AESF
RESEARCH PROJECT 68
Electrodeposition is localized
by applying a high peak pulse current density and directing
a small insoluble anode toward a large, unmasked cathode.
Experimental data for pulse plating with a neutral phosphate
and hard gold baths and with both fluoborate and sulfonate
tin-lead solder plating solutions showed that highly localized
deposits were obtained at (1) an anode peak current density
of 0.5 to 10 A/cm2, (2) a pulse period of 0.01
to 0.1 sec, (3) a duty cycle of 1 to 10 percent and (4)
an anode-cathode spacing of less than one anode radius.
In selective, spot plating,
the localization of metal deposition is generally achieved
by the use of cathode masks, photo-resist coatings and brush
plating. These techniques require a series of pre-plating
and post-plating treatment procedures, resulting in a high
production cost in the overall finishing process.
Several theoretical investigations1-3 have shown that high peak current
densities on the order of 10 to 100 A/cm2 can
be used with short pulse periods and small duty cycles without
depleting the concentration of metal ions at the cathode
surface. The primary current distribution prevails at high
plating current densities and the throwing power is worse
than the secondary current distribution. This suggests that
selective plating can be achieved by (1) directing a small,
insoluble anode toward a large unmasked cathode, (2) keeping
the anode-cathode spacing small and (3) applying a high
peak pulse current density to localize electrodeposition
at the selected area facing the anode. These principles
were used to carry out selective pulse plating of copper
from an acid copper sulfate bath.4,5 Pulsed current produced a more
localized copper deposit than direct current plating.
This report describes a study
of the feasibility of using pulse plating to localize the
electrodeposition of gold and a 60 percent tin-40 percent
lead alloy on an unmasked cathode. The metal thickness distributions
were experimentally measured after using different peak
pulse current densities, duty cycles and pulse periods.
The results are compared to those obtained with direct current
at the same average current densities and anode-cathode
spacing.
The details of the experimental
procedure and instrumentation were previously described
in several publications. 4-6 The plating cell used in the study
reported here contained a large, unmasked cathode and a
small, insoluble circular anode. The cathode was supported
vertically in the cell. The circular anode with its exposed
plane parallel to that of the cathode was introduced from
one side of the cell. The anode-cathode spacing was adjusted
by sliding the anode support rod through a water-tight fitting
on the cell wall. The entire cell was immersed in a constant
temperature water bath. A thermometer inserted through the
cell cover was used to measure the temperature of the electrolyte.
A combination pH electrode was used to monitor the pH of
the plating bath. A Teflon-coated magnetic stirrer served
to circulate the electrolyte during plating experiments.
Pulse plating was controlled
galvanostatically. A mini-Pulsir3 connected to a dc source
was used to generate a rectangular pulse voltage signal
of a desired magnitude, duty cycle and pulse period. The
voltage signal was fed to a high power potentiostat and
converted to pulse current by operating the potentiostat
in the galvanostatic mode. A microcomputer was used to interface
a digital data acquisition unit to monitor peak pulse current,
the time averaged dc current, the anode-cathode cell voltage
and the charge passed during each experiment.
Gold was pulse-plated with a neutral
phosphate bath that contained no additives.7 A commercial acid
bath that contained cobalt chelates and wetting
agents was used to obtain hardened gold deposits. Bath compositions,
pH values and operating temperatures are given in Table 1. A 6-mm-diameter
stainless steel disk mounted on a Teflon sleeve was the
anode. Copper cathodes with dimensions of 0.038 x 5 x 5
cm were polished with 0.3 
m-alumina polishing compound, degreased with methanol
and rinsed in distilled water before each experiment. A
1.0-m layer of bright nickel was followed by 0.025 to 0.05
m of gold from an alkaline strike.
Pulse plating experiments were conducted
over the 0.01 to 10 A/cm2 range of anodic peak
current densities with duty cycles of 1 to 100 percent,
pulse periods of 1 to 1000 msec and bath temperatures of
21° to 55°C. The anode-cathode gap was equal to one anode
radius unless otherwise noted. The experiment was terminated
after a desired charge (based on the time-averaged direct
current). A total charge of 10 coulombs was adopted for
experiments with the neutral phosphate bath, whereas a total
charge of 25 coulombs was used for the acidic hard-gold
bath. The gold thickness was measured with a coulometric
thickness gage and the morphology of the deposits was examined
by scanning electron microscopy (SEM).
The 60-40, tin-lead alloy deposits
were obtained with (1) a commercial fluoborate bath and
(2) a commercial sulfonate bath. The bath compositions are
listed in Table 2. The anode was a circular,
6-mm-diameter stainless steel disk. Each 5 x 5 cm,
uncoated brass cathode was polished to a 0.3 m finish, degreased in methanol, dipped in dilute
HCI and rinsed in water before transfer to the cell. The
anode-cathode spacing was one anode radius. Bath temperature
was 21°C. Pulse peak-current densities ranged from 0.025
to 10 A/cm2, duty cycles from 1 to 100 percent
and pulse periods from 1 to 1000 msec. Plating was terminated
after a 15-coulomb charge. Deposit thickness was measured
with a coulometric thickness gage calibrated with a standard
coating of 60 percent tin and 40 percent lead. It was assumed
that deposit composition did not vary with plating conditions
or from one location on the cathode to another. The morphology
of the alloy deposits was examined by SEM.
To establish a basis for comparison
with pulse plating, the characteristics of the gold baths
were determined with dc plating. Figure
1 shows the cathode current efficiency for dc gold plating
over the range of 0.002 to 1.0 A/cm2. The efficiency
data were obtained with a copper rotating disk cathode that
had an exposed area of 0.0707 cm2. The disk was
rotated at 2500 rpm and the weight gain after plating was
compared with the theoretical amount to calculate the efficiency.
The phosphate bath was examined at 40°C with gold concentrations
of 0.05, 0.1 and 0.2M; the hard-gold solution was evaluated
at 21° and 55°C. Figure 1 exhibits general trends of:
(1) increasing current efficiency with increasing current
density in the low current density regime, (2) reaching
a maximum efficiency at a current density of about 0.05
to 0.1 A/cm2 and (3) decreasing efficiency with
a further increase in current density.
The current efficiency of the
phosphate bath, which reached a maximum of 95 percent, was
higher than that of the hard-gold bath. Increasing the temperature
of the hard-gold bath from 21° to 55°C increased the maximum
current efficiency at 0.05 A/cm2 from 30 to 50
percent. The current efficiency behavior has a significant
effect on the metal distribution, which tends to be more
localized if the current density is on an ascending portion
of the current efficiency curve. On the other hand, metal
distribution is more uniform if the cathode current density
is on the descending part of the current efficiency curve.
The active cathode area where
gold was electrodeposited during our selective plating experiments
was not defined. Thus, a controlled current density based
on the finite area was used to specify the plating conditions.
According to a current distribution model presented previously,
the maximum local current density at the cathode center
facing the anode is approximately equal to 1/3 to of the
anode current density at an anode-cathode spacing of one
anode radius. Figures
2 and 3 show the metal distribution results for
dc gold plating from the phosphate and hard-gold baths,
respectively, over a range of anode current densities from
0.01 to 0.2 A/cm2. The local gold thickness is
plotted as a function of the radial distance from the center
of the deposits. The thickness of gold from both baths was
maximum at the cathode center and decreased with an increasing
radial position. Increasing the average anode current density
increased the thickness at the center, and the metal distribution
was more localized.
The effect of changing the anode-cathode
gap from 0.2 to 4.0 of the anode radius on the selective
plating of gold also was examined while the average anode
current density was kept at the same value. The results
were similar to those of previous investigations. A decrease
in the anode-cathode spacing increased the localization
of gold deposition on the unmasked cathode. Metal thickness
at the center of the deposit with a spacing equal to 0.2
of the anode radius was 10 times greater than that obtained
with a gap of 1.0 anode radius and about 20 times the thickness
with an anode-cathode spacing equal to four times the anode
radius.
The effect of pulse plating on gold
thickness distribution is shown in Figs.
4 and 5 for the 0.2M neutral phosphate bath at 40°C
and the hard-gold solution at 21°C, respectively. The anode-cathode
spacing was kept at one anode radius; the average current
density and pulse period were maintained at 0.05 A/cm2.
A 100-percent duty cycle corresponded to dc plating.
Pulsing the current increased
the thickness of gold at the center area facing the anode
and localized metal distribution more than dc plating. The
extent of localization increased with decreasing duty cycles.
At a 1-percent duty cycle, the gold deposit at the center
was 13 times thicker than that measured after dc plating
with the phosphate solution and six times the dc thickness
with the hard-gold bath.
The effects of changing the
average anode current density and pulse period on selective
gold plating also were examined. Fig
6 shows typical results for varying the average anode
current density from 0.025 to 0.2 A/cm2
for the hard-gold bath at 21C while the anode-cathode
spacing was equal to one anode radius, the pulse period
was 10 msec and the duty cycle was 10 percent. The results
were similar to those obtained with dc plating. In both
cases, increasing the average current density increased
the localization of gold thickness. With an average anode
current density of 0.2 A/cm2, the thickness in
the center of the deposit was 3.5 times greater than that
obtained with a current density of 0.025 A/cm2.
The effect of changing the pulse
period (or frequency) on metal distribution is shown in
Fig 7 for the hard-gold bath at
55°C. The data were obtained by keeping the average anode
current density at 0.05 A/cm2 and the duty cycle
at 10 percent. The pulse period was varied from 1 msec to
1 sec (the pulse frequency from 1000 to 1 Hz). A uniform
deposit was obtained with a 1-msec pulse period. Metal distribution
became more localized with increasing pulse periods (decreasing
pulse frequencies). The same trend also was observed for
the selective pulse plating of gold from the neutral phosphate
bath at 40°C. The Fig.
7 data show that the most localized deposit was obtained
with a 1-sec pulse period; gold thickness at the center
of the deposit was 20 times greater than that observed with
a 1-msec pulse.
The effect of gold concentration
on the metal distribution by pulse plating with the phosphate
bath at 40C at an average anode current density of 0.05
A/cm2 and a 10-percent duty cycle is shown in
Fig. 8. Deposition from the 0.05M-gold
bath was more localized than that from the 0.1 or 0.2M-gold
solutions, because the conductivity of the electrolyte decreased
with a decreasing gold concentration. In pulse plating with
high peak current densities, the primary current distribution
prevails and the thickness distribution is more localized
with a low-conductivity electrolyte, as described by current
distribution theory.
The morphology of gold from
the hard-gold bath at 21°C was examined by SEM. Figure 9 shows SEM photographs
of deposits at the central area facing the anode.
Figure 9a is a dc deposit obtained
at an anode current density of 0.05 A/cm2 and
Fig. 9b shows a deposit produced with a 10-msec pulse, a
10-percent duty cycle and the same average current density.
The deposits exhibited hemispherical nodular growth. The
average crystalline size of the dc and pulse-plated deposits
was 1-3 and 10-15 m in diameter, respectively.
Figure
10 shows the
effect of average anode current density on the morphology
of gold deposits obtained from the acid hard gold bath with
pulse plating at a 50-msec pulse period and a 10 percent
duty cycle. Increasing the average anode current density
from 0.05 A/cm2to 0.2 A/cm2 caused
a decrease in crystalline size from 20 m (Fig. 10a) to 1-2 m in diameter (Figs. 10b and 10c).
The dc cathode current efficiencies
of the tin-lead fluoborate and sulfonate baths are shown
in Fig. 11 as
a function of the cathode current density. The data were
obtained with a rotating copper disk electrode, similar
to that used for gold plating. Deposit composition was assumed
to be 60 percent tin and 40 percent lead for all current
densities. The current efficiency of the fluoborate bath
was nearly constant at 80 percent over the range of 0.002
to 0.2 A/cm2. The sulfonate solution had a higher
efficiency of 90 to 100 percent at current densities below
0.3 A/cm2 , but above 0.3 A/cm2, current
efficiency dropped to 4 percent.
Typical results of pulse plating on
the deposit thickness distribution are given in Figs. 12 and 13, for the fluoborate and sulfonate
baths, respectively. The average anode current density was
kept at 0.025 and 0.1 A/cm2 for the fluoborate
and sulfonate baths, respectively. The pulse period was
kept at 10 msec; the duty cycle varied from 1 to 100 percent.
Uniform metal distributions were obtained with direct current
(100 percent duty cycles). Pulse current significantly improved
the localization of the deposit and the extent of localization
increased with decreasing duty cycles. At a 1 percent
duty cycle, the thickness at the center of the deposits
was 13 and 6.5 times greater than that obtained with direct
current for the fluoborate and sulfonate baths, respectively.
The effect of changing the anode current density was similar
to that established for gold plating. Thickness distribution
became more localized with an increasing average anode current
density for a given pulse period and duty cycle.
Figure
14 shows
the effect of changes in the pulse period on the thickness
distribution of the solder alloy deposited in the fluoborate
bath at 21C. The data in this figure were obtained while
the pulse period was varied from 1 msec to 1 sec when the
average anode current density was 0.025 A/cm2
and the duty cycle was 10 percent. Thickness distribution
was more localized with an increasing pulse period (a decreasing
frequency). With a 1-sec pulse period, the deposit at the
cathode center was nearly four times thicker than that obtained
with a 1-msec pulse period.
Current pulses increased the
crystalline size of the deposits. Figure 15 shows a set of SEM photographs
of tin-lead deposits obtained from the sulfonate bath at
an average anode current density of 0.1 A/cm2
with a 10 msec pulse. The duty cycle was 100 percent for
the deposit in Fig 15b and 1 percent for that
in Fig. 15c. Crystalline size
increased from 0.5 to 3-4 to 50 m, respectively. Small crystallites resulting
from secondary growth were visible at the 1-percent duty
cycle (Fig. 15c).
Localized
electrodeposition was achieved for gold and tin-lead alloy
plating by directing a small insoluble anode toward an unmasked
cathode at a small anode-cathode spacing and applying a
high-peak pulse current density. The metal distribution
of gold deposits from a neutral phosphate bath and a hard-gold
solution and the distribution of tin-lead solder from fluoborate
and sulfonate solutions were examined over a range of peak
anode current densities from 0.025 to 10 A/cm2
with pulse periods of 0.001 to 1 sec and duty cycles of
1 to 100 percent. Thickness distribution on the cathode
became more localized with a decreasing duty cycle and an
increasing pulse period. Highly localized gold and lead-tin
deposits were obtained by pulse plating at a high peak anode
current density of 0.5 to 10 A/cm2 a high pulse
period of 0.01 to 1 sec. a short duty cycle of 1 to 10 percent
and a small anode-cathode spacing of less than one radius
of the anode. Pulsing the current also changed the morphology
of the gold and tin-lead alloy deposits. At the same average
anode current densities, the average crystalline size of
pulse-plated deposits was much larger than that obtained
with direct current. For a given pulse period and duty cycle,
the crystalline size of gold decreased with increasing average
anode current densities.
The authors gratefully acknowledge
the financial support of the American Electro platers and
Surface Finishers Society (AESF Project 68) for the study
reported in this paper.
1.
D-T. Chin, J. Electrochem. Soc. 130, 1657 (1983)
2. N. lbl, Proc. AESF 2nd Int'l Pulse Plating Symposium
(1981).
3. D-T. Chin N.R.K. Vilambi and M.K. Sunkara,
Plat. And Surf. Fin., 76, 74 (Oct 1998)
4. D-T. Chin and N.R.K. Vilambi, Proc. AESF
Sur/Fin'86, Session G (1986).
5. N.R.K. Vilambi and D-T Chin, Plat. And
Surf.Fin, 75, 67 (Jan., 1988)
6. D-T Chin and M.K. Sunkara, Proc. AESF SUR/FIN '88, Session
B (1988).
7. Y. Okinaka and F.B. Koch, Proc. Interfinish
'80, Kyoto, Japan (1980).
Dr. Der-Tau Chin is professor
of chemical engineering, Clarkson, University, Potsdam,
NY 13699-5705. He has more than 20 years research experience
in electroplating, corrosion, electrochemical energy conversion,
and industrial electrolytic processes. Prior to joining
Clarkson, he was a senior research engineer in the Electrochemistry
Department of General Motors Research Laboratories. Dr.
Chin received his PhD from the University of Pennsylvania.
M.K. Sunkara worked as a research
assistant for AESF Research Project 68. He holds a MS in
chemical engineering from Clarkson University.
