Acoustic streaming enhanced electrodeposition of nickel

Article abstract of DOI:10.1016/S0009-2614(02)01917-6

Electrochemical deposition of Ni from a Watts-type electrolyte under the influence of high frequency ultrasound at both high (250 W) and low (5-10 W) power sonication was investigated. An improvement in the material distribution of the deposited Ni in millimeter-sized groove-features on the cathode surface was observed. A theory based on mechanical interaction between organic additives adhering to the cathode surface and ultrasonically induced streaming-phenomena is presented here to account for the observed uniform filling behaviour. The present study further indicates a correlation between fringe-patterns on the surface of the deposit and near-boundary acoustic streaming.

Full text of DOI:10.1016/S0009-2614(02)01917-6

Chemical Physics Letters 368 (2003) 732–737
www.elsevier.com/locate/cplett
Acoustic streaming enhanced electrodeposition of nickel
a,b,
c,d
, Pawel Pocwiardowski , Per O.A. Persson ,
b
*
ꢀ
Jens A.D. Jensen
b
Lars Hultman , Per Møller
a
a
Department of Manufacturing Engineering and Management, The Technical University of Denmark, Kemitorvet,
b. 204, DK-2800 Kgs. Lyngby, Denmark
b
Thin Film Physics Division, IFM, Link €o ping University, S-581 83 Link €o ping, Sweden
Gdansk University of Technology, Faculty of Electronics, Telecommunications and Informatics, Gdansk, Poland
c
d
RESON A/S, Denmark
Received 3 October 2002; in final form 25 November 2002
Abstract
Electrochemical deposition of Ni from a Watts-type electrolyte under the influence of high frequency ultrasound at
both high (250 W) and low (5–10 W) power sonication was investigated. An improvement in the material distribution of
the deposited Ni in millimeter-sized groove-features on the cathode surface was observed. A theory based on me-
chanical interaction between organic additives adhering to the cathode surface and ultrasonically induced streaming-
phenomena is presented here to account for the observed uniform filling behaviour. The present study further indicates
a correlation between fringe-patterns on the surface of the deposit and near-boundary acoustic streaming.
Ó 2003 Elsevier Science B.V. All rights reserved.
1
. Introduction
providing both increased mass transport of reac-
tive species to the cathode as well as mechanically
affecting the electrochemical reactions taking place
on the cathode.
Local mass transport at the cathode surface is a
critical parameter in electrochemical deposition of
metals and alloys. The depletion of the electrolyte
near the cathode due to insufficient mass transport
can lead to unwanted side-reactions apart from the
electrodeposition and in worst case the deposition
of metal hydroxide – so called ÔburningÕ of the
electrodeposit. Ultrasonic agitation offers a level of
control of the electrochemical deposition mecha-
nism at the cathode, which is beyond most tradi-
tional agitation-techniques used today by
The effects of ultrasound on electrochemical
deposition processes have been the topic of several
recent studies. Suni et al. [1,2] studied the effects of
high frequency (0.7 MHz) high intensity (up to
2
15 W/cm ) sonication on electrodeposition of Cu.
The study showed an improvement of the filling
capacity of 1 lm deep and 0.2–2.0 lm wide in-
terconnect vias applying sonication. The size of
these vias was of the same order as the extension of
the ultrasonic boundary layer close to the sub-
strate, where micro-streaming is produced by the
high intensity ultrasound.
*
Corresponding author. Fax: +45-45-93-62-13.
E-mail address: jdj@ipl.dtu.dk (J.A.D. Jensen).
0
009-2614/03/$ - see front matter Ó 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0009-2614(02)01917-6

J.A.D. Jensen et al. / Chemical Physics Letters 368 (2003) 732–737
733
The present study adresses the effect of 1 MHz
ultrasound on the electrodeposition of Ni into
machined via-structures of millimeter-size – i.e.,
three orders of magnitude larger than the thickness
of the acoustic boundary layer.
It is a first report of the application of high
frequency ultrasound in the enhancement of nickel
electroforming. It is shown that ultrasonic stream-
ing can improve the uniformity of Ni grown in a
modified Watts electrolyte into via structures as
well as affecting the surface topography of the
Ni-deposit.
Electrodeposition at high power sonication was
performed in the transducer near field, i.e., in the
region where the pressure gradient from the
transducer does not reach a steady state value.
The cathode was placed approximately 20 cm from
the ultrasonic transducer in a plating cell measur-
3
ing 106 ꢀ 56 ꢀ 45 cm .
The low power transducers were piston-shaped
with a diameter D ¼ 2 Á a ¼ 25 mm and were op-
erated at a wavelength k ¼ 1:5 mm, giving the
ratio a=k ¼ 8:3. The acoustic pressure field or
particle velocity along the beam axis of the low
power transducer can be approximated by [3]
Electrodeposition (ED) of Ni was carried out
in a modified Watts electrolyte. Bath composi-
tion and deposition parameters are presented in
Table 1.
ꢂ
ꢂ
ꢂ
2
0sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
13ꢂ
ꢀ
ꢁ
ꢂ
2
ꢂ
p
D
2
ꢂ
pðzÞ ¼ p
ꢂ
4
@
2
A5ꢂ
0
2 sin
Á
þ z ꢁ z
;
ð1Þ
ꢂ
ꢂ
k
3
ꢂ
Brass slabs measuring 70 ꢀ 60 ꢀ 10 mm served
as substrates. Vias of 1 mm width, 20 mm length
and depths varying from 0.5 to 2 mm were ma-
chined in the substrate by milling in order to test
the filling capability of the electrodeposited Ni
with and without the use of ultrasonic streaming.
Two kinds of 1 MHz transducers were used in
the present study: a high power type, operating
with an electrical input power of 250 W and a
low power type having an electrical input power
of 2–6 W.
where ÔzÕ is the distance away from the transducer
along the beam axis and p is the pressure at the
0
transducer surface.
Electrodeposition under low power sonication
was performed in the transducer far field.
A standing wave was formed by combining two
low power transducers facing each other at a dis-
tance of about 25 cm. Positioning the substrate in
the central region between the two transducers
during electrodeposition, the cathode surface is
sonicated uniformly within the standing wave re-
gion. Near-boundary streaming occurs along the
cathode surface, characterised by the development
of local cylindrical vertices, which act as local mass
transport pumps for ions and other species in the
electrolyte [4].
For investigation of the filling capacity of the
vias on electrodeposition, light optical microscopy
was applied. As deposited samples were cut in a
direction perpendicular to the grooves. The cross
sections were polished on Silicon carbide paper,
followed by fine polishing using 1-lm diamond
paste.
2
The high power sonicator measured 15 ꢀ 15 cm
and consisted of four individual sources of ultra-
sound (transducers), operated in switch-mode.
Table 1
Electrolyte composition and electrodeposition conditions
Bath Constituent
Amount (g/l)
46.0
Nickel chloride hexahydrate
(
NiCl
Nickel suphate heptahydrate
NiSO O)
Á 7H
Boric acid (H BO
2
Á 6H
2
O)
367.5
(
4
2
3
3
)
44.0
0.1
Sodium lauryl sulphate (NaC12
Naphthalene trisulphonic acid
4
H25SO )
3.0
(
C
10
H
5
Na
3
O
9
S
Butyne diole (C
3
Á H
2
O)
H
6
O
2
)
0.3
The use of high acoustic intensity and high
frequency (1 MHz) causes Eckart streaming [5]
with a streaming speed in the range 20–40 cm/s.
The liquid in motion contained tiny air bubbles,
created by gaseous cavitation [6]. The bubbles
enhance acoustic streaming, leading to quasi
acoustic streaming [7]. In the conducted experi-
4
Operating conditions:
Temperature ¼ 323 K (50 °C)
pH ¼ 4.0–4.2
2
Current density (DC) ¼ 200 A/m
Continous filtration
(1 lm cartridge filter)

734
J.A.D. Jensen et al. / Chemical Physics Letters 368 (2003) 732–737
Fig. 1. Ni-topography caused by high power, high frequency sonication (pits are indicated by arrows).
ments gas bubbles appeared at a short distance
from the transducer (5–10 cm) and grew in size
while following the streaming direction, where
they adhered to the substrate. The pits in the
surface morphology of the finished Ni-deposit
from the right- to the left-hand side of the shown
sample. A fringe-pattern is observed, both to the
left and to the right of the machined trenches. The
width of the individual fringes was ꢂ0.5 mm, i.e.,
approximately half the wavelength of the ultra-
sonic wave.
(
Fig. 1) are attributed to persistent adherence of
gas-bubbles, which locally hinder Ni-deposition.
The geometry of the individual pits with pro-
truding edges from the deposit surface indicates
that velocity enhancement of the ultrasonic waves
has taken place at the location of the adhering
bubbles. This increased hydrodynamic activity
caused by bubbles and/or other geometrical ob-
stacles and the ultrasonic field is referred to as
microstreaming [8].
Fig. 2b shows the morphology of the deposit
resulting from the standing wave between two low
power transducers facing each other at a distance
of 25 cm. Well-defined parallel fringes are ob-
served on the surface of the deposited Ni in the
region covered by the acoustic beam. The ap-
pearance of pits in the surface topography on the
Cavitation caused by the high power transducer
hindered measurement of the exact velocity of the
streaming liquid with laser doppler anemometry.
Therefore, the flow rate was estimated by obser-
vation of the bubble-motion. The observation of
pits (Fig. 1) is believed to be attributed to two
main factors: (1) the formation of gaseous cavita-
tion, causing air bubbles to reach the cathode
surface and (2) the possible detrimental effect of
the ultrasound on the functionality of the surfac-
tant Lauryl sulphate in the electrolyte (Table 1),
which would normally hinder any bubbles in ad-
hering to the cathode surface.
For experiments with one low-power transducer
5–10 W, 1 MHz), the substrate was inserted ver-
(
tically into the electrolyte and inclined approxi-
mately 10° off the ultrasonic beam axis to allow the
soundwaves to interact with the machined trenches
in the substrate.
The resulting deposit morphology is shown in
Fig. 2a. The transducer generated a wave traveling
Fig. 2. (a) Deposit morphology caused by low power travelling
wave 1 MHz sonication. (b) Deposit morphology caused by low
power standing wave 1 MHz sonication.

J.A.D. Jensen et al. / Chemical Physics Letters 368 (2003) 732–737
735
Ni deposited under influence of a standing wave
Fig. 2b) as opposed to a seemingly pit-free surface
(
of Ni deposited under the influence of a travelling
wave (Fig. 2a) is in good agreement with the pre-
vious mentioned theory on gas bubble formation
and surfactant interruption with increased ultra-
sonic intensity, since the particle velocity in the
standing wave is twice as high as that of the
travelling wave.
Typical results of cross sections observed with
LOM are shown in Fig. 3. Whilst the Ni layer
deposited without sonication gives a thick Ni-film
in the bottom of the groove and thin walls fol-
lowed by a very thin film at the edges on the
substrate surface, the Ni layer deposited with
sonication has a uniform thickness and the sides of
the grooves are terminated by almost 90° angles.
No voids or inclusions could be detected in optical
microscopy or transmission electron microscopy
Fig. 4. Cross section TEM-micrograph of electrodeposited Ni
deposited under low power (5 W) sonication. The Ni coating
exhibits a dense, polycrystalline structure with a grain size in the
order of 200 nm.
of the cathode geometry, by adhering to active
sites for nucleation and growth. With the intro-
duction of ultrasonic streaming, the adherence of
the butyne diole is affected, so that the inhibition
effect of the additive is decreased (Fig. 3b), (2) with
the enhanced electrolyte exchange in the groove
geometry on the cathode, a more uniform func-
tionality of the additives is obtained. The last of
the two theories seems to be the most probable
given the relatively low ultrasonic intensity used,
and it is furthermore in good agreement with ob-
servations made by Jordan and Tobias [9] on the
effect of inhibitor transport on levelling in elec-
trodeposition.
(
Fig. 4).
The transducer nearfield is characterised by
abrupt sound pressure changes (Fig. 5a). With the
high power sonication system used, high acoustic
intensity could be obtained, inducing near-
boundary streaming and hereby enhanced elec-
trolyte exchange in the vicinity of the cathode
surface. The method successfully improved uni-
formity of the deposited nickel film thickness in
the millimeter-sized trenches (Fig. 3). Two possible
explanations for this have been considered: (1) in
the modified Watts electrolyte used (Table 1) the
organic additive bytyne diole normally acts by
inhibiting continued crystal growth on high cur-
rent density areas (Fig. 3a) like corners and edges
Application of this improved process for con-
trolling the effect of the inhibiting additives is ex-
pected in the electroforming of MEMS-structures
[
10–16].
The standing wave formed between two op-
posed 1 MHz transducers (Fig. 5b) caused very
strong differences in local mass transport as seen
by the patterns on the surface of the Ni-deposit
(Fig. 2b). It is interesting to see also that even the
much weaker standing waves forming by interac-
tion between the direct and the reflected sound-
waves from the grooves in the substrate also cause
clearly visible patterns in the deposit topography
(Fig. 2a).
The explanation for the observed scattering
fringes originating from the machined trenches in
Fig. 3. (a) Ni deposited without sonication, (b) Ni deposited
with low power (5 W) sonication from one transducer.

736
J.A.D. Jensen et al. / Chemical Physics Letters 368 (2003) 732–737
Fig. 5. (a) Axial pressure envelope from piston transducer vs. axial distance. (b) Standing wave from two opposite 1 MHz transducers.
Electrodeposition is carried out in the far field as indicated by the sample position.
the substrate (Fig. 2) could be the interaction be-
tween the incident and the scattered waves. As
stated by Morse [17], interaction phenomena can
take place when an obstacle on the waveÕs path is
comparable in size with the acoustic wavelength.
The machined trenches in the used substrates have
dimensions of approximately 1 mm and the
wavelength of the used ultrasound is about 1.5
mm, so the requirement for strong interaction is
fulfilled.
The separation between the fringes on the
samples shown in Fig. 2 is approximately 0.8 mm –
close to half the used acoustic wavelength, in
agreement with the near-boundary streaming the-
ory [18].
tance in future advanced electroforming applica-
tions.
The topography of the Ni-deposit is markedly
influenced by the use of ultrasound, with varying
fringe patterns forming on the surface of the de-
posit as a result of near boundary streaming and
standing wave phenomena. As such electrodepos-
ition can be used as a tool for visualising streaming
phenomena caused by ultrasound near a solid
boundary surface. Acoustic streaming enhanced
electrodeposition (ASEED) as a technique there-
fore shows great promise for various applications.
Acknowledgements
The results shown here indicate that there is a
direct relationship between near-boundary stream-
ing and the fringes developing during electrode-
position. The latter are obviously controlled by
local mass transport phenomena. As such electro-
deposition might be a useful tool to visualize com-
plex scattering behaviour of ultrasonic waves. The
observed strong effect of the ultrasonic streaming
on deposit topography and filling characteristics,
together with the improved understanding of the
process offer great promise for improved tailoring
of specific metallic microstructures.
This work was carried out under the European
fifth framework program under the targeted re-
search area Competitive and Sustainable Growth.
The authors gratefully acknowledge the financial
support given by the European Commission to
the project ULTRAPLATE – ultrasonic plating
technique for low cost, clean, high-speed, high
precision plating for electronic components and
micro-systems, contract No. G1RD-CT-2000-
00427. Professor Marcel Somers, The Technical
University of Denmark, Department of Manu-
facturing Engineering and Management is kindly
acknowledged for his critical reading of this
manuscript.
2
. Conclusion
High power high frequency ultrasound has a
marked effect on the electrodeposition of Ni from
a modified Watts electrolyte.
Filling characteristics on a 3D-patterned cath-
ode-geometry with millimeter-sized features is
improved considerably. This could be of impor-
References
[
1] I.I. Suni, C.S. Tiwari, A.A. Busnaina, H. Lin, H.V.
Reynolds, C. Neely, in: Proceedings from the 198th
Meeting of the ElectroChemical Society, Oct. 22–27 2000,
Phoenix, AZ, 2000.

J.A.D. Jensen et al. / Chemical Physics Letters 368 (2003) 732–737
737
[
2] I.I. Suni, C.S. Tiwari, A.A. Busnaina, V. Reynolds, C.
Neely, in: Proceedings from the AESF SURFIN 2001
conference, June 25–28 2001, Nasville Tennesee, 2001.
3] J. Krautkr €a mer, H. Krautkr €a mer, Ultrasonic Testing of
Materials, fourth ed., Springer, Berlin, 1990.
[10] J.P. Rasmussen, P.T. Tang, C. Sander, O. Hansen, P. Møller,
in: TransducersÕ 97, Chicago, June 16–19, 1997, p. 463.
[11] P.T. Tang, PhD-thesis: ÔFabrication of Micro Components
by Electrochemical DepositionÕ, DTU, March 1998.
[12] W. Ruythooren, K. Attenborough, S. Beerten, P. Merken,
J. Fransaer, E. Beyne, C. Van Hoof, J. De Boeck, J.P.
Celis, J. Micromech. Microeng. 10 (2000) 101.
[13] S. Abel, H. Freimuth, H. Lehr, H. Mensinger, J. Micro-
mech. Microeng. 4 (1994) 47.
[
[
4] P. Pocwiardowski, L. Bjorno, A. Stepnowski, Hydroacous-
tics, Annual Journal, Vol. 4, Gdynia 2001, 205.
5] C. Eckart, Phys. Rev. 73 (1) (1948).
[
[
6] H.G. Flynn, in: W.P. Mason (Ed.), Physical Acoustic,
part A, vol. II, Academic Press, New York, 1965, p.
[14] J. Gobet, F. Cardot, J. Bergqvist, Rudolf, J. Micromech.
Microeng. 3 (1993) 123.
265.
[
[
7] Mitome, 13th International Symposium on Nonlinear
Acoustics, Bergen, 1993, Publ. World Scientific.
[15] W. Qu, C. Wenzel, A. Jahn, J. Micromech. Microeng. 8
(1998) 279.
8] W.L. Nyborg, in: W.P. Mason (Ed.), Physical Acoustic,
part B, vol. II, Academic Press, New York, 1965, p.
[16] V. Sommer, Galvanotechnik 91 (3) (2000) 791.
[17] P.M. Morse, K.U. Ingard, Theoretical Acoustics, Mc-
Graw-Hill, New Jersey, 1986.
2
65.
9] K.G. Jordan, C.W. Tobias, J. Electrochem. Soc. 138 (5)
1991) 1251.
[
[18] H. Schlichting, Boundary Layer Theory, McGraw-Hill,
New York, 1955.
(