A customer observed assemblies which had grainy plated through-hole solder joints with poor or no wetting of the bottom sides of the holes.

The printed circuit board (PCB) design had an ENIG finish. A eutectic tin-lead solder RMA paste/flux combination was used in the assembly process with three reflow processes, the last being the Through-Hole Technnology (THT) step.

The customer requested confirmation that black pad was notthe failure mode, and, if not, to determine the actual cause.

Test Methods
The customer supplied two bare boards and a failed assembly. The first board was submitted for elemental surface analysis using Auger Spectroscopy. This was performed at both a pad area and at the annular ring of an adjacent plated through-hole. The plating thickness was measured at the location of several pads and annular rings of both boards using X-ray Fluorescence Spectroscopy (XRF).

SEM/EDS (Scanning Electron Microscopy / Energy Dispersive Spectroscopy) analysis of the solder joints from a failed assembly was performed to examine the microstructure and intermetallic make up of the solder joint along with surface analysis of a board.

Finally, a pin header from the failed assembly was cross-sectioned, mounted in epoxy, and prepared for SEM/EDS analysis.

Results
Auger Spectroscopy indicated the presence of only carbon, oxygen, nitrogen, sulfur or chlorine on the surface of the gold, at both the pad and annular ring.

XRF analysis indicated that the thickness of the underlying nickel plate at pads and annular rings was the same and averaged 123 micro inches; however, the gold top coat averaged 6.0 micro inches at the pads and 1.5 micro inches at the annular rings.

SEM/EDS indicated that the composition of the annular ring plating surface was different than that of the pad, confirming the XRF results.

There was however significant voiding observed in the cross-sectioned pin header. In general, the voids appeared non-uniform, suggesting that they were not caused by out-gassing of flux and tended to occur at the interface of either the barrel or pin.

The solder appeared to reflow properly at the header, as the solder grain structure was typical. SEM/EDS analysis of the pin at the header indicated that it was composed of copper and zinc, which was expected. SEM/EDS analysis of the barrel from pin 6 of header J-31 indicated it was composed of copper, also as expected. SEM/EDS analysis of the cracks at pin 6 from the header indicated phosphorus levels of 4-6wt% at the electroless-nickel (E-Ni) coating (Figure 3-1).

SEM/EDS analysis of both interfaces of the crack indicated that sufficient nickel-tin-intermetallic compound (Ni-Sn-IMC) was not generated at the interface between the solder and the barrel (Figures 3-1 and 3-2).

Conclusions
The poor solderability was the result of nickel oxidation, caused by thin gold plating. Nickel oxide is among the more stable metal oxides. Even aggressive flux chemistries are unable to reduce nickel oxide.

IPC-4552 recommends a minimum of 1.97 micro inches of immersion gold plating, and IPC-2221 recommends an electroless-Ni plate of 98 to 197 micro inches and an immersion gold plating of 3.14 to 9.05 micro inches. In this case, the immersion gold plating ranged from 0.01 to 3 micro inches at the annular rings.

The thin gold did not sufficiently protect the underlying bar-rier layer of nickel during the three reflows and caused the observed de-wetting. The de-wetting was confirmed by the lack of Ni-Sn-IMC formation.

The results did not indicate a black pad failure mechanism, as many of the associated characteristics were not present:

Solderability issues were limited to the bottom side of the plated through-hole suggesting that intermediate plating chemistries were not removed and became trapped within the plated through-holes which could have hindered plating of the gold.

Recommendations
To further confirm that a solderability problem is present, the EMPF recommended either steam aging or replicate reflows of a bare board and subsequent wetting balance testing as perJ-STD-003A. A properly applied ENIG finish would be expected to survive steam aging or multiple reflows.

Examination of the plating process is also recommended. Plating baths consist of a myriad of constituents, including buffers, complexing agents, and brighteners. If these become unbalanced, a thin or skipped plate can result.

For more information on these types of issues, or to see how the EMPF Helpline can aid you in resolving your own maunfacturing problems, call (610) 362-1320, or email the helpline at helpline@empf.org.