New piezoelectric polymer for air-borne and water-borne

sound transducers
Reiner Kressmanna)
Institute for Telecommunications, Darmstadt University of Technology, Merckstr. 25,
D-64283 Darmstadt, Germany

Received 13 September 2000; accepted for publication 18 January 2001

Acoustic transducers made of a charged cellular polymer called EMFi have been designed and
investigated with respect to air-borne and water-borne sound. The longitudinal transducer constant
is around 90 pC/N, strongly exceeding the values of other piezoelectric polymers. This is mainly
attributed to the very low Youngs modulus of about 2 MPa. The acoustic impedance is only 2.6
104 kg/m2 s) and results in good matching to air but strong loading under water. Due to this
strong loading, a pronounced reduction of resonance frequency from 300 kHz in air down to 17 kHz
under water is observed. The experiments indicate that fluid loading is not only mass-like but also
compliant, reducing the transducers sensitivity below the resonance frequency of about 63
dB re 1 V/Pa 0.7 mV/Pa in air to 71 dB re 1 V/Pa under water. This compliance is attributed to
the mediums compressibility. Piezoelectricity of EMFi films is limited to temperatures below
70 C; above, irreversible discharge of trapped charges takes place. Furthermore, a second type of
EMFi, called OS was investigated, having a piezoelectric constant of 15 pC/N and a Youngs
modulus of 6 MPa. In quasi-static sensor measurements, the piezoelectric constant increases with
the applied load. This nonlinearity explains the higher values reported in other publications on the
same materials. 2001 Acoustical Society of America. DOI: 10.1121/1.1354989
PACS numbers: 43.38.Fx, 43.30.Yj, 77.84.Jd, 43.30.Jx SLE

I. INTRODUCTION

Polymers are well-known materials for acoustic transducers, both for air-borne and water-borne sound applications. The most important ones are Teflon due to its charge
storing capabilities and piezoelectric polyvinylidene fluoride
PVDF. In the last few years, scientific interest on charge
retention in cellular and porous polymers18 and inorganic
porous electrets9 has grown rapidly. In this paper, the properties of transducers fabricated with two types of a novel
polymer, the so-called electro-mechanical film EMFi, provided by the Finnish company VTT, are presented and explained by a simple model. The two types are called HS
high sensitivity and OS ordinary sensitivity.

II. SAMPLES

The EMFi film is a foam based on polypropylene which

is biaxially stretched during fabrication.1 This results in the
formation of lens-like air bubbles with diameters of about 10
to 70 m and a thickness of about 5 m resulting in a quite
soft material. The manufacturer reports for the HS high sensitivity type of the material a thickness of 70 m, a density
of 330 kg/m3, a Youngs modulus of around 1 MPa, a relative permittivity of 1.2 and a sensitivity of 160 pC/N. The
respective values for the OS type are thickness, 37 m; density, 550 kg/m3; Youngs modulus, 9 MPa; relative permittivity, 1.6, and sensitivity, 30 pC/N.1 The piezoelectricity is
induced by means of corona charging resulting in charge
storage inside the film.
a

Electronic mail: kre@nt.tu-darmstadt.de

1412

J. Acoust. Soc. Am. 109 (4), April 2001

EMFi films of several cm2 size metallized on one side

have been cut, glued on brass electrodes and mounted in
aluminum housings to assure good shielding. Transducers
with diameters of 20 and 35 mm active size were fabricated.
For the underwater measurements, only carried out with the
HS type, they have been covered with a thin waterproof film
to protect them from contact with the water. A schematic of
a transducer is given in Fig. 1.
III. EXPERIMENTAL WORK

Air-borne sound measurements have been carried out in

an anechoic chamber under free-field conditions and a coupler. For the water-borne sound experiments an approximately cubic plastic tank of 1 m3 size was used. The problems caused by its small opening of only 20 cm diameter are
discussed together with the results in the following section.
Calibrated sensors were from Bruel and Kjr quarterinch microphone 4135, eighth-inch microphone 4138 and
hydrophone 8103. Since the transducers active capacitances are between 44 and 400 pF, a charge amplifier 2635
from Bruel and Kjr was used in the sensor experiments.
Furthermore, an audio analyzer UPD from Rohde and
Schwarz was used for signal generation and analysis.
A. Actuator response

The radiated sound pressure of a circular EMFi-HS actuator of 35 mm diameter driven sinusoidally with 100 V is
shown in Fig. 2. The actuator was baffled slightly unsymmetrically with a wooden plate of 2015 cm2 size. The microphone was placed 60 cm apart from the sample on the
main axis. The sound pressure increases with the square of
frequency as expected below the resonance since in this

0001-4966/2001/109(4)/1412/5/$18.00

2001 Acoustical Society of America

1412

FIG. 3. Sound pressure level of an EMFi-HS projector (2R35 mm) at 20

cm distance 25 off axis. This unusual orientation results from the geometrical limitations of the small opening of the tank.
FIG. 1. Schematic of EMFi transducers with diameter 2R.

range the deflection of the piezoelectric does not change with

frequency. From interferometric measurements of dynamic
deflection and dielectric measurements, a resonance frequency of about 300 kHz is known.10 For the linear response,
the transmitting sensitivity at 1 m distance is 4.7 mPa/V at
100 kHz.
The resonance frequency changes drastically, if the
transducer is driven under water. The result shown in Fig. 3
is for an actuator of 35 mm diameter driven with 80 V at a
distance of 20 cm. Though the boundary of the tank is
strongly reverberant, it has been proved at a frequency of 10
kHz with gated actuation that the echoes from the walls are
10 dB weaker than the direct sound. The projector was
mounted with its back close to the water surface to prevent
contact at the electrical connectors with water. Therefore, the
projector is baffled by the sound-soft water surface. Unfortunately, the adjustment of the reference hydrophone on the
main axis of the sample was not very successful. This mainly
results from the limited space for mounting transducers due
to the small opening of the tank of just 20 cm diameter.
Furthermore, the tanks upper wall does not allow the usage
of heavy supports. Hence, the high frequency response of the
sample can only be explained by assuming a deviation of 25
degrees off axis. Since this coincides with a 3 dB loss at 53
kHz it does not corrupt the data at lower frequencies.
The resonance shifts down to around 17 kHz which is
attributed to the medium load. The transmitting sensitivity in
water at 1 m distance is 8 mPa/V at 10 kHz.

FIG. 2. Sound pressure level of an EMFi-HS actuator in air (2R

20 mm) at 60 cm distance.
1413

J. Acoust. Soc. Am., Vol. 109, No. 4, April 2001

The transmitting sensitivity in air of an EMFi-OS actuator not shown of 20 mm diameter is 0.32 mPa/V at 100
kHz.
Both in water and in air, a quite strong second harmonic
is present in the actuator experiments, whereas other harmonics do not occur. The second harmonic is due to an electrostatic effect as shown later in Eq. 7. Unpoled films only
radiate the second harmonic and no fundamental. Poled
films, too, show a similar behavior after thermal discharge.
In this experiment, an EMFi actuator was placed inside an
oven, and the radiated fundamental and second harmonic
were recorded. The result is depicted in Fig. 4. The acoustic
signals have not been very stable due to refraction caused by
temperature gradients in the air between sample and microphone.
Discharge takes place between 70 C to 100 C and is
irreversible.1,3 This is not surprising, since also solid polypropylene is known to show weak charge storage
capabilities.11 The slight decrease of the sound pressure level
during cooling is attributed to the lower Youngs modulus at
elevated temperature. After the experiment it could be seen
at the first view that the sample could not further be used.
B. Sensor response

The sensitivity to air-borne sound of a transducer of 20

mm diameter is depicted in Fig. 5. At low frequencies, a
three-port coupler with a diameter of 20 mm and a volume of
3 cm3 was used. It guarantees a constant sound pressure up
to 6 kHz. In the ultrasonics range, the substitution method in

FIG. 4. Radiated sound pressure level in arbitrary units a.u. at 50 kHz

fundamental and 100 kHz second harmonic versus temperature.
Arrows indicate heating and cooling, respectively. The heating rate was
200 K/h.
R. Kressmann: Piezoelectric polymer transducers

1413

From Fig. 6 it can be seen that the resonance is shifted

down to 17 kHz as it appears in the actuator mode as well.
Furthermore, the sensitivity below the resonance frequency
is 0.3 mV/Pa, i.e., it is 8 dB lower in water than in air.
IV. THEORY
A. Mechanicalacoustical conversion

1. General remarks
FIG. 5. Sensitivity of EMFi transducers to airborne sound (2R20 mm).
The low-frequency sensitivity was recorded with a three-port coupler with
constant pressure up to 6 kHz. The sensitivity with respect to ultrasound was
measured in free field using the substitution method. The difference of the
values is attributed to the uncertainties of the measurements.

free field12 using a baffled EMFi transducer as an actuator

was used. The reference microphone and the EMFi sensor
have been placed in the actuators far field on the main axis
and carefully adjusted with respect to inclination.
The frequency response of the EMFi-HS sensor is quite
flat between 4 Hz and 6 kHz at about 0.7 mV/Pa. This corresponds to a longitudinal piezoelectric charge constant of
about 100 pC/N, exceeding the data of PVDF, for instance,
by about 10 dB. EMFi-OS sensors have a sensitivity of about
60 V/Pa yielding a piezoelectric constant of about 15 pC/N.
In the ultrasonics range, the sensors still work, but show
strong directivity due to their size. The equivalent
A-weighted noise level of the EMFi-HS microphone is 52
dB with the present setup. Even for sound pressure levels of
up to 144 dB at 300 Hz, the total harmonic distortion
THD of both sensor types does not exceed that of the
eighth-inch reference microphone B & K 4138, which is
just 3% at 168 dB. No change in the piezoelectric constant is
observed for applied pressures between 20 mPa and 340 Pa.
The sensitivity of the EMFi-HS transducer used as hydrophone was measured with the comparison method and is
shown in Fig. 6. Owing to the large reverberation time inside
the tank, several frequency responses under various geometric configurations have been monitored. In this way, it is
possible to average over the various spatial modes inside the
tank. Below 6 kHz, the sound pressure level produced by the
projector is too low to permit an accurate measurement of
the sensitivity, and the number of modes inside the tank is
too small to guarantee a successful spatial averaging.

In both air-borne and water-borne sound experiments the

actuator can be modeled as a circular piston mounted in an
infinite baffle13 radiating the power NReZmv2 with the
radiation impedance

Z m c R 2 1

J 1 2kR
S 1 2kR
j
kR
kR

and having a directivity ratio of

D

kR 2
.
1J 1 2kR /kR

In these equations, v is the velocity of the vibrating piston,

c is the characteristic impedance of the medium, R is the
pistons radius, k is the wave number, and J 1 (x) and S 1 (x)
are the Bessel and Struve function of first order,
respectively.14 Then, the sound pressure amplitude on the
main axis at distance r can be computed:
p

4 r

cND

1
R 2 2 x,
2r

where x is the deflection of the surface. The 2 increase of

the pressure is confirmed experimentally up to about 100
kHz see Fig. 2. The strong decrease of sound pressure of
the second harmonic at frequencies above 150 kHz is mainly
attributed to a decrease in sensitivity of the reference microphone.
The following calculations have been carried out for the
HS type of EMFi films: The 2 increase with frequency
holds from infra-sonic frequencies up to the mechanical resonance, above which x in Eq. 3 decreases. The mass per area
is calculated from the geometry and the density of the film to
be M 7.7103 kg/m2, in which it is considered that the
dynamic mass is only one-third of the static one since the
layer is clamped on one side. Any kind of impact of the
backing on the resonance frequency can be excluded. The
compliance is calculated from Youngs modulus Y 2 MPa
to be F3.51011 m3/N. These data result in a resonance
frequency of 307 kHz.
2. Fluid loading

For the water-borne sound measurement, carried out

with transducers of 35 mm diameter, an additional mass load
per areacomputed from the imaginary part of Eq. 1 using
a power expansion being valid for low frequencies
FIG. 6. Sensitivity of EMFi-HS under water hydrophone, 2R35 mm.
Below 6 kHz the data are corrupted by insufficient signal to noise ratio and
unsuccessful spatial averaging.
1414

J. Acoust. Soc. Am., Vol. 109, No. 4, April 2001

M s

8R
14.4 kg/m2
3

R. Kressmann: Piezoelectric polymer transducers

4
1414

data for the EMFi-OS actuator are d 1 23 m, d 2

14 m, d 3315 pm/V.
A Youngs modulus of Y 2 MPa EMFi-HS and Y
6 MPa EMFi-OS was calculated from the electrostatic
force per area A,

Y
F el
2 U 2 x
A
2d
d

FIG. 7. Schematic of simplified model of EMFi film.

has to be considered. The shifted resonance frequency would

be 7 kHz. Therefore, a further effect must be present, resulting in a slightly higher resonance and reduced sensor sensitivity at low frequencies. This effect seems to be an additional restoring force due to the medium: Considering that
the same volume which is moved as a mass M s is also compressed, the mediums compliant load times the area is
F s

8R
81012 m3/N,
3

using the compressibility of water 51010 Pa1. As a

consequence, the total compliance is reduced by a factor of
about 5 which yields a resonance frequency of 18 kHz being
in fair agreement with the experiment.
Stiffness terms in radiation impedances of fluid-loaded
vibrating plates are well known in the literature for certain
vibration modes.1517 In these experiments, however, we
have no reason to assume any kind of mode distribution on
the transducers surface, and it must be premised that the
surface vibrates piston-like. Furthermore, radiation impedances with negative imaginary part only appear in a small
frequency range,15 whereas the present experiments show,
that both mass and compliant loading from the fluid take
place over a broader frequency range. This effect might not
have been observed in the past since its impact was small.
From Eq. 4, it can be seen that mass loading increases with
increasing sensor area.

B. Electro-mechanical conversion

The electro-mechanical coupling of a stack of two dielectrics of different elasticity with a charge layer in between
which is similar to an electret microphone18 has been already
investigated19 and extended to charged cellular polymers.2
The latter model assumes a material composed of alternating
air and solid layers with a finite charge density at the
boundary. It can be further simplified as shown in Fig. 7. The
following transducer constant, identical both for sensor and
actuator mode, is deduced:
d 33

d 1
d
x
.
Y d 1 d 2 2 U

In this equation, for HS-type transducers, d 1 26 m is

the total thickness of the solid layers, 2.35 is their permittivity, and d 2 44 m is the total thickness of the air
layers. d 1 and d 2 are calculated from the total thickness d of
70 m and the density of the film. From the measured sound
pressure at the fundamental, a transducer constant of about
80 pm/V is computed from Eqs. 3 and 6. The respective
1415

J. Acoust. Soc. Am., Vol. 109, No. 4, April 2001

which results for the actuating mode in radiation of the second harmonic due to the nonlinearity with respect to the
driving voltage Fig. 2. Inserting these data into Eq. 6, a
charge density of 6104 C/m2 for EMFi-HS and of
1.6104 C/m2 for EMFi-OS are computed. The actuator
constant has been also deduced from interferometric measurement of deflection to be between 85 and 100 pm/V.
From this experiment a resonance frequency of about 300
kHz was obtained10 resulting in a Youngs modulus of 2
MPa.

V. DISCUSSION

The obtained piezoelectric constants vary up to 20%

from sample to sample, but are, on average, much smaller
than the data reported in the literature.2,20 These published
higher values are observed by applying a static load of about
104 Pa to the sample, whereas in the present sensor experiments, the load was between 20 mPa and 340 Pa. The authors experiments with static loading have shown, that in
this method a nonlinearity with respect to the loading appears. Furthermore, in the large-signal regime, the piezoelectrically induced charge increases, if just a small area of the
film is loaded. In this case, the air can flow from the bubbles
of the loaded area into the unloaded neighborhood resulting
in an apparently softer film. This also explains the slow relaxation reported in Ref. 20 which was also observed in our
experiments with quasi-static loading. The nonlinearity with
respect to loading will be discussed in more detail in a future
publication. A similar low value of 110 pm/V is reported in
Ref. 21 for the actuating mode monitoring the samples deflection under electrical load with an accelerometer.
Since the nonlinearity appears in the stressstrain relation, it affects both the measurement of piezoelectric constant and Youngs modulus. But, the product of both, related
to the surface charge at the interface see Eq. 6 is independent of this effect. Therefore, the lower Youngs modulus of
about 1 MPa reported in Refs. 2 and 20 leads to the same
surface charge.
Charged porous polymers show longitudinal piezoelectric constants being much larger than those known for other
polymers, e.g., polyvinylidene fluoride PVDF. This follows
mainly from the very low Youngs modulus of EMFi-HS
layers which also results in a very low acoustic impedance of
about Z Y 2.6104 kg/m2 s. Therefore, cellular layers are matched better to air than to water. It should be
clearly stated, that EMFi films are no piezoelectrics in the
classical sense. Polypropylene does not contain any dipoles,
and our measurement of the transversal coefficient yielded
R. Kressmann: Piezoelectric polymer transducers

1415

the very low value of d 312 pC/N. The same value is reported in Ref. 10. Hence, EMFi films have been called
pseudo-piezoelectrics.2
The observed low stability of EMFi films against heating is insufficient for many applications; it is inferior to that
of other common piezoelectrics, e.g., PVDF and PZT. This
problem might be overcome by using Teflon as a basic material known to have excellent charge storage properties.11
Charge retention in porous polytetrafluorethylene has been
already investigated,4,5 and the piezoelectricity of various
solid double layers with charge layer at the boundary was
reported,68 but no experiments on temperature dependence
and transducers are reported up to now. In the latter publication, piezoelectric coefficients up to 600 pC/N are reported
for double layers, obtained from quasi-static measurements.
It should be remembered that our own measurements were
all obtained on the simpler and mechanically more stable
single layers.
VI. CONCLUSIONS

The presented acoustic experiments on transducers for

water-borne and air-borne sound show that charged cellular
polymers are promising candidates for several applications,
mainly as hydrophones in shallow water or in air as ultrasonic sources or broadband sensors. The flexibility of the
polymer allows the fabrication of large-scale transducers
with almost arbitrary directivity pattern e.g., a true omnidirectional transducer at high frequencies, if the film is fixed
on a curved substrate which can also contain a more or less
sophisticated electrode pattern. The underwater experiments
seem to indicate that a low impedance material as the EMFi
film is loaded by the medium simultaneously mass-like and
compliant. The very weak transversal piezoelectricity of 2
pC/N see above and Ref. 10 avoids bending of actuators
glued on a substrate. For narrow-band applications, the resonance frequency can be tuned by applying an additional mass
on top of the film which might be also desirable for protection against any kind of environmental influence.
The main future scope is to fabricate layers with higher
temperature stability. Furthermore, the interaction between
EMFi films and water should be investigated more deeply.
Finally, the microscopic structureboth mechanically and
electricallyof cellular polymers is of great interest to improve material properties and to understand the nonlinear
behavior. Higher piezoelectric constants are possible by further softening films through enhanced gas content and higher
charge density.
ACKNOWLEDGMENTS

The author thanks Professor G. M. Sessler, H. Berger,

M. Fischer, J. Meyer, and F. Steffens for stimulating discussions and help with some of the experiments. Bruel and
Kjr, Copenhagen, Denmark, is gratefully acknowledged for

1416

J. Acoust. Soc. Am., Vol. 109, No. 4, April 2001

providing the hydrophone, and VTT, Tampere, Finland, for

making the EMFi films available. Financial support of the
Deutsche Forschungsgemeinschaft DFG is also acknowledged.
1

J. Lekkala and M. Paajanen, EMFiNew electret material for sensors

and actuators, Proceedings of the 10th International Symposium on Electrets, Delphi IEEE, Piscataway, NJ, 1999, pp. 743746.
2
G. M. Sessler and J. Hillenbrand, Electromechanical response of cellular
electret films, Appl. Phys. Lett. 75, 34053407 1999.
3
J. v. Turnhout, R. E. Staal, M. Wubbenhorst, and P. H. de Haan, Distribution and stability of charges in porous polypropylene films, Proceedings of the 10th International Symposium on Electrets, Delphi IEEE,
Piscataway, NJ, 1999, pp. 785788.
4
Z. Xia, J. Jian, Y. Zhang, Y. Cao, and Z. Wang, Electret properties for
porous polytetrafluorethylene PTFE Film, CEIDP, 1997 Annual Report
IEEE, Piscataway, NJ, 1997, pp. 471474.
5
R. Schwodiauer, G. Neugschwandtner, S. Bauer-Gogonea, S. Bauer, J.
Heitz, and D. Bauerle, Dielectric and electret properties of novel Teflon
PTFE and PTFE-like polymers, Proceedings of the 10th International
Symposium on Electrets, Delphi IEEE, Piscataway, NJ, 1999, pp. 313
316.
6
R. Gerhard-Multhaupt, Z. Xia, W. Kunstler, and A. Pucher, Preliminary
study of multi-layer space-charge electrets with piezoelectric properties
from porous and non-porous Teflon films, Proceedings of the 10th International Symposium on Electrets, Delphi IEEE, Piscataway, NJ,
1999, pp. 273276.
7
W. Kunstler, Z. Xia, T. Weinhold, A. Pucher, and R. Gerhard-Multhaupt,
Piezoelectricity of porous polytetrafluoroethylene single- and multiplefilm electrets containing high charge densities of both polarities, Appl.
Phys. A: Mater. Sci. Process. 70, 58 2000.
8
G. Neugschwandtner, R. Schwodiauer, S. Bauer-Gogonea, and S. Bauer,
Large piezoelectric effects in charged, heterogeneous fluoropolymer
electrets, Appl. Phys. A: Mater. Sci. Process. 70, 14 2000.
9
Y. Cao, Z. Xia, Q. Li, L. Chen, and B. Zhou, Study of porous dielectrics
as electret materials, IEEE Trans. Dielectr. Electr. Insul. 5, 5862
1998.
10
G. S. Neugschwandtner, R. Schwodiauer, M. Vieytes, S. Bauer-Gogonea,
S. Bauer, J. Hillenbrand, R. Kressmann, G. M. Sessler, M. Paajanen, and
J. Lekkala, Large and broadband piezoelectricity in charged polypropylene foam electrets, Appl. Phys. Lett. 77, 38273829 2000.
11
R. Kressmann, G. M. Sessler, and P. Gunther, Space-charge electrets,
in Electrets, 3rd ed., edited by R. Gerhard-Multhaupt Laplacian, Morgan
Hill, 1999, Vol. 2, Chap. 9, pp. 140.
12
V. Nedzelnitsky, Calibration of pressure and gradient microphones, in
Encyclopedia of Acoustics, edited by M. J. Crocker Wiley, New York,
1997, Vol. 4, Chap. 157, pp. 18691879.
13
H. F. Olson, Acoustical Engineering van Nostrand, New York, 1957.
14
M. Abramowitz and I. A. Stegun, Handbook of Mathematical Functions,
9th ed. Dover, New York, 1970.
15
M. Lax, The effect of radiation on the vibrations of a circular diaphragm, J. Acoust. Soc. Am. 16, 513 1944.
16
P. W. Smith, Jr., The imaginary part of input admittance: a physical
explanation of fluid-loading effects on plates, J. Sound Vib. 60, 213216
1978.
17
D. G. Crighton, The 1988 Rayleigh medal lecture: Fluid loadingThe
interaction between sound and vibration, J. Sound Vib. 133, 127
1989.
18
G. M. Sessler and J. E. West, Applications, in Electrets, 3rd ed., edited
by G. M. Sessler Laplacian, Morgan Hill, 1999, Vol. 1, Chap. 7, pp.
347381.
19
R. Kacprzyk, E. Motyl, J. B. Gajewski, and A. Pasternak, Piezoelectric
properties of nonuniform electrets, J. Electrost. 35, 161166 1995.
20
M. Paajanen, H. Valimaki, and J. Lekkala, Modelling the electromechanical film EMFi, J. Electrost. 48, 193204 2000.
21
J. Hillenbrand and G. M. Sessler, Piezoelectricity in cellular electret
films, IEEE Trans. Dielectr. Electr. Insul. 7, 537542 2000.

R. Kressmann: Piezoelectric polymer transducers

1416