Effect of strike deposition on nanoscale voiding in electrolytic Ni/Au

Article abstract of DOI:10.1149/1.3497293

Nanoscale voiding in electrolytic Ni/Au was studied by varying Au strike deposition conditions. Two kinds of voids were observed in electrolytic Ni/Au when electric current was not applied during strike plating. One was located along the interface between Ni and Au, which could deteriorate the adhesion between Ni and Au. The other was nanoscale bubbles coated with Ni inside Au film, which could result in porous Au layer. Our results suggest that Ni could grow on the surface of hydrogen bubbles and then the bubbles trapped in the Au film. Needlelike Au or Au whiskers were also observed on the Au surface and a line of nanoscale voids was found beneath Au whiskers when electric current was not applied during Au strike process. We propose a model that the nanoscale voiding and Au whisker formation could be related to a galvanic displacement reaction and hydrogen evolution at the early stage of Au electrodeposition.

Full text of DOI:10.1149/1.3497293

D620
Journal of The Electrochemical Society, 157 ͑12͒ D620-D623 ͑2010͒
0
013-4651/2010/157͑12͒/D620/4/$28.00 © The Electrochemical Society
Effect of Strike Deposition on Nanoscale Voiding in Electrolytic
NiÕAu
z
Yeonseop Yu, Jinseok Kim, DoKyung Lim, SeongJae Lee, MiYang Kim,
SangWon Lee, and JongSoo Yoo
Samsung Electro-Mechanics Company, Limited, Chungnam, Korea 339-702
Nanoscale voiding in electrolytic Ni/Au was studied by varying Au strike deposition conditions. Two kinds of voids were observed
in electrolytic Ni/Au when electric current was not applied during strike plating. One was located along the interface between Ni
and Au, which could deteriorate the adhesion between Ni and Au. The other was nanoscale bubbles coated with Ni inside Au film,
which could result in porous Au layer. Our results suggest that Ni could grow on the surface of hydrogen bubbles and then the
bubbles trapped in the Au film. Needlelike Au or Au whiskers were also observed on the Au surface and a line of nanoscale voids
was found beneath Au whiskers when electric current was not applied during Au strike process. We propose a model that the
nanoscale voiding and Au whisker formation could be related to a galvanic displacement reaction and hydrogen evolution at the
early stage of Au electrodeposition.
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3497293͔ All rights reserved.
Manuscript submitted January 25, 2010; revised manuscript received September 14, 2010. Published October 14, 2010.
Electrolytic Ni/Au surface finish has been commonly used in the
Experiments
electronic industry to provide both highly solderable and wire bond-
able surface for ball grid array packages. In particular, soft gold or
pure gold is primarily used for connectors and printed circuits in
In the present study, we followed a typical process sequence for
deposition of electrolytic Ni and electrolytic Au, which includes acid
clean, microetching, acid dip, Ni plating, Au strike, Au plating,
rinse, and drying.
1
-4
which gold wire bonding of integrated circuits is required. Elec-
trolytic Ni is used as a undercoat prior to plating electrolytic Au and
makes a good diffusion barrier for the base copper.
Ni was electroplated on a copper substrate in a nickel sulfamate
bath operated at 50 Ϯ 5°C and a pH of 4.0 Ϯ 0.5. The nickel sulfa-
mate bath contained 450 Ϯ 50 g/l nickel sulfamate, nickel chloride
Electrolytic Ni and electrolytic Au processes employ galvanic
electroplating which requires an applied electric current and a bus
connection incorporated into the circuit design of the printed circuit
1
2 Ϯ 4 g/l, boric acid 25 Ϯ 3 g/l, and proprietary additives. Au
5
strike plating was carried out at 40 Ϯ 5°C and a pH of 4.5 Ϯ 0.5
for about 33 s in an aqueous solution containing 0.75 Ϯ 0.25 g/l Au
board. The electric current increases the deposition rate and higher
current usually provides a denser coating.
͑
as KAu͑CN͒ ͒ with proprietary acid strike additives of JPC ͑Japan
2
In a typical electrolytic Ni/Au process, there is a special plating
process called “Au strike” or “Flash Au.” Au strike process is used
to form a very thin, typically less than 0.1 ␮m thick, plating with
high quality and good adherence to the Ni substrate. Au strike plat-
ing serves as a foundation for subsequent Au plating process and
plays an important role in securing the adhesion of the electroplated
Au over Ni. When a Au film is lifted off the surface of Ni plating
during gold wire bonding applications, the poor adhesion is fre-
quently attributed to the Au strike process.
Pure Chemical, Tokyo, Japan͒. Au was then electrodeposited in a
cyanide bath at 70 Ϯ 5°C and a pH of 6.25 Ϯ 0.15 for about 447 s.
The concentration of Au was 5.0 Ϯ 0.5 g/l as KAu͑CN͒ . Propri-
etary additive Temperesist EX ͑JPC͒ was used.
We deliberately did not apply an electric current during Au strike
plating in the standard electroplating procedure in order to study the
effect of Au strike plating on the microstructure of the electrodepos-
its.
2
The thicknesses of the Ni and Au were measured to be about 8
and 0.35 ␮m, respectively, from the cross-sectional focused ion
beam ͑FIB͒ micrographs. We investigated the microstructure and
composition of Ni and Au electrodeposits using a scanning electron
microscopy ͑SEM͒ and X-ray photoelectron spectroscopy ͑XPS͒.
We also employed a transmission electron microscope ͑TEM͒
equipped with an X-ray energy dispersive spectroscopy ͑EDS͒ de-
tector and a high angle annular dark-field detector ͑HAADF͒. TEM
specimens were prepared by using FIB operated at 30 kV. For these
investigations, we used a Tecnai F20 S-TWIN TEM ͑FEI, Eind-
hoven, The Netherlands͒ operated at 200 kV, a Strata 400S dual-FIB
Au strike plating is usually performed in a bath with low con-
centration of the metal under high current density, which reduces
galvanic displacement reactions. This is because the displacement
reactions can be suppressed by lowering the electrode potential of
metals given by
RT
nF
1
0
E = E −
lnͩ ͪ
͓1͔
n+
M
0
where E is the electrode potential, E is the standard electrode po-
tential, R is the universal gas constant, T is the absolute temperature,
n is the number of electrons transferred, F is the Faraday constant,
͑
FEI, Eindhoven, The Netherlands͒, an S4800 SEM ͑Hitachi, Japan͒,
n+
and M is the concentration of the metal.
and a Quantera XPS ͑PHI, USA͒.
Although the Au strike process is widely used in industry and
known to be responsible for the adhesion between Au and Ni in
electrolytic Ni/Au, we do not completely understand how the very
thin layer of Au strike can have a significant influence on the adhe-
sion between Ni and Au. In the present study, we investigate the
effect of Au strike deposit on the microstructure of electrolytic Ni
and Au to understand the failure mechanism caused by poor adhe-
sion.
Our finding is that improper Au strike plating is closely related to
nanoscale voiding in electrolytic Ni/Au, which results in poor adhe-
sion. On the basis of these results, we propose a model to explain
the mechanism of voiding.
Results and Discussion
We prepared two kinds of electrolytic Ni/Au surface finishes.
One was prepared by the standard procedure described in the previ-
ous section, while the other sample was prepared without applica-
tion of an electric current during Au strike plating.
Figure 1 shows the microstructures of electrolytic Ni/Au ob-
tained from two kinds of samples. We can find striking differences
between two samples, especially at the interface of Au and Ni. When
no electric current was applied during Au strike, nanoscale voiding
occurred at the interface and in the Au layer as marked by white and
black arrows, respectively, as shown in Fig. 1b. Furthermore, the
interface between Au and Ni is relatively rough compared to the
sample prepared by the standard procedure. The voids at the inter-
face do not have specific shape and are connected in some area. In
z
E-mail: yeonseop.yu@samsung.com
Downloaded on 2015-10-26 to IP 139.80.123.48 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑12͒ D620-D623 ͑2010͒
D621
Figure 3. ͑a͒ XPS depth profile obtained from the electrolytic Ni/Au pre-
pared by Au strike plating without an application of electric current. ͑b͒ XPS
spectrum of Ni acquired from the interface of Au and Ni. It indicates that the
Ni at the interface of Au and Ni is not nickel oxide but metallic Ni.
Figure 1. Cross-sectional HAADF STEM micrographs obtained from elec-
trolytic Ni/Au prepared by Au strike plating ͑a͒ with application of an elec-
tric current and ͑b͒ without application of an electric current. The dark areas
marked by the arrows are voids.
The nanoscale voiding at the interface and coexistence of Au and
Ni right below the interface can be explained by galvanic displace-
ment reactions caused by the difference in electrochemical potential
for the Ni and Au as given by
extreme case, a crack develops at the interface. On the other hand,
the voids located in the Au electrodeposits are scattered in the Au
layer in round shape and more frequently observed in the lower part
of the Au layer.
Composition analysis across a void at the interface indicates that
the void is deficient in Ni as shown in Fig. 2b. The void is not
located in the middle of the interface, but right below the interface
of Ni and Au if we assume that the interface is the crossing point of
Ni and Au profiles as shown in Fig. 2. This suggests that Ni atoms
should be dissolved into the plating solution to result in voiding. We
confirm that the void is not filled with organic materials from image
contrast and EDS analysis, which is not presented here.
+
−
Au + e ꢀ Au͑s͒, E° = + 1.83 V
Ni2+ + 2e ꢀ Ni͑s͒, E° = − 0.25 V
͓2͔
−
͓3͔
The substrate Ni can be substituted with Au by the reaction
+
0
0
2+
2
Au + Ni → 2Au + Ni , instead of the electrolytic plating reac-
+ − 0
tion by Au + e → Au as depicted in Fig. 5. The dissolution of
the Ni can lead to etching of Ni beneath the Au layer, which results
in voids and thereby eventually makes the adhesion between Ni and
Au weak.
We may suspect that nickel oxide ͑NiO͒ or nickel hydroxide
The nanoscale voids in the Au layer are also of great interest
since they can give some explanations on the origin of porosity of
͑
Ni͑OH͒ ͒ formation is responsible for the voiding at the interface.
2
However, both XPS depth profile and chemical state analysis of Ni
indicate that Ni is not oxidized but in metallic state at the interface
as shown in Fig. 3.
There are some areas in which both Ni and Au exist at the top
portion of the Ni layer near the interface as shown in Fig. 4a. The
slightly bright particlelike features at the center in Fig. 4a are the
areas where heavier element than Ni is present considering the
Z-contrast of HAADF scanning transmission electron microscope
͑
STEM͒ image. Composition profile across the interface clearly
shows that they are Au-rich areas in the Ni layer right below the
interface. Au can be found as far as about 150 nm below the inter-
face in Fig. 4a. This finding again supports the idea that Ni is dis-
solved into the plating solution and Au is deposited in the area
where Ni is dissolved out when no electric current is applied during
Au strike.
Figure 4. ͑a͒ Cross-sectional HAADF STEM micrograph showing the co-
existence of Au and Ni below the interface of Au and Ni. ͑b͒ Compositional
line profile across the interface of Au and Ni acquired from the line marked
in ͑a͒. Au exists in the bright areas in the Ni layer below the interface.
Figure 2. ͑a͒ Cross-sectional HAADF STEM micrograph showing voids at
the interface of Au and Ni. ͑b͒ Compositional line profile across the interface
of Au and Ni obtained from the line marked in ͑a͒. The region with low Ni
concentration below the interface corresponds to the void located in the
middle of the line marked in ͑a͒.
Figure 5. Schematic that shows the galvanic displacement reaction at the Ni
and Au interface when electric current is not applied during Au strike plating.
The white circles are Au and the gray circles represent Ni.
Downloaded on 2015-10-26 to IP 139.80.123.48 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D622
Journal of The Electrochemical Society, 157 ͑12͒ D620-D623 ͑2010͒
Figure 6. ͑a͒ Cross-sectional HAADF STEM micrograph showing the voids
in the Au layer. ͑b͒ Compositional line profile across the void marked by the
line in ͑a͒. The area with lower Au concentration in ͑b͒ corresponds to the
void in ͑a͒. The locations marked by the arrows in ͑b͒ are the regions where
Ni exist and correspond to the boundary of the void.
the Au electrodeposits. Unlike the voids in irregular shape observed
at the interface between Ni and Au, the voids in the Au layer are in
round shape. They look more like nanoscale bubbles of which size
are less than 50 nm. The elemental analysis by EDS indicates that
Ni exists in the voids in the Au layer as shown in Fig. 6b. Compo-
sition profile across the void in the Au layer clearly shows that the
concentration of Ni is relatively high, especially, at the boundary of
the void as marked by the arrows in Fig. 6b. Interestingly, the void
is surrounded by a thin layer of Ni of which thickness is less than
Figure 7. ͑a͒ SEM micrograph obtained from an electrolytic Ni/Au prepared
by Au strike plating without application of electric current. It shows that
needlelike features or whiskers exist on the Au surface. ͑b͒ Magnified image
of a whisker. ͑c͒ Cross-sectional bright-field ͑BF͒ TEM micrograph showing
a lower part of a whisker. The upper part of the whisker was lost during FIB
milling. ͑d͒ An EDS spectrum obtained from the position in the Au whisker
marked by the white circle in ͑c͒. It shows that the whisker consists of Au.
Cu is detected due to use of a copper TEM grid.
1
0 nm in this particular case.
This observation can be related to the hydrogen bubbles that
6
,7
have been reported by many authors in metal electrodeposition. It
is well known that hydrogen gas evolves at the cathode during elec-
trodeposition as given by Eq. 4 or 5
8
surface as shown in Fig. 7a and b. The size of the whiskers is less
than 200 nm in diameter and sometimes more than 1 ␮m in length.
The image contrast of the bright-field TEM micrograph in Fig. 7c
shows some grain structures in the whisker, and the EDS analysis
indicates that the whisker consists of Au as shown in Fig. 7d. We
can therefore conclude that the whiskers are polycrystalline Au.
Whisker formation in the electrodeposition of many metals such
+
−
2
H + 2e ꢀ H ͑g͒, E° = 0.0 V
͓4͔
͓5͔
2
−
−
2
H O + 2e ꢀ H ͑g͒ + 2OH , E° = − 0.8277 V
2
2
The dissolution of Ni mentioned above is not the only reaction
that can occur at the interface of Ni and Au during Au strike. A small
portion of electrons produced by dissolution of Ni in Eq. 3 and
supplied by the power supply during Au electrodeposition can be
consumed in the generation of hydrogen. The generated hydrogen
gas can form bubbles on the Ni surface. This is a kind of alloy
plating in which hydrogen is the codeposited element.
1
2-14
as Cu, Fe, and Ag have been known for decades.
Whiskers of
such metals were grown by adding some additives and controlling
current densities. However, Au whiskers are very rare and we could
1
5
find only one report by Teverovsky and Sharma from NASA. They
have observed Au whiskers on the Au electrodeposits which were
electroplated over the Ni electrodeposits. Their system is quite simi-
lar to ours. Typical whiskers observed by them have lengths of sev-
eral micrometers and diameters of less than 200 nm, which are a
little bigger than our observation. They have assumed that the whis-
ker growth may be caused by the contamination of Au plating with
rubidium during manufacturing.
Nanoscale hydrogen bubbles in electrodeposits have been rarely
observed, although microscale bubbles in electrodeposition coating
have already been extensively studied due to their significant influ-
9,10
ence on the quality of the metal coating.
Tsai et al. have shown
that electrochemically generated hydrogen gas may lead to the for-
mation of bubbles in electrodeposition and metal can grow directly
9
on gas bubbles, leading to the formation of voids in the coating.
In the present study, we have found some evidences that connect
the Au whisker growth to nanovoiding. Although Au whiskers and
the nanovoids in the Au electrodeposits do not seem to have any
connection at first glance, it turns out that they are closely related.
Figure 8a shows a Au whisker that is a small protrusion on the Au
surface. Interestingly, a line of voids exists beneath the Au whisker
as shown in Fig. 8a and b. Usually nanovoids in the Au electrode-
posits existed at the lower part of the Au layer near the interface
between Au and Ni as shown in Fig. 1b and 4a. However, we found
that aligned nanoscale voids exist beneath the Au whiskers.
Furthermore, the dark area inside of the Au whisker is analyzed
to be deficient in Au as shown in Fig. 8d. This indicates that the Au
whisker has a sort of hollow structure at the bottom. If we combine
two observations that a line of voids exists beneath the Au whiskers
and there are voids at the bottom of Au whiskers, we can draw a
hypothesis that the nanovoids result in Au whisker formation.
From our experimental results, a model for nanoscale hydrogen
bubble formation, the growth of Ni on the surface of the bubbles,
and Au whisker formation is devised as depicted in Fig. 9. This is
Our observation of higher Ni concentration at the boundary of the
voids suggests that Ni may grow on the surface of the hydrogen
bubbles, which agrees well with the results of Tsai et al.
The voids in the Au layer can be explained by combining gal-
vanic displacement reaction and hydrogen evolution during Au
strike. Adopting the explanation of Hsu et al. on hydrogen bubbles
11
and the growth of ramified zinc by electrodeposition, the dissolved
Ni ions in the electrolyte during Au strike are then reduced at the
interface between the solution and the Ni surface, where the hydro-
gen bubbles do not occupy during electrolytic Au deposition. Due to
the hydrogen bubbles on the Ni surface, the lateral growth of Ni and
Au is suppressed, while the growth at the tip is accelerated by the
concentrated electric field during electrolytic Au deposition. Hydro-
gen bubbles surrounded by a thin layer of Ni are then released from
the Ni surface due to their buoyancy. This finally results in inclu-
sions of hydrogen bubbles coated with Ni in the Au layer.
Another interesting observation related to the nanoscale voids in
the Au layer is the needlelike features or whiskers existing on the Au
Downloaded on 2015-10-26 to IP 139.80.123.48 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑12͒ D620-D623 ͑2010͒
D623
Figure 9. Schematic that shows hydrogen bubble formation and the growth
of Ni on the surface of the bubbles. The white circles represent hydrogen
bubbles and the black circles represent Ni. The protrusion on the Au surface
represents a Au whisker.
Figure 8. ͑a͒, ͑b͒ Cross-sectional HAADF STEM micrographs showing Au
whiskers. Au whiskers are the protrusions on the Au surface. There are many
voids below the Au whiskers. ͑c͒ Magnified image of the whisker shown in
͑
a͒ and ͑d͒ Compositional line profile across the Au whisker marked by the
We conclude that the adhesion between Ni and Au and the me-
chanical properties of the electrodeposits in electrolytic Ni/Au are
closely connected to the displacement reaction and hydrogen evolu-
tion because those electrochemical reactions can generate nanoscale
voids at the interface and in the Au film.
line in ͑c͒. The center area of the line is deficient in Au. The concentration of
Ni and Au is higher in the left because the whisker is overlapped with the Au
layer. This can happen since we observe the projected image of a sample in
TEM.
Samsung Electro-Mechanics Co., Ltd. assisted in meeting the publication
costs of this article.
based on galvanic displacement reaction and accompanying hydro-
gen evolution. Hydrogen bubbles form and adhere to the Ni surface.
Ni ions in the plating solution, which are displaced by Au, are de-
posited on the Ni surface which is not covered by hydrogen bubbles.
As Ni electrodeposition proceeds, the hydrogen bubbles are covered
by a thin layer of Ni. The bubbles may move during electroplating
of Au until the Au film becomes dense, which results in permanent
nanoscale voids. If the hydrogen bubbles covered with Ni can reach
to the surface of the Au layer, electric field is concentrated on the
bubbles, which results in Au whiskers.
References
1. I. R. Christie and B. P. Cameron, Gold Bull., 27, 12 ͑1994͒.
2
3
4
5
. Y. Okinaka and M. Hoshino, Gold Bull., 31, 3 ͑1998͒.
. T. A. Green, Gold Bull., 40, 105 ͑2007͒.
. J. Bath, Lead-Free Soldering, 1st ed., p. 250, Springer, Milpitas, CA ͑2007͒.
. C. F. Coombs, Jr., Printed Circuits Handbook, 6th ed., p. 32.13, McGraw Hill
Handbooks, New York ͑2007͒.
6. D. H. Coleman, B. N. Popov, and R. E. White, J. Appl. Electrochem., 28, 889
1998͒.
͑
7
. M. Monev, L. Mirkova, I. Krastev, Hr. Tsvetkova, St. Rashkov, and W. Richtering,
J. Appl. Electrochem., 28, 1107 ͑1998͒.
Conclusion
8
. Cs. Szeles and A. Vertes, J. Phys. F: Met. Phys., 17, 2031 ͑1987͒.
If an electric current is not applied during Au strike plating,
initial galvanic displacement reaction occurs during electrolytic Au
plating. Ni is displace by Au, which results in nanoscale voiding at
the interface between Ni and Au. Furthermore, the displacement
reaction is accompanied by hydrogen evolution which may form
nanoscale bubbles. Ni can grow on the Ni surface which is not
occupied by the bubbles and eventually surrounds the bubbles.
These Ni-coated nanoscale bubbles can act as a seed for Au whisker
formation during electrodeposition of Au.
9. W. L. Tsai, P. C. Hsu, Y. Hwu, C. H. Chen, L. W. Chang, J. H. Je, H. M. Lin, A.
Growo, and G. Margaritondo, Nature (London), 417, 139 ͑2002͒.
0. L. Zhang, Y. Zhang, X. Zhang, Z. Li, G. Shen, M. Ye, C. Fan, H. Fang, and J. Hu,
Langmuir, 22, 8109 ͑2006͒.
1
1
1. P.-C. Hsu, S.-K. Seol, T.-N. Lo, C.-J. Liu, C.-L. Wang, C.-S. Lin, Y. Hwu, C. H.
Chen, L.-. Chang, J. H. Je, et al., J. Electrochem. Soc., 155, D400 ͑2008͒.
2. P. A. van der Meulen and H. V. Lindstrom, J. Electrochem. Soc., 103, 390 ͑1956͒.
3. P. B. Price, D. A. Vermilyea, and M. B. Webb, Acta Metall., 6, 524 ͑1958͒.
4. A. V. Bondarenko, Electrochim. Acta, 29, 887 ͑1984͒.
1
1
1
1
5. A. Teverovsky and A. Sharma, SPIE: Micromachining and Microfabrication, San
Jose, CA, pp. 12–21 ͑2003͒.
Downloaded on 2015-10-26 to IP 139.80.123.48 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).