Authored By:
Shuai Shao1, Francis Mutuku
Binghamton University, SUNY, Binghamton, NY, USA
Babak Arfaei, Ph.D., Jim Wilcox, Ph.D.
Universal Instruments Corporation, Conklin, NY, USA
Summary
Drop shock reliability testing was performed on circuit boards assembled with several different lead-free solder alloys including SAC305 (Sn3.0Ag0.5Cu). The solder compositions tested range in Ag content from 0 to 3.0% by weight. Alloys with various secondary alloying elements were also included. All drop test boards were assembled such that the solder paste composition matched that of the BGA solder ball alloy to produce homogeneous solder joints of known compositions. An alternative test board design (not JEDEC standard) was used for this drop test evaluation. The test board contains a centrally located CABGA 256 package (17x17 mm body, 1 mm pitch). The board was designed with soldermask defined pads to minimize the occurrence of pad cratering failure modes in the laminate material. The test package was soldered to the drop board using either BGA or LGA interconnections to explore the effects of solder joint volume. Drop shock events were characterized with acceleration monitoring on the drop table and strain gage measurements on the mounted test boards.
All samples were dropped until electrical failure. Solder joint microstructural analysis was performed on failing parts to establish the failure modes. The dominant failure mode was observed to be solder joint failure, either in the bulk solder or cracking along the interfacial intermetallic compound on the board pad. The effect of alloy silver content on drop reliability is noted. SAC305 solder joints were found to produce the best drop performance of all alloys tested for both BGA and LGA joint formats.
Conclusions
Five Pb-free solder alloys on two PCB surface finishes were evaluated for drop shock reliability with two different solder joint volumes (LGA and BGA). Using a drop test board specifically designed to promote solder joint failures (i.e., solder mask defined board pads), the following experimental observations were made.
Of the five solder alloys evaluated, SAC305 performs the best, or nearly so, for all test conditions (board finish and solder joint volume). SN99CN is generally the second best drop performer with SN100C performing very similarly. In the BGA joint configuration, SAC-M is characterized by notably low variability in drop lifetime when used on Cu- OSP finish boards. On immersion Ag boards or in the LGA configuration, its failure rate variability is more typical.
Repetitive drop shock testing was seen to produce four distinct interconnect failure modes: bulk solder failure, interfacial IMC failure, mixed IMC/solder failure and laminate pad cratering. Different failure mode trends were observed between BGA and LGA joints. Board surface finish also played a role in determining failure mode. On Cu-OSP surface finish, SAC305 BGA joints showed mainly pad cratering failure while BGA joints of other alloys generally showed mixed IMC/solder failure. On the ImmAg finish, the results were roughly reversed; BGA joints of SAC305 showed IMC/solder failure while other alloys mostly produced pad cratering failures.
LGA joints on the Cu-OSP finish produced mainly bulk solder failures. On the ImmAg finish however, LGA joints produced examples of all four failure modes with the low Ag alloys tending to have more solder bulk failure. For BGA joints on ImmAg, alloys with lower Ag amount tended to have more pad cratering. Pad cratering failure was in general more prevalent on the ImmAg finish.
Given the integral involvement of the laminate in the drop shock failure process, the observations and conclusions made in this study should be considered applicable only to the laminate material and pad design used. If a more robust laminate formulation further suppresses the pad cratering mechanism, more solder and interfacial failures would be observed, perhaps altering the observed relative performance the alloys.
|
