it offers numerous advantages over SAC305, including much lower copper dissolution rates 1, smoother and shinier solder joint surfaces without shrinkage cracks, and considerably lower cost due to the elimination of the silver content.

2 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 PCB 2 was printed using Type 3 AIM NC254 no-clean solder paste for both alloys with a laser-cut, electropolished stencil on an MPM UP2000 printer. Components were placed with a Mydata MY12 and the boards were reflowed with a 10-zone Electrovert Omniflo oven. 50 of each assembly were run on the production line with both SAC305 and SnCuNi solder pastes. They were all fully assembled with PTH components and wave soldered with SAC305 or SnCuNi, corresponding to their SMT solder alloy. Figure 1a. Controller board used in experimental production run (Case 1a/PCB 1a). Reflow Profiles PCBs 1a and 1b ran relatively fast, cool, ramp-style reflow profiles, reaching peak temperatures of C in about 3 minutes 30 seconds, with seconds above 217 C (figure 3). In contrast, PCB #2 ran a longer, hotter, soakstyle profile that peaked at 245 C in 4 minutes 45 seconds (figure 4). Figure 1b. Display board used in experimental production run (Case 1b/PCB 1b). Figure 3. Reflow profile for PCB 1a. Figure 2. Controller board used in baseline production run. (Case 2/PCB 2). Assembly Processes PCBs 1a and 1b were printed using Type 4 AIM NC259 solder paste for both alloys with electroformed stencils for both assemblies on an EKRA X4 printer. Components were placed with ASM SIPLACE D4 and reflowed in a 7-zone BTU Pyramax oven. 30 of each assembly were run on the production line with both SAC305 and SnCuNi solder pastes. Five of each set were assembled with PTH components and wave soldered using SnCuNi. Figure 4. Reflow profile for PCB 2.

3 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 Visual inspection of PCB 1b on its original reflow profile indicated minimal wetting to some of the SOIC leads and a few unmelted solder spheres near areas of higher thermal mass. Maintaining considerations for the thermally sensitive component(s), the profile was modified slightly by slowing the conveyor and increasing one zone temperature. These changes raised the peak temperature to 237 C with an approximate time above 227 C of 60 seconds and demonstrated improved wetting upon visual inspection. The original and modified profiles for PCB 1b are shown in figures 5 and 6. found. 5 of each set of PCBs (SAC305 and SnCuNi) were fully built up with PTH components, wave soldered (all with SnCuNi) and assembled into chassis for functional test. All passed. Assembly 2 All 100 PCBAs were visually inspected, in-circuit and functionally tested. Defect rates were similar with the exception of tombstones. This PCB design has a propensity for tombstoning in its usual production process. In this build, the SAC305 assemblies averaged 3 tombstones per board, whereas the SnCuNi assemblies had no tombstones at all. 2 ADDITIONAL AOI TESTING Samples of PCBAs 1a and 1b were submitted to Mirtec for additional AOI compatibility analysis. Using the MV-3L AOI system s standard H-V light algorithm (figure 7) very little difference in the reflectivity of the joints was noted. Minor differences were noted when using non-standard horizontal or vertical lighting only (figures 8 and 9). Figure 5. Original reflow profile for PCB 1b Figure 7. AOI Image of SnCuNi solder joints Figure 6. Modified reflow profile for PCB 1b INSPECTION & TEST Assemblies 1a and 1b All 120 PCBAs were inspected with a Mirtec MV-7 Automated Optical Inspection immediately following reflow. No defects were identified. They were also 100% visually inspected by certified inspectors, with no defects Figure 8. Image of Sn100C solder joints using vertical lighting only

4 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 Figure 9. AOI Image of SAC305 solder joints using vertical lighting only. Using standard solder joint identification and inspection algorithms, the AOI systems processed both alloys without requiring any changes to the standard inspection parameters. Figure 11. SnCuNi solder joints formed in Case 1. Typical SnCuNi SMT Solder Joints SOLDER JOINT ANALYSIS Visual Appearance In addition to the professional visual inspection on the assembly line, samples from Case 1 were again inspected photodocumented. Figure 10 shows examples of the SAC305 SMT solder joints and figure 11 shows examples of the same joints formed with the Ni-modified SnCu. While the SnCuNi joints are shinier and smoother than the SAC305 joints, they do not exhibit the typical luster associated with this alloy (figure 12), presumably due to their low thermal exposure during the liquidus phase of the reflow process. Despite their relatively lackluster appearance, they exhibit acceptable wetting throughout the PCB assembly, including areas of higher thermal mass. Figure 12. joints Typical appearance of SnCuNi SMT solder X-Ray Both SMT and PTH solder joints were inspected by manually operated transmission X-ray. Case 1a was chosen for SMT because the fastest, coolest reflow profile offers the greatest opportunity for voiding. Case 2 was chosen for PTH to allow a performance comparison of the two alloys in wave soldering. Representative X-rays are shown in figures Figure 10. SAC305 solder joints formed in Case 1.

5 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 SnCuNi SAC305 Figure 13. X-ray images of 0805 resistors from Case 1a. SnCuNi SnCuNi SAC305 Figure 15. Representative X-rays of both alloys in PTH soldering from Case 2. Both solder alloys demonstrated some degree of voiding in the SMT solder joints, but well below the 25% guideline for acceptability. The Ni-modified SnCu alloy showed slightly more voiding in the SMT solder joints of Case 1; the SAC305 showed slightly more voiding in the PTH solder joints of Case 2. It is believed that the level of SMT voiding seen in Case 1 can be reduced by reflow profile modifications that add some soak time in the ramp phase to permit outgassing of volatiles without increasing the ramp rate, peak temperature or TAL dictated by the boards temperature-sensitive components. The level of voiding in Case 2 s PTH components is not addressed because wave soldering of can have varied and multiple causes of voiding, including materials and processes used in PCB fabrication, component manufacturing or PCB assembly. Microstructural Analysis Samples from Case 1a, which is the fastest, coolest profile that offers the least opportunity for wetting and intermetallic compound (IMC) formation, were cross-sectioned and examined via optical microscopy. Figures 16 and 17 show good solder wetting to gull-wing and end-terminated components for both SnCuNi and SAC305. SnCuNi SAC305 Figure 14. X-ray images of SOIC solder joints from Case 1a. SOIC 0805 Figure 16. SnCuNi solder joint wetting

6 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 SAC305 SnCuNi SOIC 0805 Figure 17. SAC305 solder joint wetting The SnCuNi joint shown in figure 18 has a continuous, uniform IMC approximately 3um thick. The IMC has a smoother, more nodular appearance due to the nickel that inhibits intermetallic growth. The SAC305 joint also shown in figure 18 also has a continuous IMC approximately 3um thick, but with a more dendritic morphology, possibly complicated by a Ag 3 Sn IMC formation. The SnCuNi bulk solder showed a more equiaxed grain structure, whereas the SAC305 showed a columnar grain structure. SnCuNi SAC305 Figure 18. Microstrucutres and IMCs of both alloys in Case 1. SAC305 Figure 19. Microstructure and IMC of both solder alloys in Case 1 after aging at 125 C for 96 hours. Thermal Cycling Samples from Case 2, the more complex and densely populated assembly, were thermally cycled from 0 to +100 C for one hour cycles. After each 1000 cycles, they were in-circuit and functionally tested. Grain structure and shear strength were also assessed. This process was repeated three times, for a total of 3000 thermal cycles. No solder-related defects were found on any of the assemblies during any of the electrical testing. No structural changes were observed in any of the cross sectioned solder joints. Shear Tests Samples from both cases were shear tested chip resistors were sheared at 15 angles on a Dage 4000 series shear tester. In Case 1, 16 of each component were tested before thermal aging and 8 of each after. No appreciable differences were noted in overall shear strength, which ranged from 4-6kg, as expected. SAC305 demonstrated slightly higher average shear strength, with slightly larger standard deviations than SnCuNi. The results are shown in figure 20. Thermal Aging The assemblies from Case 1 were aged at 125 C for 96 hours. They were again cross-sectioned and examined via optical microscopy. The grain structures and IMCs are shown in figure 19. Thermal aging had no significant effect on IMC growth. Figure 20. Shear strengths of 0805 resistors before and after thermal aging In Case 2, 10 components were tested after each of 1000, 2000 and 3000 thermal cycles. Most of the results were also in the anticipated 4-6kg range, with the exception of the SnCuNi at 1000 cycles, which averaged only 3.1 kg (figure

7 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, ). The root cause of the low values and small standard deviation in this dataset is unknown. The remaining tests at 2000 and 3000 cycles produced results in the expected ranges. PCB ASSEMBLY AND ANALYSIS SUMMARY The assembly and test process is summarized in table 1; the solder joint quality and reliability analysis is summarized in table 2. Figure 21. Shear strengths of 0805 resistors after thermal cycling. Table 1. PCB Assembly and Test Summary Process Case 1 Case 2 30 each of controller and display PCBs 50 each with SnCuNi and SAC305 Build Quantity with SnCuNi and SAC total 100 total Reflow Cool Moderate Wave solder 5 of each with SnCuNi only 50 each with SnCuNi and SAC total 100 total Visual Inspection No defects found Similar defect rates but more tombstones with SAC305 No defects found on production line Automatic Optical Inspection No program tweaks needed in laboratory analysis In-Circuit Test Pass Functional Test 5 of each pair pass Pass

8 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 Table 2. Solder Joint Quality and Reliability Analysis Characteristic Case 1 Case 2 Visual Appearance Voiding Microstructure Thermal Aging 125 C for 96 hrs Thermal Cycling C, 1 hr Shear Strength Slightly duller than typical SnCuNi surface finish Acceptable wetting SnCuNi showed slightly more voiding in SMT. Both within acceptable limits Uniform, continuous IMC approx. 3um thick in both alloys SMT joints. Equiaxed grain structure in bulk solder of SnCuNi; dendritic grain structure in bulk solder of SAC305. No effect on IMC or grain structure or either alloy Resistors average 4-6kg shear strength, no significant difference between alloys before or after thermal aging. Typical surface finish and wetting SAC305 showed slightly more voiding in PTH Both within acceptable limits No electrical test failures found after 1k, 2k & 3k cycles No structural changes observed in cross sections at 1k, 2k & 3k cycles 0805 Resistors average 4-6kg shear strength, only deviation noted was SnCuNi after 1k thermal cycles. DISCUSSION AND CONCLUSIONS Ni-modified SnCu solder alloy should be considered a viable replacement for SAC305 in nearly every SMT assembly process. The tests in Case 1 reflowed the SAC alternative in a profile that peaked at 235 C with only 60 seconds above 217 C a thermal excursion considered near the lower limit for SAC305 processing. This profile allowed only 50 seconds above the 227 C melting point of SnCuNi. Even in a reflow process previously considered too cool for the silver-free alloy, acceptable solder joints were formed. The visual appearance of the SnCuNi solder joints was a bit duller than typical SMT solder joints formed with this alloy at higher peak temperatures, but good wetting and IMC formation was confirmed by visual inspection and crosssectional optical microscopy. AOI algorithms recognized the SnCuNi solder joints using the same inspection parameters as the SAC305 joints. No program changes or tweaks were required to inspect the substituted solder alloy. The Ni-modified SnCu SMT solder joints processed on the cooler profile demonstrated higher levels of voiding than the SAC305 solder joints processed under the same profile. Both solder pastes used the same flux. Although the level of voiding is considered acceptable, minor reflow modifications can likely reduce the incidence of voids in both alloys. Solder joint strength was measured by shear testing 0805s. Their strength was in the normal range, and was not affected by thermal aging. Joint strength of the SnCuNi was comparable to the SAC305 formed under the same profile and also comparable to the joint strength of both alloys formed under the hotter, more traditional profile. Thermal cycling did not affect the electrical or mechanical integrity of the solder joints in Case 2, which factored into the decision to forego the thermal cycling tests for Case 1. The solder joints formed in Case 1 under the cooler profile all exhibited typical characteristics joint wetting, IMC formation and strength indicative of quality, reliable solder joint formation. Originally, a planned thermal cycling test for Case 1 included cross sectioning, shear testing and electrical testing at the end of the 3000 cycles. However, given the normal characteristics of the solder joints and the abundance of thermal cycling results available on these alloys 3, it was highly likely that no joints would fail and no advances in valuable knowledge would be gained by performing the tests. Therefore, they were omitted from the final analysis. Silver-free, lead-free electronic solder alloys offer substantial cost savings over silver-containing ones, reducing solder paste costs by as much as 20 percent. This potential savings is very attractive to many OEM and CEMs, particularly in price-competitive sectors such as consumer electronics, but concerns over compatibility with the established SAC305 reflow window have slowed its implementation. Case 1 provides crucial information about the Ni-modified SnCu alloy s reflow capability in the low end of the SAC reflow process. It demonstrates that the SN100C can form quality, reliable SMT solder joints even in thermal profiles on the lower margin of the SAC305

9 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 process window, and should be considered a drop in replacement for nearly any SMT assembly process that uses SAC or SAC-based solder. ACKNOWLEDGEMENTS The authors would to thank Mirtec s Brian D Amico and Robert Horowitz for their AOI analysis in support of this project. REFERNCES [1] C. Hamilton, P. Snugovsky, PhD, Celestica, M. Kelly, IBM A Study of Copper Dissolution during Pb-free PTH Rework using a Thermally Massive Test Vehicle, Proceedings of SMTA International, September 2006 [2] K. Seelig, Lead-Free Alloy Development, Proceedings of the International Conference on Soldering and Reliability, May, 2012 [3] J. Arnold, N. Blattau, and C. Hilllman, DfR Solutions, K. Sweatman, Nihon Superior, Reliability Testing of Ni- Modified SNCu and SAC305 Accelerated Thermal Cycling, Proceedings of SMTA International, October, 2008 Additional reference material includes internal reports from Shea Engineering Services titled Hayward Build Report April 30, 2013 and Hayward PCB Assembly Images Report issued April 17, 2013, available upon request from AIM Solder.

10 Originally published in the Proceedings of SMTA International, Ft. Worth, TX, October, 2013 PRODUCTION TESTING OF Ni-MODIFIED SnCu SOLDER PASTE Karl Seelig, Timothy O Neill, Kevin Pigeon & Mehran Maaleckian AIM Solder kseelig@aimsolder.com Andy Monson & Walter Machado Hayward Industries, HRI (formerly Goldline Controls) North Kingstown, RI, USA Chrys Shea Shea Engineering Services Burlington, NJ, USA

11 Outline/Agenda Introduction Case Study Descriptions Assembly & Test Solder Joint Analysis Summary Discussion and Conclusions Questions

12 Introduction Ni-Modified SnCu solder Silver-free, lead-free alloy Less expensive than SAC305 Long, successful history in lead-free wave soldering and HAL PCB final finish Limited acceptance in SMT reflow due to higher melting temperature than SAC305 Concerns of cold, poorly wetted, or otherwise unreliable solder joints when reflowed on the low end of the SAC305 process window

13 Alloy Comparison Property SnCuNi SAC305 Composition Sn-0.7Cu-0.05Ni+Ge Sn-3.5Ag-0.5Cu Liquidus Point 227 C 221 C Solidus Point 227 C 217 C Appearance Shiny Dull with primary Sn dendrites on surface Source: Nihon Superior Recommended peak reflow temperatures are typically higher than the solder alloy s liquidus temperature

14 Two Case Studies PCB assemblies with similar configurations Mix of SMT and PTH Low to moderate complexity Different reflow profiles One in the sweet spot One on the low end of the process window Comparison of results Assembly Inspection & Test Reliability Analysis

15 Test Vehicle Case 1 Case 1a Controller Board Approx. 6x6 Case 1b Display Board Approx. 3x3.5

16 Test Vehicle Case 2 Case 2 Controller Board Approx. 8x8

17 Reflow Profiles Case 1a Case 2 Peak Temp ~234C TA 217C: ~60sec Low end of reflow window Peak Temp ~245C TA 217C: ~75sec Sweet spot in reflow window PCB 1a reflowed near lower edge of process window

18 Reflow Profiles Case 1b Original Case 1b Modified Peak Temp ~234C TA 217C: ~60sec Incomplete reflow in area of higher thermal mass Peak Temp ~237C TA 217C: ~80C; TA 227C:~60sec Modified to achieve reflow in cold area PCB 1b modified slightly from production recipe

19 Other Assembly Process Info Case 1: 30 of each PCB for each alloy Printer: EKRA X4 Paste: Type 4 NC Stencil: 5mil Eform Placer: ASM SIPLACE D4 Reflow: BTU Pyramax 7-zone AOI: Mirtec MV-7 Wave solder: 5 sets of each alloy, all with SnCuNi wave alloy Case 2: 50 PCBs with each alloy Printer: MPM UP2000 Paste: Type 3 NC Stencil: 5mil laser-cut, E-polished Placer: Mydata MY12 Reflow: Electrovert Omniflo 10- zone Wave solder: All 50 PCBs, with wave alloy matched to SMT alloy

20 Solder Joint Appearance, Case 1 SnCuNi SAC305 Both alloys produced acceptable solder wetting and fillet formation

21 SnCuNi Appearance The cooler reflow process produced joints with less luster than usual: Case 1a SnCuNi Typical SnCuNi Solder Joints

22 Inspection & Test Case 1: All (120 total) PCBAs inspected visually and by AOI No defects found by inspectors or AOI machine SnCuNi SMT joints had slightly duller appearance than usual 5 of each alloy built up with PTH components, wave soldered (with SnCuNi) and assembled into chassis for functional test All 10 complete assemblies passed functional test Case 2: All (120 total) PCBAs inspected visually; no AOI Similar (undisclosed) defect rates, between alloys except for tombstones PCB design has a tendency for tombstones SnCuNi: no tombstones SAC305: avg 3 tombstones/bd All passed in-circuit test All passed functional test

23 AOI, Case 1 Performed 100% on production line No defects found on SnCuNi or SAC305 SnCuNi & SAC305 on same program Samples sent to Mirtec lab for analysis Both alloys ran under same program with standard lighting No program or parameter tweaks required

24 X-Ray Analysis SMT, Case 1 SnCuNi SAC305 The SnCuNi produced slightly more SMT voids than the SAC305, but within the acceptable 25% range. Voiding can be mitigated by adding soak to reflow profile

25 X-Ray Analysis PTH, Case 2 SnCuNi SAC305 The SnCuNi produced less voids than the SAC305 Causes of PTH voiding not investigated

26 Wetting, Case 1- Least thermal exposure SnCuNi SAC305 Both alloys showed good wetting

27 Intermetallic Compound (IMC), Case1 Both alloys show continuous IMCs approximately 3um thick SnCuNi SAC305 SnCuNi IMC is smoother and more nodular due to nickel s inhibition of intermetallic precipitation Grain structure of bulk solder is more equiaxed SAC305 s IMC is more dendritic Grain structure of bulk solder is more columnar Both alloys showed good IMC formation

28 Thermal Aging, Case 1 Assemblies aged at 125 C for 96 hours (4 days) Cross sectioned No changes in IMC Shear tested 0805 components 15 angle Dage 400 series shear tester No decrease in shear strength Further aged at 150 C for 240 hours (10 days)

29 Thermal Aging & IMC Growth SnCuNi SAC305 As Reflowed ~3µm ~3µm Age #1: 125 C 96 hrs ~3µm ~3µm Age #2: Age # C 240 hrs ~8µm ~7.5µm Initial thermal aging had no significant effect on IMC growth; continued aging at higher temperature showed similar effects

30 Thermal Aging & Shear Strength Thermal aging had no significant effect on joint strength

31 Thermal Cycling, Case 2 - Highest Complexity Assemblies cycled C 1 hour cycles Tested at 1000, 2000 and 3000 cycles Electrical test no defects found Cross-sections: no structural changes noticed in solder joints Shear tests: no significant changes in joint strength SnCuNi had unexplained lower average shear strengths at 1000 cycle interval, but typical strengths at 2000 & 3000 cycle measurements

32 Thermal Cycling & Shear Strength Thermal cycling had no significant impact on joint strength

33 Summary: Assembly and Test Process Case 1 Case 2 30 each of controller and display PCBs with SnCuNi and SAC total 100 total Reflow Cool Moderate Build Quantity Wave solder Visual Inspection Automatic Optical Inspection 5 of each with SnCuNi only 20 total No defects found No defects found on production line No program tweaks needed in laboratory analysis 50 each with SnCuNi and SAC each with SnCuNi and SAC total Similar defect rates but more tombstones with SAC305 In-Circuit Test Pass Functional Test 5 of each pair pass Pass -----

34 Quality and Reliability Analysis Characteristic Case 1 Case 2 Visual Appearance Voiding Microstructure Thermal Aging 125 C for 96 hrs Additional 150 C for 240 hrs Thermal Cycling C, 1 hr Shear Strength Slightly duller than typical SnCuNi surface finish Acceptable wetting SnCuNi showed slightly more voiding in SMT. Both within acceptable limits Uniform, continuous IMC approx. 3um thick in both alloys SMT joints. Equiaxed grain structure in bulk solder of SnCuNi; dendritic grain structure in bulk solder of SAC305. Initial aging had no effect on IMC or grain structure or either alloy. Additional aging showed similar results Resistors average 4-6kg shear strength, no significant difference between alloys before or after thermal aging. Typical surface finish and wetting SAC305 showed slightly more voiding in PTH Both within acceptable limits No electrical test failures found after 1k, 2k & 3k cycles No structural changes observed in cross sections at 1k, 2k & 3k cycles 0805 Resistors average 4-6kg shear strength, only deviation noted was SnCuNi after 1k thermal cycles.

35 Discussion & Conclusions (1) Two case studies compared performance of SnCuNi and SAC305 in moderate and cool reflow profiles Moderate profile provided performance benchmark, cool profile provided critical process characterization information SnCuNi produced acceptable solder joints, even when processed at temperatures previously considered too cool for the alloy Visual appearance of SnCuNi joints were duller than when processed in hotter reflow profile, but good wetting and IMC formation were observed in cross sections AOI recognition without program adjustment eases implementation concerns on automatic inspection

36 Discussion & Conclusions (2) Voiding SMT voiding was slightly higher with SnCuNi alloy PTH voiding was higher with SAC305 alloy Both within acceptable ranges Shear strengths were in normal range, before and after thermal aging and thermal cycling Thermal cycling (3000 cycles) had no noted effects on solder joint electrical or mechanical integrity on Case 2 Forewent thermal cycling on Case 1 boards based on failure-free performance of Case 2 and result of other Case 1 solder joint analyses

37 Discussion & Conclusions (3) Thermal aging cycle had no significant effect on IMC growth rate Silver-free solder alloy can reduce solder paste costs by as much as 20% Concerns about low temperature processing have slowed its implementation & delayed cost savings Case 1 demonstrates the SnCuNi s full compatibility with the SAC305 reflow process window, even on the low end Can be considered a viable drop-in replacement for nearly any SMT process that uses a SAC or SAC-based alloy

38 Thank You Questions? Karl Seelig Vice President of Technology AIM Solder Chrys Shea President Shea Engineering Services