THE EFFECTS OF PLATING MATERIALS, BOND PAD SIZE AND BOND PAD GEOMETRY ON SOLDER BALL SHEAR STRENGTH

THE EFFECTS OF PLATING MATERIALS, BOND PAD SIZE AND BOND PAD
GEOMETRY ON SOLDER BALL SHEAR STRENGTH

Keith Rogers and Craig Hillman
CALCE Electronic Products and Systems Center
University of Maryland
College Park, MD, USA


ABSTRACT
oxidized. The most common finish has been eutectic tin-
There has been a vast amount of literature published on the lead alloy by hot air solder leveling (HASL) method,
characterization of the shear strength of solder balls attached because it has desirable properties of an ideal PWB surface.
to ball grid array (BGA) substrates. This has allowed for the Unfortunately this coating does not meet the requirement of
establishment of a quasi-defacto specification of 1000 grams- soldering pads planarity; a fundamental factor for fine pitch
force shear strength on packages with 30 mil diameter solder components such as ball grid arrays (BGAs), and contains
balls and 25 mil pads for BGA components. There is lead, one of the toxic metals. Alternatives to the tin/lead
significantly less data on the expected shear strength values of (Sn/Pb) HASL finish include co-planar coatings such as:
the solder ball to printed wiring board (PWB) attachment. This
prevents contract manufacturers and their customers from
• Electroless nickel/immersion gold (ENIG)
benchmarking the robustness of their material sets and reflow
• Immersion matte tin (ImSn)
processes.
• Immersion silver (ImAg)

• Organic solderability preservative (OSP)
An extensive amount of shear testing has been performed to
quantify the performance of the solder ball joint at the PWB, To evaluate the reliability of the BGA interconnections, the
under a variety of materials and designs. These include strength of the solder balls (or solder bumps) attachment not
different plating materials (immersion tin, immersion silver, only to the BGA, but also to the printed circuit board is one
organic solderability preservative and electroless
of the perquisite data. To characterize the integrity or
nickel/immersion gold), plating suppliers and bond pad strength of the solder ball connection at the PWB interface,
dimensions and geometries (round vs. square). In this paper, companies have experimented with varying parameters
solder ball shear strength variation at the PWB interface as a including solder pad size on PWB, solder pad geometry (on
function of some of these parameters is presented, along with PWB), solder ball size, plating bath chemistry, surface
a discussion on the relevant failure types.
finish, cleaning method and storage condition.


Key words: solder ball, ball shear, shear strength
There are two test methods to assess the interconnect

strength of the ball to pad interface on the PWB: (1) solder
BACKGROUND
ball shear testing and (2) solder ball pull testing. Currently,
The present trend towards higher-performance, smaller and
the most popular method to evaluate the interconnect
lighter products with increased functionality has resulted in
strength is the ball shear test. In July of 2000, JEDEC
an increasing demand for smaller component packages
established a standard for BGA tests, JESD22-B117. The
and/or higher pin counts. Due to this, the semiconductor
stated purpose of this test is to determine the ability of the
packaging industry has experienced a shift from traditional
BGA to withstand mechanical shear forces that may be
peripheral leaded devices to area array technology, applied during module manufacturing and handling
specifically ball grid array (BGA) technology. The BGA
operations. The test applies to shear force testing of the
package has become one of the high performance solder balls on the BGA, prior to second level attachment to
semiconductor packages of choice for advanced applications
the PWB. This method can also be used to assess the shear
and is expected to have an increasing share in the strength of the solder ball at the PWB interface as opposed
microelectronics market in the future.
to the BGA substrate interface.


Two disadvantages of using BGAs are inspectability for
It has been observed under combined thermal cycling and
interconnection cracks and individual ball - only peripheral
vibration that the solder ball attachment can fail at either the
balls/columns can be inspected easily, and lack of individual
solder ball to PWB interface or the solder ball to BGA
solder ball re-workability.
substrate interface [1]. Figure 1 shows a photo where there

are separations in the solder ball attachment at both the
To preserve the solderability of PWBs to which the BGAs
component and PWB interface after 82 thermal cycles (-40
are attached to by reflow, it is necessary to protect the
to 125 °C) and vibration (0.10 G2/Hz) testing. The
copper surface mount pads during storage with a solderable
probability of failure at each interface depends on a number
surface finish. This is because the copper pads are easily
of geometric factors including relative size of solder pads, if
421

---

the solder pads are solder mask defined, the shape of the
To assess the strength of the ball attachment to the PWB,
solder attachment at the interface, and the presence of voids
ball shear tests were conducted using a wire pull, ball/die
at the interface.
shear tester (see Figure 3). The shear cartridge used to

evaluate the strengths measures a maximum load of 2
Given that the disadvantages of using BGAs include limited
kilograms at a resolution of 5 grams.
inspection to only peripheral balls and lack of individual

solder ball re-workability, and that separation could occur at
either interface, both interfaces of the solder balls
attachment should be analyzed.


Figure 2: Plot of the time versus temperature shows the
solder reflow profile that was used during the experiments


Figure 1: This photo shows separations in the solder ball
attachment at both the component and PWB interfaces after
82 thermal cycles (-40 to 125 °C) and vibration (0.10
G2/Hz) testing [1].

EXPERIMENTAL PROCEDURES
Most of the previous studies regarding solder shear strength
of solder balls/bumps have characterized the shear strength
of the solder ball/bump attachment at the solder ball to the
component (usually a BGA) interface. Since most of earlier
work focused on the BGA solder connection to the BGA
substrate, CALCE decided to investigate the robustness of
the other interface; the solder ball to PWB interface.

Eutectic solder balls - 63Sn/37Pb - of 25 mils ± 1 mil (635
microns ± 25 microns) were reflowed onto pads of different
sizes, geometries, plating materials, cleaning chemistries,
exposure conditions and laminate manufacturers. The balls
were reflowed using eutectic solder paste applied to the pads
on the PWBs with a stencil designed specifically for each
different test board layout and pad geometry. The stencils
had a thickness of 6 mils.

The reflow conditions were:

• Ramp up time – 180 seconds
• Time above 183oC – 50 seconds

• Max temperature - 210oC
Figure 3: Ball shear test equipment
• Time at max temperature - 10 seconds
• Cool down time - 140 seconds


Shear testing is a method of determining bond strength that
A plot of the time versus temperature for the reflow profile
has been in use for many years. A simple tungsten chisel
is given in Figure 2. This profile was used for all of the
tool is positioned behind the bump and, as the tool moves
tests.
forward at a predetermined shear height and velocity, it
422

---

performs the test (see Figure 4). The peak force reached
during the test is recorded, and the sheared pad or interface
is then examined to determine the failure mode and
mechanism.

The parameters used for the shear testing were a shear
velocity of 100 microns/sec (6mm/min) and a shear tool
height location of 50 microns above the board. This height
is specified in JESD22-B117. The height of the reflowed
solder balls was approximately 630 - 640 microns.


Figure 5: Shear strength of solder bumps, reflowed on
Au/Ni/Cu as a function of the relative shear tip height [2].


Figure 4: Shear setup showing the tungsten chisel tool
placement prior to test initiation
Research by Choi has shown that to compare ball shear
strength results with varying parameters, it is important to
keep the shear velocity constant as well as the tip of the
shear tool lower than 30% of the reflowed solder ball height
[2]. JESD22-B117 specifies that the shear tool height must
be equal to or less than 25% of the reflowed solder ball
height [4]. Choi also shows that shear strength of the solder
ball can decrease if the shear tip, positioned above the
substrate prior to the solder ball shear test, is higher than
half of the reflowed solder ball height (see Figure 5).

Tests by Choi have also revealed that shear strength is
approximately linearly related to shear velocity; increasing
as the shear velocity increases (see Figure 6). Since the
shear strength varies with shear tool tip height and shearing
velocity, it is imperative that these two parameters be kept
constant, to evaluate the effect of the other factors under
investigation.

For each parameter variation investigated, a minimum of 96
solder balls were reflowed and shear tested. Four balls were

reflowed on each of at least 28 test boards, one ball on each
Figure 6: Shear strength of solder bumps, reflowed on
corner (see Figure 7).
Au/Ni/Cu as a function of the shear velocity [2]




423

---

A. Solder balls attached to 270 µm diameter pads


Figure 7: This photo shows an example of one of the test
board designs, with each of the four reflowed solder ball
locations indicated.
EFFECT OF PARAMETER CHANGES ON SHEAR
STRENGTH
Solder pad size
Test boards that were similar except for solder pad diameter
(sizes of 270 and 475 microns) were used to evaluate the
effect of pad size on shear strength. After the solder balls

(25 mil diameter) were reflowed, the size of the reflowed
B. Solder balls attached 475 µm diameter pads
balls on the larger pads was notably larger than those on the
Figure 8: These photos show the difference in reflowed
smaller pads due to a greater volume of solder paste
solder ball size, due to different amounts of solder paste
adhering to the larger pads (see Figure 8)
stenciled unto the respective pads.

Comparing the area of the pads with different diameters

reduces to the ratio of the square of their diameters:
2
A
?r
4752
2
2
=
=
= 3.1
2
A
?r
2702
1
1

The ratio of the areas of these pads is 3 to 1. The plot of
shear strength values for these pad diameter sizes also
shows an average of ~3 to 1 (~900 vs. 300 grams of force,
see Figure 9). This suggests that shear strength is linearly
proportional to wetted contact area between the solder ball
and the pad.

Pad geometry
Two boards, with all parameters equal except for the pad
geometry, were designed so that the effect of the pad
geometry on solder ball shear strength could be evaluated.
One design had the typical circular pads while the other had

square pads, but with equals areas (see Figure 10). The
Figure 9: Shear strength variation for two different pad
diameter of the circular pads was 650 microns (area of 0.33
sizes, averaging approximately 300 and 900 grams of force
mm2), while the dimension of the length and width for the
for the 270 and 475 micron diameter pad sizes respectively.
square pads was 575 microns (area of 0.33 mm2).

424

---

The elemental analysis showed that there may be a thicker
gold layer over the nickel in sample A, as compared to
sample B, since the electron beam penetrates through the
outer layer of gold before hitting the nickel at the same
operating kV. For a given material, the depth of electron
penetration into the sample is proportional to the operating
voltage. In sample C, the organic solderability preservative
over bare copper, the electron beam penetrates through the
coating, allowing copper to be detected. A small amount of
chlorine, a possible contaminant is also detected during the
elemental analysis.

A. Circular pad design layout


E-SEM image of surface plating (sample A)

B. Square pad design layout
Figure 10: Optical photos show the two designs with equal
pad area; the circular pads in shown A, while the square
pads in shown in B.
Results of the shear testing showed that the geometry of the
pad had a negligible effect on the shear strength values. The
average shear strength value for the circular pads was 1.73
kilograms compared to 1.80 kilograms for the square pads.
This difference is around 4%.


Plating materials
EDS spectrum of surface plating (sample A)
The effect of plating materials on shear strength was
characterized by testing reflowed solder balls on three
Figure 11: E-SEM image at 5000x and EDS spectrum for
surface finishes. The surface platings investigated were (A)
the surface plating on sample A
ENIG - electroless nickel/immersion gold, (B) PNS - post

nickel strike/immersion gold and (C) OSP – organic
solderability preservative over bare copper.
Table 1: Compositional analysis of plating

Elemental weight percent
Prior to reflow, an energy dispersive spectral (EDS) analysis
Sample
Au Ni Cu Cl Si O
and a 5,000x magnification environmental scanning electron
microscope (E-SEM) image were acquired for each of the
A
39.6
60.4
- - - -
three surface finish types. The E-SEM was operated at 20
kV for both the imaging and elemental analysis of the three
B
31.1
68.9
- - - -
surface finishes (see Figure 11, Figure 12, Figure 13 and
C
- - 92.3 1.5 2.2 4.0
Table 1).
425

---

150oC are given in Table 2. The results show that except for
the organic solderability preservative, the finishes show
virtually no degradation after 8 hours of exposure to 150oC.


E-SEM image of surface plating (sample B)


E-SEM image of surface plating (sample C)


EDS spectrum of surface plating (sample B)
Figure 12: E-SEM image at 5000x and EDS spectrum for
the surface plating on sample B

A plot comparing the shear strengths of the balls reflowed
EDS spectrum of surface plating (sample C)
onto the three surface finishes is given in Figure 14.
Figure 13: E-SEM image at 5000x and EDS spectrum for

the surface plating on sample C
Stabilization bake
According to MIL 883, Method 1008.2, a stabilization bake

may be used to determine what effect storage at elevated
temperatures could have on microelectronic devices without
Table 2: Comparison of shear strength values for the
electrical stress applied. This test may also be used in a
various plating types and high temperature exposure
screening sequence or as a preconditioning treatment prior

Ave shear strength (kilograms)
to the conduct of other tests. To evaluate the effect of
Sample
As plated
8 hours at 150oC.
storage conditions on the reliability of the solder joints,
ImAg
1.46 1.50
several finishes were subjected to an 8-hour bake at 150oC.
OSP
1.45 1.30
The shear strengths of reflowed solder joints on the exposed
ImSn
1.45 1.42
surfaces of these boards were evaluated and compared to a
HASL
1.56 1.56
similar set of boards, which received no exposure. The
ENIG
1.55 1.58
solder balls were attached after the high temperature
exposure. The finishes evaluated were Immersion silver
Manufacturing process
(ImAg), Organic solderability preservative (OSP), Boards plated with immersion tin from five manufacturers
Immersion tin (ImSn), hot air solder leveling (HASL) and
were characterized to assess the difference in ball shear
Electroless nickel/ immersion gold (ENIG). The results of
strength as a function of the manufacturing process after
the shear strength values comparing the various plating
reflow. The shear strength results for the five companies
finished in the “as-plated” samples to those exposed to
represented as U, V, W, X, Y and Z are shown in Figure 15.
426

---

The four significant failure modes are:

• Ball shear – solder ball fracture through the bulk
solder. In this ductile fracture, the pad is still
covered with solder after the ball has been shear
off. This is the expected failure mode and is
typical of a robust solder joint.
• Intermetallic fracture – fracture interface is at the
intermetallic diffusion layer. This is a brittle
fracture with minimal distortion of the solder ball
and a clean separation of the solder ball at the pad
surface.
• Pad lift – solder pad lifts off the substrate. This
occurs when the peel strength of the copper pad is

lower than the bulk flow stress of the solder.
• Ball lift – ball lifts off from the pad. This is due to
Figure 14: Plots of the shear strength of the solder balls on
insufficient wetting. The pad plating is visible
the three surfaces is shown. The PNS has the lowest average
after the ball is sheared off.
strength, but also the smallest variability. The OSP finish

has the largest average shear strength but also has the largest
Acceptable failure modes are solder ball shear, intermetallic
variability.
fracture and pad lift, for which acceptable shear force
requirements have been met [4]. The other three are

unacceptable modes of failure (ball lift, solder ball sheared
above center line and tool cutting into substrate before the
solder ball is sheared).

In the tests conducted for this study, virtually all of the
solder ball failures were due to the ball shear failure mode
where the fracture is through the bulk solder. The other few
cases were due to pad lift, with shear strengths all above 1.5
kilograms.

CONCLUSIONS
Given the drawbacks of using BGAs, limited inspectability
and lack of individual reworkability, both interfaces of the

solder joint (at the component and the PWB) should be
characterized to ensure strong interfacial bonds. If there is
Figure 15: Plots of solder ball shear strength for boards
separation at either interface, the part could lose
plated with immersion tin for five different manufacturers
functionality.
The plot shows that manufacturers X and Y had lower shear

strengths values than V, W and Z. There is also more
Investigations with varying parameters suggest the
variability in manufacturers V and W, compared to Z. The
following:
average shear strength values for the five manufacturers

were 1.73, 1.89, 2.06, 2.08 and 2.10 kilograms for
1. Solder ball shear strength is linearly related to
companies V, W, X, Y and Z respectively. There is a
bonded surface area; solder pad geometry does not
difference of 3,700 grams of force from the weakest to the
appear to affect the shear strength if the pad sizes
strongest.
areas are equal

2. For ImAg, ImSn, ENIG and HASL, high
FAILURE MODES
temperature storage at 150oC for 8 hours did not
There are basically six different failure modes, two of which
cause any significant change in shear strength; less
can be attributed to setup errors. These errors due to setup
than 3%. For the OSP, the same conditions caused
are from incorrect settings of the shear tool. Setting the tool
a 10% reduction in shear strength.
height too high may result in shearing off a section of the
3. The variability in shear strength from different
ball above its center line, while too low of a setting may
manufacturers of the same surface finish can
allow the tool to cut into the substrate prior to contact with
exceed that of different finishes from one source.
the solder ball.


ACKNOWLEDGEMENT

Thanks to Mr. Sarju Patel at Capital Electro-Circuits, Inc. in

Gaithersburg Maryland for the use of his reflow equipment.
427

---

REFERENCES
[1] H. Qi, A. Ganesan, M. Osterman, and M. Pecht,
“Accelerated Testing and Finite Element Analysis of
PBGA Under Multiple Environmental Loadings,” 2004
International Conference, IEEE Business of Electronic
Product Reliability and Liability, pp. 99-106, April 27-
30, 2004.

[2] Choi, Jin-Won, Choi, Jae-Hoon and Oh, Tae-Sung,
“Ball Shear Strength of 63Sn-37Pb Solder Bump with
Test Conditions,” Department of Metallurgical
Engineering and Materials Science, Hong-Ik
University, Seoul, Korea

[3] Li, M., Lee K. Y., Chen W. T., Tan B. T., and
Mhaisalkar S, “Microstructure, Joint Strength and
Failure Mechanisms of SnPb and Pb-Free Solders in
BGA Packages,” IEEE Transactions on Electronic
Packaging, Vol. 25, No. 3, July 2002, pp. 185-192.

[4] JEDEC, BGA Ball Shear: JESD22-B117, July 2000.

[5] Hung, S. C., Zheng, P.J., Lee, S. C., and Lee, J. J., “The
Effect of Au Plating Thickness of BGA Substrates on
Ball Shear Strength Under Reliability Tests,”
Proceedings of the IEEE/CMPT International
Manufacturing Technology Symposium, 1999, pp. 7-
15.

[6] Erich, R., Coyle, R., Wegner, G., and Primavera, A.
“Shear Testing and Failure Mode Analysis for
Evaluation of BGA Ball Attachment,” Proceedings of
the IEEE/CMPT International Manufacturing
Technology Symposium, 1999, pp. 16-22.

[7] Morawska, Zofia and Koziol Grazyna, “Lead-Free
Solderability Preservative Coatings of PCBs,”
Advancing Electronics, Volume 28, No. 3, May/June
2001.
428

---

Document Outline

• MAIN MENU
• PREVIOUS MENU
• ---------------------------------
• Search CD-ROM
• Search Results
• Print

---