Microstructured metallization of insulating polymers

Article abstract of DOI:10.1016/S0013-4686(03)00369-4

Direct Ni electrodeposition on insulating polymers by the so-called PLATO technique is studied and the application of this technique for microstructured metallization is investigated. Propagation behavior, surface morphology, conductivity and thickness of a deposited metal layer are characterized using microscopy, AFM, four-point conductivity and XPS sputter measurements. Two layers are formed during metal deposition: primary layer and secondary layer. Both layers propagate with constant rates during the first 60 s and the propagation rates are influenced by the metallization potential. The primary layer has hemispherical morphology, low conductivity and an uneven thickness of about 25-100 nm. The secondary layer has the repetition morphology of the primary and higher roughness (R_a^prim = 40 nm, R_a^sec = 150 nm), higher conductivity (σ_sec/σ_prim = 10⁸/10¹⁰) and a thickness of 100-200 nm. The high lateral propagation rate of the metal strip during metal deposition offers possibilities for metallization of insulating microstructures. Routines for microstructured metallization using PLATO technique are proposed and examples for the applications are demonstrated.

Full text of DOI:10.1016/S0013-4686(03)00369-4

Electrochimica Acta 48 (2003) 3021ꢀ3027
/
www.elsevier.com/locate/electacta
Microstructured metallization of insulating polymers
a,
T.T. Mai *, J.W. Schultze , G. Staikov
a
b
a
Institut f ¨u r Physikalische Chemie und Elektrochemie, Heinrich-Heine Universit ¨a t D ¨u sseldorf, D-40225 D ¨u sseldorf, Germany
b
Institut f ¨u r Schichten und Grenzl ¨a chen, Forschungzentrum J ¨u lich, D-52425 J ¨u lich, Germany
Received 20 December 2002; received in revised form 11 April 2003; accepted 14 April 2003
Abstract
Direct Ni electrodeposition on insulating polymers by the so-called PLATO technique is studied and the application of this
technique for microstructured metallization is investigated. Propagation behavior, surface morphology, conductivity and thickness
of a deposited metal layer are characterized using microscopy, AFM, four-point conductivity and XPS sputter measurements. Two
layers are formed during metal deposition: primary layer and secondary layer. Both layers propagate with constant rates during the
first 60 s and the propagation rates are influenced by the metallization potential. The primary layer has hemispherical morphology,
low conductivity and an uneven thickness of about 25ꢀ
/
100 nm. The secondary layer has the repetition morphology of the primary
8
prim
sec
10
and higher roughness (/
Ra
/
ꢁ/40 nm, Ra
/
ꢁ
/
150 nm), higher conductivity (ssec/sprim
ꢁ
/
10 /10 ) and a thickness of 100ꢀ200 nm. The
/
high lateral propagation rate of the metal strip during metal deposition offers possibilities for metallization of insulating
microstructures. Routines for microstructured metallization using PLATO technique are proposed and examples for the
applications are demonstrated.
#
2003 Elsevier Ltd. All rights reserved.
Keywords: Metal deposition; Insulating polymers; Cobalt sulphide; Microstructuring; PLATO technique
1
. Introduction
metal deposition in microsystems. The key step in the
direct metallization methods consists of the activation of
the surfaces by a non-metallic substance which is a metal
Microstructured metallization of insulating substrates
by localization of electrical contact is an interesting
prospective in microsystem technology [1]. Until now,
the procedures used for metallization of insulating
substrates in microsystems must follow several steps:
pre-deposition of the metal by physical vapor deposition
sulphide in the so-called PLATO technique [8ꢀ10]. An
/
interesting feature of the activated surface is that the
metal sulphide is not an intact, continuous film covering
completely the polymer surface, and the degree of
activation is crucially influenced by pre-conditioning
and by the nature of the polymer [11]. Therefore, the
beginning stages of metallization do not follow the well-
studied mechanisms of metal deposition on metal
substrates. This leads to characteristic properties of the
deposited metal layer and to specific electrochemical
behaviors of the metal deposition processes in macro-
and microscales.
(
PVD) or electroless metal deposition to form a
conducting layer and subsequent electrodeposition.
The microstructuring is achieved by mask and lithogra-
phy [2,3]. Recently, new methods using a local electrical
contact for direct electrodeposition of metals onto
modified insulating substrates (direct galvanic metalliza-
tion) have been proposed [4ꢀ/10]. The application of the
In this paper, we show studies of Ni layers deposited
local contact offers possibilities for localization of the
on acrylonitrileꢀ
/
butadieneꢀstyrene (ABS) polymers
/
using cobalt sulphide activation. The electrochemical
behavior of the deposition process is examined. The
application of PLATO technique for microstructured
metallization is also reported.
*
Corresponding author. Tel.: ꢂ
112803.
E-mail address: maith@uni-duesseldorf.de (T.T. Mai).
/
49-211-8113697; fax: ꢂ49-211-
/
8
0
013-4686/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0013-4686(03)00369-4

3022
T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ3027
/
2
. Experimental
activated polymer surface and to induce electrodeposi-
tion. The contact area of the Pt tip was shielded by an
appropriate insulation with Apiezon wax so that only
†
The deposition was performed on ABS substrates
supplied by Good Fellows. The process of direct
metallization consists of three main steps: etching,
activation and metal deposition. The activation is based
on the so-called adsorption concept and includes two
steps: adsorption of cobalt(II) and sulphidication. The
conditions of the process steps are given in Table 1.
After each step, the samples were thoroughly cleaned in
the contact area was exposed. The masks for micro-
structuring were laser-structured polyimide (PI), sup-
plied by the company Bartels GmbH (Dortmund,
Germany).
3. Results and discussion
†
Millipore water.
The lateral propagation of the metal layer was studied
using optical microscope (PM-6; Karl Suss, USA). The
topography and roughness of the deposited Ni surfaces
were measured by atomic force microscopy (AFM)
using commercial equipment (PicoSPM, Molecular
Imaging, and Nanoscope E Controller; Digital Instru-
ments, USA). The conductivities were measured using a
3
.1. Characterization of electrodeposited metal layers
The electrodeposition of Ni on the activated surface
takes place by fast lateral propagation of the overlayer.
As shown in the micrograph (Fig. 1), two layers are
generally formed during the propagation: a primary and
a secondary layer. Fig. 2 shows the changes of the
distances to the Pt tip from the edge of the primary and
four-point method with the amperometer having a
ꢃ12
current limitation of 10
A. The surface composition
of the secondary layers during propagation at Eꢁ
/
ꢃ0.6
/
and the nature of the chemical states were investigated
by X-ray photoelectron spectroscopy (XPS) using a
modified ESCALAB-V (VG-Instruments, UK) with
polychromatic AlKa radiation, CAE mode, a pass
energy of 50 eV for survey scans and 20 eV for specific
element scans. The base pressure of 10ꢃ10 mbar with
polychromatic Al Ka radiation (1486.6 eV), CAE mode,
a pass energy of 50 eV for survey scans and 20 eV for
specific element scans. Ion etching for sputter profile
V. Results show that both distances increase linearly
with time after the first 60 s, indicating that the
propagations of the primary and the secondary layers
take place with constant rates V?
should be noted that after extended deposition (t ꢀ
and Vƒ
; respectively. It
180
x
x
/
s), propagation rates of the primary and the secondary
layers are no longer constant and decrease with time.
This can be explained by the increasing potential drop
under these conditions. Fig. 3 shows the influence of the
measurement was carried out by an AG-S2 ion source
ꢃ5
with an Ar pressure of 1.5ꢄ
/
10
mbar and an
acceleration potential of 4 kV. The sputter rate was
1
.6 nm/min, based on Ta O5 standard. A computer
2
program for peak deconvolution (Unifit V.2.4; Univer-
sity of Leipzig, Germany) was used to analyze quantita-
tively XPS peaks. The interface of the deposited Ni layer
and the polymer substrate was characterized by a cross
section prepared by chemical etching (acetic acidꢀ
/
perchloric acid).
Electrochemical metal deposition was performed in a
conventional three-electrode electrochemical cell with a
Hg/Hg SO /1 M Na SO reference electrode and a Ni-
2
4
2
4
plate counterelectrode. All electrode potentials are
referred to the potential of the standard hydrogen
electrode (SHE). A Pt tip was used to contact the
Fig. 1. Micrographs of the primary and the secondary layers during
metal deposition (Eꢁ
/
ꢃ
/
0.6 V, tꢁ5 s).
/
Table 1
Conditions for process steps of the technique
Process
Solution
CrO /H
PLATO Aktivierung (CoSO
Temperature (8C)
pH Time (min)
Etching
Activation
3
2
SO
4
65
ꢀ/
5
TM
Adsorption
Sulphidication
Electrodeposition
4
×
/
7H
2
O, NH , oxidizer)
3
Room temperature
Room temperature
45
10
3
1
TM
PLATO Vernetzung (Na
2
S)
150 g/l NiSO 6H O, 30 g/l NiCl
ꢀ/
4
×
/
2
2
×
/
6H O, 30 g/l H
2
3
BO
3
4
ꢀ/

T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ
/
3027
3023
ciency [11,12]. Results show that at E ꢀ
/
ꢃ
/
0.8 V, the rate
V? is larger than Vƒ and both layers can be observed in
x
x
the micrograph (Fig. 1), whereas V? is equal to Vƒ at
x
x
E B
/
ꢃ0.8 V, leading to the complete covering of the
/
secondary layer on the primary sublayer during propa-
gation and only the secondary layer is observed in
microscope. It should be noted that the lateral propaga-
tion rates, V? and Vƒ; are higher than V by a factor of
x
x
z
4
about 10 , slightly increasing with decreasing potential.
This offers possibilities for metallization of microstruc-
tures.
Fig. 4 shows typical AFM topographies of the
activated area, the primary layer and secondary layer
obtained after 60 s deposition at Eꢁ
sulphide clusters with average roughness, Ra ; of 22 nm
can be observed on the activated surface. The primary
/
ꢃ0.6 V. Cobalt
/
prim
Ni layer shows hemispheres with a height of 10ꢀ
and a diameter of 0.5ꢀ2 mm. The secondary Ni layer
owns a repetition topography of the primary one with a
/50 nm
/
prim
sec
higher roughness (/
Ra
/
ꢁ
/
40 nm, Ra
/
ꢁ150 nm), but
/
lower density of hemispheres. This repetition behavior is
expected, since the deposition of the secondary Ni layer
follows the mechanism of electrodeposition on a metal
substrate. In the absence of additives, the morphological
repetition of substrates for electrodeposited metal layers
is well-known [13]. The conductivity measurements
carried out on surfaces of the three zones are shown in
Fig. 5. The conductivity of the secondary Ni layers is
clearly higher than that of the primary (ksec/kprim
8
10 /
ꢁ
/
1
0 ). This result indicates that the secondary Ni layer is
0
1
remarkably thicker than the primary one. The conduc-
tivity of the activated zone cannot be measured due to
the fact that cobalt sulphide is in the form of discrete
clusters and not as a continuous layer on the surface
Fig. 2. (a) Schematic description of the Ni layer propagations. (b) The
change of the distance from Pt tip to the end of the primary and the
secondary layers during metal deposition (Eꢁ
/
ꢃ
/
0.6 V, tꢁ5 s).
/
[
11].
XPS sputter profiles of the primary and secondary Ni
layers are shown in Fig. 6. It should be noted that
during sputtering, the Ni layer and the polymer surface
are etched with different rates (average sputter rates of
the Ni layer n ꢁ
/
1.6 nm/min and of the polymer n ꢁ6
/
Ni
Ni
nm/min). On the unsputtered surface of the primary
layer (Fig. 6a), the concentrations of Co and S are zero,
indicating that the primary Ni layer is a continuous
layer and the activated surface underneath this layer is
not exposed. The first Co and S signals are recorded
after 15 min of sputtering, corresponding to the removal
of the thinnest part with thickness of about 25 nm. It is
interesting to note that during the first 15ꢀ60 min
/
sputtering, all Ni, Co and S signals appear with
concentrations decreasing gradually with time. This
result indicates that the thickness of the primary layer
is not evenly distributed, e.g. thinner parts were etched
away faster. After sputtering for 60 min, Ni is com-
pletely removed, corresponding to a thickness of about
100 nm. The corresponding thickness of the primary
Fig. 3. Influence of the metallization potential on the propagation
rates of the primary layer (V?); the secondary layer (Vƒ) and the
/
/
x
x
z
growth of metal in the z-direction (V ).
potential E on V? and Vƒ; and the growth of the Ni
x
x
layer in the z-direction, V , which is calculated from the
current transients and the corresponding current effi-
z
layer is thus in the range of about 25ꢀ100 nm (Fig. 6a).
/

3024
T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ3027
/
Fig. 4. AFM images of the activated surface, and the primary and the secondary Ni layers. The Ni layer is obtained after deposition with Eꢁ
/
ꢃ/0.6 V
and tꢁ60 s.
/
primary layer is thus attributed to the Ni deposition
reaction on cobalt sulphide clusters. The different sizes
of the sulphide clusters lead to the uneven distribution
of the primary thickness, e.g. the Ni layers are thicker on
larger cobalt sulphide clusters. On the other hand, the
secondary layer is formed on the primary one by the
mechanism of metal deposition on a metal substrate.
Since the propagation of the secondary layer occurs very
fast at E 5
/
ꢃ0.6 V, a thick deposited Ni layer has an
/
even thickness distribution (Fig. 7a), which is required
for metallization of insulating microstructures. At more
positive potential of E ꢀ
/
ꢃ
/
0.6 V and low degree of
activation (coverage of cobalt sulphide, u B
/40%), the
thickness decreases with distance from contact (Fig. 7b).
Therefore, for microstructured metallization, the deposi-
tion potential and activation should be optimized in
order to obtain metal microstructures with even thick-
ness.
Fig. 5. Conductivities of the activated surface, and the primary and
the secondary Ni layers. The Ni layer is obtained after deposition with
Eꢁ
/
ꢃ
/
0.6 V and tꢁ60 s.
/
3
.2. Microstructured metallization
Further sputtering yields the decreasing of Co and S
concentrations until they vanish after 180 min. Fig. 6b
shows the sputter profile of the secondary layer. The
same sputter behavior of the primary layer is also
observed for the secondary layer. However, the time
for an appearance of the first Co signal is much higher
In principle, the microstructured metallization apply-
ing PLATO technique can be carried out using post-
structuring concept. A typical example for this routine is
the damascene process (IBM) [14], in which the direct
metallization method is considered as a substitutional
method for electroless deposition and PVD. In this
study, we propose two new approaches for microstruc-
tured metallization considering the specific properties of
the direct metallization using metal sulphide (Fig. 8):
(
after 60 min sputter), corresponding to the higher
thickness range of 100ꢀ200 nm (Fig. 6b).
/
It is suggested that the primary and secondary Ni
layers are formed following different mechanisms. On
the primary layer, the appearance of hemispheres
indicates that the Ni is deposited preferentially at certain
positions. Recent studies show that the reduction of
a) homogeneous activation and selective metallization
using a mask and
cobalt sulphide at potentials E B
deposition reaction of Ni [11]. The formation of the
/
ꢃ
/
0.4 V catalyzes the
b) microstructured activation (pre-structuring) and
subsequent electrodeposition.

T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ
/
3027
3025
Fig. 7. (a) Even thickness distribution of an electrodeposited Ni layer
obtained at Eꢁ 0.6 V, tꢁ120 min. (b) Thickness profile obtained on
surface low coverage of cobalt sulphide (uꢁ40%) at Eꢁ 0.5 V and
tꢁ120 min.
/
ꢃ
/
/
/
/
ꢃ
/
/
Fig. 6. XPS sputter profile of (a) the primary layer and (b) the
secondary layer. The Ni layer is obtained after deposition with Eꢁ
.6 V and tꢁ60 s.
/
ꢃ
/
The mask, which has the shape of a Christmas tree, was
fixed on the activated ABS surface (coverage of cobalt
0
/
sulphide, uꢁ
/
84%). The metallization was carried out
Fig. 9 demonstrates the example of direct microstruc-
tured metallization using mask (routine (a) in Fig. 8).
using a local contact at the top of the tree. This
Fig. 8. Routines for microstructured metallization using PLATO technique: (a) selective activation and subsequent metallization and (b)
homogeneous activation and selective metallization using masks.

3026
T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ3027
/
through-hole metallization was carried out successfully
by placing a contact tip on the surface of the sample.
The uneven distribution of the thickness (Fig. 10) due to
the diffusion of the electrolyte in the hole can be further
optimized by applying suitable technological parameters
(additive, convection, etc.).
4. Conclusions
The electrodeposition of Ni on insulating ABS poly-
mers activated by cobalt sulphide takes place by the
propagation of primary and secondary layers. The
propagation rates of both layers are constant during
the first 60 s and are dependent on the metallization
potential. The primary layer has a topography char-
Fig. 9. Metallized microstructure of a homogeneously activated ABS
substrate (coverage of cobalt sulphide, uꢁ84%). For metallization, an
excimer laser-structured PI mask was applied (Eꢁ 0.8 V, tꢁ3 min).
/
/
ꢃ
/
/
acterized by hemispheres with different sizes (10ꢀ
/
50 nm
experiment shows that the propagation during metal
deposition can also take place in small and complex
geometries with the width of 40 mm and the length of
height and 0.5ꢀ1 mm diameter), which correspond to the
/
cobalt sulphide clusters underneath. The relatively low
conductivity of the primary layer could be related to the
insufficient electrical contact between Ni hemispheres.
several millimeters. Bended shapes with angles B308 are
/
well reproduced. However, similar to electrodeposition
through masks on metal substrates, concentration and
current distributions should be taken into account
The morphology of the secondary layer repeats the
sec
primary layer, however, with a higher roughness (/Ra
/
ꢁ
/
prim
1
50 nm compared with Ra
of hemispheres. The conductivity of the secondary layer
is much higher than that of primary layer (ksec/kprim
/
ꢁ40 nm) and lower density
/
[
15,16].
Due to the fast propagation rate and the flexibility of
ꢁ
/
the activation process, the direct metallization is a good
method for metallization of insulating microstructures
with complex geometries. Fig. 10 demonstrates a
through-hole metallization of an insulating structure
8
10
1
0 /10 ). This result is explained by the higher thickness
of the secondary layer. The high lateral propagation rate
of the overlayer offers possibilities for metallization of
insulating microstructures. This behavior can be suc-
cessfully applied for metallization of insulating polymers
with complex microstructures. Different concepts for
with aspect ratio Aꢁ/5. The activation was applied on
the surface of the sample and on the hole walls. The
Fig. 10. Cross section of the through-hole-metallized structure on ABS substrate applying PLATO technique (Eꢁ
/
ꢃ
/
0.8 V, tꢁ5 min).
/

T.T. Mai et al. / Electrochimica Acta 48 (2003) 3021ꢀ
/
3027
3027
[
[
[
[
[
3] K.K.H. Wong, S. Kaja, P.W. De Haven, IBM J. Res. Dev. 42
1998) 587.
microstructured metallization using PLATO technique
are proposed and applications of those concepts are
demonstrated.
(
4] J. Hupe, W. Kronenberg, Patent Application by Blasberg
Oberfl a¨ chentechnik, PCT No. PCT/EP89/00204, 1989.
5] D. Schattka, S. Winkels, J.W. Schultze, Metalloberfl a¨ che 5 (1997)
1
823.
6] S. Pilyte, G. Valiuliene, A. Zieliene, J. Vinkevicius, J. Electroanal.
Chem. 436 (1997) 127.
Acknowledgements
7] J. Vinkevicius, I. Mozginskie, V. Jasulaitiene, J. Electroanal.
Chem. 442 (1998) 73.
We thank the Ministerium f u¨ r Schule, Weiterbildung,
Wissenschaft und Forschung des Landes Nordrhein-
Westfalen, for the financial support of this work. The
technical supports of the companies ENTHONE GmbH
(Langenfeld, Germany) and Bartels GmbH (Dortmund,
Germany) are gratefully acknowledged.
[8] A. M o¨ bius, P. Pies, A. K o¨ nighofen, Metalloberfl a¨ che 54 (2000) 3.
9] A. M o¨ bius, J. Oberfl a¨ chetechnik 1 (2000) 14.
[
[
[
[
[
[
10] A. M o¨ bius, et al., Patent Application by Enthone OMI, PCT No.
PCT/US99/26066, 2000.
11] T.T. Mai, G. Staikov, J.W. Schultze, Sol. Stat. Electrochem.,
submitted for publication.
12] A. Brenner, Electrodeposition of Alloys: Principle and Practice,
Academic Press, New York, 1963.
13] G. Staikov, E. Budevski, W.J. Lorenz, Electrochemical Phase
Formation and Growth, VCH, Weinheim, 1996.
14] C. Andricacos, C. Uzoh, J.O. Dukovic, J. Horkans, H. Deli-
gianni, IBM J. Res. Dev. 42 (1998) 567.
References
[
[
1] J.W. Schultze, A. Bressel, Electrochim. Acta 47 (2001) 3.
2] E. Sacher, J.J. Piraeux, S.P. Kovalczyck, Metallization of Poly-
mers, American Chemical Society, Washington, DC, 1990.
[15] M. K u¨ pper, J.W. Schultze, Electrochim. Acta 42 (1997) 3033.
[16] L.T. Romankiw, Electrochim. Acta 42 (1997) 2985.