Materials and Corrosion 2013, 64, No. 1
DOI: 10.1002/maco.201106340
Corrosion under mixed flowing gas conditions of various
connector coatings
M. N. Collins*, M. Reid, E. Dalton, K. Shannon and
L. F. Garfias-Mesias
This paper presents an investigation on the corrosion performance of Ni/Au, NiP/
Au, Ni/NiP/Au and Ni/NiPd/Au electrical connector coating systems. The coating
systems were exposed to 10 days Class III mixed flowing gas and were
subsequently examined by X-ray diffraction, focused ion beam microscopy,
scanning electron microscopy and energy dispersive spectroscopy to evaluate
the performance of each coating system. The Ni/Au coating system showed the
worst performance followed by NiP/Au and Ni/NiP/Au. The Ni/NiPd/Au coated
connector materials exhibited the least surface corrosion and this was
attributed to a number of factors including a thicker coating system, more
compact gold layer with fewer defects and a reduction in the electrochemical
potential difference between layers reducing the local cell effect.
1 Introduction
Reliability of an electronic material or device is defined as the
probability that the package will perform its intended function for
a specified period of time under a given operating condition
without failure. In particular, for most electronic devices, the
reliability of the electronic package and the metallic connectors
are the most important parts of the device, when it comes to
corrosion susceptibility. Most electronic devices and systems use
metallic electrical connections that can react when exposed to an
aggressive environment. Therefore, reliable connectors are
essential to manufacture reliable electronic devices and systems
[1]. Corrosion of the connectors of most electronic systems can
manifest itself in a wide range of consequences ranging from
intermittent electrical faults to complete functional breakdown
[2–5].
The majority of connectors comprise of a copper base alloy
substrate used because of its low cost and high electrical
conductivity. Unfortunately, copper is a very reactive metal to the
atmospheric environment; therefore, a protective coating is
required to prevent corrosion products developing on the surface
M. N. Collins, M. Reid, E. Dalton
CTVR, Stokes Institute, University of Limerick, Limerick (Ireland)
E-mail: maurice.collins@ul.ie
K. Shannon
Southwest Research Institute, San Antonio, TX (USA)
L. F. Garfias-Mesias
Materials & Corrosion Technology Center, Det Norske Veritas (U.S.A.),
Inc., 5777 Frantz Road, Dublin, OH 43017-1386 (USA)
www.matcorr.com
wileyonlinelibrary.com
which can cause intermittent failures. Typically the industry uses
metallic coatings, which consist of a decorative gold top layer with
excellent electrical properties and underneath a barrier layer (in
contact with the copper) that has the main objective of impeding
copper diffusion and improve wear resistance [1–3, 6].
In recent years, the economics of using a thick gold top layer
of approximately 0.8 mm has come under closer scrutiny and a
gold flash (a thin layer less that 0.25 mm of gold) is now being
used for electrical connections [7, 8]. With the introduction of the
gold flash, interest has turned towards the effectiveness of the
barrier layer(s), in order to improve the reliability of connectors
with a gold flash and other barrier layers. The interest in barrier
layer(s) is due to the overriding evidence indicating that using a
gold flash can be detrimental to reliability [7]. Since typical gold
flash coatings are very thin, they tend to have a high level of
defects, namely surface porosity; as a result, pore corrosion is
responsible for most corrosion failures of gold coated connectors
[1–10].
Nickel and some Ni alloys with other noble materials are
typically used as a barrier layer. Although the nickel layer does
impede the copper diffusion from the substrate, it does not
prevent it. Pinnel and Bennett [11, 12] revealed that the diffusion of
copper to the gold surface can take place at relatively low
temperatures and that the nickel layer only slow the diffusion of
copper to the gold surface but do not totally prevent the diffusion
process. The copper diffuses out along the short circuit diffusion
paths, i.e. grain boundaries, dislocations, pores, etc. Alloying the
nickel layer with phosphorus modifies the microstructure of the
nickel layer. Coatings containing 10–13 wt% phosphorus are
mostly amorphous. By removing the grain boundaries, it is
possible to reduce the diffusion paths and anodic sites for
corrosion to initiate [13]. An additional layer of a nickel–
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
7
8
Collins, Reid, Dalton, Shannon and Garfias-Mesias
Materials and Corrosion 2013, 64, No. 1
palladium (NiPd) is used in certain applications, where wear and
corrosion related failures have previously been reported. The
NiPd coatings increase the hardness in those connectors, and as a
consequence it improves the wear resistance and at the same time
decreases the diffusion of copper to the surface [14].
The aim of this paper is to investigate the corrosion
performance of Ni, NiP, Ni/NiP and Ni/NiPd connector coating
systems with a gold flash finish. A Class III mixed flowing gas
(MFG) test was used to evaluate the corrosion resistance of the
different coating systems. This test enables years of service to be
mimicked by days of testing in which the temperature, relative
humidity and concentration of gaseous pollutants are carefully
defined, monitored and controlled [1–3, 6, 15–17]. The Class III
environment represents a particular harsh corrosive environment
in the North America Region and Europe, in which the corrosion
film of pure copper in one year would be between 80 and 400 nm
[15–17].
2 Experimental
2.1 Sample preparation
Pre-stamped strips of copper-zinc alloy with dimensions
25 mm 42 mm 0.50 mm were coated with a specific thickness
of different layers of nickel, nickel–phosphorous, NiPd and gold.
The materials and plating conditions are detailed in Table 1.
2.2 Mixed flowing gas exposure system
Mixed flowing gas exposure was performed at a constant
temperature and relative humidity of 30 8C and 70% RH,
respectively. The corrosive gases used in the chamber were
100 ppb of H2S, 200 ppb of SO2, 200 ppb of NO2 and 20 ppb of Cl2.
Air was used as a carrier gas. The experimental set-up and
settings are detailed elsewhere [14]. Duplicate samples of the
coupons of the different connector coating systems were exposed
to the above MFG conditions, and removed after 10 days exposure
for further detailed analysis.
2.3 Characterization
The samples were characterized before and after the MFG
exposure using optical microscopy, X-ray diffraction (XRD),
scanning electron microscopy (SEM), focused ion beam (FIB)
milling and energy dispersive X-ray spectroscopy (EDS).
The EDS spectra were collected at 20 kV, a beam current
of 0.26 nA and a working distance of 39 mm, which was
Table 1. Plating conditions
Bath characteristics
Temperature
pH-value
Current density
Plating speed
Chloride content
Boric acid content
Ni/Au
NiP/Au
Ni/NiP/Au Ni/NiPd/Au
45 8C
60 8C
60 8C
4
2.6
4
3 A/dm2
2 A/dm2 4 A/dm2
5–10 cm/s 5–10 cm/s 5–10 cm/s
–
–
8 g/L Cl
–
–
45 g/L H3BO3
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
45 8C
6
4 A/dm2
5–10 cm/s
–
–
reproducible to a high degree of accuracy. The spectra were
processed using PGT software. For XRD a filtered Cu Ka
radiation was used and diffractograms were obtained over a
scanning range 20–908 2u with a scan speed of 0.0088 2u/s.
To observe metallographic cross-sections of the coating, the
coatings were protected by epoxy and then cut with a diamond
saw and metallographically prepared to a 1 mm finish.
3 Results and discussion
3.1 As-received contacts
Figure 1a–d shows typical FIB cross-sections of the different
connector coating systems – Ni/Au, NiP/Au, Ni/NiP/Au and Ni/
NiPd/Au, respectively. The Ni/Au connector coating demonstrated a two layer structure comprising approximately 0.1 mm
thick gold outer layer and a nickel under layer. Close examination
of the nickel under layer demonstrates a uniform fine grain size.
The NiP/Au connector coating also demonstrated a two layer
structure comprised of approximately 0.1 mm thick gold outer
layer and a 1.2 mm thick nickel under layer. The nickel under layer
illustrated an amorphous microstructure owing to the P content
in the Ni layer [13].
The Ni/NiP/Au connector coating system comprised
approximately of 0.1 mm thick gold outer layer, a 1 mm thick
NiP intermediate layer and a 1.2 mm Ni under layer.
The Ni/NiPd/Au connector coating system showed a three
layer structure, comprising approximately of 0.1 mm thick gold
outer layer, a 4 mm thick NiPd intermediate layer and a 1.2 mm
nickel under layer.
Figure 2 a–d shows X-ray patterns of the different connector
coating systems. The XRD pattern of the Ni/Au coating system
(Fig. 2a) displays cubic Cu6Zn4 reflections together with the peaks
for the gold layer. The NiP/Au connector coating system displayed
identical reflections to Ni/NiP/Au (Fig. 2b and c). These XRD
patterns demonstrate strong Cu6Zn4 reflections and minor Au
reflections. The XRD pattern of Ni/NiPd/Au connector coating
system (Fig. 2d) showed the cubic Cu6Zn4 reflections together
with peaks of Ni, NiPd and Au. The overall intensity of
the Cu6Zn4 reflections decreased in comparison to the intensity
observed in the other coating systems. This is most likely due to
the presence of the thicker NiPd layer observed in Fig. 1d.
3.2 Analyses of corrosion products and degradation
mechanism
Figure 3a–d shows photographs of the different connector coating
systems – Ni/Au, NiP/Au, Ni/NiPd/Au and Ni/NiP/Au, respectively, after 10 days exposure in Class III MFG environment.
Owing to the sometimes localized nature of the corrosion
processes, the measurements of the corroded areas were
statistically studied using quantitative analysis with digital image
processing aid: ‘‘The Image Processing ToolKit’ 4.000 . The image
analysis was based on reflective light microscopy (using a Leica
Zoom 2000 stereozoom microscope) imaging acquisition. After
image acquisition the corrosion sites were highlighted and its
www.matcorr.com
Materials and Corrosion 2013, 64, No. 1
Mixed flowing gas (MFG) corrosion
Figure 1. FIB cross-section of the as-received connector coating systems: (a) Ni/Au, (b) NiP/Au, (c) Ni/NiP/Au and (d) Ni/NiPd/Au
threshold was calculated to provide a measure of the coverage of
corrosion products.
From the images in Fig. 3a–d it is clear that a significant
spread of corrosion product occurred on Ni/Au, NiP/Au and Ni/
NiP/Au coating systems, demonstrating their propensity to
corrode. The Ni/NiPd/Au, however, did not show the same
degree of corrosion over the entire surface – developing in
isolated areas but not spreading over the surface to the same
extent as observed on Ni/Au, NiP/Au and Ni/NiP/Au coating
systems. The corrosion on Ni/NiPd/Au appears to have
originated at two different sites over the surface. Firstly, as ring
defects periodically spread over the surface and secondly line
defects spreading inwards from the edge of the sample.
Table 2 summaries the average measured area of corrosion
products that developed on the different connector coating
systems.
The images depicted in Fig. 4a–d show high magnification
representative backscattered electron images of corrosion
features on the coating systems Ni/Au, NiP/Au, Ni/NiP/Au
and, Ni/NiPd/Au, respectively, after 10 days exposure in Class III
MFG environment. The original surface of the Ni/Au coating
system is only slightly visible as bright areas in Fig. 4a. The extent
of surface corrosion on the Ni/Au coating system makes it
impossible to determine the location of sites where the initial
corrosion took place. The corrosion products that developed on
the surface of the coating system were analyzed by EDS and
comprised mainly of copper, zinc and sulphur.
www.matcorr.com
On the NiP/Au and Ni/NiP/Au coating systems the corrosion
developed as sharply defined corrosion mounds. The backscattered electron images of NiP/Au and Ni/NiP/Au in Fig. 4b
and d show the development of such corrosion mounds. The
mound consisted of a well know halo and bloom effect [9]. The
bloom at the centre which has grown to a considerable height
above the surface is surrounded by a halo which encloses the
bloom at the centre. The Ni/NiPd/Au coating system in Fig. 4d
shows a backscattered electron image of the tip of a line defect
spreading inwards from the edge of the sample. Creep of the
corrosion product over the coating surface was observed from the
initial defect. Creep corrosion is generally understood as mass
transport process in which solid corrosion products migrate over
a surface [18, 19].
Although the combination of visual, SEM and EDS of the
surface provide valuable results of extent of corrosion on the
different coating systems – metallographic cross-sections were
conducted on each sample in order to get an overall picture of
failure mechanisms and corrosion layer development. Figure 5a–
d shows backscattered electron cross-section images of the
different connector coating systems – Ni/Au, NiP/Au, Ni/NiP/Au
and Ni/NiPd/Au after 10 days exposure in Class III MFG
environment.
Each of the cross-sections show the development of
corrosion products on the surface over a corrosion defect. EDS
analysis (not shown) confirmed the corrosion products that
developed on the surface were mostly compounds containing
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
9
10
Collins, Reid, Dalton, Shannon and Garfias-Mesias
Materials and Corrosion 2013, 64, No. 1
Figure 2. X-ray diffractometer traces of as-received connector coating systems: (a) Ni/Au, (b) NiP/Au, (c) Ni/NiP/Au and (d) Ni/NiPd/Au
sulphur. Over the defect the corrosion product was mainly
comprised of copper, zinc, nickel and sulphur, whereas the area
spreading away from the defect comprised mainly of copper and
sulphur. These results are in good agreement with previous
reports of Cu exposed to MFG III for several days (up to 20 days)
in which the main corrosion products found are sulphur rich [16].
Two sites on the Ni/Au coating system from which corrosion
initiated and corrosion products subsequently migrated across
the surface are shown in Fig. 5a. Since the gold layer is very thin
(approximately 0.1 mm) for these types of metallic coatings, it is
expected that a certain degree of porosity will be present in such
thin layers [8, 19]. The mechanism by which the corrosion occurs
is as follows: the area at the bottom of a defect (within the gold
layer), is oxygen depleted; and the less thermodynamically stable
nickel area becomes anodic with respect to the surrounding Aucoated layer. The gold surface at the top, surrounding the pit, has
access to the oxygen in the air and becomes the cathodic site.
Under the MFG conditions, an invisibly thin but finite, water
layer forms on the gold surface due to the high relative humidity
[20–23]. The combination of corrosive gas dissociates to different
ionic species in the thin water layer [24]. This concentrated
surface solution provides a low conductivity electrolytic path
between the gold and nickel through the surface defect. Corrosion
proceeds because of a local cell effect resulting from a potential
difference between the cathodic gold and the anodic Ni area
exposed. This mechanism is enhanced due to the cathodic
portion of the gold plate having a much larger area than the
minute nickel point at the bottom of the defect (small anode and
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
large cathode effect). Once the nickel is corroded, it exposes the
copper substrate and allows easy transport of the corrosive gases
and water. Corrosion of the substrate occurs at a very rapid rate,
with the corrosion products migrating in all directions over the
gold surface. The extent of creep corrosion observed can be
explained by the chemical species of the corrosion products, in
this case copper and sulphur. Copper sulphide when formed has
significantly higher mobility than copper oxides [19]. In the
complex corrosive environment in this study copper-sulphur
corrosion product is continuously generated from the defects
with extensive migration across the gold surface. The continued
growth of the electrical insulating corrosion products that creeps
over the otherwise uncorroded surface is likely to cause signal
intermittence on connectors and result in device failure.
Underneath the nickel layer, undercutting corrosion of the
copper substrate has occurred, where attack of the copper
substrate under the nickel is observed. The severity of the
undercutting varied. This leads to de-alloying of the zinc, by
which the more active constituent of the Cu–Zn alloy is initially
selectively removed from the alloy, leaving behind a weak deposit
of more noble copper. This effect is notable in Fig. 5b that shows
the NiP/Au coating system, where the Zn is preferentially
corroded. With the subsequent corrosion of the remaining copper
the area under the pit is significantly weakened structurally. This
allows further gases and humidity to penetrate the outer layer and
causes further corrosion. The difference between the two
phosphorous coating systems could possibly be attributed to
the nickel layer. The grain boundaries in the Ni layer may act as
www.matcorr.com
Materials and Corrosion 2013, 64, No. 1
Mixed flowing gas (MFG) corrosion
Figure 3. Photographs of connector coating systems: (a) Ni/Au, (b) NiP/Au, (c) Ni/NiP/Au and (d) Ni/NiPd/Au after 10 days exposure to exposed to
Class III MFG
Table 2. The percentage of area covered with corrosion product on
connector coating systems after 10 days exposure in Class III mixed
flowing gas environment
% Average surface area covered by corrosion product
Ni/Au (%)
NiP/Au (%)
Ni/NiP/Au (%)
Ni/NiPd/Au (%)
81
84
16
99
sites of structural discontinuity (both microstructurally and
chemically). These discontinuities and differences may lead to
preferential corrosion of specific constituents of the nickel layer,
as some grains may be more noble or active than others [10].
The Ni/NiPd/Au demonstrated significantly less pore
corrosion and creep corrosion which may be attributed to a
thicker NiPd layer which will significantly reduce pinholes than
www.matcorr.com
the other three thinner coating systems used in the study.
Additionally, the structure of the flash gold and its properties are
significantly influenced by the surface upon which it is plated.
Typically, a deposited metal will try to mimic the structure of the
surface on which it is plated. The fine microstructure of the NiPd
layer in comparison to the slightly coarser nickel–phosphorous
layer (shown in the FIB cross-sections in Fig. 1b and d,
respectively) may facilitate the nucleation of a more compact gold
layer with fewer defects.
Despite the Ni/NiPd/Au coating system significantly reducing the pore corrosion, cracking of the NiPd layer near the edge
of the coupon allowed penetration to the substrate and
subsequent corrosion. The cracking of the NiPd layer near the
edges may have been a result of stress near the edge of the coupon
causing fracture of the more brittle NiPd layer. Application of a
thinner layer may prove more protective reducing the stress in the
layer and eliminating the stress cracks.
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
11
12
Collins, Reid, Dalton, Shannon and Garfias-Mesias
Materials and Corrosion 2013, 64, No. 1
Figure 4. Backscattered electron images of corrosion products on (a) Ni/Au, (b) NiP/Au, (c) Ni/NiP/Au and (d) Ni/NiPd/Au after 10 days exposure to
exposed to Class III MFG
Figure 5. Backscattered electron cross-section images of (a) Ni/Au, (b) NiP/Au, (c) Ni/NiP/Au and (d) Ni/NiPd/Au after 10 days exposure to exposed to
Class III MFG
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.matcorr.com
Materials and Corrosion 2013, 64, No. 1
4 Conclusions
In this study, four different coating systems (Ni/Au, NiP, Ni/NiP
and Ni/NiPd) with flash gold finish were assessed under MFG
stimulus. It is strongly evident that the Ni/NiPd/Au connector is
superior to the Ni/Au, NiP/Au and Ni/NiP/Au, coating systems in
terms of surface coverage of corrosion. Moreover, the Ni/Au
coating system is inferior to the NiP/Au and Ni/NiP/Au, when
exposed to a MFG environment for 10 days. The superior
performance of Ni/NiPd/Au coating systems was attributed to a
thicker coating system, more compact gold layer with fewer
defects and a reduction in the electrochemical potential
difference between layers reducing the local cell effect. These
results would be of interest to end users concerned with high
reliability of electronic devices.
Acknowledgements: The financial support for this work is
through Science Foundation Ireland under grant number 03/
CE3/I405. The authors are indebted to Franz Gassner at Umicore
Galvanotechnik GmbH, Schwaebisch Gmuend, Germany.
5 References
[1] R. S. Mroczkowski, Presented at the 44th IEEE Holm Conference on Electrical Contacts, 1998, pp. 57.
[2] R. Martens, M. G. Pecht, IEEE Trans. Adv. Pack. 2000, 23,
561.
[3] C. Maul, J. W. McBride, J. Swingler, IEEE Trans. Compon.
Pack. Technol. 2001, 24, 370.
[4] M. Sun, M. Pecht, R. Martens, Scr. Mater. 2000, 42, 1.
[5] R. Martens, M. Pecht, J. Mater. Sci.: Mater. Elecron. 2000, 11,
209.
Mixed flowing gas (MFG) corrosion
[6] M. Sun, M. Pecht, M. A. E. Natishan, Microelectron. J. 1999,
30, 211.
[7] M. Peel, Connector Specifier 1999, 15, 9.
[8] J. Xie, M. Sun, M. Pecht, D. Barbe, J. Electron. Pack. 2004,
126, 37.
[9] M. Reid, J. Punch, G. Grace, L. F. Garfias, S. Belochapkine,
J. Elec. Soc. 2006, 153, B513.
[10] S. J. Krumbein, IEEE Trans. Parts Mater. Pack. 1969, 5, 89.
[11] M. R. Pinnel, J. E. Bennett, Metall. Trans. A 1972, 3,
1989.
[12] M. R. Pinnel, J. E. Bennett, Metall. Trans. A 1976, 7, 629.
[13] R. M. Allen, J. B. VanderSande, Scr. Metall. 2004, 16,
1161.
[14] K. M. Chow, W. Y. Ng, L. K. Yeung, Surf. Coat. Technol. 1998,
105, 56.
[15] W. H. Abbott, IEEE Trans. Compon. Hybrids Manuf. Technol.
1988, 11, 22.
[16] M. Reid, J. Punch, L. F. Garfias, K. Shannon, S.
Belochapkine, D. A. Tanner, J. Electron. Soc. 2008, 155,
C147.
[17] P. Zhao, M. Pecht, IEEE Trans. Device Mater. 2005, 5, 268.
[18] P. Zhao, M. Pecht, Microelectron. Reliab. 2003, 43(5), 775.
[19] D. W. Williams, IEEE Trans. Compon. Hybrids Manuf. Technol. 1998, 11, 36.
[20] J. W. Dini, Electrodeposition, William Andrew Publishing,
Norwich 1993.
[21] T. E. Graedel, J. P. Franey, G. J. Gualtieri, G. W. Kammlott,
D. L. Malm, Corros. Sci. 1995, 25, 1163.
[22] T. E. Graedel, K. Nassau, J. P. Franey, Corros. Sci. 1987, 27,
639.
[23] T. Aastrup, M. Wadsak, M. Schreiner, C. Leygraf, Corros. Sci.
2000, 42, 957.
[24] C. Leygraf, T. E. Graedel, Atmospheric Corrosion, Wiley, New
York 2000.
(Received: August 28, 2011)
(Accepted: October 5, 2011)
www.matcorr.com
W6340
ß 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
13