Mechanical shock causes BGA brittle fracture

In the field of BGA packaging, the use of lead-free solder is becoming increasingly popular, especially
in portable devices. This packaging method can easily cause brittle fracture failure at the interface
between the solder ball and the pad when subjected to mechanical impact. A brittle fracture at the
interface between the solder ball and the package substrate bonding pad is considered an unacceptable
defect.

In theory, the reliability of solder joints is characterized by board-level drop
testing, however, this test has several flaws. Each drop test consumes many packages
and hundreds of solder joints, resulting in significant cost increases. Additionally, unless equipped
with a high-speed, real-time data acquisition system, cracks in the solder joint may close after impact,
making potential failures undetectable. Finally, the data analysis process is time-consuming and costly.
Therefore, there is an urgent need to find an alternative method to evaluate solder joint integrity
under mechanical shock-loading conditions.

A study comparing high-speed soldering testing to board-level drop testing of BGA packages using
lead-free solder balls and different package substrate surface finishes suggests the former may be a
viable option.

Various BGA package structures were studied using different combinations of solder alloys, surface
finishes, substrate materials, solder ball sizes, and package dimensions. Typical test devices were 316
PBGA (27 mm × 27 mm) structures using Sn 4.0%/Ag 0.5%/Cu (SAC 405) solder balls and various substrate
surface finishes including electroless nickel immersion gold (ENIG) and organic solderability
preservatives (OSP).

316 PBGA samples use standard 0.76 mm diameter spheres. The package substrate consists
of BT laminate with a thickness of 0.36 mm. The solder mask pad is defined by the mask layer and has an
opening diameter of 0.635 mm. The solder balls are attached to the substrate in a hot air convection
reflow oven with a preheated lead-free soldering profile of 150°C±2°C, with a peak reflow temperature of
260°C.

The samples were divided into groups and thermally aged at 125°C to promote intermetallic compound (IMC)
formation at the package substrate/solder joint interface. The speed range for high-speed ball shear
testing is from 10 mm/s to 3,000 mm/s, and the speed range for high-speed ball tensile testing is from 5
mm/s to 500 mm/s. An advanced high-speed bond tester is used, equipped with control and analysis
software and next-generation force sensors capable of evaluating the fracture energy of solder balls in
ball shear and ball pull tests.

After the drop test, the brittle fracture surface was analyzed (Sn4.0%Ag0.5Cu+OSP, aged for 500 hours).
The second part of the study involved board-level drop testing, recording resistance, board strain, and
fixture acceleration. A detailed analysis was performed to identify failed solder joints and
corresponding failure modes. Failure modes and loading rates from solder ball shear and pullout tests
were compared with mechanical drop tests. Likewise, the energy absorption values recorded during solder
ball shear and pull tests are considered to be effective indicators for explaining solder joint failure
modes.

To accelerate the growth of IMC, thermal aging treatment was performed, with the temperature set to
125°C and the durations of 100, 300, and 500 hours respectively. After thermal aging, some PBGA samples
with solder balls were shaped, cross-sectioned, and etched, then inspected and analyzed by scanning
electron microscopy (SEM). Similar BGA samples were assembled on a test board and dropped-tested using a
dual-rail guide. As mentioned above, some board-level test samples also experienced thermal aging. All
samples are daisy-chained and monitored for real-time data collection.

Detailed SEM analysis on two complementary surfaces of brittle fracture failure in shear and tension
test specimens revealed striking similarities to the fracture interfaces generated during board-level
drop testing and high-speed shear and tension testing. The results show that brittle fracture obtained
in high-speed bond testing is a strong indicator of board-level drop test behavior. A notable feature of
the current work is the effort that resulted in a direct comparison of the physics of brittle failures
in high-speed bonding testing and board-level drop testing.

Drop Test

For the board-level drop tests conducted, the test boards were manufactured with both non-solder
mask-defined (NSMD) and solder mask-defined (SMD) pad geometries. In both cases, the pad diameter was
0.684 mm, dipped in solder. While NSMD is more common in actual production boards, the advantage of SMD
is its correlation study, which is the location of board-level drop test fractures that are more likely
to occur on the package side. This is critical because shear/pull testing of solder balls only evaluates
fractures on the package side, as the component is not connected to the PCB. The test board assembly
exhibited brittle fractures on the solder joint surface during the drop test.

The summary of the test board assembly results in the drop test (8 components per data point) shows that
the thermal aging of the OSP package substrate surface coating caused it to degrade faster than the OSP
package substrate surface coating with ENIG coating.

High-speed Bond Testing

For high-speed bond testing, Shenzhen Highqualitypcb SMT Processing: Brittle fracture failures of
ENIG surface finish specimens usually occurred between the IMC and Ni layers. For the unaged OSP
specimens reflowed twice, brittle fracture failure was observed between the Cu 6 Sn 5 IMC and the Cu
layer. Brittle fracture failure for the OSP specimens after heat aging occurred between the Cu 6 Sn 5
and Cu 3 Sn IMC phases. Brittle fractures were observed on the surface of the solder joints subjected to
high-speed shear and pull tests, as shown in Figure 2.

Previous evaluations of high-speed solder ball shear and pull tests have observed brittle fractures with
an appearance similar to the brittle fracture patterns observed in board-level drop test assemblies, but
there has been little definitive cross-sectional evidence. This is due in part to the difficulties of
such studies, both in terms of acquiring individual sheared or pulled balls and matching them to
corresponding pads and in terms of subsequent cross-sectional work.

Direct Correlation

In addition to the microstructural correlations, strong correlations were observed between the test
parameters used for high-speed bond testing and board-level drop testing. The mathematical correlations
associated with solder ball shear/pull and drop test results are complex, and these graphs relate the
percentage of brittle fractures from shear and pull solder ball tests to the drop to failure for the
specific package and drop test conditions used in this study. In short, the graphs are derived by
plotting the drop value at each time point against the equivalent data from the shear or pull test, and
then performing a power law curve fit. Each curve corresponds to one solder ball shear or pull test
speed.