When HOP failures pop up, review the reflow profile.

Recently, a customer observed intermittent opens on thin and fine pitch BGAs (TFBGA). The PCB is a Pb-free design and the TFBGAs were soldered with SAC 305 paste.

X-ray inspection of the identified failure locations on the TFBGAs indicated possible non-wetting between the paste and solder balls. This type of non-wetting (Figure 1) is referred to as head-on-pillow (HOP)¹. Confirmation of HOP failures requires destructive analysis. The failed TFBGAs were microsectioned to examine the possible failure location (Figure 2).



Optical microscopy indicated the solder paste never wetted the solder balls. Further analysis using high magnification scanning electron microscopy (SEM) imaging (Figure 3) clearly shows a gap between the solder ball and solder paste. The pads were observed to be well adhered to the substrate with minimal voiding in the solder balls. The failed ball joints showed no cracks or separations at the solder-to-pad or component-to-solder interfaces. Energy dispersive spectroscopy (EDS) analysis of the bulk solder confirmed Pb-free, Sn-based solder was present in both the ball and paste regions.


SEM/EDS analysis at the component-to-solder interface (Figure 4) detected the presence of tin and nickel intermetallic compounds (IMC). These compounds indicate sufficient heat was applied to obtain the necessary melting and dissolution required for good solder joint formation. In addition, SEM/EDS analysis at the solder-to-pad interface detected the presence of tin and copper, another indication of an IMC formation, which confirms good wetting of solder to the pads and components. Because of the lack of phosphorous and cracks in the intermetallics, it was concluded the failure mechanism was not related to black pad. Destructive failure analysis confirms the intermittent open failures are classic HOP: nonwetting between solder balls and solder paste. This is usually the result of a non-optimized reflow profile, in which solder finishes reflowing and flux completely volatizes before reflow occurs at the solder ball. The observation that only certain solder joints displayed HOP may be due to these solder joints being located near more thermally conductive traces.


In this case, we recommend reviewing the reflow profile to solve HOP failure modes. In addition, we recommend checking the paste printing variables (printing speed and pressure) to ensure the proper amount of paste is deposited on the PCB. This can be affected by the contact between the stencil and the PCB, the stencil thickness, and the aperture design.

Finally, we recommend reviewing the board designs for excessive heat transfer that may prevent proper reflow of the solder balls. A large ground plane located near the TFBGA may conduct enough heat away to interfere with proper solder ball reflow. Preheating the board and modifying the reflow profile may be necessary for the solder balls to reflow properly. A board designed for extreme conductive cooling may become a soldering challenge for the manufacturing engineer.

Reference

1. Chrys Shea, HOP-ping Mad: The Head-on-Pillow Epidemic, Circuits Assembly, July 2008.

The American Competitiveness Institute (aciusa.org) is a scientific research corporation dedicated to the advancement of electronics manufacturing processes and materials for the Department of Defense and industry. This column appears monthly.