Minimizing Voids in Solder Joints

Estimated reading time: 1 minute

Dr. Hans Bell of Rehm Thermal Systems, Henryk Maschotta of Thales Deutschland GmbH, and Dr. Heinz Wohlrabe of TU Dresden presented the results of their project on reducing void content in the solder joints of land grid arrays (LGAs) at a soldering seminar during the Technology Days of the recent SMTconnect show in Nuremberg. They also discussed important influencing factors.

For their investigation, modules were equipped with LGA32 and LGA35 (Figure 1), among others. The internal structure of this LGA and the complex two-sided assembly made X-ray analysis particularly difficult with regard to the robust evaluation of the void content in the solder joints (Figure 2).

Figure 1: Solder side (bottom) of the LGA35.

Within the scope of the project, various X-ray methods, such as 2D X-ray laminography and X-ray CT, were used for comparative purposes in order to obtain evaluable data. In addition to the variations of the PCB design and the material parameters, the final vacuum pressure has been the overall influencing factor on the minimization of the void content. With decreasing pressure in the environment of the molten solder joints, their void content decreases significantly.

Figure 2: CT image of the soldered LGA32, internal structure.

Figure 3: Dark field image of an LGA solder joint.

In general, good LGA solder joints could be produced. The solder wetting of both the LGA connections and the PCB pads was uncritical, as the metallographic investigations prove. The solder gap characteristics essentially correspond to theoretical expectations and the variance is more influenced by the design of the PCB pads than by the voids. The microsections produced reveal the details of the topography of both joining partners, the printed circuit board and the LGA.

Figure 3 shows a section of an LGA32 solder joint in darkfield contrast, on which the dimensions of the respective solder resist masks and the gap remaining between them are measured. Only these remaining 13µm gaps allow the LGA to float in (self-centering) during soldering. Tolerances in the general manufacturing process can further shrink this gap so that both solder resist masks lie on top of each other, making it impossible for the LGA to float in.