Research Network FWF S105
Photoacoustic Imaging in
Medicine and Biology
http://pai.uibk.ac.at
Comparison of optical and
piezoelectric integrating line
detectors
R. Nuster, S. Gratt,
K. Passler, H. Grün, T. Berer,
P. Burgholzer, and G. Paltauf.
February 2009
PAI Report No. 9
Comparison of optical and piezoelectric integrating line detectors
R. Nuster*a
, S. Gratta
, K. Passlera
, H. Grünb
, Th. Bererb
, P. Burgholzerb
, and G. Paltaufa
a
Institute of Physics, Karl-Franzens-University, Graz, AUSTRIA
b
Upper Austrian Research, Linz, AUSTRIA
ABSTRACT
Currently two different types of integrating line sensors are used in photoacoustic tomography (PAT). Thin film
piezoelectric polymer sensors (PVDF) are characterized by compactness, easy handling and the possibility to
manufacture sensing areas with different shape. However, they are vulnerable to electrical disturbance and to scattered
light from the illuminated sample. Also optical sensors are used as integrating line sensors in combination with some
kind of interferometric setup. For example, one arm of a Mach-Zehnder interferometer or the cavity of a Fabry-Perot
interferometer can be used as line detector. In both cases, the light wave either propagates freely in the liquid or is guided
in an optical fiber. Such sensors are quite immune against noise sources described above and suitable for high bandwidth
detection. One drawback is the limited mobility due to the complex arrangement of the setup.
This study is focused on the comparison of the different implementations of line detectors, mainly on directivity and
sensitivity. Shape and amplitude of signals generated by defined sources are compared among the various sensor types.
While the shape of the signals recorded with the optical free beam detector matches quite well to the simulation the
signals detected with the PVDF detector are affected by directivity effects. This causes a strong distortion of the signal
shape depending on the incident angle of the acoustic wave. How these effects influence the reconstructed projection
image is discussed.
Keywords: photoacoustic, optical detectors
1. INTRODUCTION
In photoacoustic tomography (PAT) the aim is to reconstruct the initial energy density from acoustic signals measured
outside the sample at certain detector positions1-4
. The excitation process is based on the thermoelastic effect, which
induces the emission of an acoustic wave after illuminating an object with a pulse of electro-magnetic radiation. For
pulses in the visible or near infrared range the illumination source is a nanosecond laser. Energy deposition and
thermoelastic pressure generation occurs preferentially in structures with enhanced light absorption. Therefore the
contrast of the image is primarily optical. The acoustic waves propagate from their initial source towards the surface to
the sensors arranged surround the sample. In summary, PAT is an imaging method for optically diffuse biological
samples, combining the advantages of optical contrast with the resolution of ultrasound imaging techniques.
To detect the propagating photoacoustic signals, the common approach is to use small aperture detectors. Various
reconstruction algorithms have been developed for point-detectors and different recording surfaces3-6
. In order to
overcome the effect of finite dimension of the sensor on the image resolution, another approach is to use large aperture
sensors, which in at least one dimension exceed by more than a factor of two the size of the imaged object7-9
. The signal
measured with an extended sensor, also called integrating sensor, is given by an integral of pressure field over the
detector area. With respect to the initial pressure, the instantaneous signal generated in any detector is an integral over an
area that is determined by the shape of the sensor. For example, a single point sensor measures signals that are given by
an integral of the initial pressure over a spherical surface, whereas a line sensor measures signals that are determined by
the integrals of the initial thermoelastic pressure over cylinders with the line detector in the central axis. Due to this
integration the signals measured with line detectors can be thought to originate from a linear projection of the initial
pressure distribution along the direction of the detector line rather than from the original distribution. From a scan of the
line detector along the surface of the object this linear projection can be reconstructed. Finally, rotation of the detector
*
direction relative to the object yields data from which a complete three-dimensional image can be reconstructed by
applying the two-dimensional inverse Radon transform.
Integrating line detectors can be implemented using either piezoelectric, optical free beam or guided beam detection
methods (Fig.1). All of them are already successfully used to record acoustic signals in PAT8,9
. However, it is essential
to compare the sensitivity, the signal fidelity and the directivity of the different line detectors which is the aim of this
study.
Fig. 1. Comparison of piezoelectric, free bream and fiber based detection
2. DETECTION SETUPS
To compare experimentally the sensitivity of the different detectors we used photoacoustic generation of broadband
acoustic pulses. The detection setups can be separated into excitation and detection parts. The former is identical for all
detection methods. To excite the acoustic transients we used 10 ns laser pulses with variable wavelength from an optical
parametrical oscillator (OPO) pumped by a frequency tripled Nd-YAG laser system. The pulses were coupled into an
optical fiber with a diameter of 0.6 mm. The fiber tip was arranged in an adjustable fiber holder combined with a lens to
create a seven times enlarged image of the fiber end-face onto the surface of an absorbing liquid (Orange-G solution)
which was filled inside a plastic cuvette arranged in a water tank. Front and back wall of the cuvette were made of
transparent 100 µm thick plastic foils, well acoustically matched to the surrounding liquid.
Each line detector was arranged parallel to the illumination plane behind the cuvette at a normal distance of 9 mm from
the position where the excitation takes place (transmission detection method). The orientation was optimized during the
measurement for obtaining the highest signal amplitude.
2.1 Piezoelectric detector
Thin film piezoelectric polymer sensors made of PVDF (polyvinylidene fluoride) are characterized by compactness, easy
handling and the possibility to manufacture sensing areas with almost arbitrary shape. Unlike the optical free beam
detection system shown below, the single PVDF sensor can be easily moved itself around a fixed object. Furthermore, an
array detection system can be realized to reduce the signal acquisition time for the application in PAT. However,
piezoelectric sensors are vulnerable to electromagnetic disturbance and generate pyroelectric signals when hit by laser
pulses. Therefore the application of PVDF sensors in PAT is problematic due to scattered light from the illuminated
sample. A further disadvantage is their opaqueness. This is impractical in the backward detection mode where it is
important to keep the detection distance small.
The setup shown in Fig.2a consists of a homemade piezoelectric line detector in combination with a unit gain active
probe. A 28µm thick PVDF film coated on the upper side with a conducting layer was directly glued on a 200µm wide
and 30mm long copper electrode which was embedded in a plastic block. The size of the electrode determines the active
area of the sensor (Fig. 2c).
2.2 Optical free beam detector
As shown in Fig. 2b one arm of a Mach-Zehnder interferometer can be used as nearly ideal line detector. A sufficient
spatial resolution was obtained by first expanding the laser beam and then focusing it at the region of interest. Thereby a
diameter of 40µm at the beam waist was achieved. The acoustic wave induces a change of refractive index followed by a
phase shift between the two arms crossing the water tank. Thereby an intensity modulation at the outputs of the
interferometer can be detected.
The free beam detection setup contains a HeNe-Laser (632.8 nm) as coherent light source, two beam splitters to separate
and recombine the optical rays, a mirror with integrated piezoelectric actuator (M+PT) and a pair of balanced photo
diodes (PD) connected to a differential amplifier with a bandwidth of 80 MHz. Residual diffuse light from the optical
excitation of sound waves was rejected using a band pass filter (BPF) in front of the photo detector. An active
stabilization circuit using the differential amplifier output was connected to the piezoelectric actuator and adjusted the
path length difference in the interferometer at the value for maximum sensitivity (PI-control). Furthermore, a high pass
filter (HPF) was used to remove the nonrelevant content of the spectra below the cut-off frequency (fc= 100kHz).
2.3 Fiber based detectors
Confining the propagating light field in a waveguide yields a compact moveable detector with constant spatial resolution.
The fiber based Mach-Zehnder interferometer (MZ) was investigated using polymer as well as glass fibers (Fig.3). The
Fabry-Perot interferometer (FPI) was realized using a single mode fiber made of glass with two integrated fiber Bragg
gratings (FBG) arranged at a distance of 11.5cm from each other. The finesse of the FPI was 17 using FBGs with 81%
reflectivity using 1550nm wavelength. To obtain a straight line the sensing part of the fibers was fixed on a moveable
fiber holder. The operating mode is similar to the free beam detection system except that the change in refractive index
occurs now in the fiber core rather than in water, giving rise to a phase change and thus a modulation of the light
intensity.
As shown in Fig.3 the light intensity at the output of both interferometric setups was measured at the end of the fibers
using the same custom made photo detector (FEMTO HCA-S) including a built-in measuring amplifier. A feed-back
loop using a micro controller stabilizes the operating point of the interferometer. Instead of changing the distance of the
mirrors as in a classical free beam interferometer, the wavelength of the detection laser was changed. For this purpose we
used a wavelength tuneable laser (Koheras Adjustic E15).
(a) (c)
(b) (d)
Fig. 2. Setup using a piezoelectric detector (a) and a free beam MZ-Interferometer (b); cross-sections showing the active
area of both detectors (c), (d)
(a) (c)
(b) (d)
Fig. 3. Setups using a fiber based optical detection: MZ-Interferometer (a), FPI-Interferometer (b), cross-section showing
the active area of the polymer fiber (c) and the glass fiber (d)
3. SENSITIVITY ESTIMATION
To compare the sensitivity of the line detectors the noise equivalent pressure length products (NEPLPs) were estimated.
Using the signal to noise (S/N) ratios of the N-times averaged low pass filtered signals (fc=50MHz) shown in Fig.4b and
the simulated amplitude of the pressure length product ( )max
lp ⋅ at the detection place.
( )
NS
Nlp
NEPLP
/
max⋅⋅
=
The absolute value of the pressure length product was obtained by simulating the acoustic wave propagation for the
given values of radiant exposure (2.9mJ/cm²) at the surface of the absorbing dye solution with the known absorption
coefficient (18mm-1
). The resulting acoustic wave was then integrated along the line at the detection place. Fig. 4a
shows the simulated temporal signal for a temporal Gaussian shape of the excitation laser pulse.
The comparison of the signals in Fig. 4b labeled with the corresponding NEPLP values shows that the sensitivity which
can be achieved with a free beam Mach-Zehnder interferometer (MZ-FB) is about 8 times higher compared to the fiber
based Mach-Zehnder interferometer using polymer fibers (MZ-Pol.) and 180 times higher compared to the glass fiber
device (MZ-Glass). Using the fiber based Fabry-Perot interferometer made of glass (FPI-Glass) the sensitivity can be
increased compared to the MZ-glass sensor, because of the multiple round trips of the light field in the resonator where it
interacts with the acoustic field. However, it is still 18 times smaller compared to the free beam interferometer. The
differences in sensitivity between the optical detection systems are mainly caused by the sensor specific elasto-optic
coupling coefficients dpdn , the mismatch of the acoustic impedanceWDA
ZZ (Fig. 2-3) and the used wavelength of the
detection laser.
The sensitivity of the PVDF-detector is about 27 times smaller compared to the free beam Mach-Zehnder interferometer.
However, it should be possible to improve it using a specially optimized amplifier instead of a unit gain active probe.
(a) (b)
Fig. 4. (a) Simulated absolute pressure length product signal, (b) averaged temporal signals measured with the different
detection setups (Fig. 2-3) labeled with the obtained noise equivalent pressure length product values (NEPLP)
4. DIRECTIVITY OF THE PVDF LINE DETECTOR
As sample we used a black sphere embedded in gelatine. Due to the high absorption and the two sided illumination from
directions parallel to the line orientation the acoustic source can be assumed as a disk with omni-directional emission in
the projection plane. The directivity was investigated by recording 200-times averaged temporal signals while moving
the source linearly in z-direction. As shown in Fig.5, this varied the angle of incidence θ. The arrangement of the sample
and the detectors in line orientation is schematically shown in Fig. 5. While the free beam detector shows an omni-
directional response because of its radially symmetrical cross-section the shape of the signals obtained with the PVDF-
detector are affected by directivity effects (Fig. 6b). The strong distortion of the signal shape with increasing incident
angle of the acoustic wave can be explained by the flat extended geometry of the detection area, the piezoelectric
response of the PVDF film and the acoustic properties of the different materials (Fig.2c). For normal incidence (θ=0) of
the acoustic wave the signals of both detectors match quite well to the simulated signal assuming an ideal line detector
(Fig. 6a).
Fig. 5. Arrangement of sample and detector to investigate the directivity
(a) (b)
Fig. 6. (a) Comparison of the measured signals with the simulated signal for normal incident of the acoustic wave. (b)
Comparison of signals obtained with the optical free beam and the PVDF-detector dependent on angle of incidence.
To answer the question how the directivity influences the quality of the projection images we recorded the temporal
signals for a complete z-scan over a range of 5cm with a step size of 50µm and reconstructed the projection images using
a two-dimensional frequency domain reconstruction algorithm 3,10
. As sample once again a black sphere 2mm in diameter
embedded in gelatine was used. The raw data and the obtained projection images are shown in Fig. 7. It can be clearly
seen that both images are blurred in scan direction because of the limited aperture effect11
.
(a) (b)
Fig. 7. Raw data and projection image of a black sphere using the optical free beam (a), using the PVDF-detector (b).
Fig. 8. Vertical and horizontal intensity profiles across the projection images of the sphere in Fig.7.
However, to identify differences between the two images vertical and horizontal profiles were taken out of the projection
images crossing the center position of the sphere (Fig. 8). While the vertical extension of the slices matches to the
original size of the sphere the horizontal slices are blurred. This blurring effect is significantly more distinctive in the
projection image obtained from the dataset measured with the PVDF-detector. In summary the directivity causes an
additional limited aperture effect.
5. CONCLUSION
In conclusion the highest sensitivity (NEPLP=5.1mbar mm) was obtained with the integrating line detector realized with
the optical free beam Mach-Zehnder interferometer. However, the detector is fixed and can not be moved. In practice
because of flexibility either PVDF or fiber based detectors made of polymer will be preferred. The directivity of the
latter influences the quality of the reconstructed images. Therefore for image reconstruction our intention is to take the
directivity into account to improve the quality.
ACKNOWLEDGMENT
This work has been supported by the Austrian Science Fund, project No. S 10502-N20.
REFERENCES
1. Wang, X. D., Pang, Y. J., Ku, G., Xie, X. Y., Stoica, G., and Wang, L. H. V., "Noninvasive Laser-Induced
Photoacoustic Tomography for Structural and Functional in Vivo Imaging of the Brain," Nature Biotechnology
21, 803-806 (2003).
2. Kruger, R. A., Reinecke, D. R., and Kruger, G. A., "Thermoacoustic computed tomography," Medical Physics 26,
1832-1837 (1999).
3. Kostli, K. P. and Beard, P. C., "Two-Dimensional Photoacoustic Imaging by Use of Fourier-Transform Image
Reconstruction and a Detector With an Anisotropic Response," Applied Optics 42, 1899-1908 (2003).
4. Xu, M. H. and Wang, L. H. V., "Photoacoustic Imaging in Biomedicine," Review of Scientific Instruments 77,
(2006).
5. Xu, M. H. and Wang, L. V., "Analytic Explanation of Spatial Resolution Related to Bandwidth and Detector
Aperture Size in Thermoacoustic or Photoacoustic Reconstruction," Physical Review E 67, (2003).
6. Xu, M. and Wang, L. V., "Universal back-projection algorithm for photoacoustic computed tomography,"
Physical Review E 71, (2005).
7. Haltmeier, M., Scherzer, O., Burgholzer, P., and Paltauf, G., "Thermoacoustic computed tomography with large
planar receivers," Inverse Problems 20, 1663-1673 (2004).
8. Burgholzer, P., Hofer, C., Paltauf, G., Haltmeier, M., and Scherzer, O., "Thermoacoustic tomography with
integrating area and line detectors," IEEE Trans. Ultrason., Ferroelect., Freq. Contr. 52, 1577-1583 (2005).
9. Paltauf, G., Nuster, R., Haltmeier M., and Burgholzer P., "Photoacoustic tomography using a Mach-Zehnder
interferometer as an acoustic line detector," Applied Optics 46, 3352-3358 (2007).
10. Paltauf, G., Nuster, R., Burgholzer, P., and Haltmeier, M. Three-dimensional photoacoustic tomography using
acoustic line detectors. 6437, 64370N. 2007. Proc SPIE.
11. Paltauf, G., Nuster, R., Haltmeier, M., and Burgholzer, P., "Experimental evaluation of reconstruction algorithms
for limited view photoacoustic tomography with line detectors," Inverse Problems 81-94 (2007).
titlepage_comparison.pdfProcSPIE2009_RobNus.pdf