Electrodeposition and properties of NiW films for MEMS application

Article abstract of DOI:10.1016/j.electacta.2005.03.059

The study reports basic investigations on the electroplating and properties of movable NiW microstructures with potential application in MEMS as temporary contacts for IC integration. The NiW layers have been deposited from nickel sulphamate electrolyte (

Full text of DOI:10.1016/j.electacta.2005.03.059

Electrochimica Acta 50 (2005) 5573–5580
Electrodeposition and properties of NiW films
for MEMS application
E. Slavcheva ,1, W. Mokwa, U. Schnakenberg
∗
Institute of Materials in Electrical Engineering I, RWTH Aachen University, Sommerfeldstr. 24, D-52074 Aachen, Germany
Received 2 December 2004; received in revised form 15 March 2005; accepted 30 March 2005
Available online 15 June 2005
Abstract
The study reports basic investigations on the electroplating and properties of movable NiW microstructures with potential application
in MEMS as temporary contacts for IC integration. The NiW layers have been deposited from nickel sulphamate electrolyte (MICROFAB
◦
NI-110, Enthone GmbH, Germany) with addition of citric acid stabilized tungsten complex at operating temperature of 50 C and pH of about
3
.0 using direct und pulse current plating regimes. The influence of the electroplating parameters, such as the tungsten concentration in the
electrolyte and the current density on the tungsten content in the electroplated alloy, the surface structure and morphology of the deposits, as
well as the mechanical properties of the alloy have been thoroughly investigated. The process allows deposition of 12 ␮m thick homogeneous
NiW layers. An essential improvement of the mechanical properties such as micro hardness, elasticity and internal stress compared to that of
the pure nickel has been achieved at comparatively low tungsten content of 3.2 wt.%.
©
2005 Elsevier Ltd. All rights reserved.
Keywords: NiW; Electrodeposition; MEMS
1
. Introduction
brittleness and internal stress [1]. Some metallic alloying
elements, such as cobalt increase the hardness by formation
of mixed crystal systems in which the nickel atoms and the
atoms of the alloying component are uniformly distributed.
Other alloying elements such as phosphorus or tungsten dete-
riorate the formation of the crystal lattice and as a result,
nanocrystalline alloys or amorphous films known also as
metal glasses can be deposited. Usually, the increase in the
hardness is accompanied by a decrease in the grains size and
the crystal dislocations in the electroplated alloy [1,2]. The
hardening of the nickel alloys by the elements from the first
group, e.g. by formation of mixed crystal lattice, is rather lim-
ited. Once the maximum value of about 450 HV is achieved,
the further variations of the alloying element content and the
electroplating parameters have negligible effect on the alloy
hardness [3]. Much harder nickel alloys can be deposited if
the alloying elements from the second transition group are
used. Tungsten is known to be the best hardening element for
thenickelalloys. ThemainproblemconcerningtheNiWelec-
trodeposition is that the ions of both elements, nickel cations
The electro deposition of nickel and different Ni alloys
is a process of essential importance as these materials have
broad and expanding technical and industrial application.
Nickel and its alloys with elements such as iron, cobalt,
phosphor or tungsten are especially appropriate for usage
in the field of micro electromechanical systems (MEMS)
where microstructures with high mechanical strength and
wear resistance coupled with high thermal resistance and
stability against plastic deformation are required. By choos-
ing a proper alloying element the mechanical features of the
alloys can be essentially influenced, the properties of pre-
dominant interest can be improved and thus, tailored alloys
can be electroplated. It is known that nearly all electroplated
nickel alloys possess higher mechanical hardness and tensile
strength compared to the pure nickel but also an increased
∗
Corresponding author. Tel.: +49 241 80 27758; fax: +49 241 80 22392.
E-mail address: slavcheva@iwe.rwth-aachen.de (E. Slavcheva).
ISE Member.
2+
2−
1
Ni and tungstate anions WO4 , respectively, are very sen-
0
013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.03.059

5
574
E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
sible to pH of the solution and cannot be mixed in any desir-
able concentration in aqueous electrolytes. In acidic media,
tungsten forms solid tungstate and polytungstate acid and it
is very difficult to dissolve back the precipitate obtained. On
the other hand, in alkaline media Ni precipitates as nickel
hydroxide [4]. Thus, during NiW electroplating one of both
precipitates can be easily formed depending on the local pH
value, unless proper complex agents are added in the solu-
tion. It is well known that ammonium form complexes with
Ni , while citric acid stabilises the tungstate anions and can
form mixed complexes with both metal ions [5]. These two
compounds most often present in the electrolytes used for
NiW electroplating as, depending on the pH one of them is
always in much higher concentration.
cracking of the deposited layers. The crystal structure, the
surface morphology and the mechanical properties such as
microhardness, Young’s modulus and elastic behaviour were
investigated as a function of tungsten content in the alloy.
2. Experimental
The nickel–tungsten electroplating bath was prepared
from a commercial nickel sulfamate electrolyte containing
80 g/l Ni as nickel sulfamate concentrate (MICROFAB NI-
110, Enthone GmbH, Germany), 4.2–20 g/l W as a citric acid
stabilized tungstate complex prepared from Na2WO4·2H2O
and C2H8O7, and 30 g/l H2BO2. The concentration of tung-
sten was varied in the range 5–20 wt.%. Four electroplating
baths with tungsten concentrations of 4.2, 8.9, 14.1, 20.0 g/l
2+
The common electrolytes for NiW electroplating use
nickel sulphate and sodium tungstate as metal precursors.
They have alkaline pH and contain high concentration of
ammonium and oxalic acids. The concentration of tungsten
can vary considerably reaching up to 85 wt.% [6]. The alloys
deposited from these electrolytes may contain up to 70 wt.%
W and their hardness is three to four times higher than that of
the pure Ni [2,6–8]. However, these electrolytes are not envi-
ronmentally friendly due to the high concentration of ammo-
nium. In addition, the range of the optimal electroplating
parameters for these electrolytes is very narrow. The required
E
W
were used. These concentrations of tungsten, C , correspond
to 5, 10, 15 and 20 wt.% W, respectively. The concentration of
the metal ions was monitored using inductive coupled plasma
spectrometer (ICP) model Atom Scan 16/25 (Thermo Instru-
ments Systems Inc., USA). The electrolyte was refreshed
periodically to avoid the depletion of the metals. The operat-
◦
ing temperature of the bath was 50 C and pH was in the range
of 2.4–3.2. The electrolyte was circulated using a centrifugal
magnetic pump.
◦
process temperatures of 85–90 C and pH values of 8–11 are
The microsprings were fabricated using the surface micro-
machining technology including photolithography, evaporat-
ing metallization, electroplating and sacrificial layer etching.
Oxidised 100 mm silicon wafers were used as substrates. A
plating base consisting of 50 nm thick titanium and a 100 nm
thick copper layer was evaporated on the polished side of the
wafer. A sacrificial 50 ␮m thick copper film was electroplated
fromCUBATHSCelectrolyte(Enthone, GmbH, Germany)at
too high for fabrication of microstructures on Si-wafers via
the processes used in the microsystem technology. In a pre-
vious study, it has been shown that if such an electrolyte is
◦
used at the operating temperature of 40 C, NiW microstruc-
tures containing up to 22 wt.% W and a hardness two times
higher than that of the pure nickel can be fabricated. Unfor-
tunately, the deposited layers were brittle, had high internal
stress and cracked easily with the increase of the thickness,
which excluded their usage for microsprings fabrication [6].
Alternative electrolytes for NiW electrodeposition are the
nickel sulfamate electrolytes [9–15]. The main advantages
2
current density of 15 mA/cm . After a photolithography step
the NiW micro springs with thickness of up to 12 ␮m were
deposited. The deposition was performed under direct cur-
rent plating (DC) condition or in a symmetrical pulsing (PC)
mode with Ton = Toff = 5 ms using pulse plating equipment
Dynatronix, Type DPR-20-5-10 and an insoluble anode. The
of these electrolytes are the moderate plating temperatures
◦
(
about 50 C) and pH (near neutral), the lack of ammonium,
2
as well as the ability to produce homogeneous films with
increased hardness and reduced internal stress. These prop-
erties are especially attractive for application in the field of
MEMS and justified our choice of NiW electroplating bath.
Therecentpaperreportsbasicresultsonelectroplatingand
properties of NiW microstructures deposited from a modi-
fied commercial nickel sulphamate electrolyte. The envisage
application of these microstructures, namely as movable con-
tacts in the field of MEMS, requires high hardness and wear
resistance of the deposited NiW films, as well as low stress,
high elasticity and flexibility, combined with relevant current
yield of the electroplating process. The research was con-
centrated on the effect of the experimental variables on the
composition of the deposited alloy. The tungsten concentra-
tion in the bath and the current density were systematically
changed in the ranges within which they affect the compo-
sition of the alloy without causing high internal stress and
mean current density was varied in the range 5–20 mA/cm .
The surface profile and the thickness of the electroplated
layers were measured using a Tencor PA-10 profiler. The
surface morphology was investigated by scanning electron
microscopy (SEM) using a ZEISS GEMINI 982 microscope.
The nickel and tungsten contents in the electrodeposited sam-
ples and their surface topography were determined by energy
dispersive X-ray spectroscopy (EDX) in a scanning micro-
scope mode. The crystal structure of the alloy was investi-
gated by X-ray diffraction (XRD) technique using a Philips
diffractometer (Cu K␣ radiation) Model PW 3010. The Vick-
ers microhardness of the electroplated microstructures was
measured with SHIMADZU HMV-2 micro hardness tester
using a 25 g load.
For determination of the Young’s modulus single-clamped
NiW-cantilevers were electroplated from electrolytes with
different tungsten concentration. The dimensions of these

E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
5575
Fig. 2. Dependence of current yield/deposition rate on the tungsten concen-
2
Fig. 1. Experimental set up for the bending tests.
tration in the electrolyte at mean current density of 10 mA/cm in DC and
PC electroplating modes.
cantilevers were as follows: width of 100 ␮m, length of
E
W
6
00–800 ␮m and thicknesses of about 12 ␮m. The Young’s
ciency decreases with the increase of C , however at pulsing
modulus values were determined using a home-made bend-
ing set-up [11] which is schematically shown in Fig. 1. The
samples, clamped between two glass plates, were adjusted
on a weighing machine measuring the contact force with a
resolution of 50 nN and a display accuracy of 10 nN. The
weighing machine was mounted on an x/y-stage with 1 ␮m
displacement resolution. A stylus was imprinted on the can-
tilever surface causing the bending by means of a linear motor
with a step width of minimum 100 nm. Two force–deflection
curves were measured on each cantilever as the distances
from the fixation point were l1 = 300 ␮m and l2 = 450 ␮m.
The calculated spring stiffness c1 and c2 were used to deter-
mine the Young’s modulus E according Eq. (1)
current mode it is slightly higher in the whole range of tung-
sten concentrations tested. The PC electroplating mode has
been used for the rest of the tests presented in this work.
The current yield in the bath containing 5 wt.% W (4.21 g/l)
is relatively high (70%) but the efficiency of the process
decreases rapidly with the increase in tungsten concentration.
E
At C = 20 wt.% (20 g/l) it drops to 17 and 12% for pulse
W
and direct current mode, respectively. The corresponding val-
ues of the deposition rate are 36 and 25 nm/min, which is
nearly an order lower than the deposition rate of 205 nm/min
typical for the MICROFAB NI-110 bath.
The applied current density, j, itself affects essentially
the efficiency of the process. This effect is illustrated in
Fig. 3 for electroplating baths containing 15 and 20 wt.%
W. The current yield increases linearly with the increase in
current density. For the 15 wt.% W electrolyte η is only 7%
ꢀ
ꢁ
ꢁ
ꢂ−3
3
4
ꢀl
3
1
1
3
3
E =
−
(1)
bt
c2
c1
2
2
at 5 mA/cm but it reaches 56% at 20 mA/cm . The values of
the current yield in the electrolyte containing 20 wt.% W are
somewhat lower in the whole range of current densities tested
but the linear correlation between η and j does not change.
The composition of the deposited alloys also depends
strongly on the current density of the electroplating. This
where ꢀl represents the distance between the measuring
points 1 and 2, b the width and t is the thickness of the mea-
sured beam. Stress and strain values were calculated using
Hook’s law according Eqs. (2) and (3):
6
Fl
2
σb =
(2)
(3)
bt
σ
b
3 ft
2 l2
ε =
=
E
where F is the applied force and l is the distance from the
fixation point.
3
. Results
The influence of tungsten concentration in the electroplat-
ing bath on the current yield, η, respectively on the deposition
rate, R , of the DC and PC plating at mean current density
d
2
of 10 mA/cm is presented in Fig. 2. The data for each point
were obtained as an average value from three experiments.
It can be seen that in both plating modes the Faradaic effi-
Fig. 3. Dependence of current yield on the mean current density in PC elec-
troplating mode.

5
576
E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
Fig. 6. Dependence of Vicker’s microhardness on the tungsten content in
the deposited layer.
Fig. 4. Influence of current density on the tungsten content in the layers
deposited in PC electroplating mode.
changes in surface structure with the applied current density,
respectively with the deposition rate, can be seen in Fig. 5
which shows SEM images of samples deposited from the
electrolyte containing 20 wt.% W at current densities vary-
effect is presented in Fig. 4 for the electrolytes with 15
and 20 wt.% W. The results show that the content of tung-
L
sten in the deposited layers, C , decreases essentially with
W
the increase in current density. The alloy deposited from
2
2
ing in the range 5–20 mA/cm . A clear tendency of increasing
the 15 wt.% W bath at current density of 5 mA/cm con-
2
L
W
the size of the grains with the increase of η is observed. The
surface morphology changes from fine globular grains with
homogeneous distribution to comparatively rough polycrys-
talline structures.
tains 4.87 wt.% W, while at 20 mA/cm C drops down to
only 0.41 wt.% W. The corresponding values of the sample
deposited from the bath containing 20 wt.% W are higher,
6
.13 and 0.83 wt.% W, respectively as the trend is very simi-
lar.
The visual inspection of the samples electroplated at all
The tungsten content in the electroplated films deposited
2
at mean current density 10 mA/cm depends linearly on the
tungsten concentration in the bath as CL = 0.16C . Fig. 6
E
W
W
current densities showed homogeneous and metallic bright
surfaces. The scanned surface profiles however, registered
a gradual increase of the roughness with the increase of
current density. The average roughness factor, Ra, of the sam-
ples deposited in 20% W electrolyte was about 50–65 nm at
presents the data for the measured Vickers hardness of the
corresponding specimens as an average value of five mea-
surements. The trend of the microhardness follows that of
L
W
C
and increases with the increase of tungsten concentra-
2
2
tion in the plating bath.
j = 5 mA/cm , while at j = 20 mA/cm it increased with more
than an order of magnitude and reached 700–800 nm. The
Fig. 7. SEM images of NiW samples with different tungsten content (shown
on the micrographs) form electrolytes with varying W concentration at mean
current density of 10 mA/cm .
Fig. 5. SEM images of NiW layers deposited in PC mode at varying current
densities from a bath containing 20 wt.% W.
2

E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
5577
Fig. 9. Design in top view of the fabricated microsprings.
Fig. 8. XRD spectra of NiW layers with different tungsten content deposited
2
at current density of 10 mA/cm in PC mode.
The SEM images of the NiW alloys deposited from baths
E
W
2
with different C at mean current density of 10 mA/cm are
presented in Fig. 7. The SEM images demonstrate smooth
topography with globular compact grains. No cracks are vis-
ible on the surface. The more tungsten contains the specimen,
the smaller are the crystal grains and the more homogeneous
is the surface of the deposit.
In Fig. 8 are depicted the θ/2θ XRD spectra of the same
samples. They show comparatively low integral intensity,
suggesting a micro cristallinity of the electroplated NiW lay-
ers. Yet, the characteristic peaks corresponding to nickel fcc
crystal lattice are clearly depicted and are labelled by their
Miller indices. The difference in peaks intensity and width is
distinct enough in order to calculate the size of the crystal-
lites using the Sherrer equation [16]. The obtained data are
summarized in Table 1.
Fig. 10. Stress–strain diagrams of the fabricated structures as a function of
tungsten content in the deposited layer.
E
type, deposited from electroplating baths with varying C at
W
2
mean current density of 10 mA/cm were measured accord-
ing the previously described procedure. The calculated data
dependedextremelystrongonthethicknessofthecantilevers.
At the same time the measured thickness showed a variation
of ± 1 ␮m over the beam length. In addition, sporadic pits
were visible on some cantilevers. In order to compensate
these effects, the obtained strain–stress curves were fitted
by varying the experimentally determined mean thickness of
the cantilevers by maximum ± 0.6 ␮m (see Eq. (1)) and are
plotted in Fig. 10.
Fig. 9 shows the design in top view of the fabricated
microsprings. The Young’s modulus of structures from that
Table 1
Peaks intensity and crystallite size obtained by XRD spectra
ꢃ
L
W
i
p
C
wt.%
Peak intensity (Ip); peak position (2θ)
Relative intensity Ip/
I
Crystallite size (nm)
ꢀ1 1 1ꢁ
ꢀ2 0 0ꢁ
ꢀ2 2 0ꢁ
ꢀ3 1 1ꢁ
ꢀ1 1 1ꢁ
ꢀ2 0 0ꢁ
ꢀ1 1 1ꢁ
ꢀ2 0 0ꢁ
0
900,0000; 44.53
2530,0000; 45.9
795,0000; 44.29
1576,0000; 44.43
841,0000; 44.45
6593,0000; 51.87
3457,0000; 51.95
279,0000; 51.71
445,0000; 51.81
240,0000; 51.85
1528,0000; 76.41
2172,0000; 76.45
219,0000; 76.25
296,0000; 76.45
225,0000; 76.43
262,0000; 92.97
388,0000; 93.2
188,0000; 92.93
228,0000; 92.87
164,0000; 92.97
0,09695142
0,2960103
0,66139767
0,61925344
0,57210884
0,71022299
0,4044694
0,23211314
0,17485265
0,16326531
34.2
31.8
25.6
18.9
12.8
33.8
28.9
15.3
12.4
10.7
0.65
1.44
2.30
3.21

5
578
E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
4
. Discussion
EDS analysis of the corresponding alloys showed a reverse
dependence of tungsten content in the deposit on the current
density (Fig. 4). The letter fact disagrees with the obser-
vation of other authors for NiW deposition from neutral
electrolytes with high tungsten content [19,24]. However,
it should be taken into account that firstly, the electroplat-
ing bath used in this research is acidic and therefore, rich
The electroplating baths used in this study are based on the
standard nickel sulfamate electrolyte, which contains 80 g/l
nickel. The high metal content ensures an excellent elec-
tro conductivity of the bath, which in turns improves both
the homogeneity of the deposits and the current yield of the
process. These features are typical for MICROFAB NI-110
and make the electrolyte especially suitable for fabrication of
microstructures, which determined our choice of the plating
bath. The current yield of the MICROFAB NI-110 electrolyte
is 97%. The tungsten was added as citric stabilised complex
ions as neither tungsten nor nickel ions in the as-prepared
mixed electroplating solution showed a tendency to precip-
itate. The gradual increase of tungsten concentration has a
reverse effect on the efficiency of the process and the cur-
rent yield, respectively the deposition rate, drops essentially
with the tungsten enrichment of the electrolyte to reach a
of hydrogen ions and secondly, the CE is much lower than
W
that of nickel and the tungsten deposition most probably pro-
ceeds under diffusion control. The results obtained imply
that the registered increase of η, respectively of the depo-
sition rate with j, is due to a favoured reduction of nickel
ions in the bath. At the same time, the reduction of the tung-
sten and/or the mixed nickel–tungsten citrate complex ions
is depressed due to intensification of the side reaction of
hydrogen evolution, which is facilitated by the low pH of
the electrolyte.
The tungsten content in the deposited alloys affects
strongly the surface topography. The layers fabricated at
value of 17% at CE = 20 wt.% (20 g/l) for the PC plating.
W
2
In accordance with the literature data the η values for DC
mode are even lower [17]. As far as very low deposition
rates are not appropriate from technological point of view,
the tungsten concentration in the plating bath was not fur-
ther increased and the rest of the tests were performed in
a PC plating mode. The observed drastic decrease of the
current yield with the increasing tungsten concentration in
the bath and the registered low values of η are not sur-
prising. According to the available literature, the Faradaic
efficiency of NiW electroplating usually does not exceed
mean current density of 5 mA/cm have maximal tungsten
content and very fine homogeneous surface with spherical,
compact and smooth grains. The increase in current den-
sity reduces the tungsten content in the alloy, which in turns
causes gradual increase of the grains size and transformation
of the fine spherical nodules in rather rough and disordered
structures.
2
The mean current density of 10 mA/cm was considered
L
W
as a reasonable compromise with respect to η and C . It
was used to investigate the dependence of tungsten content
in the alloy on its concentration in the electroplating bath.
1
2–14% and alloys with high tungsten content, never mind
E
the exact composition of the plating bath, can be deposited
mainly in the cost of decreasing the corresponding Faradaic
efficiency [18,19]. These effects are typical for electroplating
of metals such as tungsten, which cannot be deposited from
aqueous solutions unless a second metal from the iron group
such as Ni is present in the electrolyte. The phenomenon is
known as induced co-deposition and its detailed mechanism
is still not fully understood. The most widespread hypothe-
sis assumes that the NiW alloy deposition is a result of two
main cathodic reactions, namely a reduction of nickel ions
Again, in order to avoid too low current yields, CW was var-
ied in comparatively narrow range (4.2–20 g/l). The observed
distinct linear enrichment of tungsten in the deposited alloy
with the increase of CE shows that the share of the second
W
main cathodic reaction, including the reduction on of tung-
sten complexes is enhanced proportionally to the increase of
tungsten concentration in the bath. The results obtained sug-
gest fast discharge of tungsten complex ions on the cathode
and diffusion control of the process. The maximum tung-
sten content achieved was 3.21 wt.% W, which is rather low
in comparison to the values obtained in other electrolytes
[5,13,20]. Nevertheless, theenrichmentofalloywithtungsten
even at these low concentrations leads to essential hardening
of the alloy (Fig. 6). The trend of the microhardness fol-
lows that of the CL and increases linearly with C . The
(
usually as ammonium complexes) and a depolarization of
tungsten or/and mixed nickel–tungsten complexes (usually
of oxalic acids, such as the citric acid), which proceed simul-
taneously [5,20]. According to other authors deposition of
NiW alloy from nickel sulfamate electrolytes is a result from
the inclusion of metallic tungsten or tungsten anions in the
already formed nickel matrix [21]. As far as tungsten and
hydrogen have very close redox potentials, depending on pH
of the bath a more or less intensive side reaction of hydrogen
evolution always takes place on the cathode as well [15,22].
In addition, tungsten is a very good catalyst for hydrogen
evolution [23] and its deposition on the cathode accelerates
E
W
W
maximal value of 730 ± 11 HV 0.025 obtained for the sam-
ple with highest tungsten content is about 2.5 times higher
than that of the pure nickel. The strong influence of tung-
sten on microhardness is combined with essential changes
in surface structure and morphology of NiW alloy (Fig. 7).
The more tungsten contains the specimen, the smaller are
the crystal grains and the more homogeneous is the surface
of the deposit. Such an effect has been reported by other
authors, although for NiW alloy much richer of tungsten
[25]. However, our results are entirely reasonable, as it is well
known that usually, the hardening of the alloy is most pro-
the side cathodic reaction leading to the observed reverse
E
W
dependence between η(R ) and C . At equal concentration
d
of tungsten in the bath the Faradaic efficiency increases lin-
earlywiththeincreaseinthecurrentdensity(Fig. 3), whilethe

E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
5579
nounced during the initial inclusion of the alloying element
5. Conclusions
in the main metal matrix. It should be noted that the observed
changes in the surface structure and morphology with the
Moveable microstructures with application as temporary
contacts in IC integration were fabricated by electroplat-
ing of NiW layers from nickel sulfamate electrolyte doted
with citric acid stabilised tungstate complexes. An essential
improvement of the mechanical properties such as higher
hardness, lower stress and improved elastic behaviour com-
paring to the pure nickel was achieved at comparatively low
tungsten content of 3.2 wt.%. It was shown that the increase
of tungsten content can be realised either through increasing
the tungsten concentration in the bath or by usage of very
low current densities. The former approach leads to essential
decrease of the current yield. Much higher Faradaic effi-
ciencies can be achieved by increasing the current density,
which however, has a reverse effect on the tungsten content
in the deposited layer. If low current densities are used, the
time required for deposition of films with a defined thick-
ness must also be prolonged. Therefore, none of these both
approaches is ideal. As a result of the research performed,
the electroplating of NiW microstructures from a bath with
increase in tungsten content as a result of CE rising and those
W
caused by the varying current density (Fig. 5) are in good
agreement.
The θ/2θ spectra in Fig. 8 show relatively low integral
intensity with well depicted peaks characteristic for a face
centred cubic crystal lattice of isotropic nickel. The spec-
trum of the pure nickel sample is dominated by the ꢀ2 0 0ꢁ
peak. The intensity of this peak decreases with the increasing
tungsten content, accompanied by slight increase in the rela-
tive intensity of the ꢀ1 1 1ꢁ peak. The changes in the other
two peaks are negligible. As a whole the intercalation of
tungsten atoms in the nickel crystal lattice does not change
much the crystal orientation in the deposited layers. However,
the enrichment of the alloy with tungsten leads to essen-
tial decrease in the size of the crystallites as can be seen
from the data in Table 1. The crystallites decrease system-
atically with the increase of the tungsten content. A very
good correlation between the data calculated from the full
width of the half maximum (FWHM) of ꢀ1 1 1ꢁ and ꢀ2 0 0ꢁ
peaks is obtained. These results are in accordance with the
SEM images of the samples showing change of the surface
structure in direction to lower micro crystallinity with the
tungsten enrichment of the alloy. They correlate well with
the registered strong increase in the micro hardness of the
investigated samples, as it is well known that the hardenning
ofalloysgenerallyisaccompaniedbyadecreaseingrainssize
E
2
C
= 14.1 g/l (15%) at current density j = 10 mA/cm is rec-
W
ommended as an optimal compromise at the time being. A
further research is in progress, aimed to improve the deposi-
tion efficiency by optimisation of the pulsing current mode
and to elucidate the exact mechanism of the electroplating
process.
[1,2].
Acknowledgements
The fitted curves in Fig. 10 show a linear increase of
the strain with the increasing stress. The measured Young’s
modulus of the cantilevers electroplated from the nickel
bath is comparable to that obtained earlier by micro ten-
sile testing [26] and resonating beam experiments [27]. The
slope of the curves, which represents the Young’s modu-
lus, is nearly constant for all samples with a value of about
The authors would like to express their sincere thanks to
Robert Bleidiessel for preparation of the electrolytes, to Ben-
jamin Benitsch for carrying out very carefully the bending
experiments, and to Ingo Klammer for the interpretation of
the stress-strain diagrams. Financial support for this work
was given by the German Science Foundation, grant SFB
440.
1
80 GPa. At a certain stress the samples deform plastically
which results in decrease of the slope values. The plastic
deformation occurs at higher stress values at microstructures
with higher tungsten content, indicating a distinct favourable
effect of tungsten on the alloy elasticity. The values of yield
strength, Rp02 increase correspondingly, as can be seen from
thedataincludedin Fig. 10. Theelasticbehaviourofthemate-
rial under study correlates well with the registered increase
of the micro hardness and the rule that a doping element,
which increases the hardness of the material under inves-
tigation, also increases its yield and tensile strength. The
improvement in the elastic behaviour can be explained by
the grain size and the texture effects in the deposited lay-
ers. It is a common experience in material engineering that
metals with smaller grain sizes show less ductility com-
pared to those with high grain sizes. They possess a higher
concentration of grain boundaries, which when being plas-
tically deformed are obstacles for the slipping of the atomic
layers.
References
[1] A. Brenner, Electrodeposition of Alloys, Principles and Practice, vol.
I–II, Academic Press, New York, 1963.
[
[
2] M.M. Younan, M. Shoeib, Galvanotechnik 86 (4) (2002) 932.
3] D.E. Kranz, Galvanotechnik 89 (1997) 1506 ff.
[4] A.F. Holleman, N. Wieberg, Inorganic Chemistry, Walter De Gruyter,
Berlin, New York, 1985.
[
[
5] O. Younes, E. Gileady, J. Electrochem. Soc. 149 (2) (2002) C100.
6] Th. Fritz, W. Mokwa und, U. Schnakenberg, Galvanotechnik 92 (8)
(2002) 2108.
[7] T. Yamasaki, P. Schlossmacher, K. Ehrlich, Y. Ogino, Nanostruct.
Mater. 10 (3) (1998) 375.
[
[
8] P. Schlossmacher, T. Yamasaki, Microchim. Acta 132 (2000) 309.
9] C. Krueger, W. Mokwa, U. Schnakenberg, Proceedings of 17th Con-
ference MEMS, Maastricht, Netherlands, 23–25 January, 2004, p.
1
17.
10] Th. Friz, Ph.D. Thesis, Aachen University, 2002.
[11] C. Krueger, Ph.D. Thesis, Aachen University, 2004.
[

5
580
E. Slavcheva et al. / Electrochimica Acta 50 (2005) 5573–5580
[
[
[
12] Th. Fritz, W. Mokwa, U. Schnakenberg, Galvanotechnik 92 (10)
2002) 2684.
13] N. Atanasov, K. Gencheva, M. Bratoeva, Plat. Surf. Finish. 84 (2)
1997) 67.
14] M. Bratoeva, K. Gencheva, T. Tsatscheva, N. Atanasov, Galvanotech-
nik 54 (2000) 10.
[20] O. Younes, E. Gileadi, Electrochem. Solid State Lett. 3 (12) (2000)
543.
[21] T.A. Alekhina, I.A. Shoshina, A.L. Rotinyan, J. Anal. Chem. 11
(1990) 2419.
[22] M. Obradovic, J. Stevanovic, R. Stevanovic, A.R. Despic, J. Elec-
troanal. Chem. 491 (2000) 188.
(
(
[
[
15] M. Bratoeva, N. Atanasov, Met. Finish. 96 (1998) 92.
16] H.P. Klug, L. Alexander (Eds.), X-Ray Procedures for Polycrystalline
and Amorphous Materials, Wiley/Interscience, New York, 1980.
17] M. Donten, H. Cesiulis, Z. Stojek, Electrochem. Acta 45 (2000)
[23] M.M. Jaksic, Int. J. Hydrogen Energy 26 (2001) 559.
[24] B.N. Maruthi, L. Ramesh, S.M. Mayanna, D. Landolt, Plat. Surf.
Finish. 86 (3) (1999) 85.
[25] L. Zhu, O. Younes, N. Ashkenasy, Y. Shacham-Diamand, E. Gileadi,
Appl. Surf. Sci. 200 (2002) 1.
[26] T. Fritz, W. Mokwa, U. Schnakenberg, Microsyst. Technol. 9 (2002)
87.
[27] T. Fritz, M. Griepentrog, W. Mokwa, U. Schnakenberg, Electrochim.
Acta 48 (2003) 3029.
[
[
[
3389.
18] H. Cesiulis, A. Baltutiene, M. Donten, Z. Stojek, J. Solid State
Electrochem. 6 (2002) 237.
19] Y. Wu, D. Chang, D. Kim, S. Kwon, Surf. Coat. Technol. 173 (2003)
259.