Growth of metal around particles during electrodeposition

Article abstract of DOI:10.1149/1.2198090

The deposition profile of metal around particles during electrodeposition was studied by atomic force microscopy measurements. In addition to these measurements, the local metal deposition rate around particles was determined from electrodeposited nickel-iron multilayers using process archeology. A strong correlation was found between the metal deposition profile and the surface properties of the particles. Hydrophilic particles are pushed ahead by metal that deposits underneath the particles before they become incorporated. Metal deposition does not take place underneath hydrophobic particles, hence incorporating from the start of the metal deposition. For electrically conductive particles, metal deposits on the particles which causes such particles to incorporate almost immediately.

Full text of DOI:10.1149/1.2198090

C472
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
0013-4651/2006/153͑7͒/C472/11/$20.00 © The Electrochemical Society
Growth of Metal around Particles during Electrodeposition
*
**,z
L. Stappers and J. Fransaer
Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, B-3001 Heverlee,
Belgium
The deposition profile of metal around particles during electrodeposition was studied by atomic force microscopy measurements.
In addition to these measurements, the local metal deposition rate around particles was determined from electrodeposited nickel–
iron multilayers using “process archeology.” A strong correlation was found between the metal deposition profile and the surface
properties of the particles. Hydrophilic particles are pushed ahead by metal that deposits underneath the particles before they
become incorporated. Metal deposition does not take place underneath hydrophobic particles, hence incorporating from the start
of the metal deposition. For electrically conductive particles, metal deposits on the particles which causes such particles to
incorporate almost immediately.
© 2006 The Electrochemical Society. ͓DOI: 10.1149/1.2198090͔ All rights reserved.
Manuscript submitted January 10, 2006; revised manuscript received March 10, 2006. Available electronically May 4, 2006.
Incorporation of particles in metals leads to composite coatings
obtained by electrodepositing nickel–iron multilayers around glass,
polymethylmethacrylate ͑PMMA͒, and graphite particles.15 Cross-
sectioning and preferential etching of the iron layers reveals the
individual layers around the particle. As each layer corresponds with
a certain deposit time, the distance between the layers is a measure
of the rate of metal deposition. This technique, which was first em-
ployed by Schwartz et al. and which is called “process archaeology,”
shows where and when the highest metal deposition rate is
obtained.15 With this information, the incorporation behavior of dif-
ferent types of particles can be better understood.
with new and better properties.1-4 Therefore, the process to make
such coatings by electrolysis has gained interest since the late
1960s.5-8 Despite the fact that various studies give insight into the
effect of different process parameters on the codeposition of par-
ticles, the process is still not well understood. The last decade, the
relation between surface properties and particle incorporation was
intensively studied. These studies showed that hydrophilic particles
hardly incorporate while hydrophobic or conductive particles incor-
porate very well.9-12 Hence, it is the purpose of this work to inves-
tigate how particles with different surface properties incorporate in
metals by studying the form of the metal deposit around different
types of particles. The way in which metal grows around a particle is
important because the adhesion between the particle and the metal
deposit depends on the form of the metal deposit around the particle.
If the deposit is nonconformal, there will be gaps between the par-
ticle and the metal, which will lower the adhesion force. The shape
͑and roughness͒ of the metal deposit is important because the van
der Waals force which keeps the particle on the electrode during
incorporation is larger if the contact area between the particle and
electrode is larger.13 Therefore, particles can already become incor-
porated by thin metal deposits if the deposit grows around the par-
ticle, while much thicker deposits are necessary if the metal deposits
nonconformally around the particle. Also, for applications such as
cutting tools where large diamonds or other hard particles are incor-
porated in nickel, adhesion plays an important role in the perfor-
mance and life expectancy of the cutting tool.
Experimental
AFM measurements.— To study the shape of metal deposition
around a particle, a glass particle glued to an AFM cantilever was
pushed against the center of a small disk electrode ͑1 mm diam͒
with a force of about 50 nN.16,17 While the particle is in contact with
the electrode, nickel is deposited from an electrolyte composed of
180 g/L NiSO4·6H2O and 15 g/L H3BO3 with a pH of 4. The de-
posits are made at room temperature using a constant current. After
metal deposition, the particle is retracted from the electrode, and the
deposit profile at the position where the particle was located is de-
termined by scanning the deposit with a normal AFM tip ͑i.e., a
cantilever without an attached particle͒. The shape of the nickel
deposits was determined at current densities of −0.5 and
−10 A dm−2 and for metal deposit thicknesses of 500 and 2000 nm.
At each condition, three deposits were made and for each deposit,
four different cross sections through the AFM image were made at
angles of 0, 45, 90, and 135°. These four cross sections were used to
describe the metal growth profile around the particle. A circle was
fitted through the measured profile of the cavity formed by metal
deposition around the particle in order to quantify the size of the
cavity. Values of the radius and depth of the cavities shown later in
this work are average values determined from the four fitted circles.
The error on the values is the standard deviation on the four values
determined by the fitted circles.
The glass particle that was used in these AFM experiments is a
glass sphere ͑Potters-Ballotini, Valley Forge, PA͒ with a diameter
of 42.5 ␮m. The diameter was determined with a measuring ob-
jective on an optical microscope along four different directions. It
was found that the diameter did not depend on the direction of
the measurement, which indicates that the particles are spherical.
Moreover, scanning electron microscopy ͑SEM͒ of particles from
the same batch showed that the particles were spherical. The rough-
ness of particles from the same batch was determined by AFM
͑Nanoscope III, Digital Instruments, Santa Barbara, CA͒ and a value
of Rq = 0.2 nm was found.
No previous experimental studies on the growth of metal around
particles were found in literature. However, recently the metal depo-
sition profile around a nonconducting particle was studied theoreti-
cally by Lee and Talbot.14 In this study, both the primary and sec-
ondary current distributions were calculated and the level set
method was used to track the moving metal/electrolyte interface as
the metal grew around the particle. When the metal deposition is
governed by the primary current distribution, the metal deposition
rate accelerates at a radial position equal to the particle radius, caus-
ing a large shape-change of the metal/electrolyte interface with time
and resulting in the formation of a void on the sides of the particle
͑Fig. 6b in Ref. 14͒. Although the secondary current distribution is
much more uniform, formation of a void was also observed for
dimensionless exchange current densities above 1 ͑Fig. 9b in Ref.
14͒. Because the shape of the metal deposit around particles is im-
portant for their incorporation, the shape of the metal deposit around
glass particles was investigated experimentally by atomic force mi-
croscopy ͑AFM͒ measurements. In addition to these measurements,
the profile of metal deposits as a function of the deposit time was
In this work, a standard AFM liquid cell was equipped with a
ring-shaped platinum counter electrode with an inner ring diameter
of 6 mm, a nickel working electrode with a diameter of 1 mm, and
a miniaturized Ag/AgCl reference electrode. The counter electrode
was positioned at the top of the liquid cell while the working elec-
*
Electrochemical Society Student Member.
Electrochemical Society Active Member.
**
z E-mail: jan.fransaer@mtm.kuleuven.be
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
C473
Table I. Experimentally determined current efficiencies of nickel
deposition at different current densities.
i
Efficiency
͑A dm−2
͒
͑%͒
−0.5
−1
−2
−5
−10
82
83
92
99
98
trode was embedded in epoxy resin and was positioned at the bot-
tom of the cell. The electrolyte flowed through the cell using a
gravity flow setup.18 In order to interpret the AFM experiments, the
current efficiency and the current distribution in the electrochemical
AFM cell should be known. The current distribution for this elec-
trochemical cell geometry and the current efficiency were deter-
mined experimentally. This was done by depositing nickel layers at
average current densities of −0.5, −1, −2, −5, and −10 A/dm2 with
average deposit thicknesses of 2 ␮m. The profiles of these deposits
were determined by laser profilometry ͑Wyco NT 3300 profiling
system, Veeco Tucson, Inc., Tucson, AZ͒, and these profiles were
used to calculate the current distribution using Faraday’s law
Figure 2. Differential size distribution of glass particles ͑Potters-Ballotini͒
before and after sieving with a 45 ␮m sieve.
nF␳dA
i =
͓1͔
distributions at the other current densities were comparable. Figure 1
shows that the current distribution is quasi-homogeneous and hence,
the nickel deposit thickness equals the average nickel deposit thick-
ness everywhere on the electrode. Therefore, the average metal
deposition in the vicinity of the particle is not influenced greatly by
the position of the particle on the electrode. It was tested by laser
profilometry of a deposit that the presence of the cantilever to which
the particle is attached does not change the current distribution on
the electrode.
Mt
with i the current density, n the valency of the nickel ions, F Fara-
day’s constant, ␳ the density of the metal, d the thickness of the
deposit, A the area of the electrode, M the molar weight of the
metal, and t the deposition time. The experimentally determined
current distributions were used to calculate the average current den-
sities on the electrode. The values of these experimentally deter-
mined current densities were compared with the applied current den-
sities to obtain the current efficiencies ͑Table I͒. The current
distribution determined according to the above procedure at an av-
erage current density of −2 A dm−2 is shown in Fig. 1. Current
Nickel–iron multilayer deposition.— Nickel–iron multilayers
were deposited around glass, PMMA, and graphite particles. The
deposits were made at room temperature from an electrolyte com-
posed of 0.2 M Ni͑NH2SO3͒2, 0.05 M FeCl2, and 0.4 M H3BO3
with a pH of 4. Ascorbic acid ͑1 g/L͒ was added to prevent oxida-
tion of Fe2+ to Fe3+. Besides ascorbic acid, no additives were used,
although most nickel–iron plating baths cited in literature contain
saccharin to reduce internal stresses and sodium lauryl sulfate to
improve the release of hydrogen bubbles formed on the cathode.
Additives were not used as they might adsorb on the particles in the
electrolyte, altering the surface properties of the particles and hence
changing their incorporation behavior in metals. Nickel–iron multi-
layers were deposited on brass substrates ͑Messing Folien-Band
0.1 mm, Overtoom International Belgium N.V., Belgium͒ which
were degreased with an alkaline cleaning solution ͑P3-RST, Henkel,
Germany͒ and rinsed with demineralized water and ethanol. The
substrates were dried with a flow of warm air before they were
mounted in the electrochemical cell. The electrochemical cell con-
sisted of a glass cylinder with a flange which was pressed against the
brass substrate. The diameter of the glass cylinder was 4 cm and the
height 6 cm. A disk-shaped nickel anode with a thickness of 1 cm
and a diameter of 3 cm was positioned at the top of the cell. Prior to
multilayer deposition, particles were added to the electrolyte.
The glass particles used in this work were obtained from Potters-
Ballotini and had a size distribution between 0 and 50 ␮m, as speci-
fied by the manufacturer. To obtain a narrower size distribution, the
particles were sieved with a 45 ␮m sieve. The resulting size distri-
bution is shown in Fig. 2. Graphite particles with a maximum size of
50 ␮m were obtained from Goodfellows, and PMMA particles were
synthesized by solvent evaporation according to the following pro-
cedure. Bulk PMMA ͑20 g͒ was dissolved in 300 mL of dichlo-
romethane. The dichloromethane/PMMA solution was added slowly
to an aqueous solution of 1 wt % gelatin while vigorously stirring.
Figure 1. Geometry of the electrochemical AFM cell with dimensions given
in millimeters. The AFM cell contains a circular cathode with a diameter of
1 mm and a ring-shaped anode with an inner ring diameter of 6 mm. The
current distribution for this cell geometry was determined experimentally
from the profile of a nickel deposit and was found to be relatively homoge-
neous. The maximum and minimum values differ by only 15%.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

C474
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
Figure 5. Cavity formed by metal deposition around a spherical glass par-
ticle with a diameter of 42.5 ␮m. The metal deposit was made at a current
density of −0.5 A dm−2, and the thickness of the deposit was 2000 nm. The
geometry of the cavity formed during metal deposition around the particle is
conformal with the shape of the particle and hence the particle fits in the
cavity.
Figure 3. Cavity formed by metal deposition around a spherical glass par-
ticle with a diameter of 42.5 ␮m. The metal deposit was made at a current
density of −0.5 A dm−2, and the thickness of the deposit was 500 nm. The
geometry of the cavity formed during metal deposition around the particle is
not conformal with the shape of the particle, and the particle has been pushed
up by metal deposition underneath the particle.
In this way, an emulsion of dichloromethane/PMMA droplets in an
aqueous gelatin solution was formed. This emulsion was heated
from room temperature to 50°C in about 1 h while stirring with a
paddle stirrer at 200 rpm. The emulsion was kept at this temperature
for 12 h in order to evaporate the dichloromethane. Once the dichlo-
romethane evaporated, a dispersion of solid PMMA particles in an
aqueous gelatin solution was obtained. This dispersion was washed
seven times with 50°C water in order to remove the gelatin. Each
washing cycle consisted of centrifugation of the dispersion, discard-
ing the solution, and resuspending the particles in demineralized
50°C water.
The amount of particles added to the electrolyte prior to
multilayer deposition was chosen such that the surface coverage of
particles on the brass cathode was approximately 15% when all
particles had settled on the cathode. Multilayer deposition was only
started after all particles had settled. The electrolyte was not stirred
during multilayer deposition. Multilayers were made by switching
the current density between −2 and −0.2 A dm−2 to obtain nickel-
and iron-rich layers, respectively. The deposition times were chosen
such that the thicknesses of the nickel- and iron-rich layers were
1000 and 500 nm, respectively. The total thickness of the deposit
was varied between 20 and 40 ␮m, depending on the average size of
the particles which were added to the electrolyte. After multilayer
deposition, a copper layer was deposited on top of the nickel–iron
multilayer in order to protect the multilayer during preparation of
cross sections. The multilayer deposits were embedded in epoxy
resin ͑Epofix͒, cross-sectioned, ground, and polished. The polished
cross sections were etched in 2 M phosphoric acid for 1 min, which
selectively attacked iron-rich layers and hence revealed the growth
profile of the metal. Two deposits were made for each type of par-
ticle and for each deposit; about 15 cross sections were analyzed.
When no metal was deposited underneath the particle, which was
the case for PMMA and graphite particles, the cross section of the
deposit went through the center of the particle if the particle touched
the substrate on the cross section. When metal deposited underneath
the particle, which was the case for glass particles, the size of the
particle was used to determine whether the cross section went
through the center of the particle. Because glass particles have a
narrow size distribution and most particles had a size between 45
and approximately 50 ␮m ͑Fig. 2͒, it was assumed that the cross
section went through the center if the size of the particle determined
on the cross section was between 45 and 50 ␮m. Hence, the actual
particle size might differ 5 ␮m with the particle size on the cross
section. Therefore, it can be estimated that the plane of the cross
section shifted 10 ␮m at maximum from the plane through the ac-
tual center of the particle.
Figure 4. Cavity formed by metal deposition around a spherical glass par-
ticle with a diameter of 42.5 ␮m. The metal deposit was made at a current
density of −10 A dm−2, and the thickness of the deposit was 500 nm. The
geometry of the cavity formed during metal deposition around the particle is
not conformal with the shape of the particle, and the particle has been pushed
up by metal deposition underneath the particle.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
C475
Figure 6. Cavity formed by metal deposition around a spherical glass par-
ticle with a diameter of 42.5 ␮m. The metal deposit was made at a current
density of −10 A dm−2 and the thickness of the deposit was 2000 nm. The
geometry of the cavity formed during metal deposition around the particle is
not conformal with the shape of the particle, and the particle has been pushed
up by metal deposition underneath the particle.
Results
Shape of metal deposition determined by AFM measurements.—
The shape of nickel deposits around spherical glass particles was
determined for current densities of −0.5 and −10 A dm−2 and for
metal deposit thicknesses of 500 and 2000 nm. The results are
shown in Fig. 3-6. These figures show sections of deposit profiles
around glass particles determined by AFM imaging. The contour of
the spherical particle is drawn on the graphs. Due to the nonisotropic
scaling of the axes, it does not look spherical on the graphs. All
images show a cavity where the particle had been positioned during
metal deposition and hence it can be concluded that nickel has
grown around the particle. Next to the cavity, the deposits are al-
ways relatively flat. For deposits of 500 and 2000 nm, the depth of
Figure 8. Schematic representation of AFM measurements to determine par-
ticle riding. Two force curves ͑cantilever deflection vs piezo displacement͒
were measured, respectively, before and after metal deposition. The horizon-
tal shift between the two curves corresponds to thickness of the metal deposit
underneath the particle, which is defined as the riding distance.
the cavities should be 500 and 2000 nm deep. However, this was
never the case. The depth of the cavity measured on AFM images
was always smaller than the thickness of the metal deposit. This
indicates that metal deposition occurred underneath the particle and
that the particle was pushed up by the metal deposition. This phe-
nomenon is called particle riding, and the distance by which the
particle is pushed up by the metal deposit is the riding distance.19
The riding distance is also shown in Fig. 3-6. Riding of glass par-
ticles was also observed on SEM pictures of glass particles incorpo-
rated in nickel deposits ͑Fig. 7͒. To quantify particle riding, the
relation between the riding distance and the deposit thickness was
determined by AFM measurements. These measurements were done
by measuring a force curve between a glass particle and a nickel
electrode before and after metal deposition ͑Fig. 8͒. When metal
deposits underneath the particle, the particle is pushed upward and
this results in a horizontal shift of the force curve to the right side.
The distance by which the force curve shifted is equal to distance by
which the particle was pushed up by the metal deposit. Figure 9
shows the riding distance as a function of deposit thickness. The
linear part of the curve in this figure corresponds to the situation
where particle riding occurs, and the slope of this linear part is the
ratio of the metal deposition underneath the particle to the average
metal deposition on the electrode. Because the slope of the linear
Figure 7. SEM picture of a nickel deposit with incorporated glass particle.
The deposit was made by adding glass particles to a nickel electrolyte and
starting deposition of nickel after all particles had settled on the electrode.
The picture shows a glass particle that was originally positioned on the
electrode ͑white circle͒, but the particle has moved up by metal deposition
underneath the particle.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

C476
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
Figure 10. Enlargement of the lower part of the particle shown in Fig. 7.
This figure shows that a hole has formed underneath the glass particle.
which had a radius of 21.25 ␮m. Hence, the metal deposition was
not conformal, with the shape of the particle and a hole formed
underneath the particle. Such holes were also seen in coatings made
by sedimentation codeposition ͑Fig. 10͒.
Figure 9. Riding of a glass particle as a function of nickel deposit thickness.
This figure shows that at a certain metal deposit thickness, riding starts to
decrease with deposit thickness and becomes irregular: sometimes the par-
ticle continues riding and sometimes it stops riding.
Shape of metal deposition determined by nickel–iron multilayer
deposition.— The metal deposit profile as a function of deposition
time was determined by nickel–iron multilayer deposition in order
to reveal the formation of holes during riding of glass particles and
more generally, to study the incorporation behavior of glass particles
where the formation of holes occurs.15 To compare the incorporation
of hydrophilic glass particles with hydrophobic particles, the metal
deposit profile as a function of deposit time was also determined for
hydrophobic, electrically conductive graphite particles and for hy-
drophobic, electrically insulating PMMA particles.
part has a value of 0.9, the thickness of the metal that deposited
underneath the particle is almost equal to the average metal deposit
thickness on the electrode. Figure 9 also shows that riding stops
after a certain deposit thickness. However, riding does not stop im-
mediately but starts to decrease after a certain deposit thickness. The
decrease is not constant but goes in steps: sometimes there is riding
and sometimes no riding occurs. From the cross sections of AFM
deposit profiles, it was checked whether metal growth was confor-
mal with the shape of the spherical particle once particle riding
stops. This was done by fitting a circle through the cavity in the
cross sections at the place where the particle was positioned during
metal deposition. Table II gives the depth of the cavities and the
radius of a circle fitted through the cavities for the two metal deposit
thicknesses and two current densities, and it also shows values of the
riding distance. The results in Table II show that particle riding
decreases with increasing current density. This might be because
deposits are less uniform at higher current densities. It is also shown
that the depth of the cavities increases with increasing current den-
sities, which might be because higher current densities result in less
uniform deposits and hence, in lower current densities underneath
the particle. For large current densities, the radius of the cavity is
smaller than for small current densities. In fact, the radius of the
cavity is smaller than the radius of the particle used for the deposits,
Glass particles.— A typical growth profile of nickel–iron multilay-
ers around a glass particle is given in Fig. 11. To visualize the metal
growth profile as a function of deposition time, the profile of the
metal deposit for different deposition times is indicated on the pic-
tures by white lines and the position of the particle at those deposit
times is drawn on the pictures as a black circle. Figure 11a-d shows
that the growth of metal around this glass particle can be divided
into two steps. The first step is shown in Fig. 11a-c. These figures
show that planar metal deposition occurred underneath the particle.
The position of the particle for different deposition times, which is
also drawn on the figures, shows that the particle was pushed up by
this planar metal deposition. This can be concluded because the
cross section goes through the center of the particle, which was
resting on the brass substrate at the start of the metal deposition and
is no longer in contact with the substrate after metal deposition
occurred. The pushing of particles by metal deposition underneath
them is called particle riding and was also observed by AFM mea-
surements, as shown in Fig. 9. The figure shows that the riding
distance is almost equal to the deposit thickness up to a certain
deposit thickness. This means that the metal deposit thickness un-
derneath the particle is almost equal to the average metal deposit
thickness on the electrode. This is in agreement with the fact that the
deposit stays almost flat during riding ͑Fig. 11a-c͒. Although the
above situation of particle riding was observed for all hydrophilic
glass particles, it cannot be ruled out that the metal deposition un-
derneath the particle could be an artifact. During nickel–iron depo-
sition, some gas bubbles are formed which could displace or even
lift up particles. When particles which are lifted up or moved side-
ways by gas bubbles settle on the electrode after a certain deposit
time, the result would be similar to those shown in Fig. 11. How-
ever, the fact that particle riding was also observed during AFM
measurements where the particle was fixed to the cantilever gives
strong evidence that particle riding is not an artifact, because the
Table II. Geometry of the cavity formed by metal deposition
around a glass particle and riding distance of the particle.a
Riding
distance
Deposit
thickness
i
Depth
͑nm͒
Radius
͑A/dm2͒
͑␮m͒
͑nm͒
500 nm
500 nm
2 ␮m
−0.5
−10
−0.5
−10
276 26
19.7 0.2
18.4 0.4
20.9 1.9
10.3 0.5
269
115
971
882
389
9
1132 173
1267 31
2 ␮m
a The results are shown for two deposit thicknesses and two current
densities. The values in this table are determined by fitting a circle
through cross sections of AFM images at angles of 0, 45, 90, and
135°. The values in the table are the average of four values and the
error on the values is the standard deviation.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
C477
Figure 11. Growth profile of metal around a glass particle visualized by deposition of nickel–iron multilayers. The distance between the different layers is a
measure for the local current density, and the different iron-rich layers show the profile of the metal deposit at different deposition times. ͑a–d͒ Deposit profiles
at different deposition times. The position of the particle at those different deposition times is drawn on the pictures as a black circle, and the shape of the metal
deposit is marked by white lines.
particle is not displaced laterally during the AFM experiments.
Moreover, during AFM measurements, no gas bubbles were formed
because electrolyte was flown through the AFM cell during the
metal deposit. A last argument that the observed particle riding is no
artifact comes from the fact that hydrophilic particles are not easily
lifted up by gas bubbles. To lift a glass particle, a gas bubble with a
size bigger than the size of the glass particle should attach to the
surface of the glass particle. This is not very probable because glass
particles are hydrophilic and hence do not easily attach to gas
bubbles.
The second step of metal growth around a hydrophilic glass par-
ticle is shown in Fig. 11d and occurs when the particle stops riding.
This happens after a certain metal deposit thickness, which depends
from particle to particle. When the particle stopped riding, metal
started to grow around the particle. Figure 11d shows that the dis-
tance between the multilayers, which is proportional to the local
current density, is larger in the vicinity of the contact point between
particle and electrode than far away from the particle. This results in
the upward bending of the multilayers in the vicinity of the particle.
Hence, the current density at the beginning of conformal metal
growth is higher in the vicinity of the contact point between particle
and electrode.
Besides the incorporation behavior of hydrophilic glass particles
as shown in the above paragraph, another route of particle incorpo-
ration which results in the formation of a hole underneath the hy-
drophilic glass particles was frequently observed. The growth profile
for this case is given in Fig. 12. The incorporation mechanism can
be divided into four steps which are schematically shown in a–d in
this figure. The first two steps are the same as explained in the
paragraph above: particle riding occurred initially and after a certain
deposit thickness, riding stopped. While the particle was riding, the
deposition of metal underneath the particle remained nearly flat ͑Fig.
12a͒. As soon as riding stopped, metal deposition around the particle
started and also in this case, a higher current density was observed at
the contact point between the deposit and the particle ͑Fig. 12b͒. The
difference between the incorporation behavior of hydrophilic glass
particles shown in the previous paragraph ͑Fig. 11͒ and the incorpo-
ration behavior shown in this paragraph lies in the fact that the
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

C478
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
Figure 12. Growth profile around a glass particle visualized by deposition of nickel–iron multilayers. The distance between layers is a measure for the local
current density, and the different iron-rich layers show the profile of the metal deposit at different deposition times. The position of the particle at different
deposition times is drawn on ͑a–d͒ as a black circle. The shape of the metal deposit is marked by white lines.
particle started riding again after some metal had deposited around
the particle ͑Fig. 12c͒. This is in agreement with AFM measure-
ments of particle riding ͑Fig. 8͒. Particle riding increases propor-
tional to the deposit thickness up to a certain critical deposit thick-
ness. For thicknesses larger than this critical deposit thickness,
particle riding starts and stops in an irregular way. When riding
starts again after some metal has deposited around the particle,
metal deposits in the cavity formed by conformal metal growth
around the particle ͑Fig. 12c͒. Because the rate of metal deposition
is almost equal everywhere on the electrode, the cavity remains but
its radius decreases. Therefore, the particle no longer fits in the
cavity and a hole is formed underneath the particle. When riding
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
C479
Figure 13. Growth profile around a small hydrophobic PMMA particle vi-
sualized by deposition of nickel–iron multilayers. For PMMA particles no
riding was observed and the metal started to grow around the particle from
the beginning of metal deposition. The current density in the vicinity of the
particle is equal to the average current density on the electrode, as shown by
the equal distance between different metal layers.
Figure 15. Deformation of an initially spherical PMMA particle due the
pressure exerted by the metal during deposition.
the center of the particle because the particle is touching the brass
substrate on the cross section. In both figures it is seen that metal
started to grow around the particle at the beginning of metal depo-
sition and that no particle riding was observed. This is because the
PMMA particles are hydrophobic and hence have no hydration
layer. Therefore, particles residing on the electrode are in contact
with the electrode and no metal can deposit underneath the particle,
as was the case for hydrophilic glass particles with a hydration layer.
In the case of small PMMA particles, the current density in the
vicinity of the particle is equal to the average current density on the
electrode ͑Fig. 13͒. This can be concluded from the fact that the
distance between the layers, which is proportional to the local cur-
rent density, is the same everywhere on the electrode. However, for
larger particles, an increased current density in the vicinity of the
particle is almost always observed ͑Fig. 14͒ at the beginning of
metal deposition. Hence, it can be concluded that the current density
in the vicinity of the particle depends on the size of the particle. The
increased current density in the vicinity of the contact point between
PMMA particles and electrode is less pronounced than with glass
particles. Because PMMA particles are electrically nonconductive,
the metal layers are perpendicular to the surface of the particle at
the top of the particle. This deposit profile corresponds with the
equipotential lines for the primary current distribution around an
electrically insulating particle. The metal deposit grows nicely
around the particle and no holes are formed between the deposit and
the particle.
finally stops, metal grows conformal with the shape of the particle
and the particle incorporates. As a result, a hole is left underneath
the particle because metal cannot deposit there since the electric
field is totally shielded. Simulations of the growth of metal around a
particle due to the secondary current distribution show that holes
can be formed at high exchange current densities.14 In this case,
the hole does not form underneath the center of the particle but
on the sides of the particle. However, this is in contradiction with
our experiments where the hole was always located underneath the
center of the particle. Therefore, the holes formed during metal
deposition underneath the glass particle in this work are probably
formed by the intermittent occurrence of riding and conformal metal
deposition.
PMMA particles.— Two typical growth profiles of nickel–iron de-
posits around a hydrophobic PMMA particle are shown in Fig. 13
and 14. These figures clearly show that the cross sections go through
Although the PMMA particles used to make the multilayer de-
posits were spherical, often nonspherical particles are found in the
cross sections ͑Fig. 15͒. This proves that the metal exerts stresses on
the particle during incorporation and that relatively soft PMMA par-
ticles deform under the influence of these stresses.
Graphite particles.— Figure 16 shows a typical deposit profile of
nickel–iron multilayers around a graphite particle. The graphite par-
ticle contacts the brass substrate which proves that the particle did
not ride but started to incorporate immediately. Because graphite
particles are electrically conductive, particles in contact with the
electrode probably have a potential close to the potential of the
electrode. Therefore, metal deposition occurs both on the particle
and the electrode, and hence, the growth profile follows the shape of
the particle. The current density in the vicinity of the graphite par-
ticle is slightly increased in the vicinity of the particle. This is shown
in Fig. 17, which is an enlargement of Fig. 16. However, the in-
creased current density in the vicinity of the graphite particle is less
pronounced than the increased current density in the vicinity of a
glass particle.
Figure 14. Growth profile around a large hydrophobic PMMA particle vi-
sualized by deposition of nickel–iron multilayers. For PMMA particles no
riding was observed and the metal started to grow around the particle from
the beginning of metal deposition. The current density in the vicinity of the
particle is slightly higher than the average current density on the electrode.
This is shown by the fact that the distance between different metal layers is
larger in the vicinity of the particle than on the rest of the electrode.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

C480
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
because the hydrated ion diameter is of the same order of magnitude
as the thickness of the hydration layer, it can be assumed that water
exchange plays an important role. Because the water exchange rate
of coordinated water on metal ions is fast ͓in the order of 104 s−1 for
Ni2+ and 106 s−1 for Fe2+ ͑Ref. 23͔͒, the diffusion of metal ions
through the hydration layer should not be hampered by the exchange
rate. Moreover, experimental investigations showed that the fluidity
of highly confined water molecules in thin gaps between hydrophilic
surfaces is equal to the bulk fluidity of water. Thus, confinement of
water molecules should not hinder diffusion of metal ions in the
hydration layer.24
Process archeology showed that the metal deposit underneath a
particle during riding is virtually planar and hence that the current
density underneath the particle is uniform. This can be understood if
one looks to the relative importance of the primary, secondary, and
tertiary current distribution. The Wagner number Wa is a measure
for the ratio of cathodic polarization resistance to ohmic
resistance:25
Figure 16. Growth profile around an electrically conductive graphite particle
visualized by deposition of nickel–iron multilayers. From the start of metal
deposition, metal grows on the particle and hence, the particle immediately
incorporates.
␬ d␩
Wa =
͓2͔
l di
with ␬ the electrolyte conductivity, d␩/di the slope of the cathodic
polarization curve, and l a characteristic length scale which, for the
present case, equals the particle radius. The conductivity of the
nickel electrolyte used for the experiments equals 4 ⍀−1 m−1. The
value of d␩/di was about 23.8 ⍀/m2, as determined from polariza-
tion curves. Hence, for a particle radius of 20 ␮m, a Wagner number
of 476 is found. This proves that the current distribution is governed
by polarization resistance and not by the electrolyte resistance. The
secondary current distribution originates from slow electrode kinet-
ics and makes the current distribution more uniform than the pri-
mary current distribution.14,25 However, the fact that the deposit
during riding is totally flat cannot be explained by the secondary
current distribution alone. Recent calculations of the current
distribution14 around a particle show that the secondary current den-
sity should be lower underneath the particle.
However, the deposit profiles around hydrophilic glass particles
also showed that once particle riding stopped, the current density in
the vicinity of the particle was larger than the average current den-
sity on the electrode. The higher current density near the particle
indicates that there is a mechanism that enhances mass transport in
the vicinity of the particle. It was investigated whether the enhanced
mass transport could be due to natural convection. Natural convec-
tion occurs if metal is deposited on an electrode which is positioned
at the bottom of the recipient and is facing upward, while the anode
is positioned at the top of the recipient. During metal deposition, the
concentration of ions at the electrode lowers and hence the density
of the electrolyte lowers. Therefore, the density at the bottom of the
recipient becomes lower than the density at the top, which results in
an unstable situation. Due to this, natural convection cells occur.18
The critical concentration difference at which natural convection
sets in is determined by the critical Rayleigh number
Discussion
AFM measurements and deposit profiles as a function of time
show that metal deposition occurred underneath glass particles. This
was called particle riding and was typical for glass particles which
are hydrophilic. For PMMA and graphite particles which are hydro-
phobic, no particle riding was observed. Therefore, it is assumed
that riding is caused by the hydration layer on hydrophilic glass
particles. This hydration layer remains on the hydrophilic glass par-
ticle even when it is resting on the electrode, hence, the hydration
layer keeps the particle and electrode separated by a small gap. The
hydration layer is probably not more than three water layers thick.
For a water molecule with a radius of 0.14 nm,20 the thickness of the
hydration layer can hence be estimated to be about 0.84 nm thick.
This value corresponds well with the thickness of hydration layers
on silica surfaces in 1 M CaCl2 solutions, which was found to be
0.9 nm.21 Nickel and iron ions which respectively have radii of
0.069 and 0.078 nm22 can diffuse through this hydration layer into
the gap between the particle and the electrode and can reduce to
form a metal deposit in the gap. Metal deposition in the gap pushes
the particle up by a distance which is almost equal to the average
deposit thickness on the electrode. The mechanism of ion diffusion
in the hydration layer is not clear. Since metal ions are hydrated and
⌬␳
g
h3
␳
0
Racrit
=
= 1770
͓3͔
␯D
with g the gravitational acceleration, ⌬␳ the density difference, ␳
0
the bulk density, h the thickness of the diffusion layer, D the diffu-
sion coefficient, and ␯ the kinematic viscosity. The critical concen-
tration differences for nickel and iron ions are 0.05 M and 7
ϫ 10−5 M, respectively, calculated for a diffusion layer thickness of
2 ϫ 10−4 m and an electrolyte composed of 0.2 M Ni͑NH2SO3͒2
and 0.05 M FeCl2. These concentration differences are reached dur-
ing deposition of the nickel and iron layer, so there will be natural
convection during metal deposition. To investigate whether natural
convection influences the metal deposition profile around a particle,
a very thin layer of nickel ͑50 nm͒ was deposited on an electrode
which was covered with graphite particles. The particles became
Figure 17. This photograph is an enlargement of Fig. 16 and shows that the
current density in the vicinity of the particle is slightly higher than on the rest
of the electrode. This is indicated by the larger distance between the different
metal layers in the vicinity of the particle than on the rest of the electrode.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
C481
Figure 19. Fluid stream lines caused by diffusio-osmosis. The vector plot on
the right side shows that the fluid flows toward the contact point between the
particle and the electrode and hence transports fresh electrolyte to the contact
point.
2
⑀z2␨
n
ץc
ץs
Figure 18. Growth profile around a graphite particle in the absence of natu-
ral convection. To avoid natural convection, the cathode was positioned at
the top of the recipient facing down while the anode was positioned on the
bottom. The current density is higher in the vicinity of the particle, which
indicates that there is enhanced mass transport near the particle.
=
͓5͔
vs
8␩
z2i ciϱ
͚
i
with ⑀ the dielectric permittivity, z the valency of the nickel ions, ␨
the ␨-potential, ␩ the dynamic viscosity, zi the valency of all ions in
the electrolyte, ciϱ the bulk concentration of ions in the electrolyte,
and s the tangential unit vector along the surface of the particle. For
typical values of the zeta potential, viscosity, concentration, and dif-
fusion layer thickness ͑40–400 ␮m͒, the velocity of the fluid flow is
about 4 ␮m/s. Hence, diffusio-osmosis is large enough to explain
the increased current density in the vicinity of the particle. The
streamlines and velocity vectors for such a flow are shown in Fig.
19. Because the concentration near the electrode is lower than in the
bulk, the tangential fluid flow goes in the direction of the contact
point between the particle and the electrode.
The fact that the local current density in the vicinity of glass
particles is higher than for PMMA and graphite particles might be
due to two factors. First, metal growth around glass particles only
started after an extended period of particle riding and hence, a con-
centration gradient has built up by the time the metal starts to grow
around the particle. In the case of PMMA and graphite particles,
metal started to grow around the particles right from the start of
metal deposition and hence, the concentration gradient and therefore
also the diffusio-osmotic flow is smaller in this case. Second, the
zeta potential of glass particles is probably higher than for PMMA
and graphite particles. From colloidal probe AFM measurements in
NaCl electrolytes, it was determined that the surface potential of
glass particles has a value between −40 and −80 mV, depending on
the electrolyte concentration, while hydrophobic PMMA and graph-
ite particles have a surface potential which is almost zero. Because
the zeta potential does not differ a lot from surface potentials deter-
mined by AFM, it can hence be concluded that there is more electro-
osmosis and diffusio-osmosis for glass particles than for PMMA and
graphite particles, which explains that the increased mass transport
is higher for glass than for PMMA and graphite.
attached to the electrode by this thin deposit. The deposition cell
was then turned upside down. For this electrode configuration, the
concentration during metal deposition ͑and hence the density͒ be-
comes lower at the top of the recipient than at the bottom, which
results in a stable situation without natural convection. From nickel–
iron multilayer deposits made in the absence of natural convection,
it was found that the increased current density in the vicinity of the
particle was still present ͑Fig. 18͒. Therefore, the enhanced mass
transport in the vicinity of the particles is not due to natural convec-
tion.
However, the electric field present in the electrolyte during metal
deposition acts on the ions present in the double layer of the particle
and sets up a tangential fluid flow at the outer boundary of the
double layer, which is proportional to the electric field. This is called
electro-osmosis. The velocity of the tangential fluid flow is given by
⑀␨ ץ␾
=
͓4͔
vs
␩ ץs
with ⑀ the dielectric permittivity, ␨ the zeta potential, ␩ the dynamic
viscosity, ␾ the potential, and s the unit vector tangential to the
particle. To investigate how large the tangential fluid velocity is for
the conditions used to prepare nickel–iron multilayers, the potential
gradient should be known. For a current density of 2 A dm−2 and an
electrolyte conductivity of 4 ⍀−1 m−1, a potential gradient of
50 V/m is obtained. For such potential gradient, the tangential fluid
velocity due to electro-osmosis has a value about 1 ␮m/s. This fluid
flow contributes to the enhanced mass transport in the vicinity of an
electrically insulating particle. However, an enhanced mass transport
was also observed for electrically conducting graphite particles in
contact with the electrode. For conducting particles in contact with
the electrode, the particles are at the same potential as the electrode
and therefore, ץ␾/ץs is zero everywhere on a conducting particle.
Hence, in this case, electro-osmosis cannot explain the enhanced
mass transport in the vicinity of the particle. However, during metal
deposition, a concentration gradient is present due to depletion of
metal ions in the vicinity of the electrode. Because the thickness of
the double layer scales inversely with ion concentration, it is thicker
on the side of the particle closest to the electrode than on the rest of
the particle. This imbalance in double-layer thickness on the particle
sets up a tangential fluid flow at the outer boundary of the double
layer which drives electrolyte from high to low concentration.26-30
This effect is called diffusio-osmosis. The velocity of this flow is
given by8
Conclusion
In this work it was shown that the way metal grows around
particles during electrodeposition greatly depends on its surface
properties. It was observed that hydrophilic glass particles are
pushed up by metal deposition underneath the particles before they
become incorporated. Regularly, formation of a hole underneath hy-
drophilic glass particles was observed. The formation of holes was
attributed to intermittent occurrence of particle riding and incorpo-
ration. Metal deposition does not take place underneath hydrophobic
PMMA and graphite particles, which incorporate from the start of
metal deposition. It was observed that there was an increased current
density in the vicinity of particles during the beginning of particle
incorporation. The increased current distribution was especially ob-
vious for hydrophilic glass particles, which showed riding before
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use

C482
Journal of The Electrochemical Society, 153 ͑7͒ C472-C482 ͑2006͒
9. C. Dedeloudis, Ph.D. Thesis, Katholieke Universiteit Leuven, Belgium ͑2000͒.
10. M. Kaisheva and J. Fransaer, J. Electrochem. Soc., 151, C89 ͑2004͒.
11. R. Socha, K. Laajalehto, and P. Nowak, Colloids Surf., A, 208, 267 ͑2002͒.
12. V. Terzieva, J. Fransaer, and J. P. Celis, J. Electrochem. Soc., 147, 198 ͑2000͒.
13. J. Walz, Adv. Colloid Interface Sci., 74, 119 ͑1998͒.
14. J. Lee and J. Talbot, J. Electrochem. Soc., 152, C706 ͑2005͒.
15. W. Wang, S. Leith, and D. Schwartz, J. Microelectromech. Syst., 11, 118 ͑2002͒.
16. W. Ducker, T. Senden, and R. Pashley, Nature (London), 353, 239 ͑1991͒.
17. T. Senden, Ph.D. Thesis, Australian National University, Australia ͑1993͒.
18. L. Stappers and J. Fransaer, J. Appl. Phys., 92, 5543 ͑2002͒.
19. C. Dedeloudis and J. Fransaer, Langmuir, 20, 11030 ͑2004͒.
20. F. Franks, Water: a Matrix of Life, 2nd ed., Royal Society of Chemistry, Cambridge
͑2000͒.
incorporation. Therefore it was assumed that the increased current
distribution in the vicinity of the particle was caused by diffusio-
osmosis, which occurs after a concentration gradient has built up.
Acknowledgments
This research was financed by a specialization grant of the Flem-
ish Institute for Enhancement of Scientific and Technological Re-
search in Industry ͑IWT͒. J.F. acknowledges support by the Fonds
voor Wetenschappelijk Onderzoek, Vlaanderen ͑FWO͒ by the KAN
1.5.012.03 contract.
21. M. Fielden, R. Hayes, and J. Ralston, Phys. Chem. Chem. Phys., 2, 2623 ͑2000͒.
22. Y. Marcus, J. Solution Chem., 23, 831 ͑1994͒.
23. S. Samson, L. Stillings, and C. Eggleston, Geochim. Cosmochim. Acta, 64, 3471
͑2000͒.
24. J. Klein, U. Raviv, S. Perkin, N. Kampf, L. Chai, and S. Giasson, J. Phys.: Con-
dens. Matter, 16, S5437 ͑2004͒.
25. J. Newman and K. Thomas-Alyea, Electrochemical Systems, 3rd ed., J. Wiley &
Katholieke Universiteit Leuven assisted in meeting the publication costs
of this article.
References
1. G. Barkleit, A. Grahl, M. Maccagni, M. Olper, P. Scharf, R. Wagner, and H.
Warlimont, J. Power Sources, 79, 73 ͑1999͒.
2. I. Garcia, A. Conde, G. Langelaan, J. Fransaer, and J. P. Celis, Corros. Sci., 45,
1173 ͑2003͒.
3. C. Wang, D. Wang, W. Chen, and Y. Wang, Wear, 253, 563 ͑2002͒.
4. M. Musiani, Electrochim. Acta, 45, 3397 ͑2000͒.
Sons, Inc., Hoboken, NJ ͑2004͒.
26. J. Anderson, M. Lowell, and D. Prieve, J. Fluid Mech., 117, 107 ͑1982͒.
27. D. Prieve, J. Anderson, J. Ebel, and M. Lowell, J. Fluid Mech., 148, 247 ͑1984͒.
28. J. Ebel, J. Anderson, and D. Prieve, Langmuir, 4, 396 ͑1988͒.
29. A. Brenner, Electrodeposition of Alloys, Principles and Practice, Vol. II, Academic
Press, New York ͑1963͒.
5. N. Guglielmi, J. Electrochem. Soc., 119, 1009 ͑1972͒.
6. J. P. Celis, J. Roos, and C. Buelens, J. Electrochem. Soc., 134, 1402 ͑1987͒.
7. J. Fransaer, J. P. Celis, and J. Roos, J. Electrochem. Soc., 139, 413 ͑1992͒.
8. J. Valdes, Ph.D. Thesis, Columbia University, New York ͑1987͒.
30. J. Happel and H. Brenner, Low Reynolds Number Hydrodynamics, Nordhoff Inter-
national Publications, Leiden, The Netherlands ͑1973͒.
Downloaded on 2013-11-29 to IP 35.8.11.2 address. Redistribution subject to ECS license or copyright; see ecsdl.org/site/terms_use