Testing Intermetallic Fragility

Martin K. Anselm, Ph.D. and Brian Roggeman
Universal Instruments Corp.
Conklin, NY, USA

As reliability requirements increase, especially for defense and aerospace applications, the need to characterize components used in electronic assembly also increases. OEM and EMS companies look to perform characterizations as early as possible in the process to be able to limit quality related issues and improve both assembly yields and ultimate device reliability.

In terms of BGA devices, higher stress conditions, RoHS compatible materials and increased package densities tend to cause premature failures in intermetallic layers. Therefore it is necessary to have a quantitative and qualitative test methodology to address these interfaces.

Typically, solder ball shear or pull testing is employed to measure the interfacial strength, sometimes requiring very high speeds to do so. While there is no current industry accepted specification on proper test speeds, strength or energy metrics, procedures do exist which allow for relevant comparisons. These tests are always run on unassembled BGA devices, so the interaction with the PCB is completely removed. While the data is useful for the component manufacturer, the risk is that the test does not fully represent the final assembly in terms of metallurgical condition.

Specifically when BGA components using a Nickel-Gold surface finish are soldered to PCBs with a Cu-based pad (ie, Cu-OSP, ImmAg, ImmSn or HASL), there will be additional Cu dissolved into the solder joint. The addition of this copper can have an important effect on the intermetallic structure at the ENIG pad. Current mechanical solder ball testing procedures on unassembled BGA devices do not accurately duplicate the condition of this intermetallic structure. The test results on ENIG pads will then not necessarily correlate to actual manufacturing reliability.

From this research we have determined that generating an intermetallic morphology that is similar to a standard mass reflow surface mount process is not straight forward. The method used to add Cu to the ENIG pad and lead-free solder system will affect the morphologies at the electroless Ni substrate and therefore the mechanical properties of the intermetallic.

Data is presented on the intermetallic strengths and failure modes of two bond pull test methods. Specifically Hot Bump Pull (HBP) and Cold Bump Pull (CBP) testing are compared where Cu is added by the copper pins of the HBP tester or by Cu power in a second reflow followed by CBP testing.

Intermetallic morphology affects the peak load to failure distribution in CBP testing and a possible narrow distribution can be created for SnAg using a long ball attach profile. This condition is favorable for reliability predictions and mitigation of infant mortalities for lead free product.

More research must be conducted to compare the intermetallic morphologies created in these tests with those observed in standard component attach. Since individual solder joints cannot be mechanical tested in a BGA the only comparison that can be made is through intermetallic morphological comparison. An improved HBP test method may provide pull testing results that vary based upon ball attach method and solder alloy.