Masters Abstracts (1996)

KINI, NAVEEN
(November 1996), Sr. Manufacturing Engineer, Smart Module Systems, Fremont, CA

Email: nkini@smartm.com

Effect of Board Bumping, Board and Component Related Parameters on DCA Assembly Yields

Direct Chip Attach (DCA) is becoming a strategy that is being widely used in the electronics packaging arena to deal with the need for product miniaturization with concurrent product size reduction. The constraints relating to the pitch and contact pad design requirements of DCA have posed significant challenges to board and component manufacturers. These challenges often show up as variations in the pad and solder mask dimensions, which may significantly impact DCA assembly yields. Moreover, the need for low cost DCA assemblies has resulted in the development of an alternative and cheaper solder bumping technique through the use of stencil printing. This technique can be extended to solder bump formation both on the die as well as the substrate. However, the drawback of this technique is the solder volume variation that could ultimately result in coplanarity problems of the solder bumps on the die or the substrate (if board bumping is involved). The solder bump heights are a great concern in the board bumping process, especially due to the variations that could arise from the pad and solder mask designs on the boards.

This research focused on the characterization of the component and the boards used in DCA assembly. The impact of the board parameter variations such as contact pad design variations, variations in the dimensions of the pad and the solder mask, and solder mask misregistration on the solder bump heights in a board bumping process have been studied and presented with the help of case studies. Also, the impact of component related parameters such as solder bump coplanarity, solder volume variations based on the location of the die within a wafer, solder bump deformation, and voiding within the solder bumps have been illustrated.

A 'Design of Experiments' based strategy was used to evaluate the effectiveness of the board bumping process in terms of assembly yields. The experiments performed have indicated that solder paste printing parameters as well as the board parameters can cause variations in the deposited solder paste volumes. Sensitivity analyses have shown that although mask defined pads showed the minimum scatter in the deposited solder volumes through board bumping, a circular via-in-pad would result in the minimum scatter in the solder bump heights for a given solder volume distribution. Hence, a circular via-in-pad would be preferred for a board bumping scenario. Evaluation of the robustness of different pad and solder mask geometries on the board (in a no-board bumping case) has indicated that the narrowest possible trace with or without circular pads would be preferred on the board. The desired pad design on the die should range within 3.5-4 mils if enhanced assembly yields are to be obtained.