Design, Development and Characterization of. Wideband Polymer Ultrasonic Probes. for Medical Ultrasound Applications. A Thesis

Transcription

1 Design, Development and Characterization of Wideband Polymer Ultrasonic Probes for Medical Ultrasound Applications A Thesis Submitted to the Faculty of Drexel University by Vadivel Devaraju in partial fulfillment of the requirements for the degree of Doctor of Philosophy January 2003

2 ii DEDICATIONS Dedicated to MY FATHER for his gentle humanity and dedication to excellence in bringing up me to my present level by taking the sufferings all along for my well-being. In memory of MY MOTHER for all her sacrifices in life for my betterment and for her magnificent devotion to our family. MY WIFE a part of my soul for her love, support, dedication and commitment to our family value, culture and tradition. MY SON who is my strength, reflection of my thoughts and my emulation for his great affection, understanding, forbearance and standing shoulder to shoulder with me in preserving our family heritage. They are all the unbreakable four pillars of my life upon which I always stand unbowed. ACKNOWLEDGEMENTS

3 iii I express my sincere thanks and gratitude to my thesis advisor Dr.Peter A. Lewin for his advice and support. My decade of association with Dr. Lewin will be evergreen in my memory. My sincere thanks and appreciation from the bottom of my heart are to my thesis committee members Dr. Richard B. Beard, Dr. T. S. Venkataraman, Dr. Hector V. Ortega and Dr. Feroze B. Mohamed for their guidance and encouragement during the process of my dissertation. I also extend my thanks to Dr. Ryszard M. Lec for providing me the office space in his lab and Dr. Philip E. Bloomfield for his assistance. I gratefully acknowledge the financial support provided by Dr. Banu Onaral and Dr. William Freedman in the form of Calhoun fellowship. A debt of gratitude is owed to Dr. Mark E. Schafer for his support and for the help in developing the new type of hydrophone and including it as a part of this thesis. I wish to thank Dr.Hendrik Bleeker for all his help particularly while developing the low frequency calibration method. I express my gratitude to my brother, sisters and the members of their family for taking my family burden back at home, and giving me the strength to face the challenges in life. I express my thanks to all my in-laws for their timely help from time to time and to all my relatives and friends who are all along with me by extending helping hand at all times particularly during the time of need. Table of Contents

4 iv LIST OF TABLES......xi LIST OF FIGURES...xii ABSTRACT.xvi CHAPTER 1. INTRODUCTION Objectives, Outline and Scope of the Research Objectives Outline Scope of the research Motivation Significance Fundamentals of Biomedical Ultrasound Fundamentals of Medical Ultrasound Imaging Fundamentals of Medical Ultrasound Imaging Transducer Resolution Axial resolution Lateral resolution Transverse resolution Contrast resolution Bandwidth Sensitivity Backing material Electrical matching/tuning 23

5 v Mechanical matching (Acoustic interface/impedance characteristics) Acoustic energy loss Fundamentals of Array Transducer Fundamentals of Ultrasonic Hydrophone Probe..29 CHAPTER 2. BACKGROUND AND THEORY Introduction Ultrasonic Polymer Hydrophone Probe Need for wider bandwidth hydrophone probe Calibration Techniques Planar scanning technique Time delay spectrometry technique combined with substitution method Need for developing a new measurement procedure Single Element Imaging Transducer Transducer requirements Conventional piezoceramic imaging transducers Conventional polymer imaging transducers Single element, single layer transducer Single element, multilayer polymer transducer Barker code polymer imaging transducer Barker code concept Principle of operation.45

6 vi 2.5 Multielement (array) Imaging Transducer Side lobes and grating lobes Cross talk effects Electrical cross talk Acoustical cross talk Electromechanical coupling...52 CHAPTER 3. PIEZOELECTRIC MATERIALS Introduction Material Properties Piezoelectric stress constant Piezoelectric strain constant Transmitting constant Receiving constant Dielectric permittivity (constant) Electromechanical coupling coefficient Dissipation factor Mechanical loss tangent Electrical loss tangent Acoustic impedance Piezoelectric Materials for Imaging Transducers Piezoceramic Piezocomposite.59

7 vii Single crystal Piezopolymer Modes of Vibration Summary Conclusion...68 CHAPTER 4. DEVELOPMENT AND CHARACTERIZATION OF IMPROVED DESIGN OF POLYMER HYDROPHONE PROBE Introduction Synopsis Acoustic sensitivity Frequency response and bandwidth Angular response Effective aperture size Orientation effects Conclusion...72 CHAPTER 5. DEVELOPMENT OF CALIBRATION PROCEDURE Introduction Calibration Setup Calibration Procedure Initial alignment procedure Determination of the peak pressure amplitude location Calculation of pulse intensity integral.76

8 viii Cross axis scan Raster scan Radiation Force Balance Measurements Hydrophone Sensitivity Calculation Comparative Study Results Correlation and Discussion Summary Conclusions..89 CHAPTER 6. DEVELOPMENT AND CHARACTERIZATION OF SINGLE ELEMENT NON-RESONANT POLYMER IMAGING TRANSDUCER Introduction Characterization of PVDF Film Impedance / Admittance Measurements Design and Development of System Electronics and Controls Design Considerations of Single Element Polymer Imaging Transducer Material selection, thickness and active area Electroding Backing material Wear protecting front matching layer Adhesion Fabrication Process Characterization of Non-Resonant Single Element, Single Layer Transducer and Multilayer Barker Code Polymer Imaging Transducer Electrical characterization.101

9 ix Experimental system Pulse-echo response Diffraction correction Sensitivity correction Summary Conclusion.115 CHAPTER 7. DESIGN, DEVELOPMENT AND CHARACTERIZATION OF NON-RESONANT MULTI ELEMENT (ARRAY) SINGLE LAYER/MULTILAYER POLYMER IMAGING TRANSDUCER Introduction Design Consideration and Description of Array Structure Fabrication Process Design and Development of System Electronics and Control Experimental System for Performance Evaluation Performance Evaluation of the Array Transducer Single layer array transducer Pulse echo response of individual element Pulse echo response of all the four elements Uniformity Multilayer array transducer Pulse echo response Summary Conclusion.133

10 x CHAPTER 8. SUMMARY AND CONCLUSION Summary of the Research Work Conclusion Suggestion for Future Work BIBLIOGRAPHY 140 APPENDIX A: DEVELOPMENT AND CHARACTERIZATION OF IMPROVED DESIGN OF DOUBLE LAYER POLYMER HYDROPHONE PROBE VITA 183

11 xi List of Tables 1. Comparison of piezoelectric material parameters Comparative statement showing the variation of frequency response and sensitivity of hydrophones having different thickness of PVDF film Directivity data illustrating the variation of effective diameter of the hydrophones measured along the two orthogonal axes at 5, 7.5 and 10 MHz Directivity data illustrating the variation of mean effective diameter of the hydrophones measured along the two orthogonal axes Directivity data illustrating the variation of effective diameter of the hydrophones having different thickness of PVDF film measured along the two orthogonal axes.178

12 xii List of Figures 1. Schematic representation of the outline of the research work 5 2. Schematic representation of the Barker code arrangement Geometry of three important vibration modes Typical measured hydrophone waveform with the corresponding PII and frequency spectrum of 1 MHz circular piston source Beam plot along the x and y axes of the 1 MHz acoustic source. The dashed lines show the gaussian beam (theoretical) distribution with the same 6 db beamwidth as that of the measured transducer Experimental set up for rater scanning (A) and illustration of raster fashion (B) Three-dimensional representation of the intensity field produced by the 1 MHz source transducer and obtained from planar scanning End-of-cable voltage sensitivity versus frequency of the double layer PVDF hydrophone in the frequency range 0.3 to 1 MHz obtained using Planar scanning technique ( ), Time delay spectrometry method ( ) and calibration data provided by National Physical Laboratory (NPL, ), UK Plot showing the variation of admittance magnitude and phase of a 56 µm piezo film in air Plot showing the variation of impedance magnitude and phase of a 56 µm piezo film in air Pressing mechanism while in use Different stages of the fabrication process of single element transducer Completed single element transducers Plot showing the variation of admittance magnitude and phase of single layer transducer Plot showing the variation of impedance magnitude and phase of a single layer transducer.101

13 xiii 16. Plot showing the admittance magnitude of 56 µm piezo film, one layer and three layer transducer Experimental setup for pulse echo measurement Pulse-echo response (in time and frequency domain) of single element, single layer transducer at 1 cm depth in water for a monocycle sine burst excitation Experimental pulse echo response (in time domain) of a single element, three layer Barker code transducer at 1 cm depth in water for a monocycle sine burst excitation Pulse echo frequency response of single element, single layer transducer at different depths in water for a monocycle sine burst excitation Pulse-echo frequency response of single element, three-layer transducer at different depths in water for a monocycle sine burst excitation Pulse-echo frequency response of single and three-layer PVDF transducers at 1 cm depth in water for a monocycle sine burst excitation Pulse-echo frequency response of one single layer and two three layer single element polymer transducers at 1 cm depth in water for a monocycle sine burst excitation Pulse echo response (in time and frequency domain) of single element, single layer polymer transducer at different depths in water for monocycle sine burst excitation Pulse echo response of a single element, single layer transducer for different water path length at peak frequency for monocycle sine burst excitation Pulse echo response of single layer transducer at different depths in water for a monocycle sine burst excitation (with and without diffraction correction) Schematic diagram showing the capacitance loading of the polymer transducer Schematic representation of array element and bonding pad pattern at one side of the polymer film of the proposed linear array transducer Actual electrode pattern of an array layer Different stages of the construction process of 4 element array transducer Completed array transducers..123

14 xiv 32. Photograph of the newly built transmit/receive control circuit and preamplifier Experimental configuration for pulse echo measurements Experimental pulse echo responses (in time & frequency domain) for one of the elements of single layer array transducer at 1 cm depth in water for a monocycle sine wave excitation Experimental pulse echo response (in time and frequency domain) while exciting all the four elements of single layer array transducer at 1 cm depth in water for a mono cycle sine wave excitation A typical observed time domain pulse echo response of one of the elements of the single layer array transducer at 1 cm depth in water for a monocycle sine burst excitation Pulse-echo responses (in time & frequency domain) of a representative element stack of three layer array transducer at 1cm depth in water for a monocycle sine wave excitation Typical observed time domain pulse echo response of one of the elements of single layer (A) and one of the element stacks of three layer (B) array transducer at the peak frequency at 1 cm depth in water for a monocycle sine burst excitation Pulse echo frequency response of two element stacks of three layer array transducer at 1 cm depth in water for a monocycle sine burst excitation Actual view of the newly developed double layer polymer hydrophone probe Schematic representation showing the capacitance loading of the hydrophone probe Frequency response plot of a double layer hydrophone probe using 9+9 µm thick polymer film Frequency response plots of double layer hydrophones using µm and 25+9 µm polymer film Typical directional response of double layer polymer membrane hydrophones with active element size of 0.4 mm (Fig. a), 0.6 mm (Fig. b) and 1 mm (Fig. c) in diameter, measured at the frequencies of 5 MHz, 7.5 MHz and 10 MHz 175

15 xv 45. Combined directional response of double layer polymer membrane hydrophones with active element size of 0.4 mm in diameter, measured at the frequencies of 5 MHz, 7.5 MHz and 10 MHz Combined directional response of double layer polymer membrane hydrophones with active element size of 0.4 mm, 0.6 mm and 1 mm in diameter, measured at 7.5 MHz Combined directional response of double layer polymer membrane hydrophone with active element size of 0.5 mm in diameter, measured at 10 MHz in two orthogonal axes.178

16 xvi Abstract Design, Development and Characterization of Wideband Polymer Ultrasonic Probes for Medical Ultrasound Applications Vadivel Devaraju Peter A. Lewin, Ph.D. This dissertation deals with the design, development and characterization of non-resonant polymer ultrasonic probes for medical ultrasound applications. Both single element and multielement imaging transducer design having single layer and multilayer configuration were developed with the primary goal of minimizing the trade off between resolution and penetration depth. The simultaneous improvement in the transducers pulse-echo sensitivity and bandwidth was achieved by employing a multilayer structure made of thin piezopolymer films and utilizing the concept of Barker code. The results of the experiments indicated that the multilayer Barker code transducers provided the widest bandwidth in comparison with the imaging transducers made of conventional piezoelectric ceramic material. Also, they exhibited enhanced sensitivity compared to a single layer piezopolymer transducer. Specifically, the results indicated that the 6 db fractional bandwidth extending over 2 decades (20 MHz) could be achieved in the case of non-resonant transducers, whereas the conventional resonant design imaging transducer could provide only about one half of the bandwidth. The polymer array transducers also showed uniform acoustic response from element to element, which is desirable in order to obtain high quality ultrasound images.

17 xvii The double layer hydrophone probes used for characterization of the imaging transducers were fabricated using dissimilar thickness of PVDF polymer film. This technique ensured simultaneous enhancement of sensitivity and bandwidth. Also, a measurement procedure employing planar scanning technique was developed to calibrate the hydrophone probes below 1 MHz for adequate characterization of acoustic field produced by the imaging transducers. The results of this work indicate that the new class of transducers developed features significantly enhanced bandwidth. Such transducers hold promise to be capable of operating at clinically relevant frequencies and suitable for use at fundamental, subharmonic and higher harmonics imaging. It is expected that this non-resonantly operating imaging transducer would become a useful clinical tool in medical imaging and could improve diagnostic efficacy.

18 Copyright 2003 Vadivel Devaraju. All Rights Reserved

19 1 CHAPTER 1: INTRODUCTION This dissertation deals with the design, development and characterization of wideband polymer ultrasonic probes for medical ultrasound applications. The research work included both the single element and multielement (array) imaging transducers design and also hydrophone probe used in characterization of the imaging transducers. The first phase of this work involved fabrication of a double layer hydrophone using dissimilar thicknesses of polyvinylidene fluoride (PVDF) polymer film and development of associated measurement procedures. Next, the experience and knowledge gained in the fabrication process were used to design and develop the single element, multilayer imaging transducers and multielement (array), multilayer imaging transducers. The overall goal of this work was to investigate a possibility to minimize the unavoidable trade off between the resolution and penetration depth in clinical ultrasound imaging applications. The thesis is divided into eight chapters. This chapter reviews the objectives and the scope of the work performed, and the significance and motivation for the research along with a brief review of fundamentals of the ultrasound essential for understanding of the tasks involved in development of the non-resonant imaging probes. Chapter two reviews background and theory needed in the course of this work and also provides the background information including review of the work done by other researchers. Chapter three contains detailed information about different properties of various piezoelectric materials used in imaging probes and justifies the effort in exploring the applicability of

20 2 the polymer as piezoelectric material for imaging transducers. A brief review of the development and characterization of double layer PVDF polymer hydrophone probes used for characterization of the imaging transducers developed in this work is given in chapter four, whereas a more detailed description of the development process and results are presented in Appendix A. Chapter five deals with the measurement procedure developed for calibrating the double layer ultrasound hydrophone probe below 1 MHz using planar scanning technique along with the related background and the calibration results. Chapter six describes the development of a single element, single layer transducer and three layer Barker code transducers. The design and development of multielement, single layer as well as multilayer array transducers are presented in chapter seven. Finally, chapter eight presents the summary of the work accomplished together with conclusions and suggestions for future work Objectives, Outline and Scope of the Research Objectives In order to achieve the primary goal of having a single imaging ultrasound transducer capable of operating at all clinically relevant frequencies and minimizing the trade off between image resolution and penetration depth, a set of objectives was devised. The objectives of the research described in the following were to design, develop and characterize piezoelectric polymer ultrasonic probes used for medical ultrasound applications, which are capable of exhibiting simultaneous enhancement of both the bandwidth and sensitivity in the case of both imaging transducers and hydrophone probes. The objectives were accomplished by performing the following specific tasks:

21 3 Specific Task (1): Development of non-resonant single element, single layer transducer and multilayer Barker code imaging transducers In this phase the focus was on the feasibility of implementation of a non-resonant transducer with maximized bandwidth using Barker code and verification of this concept based on the experimental results. The design used 56 µm thick PVDF film, resulting in the development of a new class of non-resonantly operating wideband polymer transducers that might be usable at clinically relevant frequencies (2-15 MHz) using a single imaging transducer. The performance of the transducers was verified by characterizing them in pulse-echo mode. Specific Task (2): Design, development and characterization of non-resonant multielement single layer and multilayer linear array imaging transducers A feasibility study was performed to demonstrate the Barker code concept with simple, multielement single layer and three-layer linear array transducers by extending the work mentioned in specific task (1). The linear array transducer was developed using 56 µm thick PVDF film having four elements and its characterization included (a) measurement of the pulse-echo response of individual element (b) test of the uniformity of the response and (c) recording of the spectral response of echoes from each element and element stack of the array.

22 4 Specific Task (3): Fabrication and characterization of wideband double-layer pvdf Polymer hydrophone probes and development of measurement procedure In the initial phase of this research, double layer piezoelectric polymer hydrophone probes used for acoustic measurements of imaging transducers were fabricated using identical and different thicknesses of PVDF film. Their performance characteristics were investigated by experimentally determining the frequency response, directional response, and effective aperture size at two different orientations (planes). A measurement procedure to calibrate the hydrophone probe below 1 MHz for adequate characterization of the acoustic field produced by the imaging transducers was also developed. The procedure was based on Planar Scanning Technique [1] and the results of the measurements were verified with the data obtained by using Time Delay Spectrometry (TDS) method [2, 3, 4] and also with the data provided by National Physical Laboratory (NPL), United Kingdom [5].

23 Outline The outline of the research work is schematically represented in Figure 1 and a detailed explanation is presented in the next section namely, the scope of the research. Single Imaging Probe for all diagnostic purposes Ultimate Goal Simultaneous enhancement of bandwidth and sensitivity Specific Task 3 Specific Task 2 Design & Development of Multielement (Array) Single/Multilayer Imaging Transducer Phase 3 Development of Single Element Single/Multilayer Imaging Transducer Development of Measurememt Procedure Phase 1 Fabrication and Characterization of Double Layer PVDF Hydrophone Probe Specific Task 1 OBJECTIVES > RESEARCH EFFORT Bottom Up Phase 2 Figure 1: Schematic representation of the outline of the research work

24 Scope of the research As shown in Figure 1, the research work presented here involved: (1) development of a wideband single element single/multilayer imaging probe, (2) design and development of a wideband multielement (array) single/multilayer probe as a feasibility study, (3) development of double layer hydrophone probe and (4) development of measurement technique and procedure. All four tasks of this research work are closely interrelated. The scope of the work also involved experimental evaluation of newly developed ultrasound transducers and critical review and discussion of the various results. The primary contribution of this work is (a) development and testing of nonresonant multilayer imaging transducers in pulse echo mode. Specifically, the possibility of using the Barker code technique to enhance desirable properties of imaging transducers is demonstrated; (b) design and development of multielement (array) multilayer transducer to validate the Barker code concept. (c) development of a double layer ultrasound hydrophone probe using dissimilar thickness films that resulted in simultaneous enhancement of its bandwidth and sensitivity; and (d) optimization of the planar scanning technique for calibration of the hydrophones below 1 MHz. A transmit/receive control system and preamplifier were designed and fabricated to evaluate the imaging probes. A pressing mechanism [6] was constructed and the thickness of the bond line was optimized in laminating the multilayer structure. In summary, this thesis presents a possible solution to the problems the clinicians are facing at present. These include: the difficulty in using a single imaging transducer for all

25 7 diagnostic applications and the trade off between the image resolution and penetration depth. 1.2 Motivation As noted above, the primary motivation of this research was prompted by a need to have an imaging transducer, which would provide both adequate pulse echo sensitivity and broad bandwidth, concurrently. This would allow a single scanhead to be used in clinical diagnostic practice. The existence of such scanhead would shorten the examination time, allow immediate, on-the-site optimization of image resolution, and hence improvement in diagnostic efficacy. By studying the properties of the several types of piezoelectric material, it was decided to explore the feasibility of using piezopolymer for the imaging transducer because of the reasons outlined in the following. Although conventional piezoceramic material provides good sensitivity, it suffers from a narrow bandwidth. The PZN/PT single crystal demonstrated good pulse-echo sensitivity and relatively wide bandwidth in comparison to solid (PZT) ceramic and piezocomposite [7]. However, the widest possible bandwidth could be realized by using a piezopolymer such as PVDF. Although this piezopolymer exhibits inadequate pulse echo sensitivity, by employing the multilayer structure and utilizing the concept of Barker code, acceptable sensitivity and widest possible bandwidth could be simultaneously achieved. The literature search revealed that no one has developed the Barker code array transducer for medical ultrasound applications so far.

26 8 1.3 Significance In medical ultrasound imaging, the overall system performance is mainly limited by the imaging transducer characteristics. Consequently, there is a strong desire to further improve the sensitivity and bandwidth of such transducers. Advanced signal processing techniques, which use frequency modulated functions rely upon the transducers possessing good sensitivity and bandwidth characteristics. It is well known that the axial resolution (i.e., the ability to distinguish closely related reflecting structures, which lie in different planes parallel to the probe face) of an image is related to the center frequency and bandwidth of the probe. High lateral resolution (i.e., the ability to distinguish closely related reflecting structures, which lie in different planes perpendicular to the transducer face) requires also wide frequency bandwidth besides other design parameters. The bandwidth of conventional imaging probes made of solid or composite ceramic and used in diagnostic systems is limited due to the fact that they operate at the resonance. Therefore, to optimize the image resolution and penetration depth, multiple transducers operating at different resonance frequencies are being used at present to cover the frequency range of interest during diagnosis due to the relatively narrow bandwidth of the PZT (Lead Zirconate Titanate) ceramic transducers. The PZT transducers consist of a thin disk, operated at resonance and the bandwidth is determined by the thickness resonance frequency (λ/2) of the ceramic. Because of the brittleness of the PZT material, thin disks are difficult to use during the imaging transducer fabrication procedure. As the PVDF is available as thin film, the polymer transducers can be operated at frequencies well below the resonance (thickness resonant frequency of 9 µm thick film is about 122 MHz).

27 9 The majority of medical ultrasound systems utilize real time imaging arrays and arrays operating at frequencies up to about 15 MHz in clinical practice. Beyond 15 MHz, there is a practical limitation in manufacturing of multi-element arrays using solid PZT material, as the required element dimension is difficult to achieve in the fragile piezoceramic materials. Besides, there is the associated problem with the necessity of using matching layers. Hence, there is a need to develop new classes of wideband ultrasonic transducers for medical imaging capable of operating at clinically relevant frequencies and providing simultaneous improvement of both the bandwidth and sensitivity. It is conceivable that instead of using several ceramic transducers necessary to cover the frequency range of different clinical ultrasound applications, if properly designed, a single PVDF probe with wider bandwidth could be used. Despite this advantage, with a few exceptions, PVDF material is not used in the design of imaging transducers due to its relatively low electromechanical coupling coefficient and dielectric constant, resulting in a poor transmitting efficiency. The exceptions include a few applications in the field of skin and ophthalmology, wherein single element PVDF probes are employed. So far, multilayer non-resonant PVDF array transducers have not been used in medical imaging. Accordingly, the research described here focused on a feasibility study, which involved development of a multilayer linear array polymer imaging transducer incorporating the concept of Barker code. The study resulted in fabrication and testing of a new class of imaging transducers featuring enhanced bandwidth and acceptable pulse-echo performance.

28 10 As already noted, the transducer designed once optimized, will be potentially capable of operating at any clinically relevant frequency. So, one transducer could be sufficient for most of the diagnostic imaging. The existence of such a wideband transducer would shorten patient s examination time, allow immediate, on-the-site optimization of image resolution and lead to improvement in diagnostic efficacy. In the next section fundamentals of biomedical ultrasound, medical ultrasound imaging and imaging transducers are outlined. This is done to facilitate a better understanding of the work presented. 1.4 Fundamentals of Biomedical Ultrasound Sound is the experience of the propagation of pressure through some physical elastic medium such as liquid or air. The pressure waves are generated due to some type of mechanical disturbance. The vibrating pressure waves transfer the mechanical energy in the form of wave to the medium and to objects that the wave contacts. If the vibrational frequency is very high, human cannot hear it, as the human hearing is limited up to about 20 khz. There are mammals such as bats and whales, which can sense above 100 khz to determine distance. The ultrasound is a mechanical wave with a frequency above the limit of human hearing. In general, there is a fundamental higher limit of frequency based on the closest atomic spacing in solids, which is about Hz. Using ultrasound, the interior of an object can be examined by transmitting pulses of high frequency sound waves into the object and receiving the echoes from the internal structure by the same source transducer. The time delay for the reflected echo to return

29 11 to the source will determine the distance of the reflecting structure. The pulse echo technique was originally developed for radar and sonar. In 1917 Langevin in France used the pulse echo technique to measure the depth of water in the ocean. After World War II, J.M.Reed and J.J.Wild first adopted the pulse echo techniques to image biological structure in 1952 [8]. An ultrasound transducer made with piezoelectric material translates the applied voltage pulses into high frequency sound vibration, and then listens for the echoes. The transducer converts the echoes (pressure waves) into voltage and the waveform is converted into digital representation, which is then converted into an image by electronic devices so as to be seen on a monitor. Medical ultrasound is a process, where the sound waves are bounced into the body and the reflected sound waves are captured and transformed into an image by the ultrasound machine. The resolution of the image is fundamentally limited by the wavelength of the ultrasound. The resolution increases with increasing frequency. As the ultrasound travels through the biological material, it gets attenuated due to spreading, scattering and absorption. The rate of attenuation in tissue is proportional to the frequency of ultrasound. Therefore, there is a trade off between resolution and penetration depth. The frequency of ultrasound needs to be high enough to give good resolution and low enough to detect the echoes from the deeper structures. So the constraints on frequency control the scale of the anatomy to be examined. Currently, frequencies from 1 to 15 MHz are generally used in medical diagnosis except in some special cases such as in the field of

30 12 ophthalmology and dermatology where higher frequency single element transducers are used. Transducers operating at 250 khz to 1 MHz are commonly used in many therapeutic applications. 1.5 Fundamentals of Medical Ultrasound Imaging In a diagnostic ultrasound system, the ultrasound pulses are generated by the imaging transducer containing a piezoelectric material. The transducer is used both for generating the pulses and detecting the echoes. When the piezoelectric material is subjected to an electrical voltage, it undergoes a change in dimension. If the electrical stimuli are of very short duration, the piezoelectric material experiences very rapid fluctuations in dimension. These rapid mechanical vibrations produce ultrasonic pressure waves. Conversely, if the pressure wave strikes the piezoelectric material, it causes mechanical deformations of the piezoelectric material, which produce electrical voltage. The ultrasound beam consists of train of pulses that are emitted by the probe about a thousandth of a second apart (PRF ~ 1 khz). Each pulse travels in the beam at the speed of sound and is reflected by the structure within the body. The reflected sound beam (echoes) gives information about the position of the reflecting structure. The pulse-echo information from the reflecting structure can be displayed in different ways. One of the ways is to find the distance, which is called A-Scan. In an A-scan display, the horizontal axis represents the time required for the echo to return to the imaging transducer, which corresponds to the distance between the imaging transducer and the reflecting organ. The vertical axis shows the height of the peak amplitude, which relates to the strength of the

31 13 reflected echo. When ultrasound pulses travel through the human body, they are partially reflected at boundaries between structures that differ in their characteristic impedance. Since the difference between the impedance of one tissue with the other is small, the echoes reflected by the boundaries of two tissues are faint. The boundaries between tissue and bone return strong echoes due to large difference in acoustic impedance. The echoes from the far side of an organ are fainter than the ones from the near side, due to attenuation of ultrasound as it propagates deeper into the body. The frequency of the reflected ultrasound wave from a stationary structure is equal to the frequency of the incident wave. The echo ranging information of the A scan can be displayed in the brightness modulated form called B-Scan. In a B-Scan, echo is represented by a spot of light. The brightness of the spot of light represents the strength of echo received. The position of the spot of light represents the time required for the echo to return to the transducer and shown on a time scale. The position and the brightness of the echoes are stored in the ultrasound machine and the integrated image of all the stored signals forms a representation of structure/organ in cross section. In the case of moving structures such as blood or a beating heart, the reflected wave has a different frequency than the incident wave. The shift in frequency is due to the Doppler effect. The movement of the structure towards the imaging transducer compresses the wavelength of the reflecting wave, causing an increase in the frequency. The movement of the target away from the imaging transducer lengthens the wavelength of the reflected

32 14 wave, resulting in a decrease in the frequency. The shift in the frequency is proportional to the frequency of the incident wave and the velocity of the moving structure. For Doppler shift measurements, the ultrasound is transmitted in the form of a continuous wave or quasi-continuous wave. The imaging transducer consists of two transducers, one for emitting continuous ultrasound wave towards the moving object and another for detecting the echoes from the target and also from the stationary structures surrounding the intended target. The received signal is electronically mixed with the reference frequency of the transmitting transducer. By mixing the two sound waves, four frequency components are obtained namely frequency of the transmitted wave, reflected wave, sum and difference of the two frequencies. The difference of the two frequencies alone is filtered out which is called Doppler shift frequency. If the imaging transducer is excited while aiming towards a stationary structure, no sound will be heard. When the transducer is aimed at a moving object, the diagnostician can hear a tone/sound whose frequency is equal to the Doppler shift frequency. The blood flow can be detected from the reflected or backscattered ultrasound from the moving blood cells. For the beating heart, the sound heard will vary with heartbeats. Since the Doppler system relies on continuous wave propagation, it cannot measure the distance of the target from the imaging transducer. By adopting a technique called gated transmission by time coding and using a long pulse, distance information is obtained. The state of the art ultrasound system uses imaging arrays instead of single element transducers. The array consists of several transducing elements which are electronically switched, steered and focused both in transmit and receive mode to form the high-

33 15 resolution images. A group of elements typically four or eight are generally used in transmitting the ultrasound. The size of the group of elements excited determines the size of the aperture and the shape of the transmitted sound beam. Each element of the array is excited with a time delay signal in order to achieve a focused beam. The beam former in the ultrasound machine selects the aperture, applies time delay and steers the beam. After the echo signal leaves the receive beam former, it undergoes time gain compensation (TGC), which compensates for the loss in amplitude of the signal due to depth and also for the losses due to diffraction of ultrasound as well as the attenuation in tissue. Logarithmic compression is employed to reduce the dynamic range of the signal so that both the weak and strong echoes are within the range of 40 db in order to display the image on a monitor, which has a limited dynamic range. The next stage is the envelope detection, which demodulates the echo signal and filters out the carrier signal. The demodulated signal is sent to the scan converter. The above process illustrates one scan line and the image display requires between 48 and 196 scan lines and a frame rate of about 30 per second. Besides A-scan, B-scan and Doppler imaging, the recent advances include tissue/native harmonic imaging, contrast imaging, 3 and 4 dimensional imaging, coded excitation and elastography. 1.6 Fundamentals of Medical Ultrasound Imaging Transducer Imaging transducers are constructed from a piezoelectric material so that the element will resonate mechanically in a thickness mode at its fundamental frequency of oscillation. The thickness is chosen to be half wavelength. An ideal imaging transducer should achieve highest sensitivity and penetration, optimum focal characteristics and best

34 16 possible resolution, all at low power output, conforming to ALARA (As Low As Reasonably Achievable) principle Resolution The resolution in medical imaging consists of four components namely (a) Axial resolution (b) Lateral resolution (c) Transverse resolution (d) Contrast resolution Axial resolution The axial resolution is defined as the minimum distance between two reflecting structures, which lie in different planes parallel to the transducer face, at which they may be distinguished as separate targets. The factors that contribute the axial resolution of an ultrasound imaging systems are (i) nature of the pulsing and receiving circuit (ii) signal processing (iii) transducer characteristics. The transducer parameters, which influence the axial resolution, are (a) damping (ring down) (b) frequency (wavelength) / spatial pulse length. A brief explanation about these parameters, which are pertinent to this transducer development project, is outlined below for the sake of understanding of the readers. Damping The vibration of the piezoelectric material should be damped out as quickly as possible. As long as the material is vibrating or ringing, it is not available to receive the ultrasound echoes from various reflecting structures. There is a trade off between limiting the amount of ringing to improve the axial resolution and allowing it ring long enough to provide sufficient acoustic energy in order to obtain echoes of sufficient magnitude. The

35 17 degree of damping or ring down gives a measure of the axial resolution potential. Highly damped transducer will have lower sensitivity. Frequency The frequency of the transducer determines the axial resolution potential as well as the penetration depth and lateral resolution. Increase in frequency improves the axial resolution. It also gives better lateral resolution in view of less divergence of the beam. Increase in frequency has decreased wavelength, which results in decreased spatial pulse length. Spatial pulse length The spatial pulse length represents the actual physical space occupied by the burst of ultrasound. It is measured from the start of the ultrasound burst to the 20 db ring down point. By the time the vibrations have been damped down to 20 db, if the wavefront has already traveled say, 6 mm, then the spatial pulse length is said to be 6 mm. So this first 6 mm of the tissue cannot be visualized and also the reflecting structures separated by less than 6 mm cannot be resolved. The relationship is stated as follows: increased frequency decreased wavelength decreased spatial pulse length better axial resolution Lateral resolution If the ultrasound beam is narrow enough to pass between two reflecting structures without being incident upon either one, then the beam is capable of resolving them as two distinct structures. The lateral resolution is defined as the minimum distance between two reflecting structures on a line perpendicular to the imaging beam at which they are

36 18 distinctly visualized as separate structures. The beamwidth of the ultrasound beam is the parameter, which determines the lateral resolving capability of the imaging transducer. As beamwidth increases, lateral resolution and image details deteriorate. Since the ultrasound beams from single element transducers are circular at any one point, the beamwidth is actually the beam diameter. It is determined by measuring the width of the amplitude at 6 db level. The transducer parameters that affect the beamwidth are (i) geometry of the piezoelectric material (ii) frequency (iii) focusing (iv) distance from the probe. The effect of these parameters on beam width and the consequent influence on image quality are delineated below. Geometry of the piezoelectric material Circular elements produce cylindrical beams and hence the lateral resolution capability is uniform across any cross section of the beam. The rectangular elements as used in linear arrays produce elliptical beams and hence the lateral resolution is better in one plane of the beam than the other. The element diameter affects the width and shape of the beam. As element diameter increases, beam spread decreases, resulting in very narrow beams. The far field divergence is given by the expression sinθ = 1.22λ/d, where d is the diameter of the transducer. Small element diameter which produces narrow beam is useful in examining superficial structures because reduced diameter will have reduced beam diameter in near field and increased far field divergence, which has no consequence in superficial exam. Larger element diameter will have reduced far field divergence, resulting in improvement in imaging deeper structures.

37 19 Frequency Frequency also affects the beamwidth. High frequency beams are less divergent. Increase in frequency causes an increase in the near field distance and decrease in far field divergence. The combined effect produces an effectively narrower beam. Focusing Focusing is a means of regulating the width of the sound beam and the region of maximum sensitivity. Focusing helps to improve the image, because it creates a narrower beam over a defined focal zone. The focal point can only be adjusted within the near field, while the focal zone may extend into the far field. The radius of curvature, in combination with frequency and active element diameter, determines the focal point of the transducer. The focal zones are defined as the region of sound beam where the acoustic sensitivity falls to ½ (-6 db) of the axial maximum. The selection of focal zone for an imaging transducer is based on the depth of the tissue of interest so as to put the narrowest portion of the beam in the region under study. The active element diameter, frequency and focal zone must be taken together while selecting an imaging transducer for a specific application, as each one directly affects the beamwidth and consequently the lateral resolution and ultimately the image quality Transverse resolution The transverse resolution is the directionality of the ultrasound beam in the transverse direction and it depends on the transducer s aperture, focusing and axial resolution.

38 Contrast resolution The contrast resolution is the minimum scattering level at which adjacent region can be visualized as having a varied intensity level Bandwidth In an ultrasound beam, the distance between two points at which the acoustic pressure drops to 6 db of the maximum value is defined as the bandwidth. A broad bandwidth implies a shorter spatial pulse length, assuring excellent axial resolution. A broader bandwidth results in larger area under the frequency distribution curve and therefore more energy in the pulse, which translates to more sensitivity. The broad bandwidth probe can be more precisely matched to the electrical characteristics of the ultrasound (system) machine with which it operates in order to achieve the optimum performance. Since the beam characteristics of a transducer are the sum of the focal characteristics of each frequency component in the ultrasound pulse, the wider bandwidth allows more latitude in tuning the imaging probe and controlling the frequency spectrum. By controlling the low frequency components of the pulse, far field divergence can be reduced, resulting in better image details in the far field Sensitivity The sensitivity of the imaging system is a key parameter affecting the ultrasound image quality. The sensitivity is related to the ability to detect small targets located at a specific depth in an attenuating medium. Although it is dependent on the interaction of several aspects of the ultrasound systems such as pulser/receiver, signal processing, display unit

39 21 etc., the imaging transducer plays a critical role in the system sensitivity. The transducer variables that contribute to system sensitivity are frequency, beam geometry and energy conversion efficiency. The conversion efficiency of the transducer indicates how well the transducer converts both the applied voltage into ultrasonic pressure pulse and the received echo into electrical voltage. Sensitivity can be related to conversion efficiency. The sensitivity is the products of transmit and receive efficiencies. The factors that influence the sensitivity of the imaging transducer are the nature of the excitation pulse, type of piezoelectric material, focal characteristics of the sound beam, propagating medium, distance of the target, and the nature of the reflecting object. The piezoelectric materials have d 33 and g 33 values, which are the indices of how well the material converts the voltage signal into mechanical deformations and the mechanical stress back into electrical voltage, respectively. Although it is advantageous for d 33 and g 33 constants to be large, the commonly used materials do not necessarily have both the highest d 33 and g 33 value. The other parameters to be considered are Q value, dielectric constant, impedance and capacitance. Although a higher degree of damping results in improved axial resolution, it lowers the transducer sensitivity. Damping is to be selected to give a best compromise between sensitivity and axial resolution. Focusing is a design variable, which affects the sensitivity. Focusing regulates the width of the sound beam. The sensitivity of the transducer changes with the variation of the beamwidth. As beamwidth is varied by focusing, the cross sectional area over which the ultrasound energy of the beam is distributed also varies. In the case of a non-focused

40 22 probe, the beam is wider and the energy content of the beam is distributed over a larger cross sectional area. Hence the focused transducer possesses higher sensitivity. For a given frequency and diameter, a non-focused transducer is less sensitive than a focused one within its focal zone. Proper selection of matching layer s impedance and thickness will increase the efficiency of the transducer in transferring energy during both the transmit and receive resulting in increased sensitivity of the transducer. Poor electrical matching also results in loss in sensitivity Backing material When an imaging transducer is excited with an electrical voltage impulse, the piezoelectric material oscillates at its natural resonance frequency. If there is a mismatch in acoustic impedances of the piezoelectric material and the backing material, ringing effect occurs during the pulse echo application, which results in lengthening of the pulse duration causing lesser bandwidth. If absorptive backing material is used, it absorbs the energy from the vibration of the back face of the piezoelectric material and the ringing is suppressed. This leads to shortening the pulse duration at the expense of sensitivity, because a large portion of the energy is absorbed by the backing material. Ideally, the backing material should provide high attenuation and match with the acoustic impedance of the piezoelectric material for efficient coupling. If the backing material has higher acoustic impedance, greater sensitivity is achieved because less energy is absorbed. But the transmitted and received pulses are longer due to increased ringing, resulting in reduced bandwidth. The two important parameters of the backing material are the

41 23 acoustic impedance and the attenuation. Attenuation is the loss of acoustic energy which is mainly due to two mechanisms namely (i) scattering loss and (ii) absorption loss Electrical matching/tuning Usually the power delivered to the transducer by the voltage source of the system is transmitted over a line having a characteristic impedance of 50 Ω. Electrical tuning is often used to convert the electrical impedance of the transducer to a value closer to 50 Ω. If the imaging transducer is not perfectly matched electrically to the pulser, then all of the pulse energy is not delivered to the piezoelectric element. For maximum power transfer, the input impedance of the transducer should be real and should match with the source. A shunt inductor of the value [1/ω 2 0 C 0 ] is used to tune out the clamped capacitance of the transducer. But this arrangement reduces the bandwidth. Electrical matching is done by selecting the material and tuning the probe so that there is a good electrical compatibility between the probe and the pulsar/receiver. It involves use of single inductor, RLC circuits and transformers. If there is no proper match between the transducer and the pulsar, a portion of the excitation pulse is reflected back into the pulser. The ultrasound manufacturers develop their own unique electronic circuits for electrical matching Mechanical matching (Acoustic interface/impedance characteristics) When a sound beam travels through a material and strikes an interface or boundary, the incident sound is partially reflected back into the first material and part will be transmitted into the second material. The magnitude of the transmitted and reflected signals depends mainly on the acoustic impedance of both the materials and the angle of

42 24 incidence of the sound beam. If the sound beam is perpendicular to the interface, the amount of transmitted sound increases. If the acoustic impedance of the second material is greater than the first, then more sound will be reflected at the interface. By introducing a proper choice of material and thickness as an intervening layer, higher transmission of sound energy from the first material through the intermediate layer and into the second material will result as compared to the situation where there is no intermediate layer. From the physics of wave phenomena, the optimum face thickness of the probe is onequarter wavelength of the frequency employed. Although three-quarter wavelength will have the same ability to produce phase reversal and signal reinforcement phenomena, one-quarter wavelength is normally employed to minimize the attenuation losses within the face material. The ideal facing material impedance is the geometric mean of the acoustic impedance of the piezoelectric material of the imaging probe and the load material such as human tissue. Z ML = Z PZ M, where Z ML, Z P and Z M are the acoustic impedances of the matching layer, piezoelectric material and the propagation medium respectively. The quarter wavelength layer with proper acoustic impedance provides improved transmission and reception of the ultrasound echoes. The quarter wavelength matching layer is acting as a mechanical transformer, stepping down the impedance change more gradually and thus reducing the degree of acoustic impedance mismatch at the interface of the transducer and human body. At any center frequency (f c ), 100% transmission of ultrasound energy can be theoretically achieved if the matching layer has a thickness of d = c/4f c, where c is the velocity of the sound in the material. In practice, 100% transmission of sound is not achieved using quarter

43 25 wavelength matching layer, because diagnostic pulse-echo is a short, broadband pulse, which contain a significant amount of acoustic energy, which is not at the center frequency alone. The material characteristics of the matching layer also contribute in the transmission efficiency. Since the single matching layer is acting as a frequency filter, the use of matching layer reduces the bandwidth. To alleviate the drawback of the single quarter wavelength matching layer and to achieve improved performance, the concept of incorporating a second matching layer, which is called multiple matching layer design is generally adopted. While there is a reduction in bandwidth in using single quarter wavelength matching layer, one could achieve a broader bandwidth in using two matching layer compared to the condition in using single matching layer. The broader bandwidth leads to more energy, which results in increased sensitivity. So the sensitivity of the transducer can be increased further by applying two matching layers, which have acoustic impedance values between the values of piezoelectric material and human tissue. The stepwise transition of impedances from the transducer to the tissue allows further reduction in the impedance mismatch resulting in decreased internal reflection and increased transmission and reception over a wide range of frequency. This will have impact clinically in order to obtain better penetration due to increased sensitivity, potential for using lower acoustic power output levels, and higher frequency transducer for superior resolution.

44 Acoustic energy loss There has been energy loss between the excitation pulse and the conversion of the received echoes into electrical signal. The losses occur at many places as detailed below. Transducer loss The transducer cannot convert all of the applied electrical energy into acoustic waves due to the following factors: (a) Electrical Matching: If the transducer is not electrically matched to the pulser, then all the applied electrical energy is not delivered to the piezoelectric material. (b) Thermal loss: A part of the pulse energy is dissipated in the piezoelectric material in the form of heat. (c) Damping: Some of the acoustic energy is absorbed in the backing material of the transducer. (d) Secondary modes of vibration: Although the piezoelectric material is designed to vibrate in thickness mode, it also oscillates in Radial mode which does not contribute constructively to the ultrasound beam. (e) Mechanical matching: If the quarter wavelength matching layer is not properly selected to act as a good mechanical transformer, the acoustic energy generated by the piezoelectric material is not fully transmitted to the propagating medium such as human body. Loss in transmission medium The transmission medium (tissue) attenuates the ultrasound at the rate of 0.5 db/cm/mhz. The round trip attenuation loss diminishes the ultrasound energy returning to the imaging transducer.

45 27 Loss at the interface When the ultrasound energy encounters an interface, the energy is not fully reflected since part of the energy is transmitted or scattered resulting reduced energy entering the imaging transducer. Diffraction loss When the ultrasound wave returns to the imaging transducer after reflecting from the reflector, it diverges resulting less energy striking the imaging transducer. Loss at the electrode The existence of piezoelectrically inactive gold electrode (Z gold = 51.5 MRayls) of thickness of about 1000 Angstrom on each side of the film for each element will contribute some loss due to acoustic mismatch. Conversion loss When the echoes are received by the imaging transducer and converted into electrical signals, similar losses as in the case of conversion of electrical voltage into ultrasound pressure wave (due to electrical matching, mechanical matching, damping and thermal loss) will occur. Due to the above mentioned energy loss associated with the imaging transducer, the sensitivity of the transducer is lowered. 1.7 Fundamentals of Array Transducer There is a strong desire in the ultrasound community to have array transducers with broad bandwidth and high sensitivity particularly in view of the emerging imaging modalities such as harmonic imaging. Arrays operable at high frequency are needed for dynamic

46 28 focusing and to have increased frame rate for many clinical needs such as blood flow measurements. A typical linear array transducer consists of a long piezoelectric element strip divided into a number of closely spaced rectangular elements. The space between the elements is called kerf and the distance between the centers of two elements is called pitch. The size of a pitch in a linear array element ranges from λ/2 to 3λ/2 or greater, where λ is the wavelength in the medium of propagation [9]. Although larger pitch is not ideal, it is possible to obtain quality images with arrays having higher pitch by using a wider bandwidth transducer and by operating with high frame rate [10]. In a linear array system, the physical movement of the transducer is eliminated. The deflection of the beam is done electronically. A group of elements of an array is excited in succession so that the ultrasound beam is electronically moved across the face of the transducer, providing an image similar to that obtained by scanning a single element transducer manually. The microprocessor selects a block of elements, typically four to eight, and pulses them. The group of selected elements simultaneously transmits the sound waves and each wave will propagate the same distance from the transducer at a given time. The sound waves from the elements cross each other along a line parallel to the array resulting in formation of higher amplitude pressure waves due to constructive interference. In the other directions, the waves are not in phase, which results in destructive interference. Consequently, the beam is formed in the axial direction only. The width of the elements should be larger than the wavelength of the sound in order to

47 29 have a highly directional beam. The resultant echoes are processed and the switching circuit advances the block by one element and the process is repeated. In this manner the entire array is sequentially activated. While the sequencing is repeated rapidly about 30 frames per second, a B scan image is obtained. Linear arrays have elements ranging from 32 to 512 elements [9]. By electronically focusing the beam in the plane of the scan, the position of the focus can be changed at will in the case of array transducers. By dynamic focusing, the equipment continually adjusts the focal point to correspond with the depth the pulse has currently reached. To focus on echoes as they return from deeper in the body, the equipment adjusts automatically according to the distance the echo has traveled so that on-axis echoes are always brought in phase. Because of these advantages, array probes are commonly employed in clinical practice. The linear array can be focused in azimuth plane only. By having a mechanical lens at the surface of the transducer, focusing on elevation plane (perpendicular to the imaging plane) is usually obtained, which will facilitate in determining the thickness of the slice in the imaging plane. Due to beam broadening both at the near field and at very far field, good resolution could not be obtained. To alleviate this problem, multidimensional arrays (2D) are being developed [9]. 1.8 Fundamentals of Ultrasonic Hydrophone Probe Hydrophone probe is a device used to measure the amplitude of ultrasound wave (pressure pulse) generated by the imaging transducer. This transducer converts the mechanical energy of the pressure wave generated by the imaging transducer into an electrical voltage signal and thus produces a voltage time waveform representative of the

48 30 acoustic wave pressure as a function of time. It produces an electrical voltage in response to an applied acoustic pressure at the active element. The hydrophone probe is made from piezoelectric polymer film called PVDF. It is the only FDA approved measurement tool in accurately measuring the acoustic output of all the newly developed imaging transducers before use in medical ultrasound imaging and to obtain premarket approval [11,12]. It is the key element of the measurement system, since all the measured parameters are directly influenced by hydrophone characteristics. The use of an ultrasound imaging device greatly affects the patient treatment efficacy and possible collateral tissue damage. Hence the ultrasound output from the imaging transducer needs to be adequately characterized prior to use in clinical practice. The hydrophone probe comprises an acoustically transparent membrane using a PVDF polymer and having a small, central region piezoelectrically active. The hydrophone can scan the ultrasound field to determine the acoustic field distribution, energy content, shape and size of the beam at the focal zone. It is essential to accurately measure all the characteristics of the acoustic waveform generated by the imaging transducer in order to estimate the required output parameter as stipulated by US FDA and international standard such as IEC. The next chapter deals with background relevant to this work besides information explaining the importance of the issue, and a review of the work done previously in this field.

49 31 CHAPTER 2: BACKGROUND AND THEORY This chapter examines the background and theory relevant to the scope of this research. A detailed review about the existing hydrophone probe and the need for improvement are presented. The shortcoming in the existing procedure in calibrating the hydrophone probe is addressed along with the method proposed to resolve the issue. The existing types of single element imaging transducers with single layer and conventional multilayer transducers are discussed and the concept of the proposed Barker code imaging transducer is explained. The issues involved in implementing the Barker code concept to multielement (array) transducer are presented. 2.1 Introduction There are two kinds of hydrophone probes in practice at present namely (i) membrane type and (ii) needle type. The membrane type uses piezopolymer (PVDF) film, which is stretched and glued over a supporting ring and making a small central spot piezoelectrically active. The needle type hydrophone uses either a piezo crystal/ceramic or piezopolymer on a tip of a needle structure (ex. hypodermic needle). Each type has both advantages and disadvantages in application. The use of membrane type is the acceptable device for evaluating the newly developed imaging transducers for submission for premarket approval by FDA. However, the needle type hydrophone is acceptable for use in the case of measuring continuous wave, since the membrane hydrophone will introduce standing waves in such a situation. Almost all the imaging transducers use pulsed wave.

50 Ultrasonic Polymer Hydrophone Probe Hydrophone probes are made with quartz, ceramic and polymer. Due to inadequate bandwidth and large size, quartz and ceramic materials are not suitable for many applications. Piezopolymer exhibits a higher maximum frequency response than the piezoceramic. Consequent to the discovery of strong piezoelectric properties of PVDF by Kawaii during 1969, use of PVDF in imaging transducers was reported by Ohigashi et. al and also by Foster et. al during 1979 [13]. A new form of polymer hydrophone probe for measuring the spatial and temporal distribution of pressure within the fields from medical ultrasound equipment was first developed by Shotton [14] and by DeReggi and Harris [15] during The polymer needle type hydrophone probe was first developed by Lewin [16] during 1981 and it is known worldwide as Lewin type hydrophone. The hydrophone probe should be capable of capturing both the compressional and rarefactional portion of the waveform in order to determine the peak acoustic pressure and the integrated energy in the pulse. In order to obtain accurate information about the temporal variation in the pulse waveform produced by the imaging transducer, the hydrophone probe should exhibit a flat response in terms of pressure sensitivity against the frequency over the output spectrum likely to be encountered in the frequency range used in medical imaging. Now PVDF hydrophone probes are in widespread use throughout the world as a reference and working device for reliable ultrasonic measurements over a wide frequency range. Due to intrinsic broadband properties of

51 33 PVDF, membrane hydrophones perform linearly up to a peak pressure of 3.4 MPa [17] and measurement of maximum pressure up to 100 MPa is possible using PVDF hydrophone in cases like measuring the lithotriptor shock wave pulses. The hydrophone probe operates in thickness mode with the unbacked active element, which has acoustic impedance close to that of water. The hydrophone has a broad frequency response due to the use of thinner films with a low Q factor. The sensitivity varies smoothly with frequency and the hydrophones are free from reverberation effects. Double layer hydrophone probe with thicker PVDF film has a larger reflection coefficient than thinner hydrophones. As the film thickness decreases, the voltage sensitivity drops because the sensitivity is proportional to the thickness. With decreasing thickness, the thickness resonance increases in frequency but the peak sensitivity drops. The hydrophone sensitivity depends on both the charge produced by the active element and the overall capacitance of the device. Smaller active element will produce less charge. Thinner polymer film would have lesser pressure sensitivity since the active volume of piezoelectric material would be smaller. In the field of ultrasound, both the hydrophone probe and imaging transducer always go hand in hand. One cannot be of any use without the other. Both are part and parcel of acoustic measurement system. Hydrophone probe is used in the measurement of acoustic fields produced by the imaging transducers.

52 Need for wider bandwidth hydrophone probe It is known that the polymer hydrophone probe is the only FDA approved, standard measurement tool in accurately measuring the acoustic output of all the newly developed imaging transducers before use in medical ultrasound imaging [12]. There is an increasing demand for precise measurements to be performed on the acoustic fields emitted by medical imaging transducers. The finite amplitude distortion is a common characteristic of large amplitude pressure waves, and it is especially prevalent in focused ultrasonic fields. Because of finite amplitude effects in water or tissue, diagnostic ultrasound pressure pulses can have significant high and low frequency spectral content. To record these pressure pulses accurately, the hydrophone probe must have sufficient broadband response [18]. The two types of double layer hydrophone probes commonly used have frequency response up to 20 and 50 MHz only. Harris [18] has pointed out the need for hydrophones with much broader bandwidth in order to measure the center frequency of the imaging transducer and higher harmonics generated through nonlinear propagation effects in water. The pulse parameters determined from the data obtained from the hydrophones with inadequate bandwidth showed large errors. The pressure pulse and the hydrophone are characterized by the pulse center frequency f c and the hydrophone thickness resonance frequency f h respectively. Harris has also pointed out that the ratio of f h /f c should be greater than about 10 on order to keep the error below ±5% in calculating the pulse average intensity [18]. IEC recommends that the bandwidth of the hydrophone/amplifier combination needs to be greater than 8 times that of the center frequency of the pulse [IEC 87(Co) 6, 1988]. As per FDA 510(k) [11] and AIUM [12] guidelines, it is desirable to have the ±3 db bandwidth of the hydrophone probe wide

53 35 enough to make the acoustic measurements at the upper frequency limit of 8f c in order to have accurate measurements of high frequency/wide bandwidth transducers exhibiting nonlinear distortion. Typical double layer hydrophone probes are customarily made with either micron thick or 9+9 micron thick PVDF film, which have the measured thickness resonance of about 20 and 50 MHz respectively. Hence an improved design of hydrophone probe with enhanced bandwidth capable of characterizing the imaging transducers emitting high frequency components was considered necessary. Therefore, besides fabricating and evaluating the conventional double layer polymer hydrophone probes with similar thickness of PVDF film (25+25 µm & 9+9 µm), polymer hydrophones using dissimilar thickness (25+9 µm) were developed to maximize the bandwidth and sensitivity. The hydrophones were characterized for acoustic sensitivity, frequency response, bandwidth, angular response, effective aperture size and orientation effects. The literature search revealed that the development of improved design of double layer hydrophone probe to enhance the bandwidth and sensitivity using dissimilar thickness of (25+9 µm) PVDF film was not published by any one previously. 2.3 Calibration Techniques There are mainly two types of calibration techniques namely (i) primary calibration technique and (b) secondary calibration technique. In the primary calibration, there is a distinction whether the calibration technique used is a primary technique or the technique uses a primary standard device. The primary technique uses independent measurement of

54 36 certain parameters, which are directly traceable to primary standards. The primary calibration techniques are (i) optical interferometry, which is traceable to the measurement of length [19], (ii) radiation force balance technique, which is traceable to the measurement of force [20], (iii) calorimetry, which is traceable to the measurement of temperature [21], (iv) planar scanning technique, which is traceable to radiation force [1], and (v) reciprocity [22]. The secondary calibration technique uses the substitution method, which compares the unknown value of the hydrophone to be calibrated with a known value of a reference hydrophone that has been calibrated previously. So, the substitution technique requires the use of a reference device. The calibration technique used in this work namely planar scanning technique requires a primary standard acoustic source traceable to the radiation force approach. Hence the calibration performed by using the planar scanning technique is an absolute calibration. The output of diagnostic ultrasound equipment is regulated by the Food and Drug Administration (FDA) in United States [11] and worldwide, by IEC standards [23]. Manufacturers of ultrasound equipment, when submitting application for 510(k) process, need to provide all relevant acoustic output data including the maximum acoustic pressure amplitude and pressure distribution produced by the ultrasonic device to be marketed. However, the absolute calibration data for such hydrophones are available usually at frequencies above 1 MHz. The existing IEC standard for primary calibration of medical ultrasonic hydrophones covers the frequency range from 0.5 MHz to 15 MHz [23] and a standard concerned with frequency range from MHz is under development. Harris has pointed out that the knowledge of the frequency response of the

55 37 hydrophone probes below 1 MHz is of importance and has provided evidence that inadequate low frequency response of the hydrophone/preamplifier assembly introduces distortions in the measured pressure-time waveforms generated by ultrasound diagnostic devices [24]. He also pointed out that, consequently, a significant measurement error (exceeding 30%) could be introduced in determining values of mechanical (MI) [24]. In the following publication Harris showed that the hydrophone bandwidth must extend to at least 10 times below the diagnostic pulse center frequency to minimize the error in the measurement of peak rarefactional pressure amplitude and thus Mechanical Index (see Eq.2.1, below) to approximately 5% [25]. It might be appropriate to briefly note that Mechanical Index (MI) is defined based on a theoretical assessment of the likelihood for cavitation to occur in a medium containing cavitation nuclei with a wide size distribution. Its value can be determined as [26]: MI = p 10 r 0.015fcz f (2.1) c where p r is the peak rarefactional pressure amplitude in MPa, f c is the center frequency in MHz of the acoustic source measured in water at the axial distance z in cm where the derated pulse intensity integral (i.e. the integral of the pressure squared) exhibits a maximum [26]. The derating factor fcz is intended to account for attenuation in tissue by using an estimated average attenuation of 0.3 db/cm.mhz [12]. As the portion of the pressure waveform where p r occurs is dominated by the low frequency components, it is essential that the frequency response of a hydrophone probe below 1 MHz be known. More specifically, due to the nonlinear propagation effects in water and with diagnostic imaging transducers operating frequently at center frequencies of 2 MHz and below for harmonic imaging, the spectrum of p r may contain frequencies down to

56 khz [24]. Also, a number of ultrasound devices operate below 1 MHz and their acoustic output needs to be measured or monitored. These devices include lithotripters operating at center frequencies between khz, nebulizers, ultrasound-healing devices which utilize frequencies ranging from khz, and High Intensity Focused Ultrasound (HIFU) systems which are designed with center frequencies between khz. Also, the commercially available ultrasonometer (Lunar) used for evaluating the risk of osteoporotic fracture in postmenopausal women operates at 500 khz. In addition, many of the therapeutic ultrasound units operate at a center frequency of 1 MHz and below. Hence, there is a well-defined need to have a dependable method for absolute calibration of hydrophones in the frequency range MHz. As already noted, at present, no commercially available hydrophone probes are available with sensitivity information below 1 MHz. Harris has examined the behavior of hydrophones from 0.2 MHz 2 MHz using broadband pulse technique [27]. He observed that the frequency response of the PVDF membrane hydrophones was essentially flat below 1 MHz and his results were corroborated using the swept frequency technique combined with reciprocity technique [28]. In this work the applicability of planar scanning technique was examined as an alternative method for low frequency calibration of the probes. While measurements using planar scanning technique have been reported in the megahertz range of frequencies [1, 29, 30], it appears that this technique was not explored at frequencies below 1 MHz. A brief description of the two methods used in this work is given below:

57 Planar scanning technique In the planar scanning technique, the total acoustic power produced by the source was determined by using the hydrophone to be calibrated by performing raster scan in the far field of the acoustic source and then carrying out the spatial integration of the square of the voltage measured at the hydrophone s terminal. By using the Radiation Force Balance (RFB), the acoustic output of the source was measured. By comparing the power obtained from the raster scan measurements with the power measured with the radiation force balance (RFB), the hydrophone s sensitivity was determined [31]. This method provides hydrophone s sensitivity at discrete frequencies Time Delay Spectrometry technique combined with substitution method To mention briefly, the Time Delay Spectrometry (TDS) method [2, 3, 4], which is swept frequency (rather than a single frequency) technique works in the following way: The wideband source transducer was driven by the swept sine wave signal from the tracking generator of the HP3585A spectrum analyzer with a built-in frequency offset unit and amplified by a linear broadband power amplifier. The received signal from the hydrophone was fed into the band pass filter of the spectrum analyzer, which is swept with appropriate time delay in relation to the transmitter-driving signal. The signal corresponding to the direct propagation delay between the transmitter and the receiver alone is picked up after filtering out the other spurious signals. Briefly mentioning the substitution calibration method, it is performed in the following way: The frequency response of the given hydrophone/source combination was compared to that obtained from the reference hydrophone used with the same source transducer. First the reference

58 40 hydrophone is placed in the acoustic field at the far field of the source transducer and the acoustic signal is maximized and captured into the spectrum analyzer. The reference hydrophone is removed and the hydrophone to be calibrated is replaced at the same location; the signal is again maximized and captured into the spectrum analyzer. The difference between the two spectra is added to the known frequency response of the reference hydrophone to obtain the frequency response of the hydrophone to be calibrated. The advantage of TDS technique is that it allows the end-of-cable receiving voltage sensitivity to be determined virtually as a continuous function of frequency. This feature effectively eliminates the possibility of overlooking rapid sensitivity variation of hydrophones Need for developing a new measurement procedure The acoustic output parameters of all the newly developed ultrasonic imaging transducers are to be measured before use in clinical practice. The Mechanical Index (MI), a predictor of potential bioeffects in ultrasound imaging is displayed on the ultrasound equipment, which gives an indication of the potential for mechanical damage to exposed tissues. AIUM standard stipulates that a hydrophone with a lower frequency limit of f c /20 will provide more accurate measurement of the peak rarefactional pressure, and thus enable to determine the Mechanical Index more accurately [12, 24]. Inadequate low frequency response of the hydrophone introduces error in the measurement of peak rarefactional pressure amplitude, particularly below 1 MHz [32]. One of the problems associated with assuring an adequate low frequency response for hydrophones is the lack of available calibration techniques below 1 MHz. For example, for adequate

59 41 characterization of acoustic field produced by an imaging transducer operating at the center frequency of 10 MHz, the sensitivity of the hydrophone at frequencies in the vicinity of 500 khz should have been known. Because the portion of the pressure waveform where p r occurs is dominated by the low frequency components, it is essential that the frequency response of all newly developed hydrophone probes below 1 MHz be known. The absolute calibration data for such hydrophones are available usually at frequencies above 1 MHz. No national or international standard exists describing the procedure for absolute calibration of the ultrasonic hydrophone probes at frequencies below 500 khz. Hence there is a well-defined need to have a dependable method for absolute calibration of hydrophone probes in the low frequency range (below 1 MHz). Hence, in order to fill the gap in the knowledge base, the applicability of planar scanning technique was examined in this work as an alternative method for low frequency calibration of hydrophone probes. Although measurements using the planar scanning technique have been reported in the megahertz range of frequency, this technique has not been explored at frequencies below 1 MHz 2.4 Single Element Imaging Transducer Transducer requirements Imaging transducer requirements mainly include high sensitivity and exceptionally broad bandwidth to cover the frequency range of interest with a single transducer.

60 Conventional piezoceramic imaging transducers The conventional imaging transducer used in diagnostic systems is limited due to the inherently strong resonance of the piezoelectric ceramic material. To optimize the image resolution and penetration depth, multiple transducers operating at different resonance frequencies are being used at present to cover the frequency range of interest during diagnosis due to the limited bandwidth of the PZT (Lead Zirconate Titanate) ceramic. The PZT transducers consist of thin disks of ceramics, operated at resonance and the bandwidth is determined by the thickness resonant frequency (λ/4 or λ/2) of the ceramic. Because of the brittleness of the PZT material, thin disks are difficult to fabricate Conventional polymer imaging transducers A brief outline of conventional polymer transducer using single layer and multiple layers is given below in order to have an understanding before presenting the Barker code imaging transducer. The use of PVDF in medical imaging was first reported by Ohigashi and Foster during 1979 [32] Single element, single layer transducer The single element transducer needs to be moved physically over the area to be examined and a two-dimensional image (B Scan) is formed by combining the lines of information generated. The beam is focused using a fixed lens, which is simple and reliable but has the major drawback that the focal point cannot be adjusted. The limitation of the single layer imaging transducer is that when operated with short impulse, only small acoustic energy can be generated, which limits penetration depth.

61 Single element, multilayer polymer transducer The weaknesses of PVDF having lower electromechanical coupling factor (k t = 0.20), and dielectric constant (ε S = 6) compared to PZT can be minimized by the use of multiple layers [33]. The multilayer structure is an alternative technique to deliver more acoustic output power from a voltage source drive. In a conventional multilayer transducer, N identical PVDF films of thickness l are stacked, while the films of alternating layers have reversed polarity. All layers have the same lateral dimensions, which are large compared to the thickness of the layer. The layers are mechanically connected face to face resulting electrically in parallel and mechanically in series. The transducer is resonant at the desired operating frequency where the total stack thickness is λ/4 or λ/2 wavelength. Since the layers are acoustically in series, the resonance frequency is inversely proportional to the total thickness of the stack: 1 ω 0 α Nl (2.2) where l is the thickness of the layer and N is the number of layers. So by increasing the thickness of the stack, the resonance frequency is downshifted. In the transmit mode, multilayer structure gives increased acoustic output for a given source voltage. Since each layer is sandwiched between ground electrode and a signal electrode, the electric field direction is constant with the poling direction when the voltage is applied to the multilayer structure. Therefore, the layers expand or compress uniformly. Since all layers move together, the output pressure is increased proportionally with the number of layers. The electrical power input is

62 44 P in α V 2 1, Z el α Z el N (2.3) The advantage of N layers is to decrease the electrical impedance by a factor of N. For a particular stack thickness Nl, and source voltage input V, the total voltage induced across the stack is the result of effective voltage of NV. The acoustic output power is proportional to the square of the pressure amplitude, which leads to: P out α N 2 (2.4) Barker code polymer imaging transducer The inherent drawbacks in the conventional multilayer transducer can be overcome by the proposed design, which utilizes the concept of Barker code [34]. Prior to the presentation of the principle of operation of Barker code multilayer polymer transducer, a brief description of Barker code concept is outlined below Barker code concept Barker codes are binary codes named after R.H. Barker, who originally developed the codes to synchronize signals in digital communication systems. Barker code is used for the internal arrangement of polarization pattern of piezoelectric layers, in which pulse compression is effected [35]. This property was used to implement the pulse-echo Barker code transducer. When a Barker code transducer is excited, the transmitted acoustic wave is the superposition of the acoustic wave generated by each piezoelectric layer in the assembly. While in receive, the pressure wave echoing from the reflection and entering into the transducer is the Barker coded pulse sequence transmitted by the same transducer. The signal compression is brought about by receiving these pulse trains by

63 45 the same transducer. With this method, the sensitivity of PVDF transducer is raised by pulse compression Principle of operation Medium of propagation Oscilloscope Acoustic Backing Piezoelectric Layer Reflector V(t) 3 p(x) 0 3l/c 6l/c t 0 3l x -1-1 Figure 2: Schematic representation of the Barker code arrangement The schematic representation of the Barker code arrangement is shown in Figure 2. In this design, identical elements of thickness l, electroded on both sides are stacked in such a way that the polarization pattern is chosen according to Barker code, and are operated in their thickness mode. [35, 36]. The layers are glued on to the plane surface of an acoustically matched and absorbing material to make an ultrasonic multilayer transducer having total thickness of Nl, whose frequency response essentially corresponds to that of a transducer made of a single layer of thickness l [36]. The transmit/receive switches are provided in such a way that the layers are connected electrically in parallel in the

64 46 transmitting mode and electrically in series in the receive mode. When such a transducer is excited, each individual film generates a pressure pulse where the polarities (compression or tension) correspond to the direction of the polarization and to the sign of the applied voltage. Hence a train of pressure pulses with Barker coded structure is produced in the medium of propagation [36]. At a particular time, each of the individual pressure pulses sent out from each layer will reflect back and will reach its corresponding layer. When this occurs, the voltage generated in each layer will exhibit the same polarity, thus the output voltage produced at the terminal of the transducer will be N times larger than the voltage generated in a single layer. During the rest of the time, the voltage generated in each layer exhibits different polarity and some of the voltages will be cancelled out [37]. Hence the stress wave echoing from the reflector and entering into the transducer is the Barker coded pulse sequence transmitted by the same transducer. The resulting signal compression gives an electrical output signal consisting of a dominant sharp peak and some ripples due to the general properties of the Barker code [35]. The maximum receive sensitivity gain (the peak output voltage) compared to a single layer transducer is a factor of N, which corresponds to the Barker code length of N. The other feature of the Barker code is that the peak side lobes are less than or equal to a factor of one in magnitude. The ratio of the amplitude of the main peak and the side lobes is N:1 [36]. The side lobes, which contain the minimum energy, are uniformly distributed. Moreover, the higher receive voltage minimizes the effect of preamplifier noise and the diagnostician can select a lower receive gain setting in the system, which results in increased signal-to-noise ratio (SNR).

65 47 In contrast to a conventional piezoceramic [38] and composite material transducer in which the bandwidth is determined by the thickness resonance frequency of the active element and operates at resonance, the piezopolymer transducer is designed to operate at frequencies below the resonance [39]. In the very near field, at which the distance is very small compared to the radius of the multilayer transducer, the received signal should be proportional to the autocorrelation function of the Barker coded pulses. As distance increases, the time-derivative of the autocorrelation function would be received [35]. In the Barker code multilayer approach, bandwidth is determined by the thickness of the single PVDF film and the sensitivity is determined by their overall thickness. Hence, the pulse-echo sensitivity is increased compared to single layer structure while retaining inherent wideband characteristic of PVDF film. The Barker code concept was originally introduced to the single element transducers by Sung in 1984 [35] and Platte in 1987 [36]. Zhang [39] had extended their work by carrying out the modeling and conducting performance evaluation of the Barker code transducers in transmit-receive mode. Zhang s thesis work covered on the experimental data using two Barker code transducers, one as a transmitter and another as a receiver. In contrast, this research work focused on implementing the Barker code concept in pulseecho mode, both in the case of single element and multielement (array) polymer transducer, which is the practical application in ultrasound imaging. The development of Barker coded polymer array probe in this research work for medical ultrasound application was the first one of this kind and no one has developed this type so far.

66 Multielement (array) Imaging Transducer The lateral resolution at any plane is limited by the depth of focus. The depth of focus is defined as that region where the geometric shadow of the rays is equal to the diffraction limited size and is given by [40]: d f λf = d λf i.e. = 2 d 2 2λf (2.5) d 2 2 = 2 where is the distance on either side of focus and thus 2 is the total depth of focus. The width of the main central lobe is determined by the width (X) of the array. If the width of the element (W) is an appreciable percentage of the combined size of width of the element and the kerf (S), then the side lobes are automatically attenuated significantly. By using large number of elements, where X/S is a relatively large number, the side lobes can be reduced but at the expense of increased complexity [40]. The array can be focused in the near field to obtain the diffraction limited resolution by using a lens in the X, Y plane and the focal length in such case is given by [40]: R f = ( ) n 1 n 2 (2.6) where, R is the radius of curvature of the lens surface and (n 1 -n 2 ) is the difference in the refractive index between the lens material and its surround. The lens is required because electronic focusing focuses the beam in the direction perpendicular to the elements only. The resolution of an array is usually defined as equal to λ f n, where f n is the focal number defined as Z/D, while D is the total length of the array (aperture) and Z is the focal distance.

67 Side lobes and grating lobes The side lobes are present next to the main lobe of the response. If the diameter of a single element transducer is large, a directional beam will be formed due to predominance of the constructive interference effect. There are weak areas of constructive interference in some other directions at certain angle to the main beam resulting in low intensity subsidiary beams, which are called side lobes. In the case of array transducers, they are called grating lobes, which appear due to the discreetness of the array. The side lobes and grating lobes introduce uncertainty in the origin of the echoes. The ultrasound image is based on the display of echoes from the reflecting structures due to the main beam only, which should be narrow both in transmission and reception. The side lobes which occur at certain angle due to constructive interference in view of the regular spacing of the element is called grating lobes, which is given by [9]: φ g 1 nλ = sin (2.7) g where n is an integer number 1,2,3, g is the pitch and φ g is the angle at which the grating lobe occurs. For the grating lobes to occur at angles greater than 90 0, the pitch of the array (g) should be smaller than λ/2. The amplitude of the grating lobe is determined by the width of the element. If the width is small, the amplitude of the grating lobe will be larger compared to the main lobe. The angle of grating lobe and the main beam is given by: sinϕ = sinθ + p λ (2.8) where ϕ is the angle of grating lobe, θ is the angle of main lobe, λ is the wavelength and p is the pitch. The effect of side lobes will be the distortion of the image. The side lobes

68 50 are located approximately at a distance 1.5λZ/D from the center of the response, where Z is the focal length and D is the aperture length. The grating lobes are spaced in the x direction by a distance of λz/d, where d is the center-to-center spacing of the elements Cross-talk effects The cross-talk effect is an important aspect in evaluating the performance of the array transducer, because it affects the ring down through delayed signals from the adjacent elements and angular dispersion by increasing the effective element size. In a linear array transducer, the voltage generated by the receiving element should be related to the incident pressure wave on the receiving elements only or to the excitation voltage applied to the corresponding elements only. But in practice, if a single or a group of elements are fired during transmit mode, not only the specified elements generate an acoustic signal but also it causes the adjacent elements to generate the acoustic signal. In the same way, the echoed pressure pulses from the reflecting target, not only causes generation of electrical voltage from the respective transmit elements, but also some additional signal from the adjacent elements. This phenomenon of generation of additional pressure wave/voltage unconnected with the excitation events of the elements of an array is called cross talk effect and the mechanism of the cross talk is usually classified mainly as electrical and acoustical. The cross talk effect will arise due to interfering waves propagating along the inter element. The cross talk leads to loss in sensitivity due to large parasitic capacitance within the array structure. If the cross talk effect is sufficiently strong, then the directivity pattern of the individual element will differ from the ideal one.

69 Electrical cross talk The electrical cross talk is related to the energy coupling between adjacent array elements, which may be due to the generation of fringe electric fields along the edge of the electrodes and influencing the neighboring elements. The cross talk may also be caused by parasitic capacitance between the adjacent elements Acoustical cross talk Acoustical cross talk arises due to lateral acoustical coupling among the adjacent elements by the surface wave and Lamb waves. If a longitudinal wave reaches the surface of a medium with an angle of incident other than normal, then a surface acoustic wave is generated and it will propagate along the surface of the medium. The resonance frequency is determined by the periodic loading of the surface by the array elements. So, when the reflected pulse wave is incident on the surface of an element in an array, the surface wave will be generated and travels to the adjacent elements. A Lamb wave will be generated if the incident wave hits a thin layer between two other layers, and it will propagate along the thin layer. The Lamb wave causes cross coupling between adjacent elements similar to surface wave. The surface wave and Lamb wave emerge due to parasitic modes of vibration and cause additional energy dissipation in an array.

70 Electromechanical coupling Since the array elements are in the same sheet of PVDF, the mechanical vibration caused by the incident pressure wave on a particular element may be coupled to the neighboring elements, which will cause output voltage generated across these elements. Similarly, in transmit mode, the applied voltage on a particular element not only generates longitudinal strain in the thickness direction but also in the transverse direction due to transverse piezoelectric coupling. This vibration may induce voltages in the neighboring elements. Cross talk effects in imaging transducers are not desirable. They introduce a phase shift to the array element causing loss of phase coherence, resulting in increased insertion loss and loss in resolution affecting the image quality. The next chapter deals with the critical analysis of material properties and suitability of piezoelectric materials for the use in imaging transducers.

71 53 CHAPTER 3: PIEZOELECTRIC MATERIALS A brief description of different properties of various piezoelectric materials used in imaging transducers are presented here in order to arrive at a decision about the suitability of using a particular material for the imaging transducer. 3.1 Introduction The phenomenon that the application of voltage on a piezoelectric material makes changes in its physical dimension and the application of pressure on a piezoelectric material produces electric field is called pressure-electric effect or piezoelectric effect. The piezoelectric effect was first discovered by Pierre and Jacques, French Physicists during 1880 [9]. The piezoelectric material consists of randomly placed numerous electric dipoles. By polarizing the piezoelectric materials, the dipoles are aligned uniformly in the direction of polarization resulting in change in the thickness of the material. One method of polarization is carried out by applying a strong electric field along the direction in which the piezoelectric effect is preferred and simultaneously heating it just above its Curie temperature and then cooling it slowly to the room temperature by still keeping the applied electric field. 3.2 Material Properties The piezoelectric material has several important properties, which will have impact on the performance of the imaging transducer. A brief description of some of the relevant

72 54 properties is given below in order to have an understanding as to how each parameter affects the performance of the imaging transducer Piezoelectric stress constant (e 33 ) The stress produced in a piezoelectric material upon the application of a unit electric field, while being clamped, is called the piezoelectric stress constant e. The unit is Newton/V-m or Coulombs/m 2. The larger the magnitude of the stress constant, the greater the coupling between elastic and electrical effects Piezoelectric strain constant (d 33 ) The piezoelectric crystal is defined in x, y, z axis or direction and indicated by 1, 2, 3. The polarization direction is the one in which a stress produces an electric field. The piezoelectric crystal when it is cut with its surface perpendicular to x-axis, is called x cut and so on. The z direction is normally used to indicate the polarization direction of a crystal. The strain produced in the z direction while applying the electric field with no external stress is called piezoelectric strain constant, d Transmitting constant The strain produced per unit of applied electric field with no external stress is known as the transmitting constant, d. The unit is Coulombs/newtons or meters/volt. The relationship between stress constant and transmitting constant is e = c E.d, where c E is the elastic constant of the material.

73 Receiving constant For a unit applied stress, the open circuit electric field produced is called the receiving constant, g. The unit is V-m/Newton. This relates the electric field to the stress as g h = d h /ε T, where d h is the hydrostatic piezoelectric coefficient and ε T is the free dielectric constant Dielectric permittivity (constant) The dielectric constant determines the electrical impedance of the piezoelectric material. A large dielectric constant is important for enabling a good electrical impedance match to the system electronics. While the piezoelectric material is clamped in such a way that it cannot move while applying a voltage or if there is no strain, the measured dielectric constant is denoted as clamped dielectric constant, ε S. While the piezoelectric material is kept in such a way as to move freely without any restriction, the measured dielectric constant is denoted as free dielectric constant, ε T. The relationship between the transmitting and receiving constant is defined as ε T = d/g. If the dielectric constant of a piezoelectric material is low, then the input electrical impedance will be high, necessitating higher drive voltage to generate the required output acoustic power Electromechanical coupling coefficient The electromechanical coupling factor determines the efficiency of the piezoelectric material in converting the mechanical energy into electrical and vice versa in a single thermodynamic cycle. A failure to convert a large fraction of energy in a single cycle will result in loss of sensitivity and loss of bandwidth if the energy is converted in a later

74 56 cycle resulting in spreading the response time. High coupling factor leads to improved axial resolution, broader bandwidth and higher sensitivity. The electromechanical coupling coefficient (ECC), which is a measure of the performance of a material as a transducer is expressed as [9]. T ε ECC = 1 S ε (3.1) The piezoelectric material can have resonance in both lateral and thickness modes depending upon its dimension. The thickness mode ECC is designated as k t, which is the measure of the ratio of the transferred electromechanical energy in the thickness mode to the total input energy. This electromechanical coupling factor (k t ), which is a dimensionless quantity, is one of the important factors in obtaining higher acoustic output power, which gives the better measure of the acoustic radiating power of the transducer. The pulse echo response of an imaging transducer depends on k t, while larger k t contributes to the increase in response amplitude. The larger the value of k t, the higher the piezoelectric coupling between the acoustic and electrical properties of the material Dissipation factors Mechanical loss tangent The larger mechanical loss decreases the response time resulting in improvement in the sharpness of the response, although it reduces the response amplitude considerably. The mechanical quality factor Q m is the reciprocal of the mechanical loss tangent; i.e. Q m = 1/tanδ m.

75 Electrical loss tangent With increase in electrical loss tangent, conversion loss will increase due to dissipation of energy within the transducer. The dielectric loss will not affect the response time but the amplitude will be decreased slightly. The electrical quality factor is the reciprocal of the electrical loss tangent i.e. Q e = 1/tanδ e. A minimal electrical loss tangent reduces the amount of heat generated by the material on transmit and improves the signal to noise ratio during receive mode Acoustic impedance The acoustic impedance matching of the piezoelectric material of the imaging transducer matching with that of the propagation medium such as water or tissue is a very important and sensitive factor on the response time of the waveform. The response time is strongly dependent on Z. When operating into water or tissue, low impedance is a distinct advantage as it results in relatively low reflections at the interface of the transducerwater/tissue. This will lead to broad bandwidth without requiring matching layers. 3.3 Piezoelectric Materials for Imaging Transducers Piezoceramic Although piezoelectricity was discovered in 1880, the fact that piezoelectricity could be observed in ceramics was not published until The piezoelectric properties were discovered in barium titanate in A family of ceramics made from lead zirconate and lead titanate are collectively known as PZT.

76 58 Lead Zirconate Titanate (PZT) has been used in medical ultrasonic transducer design since the 1970s. The advantage of PZT is the high electromechanical coupling coefficient in thickness mode (k t = 0.50), high relative dielectric constant (600) and low mechanical loss tangent (tanδ m = 0.004), and dielectric loss tangent (tanδ e = 0.002). For medical ultrasound transducers, the electromechanical coupling factor k t is one of the most important factors, which characterizes the energy conversion in the thickness mode vibration, resulting in large acoustic power. Since the pulse echo response amplitude strongly depends on k t, the larger k t contributes to the increase in the response amplitude to a great extent. The major disadvantage of PZT is its brittleness. Ceramic transducers have struggled with reproducibility, and fabrication difficulties. Because of the high acoustic impedance of piezo-ceramic (34 MRayls) compared to human tissue (1.7 MRayls), there is an acoustic mismatch. Because of this mismatch, acoustic waves leaving the ceramic and entering the water/tissue are highly reflected causing reverberations in the ceramic leading to very sharp resonant peak, resulting in narrow bandwidth. It also leads to an impulse response ringing for several cycles, and inefficient power transfer. To improve these characteristics and increase the efficiency of the transducer, matching layers having intermediate impedances need to be placed between the ceramic and water/tissue, which further limits the bandwidth. The typical bandwidth attainable with PZT is about 60%. PZT could not be used in imaging transducer above 15 MHz due to difficulties in making and working with the very thin plate, which is very brittle.

77 Piezocomposite By combining the piezoelectric ceramic with a piezoelectrically inert polymer, many composite configurations are made, which provide new properties that could not be realized with either one alone. There are three types of piezocomposite that are commonly used in imaging transducers. They are (a) 1-3 piezocomposite (b) 2-2 piezocomposite (c) 0-3 piezocomposite. In a piezocomposite, one component is piezoelectrically active material and the other is an inert (insulator) material. The term connectivity means the number of directions through which the material is continuous. The X, Y, Z axes are also called 1, 2, 3 directions. If the ceramic connectivity is 1, and the polymer connectivity is 3, then the resulting material is called 1-3 composite The 1-3 piezocomposite consists of arrays of piezoelectric ceramic rods arranged in a polymer matrix. The 2-2 composite have arrays of piezoelectric ceramic strips separated by inert polymer strips or air. The 0-3 piezocomposites have piezoelectric ceramic particulates dispersed in a polymer matrix. The low-density polymer assures that the piezoceramic has a high electromechanical coupling coefficient. Since the density and velocity of sound of the composite is reduced, the acoustic impedance of the piezocomposite is lower compared to ceramic. In a piezocomposite, only the thickness mode is excited in order to obtain the compact

78 60 impulse response. The composite material can be made flexible, having electromechanical coupling coefficient higher than PZT. A mixture of polymer material and PZT (1-3 composite) is used to modify specific feature of the ceramic for applications in the 1 to 7.5 MHz range [9]. These 1-3 composites, consisting of small PZT rods embedded in a low density polymer having a lower acoustic impedance (4 to 25 MRayls) than conventional PZT, better match the acoustic impedance of human skin. Besides low acoustic impedance, the composite material has the advantage of mechanical flexibility, improved electro-acoustic efficiency with an electromechanical coupling coefficient higher than that of PZT. The piezocomposite makes a short pulse with increase in amplitude compared to PZT resulting in increased bandwidth. The typical bandwidth attainable with piezo composite is about 80% [33] Single crystal Ever since the discovery of the piezoelectric effect in barium titanate in 1952, only the ceramics are used for the piezoelectric activity, primarily due to high coupling factor of PZT ceramic (> 0.75) over any known crystal. Besides, the single crystal of usable size could not be grown until 1982, when Kuwata et. al discovered high piezoelectric coupling (> 0.9) in a solid solution of lead zirconate niobate (PZN) and lead titanate (PT) known as PZN/PT [41].

79 61 Lithium Niobate (LiNbO 3 ) A typical piezocrystal is Lithium Niobate (LiNbO 3 ), which is one type of single crystal, which has a high thickness mode coupling coefficient (k t = 0.49) and low relative clamped dielectric permittivity (ε S /ε 0 = 28) [42]. The low relative permittivity leads to increase in electrical impedance. The high acoustic impedance of LiNbO 3 (34 MRayls) poses problems as PZT in acoustic impedance matching. The measured 6dB bandwidth varied from 50% to 70% [42]. Potassium Niobate (KNbO 3 ) Potassium Niobate (KNbO 3 ) possesses a low dielectric constant, and high coupling coefficient k t = 0.68). The high coupling factor leads to improved axial resolution and high sensitivity. A bandwidth of 64% was achieved [43]. PZN/PT and PMN/PT Single Crystal Single crystals are made from the solid solution of Pb(Z n1/3 Nb 2/3 )O 3 abbreviated as PZN and PbTiO 3 abbreviated as PT. The PbTiO 3 is a typical piezocrystal. Kuwata et.al has reported a very high electromechanical coupling coefficient (k 33 = 0.92) in PZT/PT single crystal and the imaging probe fabricated using PZN/PT 1-3 composite revealed a bandwidth ranging from 74 to 141% [7]. Since the sound velocity of PZN/PT is less than 70% of PZT, PZN/PT offers lower acoustic impedance. Researchers have also developed a single crystal with 0.91Pb(Z n1/3 Nb 2/3 )O PbTiO 3 [41].

80 62 There are several issues in using single crystals for imaging transducers. There is a practical difficulty in growing the crystal without defects and consistent properties within a piece, within a batch and from batch to batch. The other issues are maintaining proper crystal orientation, avoiding chipping and cracking, achieving good electrode and adhesion, problems of depoling and repoling etc. Although these materials can transmit a large amount of energy, they cannot transmit high bipolar voltage or voltages with incorrect polarity. Once the above practical problems are resolved, single crystals may dominate the ultrasound industry due to their capability of giving simultaneous broad bandwidth and high sensitivity, which is the prime demand of the ultrasound world Piezopolymer PVDF is a semicrystalline polymer. It was patented by Ford and Hanford in 1948 [44]. In 1969, Kawai reported the discovery of a large remnant polarization in oriented films of PVDF [44]. PVDF exhibits non-polarizable form after cooling from the melt. In order to make it piezoelectric, it is to be converted into polarizable form by stretching the film to approximately 4 to 5 times its lateral dimensions. The lateral stretching induces a high degree of molecular orientation in the direction of stretching. The films can be stretched in both lateral direction called biaxially stretched and also in only one lateral direction called uniaxially stretched. For uniaxially stretched PVDF film, the lateral strain constant in the direction of stretch is high (i.e. d 31 >> d 32 ). For the biaxially stretched film, the lateral strain constant is almost equal (i.e. d 31 = ~d 32 ). The stretching changes the mechanical properties and density of the material. The dielectric and piezoelectric properties of PVDF are temperature dependant. While the electromechanical coupling

81 63 coefficient k t is independent of temperature, the dielectric properties and mechanical loss tangent are temperature dependant. When used in imaging transducer, the center frequency and bandwidth vary with temperature. The acoustic velocity of PVDF decreases approximately linearly with temperature. So the acoustic impedance and the resonance frequency of the imaging transducer made with PVDF decrease similarly with temperature. PVDF is highly flexible and thus allows application on curved structures. It can achieve large strain because of its high depolarization field and high field strength against electrical breakdown. One of the polymers used in transducer application is polyvinylidene fluoride (PVDF), which is semicrystalline. PVDF is commercially available as thin films ranging from µm thick except in special cases where much thinner/thicker films are made. Owing to poling difficulties, the thickest available film is usually 110 µm except in special cases where higher thickness films are made. The advantages of this material are that it is wideband, flexible, and inexpensive. Generally, the use of matching layers limits the bandwidth. Because the acoustic impedance of PVDF (4 MRayls) is close to that of water (1.5 MRayl) and muscle (1.7 MRayl), the reflection at the interface between imaging transducer and the medium of propagation is minimized. Hence matching layers are not necessary, thus preserving the intrinsic broadband property (mechanical quality factor Q m = 1/tanδ m ) of PVDF. The property of large mechanical internal loss (mechanical loss factor tanδ m = 0.10) of PVDF decreases the response time considerably and improves the sharpness of the response, thus contributing to widening the frequency bandwidth of the polymer transducers [45]. The short impulse response leads to high

82 64 spatial resolution capabilities in medical imaging. Their low radial/lateral mode coupling reduces the effects of edge waves and near field distortions, which is typical of piezo ceramic transducers. Research has shown that planar PVDF transducers can produce plane wave performance in the acoustic near field, which is superior to that of piezoceramics [44]. The disadvantages are (i) that it has a low thickness mode electromechanical coupling factor (k t = 0.20), which governs the transmitting efficiency. The square of this number is a measure of the ratio of the transformed electromechanical energy to the total input energy. Therefore, with equal amount of available input electrical power, the PVDF transducer will generate less acoustic output compared to PZT. Larger k t contributes to the increase in the response amplitude to a great extent. (ii) It has large dielectric loss factor (tanδ e = 0.25), which will result in an appreciable amount of electric power being dissipated in the transducer itself. (iii) It has low relative dielectric constant (6), which will result in very high input electrical impedance in comparison to that of a PZT. Therefore, to generate equal amount of output acoustic power, a higher drive voltage is required for PVDF transducer. The capacitance of the transducer is proportional to the dielectric constant, which implies that the voltage developed across the transducer in the receiving mode is inversely proportional to its dielectric constant. Low dielectric constant is advantageous in terms of developing large voltage signal.

83 Modes of Vibration Although the performance of the imaging transducer depends mostly on the kind of piezoelectric material, other factors such as frequency, backing material, matching layer and the geometry of the active element are also have their influences on the efficiency of the transducer. The acoustic energy from the piezoelectric material is due to the vibrations of material. The mode of vibration is related to the geometry of the piezoelectric material. Normally the vibration is required in one direction only. The vibration in other directions if any will cause artifacts and energy loss. Usually, there are three modes of vibration namely (i) plate mode (ii) thickness mode and (iii) bar mode. The schematic illustrations of three important vibrational modes are shown in Figure 3. l t l w t w t l w Plate mode w, l >> t Thickness mode l >> w Bar mode t >> l, w Figure 3: Geometry of three important vibration modes

84 66 Plate mode In the plate mode, the width (w) and length (l) are much larger than the thickness (t) of the piezoelectric material. If the vibration is in the width direction, then it is called lateral mode. Thickness mode In the thickness mode, the length (l) is much larger than the width (w) and thickness (t). For optimum efficiency, the ratio of width (w) and thickness (t) should be smaller than 0.7, if the thickness vibrational mode alone is required. Bar mode In the bar mode, the thickness (t) is much larger than the length (l) and width (w). In all the vibrational modes, the poling is done in the thickness direction and the metalization is normal to the poling direction. The velocity in the piezoelectric material is related to its elastic properties. The acoustic velocity is different for different modes of vibration. The piezoelectric ceramic used should satisfy certain relationship between length, width and thickness in order to have only one of the vibrational modes. If the prescribed relationship is not maintained, then there will be more than one vibrational mode, which results in poor performance. All imaging transducers are made to operate in thickness mode. 3.5 Summary In order to achieve broad bandwidth and high sensitivity simultaneously, the following properties of the piezoelectric materials should be considered: (a) coupling factor (b)

85 67 dielectric constant (c) losses (d) depoling temperature and (e) velocity. The depoling occurs whenever the piezoelectric material approaches the Curie temperature. The piezoelectric materials with high velocity give high impedance. The various parameters of the piezoceramic and piezopolymer materials are compiled from various publications and given in Table 1 for easy comparison. Table 1. Comparison of piezoelectric material parameters SL No. Material Property PiezoCeramic (PZT5H) 1 Transmitting Constant: d (10-12 C/N) K Receiving Constant: g (10-2 Vm/N) K Electromechanical coupling coefficient: k t K Free dielectric constant: ε T (10-11 F/m) K Clamped dielectric constant: ε S (10-11 F/m) L Acoustic velocity: c (m/sec) K Density ρ (kg/m 3 ) K Acoustic Impedance (Z) kg/m 2 -s*10 6 P Electrical loss tangent: tanδ e L Mechanical loss tangent: tanδ m L Curie Temperature 0 C. K Typical 6dB fractional bandwidth K 60% PiezoPolymer (PVDF) K 15 K 14 F ( ) K 9.7 B 5 B 2200 O 2260) O 1780 O 4.02 O 0.25 O 0.1 K 100 D >150% Note: These data were compiled by the author from various sources of literature: K From[9], O From [45], B From [44], F From [32], D From [46], L From [47 ] To compensate for the high electrical impedance of the piezopolymer, use of multilayer structure, passive tuning, placing an impedance transformer or active preamplifier near the element are the possible options.

86 Conclusion By studying the properties of the several types of piezoelectric material, it was decided to explore the feasibility of using piezopolymer for the imaging transducer because of the following reason. The piezoceramic provides good pulse echo sensitivity but relatively narrow bandwidth. Single crystal demonstrates best sensitivity and wide bandwidth compared to piezoceramic. But the piezopolymer provides the highest bandwidth but poor sensitivity. By employing multilayer structure and utilizing the concept of Barker code, high sensitivity and highest bandwidth can be simultaneously achieved using piezopolymer. The next chapter deals with the development of double layer polymer hydrophone probe, which not only aimed in achieving simultaneous improvement of bandwidth and sensitivity but also provided experience in fabricating complex multilayer Barker code transducer as described in chapter VI.

87 69 CHAPTER 4: DEVELOPMENT AND CHARACTERIZATION OF IMPROVED DESIGN OF POLYMER HYDROPHONE PROBE A detailed description of the development process, performance evaluation, results and discussion about the newly developed double layer polymer hydrophone probe is presented in Appendix A. However for the purpose of continuity for the readers and sake of completeness, a brief description of the characteristics of the double layer hydrophone probe is presented in this chapter. 4.1 Introduction As a part of research oriented product development work, several wideband double-layer ultrasonic polymer hydrophone probes with different thicknesses of PVDF film having different active diameters were fabricated and tested for several of its performance characteristics such as acoustic sensitivity, bandwidth, frequency response, angular response, effective aperture size and orientation effects at Perceptron, Inc. in order to explore the fundamental improvement in the existing design [48, 49]. The hydrophone probes were tested at the in house laboratory. The development of this polymer-based hydrophone is specifically intended in evaluating the diagnostic ultrasound imaging transducers.

88 Synopsis Acoustic sensitivity The sensitivity of a double layer hydrophone is determined by the resonance of the membrane film, and the electric and piezoelectric properties of the PVDF film. The sensitivity of a particular hydrophone is determined by its active area and the capacitive loading of the element, electrode leg and connecting cable. The evaluation of the performance characteristics of several hydrophones revealed that larger spot size hydrophones exhibited better sensitivity Frequency response and bandwidth The double-layer hydrophone probes made by employing 25 µm and 9 µm thick PVDF films having 0.4 and 0.6 mm of geometrical spot diameter were tested for their performance. The results of the randomly chosen hydrophone probes were also verified at National Physical Laboratory (NPL), UK. The frequency response of a double-layer hydrophone probe fabricated with (9+9) µm thick PVDF film measured at NPL showed that the sensitivity is being constant to ±3 db between 1 to 50 MHz. In the case of hydrophone probe fabricated with µm thick PVDF film, the sensitivity variation in the frequency range from 1 to 20 MHz is within ±3 db of the mean value for all frequencies. The double layer polymer hydrophones developed using dissimilar thickness (25+9 µm) were demonstrated higher bandwidth compared to that of hydrophones made using similar thickness (25+25 µm) of PVDF film. A flat frequency response up to 25 MHz has been demonstrated.

89 Angular response The angular responses of several double-layer PVDF hydrophones of different thickness of PVDF film and various geometrical spot sizes were obtained by rotating the hydrophone in the far field of the transducer s plane wave and measuring the hydrophone s response at some angle of rotation. The angular response and the frequency dependency were studied Effective aperture size The effective diameter of the hydrophone was determined for different thickness and diameter hydrophones at varies frequencies as the determination of effective diameter is necessary for applying the spatial averaging corrections [50] while reporting the acoustic measurement results of medical ultrasonic transducers Orientation effects To study the apodized behavior and asymmetry of the sensitive region of the hydrophone probes, angular response of several hydrophone probes of different kinds, various geometric spot sizes and different thicknesses of PVDF film were measured in two rotational axes perpendicular to each other. The effective diameter of the hydrophone probe was determined from the measurement of its directivity pattern. From the results, it was seen that the effective diameter of the hydrophone varied significantly with reference to axis of rotation. There was a marked asymmetry in the effective diameter measured along the two orthogonal axes. The cause and effect of such asymmetry were investigated.

90 Conclusion This improved design of double-layer hydrophone probes to enhance the bandwidth using dissimilar thicknesses of (25+9 µm) PVDF film developed in this work was the first one of this kind and no one had developed this type previously. The new design hydrophone exhibited higher bandwidth compared to µm thick film and possessed better sensitivity and robustness compared to 9+9 µm thick PVDF film. This experience helped the furtherance of understanding of the behavior of polymer probes, as explained in the previous sections. The drop method of adhesion procedure used in fabricating the hydrophone was adopted in developing multilayer imaging transducers after optimizing the glue thickness. The next chapter deals with the new measurement technique developed for calibrating the double layer hydrophone below 1 MHz, which is essential in accurately characterizing the newly developed imaging probes before use in clinical practice.

91 73 CHAPTER 5: DEVELOPMENT OF CALIBRATION PROCEDURE A detailed description of the development of a new measurement technique along with the results and discussion are presented here. Development of a new measurement technique is essential in view of the non-existence of such a procedure in calibrating the hydrophone below 1 MHz, which is important in characterizing the newly developed imaging transducers for assessing the potential bioeffects while in clinical use. 5.1 Introduction The primary motivation of this work was to develop and optimize a rapid and reliable ultrasonic hydrophone calibration procedure suitable for use in the frequency range from 100 khz 1000 khz and to investigate the feasibility of using a planar scanning technique to determine the hydrophone s sensitivity in this frequency range. The outcome will facilitate establishing the foundation for the implementation of low frequency calibration technique for diagnostic and therapeutic ultrasound equipment, including High Intensity Focused Ultrasound (HIFU) transducers. 5.2 Calibration Setup Two piezoelectric (PZT) transducers (Panametrics and KB-Aerotech) were used as acoustic sources; Panametrics source operated at the center frequency of 0.5 MHz and the KB-Aerotech s at 1 MHz. The results were obtained in the following way. The source transducer was placed in a tank containing deionized and degassed water at room temperature with the transducer face oriented downwards. The use of deionized water is advisable to minimize the effect of possible corrosive interaction between the water and

92 74 hydrophone housing and metal electrodes. IEC standard recommends that the water conductivity should be less than 10 µsiemens/cm [23]. The degassed water was used to minimize formation of micro bubbles on the hydrophone and source surfaces, which not only perturb the ultrasound field during measurements but also lead to cavitational effect. An automatic XYZ scanning system, controlling the hydrophone and the source transducer movements through stepping motors was used. The hydrophone under test was supported by a holder featuring independent translations in the X, Y and Z directions. The source was also mounted on a XYZ manipulator and could be tilted and adjusted about the vertical Z-axis. The oscilloscope and stepper motor controller were interfaced to a personal computer through an IEEE-488 bus for capture and storage of the hydrophone signals [51]. During the calibration, the source transducer was excited with a 10 cycles tone burst generated by a HP 8116A signal generator and amplified by a linear wideband power amplifier (ENI 350L). Tone burst transmission was used to avoid the undesirable effects of standing waves and multiple reflections from the hydrophone and the walls of the tank. Nonlinear propagation effects were avoided by keeping the transmitter driving voltage sufficiently low so that the second harmonic of the received signal was at least 30dB below the fundamental (center) frequency. The 0.5 MHz source transducer was driven at 0.3, 0.35, 0.4, 0.5, 0.6 and 0.65 MHz whereas the 1 MHz transducer was driven at 0.65, 0.7, 0.75, 0.8, 0.9 and 1 MHz to cover the whole frequency range considered. Thus at the frequency of 650 khz, it was possible to obtain data produced independently by the two sources. As evidenced from the results, the 650 khz data are in good agreement (to

93 75 within ± 0.7 db), which further enhances confidence in the robust nature of the planar scanning technique. 5.3 Calibration Procedure Initial alignment procedure One of the double layer hydrophone probes fabricated by me, which has a preamplifier, was calibrated at the National Physical Laboratory (NPL), UK at 0.25, 0.5, 0.75 and 1 MHz and was used as a reference hydrophone. The working hydrophone used in the measurements described below was another double layer hydrophone but without preamplifier. The working hydrophone was cross calibrated against the above reference hydrophone at discrete frequencies by substitution. The calibration by substitution was performed in the far field of the source transducers. The hydrophone was placed in the water tank approximately 30 minutes before the measurements began. The position of the source transducer was adjusted to maximize the output signal from the hydrophone, which was displayed on the Tektronix 2440 digital oscilloscope. Once the actual acoustic axes of the source and the hydrophone were aligned, the pressure-time waveform corresponding to the maximum output signal was captured and Fourier transformed to determine the center frequency (arithmetic mean of the upper and lower half power frequencies) [26]. The center frequency was used to determine the working hydrophone sensitivity, corrected for the oscilloscope input impedance (1 MΩ 15 pf).

94 Determination of the peak pressure amplitude location Next, the working hydrophone was scanned axially at 1 mm increments along the acoustic axis in the far field region. The received signal was displayed on digital oscilloscope (Tektronix 2440) to detect the peak voltage and the influence of noise present in the measurement system was minimized by using the signal averaging facility of the oscilloscope. Once the axial location that produced the highest in-water intensity was determined by computer analysis, the working hydrophone was repositioned in the location corresponding to the far field maximum pressure amplitude. The working hydrophone signal was re-maximized and the pressure-time waveform was captured. Based on this waveform, the intensity values and other relevant parameters were calculated. Before each measurement, the peak-to-peak driving voltage was measured at the source terminal to ensure identical excitation conditions Calculation of pulse intensity integral The Pulse Intensity Integral (PII) (see Fig 1b) was calculated using the following expression [14, 45]: PII = 2 p ( t) dt ρc (5.1) where p (t) is the instantaneous acoustic pressure, ρ is the density, c is the speed of sound and the integration limits bound the pulse. The value of the PII is a measure of the total energy in the pulse. The pulse duration was determined by finding the time for which the PII is between 10 and 90 percent of its final value. This time was multiplied by 1.25 as called for in the NEMA Standard [12]. Knowing the PII, the pulse duration, and the

95 77 pulse repetition frequency, the pulse average and time average intensity were calculated [12]. An example of the representative plots showing the hydrophone received voltage waveform and corresponding pulse intensity integral together with the frequency spectrum of the signal received from the working hydrophone are shown in Figure 4. These plots correspond to a tone burst excitation at 1 MHz. A similar set of plots was obtained for each frequency considered here. Figure 4: Typical measured hydrophone waveform with the corresponding PII and frequency spectrum of 1 MHz circular piston source.

96 Cross axis scan Cross axis scans (lateral beam profiles) shown in Figure 5 were conducted to determine beamwidth of the acoustic sources so that appropriate sampling interval and number of sampling points sufficient to measure the beam down to the 26 db level could be identified while performing the raster scan [51]. The procedure involved spatial integration of the beam intensity values in the acoustic field along two perpendicular directions. The profiles of Figures 5a and 5b were obtained in the following way. The hydrophone was moved in a plane perpendicular to the acoustic axis of the source from the spatial peak location to the -26 db point location. Next, the hydrophone was scanned back through the maximum of the beam to the -26 db point on the other side of the beam. The procedure was repeated to obtain a beam width in the plane perpendicular to the one already scanned. The signals were averaged to improve the signal-to-noise ratio, and the pressure-time waveform was captured at each measurement point, transferred to the computer and analyzed. The plots shown in Figure 5 indicate that the measured 6 db beamwidths (solid line) are almost identical with the theoretically predicted Gaussian shaped ones (dashed line). However, due to the finite aperture of the active transducer element and the influence of the side lobes in practice, below the 6 db level, the Gaussian shape of the overall radiation pattern is not retained.

97 79 Figure 5: Beam plot along the x and y axes of the 1 MHz acoustic source. The dashed lines show the gaussian beam (theoretical) distribution with the same 6 db beamwidth as that of the measured transducer Raster scan The raster scan was performed to calculate the total acoustic power produced by the source. The scan plane was scanned in a two-dimensional raster pattern, parallel to the sinusoidally excited plane piston source transducer over an array size of 30x30 with a sampling interval of 1.5 mm. This sampling interval (step size) was chosen based on the results of the cross axis scan measurements as described above. The distance between the sampling point is usually less than or equal to one wavelength or the hydrophone diameter, whichever is higher. The sampling was performed down to the 26 db level for the maximum measured in the given plane to ensure that the active cross section of

98 80 the source could be identified [51]. At each sampling point, the hydrophone output voltage was captured, peak detected and transferred to the computer memory. Concurrently the corresponding pulse intensity integral (PII) was computed. The resulting data set was stored as a two-dimensional matrix representing the spatial pressure variation over the scan plane. The surface integral was then obtained by rectangular numerical integration [30]. The experimental setup for the raster scanning and the illustration of the raster fashion are shown in Figure 6.

99 Figure 6: Experimental set up for raster scanning (A) and illustration of raster fashion (B) 81

100 82 Raster scan beam intensity distribution for the 1 MHz sources is shown in Figure 7. 0 Relative Intensity in db Y-axis X-axis 30x30 Samples at 1.5 mm spacing Figure 7: Three-dimensional representation of the intensity field produced by the 1 MHz source transducer and obtained from planar scanning The set of measurements described above and needed to obtain the total acoustic power produced by the two sources used was made at 11 frequencies (i.e. 300, 350, 400, 500, 600, 650, 700, 750, 800, 900 and 1000 khz). This number of measurement points was considered appropriate to obtain an adequate sensitivity versus frequency response in the range khz. As already noted, since a single wideband source in the frequency range 0.1 MHz to 1 MHz was not commercially available, two transducers having nominal center frequencies of 0.5 and 1 MHz were used as acoustic sources.

101 Radiation Force Balance Measurements The Radiation Force Balance (Cahn C-32) measured the radiation force of the beam striking a totally absorbing target (SOAB and HFMA). The diameter of the target was larger than that of the source transducer and in this case, the total beam power could be determined directly by measuring the force exerted on the target. The radiation force pressure F, was determined as: F = P cosθ c (5.2) where θ is the angle of incidence of acoustic beam, P is the magnitude of the time averaged, spatially integrated power intercepted by the target and c is the speed of sound in water. With θ = 0 0 (normal incidence) and for an absorbing target, F/P= 67 mg/w. i.e. one watt of ultrasound power induces a force equivalent to that of 67-milligram weight [52]. In measurement practice, a settling time of an hour after filling the balance with degassed water and inserting the transducer is usually recommended to ensure that any possible convection currents have been minimized and air bubbles, if any, on the target or transducer have been reabsorbed into the water. The power measurements were repeated 5 times to take into account thermal drifts and minimizes uncertainty of the measurements. The next section describes calibration of the selected hydrophone in terms of µv/pa based on the measurements of total acoustic power and raster scan beam distribution.

102 Hydrophone Sensitivity Calculation The basic procedure underlying the planar scanning method of calibration involves in producing a known pressure field at the hydrophone and measuring the hydrophone s electrical output. The core of this procedure is to assess the pressure level of a known acoustic power of the source transducer independently by using the hydrophone. The intensity response factor is one form of expression for the sensitivity of the hydrophone, which is the relationship between the plane wave intensity at the hydrophone and the square of the voltage produced at the hydrophone terminal as a result of that intensity. Since the planar scanning method and the radiation force balance both measure the acoustic output of the source, by comparing the power obtained from the raster scan measurements with the power measured with the Radiation Force Balance (RFB), the intensity response factor ( K ) of the hydrophone was calculated as [30]: 2 f K 2 f = PRF T 2 V TP ( t ) dt M N 0 x y 2 * * RFBPower VTP ( xi, y j ) 2 VTP (0,0) i= 1 j= 1 (5.3) where M = Number of linear scans N = Number of sample points per linear scan x = Distance between linear scans y = Distance between sample points V TP = Temporal peak voltage PRF = Pulse Repetition Frequency

103 85 Equation (4) indicates that the units of K 2 f are V 2 W -1 cm 2. Since hydrophone probe sensitivity M h is usually given in terms of V/Pa or db re 1µV/Pa, the values of K 2 f were converted into the corresponding values of M h using the following relationship [53]. K 2 f = 1.5 * * [10 (M (db)/10) ] ( V 2 W -1 cm 2 ) (5.4) 2 K f M h = 10 log ( *10 ).. (db re 1V/µPa) (5.5) It should be noted that this relationship holds for plane waves. To comply with this requirement, the measurements were carried out in the far field of each source. The results obtained are summarized in the next section. 5.6 Comparative Study To facilitate a comparison and discussion of the calibration results obtained using the planar scanning technique, calibration of the same hydrophone using Time Delay Spectrometry (TDS) technique combined with substitution calibration method [2-4] was performed. The results were also verified with the data provided by National Physical Laboratory (NPL), UK, using independent calibration.

104 Results Correlation and Discussion The plots of Figure 8 summarize the end of cable (EOC) voltage sensitivity versus frequency for the working membrane hydrophone. The calibration results obtained using the planar scanning technique are presented together with those provided by the independent national laboratory (NPL, UK) and also with those determined by substitution using the swept frequency technique [28, 54]. Since the swept frequency calibration technique yields virtually continuous frequency response, for clarity and consistency the error bars corresponding to discrete planar scanning technique data and NPL data are not shown in Figure Voltage Sensitivity (db re 1V/uPa) NPL Calibration data Planar Scanning data Swept Frequency Calibration data Frequency (MHz) Figure 8: End-of-cable voltage sensitivity versus frequency of the double layer PVDF hydrophone in The frequency range 0.3 to 1 MHz obtained using Planar scanning technique ( ), Time delay spectrometry method ( ) and calibration data provided by National Physical Laboratory ( ).

105 87 Based on the results presented in the previous section, the experimental data indicated that the sensitivity variation of the double layer hydrophone probe by employing planar scanning technique was approximately ±1 db from 0.3 to 1 MHz [5]. The calibration results of the working hydrophone by using the TDS technique combined with substitution method [3, 4] agreed well with those obtained here using planar scanning approach. The overall uncertainty of the TDS calibration was estimated to be better than ±1.5 db in the frequency range considered. The data provided by NPL at discrete frequencies using independent calibration method exhibited a sensitivity variation of approximately ±1 db, which agrees very well with the calibration results obtained by employing planar scanning technique and TDS technique in this work. Such uncertainty is considered to be acceptable in the underwater acoustic measurements carried out in a similar frequency range [55]. Hence the planar scanning approach developed in this work offers an alternative to other primary calibration techniques such as reciprocity or broadband pulse technique. Planar scanning technique provides discrete frequency calibration and it would be desirable to produce additional data in the frequency range overlapping the upper bandwidth of the 0.5 MHz source and the lower bandwidth of the 1.0 MHz source. However, the transmitting voltage sensitivities of both sources were insufficient to provide adequate signal to noise ratio. On the other hand, as already noted in section 2.2, the data obtained at 650 khz were in good (to within ±0.7 db) agreement, which further enhances confidence in the robust nature of the planar scanning technique. In general, it would be desirable to use a single wideband transducer source capable of producing a signal in the whole frequency range considered. The availability of such

106 88 transducer would facilitate implementation of rapid hydrophone calibration. While such wideband source, producing a signal with adequate signal to noise ratio, is not commercially available, it was built and described in [54], where the TDS technique was used to provide a near continuous frequency response of both membrane and needle hydrophone probes in the frequency range 0.25 to 2.5 MHz [55]. In this work it was desirable to obtain calibration data down to 100 khz, however, the calibration could not be performed below 300 khz due to the limitation of the available equipment (Power Amplifier ENI 350L) and inadequate signal to noise ratio achieved with a low frequency source operating at frequencies below 0.3 MHz. The results obtained were also examined to determine the possible spatial averaging effects. As the effective diameter of the hydrophone increases with decreasing frequency, the spatial averaging correction was not required in the measurements reported here because the ultrasound beam was sufficiently wide to cover the active element of the working hydrophone in the frequency range considered. 5.8 Summary To address the issue of non-existence of calibration standard and the need for a new calibration procedure, a planar scanning technique was developed to determine the frequency response of ultrasonic polymer hydrophone probes below 1 MHz and was compared with the sensitivity data determined by TDS calibration technique combined with the substitution method [31] and the data obtained by using independent calibration method at National Physical Laboratory (NPL), UK, which is an internationally recognized measurement laboratory. The experimental data indicate that the sensitivity

107 89 variation of the hydrophone probe determined by employing the planar scanning technique [5] is approximately ±1 db, which agrees well with the results obtained by using Time Delay Spectrometry [TDS] technique [28, 54] and with the data provided by NPL. The double layer hydrophone studied had bandwidth extending below 0.25 MHz. The results are shown in Figure 8. Harris and Gammell measured the voltage sensitivity of piezoelectric hydrophones in the frequency range from MHz using a broadband pulse technique and reported that the sensitivity of the bilaminar membrane hydrophone [27] was essentially flat. Their results were supported by the results obtained in this work by using planar scanning technique [5] and TDS technique [28, 54]. The planar scanning technique is suitable for absolute calibration of hydrophone probes in the frequency range below 1 MHz to within ±1 db and offers an alternative to other primary calibration technique such as reciprocity or broadband pulse technique [5]. So far, no one has explored this planar scanning technique for calibrating the hydrophones below 1 MHz. 5.9 Conclusions In conclusion, a calibration method using planar scanning technique was developed to determine the low frequency response of the PVDF ultrasonic hydrophone probes. This method is suitable for absolute calibration of hydrophone probes in the frequency range below 1 MHz to within ±1 db and offers an alternative to other primary calibration techniques such as reciprocity [5, 55] or broadband pulse technique [27]. The method developed could also be used while characterizing the newly developed ultrasonic transducers for medical imaging in accurate measurement of peak rarefactional pressure at frequencies down to f c /20 as required by federal regulation and thus in arriving precise

108 90 determination of Mechanical Index (MI). It would appear that the method developed could also be used in characterization of medical and therapeutic ultrasound devices operating at frequencies below 1 MHz and for periodical spot check of working and reference hydrophones. The next chapter addresses all the aspects of the development and testing of single element, single and multilayer imaging transducers and one of the research goals of simultaneous improvement in sensitivity and bandwidth is dealt with.

109 91 CHAPTER 6: DEVELOPMENT AND CHARACTERIZATION OF SINGLE ELEMENT NON-RESONANT POLYMER IMAGING TRANSDUCER The procedure devised and process developed for fabrication, the related design and development issues, and the results and analysis of the polymer imaging transducers are presented in this chapter in detail starting from the material characterization. 6.1 Introduction In medical ultrasound imaging, fundamental improvements should result in both the improved image resolution and increased penetration depth. This depends primarily on simultaneously achieving broader bandwidth and higher sensitivity in a single transducer. In order to achieve this goal of simultaneous improvement of bandwidth and sensitivity, piezopolymer was explored as a transducer material for imaging transducer. The multilayer structure with Barker code approach was implemented in fabricating the transducer and the results of multilayer structure were compared with a single layer transducer. 6.2 Characterization of PVDF film - Impedance/Admittance Measurements To validate the resonance and non-resonant behavior of the piezo film and PVDF transducers respectively, the fundamental thickness mode frequency measurements were performed using an Agilent 4395A Network Analyzer. The electrical input admittance and impedance measurements were carried out separately (for the sake of counter verification) on an electroded PVDF film of 56 µm thick and having 3/8 diameter. The half wavelength fundamental thickness mode resonance frequency of PVDF film was

110 92 observed to be 19.8 MHz while the theoretical prediction was (2200 m.sec -1 /(2*56 µm)) 19.6 MHz. The resonant behavior of the PVDF film is clearly seen from the admittance and impedance plot shown in Figure 9 and Admittance Phase plot Admittance magnitude in Mhos Admittance magnitude plot fr=19.8mhz 60 Phase in Degree Impedance Magnitude in Ohms Impedance phase plot Impedance magnitude plot fr = 19.8 MHz Impedance Phase in Degrees Frequency in Hz Frequency in MHz -80 Figure 9: Plot showing the variation of admittance magnitude and phase of a 56 µm piezo film in air Figure 10: Plot showing the variation of impedance magnitude and phase of a 56 µm piezo film in air. 6.3 Design and Development of System Electronics and Controls A common problem with all pulse-echo imaging systems is the need to couple the transducer to both the high voltage excitation pulse and the sensitive receiving electronics. A transmit-receive (T/R) switching circuit by using two pairs of crossed diodes was built, in which the diode pairs were closed during the application of higher voltage (> 0.5 V) pulse but open while the probe was receiving energy. During the excitation pulse, a clear signal path was established between the excitation source and the transducer, while the input to the preamplifier is shorted to ground. After the excitation pulse has passed, both pairs of crossed diodes become non-conducting. A high frequency

111 93 (200 MHz) National Semiconductor current feedback amplifier (CLC 411) was selected for the preamplifier and the circuit built and tested for satisfactory operation. 6.4 Design Considerations of Single Element Polymer Imaging Transducer Material selection, thickness and active area PVDF was selected as the basis for a wideband transducer due to the relatively high operating frequency, wide potential bandwidth, reasonable acoustic impedance matching to water and tissue, mechanical flexibility and robustness. The thickness and area of an active piezoelectric polymer resonator are important criteria for transducer operating frequency, bandwidth and insertion loss. The operating frequency of the transducer is determined by the thickness of the film and the boundary condition. The thickness is usually chosen as λ/2 and it should be at least an order of magnitude smaller than the lateral dimension. A half wavelength resonator design was selected to maximize the bandwidth and it was achieved by using a low impedance backing material. The thickness of the PVDF material was selected as 56 µm for the desired frequency range of 2-15 MHz. Circular PVDF transducers were proposed for the single element transducer in order to have minimal edge diffraction and hence may be considered as an acoustic source producing a plane wave thereby minimizing possible response variation due to aperture diffraction [57]. Although a large active area will have a better electrical impedance match between the transducer and the driving electronics, smaller aperture with large depth of field is generally preferred in medical ultrasound imaging. But the small aperture will lead to higher impedance resulting in poor sensitivity of the probe.

112 Electroding The single element imaging transducer was designed with a circular electrode having a diameter of 3/8 (using a thin 56 µm film) of Ni+Cu+Ni (1000 Angstrom thick), which provide adequate performance without acoustically loading the piezoelectric polymer much. The film was electroded on both sides. The active area of the film is determined by the overlapping electrodes. The electrode patterns on opposing surface of the films are defined in such a way that by bonding the film, a multilayer structure is created. Due to mass loading effects, thickness layer of electrodes has some loading and damping on PVDF and as a result one expects lowering of the resonant frequency of the transducer Backing material Selection of the proper backing material for a piezoelectric polymer ultrasonic transducer demands careful consideration of such parameters as operating frequency, bandwidth, insertion loss, and operational environment. While using brass as backing material for PVDF which has an impedance of ~32 MRayals, acoustic reflections at the brass-pvdf interface have phase shift relative to the reflections occurring in probe using ceramic as piezoelectric material. Therefore, PVDF transducer with brass backing resonates in a λ/4 thickness mode. When PVDF is backed by a material having an acoustic impedance equal to or lower than the PVDF, the transducer will resonate in λ/2 mode, just like ceramic probes. In order to achieve a broadband operation, the Q factor must be low. The Q factor depends on the impedance of the backing and propagation medium. By choosing a lossy backing material of appropriate impedance, the Q factor can be lowered. In general, the acoustic requirements for an absorber backing are (i) it must have an

113 95 acoustic impedance that is close to that of the piezoelectric polymer and (ii) it must have sufficient acoustic attenuation to prevent unwanted acoustic reverberations (i.e. back wall reflections). Other important requirements for the backing material include, that it (i) must be able to adhere to the piezoelectric material and (ii) must be available in sufficiently thick substrate form and have a high surface quality (e.g. polished). Since the acoustic impedance of PVDF material is about 4 MRayls, backing for the piezoelectric film is provided by using a non-piezoelectric rod made from Kynar polyvinylidene fluoride (PVDF) resin having a measured acoustic impedance of 3.81 MRayls. The low impedance backing moves the resonance frequency to a higher value near towards the value for half wavelength film thickness, compared to high impedance backing. The backing block was carefully shaped and polished on the front surface, whereas it was roughened on the back surface to minimize internal reverberation Wear protecting front matching layer The imaging transducers generally require a front layer for wear protection. For the transducer developed in this work, which is operating at half wave resonance frequency, the thickness of the front layer should be λ/4 at its resonance frequency, which is estimated as: t FL = t PVDF * c 2c FL PVDF (6.1) where, t FL and c FL are the required thickness and the longitudinal speed of sound in the front layer material and t PVDF and c PVDF are the thickness and speed of sound in the PVDF film respectively. The proper front layer material will improve the sensitivity but decrease the bandwidth of the transducer slightly. Since the acoustic impedance of the

114 96 PVDF and tissue are about 4 MRayls and 1.6 MRayls respectively, the optimum front layer material of the imaging transducer should have an impedance of about 2.5 MRayls ( 4 *1. 6 ). The front layer can be designed in such a way that it will serve both for impedance matching and as well as focusing. To avoid quarter wavelength filtering effects, the front surface was not covered with any protective coating to the transducers developed in this work Adhesion The adhesive used for gluing the PVDF film and the backing must have acoustic impedance close to that of PVDF. This will prevent reflection at the boundary of the acoustically mismatched material. The adhesive layer must be extremely thin (very much less than the wavelength = ~1/10 of wavelength) such that the capacitance of the adhesive layer is many times greater than that of the active piezoelectric layer. Otherwise, the adhesive layer introduces an undesirable series capacitance voltage divider and undesirable acoustic reverberations and losses. The thickness of the glue will have an impact on the resonance frequency, as the resonance frequency is determined by: c f λ / 2 = (6.2) 2 * ( t + ) film t adhesive where c is the speed of sound in the PVDF material. 6.5 Fabrication Process The knowledge gained and the technique of lamination adopted in the fabrication of double layer ultrasonic hydrophone probes were applied in developing the multilayer polymer acoustic transducers. Three layer polymer transducers were fabricated using 56

115 97 µm thick PVDF film in such a way that it produces Barker coded pulses when excited. This was achieved by arranging the polarization pattern of the PVDF films according to the Barker code length 3. The fabrication of Barker code multilayer transducers facilitated in validating the Barker code concept experimentally. A single layer transducer was developed to demonstrate the wideband characteristics and compared with a multilayer transducer. The brief description of the fabrication process is given below. A uniaxially stretched and poled PVDF sheet of 56 µm thickness and having circular electrode of 10 mm diameter, which had been sputtered on to the PVDF film with Ni+Cu+Ni (1000 Angstrom thick) was acquired from Measurement Specialities, Inc., Norristown, PA. The electrical contacts were established on the electrodes by bonding a thin wire with a conductive silver epoxy. A non-piezoelectric cylindrical rod of polymer (Kynar) was polished smoothly on the gluing side and the other side was roughened to prevent acoustic reverberation. Polymer backing was chosen in order to obtain fundamental thickness resonance of half wavelength. After experimenting with several brands of adhesives, a non-viscous epoxy RBC 3200 from RBC Industries, Inc. was chosen for gluing the film, which has a low viscosity of 300 cps. Manual applying of epoxy on one side of each film was attempted, but it did not yield the desired thin and uniform bondline.

116 98 In order to produce a uniform and thin bonding layer, a bonding technique called drop method that was used for the lamination of the double layer hydrophone was adopted in bonding the multilayer structure of the imaging probe [6]. A pressing mechanism was designed and developed, in which a super ball of about 4.6 cm in diameter was used to press the layered structure of PVDF film during bonding. In this method, the hand operated press was used to press the two layers of film in order to obtain a thin glue layer in such a way that the applied pressure makes the drop of epoxy, kept at the middle of the film, spread radially. This ensured that the epoxy is spread radially away from the center of the film so that no air pockets will be formed. The thickness and uniformity of the bond line was optimized after conducting series of several bonding trials. The average thickness of bonds in the range of 1-2 µm was achieved. The Figure 11 shows the pressing mechanism while in use. Figure 11: Pressing mechanism while in use

117 99 By using a small drop of non-conductive epoxy (RBC#3200, RBC Industries, Inc.), the PVDF layer was glued with the backing materials and kept pressed by the pressing mechanism overnight so as to cure the adhesive at room temperature. Care was taken to achieve a thin and uniform layer of bonding. If attempt is made to make the glue too thin, there is a possibility of epoxy being washed out. The surface to be glued was fully cleaned and the epoxy was mixed very carefully to avoid producing bubbles. The electrical leads were secured to the long grooves made along the backing and housed in a stainless steel housing, which was also grounded, thus providing effective electrical shielding before potting it with nonconductive potting compound. The other end of the leads was soldered to a BNC connector through diode pairs. By using the conductive silver epoxy (#2902, Tra-Con, Inc.), the front surface was connected to the stainless steel housing, which was grounded. One single layer and two three layer transducers were constructed. Figure 12 shows the parts in different stages of fabrication process and Figure 13 shows the completed transducers.

118 100 Figure 12: Different stages of the fabrication process of single element transducer Figure 13: Completed single element transducers

119 Characterization of Non-Resonant Single Element-Single Layer Transducer and Multilayer Barker Code Polymer Imaging Transducer The criteria examined in characterizing the transducers included bandwidth, frequency response and sensitivity. The parameter studied includes the epoxy bond thickness Electrical Characterization The typical input admittance and phase spectra for the single layer and three layer transducer measured in air are shown in Figure 14 to Admittance phase plot Impedance phase plot Admittance Magnitude in Mhos Admittance Phase in Degrees Impedance Magnitude in Ohms Impedance magnitude plot Impedance Phase in Degrees Admittance magnitude plot Frequency in MHz Frequency in MHz -80 Figure 14: Plot showing the variation of admittance magnitude and phase of single layer transducer. Figure 15: Plot showing the variation of impedance magnitude and phase of single layer transducer Admittance Magnitude in Mhos micron PVDF film Resonance frequency at 19.8 MHz Single layer transducer Three layer transducer Frequency in MHz Figure 16: Plot showing the admittance magnitude of 56 µm piezo film, one layer and three layer transducer.

120 Experimental system The experimental configuration is shown in Figure 17. Tektronix 2430 Digital Oscilloscope CH 1 CH 2 Sync HP 8116A, 50 MHz Function Generator 50 Ω Output ENI RF Power Amplifier 20 db Gain Stepper Motor Digiplan Motor Controller Preamplifier Coaxial cable interconnect T/R Control Circuit Probe under test Deionized water bath at room temperature Stainless Steel plate target at 1 cm distance Figure 17: Experimental setup for pulse echo measurement

121 103 The transducer was mounted on a manipulator and the measurement system was capable of adjusting the transducer in three orthogonal axes. The transducer was used as a transmitter-cum-receiver so that it functioned in the pulse-echo mode. A voltage-source drive was obtained by using the 50 Ω output port of a 50 MHz HP (Agilent) function generator and was used to generate a monocycle sine burst excitation over the frequency spectrum of interest. The peak amplitude produced by the function generator without connecting the transducer was measured with the oscilloscope set at 50 Ω coupling. The transducer was pulsed into deionized water. A high impedance polished stainless steel reflector kept at a distance of 1 cm was used as the pulse reflector. The single cycle strain echo returned from the reflector induces electric field distribution in the transducer via piezoelectric action. The output of the transducer was taken through the specifically built T/R (Transmit-Receive) switch and amplified in the custom built preamplifier and viewed by the Tektronix 2430 digital oscilloscope. The pressure level depends on the position of the observation point in the acoustic field. At the short axial distance from the transducer face, the transmitted and the echoed signals were assumed to be proportional to the average pressure falling upon the surface of the transducer [35]. In the case of both the technical and practical measurement point of view, one cm separation distance was chosen to evaluate the performance of the transducer, although much closer distance is preferred. Throughout the experiment, the temperature was monitored and found constant with the variation of less than ±1 0 C.

122 Pulse-echo response The performance of the polymer transducers was evaluated by measuring the pulse-echo response and the pulse spectrum. The pulse echo response was measured by recording the signal received from a flat stainless steel reflector placed at a distance of 1 cm from the transducer. The pulse-echo response was used to quantify the receive sensitivity of the transducer. The pulse echo response was evaluated using two parameters. One was the response amplitude, which was defined as the maximum value of the output voltage, and the other is the response time t p, defined as the time duration between the times at which the envelope of the peak amplitude is 40 db. Figures 18 and 19 show the pulse echo voltage received by the single layer and three layer Barker code transducers, respectively developed in this work with the stainless steel reflector kept at 1 cm distance from the transducer. From the time domain pulse echo responses, it may be seen that higher sensitivity (about 2 fold increase) was achieved in the case of three layer Barker code transducer compared to single layer transducer. 0 Pulse echo response of single layer transducer at 1 cm Pulse echo amplitude in db re peak Peak at 7 MHz Horizontal scale: 100ns/div Vertical Scale: 20mV/div Frequency in MHz Figure 18: Pulse-echo response (in time and frequency domain) of single element, single layer transducer at 1cm depth in water for a monocycle sine burst excitation.

123 105 Horizontal Scale: 50ns/div Peak at 10 MHz Vertical Scale : 30mV/div Figure 19: Experimental pulse-echo response (in time domain) of a single element, three layer Barker code transducer at 1 cm depth in water for a monocycle sine burst excitation. Figure 20 and 21 show the pulse echo response of single and three layer transducers at different depths in water, respectively. It was seen that the bandwidth of the transducer decreased as the ultrasound pulse penetrates deeper. 0 1 cm depth 0 1 cm depth Pulse echo amplitude in db re peak cm depth 20 cm depth Pulse echo amplitude in db re peak cm depth 20 cm depth Frequency in MHz Frequency in MHz Figure 20: Pulse-echo frequency response of single element, single layer transducer at different depths in water for a monocycle sine burst excitation Figure 21: Pulse-echo frequency response of single element, three layer transducer at different depths in water for a monocycle sine burst excitation

124 106 Figure 22 shows the pulse echo spectral response of a single layer and three layer transducer at 1 cm depth. 0 Single layer transducer -2 Three layer transducer Received pulse echo signal in db re peak L 3L Peak frequency: 7 MHz 10 MHz Center frequency: 15 MHz 12 MHz -6 db bandwidth: 176% 154% Frequency in MHz Figure 22: Pulse-echo frequency response of single and three layer PVDF transducers at 1 cm depth in water for a monocycle sine burst excitation. From the spectral magnitude response, it may be observed that the polymer transducers showed no discernible electrically excitable mechanical resonance in their spectra. The results show that excellent wideband performance was achieved with the 6 db bandwidth extending about 2 decades of frequency [46]. It indicated a peak frequency of 7 MHz, center frequency of about 15 MHz and 6 db fractional bandwidth between 2 to 28 MHz (176%) for the single layer transducer and a peak frequency of 10 MHz, center frequency of about 12 MHz and 6 db fractional bandwidth between 2.5 to 21 MHz (154%) for the three layers Barker code transducer, respectively. Such bandwidth is not

125 107 available using conventional, resonant transducer design. The extremely broadband response is typical of non-resonant PVDF transducer, which has low impedance backing. From the frequency response plot, a general trend may be seen that the center frequency declines with increasing number of layers. The measured center frequency of the single layer transducer was 15 MHz while the three-layer transducer was 10 MHz. The difference in acoustic impedances of the glue layer and PVDF, and the thickness of the glue layer are some of the causes for downshifting of the primary thickness mode resonance of the transducer. The increased thickness of the glue layer is one of the causes for the decreased sensitivity of the transducer The effect of epoxy bond between the layers of PVDF film was examined. It was found that thicker bond degraded the transducer performance. The thickness of epoxy bond between the layers and between the backing and the layer stack must be thin to optimize the performance. Figure 23 shows the frequency response of one single layer and two three layer transducers. It may be seen that the frequency response of three layer transducer number 2 is improved due to lesser glue thickness compared to three layer transducer number 1. If the thickness of the intermediate glue layers is maintained as very thin (~1/10 of the shortest wavelength) and constant, then the transducer s center frequency will remain unchanged and its sensitivity will improve linearly with the number of active layers.

126 108 0 Single layer transducer -2 Received pulse echo signal in db re peak Three layer transducer I Three layer transducer II Frequency in MHz Figure 23: Pulse-echo frequency response of one single layer and two three layer single element polymer transducers at 1 cm depth in water for a monocycle sine burst excitation Diffraction correction The acoustic field from uniformly vibrating transducer consists of two components namely a plane wave and an edge wave [57]. The plane wave propagates with its wavefront parallel to the transducer surface, and the edge wave radiates outward from the transducer edges. The edge wave interferes with the plane wave. The calculation of finding the average pressure on the receiver and relating it to the pressure of an ideal plane wave on the surface of the radiator is called diffraction correction. A.S. Khimunin has provided a table [58], which will relieve the necessity of carrying out the calculation of the corrections by means of the known formulae. The table presents

127 109 numerical values of the modulus ( p p ) of ratio of the average pressure to the pressure of an ideal plane wave on the surface of a circular transducer propagating into an uniform isotropic medium without absorption verses two independent variables namely, the 2πz dimensionless distance S = 2 ka and the wave parameter ka, where k is the wave number (ω/c), a is the transducer radius, and z is the distance between the radiating and receiving surfaces. The values provided in the table are for amplitude diffraction correction. The procedure for incorporating the diffraction correction to the experimental data consists of the following sequence: The wave parameter ka is determined for the given radius of the transducer and appropriate frequency/wavelength. The column with the nearest wave parameter value is found from the tables of diffraction corrections. Such procedure would be faultless only in cases where the effective radius of the transducer is equal to the geometric radius. Practically, the piezoelectric transducer will have effective radius different from the geometric radius because of the disparity in the mechanical tension between the periphery and the center of the source. So the correction effected had some systematic error. The diffraction correction is important at the lower frequencies and it decreases with increasing frequency. Hajime Seki et.al. has pointed out that one decibel per distance of 2 a λ provides a rough criteria and a general estimate of the attenuation due to diffraction and they have given the expression for the attenuation due to diffraction (α d ) as [59]:

128 110 c α d = 1.7 db/cm. (6.3) 2 a f The Figure 24 shows the pulse echo response (both in time domain and frequency domain) of a single element, single layer transducer at different depths for a monocycle sine burst excitation. It is seen that the spectrum of pulse varies as it penetrates deeper into water because the attenuation in water/tissue is frequency dependant. It also revealed that the center frequency and bandwidth of the ultrasonic pulse decrease as the ultrasound pulse penetrates deeper. This implies that the axial resolution of the beam will worsen as the ultrasound beam penetrates deeper into tissue unless time gain compensation (TGC) is applied in clinical use. Moreover, the results show that at close distance, more or less unipolar pulses are produced. With increasing separation the pulses degrade towards bipolar behavior. The unipolar degradation observed at close distance is due to the loss of low frequency content because of diffraction losses. The consequence of this diffraction loss is a degraded signal/noise ratio at the lower end of the frequency spectrum.

129 111 0 Pulse echo amplitude in db re peak Frequency in MHz Peak at 7 MHz Horizontal scale: 100ns/div Vertical Scale: 20mV/div At 1 cm depth 0 Pulse echo amplitude in db re peak Frequency in MHz At 6 MHz Horizontal scale: 100ns/div Vertical Scale: 10mV/div At 20cm depth Figure 24: Pulse echo response (in time and frequency domain) of single element, single layer polymer transducer at different depths in water for monocycle sine burst excitation

130 112 1 Normalized echo signal Water path length in cm Figure 25: Pulse echo response of a single element, single layer transducer for different water path length at peak frequency for monocycle sine burst excitation. The diffraction loss corrections were applied to the measured frequency spectra in the pulse-echo mode. The Figure 26 illustrates the frequency spectra of a single layer transducer with and without diffraction correction. It may be seen that there is an improvement in the low frequency arena and consequently there is a change in the estimated 6 db bandwidth due to the diffraction effects. 0 1 cm depth - without diffraction correction -2 1 cm depth - with diffraction correction Pulse echo amplitude in db re peak cm depth - without diffraction correction 8 cm depth - with diffraction correction Frequency in MHz Figure 26: Pulse echo response of a single layer transducer at different depths in water for a monocycle sine burst excitation (with and without diffraction correction).

131 Sensitivity correction In the receive mode, the active element of the PVDF transducer acts like a voltage source in series with a small capacitor which depending upon the construction of the transducer, will vary. The source is loaded by the cable from the active element to the measurement device (Oscilloscope). So the receive voltage that obtained is in term of effective voltage. In order to know the exact voltage produced by the active element of the transducer at its terminal, correction for the capacitance loading can be performed. The schematic diagram of the transducer acting as a voltage source is shown in Figure 27. Probe Coaxial interconnect cable Preamplifier Oscilloscope Voc Vec Veff Ce Voc Cs Cc CL Figure 27: Schematic diagram showing the capacitance loading of the polymer transducer

132 114 V oc represents the open circuit voltage V ec represents the end of cable voltage V eff represents the effective voltage measured. The voltage source V oc, produced by the active element of the probe having a capacitance C e is driving the coaxial cable with a capacitance of C c, the stray capacitance, C s and the measuring apparatus (Oscilloscope) with a capacitance of C L. The open circuit voltage of the active element of the transducer is estimated as [31]: V measured = V oc C T C T + CIC + CS (6.4) where C T is the capacitance of the transducer element C IC is the capacitance of the interconnecting cable C S is the stray capacitance The change in sensitivity caused by the change in the capacitive loading between the transducer element and the measurement device is a factor of C T CT + CIC + CS. It may be seen that a change in the interconnecting cable capacitance would directly affect the sensitivity of the transducer. 6.7 Summary The single element, single layer transducer and multilayer Barker code transducers were developed. The experimental results indicated that the polymer transducer could be operated in the clinically relevant frequencies using a single transducer. The pulse echo

133 115 spectral response of a single layer and three layer transducer indicated a 6 db fractional bandwidth between 2 to 28 MHz (176%) for the single layer transducer and 2.5 to 21 MHz (154%) for the three layer Barker code transducer, respectively. It has been reported in literature that the temperature dependence of the PVDF transducer performance for the 40 to 80 0 C range showed an approximate linear decrease in center frequency and increase in fractional bandwidth with increasing temperature. Since the speed of sound decreases approximately with increasing temperature, PVDF transducer would have a temperature dependent peak in its performance. The water temperature was monitored during the period of testing and the variation of temperature was ±1 0 C during the measurement. The measurements were carried out at room temperature. 6.8 Conclusion It is expected that this wideband non-resonant transducer will become a useful clinical tool with expanded diagnostic ability and minimize the trade off between bandwidth and penetration depth. This wideband polymer transducer, once optimized, will be useable in all clinically relevant frequencies between 2 to 15 MHz using a single transducer and will improve the diagnostic efficacy. This single element transducer will be well suited for high frequency imaging such as ultrasound backscatter microscopy. High frequency single element transducers are currently used only in some specific applications such as in the field of ophthalmology and dermatology. Since most of the diagnostic applications use multielement transducers, it was decided to explore the feasibility of implementing the Barker code concept in a linear array transducer.

134 116 The next chapter deals with the design and development of a multielement (array), single layer transducer and a multilayer transducer in which the Barker code concept used in single element was extended to multielement structure as a feasibility study.

135 117 CHAPTER 7: DESIGN, DEVELOPMENT AND CHARACTERIZATION OF NON-RESONANT MULTIELEMENT (ARRAY), SINGLE LAYER AND MULTILAYER POLYMER IMAGING TRANSDUCERS This chapter deals with the design and development process of multielement transducer incorporating the concept of Barker code and the results obtained as a part of feasibility study. 7.1 Introduction The properties and techniques developed in the preceding sections on single element transducer were applied to multielement (array) transducer. Several engineering approaches such as electrode patterning design, gluing and lamination techniques, assembly procedures and mounting arrangements were adopted in the array development. 7.2 Design Consideration and Description of Array Structure The multilayer, multielement linear array transducers were designed and developed by stacking 56 µm thick PVDF film according to Barker code. Design requirements imposed by the intended clinical application and system electronics normally determine the number, dimensions, and spacing of the array element. The length (L) of single element of the array is generally restricted by anatomical considerations; typical values range from a few mm to few cm. The frequency response of a single element is the Fourier transform of the excitation voltage across the aperture. As the element width increases, the Fourier Transform narrows and the element response decreases more rapidly with angle [10]. The design of a transducer depends on the depth of view

136 118 (penetration) imposed by the application and the targeted organ and the footprint (aperture) of the transducer. The aperture describes its lateral dimensional area. For higher frequencies, smaller aperture is preferred in order to focus at shallow depth. Theoretically, the element width (W) of an array should be less than or equal to half wavelength in water to avoid grating lobes in imaging. The array should have large number of elements in order to obtain a high-resolution image. Arrays having elements ranging from 32 to 512 are currently available [9]. If the number of elements is reduced, the quality of constructed images may be lowered. There are trade-offs between the quality of constructed images and the number of elements (or the element spacing) of an array. The probe output response should have Gaussian like shape for good imaging. Four element array design was chosen in this work since a group of four elements are usually fired together in the case of imaging arrays having larger elements in practice. Figure 28 shows the schematic view of the array element and bonding pad design at one side of a PVDF film of the linear array transducer.

137 (Not to scale) Figure 28: Schematic representation of array element and bonding pad pattern at one side of the polymer film of the proposed linear array transducer. It is a 4-element device with element widths (W) of 1.7 mm and an inter element spacing of 0.3 mm and total length (L) of 10 mm. The array aperture is 7.7 mm in azimuth and 10 mm in elevation. This approximates a square transducer source. The electrode pattern on opposing surface of the film is so defined in such a way that by laminating the films, multilayer structure is created. It has a unique design of bonding pads specifically suitable for implementation of Barker code pattern for a three-layer four element linear array. The bonding pads are spread on all four sides of the film. The bonding pads have dimensions of 1.5*1.5 mm, which facilitated making direct epoxy (Tra-Con 2902) connection of the pads to the electrical leads. Every element in the array has one interconnect for the electrical signal and another for ground. To reduce the number of vias in the multilayer structure, via was shared by the adjacent layers by capacitive

138 120 coupling. The unelectroded area of the film prevented the shorting of the signal electrode to ground wire and vice versa. The electrode legs of each element on each side of the three films of the proposed three layer transducer are designed in such a way that they lie at different locations of the four sides and not overlapping in any area to avoid possible electro-acoustic and capacitive coupling between each element of the array and to the adjacent layer. 7.3 Fabrication Process Most of the processes were developed by using several logical sets of investigative steps. For example research was done to determine the candidate method and a test method was adopted and the best method was implemented. The electrode pattern (on one of the sides of the layer) for developing the three layer Barker code array transducer shown in Figure 28 was used in fabrication of the four element array. The Figure 29 shows the photograph of designed electrode pattern (on one side of one of the layers).

139 121 Figure 29: Actual electrode pattern of an array layer The sputtering mask was made by Electro Discharge Machine (EDM) according to the designed array pattern by using a slightly curved thin stainless sheet of thick, which facilitated in holding the poled polymer film tightly when the mask was pressed against the film that resulted in good edge definition of the electrode during sputtering process. The electrode and electrical leads were made on the uniaxially stretched polarized PVDF film of 56 µm thick by vacuum deposition by using the metallic mask by the local company, Measurement Specialty Inc., Norristown, PA. The electrodes were sputtered onto the faces of the polarized PVDF films with 200 Angstrom thick of Ni (V)

140 122 {vanadium (7%) mixed Nickel (Vanadium was used to make the alloy non magnetic)} as base metal and 800 Angstrom thick of gold (Au) on top of the Nickel which serve as electrode. Before sputtering, the piezofilms were given thermal and corona treatment for thermal stability and good metal adhesion, respectively. The leads were connected permanently by means of a conductive silver epoxy (Tra-Con 2902). A novel method of array stack architecture employing the Barker code design was used. The stacking and bonding technique used in developing the single element multilayer transducer was adopted in fabricating the multielement linear array transducer. A non-piezoelectric polymer rod of square shape made from (Kynar) polyvinylidene fluoride (PVDF) resin and was machined and polished smoothly on the gluing side and roughened on the other end to prevent acoustic reverberation. Polymer backing was chosen in order to obtain fundamental thickness resonance of half wavelength. Mating grooves were made on the sides of the backing to secure the leads. Several trial pressings were conducted in order to optimize the uniformity and thickness of the glue layer. The PVDF films were bonded together by using a slow curing, low viscosity epoxy (RBC 3200) and by using the pressing mechanism shown in Figure 11. Pressure was applied during bonding in order to achieve a glue line of less than 2 µm thick. The bonded assembly was kept in the pressing mechanism overnight until the epoxy was cured at room temperature. After bonding the films on to the backing, the entire structure was placed in the stainless steel housing. Figure 30 shows the main components of the array at different stages while construction and Figure 31 shows the completed array transducers.

141 123 Figure 30: Different stages of the construction process of four element array transducer Figure 31: Completed array transducers.

142 Design and Development of System Electronics and Control A complete transmit system for the transmit circuitry and an operational amplifier for the receive circuitry were built in order to in evaluate the array transducer in pulse echo mode. Figure 32 shows the transmit/receive switching box and the preamplifier. Figure 32: Photograph of the newly built transmit/receive control circuit and the preamplifier 7.5 Experimental System for Performance Evaluation The experimental arrangement used was similar to that described earlier and used for testing the single element wideband PVDF transducer. Figure 33 shows the water tank experimental system setup used to measure the pulse echo response of the array transducer.

143 125 Figure 33: Experimental configuration for pulse echo measurements By employing the specially designed and fabricated dual transmit/receive (T/R) switch, each element in the array could be connected both to the pulsing unit through diode pairs and to the receiving unit (Oscilloscope) through the custom designed preamplifier. In this way an individual element of the array and all four elements of the array could be energized and tested in the pulse echo mode. The element was excited with a monocycle sine burst. The echo signal was reflected from a stainless steel reflector kept at 1 cm depth (to reduce attenuation) away from the array in the water, fed into the corresponding element of the array and was amplified by the preamplifier before sending it to the Tektronix 2430 digital oscilloscope.

144 Performance Evaluation of the Array Transducer The ultimate test of an array is the ability to obtain a good image. In an ultrasound imaging system, the performance of the transducer is generally masked by system electronics used to drive the array elements. Therefore the evaluation of individual elements is the best benchmark to assess the performance of any array. The criteria examined in evaluating the performance of the transducers fabricated in the course of this research include bandwidth, frequency response, and uniformity and pulse echo sensitivity. This section describes the experimental measurements performed on the array transducer Single layer array transducer Pulse-echo response of individual element The pulse echo experiment was designed in order to test the performance of the array. The time domain response of each element in pulse echo mode is the widely used measure in array performance. The pulse echo response was measured in order to determine the bandwidth, sensitivity, pulse length and center frequency of each element of the array. Each element of the array transducer was excited with a monocycle sine burst and the pressure wave, reflected from the stainless steel reflector placed in the water tank 1 cm away from the face of the array, was received by the same element and fed into the receiving circuit. A custom built preamplifier with a gain of 6 db was used in order to have a satisfactory display of the echo amplitude on the oscilloscope. The highest peak-to-peak amplitude of the echo signal was recorded over the frequency range of 1 to 40 MHz, and the center frequency, -6 db bandwidth, sensitivity and the pulse length were

145 127 determined. The highest amplitude was used to assess the sensitivity of the measured element. The pulse length was determined from the first and last points of the waveform that were 40 db down in amplitude relative to the peak [60]. Figure 34 shows the experimental frequency response recorded for one of the elements of the single layer array, which is representative of a typical response observed during testing. The response indicates a 6 db bandwidth of 155% (between 4 to 32 MHz). At 14 MHz Horizontal Scale: 50ns/div Vertical Scale: 10mV/div 0 Pulse echo amplitude in db re peak Peak frequency = 14 MHz Center frequency = MHz -6dB bandwidth = 155% Frequency in MHz Figure 34: Experimental pulse-echo responses (in time & frequency domain) for one of the elements of single layer array transducer at 1cm depth in water for a monocycle sine wave excitation

146 Pulse echo response of all the four elements All four element of the array were excited simultaneously and the pulse echo response was measured. This is a typical case in reality in medical imaging, as a group of four elements are usually fired at a time and then sequentially advanced further in the case of arrays having larger number of elements. Figure 35 shows the experimental pulse echo response. Peak at 12 MHz Horizontal Scale: 50ns/div Vertical Scale : 20mV/div 0 Pulse echo amplitude in db re peak Peak frequency: 12 MHz Center frequency: 17.5 MHz -6dB bandwidth : 166% Peak frequency: 12 MHz Center frequency: 17.5 MHz -6 db bandwidth: 166% Frequency in MHz Figure 35: Experimental pulse-echo responses (in time & frequency domain) while exciting all the four elements of single layer array transducer at 1cm depth in water for a monocycle sine wave excitation

147 129 The pulse echo response of all four elements of single layer array transducer when excited together indicated a 6 db fractional bandwidth of 166% (between 3 MHz to 32 MHz). From the time domain signal, it may be seen that the transducer radiates a clean acoustic wave almost without any ringing. This is desirable as shorter impulse response implies shorter acoustic pulse, which allows better axial resolution in ultrasound imaging Uniformity The pulse echo response of the individual elements was also examined. A uniform ultrasonic response of individual elements throughout the array was observed. A typical recorded time domain response of one of the elements of the single layer transducer is shown in Figure 36 (Fig.34 reproduced for convenience). The arrays constructed using PVDF were having uniform acoustic response from element to element. The waveforms for all the elements were virtually identical and the response amplitude typically varied within ±1 db. This property is important for the array to produce high quality ultrasound images. At 14 MHz Horizontal Scale: 50ns/div Vertical Scale: 10mV/div Figure 36: A typical observed time domain pulse echo response of one of the elements of the single layer array transducer at 1cm depth in water for a monocycle sine burst excitation.

148 Multilayer array transducer Pulse echo response Each stack of elements of the three layer array transducer was excited with a mono cycle sine burst and the pulse echo response from the stainless steel reflector placed at 1cm depth in water was measured. Figure 37 shows the experimental pulse echo response of one of the stacks of the three layer transducer. It shows a 6 db bandwidth of 141% (between 4 to 23 MHz) Peak at 10 MHz Horizontal Scale: 50ns/div Vertical Scale: 20mV/div 0 Pulse ech amplitude in db re peak Peak frequency: 10 MHz Center frequency: 13.5 MHz -6 db bandwidth: 141% Frequency in MHz Figure 37: Pulse-echo responses (in time & frequency domain) of a representative element stack of three layer array transducer at 1cm depth in water for a monocycle sine wave excitation.

149 131 Figure 38 depicts the time domain pulse echo response of single and three layer array at similar excitation condition. It may be seen that higher sensitivity (about a 2 fold increase) was achieved in the case of three layer transducer compared to a single layer array. Peak at 14 MHz Horizontal Scale: 50ns/div Vertical Scale: 10mV/div (A) Peak at 10 MHz (B) Horizontal Scale: 50ns/div Vertical Scale: 20mV/div Figure 38: Typical observed time domain pulse echo response of one of the elements of single layer (A) and one of the element stacks of three layer (B) array transducer at the peak frequency at 1cm depth in water for a monocycle sine burst excitation.

150 132 Figure 39 shows the over layed pulse echo response of two stacks of elements of the three layer array transducer, which demonstrated the excellent uniformity in the pulse echo response. 0-5 Pulse echo amplitude in db re peak Frequency in MHz Figure 39: Pulse echo frequency response of two element stacks of three layer array transducer at 1cm depth in water for a monocycle sine burst excitation. 7.7 Summary Recording the spectral response of echoes from the element stack of the array it was found that the experimentally achievable fractional bandwidth was 155% in the case of single layer polymer array and 141% in the case of multilayer Barker code array transducer. All the elements of the single layer array displayed greater than 150% bandwidth, and the element stack of three layer transducer displayed a bandwidth greater than 140%, which is one of the design goals of this research. The three layer array exhibited increased pulse echo sensitivity compared to single layer transducer.

151 Conclusion Single layer and multilayer array transducers have been constructed. The construction of multilayer array required critical alignment during bonding. The electrical connection was one of the challenging tasks. The array has been tested in the pulse echo mode and the bandwidth, sensitivity and uniformity were examined. The bandwidth and uniformity of the elements/element stack were found to be excellent and the sensitivity of the three layer array was found to be higher than that of the single layer array. This simple array design gave a satisfactory performance and extension to arrays having large number of elements would be straightforward. It is expected that this kind of array probe will find application in practical imaging systems. A full understanding of the performance of the newly designed and developed array would require extensive modeling. However, the results presented here will help for future work The next chapter presents a consolidated summary of the research work accomplished, conclusions and suggestions for future work based on the outcome of this work.

152 134 CHAPTER 8: SUMMARY AND CONCLUSION 8.1 Summary of the Research Work The simultaneous improvement of both the sensitivity and bandwidth, in the case of imaging transducers and hydrophone probes has been explored in this thesis. The research work presented provides useful information on the performance of PVDF probes for medical ultrasound application. The results presented in each of the preceding chapters lay important ground work for further advancement in employing PVDF technology and incorporating the Barker code concept to the imaging transducer design. The basic ultrasound imaging requirements, material properties and selection, design criteria, fabrication process, and the performance evaluation procedure were described in detail. Double layer hydrophone probes using dissimilar thickness of PVDF film were fabricated and the probes parameter such as sensitivity, frequency response, bandwidth, angular response, effective aperture size and orientation (in two planes) effects were determined. This design exhibited simultaneous enhancement of sensitivity and bandwidth. The development of this type of polymer hydrophone probe using dissimilar thickness of PVDF film was the first one of this kind made and commercially available and no such development was reported in the literature as to the best of the knowledge of this author.

153 135 To address the issue of non-existence of calibration standard below 500 khz and the need for a new procedure, a measurement method employing planar scanning technique was developed to determine the frequency response of double layer polymer hydrophone probe below 1 MHz. The method developed could also be used while characterizing the newly developed ultrasonic transducers for medical imaging for accurate measurement of peak rarefactional pressure used in predicting the potential biological effects in ultrasound imaging. This method using planar scanning technique for the low frequency calibration of hydrophone has not been explored by any one previously as to the best of the knowledge of the author. Transmit and receive electronic systems for the single element and multielement transducer were designed and developed. Single element, single layer transducer and multilayer imaging transducers using Barker code concept have been fabricated and the performance of the multilayer was compared with the single layer in terms of bandwidth and sensitivity. The results show that excellent wideband performance was achieved with the 6 db bandwidth extending about 2 decades (~20 MHz) of frequency. It indicated a 6 db fractional bandwidth of 176% (between 2 to 28 MHz) for the single layer transducer and 6 db fractional bandwidth of 154% (between 2.2 to 21 MHz) for the three layer Barker code transducer respectively. Such bandwidth is not available using conventional, PZT transducer design. If a commercial version of the enhanced bandwidth imaging transducers were available, an effective trade-off between the penetration depth and resolution capability could be

154 136 achieved in clinical practice as the operator will be able to control the frequency with out changing the imaging transducer. The enhanced bandwidth would also allow the transducer to be used for fundamental and higher harmonics imaging. As single element transducers are seldom used at clinically relevant frequencies between 2 15 MHz, a simple four-element array was designed with 56 µm thick PVDF film to test the feasibility of a more desirable transducer implementation. Single layer array transducer and multilayer Barker code array transducers were developed and the performance was evaluated in pulse echo mode. The results indicated a fractional bandwidth of about 155% (between 4 to 32 MHz) in the case of single layer polymer array. All the elements of the array displayed greater than 150% bandwidth and showed very good uniformity of impulse response. The three layer Barker code array transducer was also tested in pulse echo mode and exhibited a fractional bandwidth of 140% (between 4 to 23 MHz) and excellent uniformity. The time domain pulse echo response of three layer array transducer revealed increased sensitivity (about 2 times) compared to a single layer array transducer. 8.2 Conclusion The work presented in this thesis has concentrated maximizing the bandwidth of piezoelectric transducers for medical ultrasound applications utilizing PVDF. The work carried out has shown that the single element and multielement (array) transducers constructed using PVDF demonstrated excellent operating characteristics such as broad bandwidth and uniformity of response. The relatively low pulse echo sensitivity of

155 137 PVDF polymer transducer was partly overcome by adopting the Barker code technique. The pulse echo response of the transducers was studied, which is an important aspect directly relevant to clinical application of ultrasound imaging. The imaging transducer developed using piezopolymer has demonstrated excellent bandwidth with simultaneous improvement in pulse echo sensitivity. One of the problems in fabricating the multilayer transducer was the uniformity and thin adhesive layer in bonding the polymer film. Still more process control is needed on the quality of epoxy adhesion and uniformity of glue layer in order to improve the performance of the transducer. The transducer design described could be operated at frequencies well below fundamental thickness mode resonance of a single layer and could operate in the whole frequency range 2-15 MHz, which is relevant for current clinical applications. Consequently, one transducer could be sufficient for almost all diagnostic imaging applications. Such a wide bandwidth is not available using conventional resonant transducer design. Once fully developed, such an enhanced bandwidth transducers would minimize the trade off between bandwidth and penetration depth and shorten examination time, allow immediate, on-site optimization of image resolution and therefore lead to improvement in diagnostic efficacy. The literature search revealed that no one has developed the Barker code multilayer PVDF array transducer so far. It has also been shown that the use of PVDF in imaging transducers provides enhanced bandwidth without the need for quarter wave matching layer as required in the case of PZT material. The good acoustic impedance matching of piezopolymer with human

156 138 tissue and water makes the polymer a good candidate in efficient energy transfer from the imaging transducer into the propagation medium. Since polymer materials inherently possess low mechanical quality factor (Q m ), broader bandwidth and short impulse response can be achieved, which results in better image resolution and flexibility in optimizing the image of structures at various depths. Since the piezopolymer is a flexible material, it is easy to fabricate probes with curved surfaces. However, due to low electromechanical coupling coefficient and low dielectric constant, the piezopolymer is disadvantageous compared to piezoceramic in terms of sensitivity and transmission efficiency. By employing the multilayer structure using Barker code concept, it has been demonstrated that the sensitivity could be improved. The polymer transducer has the inherent drawback in electrical impedance match with the pulser due to its low dielectric constant. However, this can be resolved to some extent by electrical tuning. In such cases, we have to sacrifice the bandwidth to some extent. Anyhow, as long as we obtain a good sensitivity and reasonable bandwidth useable in clinical range, piezopolymer will be a good candidate for ultrasound imaging applications. 8.3 Suggestion for Future Work The development of PVDF polymer ultrasonic transducers is an exciting area of research. The experience gained in the course of development of multilayer polymer transducer suggests that the design presented could be further improved. The alignment of layers, use of some other lower viscosity bonding adhesive, the thickness of the bonding adhesive, and by designing a viable way to implement electrical connections to each of the individual array elements in the multilayer transducer design. This is a challenging

157 139 task as a typical imaging array has more than two hundred elements. Placing the transmit and receive electronics circuitry and the preamplifier closest to the array element needs to be explored in order to improve the sensitivity still further so that the array element will drive the preamplifier instead of driving the connecting coaxial cable. An integrated design approach in developing the transducer and its compatible pulsing and receiving units taking into consideration of impedance matching is worth considering. Finally in the author s opinion, innovation by combining the single crystal and piezopolymer material (may be designated as polymer /crystal composite) will be foreseen.

158 140 BIBLIOGRAPHY [1] S.S. Corbett III, The Influence of Nonlinear Fields on Miniature Hydrophone Calibration Using the Planar Scanning Technique, IEEE Trans. UFFC, vol. 35(2), pp , March [2] R.C. Heyser, Acoustic Measurements by Time Delay Spectrometry, J. Audio Eng. Soc., vol. 15, pp , [3] P.A. Lewin, Calibration and Performance Evaluation of Miniature Hydrophone Probes Using Time Delay Spectrometry, Ultrasonics Symposium Proceedings, 1981, pp [4] P.C. Pederson, P.A Lewin and L. Bjorno, Application of Time Delay Spectrometry for calibration of Ultrasonic Transducers, IEEE Trans. UFFC, vol. 35(2), pp , March [5] V. Devaraju, P.A. Lewin and H. Bleeker, Determination of Sensitivity Versus Frequency Characteristics of Miniature Ultrasonic Hydrophones Below 1MHz Using Planar Scanning Technique, Journal of Ultrasound in Medicine, vol. 21, pp , March [6] E.P. Papadakis, Nonuniform Pressure Device for Bonding Thin Slabs to Substrates, J.Adhesion, vol. 3, pp , [7] T. Ritter, X. Geng, K.K. Shung et.al., Single Crystal PZN/PT Polymer Composites for Ultrasound Transducer Applications, IEEE Trans. UFFC, vol. 47(4), pp , July [8] V.L Newhouse, Introduction to special issue in honor of John M. Reid, Ultrasound in Med. & Biol.. vol. 20(7), pp , [9] K.K. Shung and M. Zipparo, Ultrasonic Transducers and Arrays, IEEE Engineering in Medicine and Biology, pp , Nov./Dec

159 141 [10] T.A. Ritter, T.R. Shrout, R. Tutwiler, K.K. Shung, A 30MHz Piezo-Composite Ultrasound Array for Medical Imaging Applications, IEEE Trans. UFFC, vol. 49(2), pp , February [11] FDA: 510(k) Guide. Information for Manufacturers seeking marketing clearance of diagnostic ultrasound systems and transducers. Food and Drug Administration, Center for Devices and Radiological Health, [12] Acoustic Output Measurement Standard for Diagnostic Ultrasound Equipment, AIUM/NEMA Standards Publication, UD 2, Rev. 2, [13] F.S. Foster, K.A. Harasiewicz and M.D. Sherar, A History of Medical and Biological Imaging with Polyvinylidene Fluoride (PVDF) Transducers, IEEE Trans. UFFC, vol. 47(6), pp , November [14] K.C. Shotton, D.R. Bacon and R.M. Quilliam, A pvdf membrane hydrophone for operation in the range 0.5 MHz to 15 MHz, Ultrasonics, pp , May [15] A.S. DeReggi and G.R. Harris, Polymer Hydrophone Probes, IEEE Ultrasonic Symposium 1980, pp [16] P.A. Lewin, Miniature piezoelectric polymer ultrasonic hydrophone probes, Ultrasonics, pp , September [17] D.G. Shombert and G.R. Harris, Use of miniature hydrophones to determine peak intensities typical of medical ultrasound devices, IEEE Trans. UFFC, vol.33, pp , [18] G.R. Harris, A Model of the Effects of Hydrophone and Amplifier Frequency Response on Ultrasound Exposure Measurements, IEEE Trans. UFFC, vol. 38(5), pp , September [19] D.R. Bacon, Primary Calibration of Ultrasonic Hydrophones Using Optical Interferometry, IEEE Trans. UFFC, vol. 35(2), pp , March 1988.

160 142 [20] P. Wach, F. Holzer, N. Leitgeb and S. Schuy, On the Theory of Acoustic Radiation Force and its Application in Ultrasonic Power Measurements, Acoustica, vol. 49, pp , [21] E.B. Miller and D.G. Eitgen, Ultrasonic Transducer Calibration at the NBS, IEEE Trans. Son. Ultrason., vol. SU-26, pp. 28, [22] J.M. Reid, Self-reciprocity calibration of echo ranging transducers JASA, vol. 55(4), pp , April [23] Measurement and Characterization of Ultrasonic Fields using Hydrophones in the Frequency Range 0.5MHz to 15MHz, International Eletrotechnical Commission, Geneva, Switzerland. Publication IEC 61102, [24] G.R. Harris, Are current hydrophone low frequency response standards acceptable for measuring mechanical/cavitational indices?, Ultrasonics, vol. 34, pp , [25] G.R. Harris, Pressure Pulse Distortion by Hydrophones Due to Diminished Low Frequency Response, IEEE Trans. UFFC, vol. 42(6), pp , November [26] Standard for Real-Time Display of Thermal and Mechanical Acoustic Output Indices on Diagnostic Ultrasound Equipment, AIUM/NEMA Standards Publication, UD 3, [27] G.R. Harris, and P.M. Gammell, Sensitivity measurements of piezoelectric polymer hydrophones from MHz using a broadband-pulse technique, J. Acoust. Soc. Am., vol. 105(2), pp , February [28] P.A. Lewin, G. Lypacewicz, R. Bautista and V. Devaraju, Sensitivity of ultrasonic hydrophone probes below 1MHz, Ultrasonics, vol. 38, pp , [29] S.M. Jones, P.L. Carson, R.A. Banjavic and C.R. Meyer, Simplified technique for the calibration and use of a miniature hydrophone in intensity measurements of pulsed ultrasound fields, JASA, vol. 70(5), pp , November 1981.

161 143 [30] B.A. Herman, and G.R. Harris, Calibration of miniature ultrasonic receivers using a planar scanning technique, JASA, vol. 72, pp , November [31] M.E. Schafer, Techniques of hydrophone calibration, in Ultrasonic Exposimetry. Boca Raton, FL: CRC Press, 1993, M.C. Ziskin and P.A. Lewin (eds)., pp [32] F.S. Foster, K.A. Harasiewich and M.D. Sherar, A History of Medical and Biological Imaging with Polyvinylidene Floride (PVDF) Transducers, IEEE Trans. UFFC, vol. 47(6), pp , November 2000, [33] R.L. Goldberg and S.W. Smith, Multilayer Piezoelectric Ceramics for Two- Dimensional Array Transducers, IEEE Trans. UFFC, vol. 41(5), pp , September [34] R.H. Barker, Group synchronizing of binary digital systems, in Communication theory, Ed. W. Jackson, Butterworth, London, pp , [35] K.M. Sung, Piezoelectric multilayer transducers for ultrasonic pulse compression, Ultrasonics, vol. 22, pp , [36] M. Platte, PVDF Ultrasonic Transducers Ferroelectrics, vol. 75, pp , [37] P.A. Lewin and Q. Zhang, Nonresonant transducers for ultrasound imaging, J.Acoust. Soc. Jpn. (E), vol. 18(5), pp , [38] G.R. Lockwood, D.H. Turnbull and F.S. Foster, Fabrication of High Frequency Spherically Shaped Ceramic Transducer, IEEE Trans. UFFC, vol. 41(2), pp , March [39] Q. Zhang, "Wideband and effective multilayer ultrasound transducers, PhD Thesis, Drexel University, [40] Glen Wade, Ed.. Acoustic Imaging. Plenum Press, New York, 1976.

162 144 [41] S. Saitoh, T. Kobayashi, K. Harada et. al., A 20 MHz Single-Element Ultrasonic Probe Using 0.91Pb(Zn1/3Nb2/3)O PbTiO 3 Single Crystal, IEEE Trans. UFFC, vol. 45(4), pp , July [42] J.M. Cannata, T.A. Ritter, W.H. Chen and K.K. Shung, Design of Focused Single element (50-100MHz) Transducers using Lithium Niobate, IEEE Ultrasonic Symposium, 2000, pp [43] N.M. Kari, T.A. Ritter, S.E. Park, T.R. Shrout and K.K. Shung, Investigation of Potasium Niobate as an Ultrasonic Transducer Material, IEEE Ultrasonic Symposium, 2000, pp [44] L.F. Brown, Design Considerations for Piezoelectric Polymer Ultrasound Transducers, IEEE Trans. UFFC, vol. 47(6), pp , November [45] H. Ohigashi et.al Analysis of Frequency Response Characteristics of Polymer Ultrasonic Transducers, Japanese Journal of Applied Physics, vol. 27(3), pp , March [46] V. Devaraju, P.A. Lewin and P.E. Bloomfield, Development of non-resonant transducers for ultrasound imaging, Proceedings of IEEE 28 th Annual Northeast Bioengineering Conference, April 20-21, [47] C.T. Lancee, J. Souquet, H. Ohigashi, et. al., Ferro-electric ceramics versus polymer piezoelectric materials, Ultrasonics, pp , May [48] G.R. Harris, R.C. Preston and A.S. DeReggai, The Impact of Piezoelectric PVDF on Medical Ultrasound Exposure Measurements, Standards and Regulations, IEEE Trans. UFFC. vol. 47 (6), pp , November [49] S. Robinson, R. Preston, M. Smith and C. Miller, PVDF Reference Hydrophone Development in the UK From Fabrication and Lamination to Use as Secondary Standards, IEEE Trans. UFFC, vol. 47(6), pp , November [50] R.C. Preston, D.R. Bacon and R.A. Smith, Calibration of Medical Ultrasonic Equipment-Procedures and accuracy Assessment, IEEE Trans. UFFC, vol. 35 (2), pp , March 1988.

163 145 [51] M.E. Schafer and P.A. Lewin, A Computerized System for Measuring the Acoustic Output from Diagnostic Ultrasound Equipment, IEEE Trans. UFFC, vol. 35(2), pp , March [52] P.A. Lewin, Practical implementation and technology of measurement devices, in Ultrasonic Exposimetry. Boca Raton, FL: CRC Press, 1993, M.C. Ziskin and P.A. Lewin (eds)., pp [53] G.R. Harris, A Discussion of Procedures for Ultrasonic Intensity and Power Calculations from Miniature Hydrophone Measurements, Ultrasound in Med. & Biol., vol. 11(6), pp , [54] P.A. Lewin, R. Bautista and V. Devaraju, Voltage sensitivity response of ultrasonic hydrophones in the frequency range MHz, Ultrasound in Med. & Biol., vol. 25(7), pp , [55] R.J. Bobber, Ed. Underwater Electroacoustic Measurements. Government Printing Office, Washington, DC: July [56] A. Gachagen, G. Hayward, S.P. Kelly and W. Galbraith, Characterization of Aircoupled Transducers, IEEE Trans. UFFC, vol. 43(4), pp , July [57] H.D. Mair, L.F. Bresse and D.A. Hutchin, A study of Diffraction effects of Planar Transducers using a numerical expression for edge waves, IEEE Ultrasonic Symposium, 1987, pp [58] A.S. Khimunin, Numerical Calculation of the Diffraction Corrections for the Precise Measurement of Ultrasound Absorption, Acoustica, vol. 27, pp , [59] H. Seki, A. Granato and R. Truell, Diffraction Effects in the Ultrasonic Field of a Piston Source and Their Importance in the Accurate Measurement of Attenuation, JASA, vol. 28(2), pp , March [60] J.M. Cannata, T.A. Ritter and K.K.Shung, A 35 MHz Linear Array for Medical Imaging, IEEE Ultrasonics Symposium, 2001, pp

164 146 [61] Characteristics and calibration of hydrophones for operation in the frequency range 0.5MHz to 15MHz, International Electrotechnical Commission, Geneva, Switzerland, Publication IEC 60866, [62] B. Zeqiri and A.D. Bond, The influence of waveform distortion on hydrophone spatial-averaging corrections Theory and measurement, J.Acoust. Soc. Am. vol. 92(4), pp , October1992. [63] G.R. Harris and D.G. Shombert, A pulsed Near-Field Technique for Measuring the Directional Characteristics of Acoustic Receivers, IEEE Trans. UFFC, vol. 32(6), pp , [64] P. Lum, M. Greenstein, C. Grossman and T.L. Szabo, High-Frequency Membrane Hydrophone, IEEE Trans. UFFC, vol. 43(4), pp , July [65] Acoustic Output Measurement and Labeling Standard for Diagnostic Ultrasound Equipment, AIUM Standards Publication, [66] L.E. Kinsler, A.R. Frey, A.B. Coppens and J.V. Sanders, Eds. Fundamentals of Acoustics. New York: 3 rd ed., John Wiley, 1982, p.107. [67] D.G. Shombert, S.W. Smith, and G.R. Harris, Angular response of miniature ultrasonic hydrophones, Medical Physics, vol. 9(4), pp , July/Aug [68] G.R Harris, Hydrophone Measurements in Diagnostic Ultrasound Fields, IEEE Trans. UFFC, vol. 35(2), pp , March [69] D.R. Bacon, Properties of a PVDF Hydrophone with 100MHz bandwidth for studying Medical, Non-linear and other Fields, Ultrasonic Symposium 1980, pp

165 147 APPENDIX A: DEVELOPMENT AND CHARACTERIZATION OF IMPROVED DESIGN OF DOUBLE LAYER POLYMER HYDROPHONE PROBE a. Introduction In diagnostic ultrasound imaging, there is always concern about the collateral tissue damage due to exposure to ultrasound. The ultrasound device sometimes affects the efficacy of patient treatment and possible side effects. The ultrasound field produced by the newly developed imaging transducers need to be characterized by performing acoustic measurements using a device called hydrophone as recommended by FDA [11], AIUM/NEMA [12], and IEC [61] before use in clinical practice. It measures the spatial distribution of the acoustic field (including its size, shape and energy) produced by the medical imaging transducers and it should provide reliable measurements and comply with the requirement of international standards. Hydrophones are made using either piezoceramic or piezopolymer. The piezoceramic hydrophones have large physical dimension and inadequate bandwidth resulting in temporal distortion and spatial averaging of the waveform. Hence they are not suitable for complete quantitative measurements of medical ultrasonic fields. In view of its small sizes of sensing element, inherent broadband properties, acoustic transparency, and history of successful usage in characterizing the ultrasonic field produced by medical imaging transducers, polymer hydrophones are preferred and accepted for evaluating the imaging probes.

166 148 The primary criteria for such a polymer hydrophone are that the hydrophone should possess adequate sensitivity and broad bandwidth in order to determine the temporal variation of the pulse waveform accurately during the characterization of diagnostic ultrasound transducer. Since the sensitivity of the hydrophone depends on the voltage produced by the active element, thinner element will have lower sensitivity, as the active volume of the piezoelectric material is smaller. The sensitivity could be improved by having a large size of sensing element, since this increases the hydrophone s capacitance. But if the hydrophone s active element is large compared to the beam dimension, it will introduce pressure-averaging effect over the active element causing under estimation of the acoustic pressure and result in error in the derived acoustic quantities such as I spta, MI etc. Although numerical corrections to the spatial averaging effect could be made [62] to compensate for the error caused, such corrected acoustic parameters from the actual measured data will not be accurate. So the active element size could not be increased in order to obtain improved sensitivity. However, the sensitivity of the hydrophone could be increased by increasing the thickness of the PVDF film but the bandwidth will be reduced by using thicker film, as the resonance frequency is inversely proportional to the thickness of the PVDF film. Harris [63] has discussed the necessity of having wider bandwidth hydrophones in order to characterize the imaging transducers generating higher harmonics through non-linear propagation in water. He pointed out that the acoustic parameters derived from the data measured using hydrophones with inadequate bandwidth showed large error (exceeding 30%) in determining the value of Mechanical Index (MI) [24].

167 149 Double layer PVDF hydrophones are conventionally made using two similar thickness of PVDF film (25+25 µm and 9+9 µm) and have the measured fundamental thickness resonance of approximately 20 and 50 MHz respectively. The thickness of the two layers determines the resonance frequency of the hydrophone (fr = c/λ = c/2t, where f r is the resonance frequency, c is the speed of sound in PVDF, t is the thickness of the two layers of film and λ is the wavelength). Thinner film provides higher resonance but shows comparatively lower sensitivity for a given spot size. The double layer polymer hydrophone design using similar thickness of PVDF film was originated at the National Bureau of Standards [15]. Hydrophone using dissimilar thickness of PVDF film (25+9 µm) was developed incorporating better electrical shielding and cable termination. In order to have a robust device that will provide simultaneous improvement of both the bandwidth and sensitivity, a new approach was adopted in this work, which involved laminating a thicker active PVDF film with a thinner ground film, while keeping the standard element size of 0.4 or 0.6 mm diameter. It is worth mentioning here that the development of the double layer membrane hydrophone using different film thickness (25+9 µm) at the beginning of the year 1999 was the first one of such kind, which demonstrated enhanced bandwidth compared to that of hydrophone made with µm thick PVDF film. The polymer hydrophones developed using this improved design were sold by my former employer (Perceptron, Inc.) to several researchers and manufacturers of medical ultrasound equipment around the world for use as a reference hydrophone to perform on the site calibration of other hydrophones and also as working hydrophones for accurate acoustic measurements. My former employer (Perceptron, Inc) was the only

168 150 commercial manufacturer of double layer membrane hydrophone in USA and second in the whole world (the other manufacturer being Marconi in UK). Fabrication of several hydrophones with similar thickness of PVDF film has provided me the opportunity in producing a hydrophone with enhanced performance by using dissimilar thickness of PVDF film. Several PVDF double layer hydrophones with similar thickness (25 µm+25 µm and 9 µm+9 µm) and dissimilar thickness (25 µm+9 µm) of PVDF film were developed. In order to provide the physical basis of the behavior of the hydrophone, experimental determination of the different properties were investigated. Besides determining the frequency response, bandwidth and sensitivity, the directional response was also measured and the variation of the effective spot size of the hydrophones having different geometrical spot size, different film thickness, were studied at different frequencies and at different orientation. b. Background In order to reproduce the acoustic pressure faithfully, the effective diameter of the hydrophone should be smaller than the acoustic wavelength. In measurement practice, when a membrane hydrophone is used in continuous wave (CW) or long tone-burst excitation conditions, acoustic reflections will occur at the membrane surface allowing a standing wave to develop. To eliminate the standing wave, the hydrophone is tilted so that the source transducer and the hydrophone are not in parallel. By doing this, the received hydrophone signal is decreased, as the voltage response of the hydrophone is

169 151 dependent on the angle at which the ultrasound wave is incident on the hydrophone. Hence the angular response (the dependence of the amplitude sensitivity of hydrophones upon the incident wave front direction) of the hydrophone is to be known to assess the error introduced and to make appropriate directivity correction. Further, the knowledge of the effective diameter of the hydrophone is important in assessing the spatial averaging effect of the hydrophone [63]. This is a routine measurement practice while measuring the acoustic output of ultrasonic transducers, particularly when using the higher frequency transducers that are currently used in diagnostic imaging (2-15 MHz). Some ultrasound imaging applications demand higher frequency transducers to image at closer distances. Although conventional diagnostic imaging transducers operate from 2-15 MHz range, very high frequency (VHF) transducers are used for eye, skin, intraoperative and endoluminal applications. For instance, for a transducer having 1 cm diameter and operating at 20 MHz, the 6 db beamwidth at 1cm distance would be 0.08 mm. The geometrical diameter of the hydrophone commonly used in acoustic measurement of ultrasonic probes is 0.4 mm. Since the diameter of the hydrophone is large compared to the 6 db beamwidth, spatial averaging will take place resulting in substantial underestimate of the true acoustic pressure during measurements. To alleviate this problem, peak pressure amplitude at the acoustical axis and at a distance of one hydrophone radius off-axis is measured at all the four directions from the acoustical axis. From the measured data, the true on-axis acoustic pressure is estimated by applying corrections to the measured acoustic parameters [62]. In order to perform this measurement, the effective diameter of the hydrophone needs to be known preciously.

170 152 Moreover, the actual sensing element of the hydrophone will not be exactly circular, although the physical construction is circular in shape. There will be asymmetry in the shape due to fringe field effect during electrical poling of the hydrophone probe. Therefore, the exact active diameter of the hydrophone probe and its corresponding orientation setup as used while determining the effective diameter are to be known for accurate measurement of acoustic pressure radiated by the ultrasound transducers. To investigate the variation in effective diameter of hydrophones and to predict the directional behavior and dependency on frequency, amplitude directivity measurements were conducted by employing several PVDF membrane hydrophones of different kind, various geometrical spot size and dissimilar thickness of PVDF film in two orthogonal axes of the hydrophone. The angular response and the effective diameter of the hydrophone are usually determined from the measurement of its directivity pattern. The directivity pattern was obtained by rotating the hydrophone in the far field of the transducer's plane wave, and by measuring the response at some angle of rotation θ and plotting the response as a function of θ [31]. From the directivity patterns, the effective diameter of the hydrophone is determined (details provided in the later section). In order to ascertain the sensitivity and bandwidth, the newly fabricated hydrophones were evaluated by using Time Delay Spectrometry (TDS) technique combined with substitution method [2, 3, 4].

171 153 c. Physical Description The polymer hydrophone was made from acoustically transparent PVDF material with a small central portion made piezoelectrically active. Hence, the hydrophone being an unbacked membrane, operates in the thickness extensional mode and the fundamental resonance frequency is determined by the thickness of the PVDF film. One of the two layers has a poled central circular positive electrode on one side and the negative ground plane on the other side. The second layer does not have any active element but simply provide protection to the active element besides reinforcing the grounding. The standard double layer hydrophones usually have active diameter of 0.4 or 0.6 mm, with and without preamplifier. If preamplifier is not included, then the coaxial cable length is usually about 75 cm. The hydrophone has an overall dimension of 11 cm in diameter with the exposed membrane having 7 cm in diameter. d. Design Consideration It is essential that the hydrophone is capable of capturing both the compressional and rarefactional portion of the waveform in order to deliver the integrated energy in the ultrasonic pulse accurately. An ideal hydrophone needs to possess: - sufficient sensitivity for use in medical ultrasound applications - uniform frequency response - smallest possible active element in order to provide good directional response and minimize spatial averaging effect - good spatial resolution - good signal fidelity

172 154 - linearity in pressure level - long-term stability in sensitivity - known temperature dependence Selection of piezofilm Resonance frequency In order to have a wider frequency response, thin film should be used. A thinner film will have smaller active volume of piezoelectric material compared to thicker film resulting in reduced sensitivity. The thickness mode resonant frequency (f r ) is determined by: f r = c 2t (A1) where, c is the speed of sound in PVDF film and t is the total thickness of the films. Electrical Impedance The capacitance (C 0 ) and the electrical impedance (Z) of the sensing element of the hydrophone is given by: C 0 = S ε A ; Z = t 1 jωc 0 = t S jωε A (A2) where, t is the thickness A is the area of the active element ε S is the clamped dielectric constant

173 155 It may be seen that the electrical impedance varies inversely proportional to the area of the element size and directly proportional to the thickness of the PVDF film. Electrode dimension IEC 1102 specifies that the maximum effective radius a max should be [23] a max = λ (F ) 1/2 (A3) where F is the focal number (ratio of separation distance between transducer and hydrophone to the diameter of the probe). If the ultrasonic pulse contains harmonics, a smaller spot size is required and the effective radius in such case should be [23]: a max = λ F 4 n -1/2 (A4) Taking into consideration of the above aspects, the design approach was formulated to provide a rugged device capable of giving reasonably high sensitivity and broad frequency response. The design involved laminating two layers of PVDF film in such a way that the electrode and its leg are inside the lamination. Most of the outer side of both the layers is metalized with gold and connected to the ground, which acts inherently as a shielding to the live electrodes, thus reducing the noise pick up through radio frequency interference. The laminated double layer structure has excellent noise immunity. The live electrode was chosen as a small circular dot having geometrical diameter of 0.4 or 0.6 mm, which was defined by vacuum deposited gold on top of chromium (chromium was used for better adhesion) on one side of one layer. The specific central circular

174 156 region (element spot) was poled so as to make it piezoelectrically active. The voltage signal generated by the piezoelectrically active dot electrode was taken via a thin strip of about 0.2 mm wide, gold deposited electrode leg, which extended to a point at the periphery of the PVDF film and served as an electrical signal lead. The thickness of the electrode and electrode-leg materials were chosen to be sufficiently thick to provide durability and long term stability of the electrode but thin enough to avoid acoustic loading of the PVDF film, thereby reducing the bandwidth. The size of the element leg was also carefully chosen keeping in mind the effect of lead resistance. The electrode was evaporated on one surface of the one layer, which is called active layer. The outer surface of this active layer was divided into four quadrants and coated with gold in such a way that a small portion of the electrode leg was visible, which facilitated proper orientation of the hydrophone (the active element side of the hydrophone facing the acoustic source) during measurement. The other layer, which is called ground plane, was coated completely with gold. Although this protective coating has some drawbacks, it was used in order to have a better signal fidelity by way of achieving good grounding. It provides effective shielding to the live electrode from the water in the test tank so that the capacitive loading of the element is reduced. Although the coated double layer hydrophone reduces the electrical pickup, it has larger reflection coefficient compared to single layer non-coated (coplanar) hydrophone. Since the hydrophone has capacitive output impedance, it needs to be connected to an amplifier having high input impedance in order to record the pressure waveform

175 157 accurately. The preamplifier needs to be placed as close to the hydrophone as possible to avoid signal loss in the cable. It maintains the voltage output of the hydrophone unaltered over the frequency bandwidth of interest and it also acts as an impedance matching network. Although the element size was 0.4 or 0.6 mm in diameter, the overall size of the PVDF film was designed to have 7 cm in diameter, which is fairly larger than the acoustic field to be normally measured, so as to allow the beam from the imaging probe to pass through freely thus facilitating acoustic transparency. The spot poling helped in preventing false signals from acoustic perturbations and the absence of backing prevented reverberation effects. Since the impedance of PVDF (4 MRayls) is close to that of water (1.5 MRayl), the acoustic reflection coefficient at the surface of the PVDF film will be low for f < f r. Moreover, since the transverse and radial mode of vibration are related to the dimension of the entire hydrophone instead of the dimension of the active element alone, these modes of vibrations are limited to low frequency level. e. Fabrication Process Electrode patterning was done using shadow masking procedure. In this method, a metal mask covers the entire portion except the intended electrode pattern. A set of twopiece sputtering mask was made from a thin stainless steel sheet by Electro Discharge Machine (EDM). A thin plate was required for the mask to avoid shadowing of the intended electrode area by the edge of the mask.

176 158 A biaxially stretched and unpoled PVDF film was procured and kept between the two plates of the mask and pinned and held firmly. The electrode and electrical leg was vacuum deposited through metallic mask with a thin layer of chromium for better adhesion and followed by slightly thicker layer of gold for good conduction. One side of the 25 µm film was patterned for the active spot electrode and electrode leg and the other side of the film was patterned with four quadrants for the ground plane. The second film with 9 µm thick PVDF film was evaporated with gold on one surface, which is called the second ground plane. Both the ground layer and the active layer were separately stretched over supporting rings. The protective gold coating improves the life and signal/noise level. In order to have low contact resistance, gold plated surface mounting contacts were fixed directly on to the film surface, making permanent electrical connection with the deposited electrode leg by using conductive silver epoxy. Poling was performed on the clamped active layer by heating it to a temperature well below its Curie/melt temperature while simultaneously applying a high voltage for certain period of time, and then allowing it to cool down to room temperature in the presence of the applied electrical field. The poling took place in an air oven. The direction of the applied high voltage decides the direction in which the piezoelectric effect is required. The poling field was applied perpendicular to the film surface. This direction is defined as z coordinate, while the stretching direction is taken as x, y axes. This polarization process rotates and aligns the dipole moments all in the same direction, which is along the direction of polarization. After spot poling, the small central region became piezoelectrically active.

177 159 The assembled and poled active layer and the unpoled ground layer were then laminated together by applying a pressure with a dedicated press and keeping small drops of nonconductive low viscosity epoxy between the two layers. It is called drop method of adhesion. The laminating structure was kept pressed in the pressing mechanism until the epoxy was cured at a set elevated temperature. The contacts of the electrode were soldered to a standard 50 Ω coaxial cable. The outer sides (ground plane) of the laminated films were connected to the screen of the coaxial cable, which was connected to the ground, which minimizes the electrical radio frequency (RF) pickup. The laminated films were mounted on to a hoop machined from ABS (Acrylonityile Butadiene Styrene) material having an outer and inner diameter of 11 and 7 cm, respectively and sealed effectively with pre-tapped mounting holes. Since the voltage output from the active element was reduced due to the capacitance of the cable, an unity gain preamplifier with high input impedance of 10 MΩ and 6 pf having an output impedance of 50 Ω was fabricated in order to drive the electrode leg and a short length (about 3 cm) of output cable. The preamplifier was assembled and potted separately and connected close to the hydrophone using a removable connection. This was done taking into account the manufacturing difficulty and practical usage, although keeping the amplifier closest to the hydrophone is preferred to avoid signal loss in the cable. If longer cable is used, the electrical reflections from each end of the cable will distort the frequency response. Figure 40 shows the view of typical double layer hydrophone developed.

178 160 Figure 40: Actual view of the newly developed double layer polymer hydrophone probe f. Measurement Setup Frequency Response Measurements The details of the measurement setup used for the measurement of frequency response and sensitivity were described in references 2 and 3. Directivity Measurements A broadband PZT (Lead Zirconate Ttitanate) transducer was mounted in a tank containing deionized and degassed water at room temperature. The hydrophone under test was placed facing the source at a known separation distance in the far field. It was rigidly held by a mechanical holding device ensuring that the probe is positioned perpendicular to the plane of the holder. The holding device enabled the hydrophone to be rotated about an axis perpendicular to the beam alignment axis. The mounting arrangements featured independent translation in the X, Y and Z directions besides

179 161 allowing the hydrophone to rotate its acoustical axis, and tilt and adjustment about the vertical axis. The hydrophone movement was controlled through a stepping motor. The spectrum analyzer and the stepper motor controller were interfaced to a personal computer through an IEEE-488 bus for capture and storage of the hydrophone signals. g. Measurement Procedure Frequency Response Measurements The frequency response and the sensitivity were measured using Time Delay Spectrometry (TDS) combined with substitution method [3, 4]. The measurement procedure is already outlined in the main text of the thesis. For Directivity Measurements The hydrophone was placed in water tank approximately 30 minutes before the measurement began. The wideband source transducer was initially aligned visually, directly facing the hydrophone such that the acoustic beam axis was aligned with the hydrophone axis. The directivity measurements were performed by employing Time Delay Spectrometry (TDS) technique [31, 52]. The hydrophone rotational geometry was set up by using a special manipulator to rotate the hydrophone about its face until the face of the hydrophone lay approximately perpendicular to the line of the transmitter acoustic axis, and then rotating it slightly to maximize the output signal from the hydrophone, which was displayed on the spectrum analyzer. This position was defined as zero degree rotation. The hydrophone mount was rotated up to 50 0 on either side of its acoustical axis and the hydrophone response was captured at every one-degree intervals. The

180 162 hydrophone was rotated such that the center of its active element was kept at the same place in the ultrasound field. All measurements were performed in the far field of the wideband acoustic source. The relative amplitude (db) of the hydrophone response measured as a function of frequency at a constant angle of rotation and for several values of the angle of rotation was displayed on the spectrum analyzer. The displayed spectrum of the signal represents the directivity patterns of the receiver as a continuous function of frequency. The spectral data from the hydrophone measurements were transferred from the spectrum analyzer to the controlling computer. From the spectra corresponding to different incident angles, the frequencies corresponding to 3 db and 6 db points were found. The data obtained was rearranged to yield the hydrophone output voltage as a function of the angle of rotation at a constant frequency of interest and analyzed by displaying all the angular data relative to 0 0 at the specific frequency. The angles to the 3 db and 6 db points were used to find the effective diameter of the hydrophone. The measured 3 db and 6 db half angles of the hydrophone response at a particular frequency gave the effective radius. If the effective radius a 3 and a 6 are equal to each other within ±10% of the maximum value, then the mean value shall be used as the effective radius. If not, the value of the two whose corresponding angle θ is closer to 10 0 shall be used [61]. The mean of those two values resulted the effective radius of the hydrophone at that frequency. If θ 6dB is larger than 30 0, then θ 3dB was used to find the effective radius. In this study, the angular response of the hydrophones was analyzed at the frequency of 5 MHz, 7.5 MHz and 10 MHz. The effective radii determined at different frequencies (e.g. 5 MHz, 7.5 MHz and 10 MHz) were averaged in order to determine the mean effective diameter.

181 163 h. Theory and Experiments Acoustic Sensitivity The sensitivity of a hydrophone depends on the thickness resonance of the film and the electric and piezoelectric properties of the polymer material used. The sensitivity is determined by the active area of the element and the capacitance loading of the spot electrode, electrode leg and the connecting cable over a frequency range. Since the thickness resonance is inversely proportional to the thickness of the PVDF film, there is an initial gradual increase in sensitivity with frequency. At the same time, the electrode capacitance decreases with increasing frequency due to decrease in the dielectric constant of PVDF material, causing significant electrical loading of the active element. Hence the sensitivity of the hydrophone increases slowly with increasing frequency [49]. If the acoustic wave having a frequency below the resonance is normally incident on the active element, the proportionality of the output voltage of the hydrophone is given by: V C gtpc e e + C c (A5) where, p is the acoustic pressure incident on the active element g is the receiving constant of the PVDF film C e is the element capacitance C c is the cable capacitance, including the shielding film capacitance t is thickness of the active layer The schematic representation showing the capacitance loading of the hydrophone is given in Figure 41.

182 164 Hydrophone Sensing Element Coaxial Cable Measuring Instrument (Spectrum Analyzer/Oscilloscope) Moc Mec Meff Ce Voc Cs Cc CL Figure 41: Schematic representation showing the capacitance loading of the hydrophone probe where, C e is the element capacitance C s is the stray capacitance C c is the cable capacitance, including the shielding film capacitance C L is the load capacitance (measuring instrument) M OC is the open circuit sensitivity M EOC is the end of cable sensitivity M eff is the effective sensitivity As shown in the Figure 41, the element capacitance C e is loaded by the stray capacitance C s and the cable capacitance C c, including the shielding film capacitance. In order to obtain broader bandwidth, thinner films are used. Thinner the film, the element capacitance is more, resulting in lower impedance.

183 165 The active film element and the shielding film capacitances and the active film impedance are given, respectively by: C e = s ε A, C p = t s ε A t p, Z = t S jωε A (A6) where ε S is the clamped dielectric constant of PVDF t and t p are thicknesses of the active and shielding films, respectively A is the area of the electrode Better resolution/directionality can be obtained by using a small element area and the electrical impedance is increased as the element diameter decreases. The end of cable open circuit sensitivity (M EOC ) of the spot poled hydrophone is given by the ratio of the voltage developed and the acoustic pressure incident on the active element, which can be theoretically calculated by using the following expression [64]: M EOC = p V = C gtc e e + C c (A7) where, V is the end of cable open circuit voltage p is the acoustic pressure in Pascal incident on the hydrophone t is the thickness of the active layer sensing element g is the piezoelectric receiving constant of the PVDF material C e is the capacitance of the active element of the hydrophone C c is the cable capacitance, including the shielding film capacitance

184 166 The capacitance of the element was measured at low frequency far below the resonance (i.e. 1 khz) by using capacitance meter. Since the open circuit voltage sensitivity (M OC ) (which is the actual output of the hydrophone) is difficult to measure, it can be calculated from the measured end of cable sensitivity (M EOC ) of the hydrophone, by measuring the capacitance of the element using an impedance analyzer. For the hydrophone having no preamplifier, it is only a capacitive load and the relationship is given by: M EOC = M OC C e Ce + Cs + Cc (A8) The effective sensitivity (M eff ) is given by: M eff = M OC C e Ce + Cs + Cc + CL (A9) It may be seen that the capacitance of the cable greatly influences the voltage sensitivity of the hydrophone. To alleviate this problem, a preamplifier was placed as close to the hydrophone output terminal as possible. While using the hydrophone with preamplifier, the relationship is given by: M eff = M OC [ R e ( Z L R ( Z e ) + R e L ) ( Z)] I m + [ I ( Z m L ( Z L ) 2 ) + I m ( Z)] 2 (A10) where, Z is the measured complex impedance of the hydrophone Z L is the input impedance of the measurement device (e.g. Oscilloscope) R e and I m denote the real and imaginary parts of the complex impedance.

185 167 The preamplifier of the hydrophone usually has an output impedance of 50 Ω to drive the 50 Ω cable. If the input impedance of the measuring instrument has 50 Ω load, it matches with the output impedance of the preamplifier of the hydrophone. Under such condition, the end of cable sensitivity can be recorded as it is without making any corrections using the expression given in A10. It should always be ensured that the electrical loading condition at the time of calibration is the same as during the time of use so that there will not be any change in sensitivity. The evaluation of the performance characteristics of several hydrophones revealed that larger diameter hydrophones exhibited better sensitivity. It was also found that sensitivity varies from one hydrophone to another within the same type, due to variability in the fabrication process, contributing different capacitive load and also due to the variation of effective aperture size due to fringe field effects during electrical poling. Table 2. Comparative statement showing the variation of frequency response and sensitivity of hydrophones having different thickness of PVDF film. Type of Construction Theoretical Typical Typical Hydrophone using: µm thick PVDF film Thickness Measured Sensitivity Resonance Frequency db re 1V/µP Frequency Response (0.4 mm spot size) (with preamp) 22 MHz 1-20 MHz µm thick PVDF film (0.4 mm spot size) (with preamp) 32 MHz 1-25 MHz µm thick PVDF film (0.4 mm spot size) (with preamp) 61 MHz 1-50 MHz -262

186 168 From the details provided in the table above, it is seen that increase in sensitivity has been demonstrated using 25+9 µm thick PVDF film compared to 9+9 µm thick PVDF film. Frequency Response and Bandwidth The resonance frequency is inversely proportional to the thickness of the PVDF film and the frequency response is relatively constant up to its resonance frequency. Different configuration of double-layer hydrophone probes made by employing 25 and 9 µm thick PVDF films having 0.4 and 0.6 mm of geometrical spot diameter were tested for their performance. The frequency responses of the hydrophones were investigated using two types of measurements: (i) one was the absolute calibration of randomly selected hydrophone at National Physical Laboratory, UK and (ii) the second involved comparing the frequency response of the newly developed hydrophone with a reference hydrophone which was already calibrated at NPL. The results of the randomly chosen hydrophone probes were also verified at National Physical Laboratory (NPL), UK. The frequency response of a double-layer hydrophone probe fabricated with (9+9) µm thick PVDF film measured at NPL showed that the sensitivity is being constant to ±3 db between 1 to 50 MHz is shown in Figure 42, which is an example of wider bandwidth using a thinner PVDF film.

187 End-of-Cable Sensitivity (db re 1V/uPa) micron thick PVDF film Frequency in MHz Figure 42: Frequency response plot of a double layer hydrophone probe using 9+9 µm thick polymer film. In the case of hydrophone probe fabricated with µm thick PVDF film, the sensitivity variation in the frequency range from 1 to 20 MHz is within ±3 db of the mean value for all frequencies. The polymer hydrophones developed using dissimilar thickness (25+9 µm) have demonstrated higher bandwidth compared to that of hydrophones made using similar thickness (25+25 µm) of PVDF film as seen from the data provided in Table 2. A flat frequency response up to 25 MHz has been demonstrated as shown in Figure 43. The downshift of the resonance frequency of hydrophone using 25+9 µm thick PVDF film from the theoretical calculation of 32 MHz was due to the thickness of bonding layer, damping effect of the electrode material and the attenuation in the PVDF.

188 Voltage sensitivity (db re 1V/uPa) micron thick PVDF film 25+9 micron thick PVDF film 25+9 micron thick PVDF film Frequency in MHz Figure 43: Frequency response plots of double layer hydrophones using µm and 25+9 µm polymer film Hydrophone Effective Diameter Theory If the effective diameter of the hydrophone is large compared to the acoustic wavelength and ultrasound beam dimension, the measured acoustic pressure amplitude will be less than the actual value. This is because the hydrophone will respond to the pressure averaged over its active element and hence errors in spatial-peak pressure measurement will occur. This can lead to underestimates of the true acoustic pressures and derived intensities, an effect that has been termed as spatial averaging [50]. Spatial distortion of the beam profile will result if the effective diameter of the hydrophone is large in comparison with the beam dimension. It will be under sampled, and the information carried in spatial harmonics above a certain frequency would be lost. A correction factor

189 171 has to be used to compensate these errors in the measurements data. As per the AIUM/NEMA [65], the guideline for choosing the effective hydrophone diameter, d e is: d e < λz 2ds if z/d s 1 and d e < 2 λ if z/ds < 1 (A11) where d s is the source diameter z is the distance from the hydrophone to the source λ is the acoustic wavelength. The hydrophone diameter is found based on a comparison of the hydrophone's directional response with theoretical directional response of a uniform receiver with circular aperture [65]. Hydrophone's response in the form of the amplitude of the received waveform is plotted as the function of the angle of rotation θ, with respect to the geometrical axis of the hydrophone. This angle is related to the effective diameter [66]. The directional response of the hydrophone was modeled taking into account the film thickness and Lamb wave propagation. It is approximated to the first order Bessel function of a stiff disc receiver. For circular hydrophone apertures of radius a, incident waves of wavelength λ, and measured half angle θ, the far field pressure directivity function is given by [67] P(r,θ) = p(r) 2 J 1 ( sinθ ) (A12) kaka sinθ The hydrophone voltage response in the receive mode is [67] 2 V (θ) = V (0) J 1 ( sinθ ) (A13) kaka sinθ

190 172 where V (0) is the on axis output voltage a is the radius of the hydrophone r is the distance from the center of the source to the point of measurement θ is the angle between r and the source axis J 1 is the first order Bessel function. k is the acoustic wave number ( c ω ) The directivity of an ideal stiff disc hydrophone of radius a is given by [64]: D (θ, z) = jπa ( πz 2 2J1( kasinθ ) ) (A14) kasinθ where z is the axial distance from the source transducer. In practical reality, the directivity of the hydrophone will deviate from the ideal condition. The effective radius a e in mm of the active element of the hydrophone is determined from the measured directional response by inserting the measured half angle θ of the directivity function at the 3 db and 6 db points into the theoretical directivity function and calculating the two values of the effective radius at a particular frequency employing the following relation [61]. a 3dB = 1.62c 2πf sin θ 3dB and a 6dB = 2.22c 2πf sin θ 6dB (A15) where f is the frequency of the angular measurements c is the speed of sound in water at the particular temperature.

191 173 θ 3dB, θ 6dB are the measured half angle at which the hydrophone response is 3 db and 6 db down from its reference level at θ = 0 0 incidence. The two measured half angles (θ 3 and θ 6 ) at one frequency yield two theoretical values of the hydrophone radius (a 3 and a 6), and the mean of those two values yields the effective radius at that frequency [65]. Experimental Results The directivity measurements were used to determine the effective size of the sensitivity element of the hydrophone. It was seen that the effective diameter of the hydrophone varied significantly with frequency and with reference to axis of rotation. It is essential that the effective diameter be known as a function of frequency at specified interval in order to apply the spatial averaging corrections [50, 62] while reporting the acoustic measurement results on medical ultrasonic transducers. Tables 3 and 4 contain data, which illustrate the effective diameters, measured along two orthogonal axes on different kind, geometrical spot size, and having different thickness of PVDF film hydrophones. The tables show the values of effective diameter of the hydrophones d (0) and d (90) for two scan orientation, 0 0 and The d (90) was the result of directional response measurements conducted with the hydrophone being collinear with the electrical leads. Normalization with respect to the hydrophone s geometrical diameter was accomplished for easy comparison. From the results, it was seen that the effective diameter decreased with increasing frequency.

192 174 Table 3. Directivity data illustrating the variation of effective diameter of the hydrophones measured along the two orthogonal axes at 5, 7.5 and 10 MHz. SLID Active At 5 MHz At 7.5 MHz At 10 MHz Diameter Measured Diameter in mm Measured Diameter in mm Measured Diameter in mm in mm Normalized Value Normalized Value Normalized Value d (0) d (90) d(90) d (0) d (90) d(90) d (0) d (90) d(0) d(0) d(90) d(0) 1 A A A A A A B C Table 4. Directivity data illustrating the variation of mean effective diameter of the hydrophones measured along the two orthogonal axes. SL Identification Active Diameter in mm da Mean value among 5 MHz, 7.5 MHz & 10 MHz Normalized Value Measured Diameter in mm d (0) d (90) d(0) da d(90) da 1 A A A A A A B C

193 175 Typical directivity pattern of a hydrophone having 0.4 mm, 0.6 mm and 1 mm geometrical spot sizes are presented in Figure 44 for comparison. (a) (b) (c) Figure 44: Typical directional response of double layer polymer membrane hydrophones with active Directional element Response size of 0.4 mm (Fig. a), 0.6 mm (Fig. b) and 1 mm (Fig. c) in diameter, measured at the frequencies of 5 MHz, 7.5 MHz and 10 MHz.

194 176 Directional Response The membrane hydrophone could be susceptible to Lamb wave propagation across the PVDF film surface (causing a build up of acoustic pressure at the active element) resulting in side lobes when the sound is incident at the critical angle of 50 0 [68]. The directional responses of membrane hydrophones are affected by the variation of piezoelectric sensitivity of PVDF with the direction of acoustic stress and by Lamb wave propagation in the membrane, which causes a build up of acoustic pressure at certain angles of incidence. Both these effects cause the directivity of a hydrophone to deviate from that of a stiff plane disc [69]. The angular responses of several double-layer PVDF hydrophones of different thickness of PVDF film and various geometrical spot sizes were obtained by rotating the hydrophone in the far field of the transducer s plane wave and measuring the hydrophone s response at some angle of rotation. The plots are shown in Figure 45 and 46. It was seen that the Lamb wave propagation appears on the directional response plots in the form of side lobes. It was also seen how the presence of Lamb wave affects the main lobe at higher frequencies and also in larger diameter hydrophones. At lower frequencies, the side lobes dwarf the main lobe. As the frequency increases, the main lobe strengthens and becomes narrower in angular width as expected. The main lobe gradually scaled down in amplitude at higher frequencies. It also revealed that the peaks of the directional response curves at different frequencies are collinear. This means that one could minimize the error while making measurements if the angular alignment of hydrophone is optimized at the highest frequency.

195 177 5 MHz 7.5 MHz 10 MHz Figure 45: Combined directional response of double layer polymer membrane hydrophones with active element size of 0.4 mm in diameter, measured at the frequencies of 5 MHz, 7.5 MHz and 10 MHz. 0.4 mm dia spot size 0.6 mm dia spot size 1mm dia spot size 1 mm dia spot size Figure 46: Combined directional response of double layer polymer membrane hydrophones with active element size of 0.4 mm, 0.6 mm and 1 mm in diameter, measured at 7.5 MHz.

196 178 Orientation effects To study the apodized behavior and asymmetry of the sensitive region of the hydrophone probes, angular response of several hydrophone probes of different kinds, various geometric spot size and different thickness of PVDF film were measured in two rotational axes perpendicular to each other. The effective diameter of the hydrophone probe was determined from the measurement of its directivity pattern. Sensitive region (0.4mm diameter) Gold deposited electrical lead (0.2mm width) Hydrophone d (0) Orientation d (90) Figure 47: Combined directional response of double layer polymer membrane hydrophone with active element size of 0.5 mm in diameter, measured at 10 MHz in two orthogonal axes. Table 5. Directivity data illustrating the variation of effective diameter of the hydrophones having different thickness of PVDF film measured along the two orthogonal axes. Type Membrane Geometrical Effective Spot Size % variation Thickness Spot Size d (0) d (90) d (0) d (90) Bilaminar µm 0.4 mm % 96% Bilaminar 25+9 µm 0.4 mm % 55% Bilaminar 9+9 µm 0.4 mm % 52%