In this paper a new class of optical Fabry-Perot-based ultrasound detectors using low acoustic impedance glancing angle deposited (GLAD) films is demonstrated. GLAD is a single-step physical vapor-deposition (PVD) technique used to fabricate porous nanostructured thin films. Using titanium dioxide (TiO2), a transparent semiconductor with a high refractive index (n = 2.4), the GLAD technique can be employed to fabricate samples with tailored nano-porosity, refractive index periodicities, and high Q-factor reflectance spectra. The average acoustic impedance of the porous films is lower than bulk materials which will improve acoustic coupling, especially for high acoustic frequencies. For this work, two filters with high reflection in the C-band range and high transparency in the visible range (~80%) using GLAD films were fabricated. A 23 µm Parylene C layer was sandwiched between these two GLAD films in order to form a GLAD Fabry Perot Interferometer (GLAD-FPI). A high speed tunable continuous wavelength C-band laser was focused at the FPI and the reflection was measured using a high speed photodiode. The ultrasound pressure modulated the optical thickness of the FPI and hence its reflectivity. The fabricated sensor was tested using a 10 MHz unfocused transducer. The ultrasound transducer was calibrated using a hydrophone. The minimum detectable acoustic pressure was measured as 80 ± 20 Pa and the −3dB bandwidth was measured to be 18 MHz. This ultra-sensitive sensor can be an alternative to piezoelectric ultrasound transducers for any techniques in which ultrasound waves need to be detected including ultrasonic and photoacoustic imaging modalities. We demonstrate our GLAD-FPI for photoacoustic signal detection in optical-resolution photoacoustic microscopy (OR-PAM). To the best of our knowledge, this is the first time that a FPI fabricated using the GLAD method has been used for ultra-sensitive ultrasound detection.
©2013 Optical Society of America
Ultrasound detectors are found in a range of applications from humidifiers, flow-meters, sonar, and medical imaging to non-destructive test systems. Most ultrasound imaging techniques employ piezoelectric receivers to detect the ultrasonic signals. A number of optical detectors have been under investigation as alternatives to piezoelectric transducers and are beginning to offer exceptional receiver sensitivity compared to piezoelectric transducers .
For example, all-optical detection is becoming attractive in applications such as intra-operative, laparoscopic, and endoscopic ultrasound image guidance systems. In these cases, electrical interconnects must be minimized to make the imaging catheter small and flexible enough to navigate through small orifices and vessels. As well, high voltages present electrical safety hazards to patients. Additionally, legislative trends in some countries include restricting hazardous substances used in electrical and electronic equipment including lead. Hence there is significant opportunity for transducer technology which is lead-free. Finally, many invasive applications also require either sterilizable or disposable catheters. Hence the need for alternative transducer technologies makes optical-based transducers viable candidates.
Optical transducers have been under intense investigation for use in photoacoustic imaging systems where laser pulses excite tissues. In these systems, absorbed optical energy is first converted into acoustic signals by thermo-elastic expansion and then received by ultrasound transducers to be processed into final images . Unlike ultrasound images, photoacoustic images provide optical contrast, enabling estimation of blood oxygen saturation, imaging of optical contrast agents, and imaging of gene expression, among other novel and promising applications . Integration of optical-detectors with photoacoustic imaging is natural since the excitation source is optical as well.
Several teams have made considerable progress in this area, with work being done on Fabry-Perot etalons, micro-ring resonator devices, and fiber-Bragg grating approaches [4–7]. These detectors offer the high sensitivity and broad bandwidth important for photoacoustic imaging applications. For example, Zhang et al report a FPI sensor with 310 Pa sensitivity over a bandwidth of 39 MHz . They showed that the sensitivity of these FPI optical detectors could be significantly higher than piezoelectrics. Xie, et al demonstrated microring resonators with sensitivity reaching 29 Pa .
Our team recently introduced an optical-resolution photoacoustic micro-endoscopy system based on a sub-mm-footprint image-guide fiber bundle [8,9]. The system, however, required an external ultrasound transducer to detect signals from a scanned micron-sized laser spot. The present paper was in part motivated by the need to create a photoacoustic micro-endoscope system which includes all-optical detection.
In this paper we demonstrate a novel Fabry-Perot etalon- based ultrasound detector fabricated using GLAD nanostructured thin films. In previous studies FPI’s were fabricated by sputtering two stacks of alternate λ/4 thick layers of ZnS and Na3AlF6 on to a PMMA backing stub and Parylene C polymer film spacer [4,9–11]. As an alternative process to creating alternating material thin film stacks, the single-step GLAD technique can be used to create sensitive FPI-based ultrasound detectors using porous materials. As will be discussed later, the nanostructured GLAD films offer some practical and performance advantages over previous methods.
GLAD is a single-step physical vapor-deposition (PVD) technique used to fabricate nanostructured thin films [12–15]. By employing substrate motion and obliquely incident vapor flux, characteristic porous arrays of columnar structures can be produced from a range of organic, semiconductor, and dielectric materials. Devices such as optical filters, chromatographic plates, liquid crystal scaffolding, and solar cells have been created using the GLAD technique [16–19]. Using materials such as titania (TiO2), a wide band gap and transparent semiconductor with a high refractive index (n = 2.4), the GLAD technique can be used to fabricate samples with tailored refractive index periodicities with a high level of control and hence provide high Q-factor reflectance spectra. By periodically altering the deposition angle, and thus the local density and refractive index of deposited columns sub-layers, GLAD photonic crystals (PCs) can be created with periodic wavelength-scale structure and resonant Bragg reflection [20–22].
Additionally, because the GLAD films are porous with a typical density of 30 to 60% of bulk, the average acoustic impedance of the films (estimated as Z~20 MRayl in our case) can be lower than bulk materials (e.g. Z~38 MRayl for TiO2, Z~24 for ZnS) which will improve acoustic coupling to the Parylene C spacing material and to tissue, especially advantageous for high acoustic frequencies (>30 MHz) where GLAD-film thicknesses could be comparable with a significant fraction of a wavelength. Using Eq. (1) from  the estimated improvement in intensity transmission efficiency from water into the Parylene C etalon with a GLAD-film interface layer (Z~20 MRayl, thickness of 2.1μm, c(TiO2) = 8800m/s) is 12% at 15MHz compared with bulk TiO2. The improvement is calculated to be hundreds of percent for frequencies above 30MHz in future higher-bandwidth GLAD-FPI designs.
where Z1,Z2, Z3 and l corresponds to acoustic impedance of water, GLAD film, Parylene C and thickness of GLAD layer respectively. The sensitivity of the 18 MHz-bandwidth glass-backed GLAD-FPI is measured as ~80 ± 20 Pa. Using glass as the backing material makes this sensor a suitable choice for mounting on different optical components including lenses, optical fibers, etc. Previous FPI-sensors were designed for near transparency in the 700-900 nm NIR window with high reflectivity in the C-band for optical interrogation.
For many PAM and OR-PAM microvascular imaging applications visible rather than near infrared light is preferred due to the two orders of magnitude higher hemoglobin optical absorption leading to improved signal-to-noise. Non-trivially, the GLAD FPI is the first FPI sensor which is almost transparent at 532 nm, and highly reflective in the C-band. This is important since many laser systems (including frequency doubled Nd:YAG, Nd:YLF, Ytterbium-doped fiber lasers, and microchip lasers) operate in this wavelength region, and since hemoglobin absorption is high in the visible band. The GLAD FPI sensor also has high reflectivity in the C-band where a tunable 1550 nm interrogation beam can probe ultrasound-modulated etalon reflection dynamics. Availability of low cost lasers at the C-band communication range is another advantage of this sensor.
2. Fabry Perot Interferometer fabrication
The first goal of this work was to create a GLAD nanostructured layer with maximum transmission in the visible spectrum near 500 nm and high reflectance in the IR near 1550 nm. For this work we selected TiO2 as the deposition material and a simple high/low refractive index alternating stack (deposition angles α = 60° and 80° respectively) as the components of the one-dimensional photonic crystal structure. At α = 80° a more porous columnar structure is grown than at α = 60°, leading to a lower effective refractive index. To determine the optimal deposition parameters for the GLAD nanostructure we created a model of the optical system and iteratively adjusted the simulated thicknesses of the α = 60° and 80° layers until the required spectrum was found. Effective refractive indices of these layers (n60 = 1.93 and n80 = 1.42) were determined from established relationships between oblique deposition angle and effective refractive index for our deposition system and through comparison of simulated (Ts) and actual transmission (Ta) spectra . The expected spectrum for the GLAD thin film was created by modelling the effect of TM (p-polarization) light travelling through the multilayer structure using the characteristic matrix method . The structure has N sub-layers (j = 1 .. N), each described optically by a characteristic matrix (Mj) over a range of wavelengths (λ) as shown in Eq. (2).25,26]. The characteristic matrix of the complete structure is then given by Eq. (3). The transmission coefficient (t) of the entire optical system was then calculated using Eq. (4) and the transmittance spectra was found employing by Eq. (5).
Figure 1 shows the configuration of Fabry-Perot interferometers. The first step of the FPI fabrication process involved depositing the GLAD thin film layer which would act as the first mirror. The details of the GLAD thin film depositions carried out for the present work have been described in other work . A summary of the sample preparation steps is provided here. An amorphous TiO2 (Cerac, Inc., rutile phase 99.9% pure) structured thin film was deposited by electron-beam physical vapor evaporation (Axxis, Kurt J. Lesker Inc.) onto glass substrates (Schott B270, 1 inch2 x 0.04 inch thickness, S.I. Howard Glass Co. Inc.). The GLAD films were also concurrently deposited onto cleaved p-doped (100) silicon wafer pieces (University Wafer Inc.) for scanning electron microscope (SEM) characterization.
The substrates were affixed to a metal chuck which was in turn connected to two rotation motors. The deposition angle was controlled by the alpha motor, while the rotation was controlled by the phi motor. The substrates were first maintained at the oblique angle of α = 60°, while a computer-controlled motor rotated the substrates by 10 complete rotations in order to create a vertical columnar microstructure of 270 nm thickness. A flux deposition rate of 1 nm s−1 was maintained by manually monitoring the calibrated deposition rate returned by a quartz crystal microbalance oscillator (Maxtex, SC-105 Aluminum at 6 MHz) and adjusting the electron-beam current used to heat and evaporate the TiO2 melt. The deposition angle was then changed to α = 80° and another 10 rotation nanostructured layer of 170 nm were deposited. This computer-controlled deposition process was repeated until a total four layers of dense (tα = 60° = 270 nm) and porous (tα = 80° = 170 nm) sub-layers, with a final dense capping layer (tα = 60° = 270 nm), were fabricated. This number of layers provided a film with high transmission in the visible (95%) and good reflection (80%) in the IR. To promote TiO2 stoichiometry, O2 gas was introduced during deposition (pressure maintained at 3 × 10−5 Torr); after the deposition the glass samples were also annealed at 150 °C for 24 h [27,28]. In all cases, the system base pressure was below 1 x 10−6 Torr.
A 23 µm thick Parylene-C polymer was deposited at the top of the first GLAD film by vapor deposition of the Parylene polymer Parylene-C (PDS 2010, Lab Coater 2, University of Alberta Nanofabrication Facility). Following deposition of the Parylene-C layer, samples were placed back into the Axxis vacuum system for the second GLAD deposition. Again alternating layers of α = 60° and 80° columnar films with respective layer thicknesses of 270 nm and 170 nm were deposited. As with the first GLAD film, four sets of alternating dense and porous sub-layers, with a dense capping layer, were produced. During deposition O2 gas was introduced (pressure maintained at 3 × 10−5 Torr). Following deposition, annealing was carried out at a lower temperature of 80°C for 24 h to promote stoichiometry while not affecting the Parylene-C layer. Figure 2 shows an SEM image of a GLAD film sample, with four sets of alternating dense and porous layers, and the final capping layer, visible.
Top-down and side-view SEMs of the silicon witness samples were taken on a Hitachi field emission S-4800 SEM. Optical characterization was carried out on a Perkin Elmer 900 UV-VIS-NIR spectrophotometer. Transmittance was measured for normally-incident light at wavelengths from 490 to 1700 nm, in 1 nm increments.
A thin layer (4 µm) of Parylene-C was used to encapsulate the entire sample to protect it from humidity and environmental effects known to shift resonant peaks [4,22]. Film stress could be a major concern and a source of etalon non-uniformity should the reflective layers buckle. Etalon non-uniformity in turn will mean that each location on the etalon may have a slightly different resonant peak, resulting in the need for laser tuning at each interrogation location. Top-down SEMs were used to verify that the deposition of Parylene-C on top of a GLAD thin film or the deposition of a GLAD layer on top of Parylene-C did not result in film cracking suggesting low stress. The porous nature of the GLAD films prove advantageous compared to bulk-materials for reducing film stress, which may in turn improve etalon uniformity but could result in unwanted optical scattering. However, the columnar features in each layer as seen in Fig. 2 are sub-wavelength, so scattering should be minimal. The structure of the GLAD films deposited on Parylene-C was as similar to the GLAD films deposited on glass as validated by SEM and optical transmission spectra.
Figure 3(a) shows the transmission profile of a single GLAD layer filter with high transmission at 532 nm and high reflection around 1550 nm. It’s shown that both simulation and experimental work are in a good agreement. Figure 3(b) shows experimental results of the FPI peaks formed when the Parylene-C layer is sandwiched between the two GLAD layers. The experimental result of transmission spectrum of the GLAD-FPI indicates about 70% transparency near the excitation wavelength (Fig. 3(c)). Figure 3(d) shows the measured reflectance spectra for 1 peak using the optical setup (Fig. 4 ). The integration beam was focused tight at the surface of the FPI and the wavelengths were scanned while the reflected light was measured. The Q factor of this peak was measured as 620.
3. Result and discussion
Figure 4(a) (setup 1) shows the experimental setup used to test the GLAD-FPI. We employed a 10 MHz unfocused transducer as a transmitter to test the receiver sensitivity. A customized holder for the transducer was engineered in order to keep the water between GLAD- FPI and the transducer. Water was used for ultrasound coupling. A tunable continuous wavelength (CW) C-band laser (TLK-L1550R, Thorlabs Inc., New Jersey) was used in order to tune the interrogation laser wavelength to the point of maximum slope on the FPI peaks. The light at the laser aperture was coupled to a single mode fiber and collimated. This collimated interrogation beam was passed through a polarized beam splitter (VBA05-1550, Thorlabs Inc., New Jersey) and λ/4 zero order wave plate (Thorlabs Inc., New Jersey), onto the sample via an focusing objective lens, and back through the wave-plate creating 90° polarization which then reflects at the polarizing beam-splitter in order to guide the maximum possible intensity of reflected light to a 150 MHz-bandwidth InGaAs photodiode (PDA10CF, Thorlabs Inc., New Jersey). An objective lens (518125, LEICA, Germany) was used in front of the photodiode in order to refocus all possible reflected interrogation light to the small photodiode element. The 30 mW interrogation beam was focused on the GLAD-FPI. The output of the photodiode was amplified (Olympus 5900PR) and digitized using an 8-channel PCI digitizer (Gage card CS8289) at a sampling rate of 125 MSamples/s. The ultrasound transducer output was calibrated using a needle hydrophone (HNP-0400, ONDA, Sunnyvale). Figure 4(b) shows the fabricated FPI. The minimum detectable acoustic pressure was measured as ~80 ± 20 Pa and represents the pressure detected with a signal-to-noise ratio of 1.
In order to demonstrate the capability of the sensor for 532 nm excitation optical-resolution photoacoustic microscopy (OR-PAM), we performed photoacoustic imaging on a network of 7 µm diameter carbon fibres, shown in Fig. 5 (using setup 2 of Fig. 4). Nanosecond-pulses from a Ytterbium-doped fiber laser with repetition rates of up to 600 kHz (YLP-G, IPG Photonics Corporation) were used as the source. A beam splitter was used to deflect a small amount of light into a high speed photodiode to detect laser pulses and trigger the high speed data acquisition card. The excitation laser beam passes through a 2D galvanometer scanning mirror system (Thorlabs). The mirrors are controlled by a two channel function generator (Tektronix AFG3022B). An objective lens with 4 mm focal length was used to focus the excitation light into the target. The carbon fiber networks are located ~2 mm away from FPI. The optical lateral resolution was estimated ~7 µm as shown previously . The −3 dB bandwidth was measured as ~18 MHz by imaging carbon fiber network with ~7 µm diameter as shown in Fig. 6 .
As expected, since glass (with a high acoustic impedance compared to Parylene) is the backing material, our bandwidth is about half of similar FPI sensors with a polymer (low acoustic impedance) backing and sensitivity is roughly double . Polymer-backed sensors could be investigated by others, however in future work we aim to coat different glass-based optical components such as fiber optics, lenses, and prisms. To the best of our knowledge, this is the first time that a FPI fabricated using the GLAD method has been used for ultra-sensitive ultrasound detection. Other work using GLAD films predicts that ultra-high reflectivity is possible , which should lead to very high quality factors (Q-factors).
One of the other important factors of our design is that the interrogation wavelength and focused spot on the GLAD-FPI were fixed during the imaging session. However for a larger field of view, it may be more appropriate to scan the interrogation beam along with excitation beam. Etalon uniformity has been problematic for other FPI sensors, meaning that interrogation wavelengths must be tuned at each spatial location on the etalons. Etalon uniformity was studied by quantifying the changes in reflectivity over different spatial interrogation locations. The optimum interrogation wavelength (corresponded to the steepest slope on the reflectivity spectra) varies by only less than 1nm over a 1 mm × 1 mm area, suggesting locally high uniformity. Over distance scales of more than 2 cm, resonant peaks were shifted several nm. We attribute the locally high etalon uniformity to the Parylene C deposition uniformity (~5 nm over an area of 1 cm2, similar to previous work ) and the low-stress nature of the GLAD films. The locally high etalon uniformity means interrogation beams need not be tuned at each interrogation spot, and could eventually pave the way for less expensive interrogation lasers with less stringent tuning requirements and for simultaneous recording of pressure signals at multiple different spatial locations with a single laser. Present work involving OR-PAM does not involve scanning the interrogation spot at all, so non-ideal etalon uniformity is well-tolerated.
For future work reflection mode OR-PAM can be implemented to take advantage of high transmission GLAD filters at 532 nm wavelengths. The reflection mode setup can be utilized using a beam combiner and different optical setup  to perform in vivo imaging. This ultra-sensitive sensor can be a viable alternative for piezoelectric ultrasound transducers for any techniques in which ultrasound waves need to be detected. The GLAD-FPI with high transmission at 532 nm will be a powerful and flexible technology for next-generation all-optical OR-PAM and photoacoustic micro-endoscopy systems [8,9].
The process of designing an ultra-sensitive all-optical ultrasound detector using glancing angle deposition nanostructured films with glass backing is demonstrated. GLAD-based filters with high reflection in the C-band range and transparent in the visible range were fabricated on either side of a 23 µm Parylene C layer to form a Fabry Perot Interferometer. The GLAD method allows low acoustic impedance FPI device fabrication for highly sensitive ultrasound detection. A tunable CW C-band laser was focused at the FPI and the reflection was measured using a high speed photodiode. High uniformity of the GLAD-FP enables OR-PAM imaging using a fixed interrogation wavelength. The GLAD-FP sensor has Q-factors better than 620 and 80 ± 20 Pa sensitivity over an 18 MHz bandwidth. Utility for laser scanning optical resolution photoacoustic microscopy is demonstrated while the integration laser was focused and fixed at the GLAD FPI.
The first author gratefully acknowledges funding from an Alberta Innovates Graduate Student Scholarship and an SPIE Scholarship in Optics & Photonics. We also gratefully acknowledge funding from NSERC (IRCPJ 203978-08, 355544-2008, 375340-2009, STPGP 396444), Terry- Fox Foundation and the Canadian Cancer Society (TFF 019237, TFF 019240, CCS 2011-700718), the Alberta Cancer Research Institute (ACB 23728), the Canada Foundation for Innovation, Leaders Opportunity Fund (18472), Alberta Advanced Education & Technology, Small Equipment Grants Program (URSI09007SEG), Alberta Innovates: Technology Futures, Micralyne, and Microsystems Technology Research Initiative (MSTRI RES0003166).
References and links
2. P. Hajireza, W. Shi, P. Shao, S. Kerr, and R. J. Zemp, “Optical-resolution photoacoustic micro-endoscopy using image-guide fibers and fiber laser technology,” Proc. SPIE 7899, 78990P, 78990P-6 (2011). [CrossRef]
4. E. Zhang, J. Laufer, and P. Beard, “Backward-mode multiwavelength photoacoustic scanner using a planar Fabry-Perot polymer film ultrasound sensor for high-resolution three-dimensional imaging of biological tissues,” Appl. Opt. 47(4), 561–577 (2008). [CrossRef] [PubMed]
6. S. Ashkenazi, Y. Hou, T. Buma, and M. O’Donnell, “Optoacoustic imaging using thin polymer etalon,” Appl. Phys. Lett. 86(13), 134102 (2005). [CrossRef]
7. S. Huang, S. Ashkenazi, Y. Hou, R. S. Witte, and M. O'Donnell, “Toward fiber-based high-frequency 3D ultrasound imaging,” Proc. SPIE 6437, 643728, 643728-8 (2007). [CrossRef]
9. E. Zhang and P. C. Beard, “A miniature all-optical photoacoustic imaging probe,” Proc. SPIE 7899, 78991F, 78991F-6 (2011). [CrossRef]
10. E. Z. Zhang, J. Laufer, R. B. Pedley, and P. Beard, “3D photoacoustic imaging system for in vivo studies of small animal models,” Proc. SPIE 6856, 68560P, 68560P-8 (2008). [CrossRef]
11. E. Z. Zhang, J. Laufer, B. Považay, A. Alex, B. Hofer, W. Drexler, and P. Beard, “Multimodal simultaneous photoacoustic tomography, optical resolution microscopy, and OCT system,” Proc. SPIE 7564, 75640U, 75640U-7 (2010). [CrossRef]
12. K. Robbie and M. J. Brett, “Sculptured thin films and glancing angle deposition: growth mechanics and applications,” J. Vac. Sci. Technol. A 15(3), 1460–1465 (1997). [CrossRef]
13. A. Lakhtakia and R. Messier, in Sculptured Thin Films: Nanoengineered Morphology and Optics (SPIE, 2004).
14. R. Messier, V. C. Venugopal, and P. D. Sunal, “Origin and evolution of sculptured thin films,” J. Vac. Sci. Technol. A 18(4), 1538–1545 (2000). [CrossRef]
15. M. M. Hawkeye and M. J. Brett, “Glancing angle deposition: fabrication, properties, and applications of micro- and nanostructured thin films,” J. Vac. Sci. Technol. A 25(5), 1317–1335 (2007). [CrossRef]
16. J. Q. Xi, M. F. Schubert, J. K. Kim, E. F. Schubert, M. Chen, S. Y. Lin, W. Liu, and J. A. Smart, “Optical thin-film materials with low refractive index for broadband elimination of Fresnel reflection,” Nat. Photonics 1, 176–179 (2007).
17. S. R. Jim, M. T. Taschuk, G. E. Morlock, L. W. Bezuidenhout, W. Schwack, and M. J. Brett, “Engineered anisotropic microstructures for ultrathin-layer chromatography,” Anal. Chem. 82(12), 5349–5356 (2010). [CrossRef] [PubMed]
18. J. C. Sit, D. J. Broer, and M. J. Brett, “Liquid crystal alignment and switching in porous chiral thin films,” J. Adv. Mater. 12(5), 371–373 (2000). [CrossRef]
19. H. Y. Yang, M. F. Lee, C. H. Huang, Y. S. Lo, Y. J. Chen, and M. S. Wong, “Glancing angle deposited titania films for dye-sensitized solar cells,” Thin Solid Films 518(5), 1590–1594 (2009). [CrossRef]
21. M. M. Hawkeye and M. J. Brett, “Narrow bandpass optical filters fabricated with one-dimensionally periodic inhomogeneous thin films,” J. Appl. Phys. 100(4), 044322 (2006). [CrossRef]
22. M. M. Hawkeye, K. M. Krause, and M. J. Brett, “Ambient humidity monitoring using a 1D photonic crystal sensor fabricated with glancing angle deposition,” Proc. SPIE 7356, 73560G (2009). [CrossRef]
23. S. C. Richard, Cobbold, in Foundations of Biomedical Ultrasound (Oxford University, 2006).
24. M. M. Hawkeye, “Engineering optical nanomaterials using glancing angle deposition,” PhD Thesis, UofA (2010).
25. H. A. MacLeod, in Thin Film Optical Filters (American Elsevier, 1969).
27. A. C. van Popta, J. C. Sit, and M. J. Brett, “Optical properties of porous helical thin films and the effects of post-deposition annealing,” Proc. SPIE 5464, 198–208 (2004). [CrossRef]
28. K. N. Rao, “Influence of deposition parameters on optical properties of TiO2 films,” Proc. SPIE 41, 2357–2364 (2002).
29. B. T. Cox, E. Z. Zhang, J. G. Laufer, and P. C. Beard, “Fabry Perot polymer film fibre-optic hydrophones and arrays for ultrasound field characterization,” J. Phys. 1, 32–37 (2004).