SOLDER MATERIALS
The Interaction of 2 Solder Paste Alloys with 5 Surface Finishes
For spread and wetting performance, certain finishes stand out. by Pritha Choudhury, Ph.D., Morgana Ribas, Ph.D., John Fudala and Mitch Holtzer

Electronic assemblers have myriad material and process choices to make, not limited to board materials, solder masks, laminate Tg’s, components, surface finishes, assembly materials and design for manufacturing (DfM) process conditions. High-reliability alloys such as Innolot are designed to meet harsh automotive conditions and extend service life of the solder joint. Applications requiring higher operating temperatures and increased number of cycles to failure have benefited by implementing that alloy. While solder alloy selection is an important factor in determining reliability of the solder joint, considerations should be made for surface finish selection to further enhance performance. This study explores surface finish factors such as IMC formation, voiding and solder spread that contribute to reliability.

Each choice can have a significant impact on the in-service reliability and commercial success of the assembly. This multi-part article will focus on data developed from an extensive study of surface finishes and solder pastes used by many global, high-reliability assembly manufacturers. The study included two commonly used solder alloys in paste form:

  1. SAC 305 (96.5%Sn, 3%Ag, 0.5%Cu) powder size distribution (PSD) type 4 with novel “CVP-390” paste flux
  2. Innolot (91.95%Sn, 3.8%Ag, 0.7%Cu, 3.0%Bi, 1.4%Sb, 0.15%Ni) PSD type 4 with the novel paste flux and five variations of surface finishes, including
    1. Organic solderability preservative (OSP) (MacDermid Enthone Entek Plus HT) using two thickness levels
    2. Immersion tin (Ormecon CSN)
    3. Immersion silver (MacDermid Enthone Sterling)
    4. Electroless nickel/immersion gold (ENIG) (MacDermid Enthone Affinity).

Characterizations of void creation, solder spread and initial intermetallic compound (IMC) thickness are discussed here. Details of IMC growth after thermal cycling, and the effect of the IMC thickness versus solder joint shear strength for each combination of solder alloy and surface finish, will be covered in the future. Results of data using multiple surface finishes and low-temperature soldering alloys are being developed and will be presented in a follow-up article.

Background

Metal-to-metal interconnects can be made using any of three common processes: welding, brazing or soldering. Welding requires reaching the melting temperature of the joined metal(s). Copper has a melting point of 1,085°C. Although welding copper to copper creates a very strong connection with no intermetallic compounds, the very high temperature eliminates its practicality for assembling components to an epoxy/copper-laminated substrate.

Brazing, according to the American Welding Society A3.0M/A3.0:2020 standard, requires reflow temperatures in excess of 450°C. Brazing is commonly used to bond copper-to-copper tubing used in water supply lines and HVAC systems. Many alloys of brazing materials are available; 80%Cu/15%Ag/5%P is a common copper-to-copper brazing alloy. Very strong joints are formed; however, the high temperatures required for brazing and the cost of a 15%Ag brazing alloy are prohibitive for circuit assembly applications.

Solder joints need to create a reliable mechanical bond with low electrical resistivity between component I/O joints and the copper pads on a printed circuit board. Copper pads on a PCB require a surface finish to prevent oxidization during the period between circuit board production and the assembly of surface mount components.

Two solder alloys and four common copper surface finishes were used in this study. Hot air surface leveling (HASL) copper surface finish was excluded from this study because of the issue with poor coplanarity. This became an issue when the RoHS (lead-free solder) and sub-0.8mm BGA pitch technologies converged in 2006, and has become more problematic as assemblers began placing BGA devices with 0.5mm and 0.3mm pitches. Use of HASL finishes has declined dramatically in all but non-ROHS compliant assemblies using larger (1.0mm and above) BGA components.

Experimental Design

This article discusses voiding resistance, spreading/wetting of the solder alloy on the surface finishes, and initial IMC thickness of paste alloys and surface finishes. Two different OSP coating thicknesses were used: 0.4µm and 0.6µm. Both zero and one reflow preconditioning cycle were tested for each combination of materials. Five replicate test vehicles were measured for each condition. Convection reflow was performed using a Heller 12-zone reflow oven. The reflow profile, shown in FIGURE 1, was relatively hot, with no nitrogen used for the preconditioning reflow. The same profile using <1,000ppm O2 was used for final reflow.

Voiding inspection used a Phoenix Micromex system with a slightly different test vehicle. The print and reflow parameters were identical with all combinations of solder paste alloy and surface finishes.

Initial IMC thickness was measured. Results of IMC growth and shear testing differences after 2,000 thermal cycles will be reported in the second installment of this article.

Figure 1. -155°C to 175°C, 70 sec. soak, 240°C peak, 70 sec. TAL
Figure 1. -155°C to 175°C, 70 sec. soak, 240°C peak, 70 sec. TAL.
Figure 2. Voiding main interaction plots
Figure 2. Voiding main interaction plots.
Results and Discussion

Formation of a strong solder bond via the interfacial IMC is essential for the functionality and reliability of the package. With smaller solder joints, the influence of the IMC layer on joint reliability is more significant.1 Solder joint reliability depends not only on the solder alloy but the component, PCB finishes, and the IMC formed within and at the solder/substrate interface.2,3,4 The PCB surface finish forms a critical interface between the bare copper on the PCB and the component to be assembled and therefore is an important factor in the reliability of solder joints.2,5 The most important function of the PCB finishes is to increase the solderability of the substrate by preventing oxidation of the copper pads on the PCB, even after extended times between the PCB fabrication and SMT assembly.

Voiding. A summary of the voiding results as a function of solder alloy, surface finish component type with zero and one prior reflow is shown in FIGURE 2. The results show the surface finishes examined had little effect on the measured voiding. The high-reliability Innolot alloy-based solder paste generated slightly more voids on average than the SAC 305 paste. Previous internal studies have shown reducing the peak reflow temperature and increasing the time above liquidus (TAL) reduces voiding with Innolot and the novel flux, but this was beyond the scope of this study.

A second result was a clear increase in voiding with the land grid array (LGA-256) device. Voids are reduced as volatile gasses from the solder paste flux escape from the solder joint during preheat and especially during the TAL portion of the reflow profile.

Low-offset LGAs can inhibit this outflow of vapors, leaving them entrapped in the solder joint. While LGA voiding was relatively low (<12%), this device clearly had measurably higher levels of voiding among the components tested. One-time prior reflow had a minor effect on voiding levels, especially with the LGA-256 device.

Figure 3. Cross-print spread results of SAC 305 and Innolot with each surface finish
Figure 3. Cross-print spread results of SAC 305 and Innolot with each surface finish.
Figure 4. Tin-copper phase diagram
Figure 4. Tin-copper phase diagram.
Figure 5. Micrographs of measured Cu3Sn and Cu6Sn5 intermetallic layers
Figure 5. Micrographs of measured Cu3Sn and Cu6Sn5 intermetallic layers.

Wetting. Formation of the interfacial IMC is strongly influenced by the processing parameters during reflow because of its effect on wetting and microstructure.2,6 ENIG is unique in that the high-tin alloys tested interact with nickel to form (Cu, Ni)6Sn5. Some surface finishes could act as a barrier layer to reduce interdiffusion between the solder and copper base and thus reduce the formation of IMCs.5,7

Solder spread and wetting were measured using an internal test method developed by MacDermid Alpha known as the cross-print spread test. In this procedure, solder paste is printed and reflowed on parallel traces of the surface-finished copper. The distance between the traces is spaced in 0.1mm intervals ranging from 0.6mm to 0.9mm. The ability of the paste to bridge a larger gap is an indication of good spread (FIGURE 3).

R1206 passive components with tin-coated nickel terminations and multiple-sized BGA components with SAC 305 spherical interconnects were used throughout the study.

A DEK Horizon printer and a 0.004″ thick laser-cut stainless steel with no nanocoating was used to print the solder paste deposits. Each condition used the same print process settings (2″/sec. squeegee speed, 15 lb. of pressure and stencil snapoff speed @0.2″/sec.). A Fuji NXT-II pick-and-place machine was used to place the components into the paste deposits.

Cross-print spread results. In general, metallic surface finishes (ENiG, ImSn, ImAg) gave better results in the cross-print spread test (FIGURE 3). The interesting finding is that a thicker OSP coating (0.6µm versus 0.4µm) showed better solder spread. The increased resistance to copper oxidation with the thicker OSP coating may be the explanation.

As-reflowed condition. Soldering of copper substrate involves eutectic melting (reflow) of solder bump and (ii) reaction of molten solder with substrate, resulting in the formation and growth of one or two intermetallics; i.e., Cu3Sn (ε-phase) and Cu6Sn5 (η-phase). Mechanical bonding is mainly provided by the η-phase that has a peculiar scallop-like morphology.8 The η-phase (FIGURES 4 and 5) changed from a scallop to a flat structure in the presence of as little as 0.05 wt% Ni.9

Figure 6. Initial IMC thickness in R1206 on different surface finishes with SAC 305 and Innolot pastes
Figure 6. Initial IMC thickness in R1206 on different surface finishes with SAC 305 and Innolot pastes.
Conclusions

The lack of increased voiding after thermal cycling indicates neither intermetallic micro-cracks nor Kirkendall voiding was an issue with any of the combinations of solder paste alloys or surface finishes. Bulk voiding in the solder joint would not be expected to increase or decrease as a function of thermal cycling.

Innolot was prone to a slightly higher level of voiding. Reflow profile adjustment has proven to mitigate this issue.

A clear trend showed higher spread and wetting performance using the metallic (ImSn, ENIG, ImAg) pad finishes versus OSP. The thicker OSP finish (0.6µm) resulted in higher wetting performance with both the SAC 305 and Innolot-based pastes versus the 0.4µm thick OSP pad finish.

Each combination of solder alloys and surface finishes created a measured IMC between 0.9µm and 2.7µm. The data show the process capability (CpK) of the ENIG finish appears lower than the other combinations. The results of IMC growth after thermal cycling will be discussed in the next article in this series.

The follow-up article will explore the effect of solder alloy on each of the primary final finishes discussed here when exposed to harsh thermal cycling requirements (-40°/160°C).

References
  1. J.W.C. de Vries, M.Y. Jansen and W.D. van Driel, “On the Difference between Thermal Cycling and Thermal Shock Testing for Board Level Reliability of Soldered Interconnections,” Microelectronics Reliability, vol. 47, 2007.
  2. Jianbiao Pan, Jyhwen Wang and David M. Shaddock, “Lead-Free Solder Joint Reliability – State of the Art and Perspectives,” Journal of Microelectronics and Electronics Packaging, vol. 2, no.1, January 2005.
  3. A. Siewiorek, A. Kudyba, N. Sobczak, M. Homa, Z. Huber, Z. Adamek and J. Wojewoda-Budka, “Effects of PCB Substrate Surface Finish and Flux on Solderability of Lead-Free SAC305 Alloy,” Journal of Materials Engineering and Performance, vol. 22, no. 8, August 2013.
  4. Anupam Choubey, Hao Yu, Michael Osterman, Michael Pecht, Fu Yun, Li Yonghong and Xu Ming, “Intermetallics Characterization of Lead-free Solder Joints under Isothermal Aging,” Journal of Electronic Materials, vol. 37, August 2008.
  5. James Webster, Jianbiao Pan and Brian J. Toleno, “Investigation of the Lead-free Solder Joint Shear Performance,” Journal of Microelectronics and Electronic Packaging, vol. 4., no. 2, second quarter 2007.
  6. A. C. Spowage, L. M. Sim, C. M. Thong, P. A. Collier and G. Y. Li, “Investigation of the Effects of Surface Finish and Simulated Service Aging on the Characteristics of the Intermetallic Layer Formed in Various Lead-Free Solder Joints,” STR/04/033/PM.
  7. Wonil Seo, Yong‑Ho Ko, Young‑Ho Kim and Sehoon Yoo, “Void Fraction of a Sn-Ag-Cu Solder Joint Underneath a Chip Resistor and its Effect on Joint Strength and Thermomechanical Reliability,” Journal of Materials Science: Materials in Electronics, August 2019.
  8. Oleksii Liashenko, Andriy M. Gusak and Fiqiri Hodaj, “Phase Growth Competition in Solid/Liquid Reactions between Copper and Cu3Sn Compound and Liquid Tin-Based Solder,” Journal of Materials Science, Materials in Electronics, no. 10, October 2014.
  9. Per-Erik Tegehall, “Review of the Impact of Intermetallic Layers on the Brittleness of Tin-Lead and Lead-Free Solder Joints,” IVF Project Report, June 2007.
Pritha Choudhury, Ph.D., is senior scientist R&D; Morgana Ribas, Ph.D., is manager of Metals Technology Group – R&D; John Fudala is senior process specialist; and Mitch Holtzer is technical knowledge leader, Global R&D at MacDermid Alpha Electronics Solutions (macdermidalpha.com).