Characterization of an optical multilayer hydrophone with

constant frequency response in the range from 1 to 75 MHz

Volker Wilkensa)
Physikalisch-Technische Bundesanstalt, Fachlabor Ultraschall, Bundesallee 100, 38116 Braunschweig,
Germany

共Received 29 July 2002; revised 13 December 2002; accepted 17 December 2002兲

The performance of an optical multilayer hydrophone for ultrasound measurements is investigated
both in theoretical terms and experimentally. The optical measurement system comprises a thin
high-finesse dielectric interference filter structure that is deposited onto a plane glass plate. An
incident acoustic pressure wave deforms the layer system, and the induced variation of the optical
reflectance is determined. Applying the concept of an optical off-axis detection scheme offers good
sensitivity and a simple and low-cost setup. A primary interferometric calibration technique is
applied to experimentally determine the pressure–voltage transfer function in the range from 1 to 75
MHz. Within the measurement uncertainty a constant transfer factor is obtained for the whole
frequency range. Measurements of broadband ultrasound pulses are influenced neither by acoustic
resonances of the very thin sensing element nor by diffraction phenomena that are known to cause
waveform distortions in small probe hydrophone measurements. High temporal and spatial
resolution is combined with high durability of the probe, which is why the optical multilayer
hydrophone is well suited for use as a reference for secondary hydrophone calibration. © 2003
Acoustical Society of America. 关DOI: 10.1121/1.1553457兴
PACS numbers: 43.38.Zp, 43.35.Yb, 43.30.Yj 关YHB兴

I. INTRODUCTION reference detector for secondary calibrations of hydrophones

Optical measurement systems have proved to be useful
tools for the detection of ultrasound, in particular, in the
high-frequency range. The use of thin optical dielectric lay-
ers as a sensor element in both a fiber-optic hydrophone and
an optical detection array configuration has been reported on
previously.1,2 Acoustic measurement here is based on the
elastic deformation of the multilayer system by an incident
ultrasound wave. The thin layer system is designed to yield a
large change in optical reflectance at the operational light
wavelength produced by this mechanical deformation. The
change in reflectance is detected using a simple optical setup,
and high temporal and spatial resolution of the ultrasound
measurement can be achieved. The frequency response of the
fiber-optic version of the multilayer hydrophone is strongly
influenced by the acoustic properties of the fiber end onto
which the layer system is deposited. It can be understood as
the result of the superposition of longitudinal, edge diffrac-
tion, and lateral waves with a resonant vibration mode of the
fiber body representing an elastic rod.3 The pressure wave-
forms of broadband ultrasound pulses measured with such a
fiber-optic multilayer hydrophone are strongly distorted as a
result of the nonuniform frequency response. If a probe is
used that comprises a similar layer system deposited onto a
plane glass plate with a diameter larger than the lateral sound
field dimension, such distortions do not occur since edge and
resonance effects are avoided.2 In addition to the uniform
frequency response, this sensor concept basically offers high
temporal and spatial resolution and high durability of the
probe, which is why the system is well suited for use as a
a兲
Electronic mail: Volker.Wilkens@ptb.de
and fiber-optic ultrasound sensors, especially for the high-
frequency range.
In this paper, the concept of an optical off-axis detection
scheme for the multilayer hydrophone is investigated theo-
retically and by experiments. It offers both, good sensitivity
and a simple and low-cost setup, and adjustment to the op-
erating point of the optical multilayer resonator can be
achieved by variation of the illumination angle.4,5 A similar
adjustment technique has been demonstrated to be useful for
high-finesse etalon-type ultrasound sensors6 and more re-
cently for low-finesse polymer film Fabry–Perot-type ultra-
sound sensors.7 The acoustic performance of the multilayer
hydrophone is experimentally analyzed. The characterization
given comprises an estimation of the lateral resolution, the
frequency response determined by a primary interferometric
calibration in the range from 1 to 75 MHz, the sensitivity, i.e.
the minimum detectable sound pressure level of the present
system, the directional response, and the stability.
II. SENSOR DESIGN, PRINCIPLE OF OPERATION,
AND THEORETICAL DESCRIPTION
The sensor element of the optical multilayer hydrophone
consists of a dielectric layer system acting as a microinter-
ferometer deposited on a plane glass plate 共Fig. 1, insert兲.
The high-finesse dielectric interference filter structure is of a
design well suited for use as a pressure sensing element at
ultrasound frequencies.1,5 Two high-reflection subsystems,
both consisting of several ␭ D/4 layers, are connected by a
central ␭ D/2 spacer layer, where ␭ D denotes the optical de-
sign wavelength at which the interference filter offers maxi-
mum transmission. This structure combines a high optical

J. Acoust. Soc. Am. 113 (3), March 2003 0001-4966/2003/113(3)/1431/8/$19.00 © 2003 Acoustical Society of America 1431

finesse of the Fabry–Perot interferometer with a very small
overall thickness, and good acoustic sensitivity together with
a large detection bandwidth is obtained.
The principle of ultrasound measurement is based on the
elastic deformation of the layer system by an incident pres-
sure wave p in and the measurement of the induced change in
optical reflectance ⌬R by an optical detection scheme 共Fig.
1兲. The optical arrangement for a single element detector can
be considerably simplified if the oblique incidence of light is
utilized instead of perpendicular illumination in the setup
previously described.2 Only very few optical components
comprising the laser source, two lenses, the sensor probe,
and a photodetector are necessary. Focusing of the laser
beam yields high spatial resolution of the ultrasound mea-
surement. Since backreflection into the laser source does not
occur due to the off-axis detection scheme, optical isolation
is not necessary. Backreflections from the rear of the sub-
strate are laterally separated from the light reflected at the
layer system and can easily be masked to avoid parasitic
interference. Furthermore, a beamsplitter needs not to be in-
serted for the detection of the light reflected at the layer
system, which offers highly efficient use of light, and a laser
source with comparatively low output power is sufficient to
achieve good sensitivity of the optical hydrophone. For the
monochromatic light source here a HeNe laser is preferred to
a laser diode because mode competition noise typical of laser
diodes would give reason to increased intensity noise in the
reflected part of light.5
For a theoretical description the system of N dielectric
layers is assumed to be optically nonabsorbing, homoge-
neous, and isotropic. The optical reflectance R of the system
for oblique incidence of light with wavelength ␭ can then be
calculated using the matrix formalism for optical multilayer
systems,8 where the single ith layer with the physical thick-
ness d i , refractive index n i , and angle of light propagation
␪ i is described by its angle-dependent phase thickness:
2␲
␾ i⫽ n d cos ␪ i , i⫽1,...,N, 共1兲
␭ i i
1432 J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
FIG. 1. Optical multilayer hydrophone
setup; insert: multilayer system depos-
ited onto a plane glass substrate; i
⫽0: substrate, i⫽1,...,N: dielectric
layers, i⫽N⫹1: surrounding water,
N: total layer number, H, L: high-
index and low-index optical ␭ D/4
layer, respectively, p in : incident pres-
sure wave, I in : incident, and I ref : re-
flected optical intensity, n i , ␪ i : re-
fractive index and light propagation
angle, respectively, in the ith medium.
and its characteristic matrix:
冋 册
冑⫺1 sin ␾ i
cos ␾ i
Mi ⫽ ␥i , i⫽1,...,N. 共2兲
冑⫺1 ␥ i sin ␾ i cos ␾ i
Here, ␥ i denotes the optical admittance of the layer that de-
pends on the state of polarization of the incident light:
␥ i⫽ 冑␮⑀ ␪ 0
0
n i cos i, for perpendicular 共 s 兲 polarization;
冑␮⑀ 冒
共3兲
0
␥ i⫽ ni cos ␪ i , for parallel 共 p 兲 polarization,
0
with the absolute dielectric constant and permeability of the
free space ⑀ 0 and ␮ 0 , respectively. The whole layer system
is described by the product of the N individual characteristic
matrices taken in the correct order:
冋 册
N
m 11 m 12
M⫽ 兿 Mi ⫽
i⫽1 m 21 m 22
. 共4兲
The complex amplitude coefficient of reflection is then given
by the four components m 11 ,...,m 22 of M and the optical
admittance of the fiber substrate ␥ 0 and of the surrounding
water ␥ N⫹1 (n 0 ⫽1.48, n N⫹1 ⫽1.329):
␥ 0 m 11⫹ ␥ 0 ␥ N⫹1 m 12⫺m 21⫺ ␥ N⫹1 m 22
r⫽ . 共5兲
␥ 0 m 11⫹ ␥ 0 ␥ N⫹1 m 12⫹m 21⫹ ␥ N⫹1 m 22
The intensity reflectance is R⫽r•r * , the asterisk denoting a
complex conjugate.
In Fig. 2 the calculated reflectance R is depicted in de-
pendence on the normalized design wavelength ␭ D /␭ for a
19-layer system with n⫽2.30 for the high index (Nb2 O5 )
and n⫽1.48 for the low index material (SiO2 ), for three
different angles of light incidence in air ␪ in⫽0°, 25°, and
35°. For normal incidence of light, the high-finesse interfer-
ence filter design produces a narrow transmission peak at ␭
⫽␭ D⫽2n 10d 10⫽4n i d i (i⫽1,...,9,11,...,19). An incident
pressure wave causes modulation of the optical thicknesses
Volker Wilkens: Optical multilayer hydrophone

FIG. 2. Optical reflectance of the 19-layer system in dependence on the
normalized optical thickness ␭ D /␭ for three different angles of light inci-
dence ␪ in for parallel 共p兲 and perpendicular 共s兲 polarization. Operational
points with maximum ⌬R/p in at the edge of the transmission peaks 共䊉兲.
of the layers and, thus, a modulation of optical reflectance of
the whole layer system. The operating point of maximum
reflectance slope ⌬R/p in is located at the edge of the trans-
mission peak, where the reflectance is about R 0 ⬇0.25.1,5 An
alteration of ␪ in gives rise to a change of the light propaga-
tion angles ␪ i in the substrate, the layers, and the water 共cf.
Fig. 1兲 to be calculated by an appropriate application of
Snell’s law. Due to the changes in the phase thicknesses ␾ i
of the layers 关Eq. 共1兲兴, a shift of both the transmission peak
and the operating point toward higher values of ␭ D /␭ is
induced. This behavior offers an important advantage over a
setup in which the light incidence is restricted to the perpen-
dicular direction: By choice of the appropriate angle ␪ in that
depends on the laser wavelength ␭ and the design wave-
length ␭ D of the specific interference filter sample used, the
microinterferometer can be precisely adjusted to the operat-
ing point for ultrasound detection without tuning of the laser
wavelength. In general, an adjustment is necessary because
thickness deviations of the central spacer layer that in the
manufacturing process are not completely avoidable cause a
shift of the design wavelength from the desired value.1 In
addition to the spectral shift, a variation of the width of the
transmission peak of the interference filter in dependence on
␪ in can be seen in Fig. 2, which is caused by the angle- and
polarization-dependent optical admittances 关Eq. 共3兲兴. For p
polarization the finesse decreases with increasing ␪ in due to
decreasing absolute values of the complex Fresnel reflectiv-
ity coefficients for each layer boundary, whereas for s polar-
ization, the finesse increases due to increasing absolute val-
ues of the coefficients.
For a quantitative description of the reflectance slope
⌬R/ p in it is assumed that a perpendicularly incident plane
ultrasonic wave produces a pure longitudinal wave within
the sensor with a displacement w(z) in the layer system and
the substrate in the axial direction only. Negligible radial
displacement can be assumed on the axis of the acoustic field
where the small laser spot is located. Because of the radial
symmetry with respect to the z axis, no torsional component
occurs. Since the layer thicknesses d i (i⫽1,...,N) in all
practical cases are small with respect to the acoustic wave-
lengths ⌳ in the layer materials, the axial strain ⳵ w(z)/ ⳵ z
can be assumed constant along the z axis within each layer.
J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
For realistic layer numbers N, the thickness of the whole
layer system is also small compared with ⌳ in the system,
and the axial stress can be assumed constant within the
whole layer system and is given by the negative value of the
pressure p in the water in front of the sensor. Because of the
acoustic impedance mismatch at the surface of the sensor, a
reflected acoustic wave occurs and the pressure p in front of
the sensor is about 1.8 times the free-field pressure of the
incident wave p in .
The deformation leads to a change ⌬d i in the physical
thickness of the ith layer:
⌬d i ⫽ 冉 冊
⳵w
⳵z i
d i ⫽⫺
p
d ,
␳ i v 2i i
共6兲
where ␳ i and v i denote the density and longitudinal wave
velocity, respectively, of the ith layer. Since no exact material
parameters for the sputtered layers were available, the bulk
parameters are used for approximate calculation instead
(SiO2 : ␳⫽2250 kg m⫺3, v ⫽5800 m s⫺1; Nb2 O5 : ␳⫽4500
kg m⫺3, v ⫽6220 m s⫺1兲.5 However, a modulation of the re-
fraction coefficient n i of each layer and the substrate (i
⫽0,...,N) results from the strain according to the elasto-optic
effect:9
冉 冊
1 ⳵w 1 p
⌬n i ⫽⫺ n 3i p 12i ⫽ n 3i p 12i 2 , 共7兲
2 ⳵z i 2 ␳ iv i
where p 12i denotes the relevant strain-optic coefficient of the
ith medium. Here the value for optical quartz fibers p 12
⫽0.27 taken from Ref. 9 is used for both materials.3 The
effects of strain-induced birefringence are neglected. In ad-
dition to the effects of layer deformation, the modulation of
the refractive index n N⫹1 in the water by the pressure p in
front of the sensor contributes to a variation of the optical
reflectance ⌬R of the multilayer system:10
⌬n N⫹1 ⫽ p⫻1.5⫻10⫺4 MPa⫺1 . 共8兲
However, in the case of the 19-layer system, the latter con-
tribution is very small compared with the interference effect
in the layer system.
In Fig. 3共a兲 the maximum reflectance slope D⫽⌬R/p in
of the 19-layer system at the operational wavelength is de-
picted in dependence on the angle of light incidence ␪ in cal-
culated with Eqs. 共1兲–共8兲, where plane optical waves were
assumed within the layer system. With increasing angle the
maximum reflectance slope increases for s polarization and
decreases for p polarization, depending on the variations of
the optical finesse of the multilayer system described above.
Thus, s polarization should preferably be used to achieve
high sensitivity. Figure 3共b兲 shows the shift in the operational
optical wavelength ␭ of maximum reflectance slope in de-
pendence on the angle of light incidence. For a given design
wavelength ␭ D the operational wavelength ␭ decreases with
increasing angle ␪ in .
Since the thickness modulations are small compared
with the optical layer thicknesses of the undeformed system,
a linear relationship between pressure amplitude and change
in optical reflectance is realized. The calculated change in
reflectance ⌬R in dependence on the positive peak pressure
Volker Wilkens: Optical multilayer hydrophone 1433

FIG. 3. Maximum reflectance slope D⫽⌬R/p in 共a兲 of the 19-layer system at
the appropriate operational wavelength 共b兲 in dependence on the angle of
light incidence ␪ in for parallel 共p兲 and perpendicular 共s兲 polarization.
of the incident wave p in for ␪ in⫽35° and s polarization using
the same assumptions as before is shown in Fig. 4. Excellent
linearity is found up to 20 MPa. A 5% decrease in D occurs
at 37 MPa, and a 10% decrease at 53 MPa. Using different
total layer numbers N, the linear range of the system can, in
principle, be adjusted to the measurement task. It should be
noted that the absolute values of D depend on the assump-
tions made for the material parameters that are expected to
depend on the manufacturing process of the layers.
As mentioned above, all calculations were carried out
for plane optical waves within the layer system, i.e., for
single angles of light incidence. The results therefore are
directly applicable to the use of a collimated laser beam in
the optical detection scheme. However, to achieve high lat-
FIG. 4. The calculated change in reflectance ⌬R of the 19-layer system
in dependence on the acoustic positive peak pressure p in ; the range of
linearity.
1434 J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
eral resolution in the acoustic measurement, the laser needs
to be focused. In this case the reflectance slope function ob-
tained for a certain range of angles in the vicinity of ␪ in and
constant operational wavelength has to be weighted with the
angle-dependent profile of the light intensity and integrated
over this range of angles. The effect of this integration is an
increase in R and a decrease in D with increasing divergence,
which becomes significant at large ␪ in and high optical fi-
nesse, in particular. In the practical realization the maximum
reflectance slope achievable using large layer numbers N is,
therefore, restricted by the oblique and focused illumination
cone used.
III. EXPERIMENTAL CHARACTERIZATION
Experiments were carried out to investigate the acoustic
performance of the optical hydrophone. The multilayer sys-
tem manufactured was made of N⫽19 layers with alternat-
ing refractive indices of n⫽2.3 (Nb2 O5 ) and n
⫽1.48 (SiO2 ). The sputtering technique was employed to
obtain a robust coating that can be used in water without
swelling of the layers. The sensor element has an overall
thickness of d total⬇1.9 ␮ m. The design wavelength as mea-
sured by a grating spectrometer was ⬃675 nm, and a HeNe
laser of 633 nm wavelength and ⬃2 mW output power was
used in the optical detection setup. So, the angle of light
incidence was adjusted to ␪ in⬇35° 关cf. Fig. 3共b兲兴. Perpen-
dicular polarization of light was used and the focal length of
the lens was f 1 ⫽50 mm 共cf. Fig. 1兲. The measurement time
is restricted to pulse or tone burst lengths up to ⬃2.2 ␮s,
because after this transit time acoustic reflections from the
rear side of the sensor substrate 共thickness: 6.5 mm兲 may
cause distortions. For longer measurement times a thicker
substrate, a substrate with stronger acoustic absorption, or an
additional, appropriately mounted absorber can be used.
A. Lateral resolution
The lateral resolution capability of the sensor system is
mainly determined by the optical spot size of the focused
light beam within the multilayer system. For an estimation of
the spot dimension, a knife edge was fixed to the layer sys-
tem, which was then moved in the horizontal direction per-
pendicular to the edge. A successive evaluation of the part of
light transmitted in dependence on the edge position in the x
and y directions showed a Gaussian intensity profile with full
widths at 1/e of the maximum of ⬃60 ␮m in the x direction
and ⬃45 ␮m in the y direction. The increased width in the x
direction is first due to the oblique incidence of light, result-
ing in an elliptical cross section of the illuminating spot. A
further enlargement of the width in the x direction is then
brought about by the multi-interference process in the layer
system.
It should be noted that the lateral distribution of the
acousto-optic transduction efficiency within the focused op-
tical spot depends not only on the Gaussian light intensity
distribution but also on the optical angular spectrum, as de-
scribed in Sec. II, and can be quite inhomogeneous in the x
direction, in particular. Furthermore, in applications of ultra-
sound imaging using the sensor as a two-dimensional detec-
tion array, the lateral resolution may also be restricted by
Volker Wilkens: Optical multilayer hydrophone

FIG. 5. The frequency response of the optical multilayer hydrophone
as determined by the primary interferometric calibration; measurement
repeated after one year; linear regression; uncertainty bars: standard un-
certainties.
surface acoustic waves 共cf. Sec. III D兲, a mechanical cross-
talk phenomenon, which is known, for instance, from capaci-
tive micromachined ultrasonic transducer arrays.
B. Frequency response
A technique for primary calibration of hydrophones
based on a Michelson interferometer was applied to experi-
mentally determine the frequency response of the optical
multilayer hydrophone. As a detailed description of the cali-
bration setup and the evaluation procedure is given in Ref.
11, a brief description will be sufficient here. A stabilized
interferometer detected the displacement of a pellicle 共thick-
ness: 2 ␮m兲 mounted on the surface of the water contained in
a tank, while ultrasound bursts from a focusing PVDF source
transducer 共diameter: 5 mm, focal length: 12.5 mm, center
frequency: 50 MHz兲 were incident from the bottom. The
pressure amplitudes p in were calculated from the displace-
ment amplitudes measured in the focal plane for discrete
frequencies in the range from 1 to 75 MHz. In a second
course, the known acoustic field at each frequency was mea-
sured using the optical multilayer hydrophone and the same
excitation conditions of the source transducer as in the inter-
ferometric measurements. The transfer function M ( f )
⫽U/p in was then calculated by a comparison of the signal
voltage amplitudes U with the reference measurement re-
sults. A data correction was carried out to compensate the
acoustic transmission characteristic of the foil and the fre-
quency response of the photodetector in the interferometric
reference setup. In addition, a correction was performed to
compensate the frequency response of the photodetector 共in-
cluding amplifier兲 of the multilayer hydrophone determined
in a separate optical mixing measurement.11 Since the optical
spot size of the two detection systems was approximately the
same, corrections for the influence of spatial averaging of the
focused sound field could be omitted.
The pressure–voltage transfer function obtained is
shown in Fig. 5. The measurement uncertainties 共66% con-
fidence level兲 depicted for several frequencies are derived
from the uncertainties of the voltage measurements 共includ-
ing the signal-to-noise ratio兲 and sensor positioning, from an
estimation of sound field inhomogeneities, and from the un-
certainties of the different correction procedures referred to
J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
above.11 Within the uncertainty of the calibration, a constant
transfer function of M ( f )⫽9.0 mV/MPa is obtained for the
whole frequency range investigated. Due to the small thick-
ness of the sensing element (d total⬇1.9 ␮ m), no thickness
resonances occur 共subresonant mode兲 that are known to
bring about a strong variation of the transfer function, for
instance, in the case of membrane hydrophones with a typi-
cal foil thickness of 25 ␮m in the frequency range of about
40 MHz.11 Thickness mode resonance distortions are also
known from other optical ultrasound detectors like low-
finesse polymer film Fabry–Perot-type12,13 and high-finesse
etalon-type sensors.6 Although the acoustic frequency re-
sponse has not been investigated as a function of the etalon
thickness in Ref. 6, resonances can be expected according to
the acoustic sensitivity of both the front and the rear face of
the sensor element, which can be observed in the impulse
measurements performed. Since the diameter of the optical
multilayer hydrophone sensor probe is much larger than that
of the sound field, edge diffraction, lateral waves, and radial
resonances within the sensor element do not arise, whereas
these effects are known to cause distortions of the frequency
response of small probe hydrophones like, for instance,
needle-type hydrophones14,15 and, in particular, fiber-optic
multilayer hydrophones.3
It is appropriate to compare the experimentally obtained
transfer factor with the theoretical result according to the
analysis in Sec. II. The maximum reflectance slope at ␪ in
⫽35° for s polarization was determined to be D⫽3.87
⫻10⫺3 /MPa 关cf. Fig. 3共a兲兴. The diameter of the Gaussian
profile laser beam incident on the focusing lens ( f
⫽50 mm), measured by a procedure similar to that described
in Sec. III A, was d⫽550 ␮ m 共full width at 1/e of maxi-
mum兲. Numerical integration of the reflectance function and
the reflectance slope function, respectively, weighted with
this geometrical illumination cone incident at ␪ in⫽35° yields
R cone⫽1.06⫻R 0 and D cone⫽0.72⫻D. Using D cone
⫽⌬R cone /p in together with ⌬R cone /R cone⫽2.8⫻U/U 0 ,
where U 0 ⫽2.35 V is the direct current 共dc兲 voltage of the
photodetector corresponding to the reflectance R cone at the
operating point and 2.8 the ratio of electrical amplification of
the dc output to the alternating current 共ac兲 output of the
photodetector, the theoretical pressure–voltage transfer fac-
tor is given by
U UD cone D coneU 0
M th⫽ ⫽ ⫽ ⫽8.8 mV/MPa, 共9兲
p in ⌬R cone 2.8R cone
which is in good agreement with the experimental value.
Note that the absolute values given in Fig. 5 on the
ordinate depend on the amplification chosen for the photo-
current and in this sense are arbitrary and cannot be com-
pared with values known from PVDF hydrophones to esti-
mate the sensitivity. This comparison is dealt with in the
following paragraph.
C. Sensitivity
The sensitivity of the optical multilayer hydrophone sys-
tem was determined by measurement of the noise voltage
U noise of the photodetector output. From the noise voltage, as
Volker Wilkens: Optical multilayer hydrophone 1435
TABLE I. Sensitivity in terms of the noise-equivalent peak pressure values
measured for the present optical multilayer hydrophone system, a PVDF
membrane hydrophone/amplifier combination, and a small diameter needle-
type PVDF hydrophone.
Noise-equivalent peak pressure p min
共25 MHz detection bandwidth;
Measurement system without averaging兲
Optical multilayer hydrophone 30 kPa
0.5 mm diameter PVDF 0.6 kPa
membrane hydrophone
0.075 mm diameter PVDF 55 kPa
needle-type hydrophone 共Ref. 13兲
measured by an oscilloscope, the noise-equivalent pressure
was obtained using the transfer factor from the experimental
calibration: p min⫽Unoise /M . The noise-equivalent acoustic
pressure of the multilayer hydrophone for an analog detec-
tion bandwidth of the oscilloscope of ⬃25 MHz 共sampling
rate 250 MS/s兲 without signal averaging is given in Table I,
in comparison with the corresponding value of a typical 0.5
mm diameter PVDF membrane hydrophone/amplifier sys-
tem. In addition, the value taken from Ref. 13 for a 75 ␮m
diameter PVDF needle-type hydrophone obtained in a simi-
lar way is listed. It should be noted that the noise sources of
the different measurement systems are of a different kind.
While the noise during the PVDF hydrophones measurement
is dominated by that from the amplifier used 共separately or
within the oscilloscope兲, the noise during the optical mea-
surement is mainly made up of light intensity noise from the
laser source and shot noise from the optical detection pro-
cess. The results show that on the one hand the sensitivity of
the present optical multilayer hydrophone is significantly
lower 共1.5 orders of magnitude兲 than that of the membrane
hydrophone that is 0.5 mm in diameter. On the other hand,
the PVDF needle-type hydrophone with a sensor diameter
comparable to that of the optical hydrophone shows a sensi-
tivity smaller by approximately a factor of 2. In general, the
sensitivity of smaller diameter PVDF hydrophones is lower
due to the lower capacitance. In principle, this trade-off be-
tween sensitivity and lateral resolution can be circumvented
when using the optical multilayer hydrophone. Here the sen-
sitivity depends on the reflectance slope of the layer system
and on the power and noise characteristics of the light source
but not on the size of the sensing element. This applies simi-
larly to sensitivity versus frequency response: To achieve a
large frequency range with a flat transfer function, it is nec-
essary to use small foil thicknesses for the PVDF hydro-
phones 共subresonant mode兲, which, in turn, produce low sen-
sitivity. This second trade-off can also be circumvented when
using the optical hydrophone that inherently uses a very
small overall thickness, even for large layer numbers.
D. Directional response
To determine the directional response of the multilayer
hydrophone a focusing source transducer 共diameter: 6 mm;
focal length: 20 mm兲 was turned about the measurement
point at the water surface, since the present detection system
does not allow the detector to be turned. For three different
1436 J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
FIG. 6. Directional response of the optical multilayer hydrophone measured
perpendicular (yz plane, ␣兲, 共a兲 and parallel (xz plane, ␤兲, 共b兲 to the plane
of light incidence 共cf. Fig. 1兲 for different ultrasound frequencies; U: voltage
amplitude, U 0 : voltage amplitude at normal incidence.
frequencies 共10, 15, 25 MHz兲 tone bursts were measured in
angular steps of 2° and the voltage amplitudes were related
to the amplitude at normal sound incidence. Measurements
were successively made for turning in the xz plane and in the
yz plane 共cf. Fig. 1兲. The results obtained are shown in Fig.
6. Obviously, the directional response does not significantly
depend on the ultrasound frequency in the frequency and
angular ranges investigated. This is in contrast to the direc-
tional response known, for instance, from membrane hydro-
phones, which is, with the exception of the impact of Lamb
waves at large angles 共⬃55°兲 and low frequencies, well de-
scribed by the diffraction theory of a plane rigid circular disk
receiver.16 The directional response of the optical multilayer
hydrophone is not dominated by diffraction but likely by the
impact of Rayleigh waves excited on the sensor surface. In
the range from 20° to 30°, the excitation is most effective
because here the propagation constant components parallel to
the sensor surface of the incident wave and of the surface
wave are matched, which results in a strong increase of the
response in this range. The asymmetric profile found for the
xz plane is a result of the asymmetric variation of the reflec-
tance slope in dependence on the angle of illumination for
the optical cone at oblique incidence 共cf. Sec. II兲. In contrast,
the variation of the reflectance slope within the cone of illu-
mination in the perpendicular direction is very small and
symmetric to the z axis, leading to a symmetric directional
response in the yz plane.
E. Stability
The short-term stability was investigated by repeated
measurements of ultrasound pulses from a focusing source
transducer 共center frequency: 8 MHz, diameter 6 mm, focal
length: 15 mm兲 driven by a pulse generator 共peak voltage:
350 V兲. After 60 min of warming up for the laser source, the
sensor, which was at the same temperature as the water, was
positioned at the water surface and the peak voltage of a
Volker Wilkens: Optical multilayer hydrophone
20-times-averaged waveform was immediately acquired. The
measurement was randomly repeated over several hours. The
maximum relative deviation of the peak voltage was 2% and
a systematic tendency with time could not be identified. The
deviation of the laser power simultaneously measured was in
the range of 1%. With respect to other statistical deviations
caused, for instance, by the source transducer and the pulse
generator used, better stability could not be expected. Sig-
nificant swelling of the layers that would have changed the
optical reflectance was not observed.
The long-term stability of the optical multilayer hydro-
phone was confirmed when the interferometric calibration
was repeated after one year, and so far no significant devia-
tion due to degradation of the sensor could be observed
共Fig. 5兲.
IV. CONCLUSIONS
An optical multilayer hydrophone for broadband ultra-
sound detection was theoretically described and experimen-
tally characterized. The acoustic measurement is based on
the elastic deformation of the very thin optical layer system
deposited onto a glass substrate by an incident pressure wave
and the detection of the induced change in optical reflec-
tance. Using an oblique incidence of light to detect the re-
flectance variation yields good sensitivity and a very simple
and low-cost setup. For the theoretical description some sim-
plifying assumptions and material parameter estimations
were made to study the sensor performance and the opportu-
nity to adjust the operational optical wavelength of a given
layer system by variation of the angle of light incidence.
From this several design criteria for the detection arrange-
ment could be derived.
A concrete optical multilayer hydrophone arrangement
was experimentally characterized. The lateral resolution, the
frequency response, the sensitivity, the directional response,
and the stability were determined. Because of the very small
sensor element thickness and the absence of edge and radial
effects, the frequency response obtained is constant in a re-
markably broad frequency range, which was confirmed for
the range from 1 to 75 MHz. A comparison of the experi-
mental pressure–voltage transfer factor with the theoretical
estimation showed very good agreement. The directional re-
sponse is dominated by the impact of Rayleigh waves that
are excited when the sound wave obliquely strikes the sensor
surface. The sensitivity of the present system is comparable
to that of available PVDF hydrophones with similar lateral
dimensions of the sensor element.
The results, in particular, the constant response in a wide
frequency range, show that the optical multilayer hydro-
phone meets very well the demands placed on a reference
measurement system for secondary calibrations. By measure-
ment of single broadband ultrasound pulses 共wave fields rich
in harmonics兲 with the optical hydrophone and in a second
step with a measurement system to be calibrated, the fre-
quency response can be determined by a comparison of the
relevant frequency spectra of the measured pulses. The
complex-valued frequency response is obtained with high-
frequency resolution in a very fast and efficient way and can
be used for correct data evaluation and impulse deconvolu-
J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
tion. This technique has already been successfully applied to
the calibration of fiber-optic multilayer hydrophones using a
focused source transducer with a center frequency of 8 MHz,
and evaluation up to the fifth harmonic provided the fre-
quency response up to 40 MHz.5,17 The technique may be
extended also to the calibration of PVDF hydrophones using
broadband pulses.18 –21 Furthermore, a direct comparison
with the constant frequency response of the optical
multilayer hydrophone would eliminate the need for ex-
tended sound field calculations, including idealized assump-
tions for the source transducer, in broadband calibration tech-
niques without a reference receiver using nonlinear sound
propagation.22 The optical multilayer hydrophone may also
be regarded as a possible independent method to confirm the
existing interferometric calibration techniques, in particular,
as regards their extension to the high-frequency range re-
cently developed.11,23 From the metrological perspective, the
comparison of piezoelectric hydrophones with the acoustic
pressure detecting optical hydrophone is more advantageous
than a comparison with interferometers detecting particle
displacement, since for a given acoustic pressure, the particle
displacement for a plane wave is reciprocally related to the
frequency, which imposes more severe requirements for an
acceptable signal-to-noise ratio at high frequencies.23
Another application field for the detection system pre-
sented may lie in high-frequency ultrasound imaging, where
detection arrays with a small element size are called for.2,24,25
A promising concept in this field of nondestructive testing is
the application of broadband optical ultrasound detection
systems in combination with laser-induced photoacoustic
pulse generation.7,26,27
ACKNOWLEDGMENT
The author wishes to thank Walter Molkenstruck for
support with regard to the interferometric calibration.
1
V. Wilkens and C. Koch, ‘‘Fiber-optic multilayer hydrophone for ultra-
sonic measurement,’’ Ultrasonics 37, 45– 49 共1999兲.
2
V. Wilkens and C. Koch, ‘‘Optical multilayer detection array for fast ul-
trasonic field mapping,’’ Opt. Lett. 24, 1026 –1028 共1999兲.
3
W. Weise, V. Wilkens, and C. Koch, ‘‘Frequency response of fiber-optic
multilayer hydrophones: experimental investigation and finite element
simulation,’’ IEEE Trans. Ultrason. Ferroelectr. Freq. Control 49, 937–946
共2002兲.
4
V. Wilkens and C. Koch, ‘‘Optische Vielschichtsensoren: eine Alternative
zu piezoelektrischen Ultraschallhydrophonen,’’ Fortschritte der Akustik—
DAGA 2000, Proceedings of the 26th German Annual Conference on
Acoustics, Oldenburg, Germany, ISBN 3-9804568-8-9, 680-681, 2000 共in
German兲.
5
V. Wilkens, ‘‘Dielektrische optische Interferenzschichthydrophone zur
zeitlich und räumlich hochauflösenden Messung von Ultraschallfeldern,’’
Ph.D. dissertation, University of Oldenburg, Germany, 2001, and PTB
Report No. PTB-MA-67, 2001, ISBN 3-89701-750-4 共in German兲.
6
J. D. Hamilton, T. Buma, M. Spisar, and M. O’Donnell, ‘‘High frequency
optoacoustic arrays using etalon detection,’’ IEEE Trans. Ultrason. Ferro-
electr. Freq. Control 47, 160–169 共2000兲.
7
P. C. Beard, ‘‘Photoacoustic imaging of blood vessel equivalent phan-
toms,’’ Proceedings of the SPIE, BIOS 02, San Jose, CA, 2002, Vol. 4618,
pp. 54 – 62.
8
H. A. Macleod, Thin-Film Optical Filters 共Institute of Physics, Bristol,
2001兲, Chap. 2, pp. 37– 45.
9
G. B. Hocker, ‘‘Fiber-optic sensing of pressure and temperature,’’ Appl.
Opt. 18, 1445–1448 共1979兲.
10
J. Staudenraus and W. Eisenmenger, ‘‘Fibre-optic probe hydrophone for
Volker Wilkens: Optical multilayer hydrophone 1437

ultrasonic and shock-wave measurements in water,’’ Ultrasonics 31, 267–
273 共1993兲.
11
C. Koch and W. Molkenstruck, ‘‘Primary calibration of hydrophones with
extended frequency range 1 to 70 MHz using optical interferometry,’’
IEEE Trans. Ultrason. Ferroelectr. Freq. Control 46, 1303–1314 共1999兲.
12
P. C. Beard, F. Pérennès, and T. N. Mills, ‘‘Transduction mechanisms of
the Fabry–Perot polymer film sensing concept for wideband ultrasound
detection,’’ IEEE Trans. Ultrason. Ferroelectr. Freq. Control 46, 1575–
1582 共1999兲.
13
P. C. Beard, A. M. Hurrell, and T. N. Mills, ‘‘Characterization of a poly-
mer film optical fiber hydrophone for use in the range 1 to 20 MHz: a
comparison with PVDF needle and membrane hydrophones,’’ IEEE Trans.
Ultrason. Ferroelectr. Freq. Control 47, 256 –264 共2000兲.
14
B. Fay, G. Ludwig, C. Lankjaer, and P. A. Lewin, ‘‘Frequency response of
PVDF needle-type hydrophones,’’ Ultrasound Med. Biol. 20, 361–366
共1994兲.
15
Y. Nakamura and T. Otani, ‘‘Study of surface elastic wave induced on
backing material and diffracted field of a piezoelectric polymer film hy-
drophone,’’ J. Acoust. Soc. Am. 94, 1191–1199 共1993兲.
16
S. Robinson, R. Preston, M. Smith, and C. Millar, ‘‘PVDF reference hy-
drophone development in the UK—from fabrication and lamination to use
as secondary standards,’’ IEEE Trans. Ultrason. Ferroelectr. Freq. Control
47, 1336 –1344 共2000兲.
17
V. Wilkens, C. Koch, and W. Molkenstruck, ‘‘Frequency response of a
fiber-optic dielectric multilayer hydrophone,’’ Proceedings of the 2000
IEEE Ultrasonics Symposium, San Juan, Puerto Rico, 2000, pp. 1113–
1116.
18
R. Reibold and W. Molkenstruck, ‘‘Investigation of pulse-excited hydro-
phones for ultrasonic field measurements using laser interferometry,’’ Ul-
trasonics 25, 114 –118 共1987兲.
1438 J. Acoust. Soc. Am., Vol. 113, No. 3, March 2003
19
H. Djelouah, J. C. Baboux, and M. Perdrix, ‘‘Pulsed calibration technique
of miniature ultrasonic receivers using a wideband laser interferometer,’’
Ultrasonics 27, 80– 85 共1989兲.
20
G. R. Harris and P. M. Gammell, ‘‘Sensitivity measurements of piezoelec-
tric polymer hydrophones from 0.2 to 2 MHz using a broadband pulse
technique,’’ J. Acoust. Soc. Am. 105, 725–731 共1999兲.
21
R. A. Smith and D. R. Bacon, ‘‘A multiple-frequency hydrophone calibra-
tion technique,’’ J. Acoust. Soc. Am. 87, 2231–2243 共1990兲.
22
H. J. Bleeker and P. A. Lewin, ‘‘A novel method for determining calibra-
tion and behaviour of PVDF ultrasonic hydrophone probes in the fre-
quency range up to 100 MHz,’’ IEEE Trans. Ultrason. Ferroelectr. Freq.
Control 47, 1354 –1362 共2000兲.
23
T. J. Esward and S. P. Robinson, ‘‘Extending the frequency range of the
NPL primary standard laser interferometer for hydrophone calibration to
60 MHz,’’ IEEE Trans. Ultrason. Ferroelectr. Freq. Control 46, 737–744
共1999兲.
24
J. D. Hamilton and M. O’Donnell, ‘‘High frequency ultrasound imaging
with optical arrays,’’ IEEE Trans. Ultrason. Ferroelectr. Freq. Control 45,
216 –235 共1998兲.
25
P. C. Beard and T. N. Mills, ‘‘A 2D optical ultrasound array using a
polymer film sensing interferometer,’’ in Ref. 17, pp. 1183–1186.
26
G. Paltauf and H. Schmidt-Kloiber, ‘‘Measurement of laser-induced
acoustic waves with a calibrated optical transducer,’’ J. Appl. Phys. 82,
1525–1531 共1997兲.
27
K. P. Köstli, M. Frenz, H. P. Weber, G. Paltauf, and H. Schmidt-Kloiber,
‘‘Optoacoustic tomography: time-gated measurement of pressure distribu-
tions and image reconstruction,’’ Appl. Opt. 40, 3800–3809 共2001兲.
Volker Wilkens: Optical multilayer hydrophone