-
Novel Multilayer Electrostatic Solid-State Actuators with
Elastic Dielectric
Helmut F. Schlaak, Markus Jungmann, Marc Matysek, Peter Lotz
Institute for Electromechanical Design, Darmstadt University of
Technology,
Merckstr. 25, D-64283 Darmstadt, Germany
ABSTRACT
Solid state actuators provide deformation and actuation forces
mainly excited by electric fields. Piezoelectric actuators are well
established providing high forces at low strain due to their
material characteristic. Electrostatic solid-state actuators
consist of elastic dielectric layers between compliant electrodes.
Applying electric fields of up to 100 V/m at the electrodes the
dielectric contracts due to electrostatic forces and expands in
orthogonal direction. We use high elastic silicone elastomers with
thin graphite powder electrodes. In order to increase the absolute
strain values at limited voltage, we have developed a novel
multilayer process technology to fabricate elastomer stack
actuators with up to 100 layers. The electromechanical properties
of the actuators have been evaluated theoretically and
characterised experimentally. Maximum strain values up to 20 % for
prestressed multilayer films have been achieved. The novel
multilayer fabrication technology provides multilayer stack
actuators with various electrode patterns like universal linear
actuators or matrix arrays for a wide range of applications as
tactile displays for telemanipulation or Braille displays.
The strain in vertical direction versus driving voltage shows a
hysteresis due to viscous friction in the elastomer layers. These
measurements correspond to a viscoelastic theoretical model. The
mechanical stress versus strain characteristic shows a strong
nonlinearity for strains > 30 %. The dynamic characteristic has
been evaluated by measuring the mechanical impedance in the
frequency range of 2 to 1000 Hz.
Keywords: electroactive polymer, multilayer, electrostatic
actuator, dielectric elastomer, tactile display
1. INTRODUCTION
Almost all technical products contain actuators. The variety of
actuators is increasing while their physical dimensions are
shrinking. This fact enables totally new fields of applications.
During the last years piezo-electric actuators captured more and
more attention. Meanwhile they are very well analyzed concerning
available materials, characterization and dimensioning. Even todays
available piezo-electric ceramics (e.g. PZT) and polymers (e.g.
PVDF) respectively achieve a maximum strain of about 0.2 % and 0.1
% respectively.1,2 However they provide high forces. To reduce the
driving voltage multilayer cofired piezoelectric ceramic actuators
with thin layers (down to 20 m thickness) have been developed.3 The
transverse piezo-electric effect is used with multilayer bending
beam actuators. Stack actuators are using the longitudinal
piezo-electric effect in contrast.4 Often an additional mechanical
transmission is required to achieve attended strain. One way to
higher strain ratios is given by lower stiffness: the material can
easily be distorted. Foams for example have a very low stiffness
and therefore a high strain at low maximum forces. Predominantly
they are used as sensors (e.g. in electret microphones) because
their maximum load is quiet low.5 At this certain point the
electrostatic solid-state actuators are concerned. They can be
realised by an elastic dielectric layer between compliant
electrodes. Applying a voltage at the electrodes the dielectric
contracts due to electrostatic forces and expands in lateral
direction (Fig. 1). Due to the large variety of materials that can
be used as dielectric, restoring spring and as substrate carrying
the electrodes at the same time the compromise between high
pressure and high strain can easily be met. Another feature of
these materials is their high structural flexibility and their low
density. Using high elastic silicone elastomer with thin graphite
powder electrodes a relative thickness strain up to 30 % can be
achieved.6
Invited Paper
Smart Structures and Materials 2005: Electroactive Polymer
Actuators and Devices (EAPAD), edited by Yoseph Bar-Cohen,
Proceedings of SPIE Vol. 5759 (SPIE, Bellingham, WA, 2005)
0277-786X/05/$15 doi: 10.1117/12.604468
121
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
Fig. 1: Deformation of an elastic dielectric film under
electrostatic pressure
The electrostatic pressure causing the deformation is
22
0 02r rVp T Ez
= = = (1)
where 0 is the relative permittivity in air, r the permittivity
of the dielectric, z its thickness and V the applied voltage. Since
polymers are nearly incompressible, the volume remains constant
during deformation. By reducing the voltage the dielectric returns
to its initial shape and can produce forces due to the stored
elastic energy. To realise a sufficient strain with applicable
voltages the thickness of the dielectric has to be in the range of
a few microns, causing a further loss of absolute strain. The
electrodes have to be very compliant not to constrain the
deformation. Pelrine et al. have showed more than 30 % relative
strain in thickness and electrostatic pressures of more than 1 MPa
on prestrained silicone elastomer dielectric films between carbon
electrodes.6 In order to increase the absolute strain values at
limited voltages, certain arrangments are essential. Fig. 2 shows
the three major assemblies.
Fig. 2: Configuration and function of electrostatic
elastomer-actuators
A bending actuator (Fig. 2 a) consists of dielectric layers
carrying electrodes on top of a passive layer. In operation the
dielectrics elongation generates a tensile stress onto the passive
layer causing the deflection. Hollow and solid cylinders (so called
roll-actuators Fig. 2 b) are fabricated by a combination of only
two dielectric layers and two electrodes convoluted as cylinder.
The low number of layers reduces expenditure in manufacturing but a
parallel fabrication of roll-actuators as an array is nearly
impossible. As shown before stack actuators (Fig. 2 c) can be
fabricated in different sizes and arrangements. Therefore the
handling of a high number of thin dielectric and electrode layers
is quiet challenging. We have developed a multilayer process
technology to fabricate elastomer stack actuators with more than
100 layers. The according design of the electrodes allows even
parallel fabrication of stack actuators in a matrix
configuration.
2. TECHNOLOGY
2.1 Requirements There is a huge variety of parameters to be
kept in mind while choosing the right elastomer for your special
application with multilayer actuators. If you decide to use an
elastomer with a high grade of hardness you can apply quiet large
forces causing a low prestrain before electrostatic excitation. The
remaining displacement is close to its idle-maximum. Therefore a
high electrostatic pressure is required. The high electric field
may lead to a breakdown. On the
122 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
other hand you may choose a very soft elastomer. The required
voltage for maximum displacement is relatively low but the
prestrain causes high deformation and minimizes possible
contraction.
Beside these considerations the elastomer material should have a
low dynamic viscosity not to loose maximum displacement with
increasing frequency. Furthermore the mechanical parameters of the
dielectric should have a high disruptive strength and due to eq.
(1) a high permittivity to maximize actuator performance.
Technological demands are a low viscosity to realize thin
dielectric films by spin coating and an addition curing silicone,
not to get fission products. Dilution of the components is possible
but requires a certain time to degasify, otherwise bubbles would be
the consequence. A low pot life leads to an acceptable production
time each dielectric layer has to cure before the electrode is put
onto it. Often curing can be accelerated by thermal or optical
radiation. Table 1 shows several silicones and their main
parameters.
Table 1: Properties of applicable silicones
Manufacturer Name shore hardness viscosity, uncured [Pa s]
permittivity
Bayer IS 5663/20 15 8,5 3 Dow Corning 96-082 31 1,1 3,14 Wacker
Elastosil RT 607 55 10 3,7 Wacker Elastosil RT 675 80 35 6,1
The importance of the electrodes material is given by its
conductivity. The electrical cut-off frequency is dominated by the
(series-) resistance of the feed line. The sheet resistance of the
feed line and the electrode has to stay low during planar expansion
otherwise the time constant would change immensly and parts of the
electrode may get lost due to interceptions. Furthermore the
surface density of the electrode should stay high even under
expansion to assure high effective electrostatic pressure. The
mechanical requirements are influenced by the deformation of the
electrode layer. The ideal electrode would have an infinite
compliance, is thin compared to the dielectric layer and is
patterned with high resolution.7 A good adhesion onto the
dielectric layer is needed to prevent partial wash off effects
reducing conductivity. The manufacturing process shouldnt be too
complex, otherwise the time for fabrication of a stack would be too
long.
2.2 Realisation We have developed an automatic fabrication
process of multilayer actuators which is controlled by a computer.
A special LABVIEW program is used to set the parameters and drive
all the process sequences. Fig. 3 shows an overview of one
cycle.
Figure 3: Process steps of one cycle of the automated
fabrication for elastomeric actuators
Proc. of SPIE Vol. 5759 123
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
By spinning on silicone elastomeric dielectric films are
realized. Spin coating is used to achieve homogeneous coating
thickness of only a view microns. As mentioned in subsection 2.1 an
addition curing silicone consisting of two components is used to
avoid decomposition products on the surface while curing. The two
components are stored in cartridges and are squeezed out by
stepping motors. Flexible tubes head the components into a static
mixer where they get mixed consistently without embedding any
bubbles. The use of materials without any catalytic poison (e.g.
plasticizer) is very important, otherwise curing may be disturbed
or prevented. The volume of the mixer and the pot life have to be
coordinated that way, that already mixed elastomer within the mixer
will not be cured until its displacement after the further process
cycles. The spun on silicones curing is accelerated by thermal
radiation. Finally cooling down the elastomer and the spincoater
prevents the new silicone layer from curing before the end of the
spin coating process. The electrodes are deposited onto the
elastomeric film. To contact the actuator to its supply voltage
easily it is necessary to turn the leads of every second electrode.
Fig. 4 shows the schematic cross section of an actuator stack with
a parallel interconnection. The fabrication device for a structured
spray-coating of conductive powder on elastomeric layers has to
realize three major cycles:
masking the dielectric layer dosing the graphite powder spraying
on the pressurized air-graphite powder mixture
Masking is realised by a spray head carrying a shadow mask
patterned by photolithography. The graphite powder is stored in a
pressure vessel. By whirling up the particles the mixture is
generated in the upper parts of the vessel. Out of a nozzle the
air-graphite powder mixture is sprayed onto the mask.
2.3 Results Prototypes with up to 200 dielectric layers have
been fabricated with silicone elastomer Wacker Elastosil P7670.
Various electrode patterns, e.g. single electrodes as well as
electrode matrix arrays have been realised. The minimum thickness
of dielectric layers has been limited to approximately 25 m due to
the decreasing yield for thicknesses below 20 m. The
particle-electrode thickness is about 5 m while the primary
particle size of the graphite powder is 2 m. Fig. 5 shows a
micrograph of a 100-layer actuator stack. If you compare the
realised actuator stack to the schematic assembly (Fig. 4) you can
see, that the electrodes or graphite particles do not immerse into
the silicone layer nor a wash-off effect of the particles occurs.
The reproducibility of the electrodes edge is within a few microns
as the graphite stands out against the clear silicone layers (Fig.
5 a).
Figure 5: Micrograph: a) boundary between supply region and
active region (scaling bar: 50m); b) active region (scaling bar:
20m)
a) b)
Figure 4: schematic cross section of an actuator stack
124 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
In addition we investigated the reproducibility of the thickness
of the spun on silicone layers. With a line width measuring system
the thickness of each dielectric layer of a 50 layer stack has been
determined (Fig. 6 a). It can be seen that the first spun on layers
are thinner than the last ones: the mean value is slightly rising
by the number of layers.
10
15
20
25
30
35
1 6 11 16 21 26 31 36 41 46film n
film
th
ickn
ess
[m
]
measured valueslinear fit
-40%-30%-20%-10%
0%10%20%30%40%
1 6 11 16 21 26 31 36 41 46film n
rela
tive
de
via
tion
Actuator 1Actuator 2Actuator 3
a) b) Figure 6: a) Measured absolute film thickness. b)
Deviation of film thickness from mean thickness for three different
actuators
To verify this effect another three actuators have been measured
(Fig. 6 b). The effect of a growing film thickness has been proven.
As mentioned before the electrode and its conductivity are of high
importance. Not only the material but also the way of deposition
affects its quality. Since two methods are applicable spraying and
brushing both results have been analysed in Fig. 7. Even if both
both types have low starting sheet resistance (< 10 k/cm), the
sprayed electrode has more than two times greater resistance. This
difference is rising with transverse strain that is applied
manually by a mechanical device. Actuators with a high strain
should carry brushed electrodes even if there are still
difficulties by brushing graphite powder onto a micro patterned
shadow mask (and not under it).
y = 7,21x + 10
y = 3,28x + 7
020406080
100120140160
0 5 10 15 20 25transverse strain Sx,y [%]
she
et r
es
ista
nc
e [k
/cm
]
graphite, sprayedgraphite, brushed
Figure 7: Sheet resistance of graphite electrodes
The two micrographs in Fig. 8 show the surface of a brushed
(Fig. 8 a) and a sprayed (Fig. 8 b) electrode. The sprayed
electrode shows an inhomogenous surface that influences the
conductivity obviously. The homogeneous layer of the brushed
graphite may be the reason for a better conductivity for
prestrained electrodes.
Proc. of SPIE Vol. 5759 125
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
a) b) Figure 8: SEM Micrographs of a) brushed and b) sprayed
graphite electrode
3. MODELLING
Due to high strain of the dielectric films of an actuator stack
there is a nonlinearity of the stress-strain-characteristic.
Furthermore viscoelastic behavior is typical to elastomeric
materials, hysteresis and damping effects are the consequence. In
this section a model is derivated, characterizing the actuators
behavior under the mentioned restrictions.
3.1 Geometrical Nonlinearities Geometrical nonlinearities are
typical characteristics of elastomers, caused by a very low
compressibility. Therefore the volume V of a deformed elastomer
stays nearly constant8. If we assume a constant volume, vertical
strain of an elastomer probe caused by the force F deforms the
probe (Fig. 9). The relation between vertical strain dz and area
strain dAz is given by 0 0 0 0 0 0 0( ) ( )Z Z ZV const x y z A z A
dA z dz= = = = + (2)
Figure 9: Geometrical proportions of a probe deformed by force
F
The transverse strain Sx and the area strain SA of an uniaxial
compressed elastomer volume are given by
0
1 11X Z
dxSx S
= =
; 0 1Z Z
AZ Z
dA SSA S
= =
(3)
Fig. 10 shows the transverse and area strain as a function of
the vertical compression. Exceeding 61.9 % compression strain the
transverse strain is larger than the compression strain.
At quasi-static deformation the stored mechanical energy W
caused by the Force F complies with the sum of the three expansion
energies in each dimension:
z y xdW F dz F dz F dy F dx= = + + (4) Here Fi = Ti Ai
corresponds to the force acting onto the boundary face.
126 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
00,5
11,5
22,5
33,5
4
0 0,2 0,4 0,6 0,8 compression strain Sz
s
tra
in S x
, S A
transverse strainarea strainSx=Sz
Fig. 10: Transverse and area strain depending on the compression
strain
Converting eq. (2) and (4) leads to Z Y XT T T T= (5) TY and TX
are negative because they are tensile stress assuming TZ as
compression. If we assume isotropic material properties and a
constant, isotropic Youngs modulus Y = Ti/Si we get the externally
affecting stress T (using Sn from eq. (3)) causing the compression
SZ :
1( 2 ) 2 11Z X Z nZ
T Y S S Y S Y SS
= + = + =
(6)
3.2 Viscoelastic Performance The viscoelastic performance is
another important characteristic of elastomers. Abrupt distention
of these materials leads to an excursive increasing of the
mechanical stress followed by asymptotic decreasing to the minimum.
This effect is called relaxation. If there is a jump of stress,
strain increases gradually asymptotic to its maximum. This effect
is called creeping. Viscoelastic behaviour is represented by the
Thompson model (Fig. 11), consisting of stiffnesses c1,2 and the
viscous friction d2. The model is upgraded by another damper d1 and
a mass element m.
Figure 11: Modified Thompson model
The mechanical impedance of this equivalent network is given by
the following complex equation 1
1
2 2
11mech
cFZ d j mjv jc d
= = + + ++
(7)
Including accounted corpus geometry (face A0, height z0),
concentrated parameters (d1,2, c1,2, m) can be converted into
material specific terms (Youngs modulus Y1,2, viscosity-modulus 1,2
and inertia-modulus M): 0
1 10
2 2
1 11 1mech Z
z TZ Y j MA jS j
Y
= = + + ++&
(8)
Multiplying eq. (8) by j the stress-strain-transfer function G
can be calculated (for small S):
Proc. of SPIE Vol. 5759 127
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
21 1
2 2
11 1
T TG j j Y MS S j
Y
= = = + +
&
(9)
Assuming isotropic material parameters and keeping geometrical
nonlinearities in mind (equation (5) and (6)) we get the complex
stress-strain-characteristic:
nT G S= (10)
3.3 Electrical model The equivalent electrical circuit of an
electrostatic multilayer actuator is shown in fig. 12. RS
represents the resistance of the feed line of each electrode, RP
the resistance of the dielectric layer (leakage current) and C the
strain dependant capacity (C(SZ)).
Figure 12: Equivalent circuit of an electrostatic multilayer
actuator
The effective voltage drop across the dielectric layer is given
by
01
11 2 ( )C
S ZP
V VR j C S
R
=
+ +
(11)
and the strain dependant capacity C(SZ) is given by
0 00 0 0
0 0
(1 )( ) (1 )A
Z r r rZ
A dA A SAC Sz z dz z S
+ +
= = =
. (12)
With eq. (3) we can simplify the term, yielding
00 2
0
1( ) (1 )Z r ZAC Sz S
=
. (13)
Combining eq. (11) and (13) we get:
00
0 20
11 11 2 (1 )
C
S rZ P
V VAR jz S R
=
+ +
(14)
By equating relation (1) for the electrostatic pressure and
relation (10) for the mechanical stress we get the voltage VC,
needed to achieve the strain SZ:
00 0
1(1 )C z nr r
TV z z S G S
= = (15)
The coupling of mechanical and electrical model is given by eq.
(14) and eq. (15). With known model-parameters quasi-static and
dynamic performance can be developed. During deformation of an
active stack the dielectric layer is getting thinner. This causes a
faster increase of the electrostatic pressure (by the electric
field) than of the mechanical stress. At strain of SZ = 38.8 %
electrostatic pressure becomes larger than the mechanical stress.
This causes a pull-in effect until the destruction of the actuator
due to dielectrical breakdown.
128 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
4. CHARACTERISATION
To characterise the dynamic behaviour of stack actuators special
prototypes have been realised. Fig. 13 shows the prototype design
and its dimensions. The thickness of the dielectric layer
(silicone: Wacker Elastosil P 7670) is 25 m, electrodes are 5 m
high. The permittivity of the elastomer is r = 3. Stacks with 22
and 52 layers have been investigated the two layers on the surfaces
of the stack are to protect the electrodes. Material parameters are
ascertained by cast cylindrical probes.
Figure 13: Actuator design
4.1 Quasi-static behaviour The stress-strain characteristics is
used to determine the quasi-static material parameters. Therefore a
silicone probe ( 10 mm; height 10 mm) has been compressed unaxially
and the force-deflection characteristic has been determined. The
proceeding velocity is 10 m/s. Fig. 14 shows the
stress-strain-characteristic compared to the theoretical model (eq.
(6)). Stress values concern the face of the uncompressed probe (A0)
the systematic error for small deflection is quiet low.
0
0,1
0,2
0,3
0,4
0,5
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7strain Sz
str
es
s T
m [M
Pa]
modelmeasurement
Figure 14: Stress-strain characteristic
We can see the hysteresis in the measurement, due to
viscosity-modulus (1,2) and the friction on the contact surface.
This friction is minimized by using vaselin. To eliminate the
hysteresis for calculation, the model has been approximated to a
minimum mean error of 7.8 %. The determined Youngs modulus is Y1=
123.717 kPa. This Youngs modulus can be used to calculate (eq.
(15)) the strain of an actuator for an applied voltage. The real
strain as function of the supplied voltage is determined with a 20
layer actuator. To measure the deflection the actuator is put into
a mechanical frame to prevent several sources of error (e.g.
cambers). The deflection is measured by an optical
surface-profilometer. This characteristic is shown in Fig. 15.
Proc. of SPIE Vol. 5759 129
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
00,05
0,1
0,15
0,2
0 500 1000 1500 2000voltage [V]
str
ain
S z
measuredmodelcorr. Modell
Figure 15: strain-voltage-characteristic
We obtain a similar characteristic between model and
measurement, the latter shows a gradient. For this purpose we
assume the following reasons:
the outer silicone layers are not under electrostatic pressure:
they impede transverse strain reducing the linear compression
the passive elastomer surrounding the electrodes constricts
plane expansion and impedes linear compression, too
the graphite-powder layers are non-ideal electrodes : incomplete
coverage and contact resistances between particles may cause
field-free areas ending up in a lower effective electrostatic
pressure
the Youngs modulus is determined with a larger probe: the
stiffness of a thin actuator slice is clearly higher
By establishing a constant correction factor (kf = 0.14) the
model characteristics can be fitted to the measurement. This leads
to a mean error of only 2.9 %. At high strain (SZ > 12 %) there
is an increasing deviation, probably caused by the decreasing
charge density with rising strain. The corrected model is based on
a (mathematically determined) electrostatic pressure of only 14 %.
This corresponds to a reduced effective field strength of 37.4
%.
4.2 Dynamic behaviour Dynamic material parameters have been
determined by measuring the mechanical impedance. The impedance
measurement head consists of an electrodynamic oscillator, a force
sensor and an acceleration sensor. This head is pressed onto the
silicone probe with a force F0 = 0.5 N. The oscillation of the head
causes the measured acceleration, integrated to its corresponding
velocity. The resulting velocity-force-ratio delivers the
mechanical impedance. Fig. 16 shows the absolute value of the
measurement at frequencies of 10 Hz to 1000 Hz
0,1
1
10
100
10 100 1000frequency [Hz]
lZ
me
chl [N
s/m
]
measurementmodelcorr. model
Figure 16: Mechanical impedance compared to calculated model
130 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
The model parameters (eq. (8)) were determined by a
curve-fitting using MATHEMATICA. The static parameter Y1 is already
known. It has to be mentioned, that the resulting mass of the
calculation includes the mass of parts of the measuring equipment.
This mass can be determined by an idle-measurement and is
substracted from the calculated mass (corr. model in Fig. 16).
Table 2 shows the resulting specific material parameters.
Table 2: determined specific material parameters (Elastosil
P7670) Y1 Y2 1 2 M
138948 Pa 676204 Pa 198 Pa s 13487 Pa s 0,1 Pa s2
These parameters are used in eq. (9) to determine the transfer
function of strain S and mechanical stress T. To get the frequency
response of a 50-layer actuator stack it has been supplied with a
sinusoidal AC voltage superposed with a DC voltage at frequencies
from 10 Hz upto 1000 Hz. Fig. 17 shows the measured frequency
response and the qulitative comparison of the
force-deflection-frequency resonse.
1E-5
1E-4
1E-3
1E-2
1E-1
1E+0
1E+1
10 100 1000frequency [Hz]
z/F
z/
U
model z/Fmeasurement z/U
Figure 17: Frequency response of an actuator and
force-strain-ratio frequency response of the dielectric.
Both curves show a slight decay for low frequencies. The
actuators lower resonant frequency (compared with the model) with a
higher resonance magnification is obviously a result of the higher
mass that results from the passive silicone surrounding the
electrodes. Assuming an actuator without silicone surrounding
silicone ring the frequency response will be close to the models
course.
5. APPLICATIONS
As mentioned in section 1 electrostatic solid-state actuators
fabricated as multilayer-stacks are primarily used for linear
motion. They provide low forces at high deflection. This
combination establishes new applications particularly in
microelectromechanical systems (MEMS). As a result of the
perpendicular effective motion direction actuators can easily be
grouped close to each other. Multilayer elastomer stack actuators
can meet a variety of applications in optics (micromirros),
switches, acoustics (micro loudspeakers) and microfluidics (valves,
pumps).9 A further field of application is the human tactile sense.
The actuators performance is ideal to adapt actuators onto human
skin, even if the actual needed supply voltage is very high. As an
actuator for a braille-display (Fig. 18) there is only a 4x2
actuator matrix needed. 10
Proc. of SPIE Vol. 5759 131
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
Figure 18: Braille display as a possible application of
mulilayer actuators
But this is quite far away from the achievable potential of
elastomer stack actuators. If we think of a tactile display
integrated into a data-glove we have to meet enormous requirements.
For stimulating the whole human hand there are some hundred
actuators needed due to the high selectivity and spatial resolution
of the human sense. These actuators have to be integrated into a
compliant body enclosing human hand a complex structure. As we see,
the possibility to fabricate (stack-) actuators simultaneously in
arrays is predicted for these applications. Fig. 19 a shows a
design concept for a tactile display. The planar arrangement of the
actuators into a hexagonal shape assures equivalent distances
between all adjacent actuators. Embedded bumps between actuator and
skin realise the force transmission: passive (non excited)
actuators are displacing the skin (fig. 19 b).
Figure 19: Matrix design for tactile displays
The excitation of a single actuator inside an actuator array is
a challenge. If the actuators are arranged as a passive matrix
every single line and column can be selected and the actuator gets
charged or discharged. Fig. 20 shows the principle with
periodically switched column and row lines. Due to the slow
discharging (through the non-ideal isolator between the electrodes)
the display needs to be refreshed at certain clearances. Such a
passive matrix shows the principle disadvantage of crosstalk. If
this crosstalk leads to noticeable displacements of the surrounding
actuators an active-matrix-arrangement has to be used (e.g. known
from TFT-displays).
Figure 20: Passive-matrix-arrangement with periodic
refreshment
It is obvious to use such a matrix as sensor, too. The changing
of the capacity due to external displacement can be measured.
Current investigations are about an actuator-sensor-system by the
use of elastomeric stack-actuators with embedded
sensor-capacities.
V
132 Proc. of SPIE Vol. 5759
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms
-
6. CONCLUSIONS
An automatic fabrication technology has been developed to
provide electrostatic multilayer stack actuators with elastic
dielectric. Its function has been determined by several prototypes
with different electrode-designs and different numbers of layers.
Further developments have to investigate the material of the
electrodes and their fabrication. There are already some known
materials having better properties (eg. low contamination of the
dielectric layer, higher area coverage, lower sheet resistance) but
to determine their exact characteristics as electrode they have to
be implemented into the existing process. Otherwise reproducibility
will be very hard to achieve. To obtain a fabrication technology
suited to different applications elastomers with different
mechanical properties e.g. compliance have to be qualified. The
developed dynamic, nonlinear model has been verified for a wide
frequency range. So the model can be used to design electrostatic
polymer actuators with elastic dielectric. Theoretical examinations
of the actuator with load can be done easily by using the
determined impedance. Further investigations are necessary to
understand the deviations between model and real actuator: the
description of an effective field strength is probably the best way
to identify imperfections of the real actuator. The electrical
interconnects of the multilayer electrodes have to be improved in
order to assure the function of each layer. Three-dimensional
analysis of integrated actuator arrays have to be performed (e.g.
FEM-analysis) to describe the displacement and dynamic behaviour
using integrated field simulation as well.
7. REFERENCES
1. Heartling, G. H.: Ferroelectric Ceramics History and
Technology, J. Am. Ceramic Soc., 82, 4, pp. 77-818, 1999. 2.
Tiersten, H. F.: Linear Piezoelectric Plate Vibrations Element of
the Linear Theory of Piezoelectricity and the
Vibrations of Piezoelectric Plates, Plenum Press, New York,
1969. 3. Yao, K. et al.: Design and Fabrication of a High
Performance Multilayer Piezoelectric Actuator with Bending
Deformation, IEEE Trans. Ultrasonics, Vol. 46, No. 4, 1999. 4.
Rogacheva, N.: The Theory of Piezoelectric Shells and Plates, CRC
Press, 1994. 5. Bauer, S.; Gerhard-Multhaupt, R.; Sessler, G.:
Ferroelectrets: Soft Electroactive Foams for Transducers,
Physics
Today, 37-43, Feb. 2004. 6. Pelrine, R.; Kornbluh, R.; Joseph,
J.; Heydt, R.; Pei, Q.; Chiba, S.:High-field deformation of
elastomeric dielectrics
for actuators, Materials Science and Engineering, C11, Elsevier
Science, pp. 89-100, 2000. 7. Pelrine, R.; Kornbluh, R.; Joseph,
J.; Chiba, S.: Electrostriction of Polymer Films for
Microactuators, IEEE
Proceedings of the 10th Annual International Workshop on Micro
Electro Mechanical Systems, MEMS 97, Nagoya, Japan, pp. 238-243,
1997.
8. Carpi, F.; Chiarelli, P.; Mazzoldi, A.; De Rossi, D.:
Electromechanical characterisation of dielectric elastomer planar
actuators: comparative evaluation of different electrode materials
and different counterloads, Sensors and Actuators, A107, Elsevier
Science, pp. 85-95, 2003.
9. Kornbluh, R. et. al.: Electroactive polymers: An emerging
technology for MEMS. SPIE Vol. 5344, 2004. 10. Heydt, R.; Chhokar,
S.: Refreshable Braille Display Based on Electroactive Polymers.
Int. Display Res. Conf.
IDRC 03, SID Soc. Information Display, 2003, P7.5.
Proc. of SPIE Vol. 5759 133
Downloaded From: http://proceedings.spiedigitallibrary.org/ on
05/19/2015 Terms of Use: http://spiedl.org/terms