- Email: [email protected]
Transparent ITO coated PVDF transducer for optoacoustic depth profiling
Optics Communications 253 (2005) 401–406 www.elsevier.com/locate/optcom
Transparent ITO coated PVDF transducer for optoacoustic depth profiling J.J. Niederhauser a, M. Jaeger a, M. Hejazi
a,c
, H. Keppner b, M. Frenz
a,*
a
c
Institute of Applied Physics, University of Berne, Sidlerstrasse 5, CH-3012 Berne, Switzerland b Ecole dÕInge´nieurs de lÕArc Jurassien (EIAJ), Le Locle, Switzerland Medical Physics Department, Faculty of Medicine, Teheran University of medical Sciences, Iran Received 29 September 2004; received in revised form 22 April 2005; accepted 2 May 2005
Abstract In optoacoustic depth profiling, illumination of absorbing biological tissue with short laser pulses generates pressure transients that are recorded with an acoustic transducer. The time profile of these pressure transients mirrors the depth resolved light absorption in the tissue, a relevant parameter, for example for transcutaneous medication penetration or in diagnosis of portwine stain. We propose a novel piezoelectric polyvinylidene fluoride transducer with transparent indium tin oxide (ITO) electrodes. The possibility of laser illumination through the transducer allows backward mode optoacoustic detection, a prerequisite to all parts of the human body not accessible from two sides. The sensitivity of the ITO transducer is comparable to classical metal coated transducers. Its transparency allows direct placement on the skin and combined with homogeneous large area illumination a one-dimensional acoustic propagation model becomes valid. This eliminates the need for computational compensation of diffraction effects. The influence of illumination area was illustrated in a comparison of the two transducers with an absorbing methylene blue sample in forward and backward mode. Ó 2005 Elsevier B.V. All rights reserved. Keywords: Optoacoustic imaging; Pressure transducer; Laser-induced ultrasound; Backward mode; Stress analysis; Acoustic diffraction
1. Introduction Medical need for highly selective diagnostic methods has motivated research for optoacoustic *
Corresponding author. E-mail address: [email protected] (M. Frenz).
depth profiling which combines high optical contrast with deep acoustic penetration [1–13]. In optoacoustics absorbing structures, such as blood capillaries of port wine stains, hidden inside scattering media are illuminated with short laser pulses to generate acoustic transients by means of thermoelastic effect. Detected at the surface with an acoustic
0030-4018/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.optcom.2005.05.005
402
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
transducer, these signals provide quantitative spatially resolved information on optical properties, such as the absorption coefficient, in the tissue. For optoacoustics, most human body parts are not accessible from two sides (forward mode), therefore laser illumination and acoustic sensing must be performed on the same surface (backward mode). The commonly used piezoelectric transducers have metal electrodes, which make them opaque for the laser pulse and complicate homogeneous tissue illumination. To circumvent this drawback, many designs have been proposed using either a piezoelectric ring for through illumination [4,5,11], for guiding the light around the acoustic transducer [3,6,8,10,13] or recently to replace the metal coating by a layer of salt-water as electrodes [14]. The ring design has a small illumination area making it difficult to couple enough light into the tissue without violating the safety limits for laser exposure. The around illumination design requires large distances between the skin and the acoustic transducer leading to lower signal amplitudes and distortion due to acoustic attenuation and diffraction. In both cases acoustic diffraction urge the use of deconvolution algorithms, which complicates direct quantitative measurement. To overcome the opaqueness problem various optical detection methods have been proposed [1,2,7,9,12]. We propose a transparent piezoelectric transducer based on indium tin oxide (ITO) coated polyvinilidene fluoride (PVDF) which overcomes all the limitations mentioned above. ITO is a transparent conductive coating widely used for electrodes in liquid crystal displays (LCD). PVDF is a transparent piezoelectric material ideally suited for sensitive wideband hydrophones. The proposed pressure transducer features a high piezoelectric sensitivity, allows quantitative measurement in the acoustic near field directly on the skin and is transparent for the illuminating laser pulse.
2. Materials and methods As seen in Fig. 1, the experimental setup consisted of two piezoelectric polyvinylidene fluoride (PVDF) transducers, one coated with aluminum (Al) electrodes (a) and one coated with transparent
(a) Al forward mode
laser dye
Al coated PVDF piezo transducer
PMMA (b) ITO forward mode
laser dye PMMA
transparent ITO coated PVDF piezo transducer
(c) ITO backward mode
dye PMMA
transparent ITO coated PVDF piezo transducer
laser Fig. 1. Optoacoustic depth profiling setup. The transducers consisted of a piezoelectric PVDF film with either Al electrodes (a) or ITO electrodes (b,c). PMMA was used as a backing material. A 1 mm layer of absorbing dye was homogeneously illuminated with a short laserpulse. This generated an acoustic transient representing the light absorption, which was recorded with the transducer. In contrast to Al, ITO is transparent, which allowed illumination not only in forward mode (a,b) but also in backward mode (c). Each configuration was used with two circular laser illumination diameters of 6 and 9 mm resulting in a total of six experiments.
indium tin oxide (ITO) electrodes (b,c). ITO provides surface electrical conductivity, while still maintaining good transparency (80%) over a large wavelength range from 450 to 2000 nm [15]. Due to the high refraction index of ITO (n = 2.05 @ 550 nm) the main loss in transmission is caused by reflection at the interfaces, only about 2% of the incident radiation is absorbed in the thin ITO layer. For both pressure transducers, a 40-lm thin PVDF foil was either vapor deposited with 200 nm Al or sputtered with 200 nm ITO. A mask was used during the coating process to obtain a circular active area of 3-mm diameter. The foils were placed on a 2-cm thick plexiglass (PMMA) backing using glycerin for adhesion by capillary forces. For mechanical protection, a clingfilm was placed on top using another drop of glycerin for acoustical coupling. A differential amplifier probe connected the piezoelectric electrodes to a digital
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
oscilloscope for data acquisition at a sampling rate of 1 GHz and 15 times averaging. The acoustic transients were generated optoacoustically by illuminating an absorbing layer with short laser pulses. A frequency trebled Nd:YAG laser with an optical parametric oscillator (OPO) generated 5 ns pulses at freely tunable wavelengths of 400– 2000 nm. The pulses were coupled into a 600-lm fiber and projected onto a circular area of adjustable diameter using a focusing lens at the distal end of the fiber. This resulted in homogeneously circular illumination on the sample. For forward measurements the fiber was placed on the opposite side of the absorbing sample relative to the transducer (see Fig. 1(a) and (b)). The ITO transducer additionally allowed backward mode measurement, a requirement for many medical applications (see Fig. 1(c)). For backward measurements the fiber was placed behind the plexiglass backing on the same side as the transducer. Two sets of experiments were conducted: (i) Direct comparison of transducer sensitivity of the two coated PVDF foils in forward mode. As absorbing sample, a layer of 300lM methylene blue solution was placed on top of each transducer. The layer thickness was 5 mm to ensure sufficient absorption in forward mode and to prevent the transducer from heating by direct irradiation. For the sensitivity comparison a large 15-mm diameter spot with 1.8 mJ/cm2 radiant exposure was chosen to avoid acoustic distortion (near field detection). (ii) Depth profiling of an absorbing layered sample in forward and backward mode. For the depth profiling a 25-lm Al transducer was used to increase temporal resolution. For the ITO transducer this thickness reduction was not possible because of prototype manufacturing reasons. The absorbing layer consisted of a 300 lM solution of methylene blue in water filled into a 1-mm thick cuvette covered on both sides with a clingfilm. The cuvette was immersed in water at a distance of 2 mm to the transducer. The optical absorption coefficient of the methylene blue solution was spectroscopically determined
403
to be la = 1.65 mm 1 at the wavelength of k = 577 nm. For comparing the ITO to the aluminum transducer the three possible arrangements (Al forward, ITO forward, ITO backward), each with two illumination diameters 6 and 9 mm were investigated, resulting in six experiments. The exposure was 4.6 mJ/cm2 for the 6-mm experiments and 3.9 mJ/cm2 for the 9-mm experiments. The thermal background caused by direct light absorption and hereof heating of the electrodes was separately measured and subsequently subtracted. For this background measurement, the absorbing layer was moved 5 mm away from the transducer. No other means of reconstruction were used.
3. Results The sensitivity comparison was based on measurements of the positive amplitude of the transducer signal. Care was taken to precisely control the laser pulse energy to be sure to generate identical initial pressure distributions for all experiments. The peak amplitudes measured were 4.0 ± 0.4 mV for the Al and 3.75 ± 0.4 mV for the ITO transducer, which means that the sensitivity is independent of the electrode material used. Fig. 2 shows the results of the six depth profiling experiments. The signal of the absorbing layer is clearly visible in all plots during the time interval from 1.5 to 2.1 ls. With an acoustic velocity of 1.5 mm/ls in water this is equivalent to a dye thickness of 0.9 mm. The time delay of 1.5 ls resulted from the distance between the dye layer and the sensor. As expected from BeerÕs law, there is an exponential raise in all the forward measurements and an exponential decay in all the backward measurements caused by light absorption inside the dye. The 6 mm illumination experiments show strong diffraction effects visible as rarefraction minima at the end of the signals. With the larger illumination of 9 mm these diffraction effects were almost eliminated. To quantify the recorded signals, the exponential coefficients were determined by fitting an exponential function to the
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
voltage (mV)
voltage (mV)
10
10
0
0
–10
(b) ITO forw 6mm
–10
10
10
0
0
–10
(c) ITO backw 6mm
–10
10
10
0
0
–10
1
2 3 time (µs)
(a) Al background forward 9mm
(d) Al forw 9mm
4
voltage (mV)
voltage (mV)
(a) Al forw 6mm
(e) ITO forw 9mm
10 0 –10 (b) ITO background forward 9mm
(f) ITO backw 9mm
–10 1
voltage (mV)
404
10 0 –10
2 3 time (µs)
4
Fig. 2. Optoacoustic depth profiling results. Subplots (a–c) correspond to the experimental setups (a–c) in Fig. 1 with 6 mm illumination diameter. Subplots (d–f) correspond to the experimental setups (a–c) in Fig. 1 with 9 mm illumination diameter. The six subplots show the exponential slope caused by light absorption inside the dye layer (raise in forward mode and decay in backward mode). The exponential coefficients obtained by fitting an exponential function to the corresponding traces in the time interval between 1.55 and 2.0 ls have the values: (a) 0.68 mm 1; (b) 1.15 mm 1; (c) 2.40 mm 1; (d) 1.06 mm 1; (e) 1.32 mm 1; (f) 1.94 mm 1. Coefficients (d–e) agree better to the actual value of 1.65 mm 1 because of the reduced acoustic diffraction by large area illumination. In the 6 mm experiments (a–c) the waveform was distorted by diffraction which is visible as rarefraction periods after the exponential slope. The small peaks at the boundaries of the slopes are caused by higher optoacoustic conversion efficiency in the clingfilm separating the dye from the surrounding water. Apart from thermal background subtraction and 15 times averaging, no reconstruction, such as deconvolution, has been used.
corresponding traces in the time interval between 1.55 and 2.0 ls. This resulted in absorption coefficients of: (a) 0.68 mm 1; (b) 1.15 mm 1; (c) 2.40 mm 1; (d) 1.06 mm 1; (e) 1.32 mm 1; (f) 1.94 mm 1, respectively. An interesting detail of the experiments is that the thermal background signal of the Al coated transducer was about 5 times stronger than the thermal background of the ITO coated transducer (10 mV compared to 2 mV, see Fig. 3).
0
1
2 time (µs)
3
4
Fig. 3. Thermal background signals measured in forward mode (9-mm diameter illumination) for: (a) Al coated PVDF (10–15% absorption); (b) ITO coated PVDF transducer (2–4% absorption). The strong noise during the first microsecond is electromagnetic noise caused by the Q-switch modulator of the laser.
4. Discussion The direct comparison of the ITO and metal coated PVDF transducers showed similar sensitivity. Therefore, the theory and design of ITO coated transducers can be widely transferred from metal electrode PVDF transducers, which are well established in the literature. This simplifies the design of pressure transducers used in optoacoustic systems since transducer geometry does not influence the irradiation. Furthermore, the low absorption of the ITO coating allows to place the transducer directly onto the tissue surface without getting heated by backscattered light. Therefore, depth profiling of i.e., human skin can be performed in the near field, which eliminates signal distortions commonly associated with acoustic diffraction. This means that the acoustic signal exactly mimics the depth profile of the absorbed energy [14]. The signal comparison of Fig. 2 reveals some interesting details. The Al transducer traces (a,d) show a less steep exponential part with small peaks at the boundary. The corresponding conspicuously low exponential coefficients of Al (a,d) can be explained by backreflected light from the Al coating
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
causing an additional irradiation of the sample from the opposite side. The pronounced peaks characterizing the top and the bottom surface of the absorbing sample are caused by the locally high optoacoustic conversion efficiency in the clingfilm separating the dye from the surrounding water. These peaks are less pronounced in the ITO traces because of the low pass filter effect of the thicker (40 lm) PVDF foil and the higher resistance of the sensor leads. For the depth profiling experiments the ITO transducer thickness was not reduced to avoid deformation caused by mechanical stresses induced by the non-optimized ITO sputtering process. With an optimized process it should be possible to fabricate thinner transducers featuring a larger bandwidth. Transducers with 25 lm or even 9-lm thickness would allow investigation of skin features or medication penetration with theoretical depth resolutions of up to 6 lm. The higher resistance of the ITO sensor together with the finite amplifier input capacitance mimic a low pass filter, which resulted in a smoothing of the signals recorded with the ITO transducer. For Al the lead resistance was in the low X range, but for the ITO this resistance varied from 102–104 X depending on the sputtering process. This resulted in lower cut-off frequencies and loss of high frequency features in the ITO depth profiles. The effect of larger irradiation becomes evident in the 9 mm illumination experiments (Fig. 2(d)– (f)). The rarefraction periods caused by diffraction effect are largely reduced in the 9 mm ITO experiments (e,f) compared to the 6 mm experiments (b,c). With homogeneous large illumination we can measure a signal undistorted by diffraction if no signal from the illumination boundary can reach the transducer during the measuring period (causality). This is satisfied with the simplified irradiation condition: homogeneous irradiation radius > transducer radius + sensing depth. In our case with an irradiation radius of 4.5 mm and a transducer radius of 1.5 mm the condition was satisfied until a depth of 3 mm, which corresponds to the far end of the dye sample. Of interest are the 9 mm ITO traces (e,f) with the exponential coefficients 1.32 and 1.94 mm 1 having a mean of 1.63 mm 1, which corresponds very well with the spectroscopically estimated
405
1.65 mm 1. The systematic deviation of 0.3 mm 1 of the two exponential coefficients (e,f) can be explained by inhomogeneous irradiation artifacts or finite input resistance of the differential amplifier. The input resistance of the differential probe and the capacitance of the transducer act as a high pass filter, which leads to a negative deviation of the estimated coefficients from the true values. This signal distortion can be corrected by deconvolution with the impulse response of the whole receiving system. We did, however, not correct our signals in order to demonstrate the high quality of the raw signals, even without correction.
5. Conclusion We propose a transparent ITO coated PVDF transducer for optoacoustic one-dimensional depth profiling. While having a similar sensitivity compared to an Al coated transducer, it is 5 times less susceptible to thermal signals caused by light absorption in the electrodes and provides transparency for the illuminating laser beam in backward mode. Its utility was demonstrated on a known absorbing dye layer. The ITO transducer can be used in backward mode and reduces acoustic diffraction artifacts using large area irradiation. Therefore, the proposed transducer design is ideally suited for quantitative measurement of depth resolved optical properties of layered structures in the human skin.
Acknowledgments The support of the Swiss National Science Foundation (Project No. 20-66622-01) and the EU Grant MEDPHOT is gratefully acknowledged.
References [1] P.C. Beard, T.N. Mills, Applied Optics 35 (1996) 663. [2] J.D. Hamilton, T. Buma, M. Spisar, M. OÕDonnell, IEEE Transactions on ultrasonics, Ferroelectrics and Frequency Control 47 (2000) 160.
406
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
[3] A.A. Karabutov, E.V. Savateeva, N.B. Podymova, A.A. Oraevsky, Journal of Applied Physics 87 (2000) 2003. [4] K.P. Koestli, M. Frenz, H.P. Weber, G. Paltauf, H. Schmidt-Kloiber, Proceedings of the SPIE The International Society for Optical Engineering 3916 (2000) 67. [5] R.G.M. Kolkman, E. Hondebrink, W. Steenbergen, F.F.M. de-Mul, IEEE Journal of Selected Topics in Quantum Electronics 9 (2003) 343. [6] R.A. Kruger, Pingyu-Liu, Yuncai-Fang, C.R. Appledorn, Medical Physics 22 (1995) 1605. [7] J.J. Niederhauser, D. Frauchiger, H.P. Weber, M. Frenz, Applied Physics Letters 81 (2002) 571. [8] J.J. Niederhauser, M. Jaeger, M. Frenz, Proceedings of the SPIE The International Society for Optical Engineering 4960 (2003) 118.
[9] G. Paltauf, H. Schmidt-Kloiber, K.P. Kostli, M. Frenz, Applied Physics Letters 75 (1999) 1048. [10] J.A. Viator, B. Choi, M. Ambrose, J. Spanier, J.S. Nelson, Applied Optics 42 (2003) 3215. [11] M. Yamazaki, T. Shimada, S. Sato, T. Miya, H. Ohigashi, H. Ashida, M. Obara, Review of Laser Engineering 30 (2002) 598. [12] J.J. Niederhauser, M. Jaeger, M. Frenz, Applied Physics Letters 85 (2004) 846. [13] Geng-Ku, Xueding-Wang, G. Stoica, L.V. Wang, Physics in Medicine and Biology 49 (2004) 1329. [14] M. Jaeger, J.J. Niederhauser, M. Hejazi, M. Frenz, Journal of Biomedical Optics 10 (2005) 024035. [15] V.S. Sudarshanam, S.B. Desu, Applied Optics 34 (7) (1995) 1177.
Transparent ITO coated PVDF transducer for optoacoustic depth profiling J.J. Niederhauser a, M. Jaeger a, M. Hejazi
a,c
, H. Keppner b, M. Frenz
a,*
a
c
Institute of Applied Physics, University of Berne, Sidlerstrasse 5, CH-3012 Berne, Switzerland b Ecole dÕInge´nieurs de lÕArc Jurassien (EIAJ), Le Locle, Switzerland Medical Physics Department, Faculty of Medicine, Teheran University of medical Sciences, Iran Received 29 September 2004; received in revised form 22 April 2005; accepted 2 May 2005
Abstract In optoacoustic depth profiling, illumination of absorbing biological tissue with short laser pulses generates pressure transients that are recorded with an acoustic transducer. The time profile of these pressure transients mirrors the depth resolved light absorption in the tissue, a relevant parameter, for example for transcutaneous medication penetration or in diagnosis of portwine stain. We propose a novel piezoelectric polyvinylidene fluoride transducer with transparent indium tin oxide (ITO) electrodes. The possibility of laser illumination through the transducer allows backward mode optoacoustic detection, a prerequisite to all parts of the human body not accessible from two sides. The sensitivity of the ITO transducer is comparable to classical metal coated transducers. Its transparency allows direct placement on the skin and combined with homogeneous large area illumination a one-dimensional acoustic propagation model becomes valid. This eliminates the need for computational compensation of diffraction effects. The influence of illumination area was illustrated in a comparison of the two transducers with an absorbing methylene blue sample in forward and backward mode. Ó 2005 Elsevier B.V. All rights reserved. Keywords: Optoacoustic imaging; Pressure transducer; Laser-induced ultrasound; Backward mode; Stress analysis; Acoustic diffraction
1. Introduction Medical need for highly selective diagnostic methods has motivated research for optoacoustic *
Corresponding author. E-mail address: [email protected] (M. Frenz).
depth profiling which combines high optical contrast with deep acoustic penetration [1–13]. In optoacoustics absorbing structures, such as blood capillaries of port wine stains, hidden inside scattering media are illuminated with short laser pulses to generate acoustic transients by means of thermoelastic effect. Detected at the surface with an acoustic
0030-4018/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.optcom.2005.05.005
402
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
transducer, these signals provide quantitative spatially resolved information on optical properties, such as the absorption coefficient, in the tissue. For optoacoustics, most human body parts are not accessible from two sides (forward mode), therefore laser illumination and acoustic sensing must be performed on the same surface (backward mode). The commonly used piezoelectric transducers have metal electrodes, which make them opaque for the laser pulse and complicate homogeneous tissue illumination. To circumvent this drawback, many designs have been proposed using either a piezoelectric ring for through illumination [4,5,11], for guiding the light around the acoustic transducer [3,6,8,10,13] or recently to replace the metal coating by a layer of salt-water as electrodes [14]. The ring design has a small illumination area making it difficult to couple enough light into the tissue without violating the safety limits for laser exposure. The around illumination design requires large distances between the skin and the acoustic transducer leading to lower signal amplitudes and distortion due to acoustic attenuation and diffraction. In both cases acoustic diffraction urge the use of deconvolution algorithms, which complicates direct quantitative measurement. To overcome the opaqueness problem various optical detection methods have been proposed [1,2,7,9,12]. We propose a transparent piezoelectric transducer based on indium tin oxide (ITO) coated polyvinilidene fluoride (PVDF) which overcomes all the limitations mentioned above. ITO is a transparent conductive coating widely used for electrodes in liquid crystal displays (LCD). PVDF is a transparent piezoelectric material ideally suited for sensitive wideband hydrophones. The proposed pressure transducer features a high piezoelectric sensitivity, allows quantitative measurement in the acoustic near field directly on the skin and is transparent for the illuminating laser pulse.
2. Materials and methods As seen in Fig. 1, the experimental setup consisted of two piezoelectric polyvinylidene fluoride (PVDF) transducers, one coated with aluminum (Al) electrodes (a) and one coated with transparent
(a) Al forward mode
laser dye
Al coated PVDF piezo transducer
PMMA (b) ITO forward mode
laser dye PMMA
transparent ITO coated PVDF piezo transducer
(c) ITO backward mode
dye PMMA
transparent ITO coated PVDF piezo transducer
laser Fig. 1. Optoacoustic depth profiling setup. The transducers consisted of a piezoelectric PVDF film with either Al electrodes (a) or ITO electrodes (b,c). PMMA was used as a backing material. A 1 mm layer of absorbing dye was homogeneously illuminated with a short laserpulse. This generated an acoustic transient representing the light absorption, which was recorded with the transducer. In contrast to Al, ITO is transparent, which allowed illumination not only in forward mode (a,b) but also in backward mode (c). Each configuration was used with two circular laser illumination diameters of 6 and 9 mm resulting in a total of six experiments.
indium tin oxide (ITO) electrodes (b,c). ITO provides surface electrical conductivity, while still maintaining good transparency (80%) over a large wavelength range from 450 to 2000 nm [15]. Due to the high refraction index of ITO (n = 2.05 @ 550 nm) the main loss in transmission is caused by reflection at the interfaces, only about 2% of the incident radiation is absorbed in the thin ITO layer. For both pressure transducers, a 40-lm thin PVDF foil was either vapor deposited with 200 nm Al or sputtered with 200 nm ITO. A mask was used during the coating process to obtain a circular active area of 3-mm diameter. The foils were placed on a 2-cm thick plexiglass (PMMA) backing using glycerin for adhesion by capillary forces. For mechanical protection, a clingfilm was placed on top using another drop of glycerin for acoustical coupling. A differential amplifier probe connected the piezoelectric electrodes to a digital
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
oscilloscope for data acquisition at a sampling rate of 1 GHz and 15 times averaging. The acoustic transients were generated optoacoustically by illuminating an absorbing layer with short laser pulses. A frequency trebled Nd:YAG laser with an optical parametric oscillator (OPO) generated 5 ns pulses at freely tunable wavelengths of 400– 2000 nm. The pulses were coupled into a 600-lm fiber and projected onto a circular area of adjustable diameter using a focusing lens at the distal end of the fiber. This resulted in homogeneously circular illumination on the sample. For forward measurements the fiber was placed on the opposite side of the absorbing sample relative to the transducer (see Fig. 1(a) and (b)). The ITO transducer additionally allowed backward mode measurement, a requirement for many medical applications (see Fig. 1(c)). For backward measurements the fiber was placed behind the plexiglass backing on the same side as the transducer. Two sets of experiments were conducted: (i) Direct comparison of transducer sensitivity of the two coated PVDF foils in forward mode. As absorbing sample, a layer of 300lM methylene blue solution was placed on top of each transducer. The layer thickness was 5 mm to ensure sufficient absorption in forward mode and to prevent the transducer from heating by direct irradiation. For the sensitivity comparison a large 15-mm diameter spot with 1.8 mJ/cm2 radiant exposure was chosen to avoid acoustic distortion (near field detection). (ii) Depth profiling of an absorbing layered sample in forward and backward mode. For the depth profiling a 25-lm Al transducer was used to increase temporal resolution. For the ITO transducer this thickness reduction was not possible because of prototype manufacturing reasons. The absorbing layer consisted of a 300 lM solution of methylene blue in water filled into a 1-mm thick cuvette covered on both sides with a clingfilm. The cuvette was immersed in water at a distance of 2 mm to the transducer. The optical absorption coefficient of the methylene blue solution was spectroscopically determined
403
to be la = 1.65 mm 1 at the wavelength of k = 577 nm. For comparing the ITO to the aluminum transducer the three possible arrangements (Al forward, ITO forward, ITO backward), each with two illumination diameters 6 and 9 mm were investigated, resulting in six experiments. The exposure was 4.6 mJ/cm2 for the 6-mm experiments and 3.9 mJ/cm2 for the 9-mm experiments. The thermal background caused by direct light absorption and hereof heating of the electrodes was separately measured and subsequently subtracted. For this background measurement, the absorbing layer was moved 5 mm away from the transducer. No other means of reconstruction were used.
3. Results The sensitivity comparison was based on measurements of the positive amplitude of the transducer signal. Care was taken to precisely control the laser pulse energy to be sure to generate identical initial pressure distributions for all experiments. The peak amplitudes measured were 4.0 ± 0.4 mV for the Al and 3.75 ± 0.4 mV for the ITO transducer, which means that the sensitivity is independent of the electrode material used. Fig. 2 shows the results of the six depth profiling experiments. The signal of the absorbing layer is clearly visible in all plots during the time interval from 1.5 to 2.1 ls. With an acoustic velocity of 1.5 mm/ls in water this is equivalent to a dye thickness of 0.9 mm. The time delay of 1.5 ls resulted from the distance between the dye layer and the sensor. As expected from BeerÕs law, there is an exponential raise in all the forward measurements and an exponential decay in all the backward measurements caused by light absorption inside the dye. The 6 mm illumination experiments show strong diffraction effects visible as rarefraction minima at the end of the signals. With the larger illumination of 9 mm these diffraction effects were almost eliminated. To quantify the recorded signals, the exponential coefficients were determined by fitting an exponential function to the
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
voltage (mV)
voltage (mV)
10
10
0
0
–10
(b) ITO forw 6mm
–10
10
10
0
0
–10
(c) ITO backw 6mm
–10
10
10
0
0
–10
1
2 3 time (µs)
(a) Al background forward 9mm
(d) Al forw 9mm
4
voltage (mV)
voltage (mV)
(a) Al forw 6mm
(e) ITO forw 9mm
10 0 –10 (b) ITO background forward 9mm
(f) ITO backw 9mm
–10 1
voltage (mV)
404
10 0 –10
2 3 time (µs)
4
Fig. 2. Optoacoustic depth profiling results. Subplots (a–c) correspond to the experimental setups (a–c) in Fig. 1 with 6 mm illumination diameter. Subplots (d–f) correspond to the experimental setups (a–c) in Fig. 1 with 9 mm illumination diameter. The six subplots show the exponential slope caused by light absorption inside the dye layer (raise in forward mode and decay in backward mode). The exponential coefficients obtained by fitting an exponential function to the corresponding traces in the time interval between 1.55 and 2.0 ls have the values: (a) 0.68 mm 1; (b) 1.15 mm 1; (c) 2.40 mm 1; (d) 1.06 mm 1; (e) 1.32 mm 1; (f) 1.94 mm 1. Coefficients (d–e) agree better to the actual value of 1.65 mm 1 because of the reduced acoustic diffraction by large area illumination. In the 6 mm experiments (a–c) the waveform was distorted by diffraction which is visible as rarefraction periods after the exponential slope. The small peaks at the boundaries of the slopes are caused by higher optoacoustic conversion efficiency in the clingfilm separating the dye from the surrounding water. Apart from thermal background subtraction and 15 times averaging, no reconstruction, such as deconvolution, has been used.
corresponding traces in the time interval between 1.55 and 2.0 ls. This resulted in absorption coefficients of: (a) 0.68 mm 1; (b) 1.15 mm 1; (c) 2.40 mm 1; (d) 1.06 mm 1; (e) 1.32 mm 1; (f) 1.94 mm 1, respectively. An interesting detail of the experiments is that the thermal background signal of the Al coated transducer was about 5 times stronger than the thermal background of the ITO coated transducer (10 mV compared to 2 mV, see Fig. 3).
0
1
2 time (µs)
3
4
Fig. 3. Thermal background signals measured in forward mode (9-mm diameter illumination) for: (a) Al coated PVDF (10–15% absorption); (b) ITO coated PVDF transducer (2–4% absorption). The strong noise during the first microsecond is electromagnetic noise caused by the Q-switch modulator of the laser.
4. Discussion The direct comparison of the ITO and metal coated PVDF transducers showed similar sensitivity. Therefore, the theory and design of ITO coated transducers can be widely transferred from metal electrode PVDF transducers, which are well established in the literature. This simplifies the design of pressure transducers used in optoacoustic systems since transducer geometry does not influence the irradiation. Furthermore, the low absorption of the ITO coating allows to place the transducer directly onto the tissue surface without getting heated by backscattered light. Therefore, depth profiling of i.e., human skin can be performed in the near field, which eliminates signal distortions commonly associated with acoustic diffraction. This means that the acoustic signal exactly mimics the depth profile of the absorbed energy [14]. The signal comparison of Fig. 2 reveals some interesting details. The Al transducer traces (a,d) show a less steep exponential part with small peaks at the boundary. The corresponding conspicuously low exponential coefficients of Al (a,d) can be explained by backreflected light from the Al coating
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
causing an additional irradiation of the sample from the opposite side. The pronounced peaks characterizing the top and the bottom surface of the absorbing sample are caused by the locally high optoacoustic conversion efficiency in the clingfilm separating the dye from the surrounding water. These peaks are less pronounced in the ITO traces because of the low pass filter effect of the thicker (40 lm) PVDF foil and the higher resistance of the sensor leads. For the depth profiling experiments the ITO transducer thickness was not reduced to avoid deformation caused by mechanical stresses induced by the non-optimized ITO sputtering process. With an optimized process it should be possible to fabricate thinner transducers featuring a larger bandwidth. Transducers with 25 lm or even 9-lm thickness would allow investigation of skin features or medication penetration with theoretical depth resolutions of up to 6 lm. The higher resistance of the ITO sensor together with the finite amplifier input capacitance mimic a low pass filter, which resulted in a smoothing of the signals recorded with the ITO transducer. For Al the lead resistance was in the low X range, but for the ITO this resistance varied from 102–104 X depending on the sputtering process. This resulted in lower cut-off frequencies and loss of high frequency features in the ITO depth profiles. The effect of larger irradiation becomes evident in the 9 mm illumination experiments (Fig. 2(d)– (f)). The rarefraction periods caused by diffraction effect are largely reduced in the 9 mm ITO experiments (e,f) compared to the 6 mm experiments (b,c). With homogeneous large illumination we can measure a signal undistorted by diffraction if no signal from the illumination boundary can reach the transducer during the measuring period (causality). This is satisfied with the simplified irradiation condition: homogeneous irradiation radius > transducer radius + sensing depth. In our case with an irradiation radius of 4.5 mm and a transducer radius of 1.5 mm the condition was satisfied until a depth of 3 mm, which corresponds to the far end of the dye sample. Of interest are the 9 mm ITO traces (e,f) with the exponential coefficients 1.32 and 1.94 mm 1 having a mean of 1.63 mm 1, which corresponds very well with the spectroscopically estimated
405
1.65 mm 1. The systematic deviation of 0.3 mm 1 of the two exponential coefficients (e,f) can be explained by inhomogeneous irradiation artifacts or finite input resistance of the differential amplifier. The input resistance of the differential probe and the capacitance of the transducer act as a high pass filter, which leads to a negative deviation of the estimated coefficients from the true values. This signal distortion can be corrected by deconvolution with the impulse response of the whole receiving system. We did, however, not correct our signals in order to demonstrate the high quality of the raw signals, even without correction.
5. Conclusion We propose a transparent ITO coated PVDF transducer for optoacoustic one-dimensional depth profiling. While having a similar sensitivity compared to an Al coated transducer, it is 5 times less susceptible to thermal signals caused by light absorption in the electrodes and provides transparency for the illuminating laser beam in backward mode. Its utility was demonstrated on a known absorbing dye layer. The ITO transducer can be used in backward mode and reduces acoustic diffraction artifacts using large area irradiation. Therefore, the proposed transducer design is ideally suited for quantitative measurement of depth resolved optical properties of layered structures in the human skin.
Acknowledgments The support of the Swiss National Science Foundation (Project No. 20-66622-01) and the EU Grant MEDPHOT is gratefully acknowledged.
References [1] P.C. Beard, T.N. Mills, Applied Optics 35 (1996) 663. [2] J.D. Hamilton, T. Buma, M. Spisar, M. OÕDonnell, IEEE Transactions on ultrasonics, Ferroelectrics and Frequency Control 47 (2000) 160.
406
J.J. Niederhauser et al. / Optics Communications 253 (2005) 401–406
[3] A.A. Karabutov, E.V. Savateeva, N.B. Podymova, A.A. Oraevsky, Journal of Applied Physics 87 (2000) 2003. [4] K.P. Koestli, M. Frenz, H.P. Weber, G. Paltauf, H. Schmidt-Kloiber, Proceedings of the SPIE The International Society for Optical Engineering 3916 (2000) 67. [5] R.G.M. Kolkman, E. Hondebrink, W. Steenbergen, F.F.M. de-Mul, IEEE Journal of Selected Topics in Quantum Electronics 9 (2003) 343. [6] R.A. Kruger, Pingyu-Liu, Yuncai-Fang, C.R. Appledorn, Medical Physics 22 (1995) 1605. [7] J.J. Niederhauser, D. Frauchiger, H.P. Weber, M. Frenz, Applied Physics Letters 81 (2002) 571. [8] J.J. Niederhauser, M. Jaeger, M. Frenz, Proceedings of the SPIE The International Society for Optical Engineering 4960 (2003) 118.
[9] G. Paltauf, H. Schmidt-Kloiber, K.P. Kostli, M. Frenz, Applied Physics Letters 75 (1999) 1048. [10] J.A. Viator, B. Choi, M. Ambrose, J. Spanier, J.S. Nelson, Applied Optics 42 (2003) 3215. [11] M. Yamazaki, T. Shimada, S. Sato, T. Miya, H. Ohigashi, H. Ashida, M. Obara, Review of Laser Engineering 30 (2002) 598. [12] J.J. Niederhauser, M. Jaeger, M. Frenz, Applied Physics Letters 85 (2004) 846. [13] Geng-Ku, Xueding-Wang, G. Stoica, L.V. Wang, Physics in Medicine and Biology 49 (2004) 1329. [14] M. Jaeger, J.J. Niederhauser, M. Hejazi, M. Frenz, Journal of Biomedical Optics 10 (2005) 024035. [15] V.S. Sudarshanam, S.B. Desu, Applied Optics 34 (7) (1995) 1177.