Working with Large, Thick PCBs

Estimated reading time: 9 minutes

By Jennifer Nguyen; Robert Thalhamer; David Geiger; Dan Rooney; and Dongkai Shangguan, Ph.D., Flextronics

The process challenges of lead-free wave soldering for thick, large, and thermally challenging PCBs are discussed, with recommendations. Process factors for thru-hole fill and design to reduce quality defects during lead-free wave soldering are outlined.

As compared with SMT, there are limited studies on the reliability of thru-hole solder joints. The wave soldering process is subject to more variability than SMT, and achieving acceptable thru-hole fill on large, thick, and thermally challenging PCBs is a process challenge, even with tin/lead wave soldering. Differences in solder microstructure may affect reliability of partially and completely filled solder joints, in addition to the differences in the mechanical integrity from use of a higher solder volume. This challenge is even greater with lead-free due to the characteristics of lead-free solder alloys, currently available flux materials, operating process condition limitations, thermal shock between preheat temperature and solder pot temperature, maximum component temperature limitation, existing equipment, etc. In addition, lead-free wave soldering may raise other concerns on plated thru-hole (PTH) solder joint quality and reliability, such as voiding, orange peel, micro surface cracking, copper leaching and dissolution, and other problems.

Design of Experiment (DOE)

Two board thicknesses (2.4 and 5.0 mm) and two board surface finishes (immersion silver and immersion gold) were considered. The test vehicle (Figure 1) has up to 14 layers with a maximum of 12 copper ground planes. It was designed into different areas to study pin-to-hole (P/H) ratio, pad/annular ring dimensions, spacing, orientation, and thermal relief on hole-fill. In DOE 1, the flux type, flux amount, preheat, temperature, and contact time were considered. In DOE 2, the pot temperature, conveyor speed, number of wave nozzles (chip wave on/off), and wave atmosphere were tested.

Figure 1. Experiments focused on processing thermally challenging, thick PCBs.

Flux type, preheat temperature, contact time, solder pot temperature, wave configuration, and wave atmosphere significantly affect hole fill. Contact time and solder pot temperature were most significant. Longer contact times and higher solder pot temperatures resulted in greater hole fill. Choosing the right flux material also was important to achieve good hole fill. Turning on the chip-wave and using nitrogen helped.

Pot temperature, conveyor speed, and chip-wave on/off had a major impact on thru-hole solder bridging. The number of solder bridges decreased dramatically when chip wave was off. Lower pot temperature or faster conveyor speed help reduce PTH bridging.

Contact time and chip wave were the two important factors that significantly affected the occurrence of open solder. Turning on the chip wave was most significant to reduce the occurrence of open solder for wave soldered SMDs. This is contrary to the recommendation for reducing bridging, in which the chip wave should be turned off. Hence, the need of the chip wave will depend on the complexity of the products and the component mix.

Hole Fill

In general, hole fill increased as the pin-to-hole (P/H) ratio decreased. Statistically speaking, however, the difference in hole fill was not significant for different P/H ratios. The effect of annular ring on hole fill also was insignificant. Components with round or square pins had better hole fill than those with rectangular pins of the same P/H ratio.

The internal copper plane had a significant impact on hole fill. Copper thickness (copper weight) and the number of connected layers also had an effect. PTH components that connected to two layers of 0.5-oz copper resulted in better hole fill than those that connected to one layer of 1.0-oz copper thickness.

Figure 2. Average hole fill with different component types and processes.

Component type had a great effect on hole fill. It became more critical when thick PCBs were wave soldered using lead-free processing. All components in the 2.4-mm-thick board achieved good hole fill, 80?100%. On the other hand, the same components placed in the 5.0-mm boards achieved hole fill ranging from 30% to 100%. The hole-fill reduction was not linear with PCB thickness and was unpredictable, as it also depended on the component types on the assembly. Board thickness did make it more difficult to achieve good hole fill for certain component types, regardless of the solder alloy process technology. There was slightly better hole fill for components in tin/lead wave soldering as compared with the same components in lead-free (Figure 2).

Bridging

Design, particularly component orientation through the wave, plays a large role in PTH bridging. More bridged pins were seen at the components in which their long body was sent in parallel to the wave than if they were perpendicular to the wave. This was consistent with previous experiments,1 as well as experience with tin/lead wave soldering. Effect of component orientation through wave became more critical when the component pitch was small (2.0 mm) or when a long component such as a header or DIMM connector was soldered.

Open Solder

Orientation can result in open solder for SMDs. In the case of SOT components, 0° orientation had the least insufficient solder defects, and 90° orientation had the most insufficient solder defects. Pad shadowing should be avoided for the SMDs in wave soldering.

Voiding

X-ray inspection was performed for the thru-hole components for hole fill and voiding, and the PTH hole fill values from X-ray were in reasonable agreement with those from cross sectioning. Voids were seen on both tin/lead and lead-free solder joints. More voids were seen on thicker boards than thinner boards. The ENIG boards had less voids than immersion silver. The P/H ratio significantly affected the amount and size of voids in the PTH solder joints. The smaller the PTH ratio, or the larger the hole diameter, less voiding was seen. For example, the P/H ratio of 0.59 had many voids, whereas almost no void was seen on the solder joints with P/H ratio of 0.30.

Microstructure

Cross sections of the PTH solder joints were performed to evaluate the solder microstructure, intermetallic formation, via hole fill, and the condition of the PTH metallization and PCB dielectric at time zero (prior to thermal cycle testing). The cross sections exhibited bending of the annular ring and copper eyelets, which was caused by Z-axis expansion of the PCB dielectric during wave soldering. The Cu6Sn5 intermetallic compound (IMC) at the solder-to-copper interface on the immersion-silver-plated PCB vias exhibited a thicker and more uniform area coverage than the thin ternary (Ni,Cu)Sn IMC layer on the pins. The area and thickness uniformity of the Cu6Sn5 IMC layer on the copper PTH metallization was normal; IMC thickness was 1?3 μm. The (Ni,Cu)Sn IMC at the solder-to-nickel barrier plating interface on the pins was inherently ultra thin (less than approximately 0.50 μm). There were no visual indications of a concern with copper dissolution at the knee of the PTH. No fillet lift-off solder cracks or cracked PCB solder eyelets were found at the ingress of the eyelet to the PTH barrel at time zero.

Figure 3. Grainy surface finish solder with 30% hole fill and higher volume fraction of interdendritic structures.

The microstructure of the lead-free solder comprises Cu6Sn5 and Ag3Sn IMCs embedded in a tin-rich, ß-Sn dendrites matrix. The two Cu6Sn5 and Ag3Sn IMCs precipitate into the interdendritic spaces, typically as fine particulates during solder solidification. The cooling rate of solder during reflow is a critical factor that determines the size of and spacing of the so-called “dendrite arms&rdqou; that form during solidification. Silver content in the solder is another significant factor in determining the size of the dendritic arms, as the dendritic arms that form with Sn3.0Ag 0.5Cu have been observed to be twice as large as those with the higher silver content alloy Sn3.8Ag0.7Cu, in solder joints with large CBGA components.2 The size of the component also determines the size of the dendritic arms. Solder microstructures with higher relative abundances of interdendritic Ag3Sn particles are stronger and harder than solder microstructures with lower relative abundances. The interdendritic solder microstructure affects the surface appearance of the solder joint. In the sample with 100% hole fill and a lower volume fraction of interdendritic structures, the surface of the solder joints generally is smoother and shinier than in the board with 30% hole fill and a higher volume fraction of interdendritic structures (Figure 3). The characteristic grainy surface of lead-free solder joints is a reflection of the dominance of primary tin dendrites.

Thermal Cycle Test

Samples of 30%, 50%, and 100% hole fill on 2.4- and 5.0-mm board thicknesses with immersion-silver finish were prepared for thermal cycling test. The components were daisy chained for continuous monitoring during test. Solder joints with and without pin protrusion also were considered.

The thermal cycle test was performed in an air-to-air thermal cycle chamber, from 0°?100°C with a 15-min. dwell time at each peak temperature, and a temperature ramp rate of approximately 15°C/min. One sample exhibited an electrical open following 500 cycles. This particular sample did not have pin protrusion. X-ray inspection revealed that the actual solder coverage was less than 25%, effectively. The solder was not properly wetted to the pin and was likely mechanically contacting it rather than metallurgically bonded. This resulted in the early failure.

Corner solder cracks and cracks along the sides of the PTH barrels were found following 1,000 thermal cycles. These had not yet resulted in electrical opens. Solder fillet lift-off cracks were not found. Samples with complete hole fill and pin protrusion displayed cracks originating both at the top and bottom fillets and propagated along the side of the PTH. The extent of cracked annular rings was greater in samples with complete hole fill and pin protrusion because of the thermomechanical stress resulting from coefficient of thermal expansion (CTE) mismatch between the solder, pin base metal, and PCB materials. Samples initially showed that the Z-axis expansion from reflow had stressed the annular rings on the vias and solder eyelets, and then thermomechanical fatigue during thermal cycling cracked the copper metallization.

Conclusion

Metallurgical analyses of time-zero samples showed that 30% hole fill had a more grainy surface appearance and a higher volume fraction of interdendritic IMCs than 100% hole fill samples, which illustrates the relationship of solder microstructure and appearance. These observations suggest that differences in solder microstructure might also have an effect on the reliability of partially and completely filled solder joints in addition to the differences in the mechanical integrity from the use of a higher solder volume. Thermal cycling induced changes in the microstructure of the solder joints and the difference in microstructure between partially and completely filled solder joints was less obvious following 1,000 thermal cycles. Cracks were visible in the solder joints but these did not result in open solder joints as of 1,000 thermal cycles.

REFERENCES:

1. J. Nguyen, R. Thalhammer, D. Geiger, H. Fockenberger, and D. Shangguan, “Large and Thick Board Lead-free Wave Soldering Optimization,&rdqou; Proceeding of APEX 2007.
2. H. McCormick, P. Snugovsky, C. Hamilton, Z. Bagheri, and S. Bagheri, “The Great SAC Debate: Comparing the Reliability of SAC305 and SAC405 Solders in a Variety Of Applications,&rdqou; SMTA Pan Pacific Symposium, January 2007.

ACKNOWLEDGEMENTS:

1. M. Arra, D. Shangguan, S. Yi, R. Thalhammer, and H. Fockenberger, “Development of Lead-free Wave Soldering Process,&rdqou; IEEE Transactions on Electronics Packaging Manufacturing, Vol. 25, No. 4, Oct 2002.
2. R.A. Szymanowski, D. Casati, E. Saglia, P. Lotosky, K. Howell, and G. Hueste, “Dynamics in Lead-free Wave Soldering,&rdqou; Circuits Assembly, February 2006.
3. M. Miyaki, M. Kanai, and S. Ogata, “Soldering properties for lead-free solders of Sn/Ag/Cu system solders,&rdqou; Proc. 3rd 1999 IEMT/IMC Symp., Tokyo, Japan, 1999, pp. 125?130.
4. J. Wilcox, G. Dearing, B. Smith, T. Skoczowski, T. Sack, B. Sullivan, “Accelerated Reliability Testing and Analysis of Lead-free Solder Interconnects,&rdqou; Proceedings of IPC/APEX 2006.

Jennifer Nguyen; Robert Thalhamer; David Geiger; Dan Rooney; and Dongkai Shangguan, Ph.D., may be contacted at the Corporate Technology Group, Flextronics International, San Jose, Calif. 95131. Information from this study was presented at SMTA International 2007.