Mechanism of stabilizer acceleration in electroless nickel at wirebond substrates

Article abstract of DOI:10.1149/1.1617307

This paper examines the mechanism of stabilizer concentration at electroless nickel wirebond electrode. A one-step activation protocol was first achieved on copper substrates using acetic acid and dimethylamineborane. Thereafter nickel multilayers were grown onto the substrates using nickel sulfate heptahydrate as the source of nickel, sodium hypophosphite as the reducing agent, acetic acid as the complexing agent, and thiourea/lead acetate as the stabilizing agent. The morphology of the nickel layers and the effective concentration of the stabilizer were determined using quartz crystal microbalance, gravimetric techniques, and energy-dispersive X-ray analysis (EDX). The plating rate was obtained by measuring the thickness of the Ni-plated using X-ray fluorescence spectroscopy. It was found that lead acetate completely inhibited the plating, and a bimodal distribution was observed as the concentration of the thiourea was varied. We proposed a mechanism for the effect of stabilizer in electroless Ni baths. This mechanism was confirmed using Fourier Transform Infrared and EDX measurements.

Full text of DOI:10.1149/1.1617307

C816
Journal of The Electrochemical Society, 150 ͑11͒ C816-C822 ͑2003͒
0013-4651/2003/150͑11͒/C816/7/$7.00 © The Electrochemical Society, Inc.
Mechanism of Stabilizer Acceleration in Electroless Nickel
at Wirebond Substrates
,z
*
Hongwu Xu, Juan Brito, and Omowunmi A. Sadik
Department of Chemistry, State University of New York at Binghamton, Binghamton,
New York 13902-6000, USA
This paper examines the mechanism of stabilizer concentration at electroless nickel wirebond electrodes. A one-step activation
protocol was first achieved on copper substrates using acetic acid and dimethylamineborane. Thereafter nickel multilayers were
grown onto the substrates using nickel sulfate heptahydrate as the source of nickel, sodium hypophosphite as the reducing agent,
acetic acid as the complexing agent, and thiourea/lead acetate as the stabilizing agent. The morphology of the nickel layers and the
effective concentration of the stabilizer were determined using quartz crystal microbalance, gravimetric techniques, and energy-
dispersive X-ray analysis ͑EDX͒. The plating rate was obtained by measuring the thickness of the Ni-plated using X-ray fluores-
cence spectroscopy. It was found that lead acetate completely inhibited the plating, and a bimodal distribution was observed as the
concentration of the thiourea was varied. We proposed a mechanism for the effect of stabilizer in electroless Ni baths. This
mechanism was confirmed using Fourier Transform Infrared and EDX measurements.
© 2003 The Electrochemical Society. ͓DOI: 10.1149/1.1617307͔ All rights reserved.
Manuscript submitted March 24, 2003; revised manuscript received June 2, 2003. Available electronically October 2, 2003.
Electroless plating is widely used in a variety of applications
dimethylamineborane DMAB solutions at gold surfaces. Using mass
sensors, we have correlated the bath chemistry with the overall plat-
ing quality on industrial wirebond samples.8-10 We also tested the
ability of the sensors to monitor the bath chemistry during plating
using ten different bath formulations, and have these correlated plat-
ing rates with thickness on industrial laminate chip carrier ͑LCC͒
wirebond samples.9 In this work, we have investigated the effective
concentration of the stabilizer as well as examined the mechanism of
the stabilizer regeneration in the electroless plating deposition of Ni
on Cu/Ni substrate. The CSC was studied at relatively low ͑ϳ0.01
ppm͒ concentration of thiourea and a regeneration mechanism is
proposed.
such as electronics packaging, jewelry making, and as a suitable
finish for mounting chips to chip carriers. Nickel electroless deposits
have been reported to exhibit high corrosion resistance in various
media.1 Commercial electroless nickel solutions contain several
components including nickel salts, chelating agents, buffers, accel-
erating agents, stabilizers, and pH adjusters. Each component serves
a specific purpose and the balance of each species must be main-
tained throughout the bath lifetime. During the deposition process,
two main reactants must be continually replenished. These are the
nickel ions and the reducing agent. Much effort is going into study-
ing means of prolonging the lifetime of electroless nickel baths,
removing any by-products and replacing toxic components.1
The protective coatings of Ni deposits significantly depend on
the stabilizers added to the electroless nickel plating baths, pH, and
the substrate type.2,3 Spontaneous decomposition of electroless plat-
ing baths can be virtually eliminated or controlled by the addition of
a small amount of stabilizers or inhibitors. Stabilizers are chemical
agents in electroless plating baths required to prevent the homoge-
neous reaction that triggers the spontaneous decomposition of an
entire plating operation.1,2 The catalytic nature of the substrate to be
coated can be significantly modified using trace concentrations of
these stabilizers. At low stabilizer concentrations, substrate activity
can be significantly enhanced to produce an increase in deposition
rate. At low thiourea concentration, metal dissolution occurs. Be-
yond a critical level, the stabilizer may poison the substrate, thus
completely inhibiting the catalytic activity. Consequently, it is im-
portant to determine the effective concentration of the stabilizer for
specific bath chemistry.
The determination of the effective stabilizer concentration has
been achieved mainly by trial and error. Some of the methods that
have been reported for determining this parameter include polariza-
tion techniques and the measurement of deposition rate vs. stabilizer
concentration.4-6 Another is the measurement of the time prior to
visible formation of black precipitate by adding potassium chloride
solution.7 Using any of these methods, the effective or critical sta-
bilizer concentration was reported to vary from 0.1 ppm to several
milligrams/liter.4-7 A very sharp maximum is usually recorded by
measuring the deposition rate vs. stabilizer concentration. The criti-
cal stabilizer concentration ͑CSC͒ is indicated by the cessation of
the plating reaction and this is evidenced by the loss of hydrogen
evolution.
Experimental
Reagents.—All reagents were of analytical grade and were used
as received except otherwise stated. NiSO4•7H2O, NaH2PO2•H2O,
lead acetate trihydrate, thiourea and DMAB were obtained from
Acros, Inc. ͑Pittsburgh, PA͒, and acetic acid ͑HAc͒ from Sigma
Chemicals ͑St. Louis, MO͒. All solutions were prepared using
deionized ͑DI͒ water with resistivity of 17 M⍀ or higher. The QCM
measurement was performed at 80°C. The temperature was kept
constant with a Haake ͑Lachat Instruments, Milwaukee, WI͒ water
bath. The quartz crystal worked well at this temperature.
Instrumentation.—All microgravimetric quartz crystal microbal-
ance ͑QCM͒ experiments were performed using QCA 917 quartz
crystal analyzer ͑Seiko EG&G͒. An open-circuit system was used
for QCM measurements in which only the gold-coated, quartz crys-
tal working electrode was connected, but in the absence of auxiliary
and reference electrodes. The Au-coated quartz crystals ͑AT-cut, 9
MHz͒ having 0.2 cm2 geometric area per face were obtained from
EG&G Instruments ͑Princeton Applied Research͒. The resonant fre-
quency was determined using QCA 917 quartz crystal analyzer
͑Seiko EG&G͒. The electrodes consisted of 1000 Å of gold film on
a polished quartz with a texture of р1 ␮m and a 50 Å chromium
adhesion layer between the electrode and the quartz. The surface
morphology and qualitative analysis of the nickel deposits were de-
termined using scanning electron microscope ͑SEM, Phillips-
Electroscan, model 2020͒, which is equipped with a Link ISIS en-
ergy dispersive X-ray analysis ͑EDX͒ analyzer.
Previously, we have studied the interaction of the reactive,
boron-containing intermediates, which were generated in situ from
Bath composition.—The starting composition and concentrations
of the electroless Ni plating baths that we studied included: 6 g/L
Ni2ϩ, 30 g/L NaH2PO2•H2O, 30 g/L HAc, and 0.05 ppm thiourea.
When the effect of one component was investigated, its concentra-
tion was varied while other parameters were kept constant. NH4OH
* Electrochemical Society Active Member.
z E-mail: osadik@binghamton.edu
Downloaded on 2015-06-25 to IP 138.38.0.53 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, 150 ͑11͒ C816-C822 ͑2003͒
C817
Table I. Composition and operating conditions for nickel hypo-
phosphite bath at variable stabilizer concentrations.a
Table II. Composition and operating conditions for nickel hypo-
phosphite bath at variable nickel ion concentrations.a
Bath constituents
g/L
Bath constituents
g/L
6.0
30.0
Variable ͑1-12͒
30.0
NiSO4•7H2O
NaH2PO2•H2O
Thiourea
Acetic acid
pH
NiSO4•7H2O
NaH2PO2•H2O
Thiourea
Acetic acid
pH
Variable ͑0-2 ppm͒
0.05ppm
30.0
5.1
30.0
5.1
Temperature
82°C
Temperature
82°C
a No plating was recorded when lead acetate was used as bath stabi-
lizer.
a No plating was recorded when lead acetate was used as bath stabi-
lizer.
was used to adjust pH to 5.1. The bath temperature was kept at 82°C
and the plating time lasted 30 min. Tables I, II, and III provide a
summary of the compositions of all baths studied.
plated. The exact plating rate was obtained by measuring the thick-
ness of the Ni that was plated using X-ray fluorescence spectroscopy
͑XRF͒.
Electroless deposition.—In this study, we used utilized LCC
wirebond substrates from IBM Microelectronics, Endicott, NY. The
LCCs were circuitized with copper, and were used as the substrates
for the electroless Ni plating. Because Cu could not catalyze the
electroless Ni plating in a sodium hypophosphite bath, it was nec-
essary to first activate the copper substrate. Typically, copper is sen-
sitized using Sn͑II͒ and then activated with Pd͑II͒. Feldstein et al.11
suggested a one-step activation method for metal or alloy ͑e.g., Cu͒
substrates by depositing small amounts of Ni onto the metal sub-
strate using a Ni-DMAB developer solution. The reaction involved
Cu-catalyzed reduction of Ni2ϩ by DMAB. The resulting Ni layer
further catalyzed the reduction of the Ni in a Ni-P bath. The devel-
opment step required a higher temperature ͑ϳ40°C͒ and a Ni-
DMAB developer solution. To simplify the procedure, we made use
of only a 2-g/L DMAB solution in water at room temperature. The
LCC substrate was soaked in 10% H2SO4 solution for 10 min to
remove any rust or oxide, and thereafter, we rinsed and soaked the
substrate in 2 g/L DMAB solution for 30 min. After this step, the
activated surface was dipped into the appropriate bath to initiate
electroless plating. On copper surface, Ni͑II͒ was reduced into metal
Ni by DMAB with Cu acting as the catalyst and the deposited Ni
further catalyzed the reduction of Ni͑II͒ by sodium hypophosphite in
the Ni-P baths.
X-ray fluorescence spectroscopy.—The thickness of the plated Ni
was measured using Fisherscope X-ray system XDVM ͑Fisher
Technology, Windsor, CT͒. The sample was positioned under the
locating light and the thickness was measured by the software-
controlled system.
Results and Discussion
In acidic medium electroless Ni plating involves the reduction of
free Ni2ϩ and Ni͑Ac)2 ͑when HAc was used as the complexing
agent͒, as well as the oxidation of hypophosphite in which the latter
step was considered as rate determining.12 These steps can be rep-
resented as follows
2H2POϪ2 ϩ 2H2O → 2H2PO3Ϫ ϩ 2Hϩ ϩ 2HϪ
Ni2ϩ ϩ 2HϪ → Ni ϩ H2
͓1͔
͓2͔
Ni͑Ac)2 ϩ 2HϪ → Ni ϩ H2 ϩ 2AcϪ
͓2Ј͔
Thus the overall Ni plating reaction in acidic condition is often
represented as
Ni2ϩ ϩ 2H2POϪ2 ϩ 2H2O → Ni ͑metallic) ϩ 2H2POϪ3 ϩ 2Hϩ
Elemental composition and FTIR analysis.—Energy dispersive
spectrometry ͑EDS͒ was conducted to investigate the composition of
the plated Ni film. EDS was performed using a Princeton Gamma-
Tech IMIX X-ray microanalysis system ͑NJ͒. Fourier transform
infrared ͑FTIR͒ analysis was determined with a Nicolet Magna-IR
760 spectrometer ͑Madison, WI͒ using Kapton substrate because
the Ni was easier to strip off using this substrate. The Kapton was
sensitized in a similar manner as the LCC before being soaked in
Ni bath for 2 min. The substrates were dried and the plated Ni
layers stripped off, mixed with KBr and pressed to a pellet for FTIR
analysis.
ϩ H2
͓3͔
Ni͑Ac)2 ϩ 2H2POϪ2 ϩ 2H2O → Ni ͑metallic) ϩ 2H2POϪ3
ϩ 2Hϩ ϩ H2 ϩ 2AcϪ
͓3Ј͔
Plating is accompanied by an obvious production of hydrogen gas,
which may be inhibited through the addition of stabilizer ͑such as
thiourea͒. In that case, the stabilizer could be adsorbed onto the
surface of the substrate. This could decrease the number of catalytic
sites and depress the hydrogenation/dehydrogenation catalysis.12
This would inherently decrease the plating rate and subsequently
QCM.—The plating of Ni onto a QCM was achieved using the
QCM setup that was described above. In this setup, a layer of Ni
was electroplated onto the gold surface of the QCM, and the whole
QCM cell was immersed into the electroless Ni plating solution at
80°C. The increase in mass ͑or frequency decrease͒ was recorded
using the QCM analyzer. The plating time was 10 s for each bath
composition.
Table III. Composition and operating conditions for nickel hypo-
phosphite bath at variable reducing agents’ concentrations.a
Bath constituents
g/L
6.0
Variable ͑5-50͒
0.05 ppm
30.0
NiSO4•7H2O
NaH2PO2•H2O
Thiourea
Acetic acid
pH
Gravimetric analysis.—The determination of the microscopic
mass using QCM was correlated with classical gravimetric tech-
niques. An LCC sample was first weighed, and then soaked in a 10%
H2SO4 solution to remove oxide or rust. This was further rinsed
using DI water before being immersed into the plating solution. The
sample was not dried nor weighed before immersion into the plating
solution to avoid the formation of oxide. After plating, the sample
was rinsed, dried, and then weighed to estimate the mass of Ni
5.1
82°C
Bath Temperature
a No plating was recorded when lead acetate was used as bath stabi-
lizer.
Downloaded on 2015-06-25 to IP 138.38.0.53 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C818
Journal of The Electrochemical Society, 150 ͑11͒ C816-C822 ͑2003͒
Figure 1. The effect of thiourea on bath chemistry using gravimetry bath
compositions ͑Table I͒. LCC Cu as the substrates, plating time: 30 min.
stabilize the plating bath. We have studied the effect of stabilizing
agents, thiourea, and lead acetate using classical gravimetry and
QCM techniques.
Effect of stabilizing agent on plating.—Typical stabilizers used in Ni
baths include Group IV elements ͑e.g., Se, Te͒, oxygen-containing
compounds ͑e.g., AsOϪ2 ) and unsaturated organic compounds ͑e.g.,
thiourea and metal ions such as Pb2ϩ). The first two stabilizer types
are more toxic than the latter two. For this reason, we used only
thiourea and lead acetate in our experiments. It was found that when
lead acetate was used as the stabilizing agent, the plating of Ni was
almost completely inhibited, even at 0.01 ppm lead acetate. Only a
small portion of the LCC sample was plated. So, in the baths con-
taining HAc as the complexing agents and sodium hypophosphite as
the reducing agents, lead acetate was found to completely inhibit the
plating. Therefore, this leads to our conclusion that lead acetate
could not be used as stabilizer in the current bath composition and
should be avoided to eliminate any contamination.
Figure 3. The condition of the substrate surface as thiourea was added to the
plating bath during electroless deposition of nickel. ͑a͒ Thiourea binding
sites on the surface. ͑b͒ Catalysis process was dominant on the surface to
increase the plating rate. ͑c͒ The adsorption of thiourea was predominant to
increase the plating rate resulting in the first maximum plating rate. ͑d͒
Catalysis process was predominant again to increase the plating rate. ͑e͒
Competing process occurring between the thiourea dimer and Ni2ϩ for hy-
pophosphite, which decreased the plating rate resulting in the second maxi-
mum plating rate.
Thiourea can be strongly adsorbed onto catalytic surfaces.13,14
Solution agitation in the presence of the stabilizer can cause inhibi-
tion of the nickel deposition reaction, which can be noticeable by a
decrease in the plating rate. As stated earlier, when the stabilizer is
adsorbed onto the catalytic surface, the number of catalytic sites
available for dehydrogenation of hypophosphite is reduced. Com-
plexing agents are also required to decrease the concentration of free
metal ions such as Ni͑II͒ and hence the deposition rate.
We first utilized the weight gain method to determine the plating
efficiency of the baths as shown in Tables I, II, and III. Using the
classical gravimetric techniques, the weight gained recorded at each
LCC substrate using different thiourea concentration is shown in
Fig. 1. Two maxima were observed in Fig. 1. The first occurred at a
thiourea concentration of 0.05 ppm, while the second was noticed at
1 ppm thiourea concentration. This suggested that the weight gain
method could also be used to predict the plating rates. To our knowl-
edge, this appears to be the first time a bimodal distribution was
recorded for stabilizer concentration. Our observation was consistent
with the phenomenon reported earlier in which thiourea was first
suggested to increase the plating rate at very low concentration but
subsequently decreased the plating rate.15 Perhaps, the levels of sta-
bilizer concentrations studied in previous reports could have been
very high, which might have prevented the observation of the bimo-
dal distribution recorded in this work.5-7 This rather unexpected
finding has prompted us to conduct a detailed, systematic study of
the role of different bath species to plating efficiency. The gravim-
etry analysis was correlated to the thickness effects using XRF. As
shown in Fig. 2, the thickness measured using XRF followed a
similar trend as with the gravimetry when thiourea concentration
was varied. The complicated bimodal distribution shown in Fig. 1
Figure 2. The effect of thiourea on thickness of Ni deposit using XRF. Bath
compositions and conditions as in Fig. 1.
Downloaded on 2015-06-25 to IP 138.38.0.53 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, 150 ͑11͒ C816-C822 ͑2003͒
C819
• The dimer produced is finally reduced by H2POϪ2 as in Eq. 8. Thus,
Thiourea is regenerated following a reduction of formamidine disul-
fide by hypophosphite. It seems highly probable that the reduction
of formamidine disulfide may also proceed via a breakdown product
into two thiourea free radicals, which are then subsequently reduced
to thiourea. The radical formation in Eq. 5 is reasonable intermedi-
ate for a single electron oxidation process because the end product is
a disulfide.16 The presence of the radical formation from thiourea
has been confirmed by chemical oxidation by hydrogen peroxide
using electron spin resonance.17
and 2, which related the plating rate to the concentration of thiourea
could be explained using the following mechanism.
Mechanism of stabilizer regeneration at electroless nickel
baths.—According to the proposed mechanism, thiourea was first
oxidized into a reactive intermediate radical, which facilitated the
transfer of electrons to the Ni ions. It is this anodic reaction that
inherently accelerated the plating rate, yielding elemental Ni and
formamidine disulfide. The formamidine disulfide was then reduced
by hypophosphite to regenerate the thiourea. The elementary steps
involved in this mechanism are described from Eq. 4 through 8
On the other hand, a competing mechanism may predominate in
which the thiourea is adsorbed onto the substrate surface where it
impedes the plating reaction as illustrated in Fig. 3a. We believe that
when small amount of thiourea is added to the bath solution, the
plating rate increases because the former processes ͑the catalysis
process, Eq. 4-8͒ are predominant as in Fig. 3b and part a of Fig. 4.
When additional thiourea ͑Ͼ0.05 ppm͒ was added, more of Tu ad-
sorbed onto the substrate surface. This decreased the number of
active sites and slowed the plating rates. This explains the maximum
plating rate observed at 0.05 ppm. With continuous addition of thio-
urea, more thiourea would adsorb onto the substrate surface and
would further decrease the plating rate until all available thiourea
binding sites are filled and the plating rate reaches a minimum as
illustrated in Fig. 3c and part b of Fig. 4. When an increasing
amount of thiourea was added, the plating rate increased again be-
cause no more thiourea could adsorb onto the substrate surface ͑all
available binding sites have been filled͒ to decrease the plating rate,
and the catalytic process ͑Eq. 3.4-3.8͒ would be predominant again
to increase the plating rate as illustrated in Fig. 3d and part c of Fig.
4. As additional thiourea was added, more formamidine disulfide
was produced, and when that dimer had been generated in sufficient
amount, its ability to oxidize H2POϪ3 to H3PO3Ϫ ͑Eq. 8͒ inhibited the
͓4͔
The oxidation of thiourea follows a two-step mechanism in which
the tautomeric form of thiourea is oxidized to a thiourea free radical
͑Thiourea ͒ through a charge transfer. The thiourea will likely
react with another radical to form reactive intermediate disulfide.
This process produces a reversible mechanism that regenerates the
thiourea
16a
*
*
• Thiourea ϭ thiourea ϩ Hϩ ϩ e
*
͓5͔
reduction of Ni2ϩ by H2PO3Ϫ thus decreasing the Ni deposition rate.
So, with the competing effect of thiourea, the second maximum
plating rate at 1 ppm thiourea resulted and the rate continued to
decrease with more thiourea added as illustrated in Fig. 3e and part
d of Fig. 4.
*
• where thiourea
ϭ
A QCM study was then conducted to validate the results obtained
using the weight gain method. In QCM, the mass of a metal or other
substance deposited onto the quartz electrode can be related to the
change in frequency by using Saurbrey equation18
*
• Then, 2 thiourea → formamidine disulfide
2f20
⌬f ϭ Ϫ
⌬m
͓9͔
ͩ
ͪ
Sͱ␮␳
͓6͔
• The overall reaction with Ni2ϩ and H2PO3Ϫ will be
where ⌬f is the change in frequency of the crystal ͑which is a
measurable parameter͒, f0 is the fundamental resonant frequency of
the crystal ͑in the electrolyte͒, S is the electrode area, ␮ is the sheer
modulus of the crystal (2.947 ϫ 1011 g cmϪ1 sϪ2͒, ␳ is the density
of the quartz ͑2.648 g cmϪ3͒, and ⌬m is the change in mass at the
surface of the crystal resonator. Therefore, the adsorption/desorption
processes at the electrode surface can be monitored as a function of
the change in mass.
͓7͔
͓8͔
Downloaded on 2015-06-25 to IP 138.38.0.53 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C820
Journal of The Electrochemical Society, 150 ͑11͒ C816-C822 ͑2003͒
Figure 4. The dependence of plating rate on thiourea. ͑a͒ Rate increased
with thiourea where the catalytic process is predominant. ͑b͒ Rate decreased
where the adsorption of thiourea onto the surface is predominant; ͑c͒ Rate
increased again where the catalytic process was predominant and no avail-
able binding site thiourea. ͑d͒ Rate decreased due to the competition between
the dimer and Ni͑II͒ for hypophosphite.
Figure 6. The effect of Ni͑II͒ on bath chemistry using gravimetric technique
͑Table II͒. LCC Cu was used as the substrates with plating time of 30 min.
The increase in mass vs. thiourea concentration when Ni was
electrolessly plated onto the QCM is shown in Fig. 5. This result
showed that the relationship obtained from QCM was the same as
that generated using classical gravimetry. This further supports our
mechanism described above. Additional experiments were per-
formed to better understand the effect of Ni͑II͒ or sodium hypophos-
phite on plating efficiency. The results shown in Fig. 1, 2, and 5
indicate that the plating rate is higher when thiourea concentration is
kept at 0.05 ppm. However, for practical purposes, we decided to
use Ni plating bath having thiourea concentration of 1 ppm because
noticeable fluctuation in the plating rates are recorded at thiourea
concentration of 0.05 ppm, which makes the plating unstable.
complexed by the HAc ligand. This corresponds to the processes
occurring in Eq. 2-3 . Because the reduction of Ni͑II͒ was too fast,
Ј
we observed that the LCC samples were not plated. So, it was con-
cluded that the critical concentration required for Ni͑II͒ in this bath
was 6 g/L. Subsequently, the effect of hypophosphite was studied
using the established concentration.
The effect of sodium hypophosphite on electroless Ni plating.—The
effect of sodium hypophosphite was studied using the composition
shown in Table III. By varying the concentration of sodium hypo-
phosphite, we expected that the plating rate for the Ni bath would
change. Figure 7 shows the weight gain of Ni at varying hypophos-
phite concentrations. Results indicated that as the concentration of
sodium hypophosphite increased, the plating rate also increased, but
later leveled off at 30 g/L sodium hypophosphite. The effect of
sodium hypophosphite on Ni plating was not as dramatic as when Ni
or thiourea concentrations were varied.
The effect of Ni(II) on electroless Ni plating.—We studied the effect
of changing the concentration of Ni͑II͒ on plating rate by varying its
concentration while keeping other parameters constant ͑Table II͒.
The effect of Ni͑II͒ concentration on the deposition rate is shown in
Fig. 6 using weight gain method and XRF techniques. Because both
the weight gain method and XRF technique exhibited similar ten-
dency, subsequent results were reported using only the weight gain
method. It can be observed that the highest plating rate was recorded
between 4 to 6 g/L ͑Fig. 6͒. It was also noted that at Ni͑II͒ concen-
tration of 4 g/L, partial and uneven plating was observed on the
LCC substrate. However, at increasingly higher concentration of
Ni͑II͒, especially at 10 g/L Ni concentration, the deposition rate was
quite high such that the bath began to decompose. This was evident
by the rapid decay in the plating rate beyond 6 g/L. Therefore at
extremely high Ni͑II͒ concentration ͑12 g/L͒, the bath rapidly de-
composed because excess free Ni͑II͒ was present that could not be
EDX analysis and FTIR studies.—The elemental composition of the
electroless nickel layer was investigated using EDX and FTIR tech-
niques. A representative EDX spectrum is shown in Fig. 8. The
strong Ni peaks at 0.05 and 0.10 ppm thiourea were located at ap-
proximately 7.5 keV. Another weak Ni peak was observed at 1.0
keV. Phosphorus peaks were observed at 2.0 keV, which signified
the inclusion of either elemental phosphorus or hypophosphite anion
in the plated Ni layer. The EDX spectrum also confirmed that the
deposited Ni peaks were pure Ni phases. The ratio of Ni:P found
also examined. At less than 0.05 ppm Tu, the proportion of Ni ͑7.5
keV͒ to phosphorus ͑12.0 keV͒ was determined to be ϳ6.3. This
proportion however increased to 7.3 when Tu concentration changed
to 1 ppm. This is attributed to the relative increase in the amount of
Ni.
Figure 5. The effect of thiourea on the mass increase of the QCM immersed
in the Ni plating solution. Bath compositions and conditions as in Fig. 1
while the plating time was 10 s.
Figure 7. The effect of NaH2PO2 on bath chemistry using gravimetry ͑Table
III͒. Other conditions as in Fig. 4.
Downloaded on 2015-06-25 to IP 138.38.0.53 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, 150 ͑11͒ C816-C822 ͑2003͒
C821
Figure 9. Experimental conditions: Supplementary material 1 shows the
FTIR of the electroless Ni plated from a Ni bath without thiourea. Supple-
mentary material 2 shows that of Ni plated from a Ni bath with 1 ppm
thiourea present. Plating was achieved on Kapton substrates for FTIR experi-
ments. The Kapton substrates were sensitized by SnCl2 solution and acti-
vated using PdCl2 solution ͑Ref. 1a͒ and then soaked into the two Ni baths
separately for 2 min. One bath contained no thiourea while the other con-
tained 1 ppm thiourea. Other bath compositions and conditions are as in
Fig. 1.
erated, no traces could be detected. Confirmatory evidence resulting
from both the EDX and FTIR however, suggests that the latter
seems to be the case. The thiourea is being regenerated. Thus the
optimal conditions for an acidic electroless Ni plating bath was de-
termined to be: Ni͑II͒, 6 g/L; sodium hypophosphite, 30 g/L; acetic
acid, 30 g/L; and thiourea, 1 ppm.
Figure 8. EDX spectra recorded for plated Ni samples at varying concen-
trations of thiourea. Bath compositions include: Ni͑II͒ at g/L,
NaH2PO2•H2O at 30 g/L, HAc at 30 g/L, pH ϭ 5.1. Temperature 82°C,
plating time was 30 min and thiourea concentrations at ͑a͒ bare copper sub-
strate, ͑b͒ 0.01, ͑c͒ 0.05 ppm, and ͑d͒ 1 ppm.
6
Moreover, the EDX spectra qualitatively revealed a higher pro-
portion of Ni to P at very high Tu concentration. It could be assumed
that metallic Ni was codeposited with elemental P to form an Ni-P
layer but there was no codeposition of sulfur. The absence of sulfur
peaks in the EDX study regardless of the thiourea concentration
further confirmed our mechanism above. It appeared that stabilizer
was regenerated. Finally, the presence of the stabilizer was exam-
ined using FTIR. The rationale is that if thiourea was incorporated in
the nickel layer, certain components of these compounds ͑e.g., sul-
fur͒ or its traces could be observed from elemental analysis. The
FTIR spectra were recorded with and without thiourea present. No
sulfur peaks were observed ͑supplementary results, Fig. 9͒. This
could either be attributed to insufficient sensitivity of the instrument
for the low levels of sulfur present or that because the Tu is regen-
Conclusions
We have investigated the effect of Ni͑II͒ ͑as Ni source͒, sodium
hypophosphite ͑as the reducing agent͒, thiourea, and lead acetate
͑both as stabilizing agents͒ on the electroless nickel plating. Since
lead acetate was found to completely inhibit the plating reaction, we
have concluded that lead acetate should be avoided in this type of
bath. However, thiourea had a profound effect on the Ni plating. A
bimodal distribution was observed as thiourea concentrations were
varied. Therefore, a regeneration mechanism was observed for thio-
urea concentration and this was confirmed using EDX and FTIR
data. Thus the optimal conditions for the acidic electroless Ni plat-
ing bath was determined to be Ni͑II͒ at 6 g/L; sodium hypophos-
phite, 30 g/L; acetic acid, 30 g/L and thiourea, 1 ppm.
Downloaded on 2015-06-25 to IP 138.38.0.53 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

C822
Journal of The Electrochemical Society, 150 ͑11͒ C816-C822 ͑2003͒
4. Y. Okinaka and T. Osaka, in Advances in Electrochemical Science and Engineer-
Acknowledgments
ing, Vol. 3, H. Gerischer and C. Tobias, Editors, p. 55, VCH, Germany ͑1994͒.
5. N. Feldstein and T. S. Lancsek, J. Electrochem. Soc., 118, 869 ͑1971͒.
6. C. Gabrielli and F. Raulin, J. Appl. Electrochem., 1, 167 ͑1971͒.
7. D. Altura, in Proceedings of Interfinish 84, the 11th World Congress on Metal
Finishing, p. 312 ͑1984͒.
8. A. Sargent, O. A. Sadik, and L. J. Matienzo, J. Electrochem. Soc., 148, C257
͑2001͒.
9. A. Sargent and O. A. Sadik, Langmuir, 17, 2760 ͑2001͒.
10. A. Sargent and O. A. Sadik, J. Electrochem. Soc., 148, C413 ͑2001͒.
11. N. Feldstein, U.S. Pat. 4,305,997 ͑1981͒.
The authors gratefully acknowledge financial support from the
Integrated Electronics and Electrical Center ͑IEEC͒ at SUNY-
Binghamton. J.B. is also grateful for his NIH-funded Bridge-to-
Baccalaureate Program at SUNY-Binghamton. We also acknowl-
edge Anita Sargent of IBM Microelectronics, Endicott, NY, for
providing the wirebond Cu samples and for performing XRF thick-
ness measurements and finally, Bill Blackburn of Geology Depart-
ment is acknowledged for performing the EDX measurements.
12. G. Mallory, in Electroless Plating: Fundamentals and Applications, G. O. Mallory
and J. B. Hajdu, Editors, Chap. 1, AESF, Orlando, FL ͑1990͒.
13. G. Deschenes, D. Bernard, P. Prud’Homme, and D. Laforest, Minerals Eng., 7, 435
͑1994͒.
14. ͑a͒ C. Chen, T. Lung, and C. Wan, Hydrometallurgy, 5, 207 ͑1980͒; ͑b͒ T. Groe-
newald, Hydrometallurgy, 1, 277 ͑1976͒; ͑c͒ A. Brenner and G. E. Riddell, Plat.
Surf. Finish., 85, 54 ͑1998͒.
15. K. Han, Y. Wu, M. Zhang, and J. Wang, Trans. Inst. Met. Finish, 74, 91 ͑1996͒.
16. ͑a͒ S. Reddy and J. Krishnan, J. Electroanal. Chem., 27, 473 ͑1970͒; ͑b͒ N. Mo-
tohashi and I. Mori, J. Inorg. Biochem., 26, 205 ͑1986͒.
References
1. ͑a͒ M. Schlesinger, in Modern Electroplating, 4th ed., M. Schlesinger and M.
Paunovic, Editors, Chap. 18, John Wiley & Sons, Inc., New York ͑2000͒; ͑b͒ J.
Berkowitz and N. H. Emerson, Plating Methods: A Survey, Chap. 6, NASA, Wash-
ington, D.C. ͑1972͒.
2. ͑a͒ I. Petukhov, M. Shcherban, and V. Kichigan, Prot. Met., 35, 85 ͑1999͒; ͑b͒ K.
Chen and Y. Chen, Plat. Surf. Finish., 84, 80 ͑1997͒; ͑c͒ G. S. Tzeng, J. Appl.
Electrochem., 26, 969 ͑1996͒; ͑d͒ T. Saito, E. Sato, M. Matsuoka, and C. Iwakura,
J. Appl. Electrochem., 28, 559 ͑1998͒; ͑e͒ A. H. Gafin and S. W. Orchard, J. Appl.
Electrochem., 22, 830 ͑1992͒.
17. S. Saha and D. J. Greenslade, Bull. Chem. Soc. Jpn., 65, 2720 ͑1992͒.
18. G. Z. Sauerbrey, Z. Phys., 155, 206 ͑1959͒.
3. T. Osaka, M. Kato, J. Sato, K. Yoshizawa, T. Homma, Y. Okinaka, and O.
Yoshioka, J. Electrochem. Soc., 148, C659 ͑2001͒.
Downloaded on 2015-06-25 to IP 138.38.0.53 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).