Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Development of high frequency annular array ultrasound transducers
(USC Thesis Other)
Development of high frequency annular array ultrasound transducers
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
DEVELOPMENT OF HIGH FREQUENCY ANNULAR ARRAY ULTRASOUND TRANSDUCERS by Emanuel John Gottlieb A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSPHY (BIOMEDICAL ENGINEERING) December 2005 Copyright 2005 Emanuel John Gottlieb R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. UMI Number: 3220107 Copyright 2005 by Gottlieb, Emanuel John All rights reserved. INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3220107 Copyright 2006 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ii ACKNOWLEDGMENTS Over the past six years that I have been in graduate school I have undergone both a personal and intellectual transformation. I would like to thank my advisor, Professor K. Kirk Shung for allowing me to be apart of his research laboratory. Professor Shung provided me with an opportunity that definitely shaped my life and career as an engineer. It has been a great privilege to be one of Professor Shung’s students. I believe that the past few years have shown me that patience and determination are instrumental in studying for a Ph.D. I have learned a great deal about ultrasound imaging, acoustics, transducer design and fabrication from the many professors, researchers and students at both USC and Penn State. It has been the culmination of many experiences with these individuals that has allowed me to be where I am today. I would like to thank them for their time and effort in furthering research in the field of ultrasonics. I would like to thank my thesis committee: K. Kirk Shung, Jonathan Cannata, Jesse Yen and E.S. Kim for advising me on the preparation of this dissertation. I would like to thank Dr. Jon Cannata for mentoring me on transducer fabrication methods, characterization, and editing of publications. I would also like to thank the staff, graduate assistants, visiting scholars at the Ultrasonic Transducer Resource Center, especially Jay Williams, Joe Han, Thomas Shrout, Richard R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Tutwiler, Qifa Zhou, Ruibin Liu, Chang Hong Hu, Hyung Ham Kim and Kevin Snook for their help with this work. I would like to thank my family, especially my Mother, Florence Gottlieb, Father, Jeff Gottlieb, Brother, Matthew Gottlieb and Grandfather, Frank Zeo for their support while I have been away from home. I would like to give an honorable mention to Dr. Barbara Sonies, who first introduced ultrasound imaging and clinical research to me when I was an undergraduate student. Lastly, I would like to thank my fiancee Megan Hamilton for leaving Pennsylvania to move out to Los Angeles, CA to be with me. I am glad we were able to adapt to our new living environment. Together we have learned allot about ourselves and you have made living here a good experience. I’m looking forward to moving on to our next main event. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. iv TABLE OF CONTENTS ACKNOWLEDMENTS ii LIST OF TABLES vii LIST OF FIGURES ix ABSTRACT xv Chapter 1. INTRODUCTION TO HIGH FREQUENCY ULTRASOUND 1 1.0 General Background 1 1.1 Piezoelectric Materials 2 1.2 Single Element Transducers 6 1.3 Review of High Frequency Linear Arrays 9 1.4 Review of High Frequency Annular Arrays 10 1.5 Review of High Frequency Ultrasound Imaging 11 1.6 Thesis Aims and Goals 16 1.7 Thesis Outline 18 Chapter 2. ANNULAR ARRAY DESIGN AND MODELING 20 2.0 Annular Array Design 20 2.1 Equivalent Circuit Modeling 27 2.2 Pressure Field Modeling 30 Chapter 3. ANNULAR ARRAY FABRICATION AND 52 CHARCTERIZATION R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. V 3.0 Annular Array Fabrication Issues 52 3.1 Annular Array Flexible Circuit Design 53 3.2 Annular Array Fabrication Procedure 54 3.3 Annular Array Characterization 59 3.4 Acoustic Output Measurements 74 Chapter 4. ANNULAR ARRAY IMAGING 80 4.0 Introduction to Imaging 80 4.1 Anatomy of the Eye 86 4.2 Anatomy of the Skin 87 4.3 Wire Phantom Imaging 89 4.4 Annular Array Imaging System 95 4.5 Preliminary Ultrasound Images 96 4.6 Annular Imaging System Limitations 101 Chapters. DEVELOPMENT FEAS ABILITY OF COMPOSITE AND 102 TUNABLE COPOLYMER ANNULAR ARRAY TRANSDUCERS 5.0 Introduction 102 5.1.0 Piezoelectric Composites 103 5.2.1 PZT-5H/Epoxy 1-3 Composite Acoustic Modeling 105 5.2.2 Fabrication of 1-3 Composite using IPB 113 5.2.3 PZT-5H/Epoxy 1-3 Composite Equivalent Circuit Modeling 114 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. v i 5.2.4 1-3 Composite Annular Array Fabrication Procedure 119 5.2.5 1-3 Composite Annular Array Fabrication Issues and 122 Characterization 5.2.6 PMN-33%PT 1-3 Composite Acoustic Modeling 132 5.3.0 Tunable Copolymer Annular Array Transducer 141 5.3.1 Equivalent Circuit Modeling of a Tunable Copolymer Annular 141 Array Transducer 5.3.2 Fabrication and Characterization of a Tunable Copolymer 145 Annular Array Transducer Chapter 6. SUMMARY AND FUTURE WORK 154 6.0 Summary of Work 154 6.1 Future Work 158 BIBLIOGRAPHY 161 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Table 1.1. Table 1.2. Table 1.3. Table 2.1. Table 2.2. Table 3.1. Table 3.2. Table 3.3. Table 3.4. Table 3.5. Table 4.1. Table 4.2. Table 4.3. Table 5.1. v ii LIST OF TABLES Material properties (*Snook, 2000; 2Cannata, 2000; 3 Zipparo, 4 1996; 4 Kari, 2001,5Snook, 2004) Transducer performance (From Snook et al., 2002) 7 Frequency, Axial Resolution, Lateral Resolution, Depth o f 9 Penetration (Thiboutot, 1999) Geometry for the annular array 25 Calculated impedance, capacitance, and focus for the center 27 element Properties of the 75 Q . coaxial cable (PI #171-1019-XX, Precision 58 Interconnect, Portland OR) characterized at 50 MHz, used to impedance match annular array elements Impedance measurement of annular array elements before and 61 after matching at 55 MHz Pulse echo response of annular array elements with impedance 65 matching Annular array insertion loss measurement compensated for 69 attenuation in water and diffraction effects Acoustic output measurements 79 Skin Structure dimensions (from Thiboutot, 1999) 88 Properties of the 50 Q coaxial cable (General Cable RG/58 89 C5779.18.10) characterized at 50 MHz, used in the fixed focus transmit beamformer. Calculated time delays and corresponding cable lengths for a 90 single focus (12 mm) transmit beamformer Properties of fine grain PZT-5H (TRS600FGHD) and epoxy 106 (Epotek 301) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Table 5.2. Table 5.3. Table 5.4. Table 5.5. Table 5.6. Table 5.7. Table 5.8. Table 5.9. Table 5.10. Table 5.11. Table 5.12. Table 5.13. Table 5.14. v iii Calculated properties of PZT-5H/Epoxy 1-3 composite 112 Impedance calculation of matching layers from Desilets et al., 115 1978 Properties of the 50 Q . coaxial cable (N12-38T-125-1, New 121 England Electric Wire Company, Lisbon, NH) characterized at 30 MHz, used to impedance match annular array elements Transducer material comparison of calculated impedance values 123 based on thickness Impedance measurement of annular array elements before and 124 after matching at 30 MHz Pulse echo response of annular array elements with impedance 126 matching Annular array insertion loss measurement compensated for 130 attenuation in water and diffraction effects Properties of PMN-33%PT from Zipparo et al., 2001 134 Calculated properties of PMN-PT/Epoxy 1-3 composite 140 Properties of the 72 Q. coaxial cable (PI # 171 -1019-XX, Precision 146 Interconnect, Portland OR) characterized at 30 MHz, used to impedance match annular array elements Impedance measurement of annular array elements before and 148 after matching at 30 MHz Pulse echo response of annular array elements with impedance 149 matching Annular array insertion loss measurement compensated for 152 attenuation in water and diffraction effects R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Fig. 2.1. Fig. 2.2. Fig. 2.3. Fig. 2.4. Fig. 2.5. Fig. 2.6. Fig. 2.7. Fig. 2.8. Fig. 2.9. Fig. 2.10. Fig. 2.11. Fig. 2.12. ix LIST OF FIGURES Path difference Ad = di-do 20 Simulated time domain pulse echo response (normalized). The 28 time domain pulse echo response has a -20 dB pulse length of 65 ns Simulated frequency spectrum of the pulse echo response has a - 6 29 dB bandwidth of 45 % Simulated electrical impedance magnitude (solid line) and phase 30 angle (dotted line) for an array element Annular array aperture used in Field II simulations. Each annulus 32 in the array is subdivided into the equal aperture elements. A X kerf width separates each annulus from its nearest neighbor Simulated annular array two-way beam profile with no focusing 34 Simulated annular array two-way beam profile with dynamic 35 receive focusing only Simulated annular array two-way beam profile with dynamic 36 transmit focusing only Simulated annular array two-way beam profile with both dynamic 37 transmit and receive focusing Simulated annular array two-way beam profile with a 12 mm 38 transmit focus and dynamic receive focusing. Simulated annular array two-way beam profile with a 10 mm & 39 12 mm transmit foci and dynamic receive focusing Simulated annular array two-way beam profile with 4 transmit 40 zones: 8mm, 10 mm, 12 mm, 14 mm transmit foci and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. X Fig. 2.13. Simulated annular array two-way beam profile with 7 transmit 41 zones: 6 mm, 8mm, 10 mm, 12 mm, 14 mm, 16 mm, 18 mm transmit foci and dynamic receive focusing Fig. 2.14. Simulated annular array two-way beam profile with 13 transmit 42 zones: 6 mm, 7 mm, 8mm, 9 mm, 10 mm, 11 mm, 12 mm, 13 mm, 14 mm, 15 mm, 16 mm, 17 mm, 18 mm transmit foci and dynamic receive focusing Fig. 2.15. Simulated phantom images: no focusing (left), one transmit focus 44 and dynamic receive focusing (middle), both transmit and receive dynamic focusing (right) Fig. 2.16. Simulated lateral resolution (- 6 dB) from wire targets before and 45 after dynamic focusing Fig. 2.17. Simulated two-way radiation patterns for annular array with 46 increasing number of elements at 12 mm from the transducer Fig. 2.18. Simulated two-way radiation patterns for annular array with 48 increasing aperture size at 12 mm from the transducer Fig. 2.19. Simulated two-way radiation patterns for annular array with 49 increasing frequency at 12 mm from the transducer Fig. 2.20. Field II simulated directivity pattern for each annular array element 50 at 12 mm from the transducer. As expected, due to a decrease in width, grating lobe levels were higher in amplitude for the outer annuli. Fig. 2.21. Field II simulated directivity pattern for the 8 element annular array 51 transducer at a 12 mm focus Fig. 3.1. A CAD of the annular array flexible circuit 54 Fig. 3.2. Annular array design cross section. 56 Fig. 3.3. The annular array transducer shows the copolymer film attached to 57 the front surface of the flex circuit and housing R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. x i Fig. 3.4. Electrical impedance magnitude (top) and phase (bottom) for each 60 annulus in water. Each element was impedance matched with a 1.2 pH inductor and a A/4 75 Q coaxial cable Fig. 3.5. Pulse echo analysis setup 62 Fig. 3.6. Measured time domain pulse echo response of the center array 63 element with a -20 dB pulse length of 60 ns Fig. 3.7. Normalized frequency spectrum of the center array element with a 64 - 6 dB bandwidth of 49 % Fig. 3.8. Insertion loss measurement setup 67 Fig. 3.9. Annular array element and its mirror image used to calculate the 68 diffractive loss (From Snook, 2004) Fig. 3.10. Cross talk measurement setup 70 Fig. 3.11. Crosstalk measurement between adjacent elements 71 Fig. 3.12. Crosstalk model from Guess et al., 1995 72 Fig. 3.13. End of cable loaded sensitivity for needle hydrophone 75 Fig. 3.14. Measured received pulse wave form from the hydrophone 78 Fig. 3.15. Pulse intensity integral 79 Fig. 4.1. Lateral resolution as a function of frequency for varying f/#’s 81 Fig. 4.2. Axial resolution as a function of bandwidth in MHz 82 Fig. 4.3. Attenuation for eye, skin, and blood as a function of frequency 84 Fig. 4.4. Maximum penetration as a function of frequency 85 Fig. 4.5. Anatomy of the Eye (Eyesight Insight, [web page] Feb 2000) 86 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. x ii Fig. 4.6. Anatomy of the Skin (University of Maryland Medicine, [web page] 87 January 2001) Fig. 4.7. Measured wire target phantom images (48 dB dynamic range): 92 before dynamic focusing (left) and after (right) dynamic focusing Fig. 4.8. The lateral beam profile measured from a 20 pm wire target at the 93 focus before and after dynamic focusing (DF) Fig. 4.9. Measured and simulated lateral resolution (- 6 dB) from wire targets 94 before and after dynamic focusing Fig. 4.10. Annular array imaging system (From Cao et al., 2003) 95 Fig. 4.11. Wire target image using annular array 97 Fig. 4.12. Ultrasound image of rabbit’s cornea in-vitro using the annular array 98 Fig. 4.13. Wire target image using single element transducer 99 Fig. 4.14. Ultrasound image of rabbit’s cornea in-vitro using the single 100 element transducer Fig. 5.1. Variation with volume fraction of piezoelectric ceramic, v, of a 107 composite’s density, p Fig. 5.2. Variation with volume fraction of piezoelectric ceramic, v, of a 108 composite’s relative clamped dielectric constant, sS 33/so Fig. 5.3. Variation with volume fraction of piezoelectric ceramic, v, of a 109 composite’s coupling coefficient, Kt Fig. 5.4. Variation with volume fraction of piezoelectric ceramic, v, of a 110 composite’s longitudinal velocity, V l Fig. 5.5. Variation with volume fraction of piezoelectric ceramic, v, of a 111 composite’s acoustic impedance, ZA R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Xlll Fig. 5.6. Simulated pulse echo response from 1-3 composite 116 Fig. 5.7. Simulated frequency response from 1-3 composite 117 Fig. 5.8. Simulated pulse electrical impedance from 1-3 composite, 118 magnitude (solid line), phase (dashed line) Fig. 5.9. Voltage divider between composite and epoxy bond line 122 Fig. 5.10. Measured electrical impedance in water for all annuli: magnitude 125 (top) and phase (bottom). Each annulus was impedance matched with a 1.2 pH and a X/4 50 Q coaxial cable Fig. 5.11. Measured time domain pulse echo response for the center array 128 element. The -20 dB pulse length was 120 ns Fig. 5.12. Measured normalized frequency spectrum for the center array 129 element. The measured center frequency was 27 MHz with a - 6 dB bandwidth of 46 % Fig. 5.13. Composite annular array crosstalk measurement between adjacent 132 elements Fig. 5.14. Variation with volume fraction of piezoelectric ceramic, v, of a 135 PMN-PT composite’s density, p Fig. 5.15. Variation with volume fraction of piezoelectric ceramic, v, of a 136 PMN-PT composite’s clamped dielectric constant, £S 33/£o Fig. 5.16. Variation with volume fraction of piezoelectric ceramic, v, of a 137 PMN-PT composite’s coupling coefficient, Kt Fig. 5.17. Variation with volume fraction of piezoelectric ceramic, v, of a 138 PMN-PT composite’s longitudinal velocity, V l Fig. 5.18. Variation with volume fraction of piezoelectric ceramic, v, of a 139 PMN-PT composite’s acoustic impedance, Za Fig. 5.19. Modeled time domain pulse echo response for the center array 142 element. The -20 dB pulse length was 125 ns R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. x iv Fig. 5.20. Fig. 5.21. Fig. 5.22. Fig. 5.23. Fig. 5.24. Fig. 5.25. Modeled normalized frequency spectrum for the center array 143 element. The measured center frequency was 36 MHz with a - 6 dB bandwidth of 36% Modeled electrical impedance magnitude (solid line) and phase 144 (dashed line) for center annulus in water. The element was impedance matched with a 3.9 pH inductor, step down transformer N = 1.5 and a 1/4 72 Q coaxial cable Measured electrical impedance in water for all annuli: magnitude 147 (top) and phase (bottom). Each annulus was impedance matched with a 3.9 pH inductor, transformer (N = 1.5) and a X/4 72 Q coaxial cable Measured time domain pulse echo response for the center array 150 element. The -20 dB pulse length was 175 ns Measured normalized frequency spectrum for the center array 151 element. The measured center frequency was 32 MHz with a - 6 dB bandwidth of 29 % Tuned copolymer annular array crosstalk measurement between 153 adjacent elements R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. XV ABSTRACT The advantage of ultrasonic annular arrays over conventional single element transducers has been in the ability to transmit focus at multiple points throughout the depth of field, as well as receive dynamic focus. Today, annular, linear and multidimensional array imaging systems are not commercially available at frequencies greater than 20 MHz. The fabrication technology used to develop a high frequency (> 50 MHz) annular array transducer is presented. A 9 pm P(VDF-TrFE) film was bonded to gold annuli electrodes on the top layer of a two sided polyimide flexible circuit. Each annulus was separated by a 30 pm kerf and had several electroplated micro vias that connected to electrode traces on the bottom side of the polyimide flexible circuit. The array’s performance was evaluated by measuring the electrical impedance, pulse echo response and crosstalk measurement for each element in the array. In order to improve device sensitivity each element was electrically matched to an impedance magnitude of 50 £2 and 0“ phase at resonance. The average round trip insertion loss measured for the array and compensated for diffraction effects was - 33.5 dB. The measured average center frequency and bandwidth of an element was 55 MHz and 47 % respectively. The measured crosstalk between adjacent elements remained below - 29 dB at the center frequency in water. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. x v i A vertical wire phantom was imaged using a single focus transmit beamformer and dynamic focusing receive beamformer. This image showed a significant improvement in lateral resolution over a range of 9 mm after the dynamic focusing receive algorithm was applied. These results correlated well with predictions from a Field II simulation. After beamforming the minimum lateral resolution (-6 dB) was 108 pm at the focus. Preliminary ultrasound B-mode images of the rabbit eye using this transducer were shown in conjunction with a multi channel digital beamformer. A feasibility study of designing and fabricating tunable copolymer annular array transducers and annular array transducers using 1-3 composites was carried out. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. 1 CHAPTER 1 INTRODUCTION TO HIGH FREQUENCY ULTRASOUND 1.0 General Background Ultrasound imaging is one of the most widely used modalities in medical imaging today. The popularity of ultrasound is due mostly to its noninvasive nature. Other contributing factors such as low cost, portability, and soft tissue differentiation make ultrasound an attractive tool for imaging. Commercial ultrasound imaging systems are available in the operating frequency range of 1 MHz to 20 MHz. Within the last decade advances in electronics and transducer technology have made possible the development of imaging systems at frequencies greater than 20 MHz. This advancement in technology is not wide spread commercially for applications other than intravascular imaging. In the future it is possible that this technology will migrate into most medical disciplines. The medical fields of dermatology, ophthalmology, developmental biology, cardiology and genomics will be most benefited from the technology afforded by high frequency ultrasound. Spatial resolutions between 20 and 100 pm are achieved in the frequency range of 20 to 100 MHz. Disease diagnoses and prevention are the most prevalent goals for applications in high frequency ultrasound. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 1.1 Piezoelectric Materials Ultrasound transducers consist of either a single element or arrays made up of many elements. Array transducer elements can either be slotted from a piezoelectric material or surface micromachined from silicon nitride suspended over a silicon substrate (Ladabaum et al., 1998). Piezoelectric array elements undergo a deformational change when an electric field is applied. When an applied voltage is distributed across the element the dipoles in the material align and cause a strain in the material. A mechanical wave propagates through a medium as a result of this strain in the material. The piezoelectric effect describes the fundamental behavior of a transducer element. The piezoelectric constitutive equations used are: D = eJE + dT (2.1) S = dE + s ET where strain (S) and electric field (E) are the independent variables, s is the dielectric constant or permittivity, s is the elastic compliance constant, d is the piezoelectric strain constant, T is stress and D is electric displacement. The superscript (T) above s represents T = 0 i.e., the piezoelectric material in the free status. The superscript {E) above 5 represents that the material is not under any electric field. Active material properties are crucial in determining transducer performance thus affecting image quality. Polyvinylidene fluoride (PVDF), lithium niobate R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3 (LiNbCb), lead titinate (PbTi0 3 ), potassium niobate (KNbCf) and lead zirconate titanate (PZT-5H) are piezoelectric materials that have been used for high frequency transducer fabrication (Snook, 2000; Cannata, 2000; Zipparo, 1996; Kari, 2001). Naturally occurring piezoelectric materials are crystalline such as crystal quartz (Si02). Ferroelectric materials are not naturally piezoelectric but can become piezoelectric when poled. Ferroelectric materials such as lead zicronate titanate (PZT) and lead titanate (PbTiOa) have to be heated above room temperature and poled in order to exhibit piezoelectric properties. Poling is a process where an electrical field is applied to the ferroelectric in order to align dipoles in one direction. The piezoelectric properties of some piezoelectric and ferroelectric materials are listed in Table 1.1. Polyvinylidene fluoride (PVDF) is a long-chain semi-crystalline polymer containing the vinylidene fluoride monomer [-CH2-CF2-] that goes through an alignment restructuring if the material is stretched. The chains align along with the axis of stretching creating a net dipole moment within the units of the chain. If the material is heated to near its Curie temperature and then cooled during stretching, the units within the chain can be locked into position, thus preserving the net dipole moment. PVDF has low acoustic impedance, low coupling coefficient, high electrical and mechanical losses. Transducers made with PVDF have exhibited bandwidth’s greater than 100% (Foster et al., 2000b) but have poor sensitivity because of PVDF’s low coupling coefficient, kt. The kt is defined as the capability R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 of a piezoelectric material to convert mechanical energy into electrical energy and vice versa. P(VDF-TrFE) films have a significantly higher coupling coefficient and therefore can be used to produce devices with superior bandwidth and sensitivity when compared to devices made with PVDF. A material with a high kt is advantageous for good sensitivity and bandwidth. Material kt Density (kg/m3 ) Velocity (m/s) Q m es/£o Za (Mrayls) ‘PVDF 'P(VDF- 0.13 1800 2200 10 1 1 4 TrFE) 0.25 1820 2400 12 8 4 3 PZT-5H 0.53 7500 3900 60 1200 35 5 PbTi03 0.49 6900 5200 120 200 34 2 LiNb03 0.49 4640 7300 10000 40 34 4 KNb03 0.68 4500 8000 40 41 36 Table 1.1. Material properties ('Snook, 2000; 2Cannata, 2000; 3 Zipparo, 1996; 4Kari, 2001,5Snook, 2004). P(VDF-TrFE) became commercially available in the late 1980’s (Brown et al., 2000). The ratio composition of VDF to TrFE is 3:1, P(VDFo.7 5-TrFEo.25), which has been shown to exhibit the best piezoelectric properties for this material. The low acoustic impedance of P(VDF-TrFE) is a major advantage over many other R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 5 piezoelectric materials. Because of this no matching layers are necessary to achieve broad bandwidth when using P(VDF-TrFE) since the acoustic load of human tissue is 1.5 Mrayls. P(VDF-TrFE) has a Curie temperature of 135°C and a melting temperature of 150°C (Brown et al., 2000). LiNbC>3 (crystal, 36° Y-cut) has a high coupling coefficient and low dielectric constant which makes it an excellent choice for single element transducer designs (Cannata, 2000). LiNbC>3 has high acoustic impedance therefore A ./4 acoustic matching layers are needed to acoustically the transducer to tissue. Broadband 20 to 80 MHz LiNbC>3 transducers have been developed by Cannata et al. (2003) for medical ophthalmology applications. PbTiC>3 has a high coupling coefficient, high acoustic impedance, and low dielectric constant which also makes it suited for single element transducer development. Lead zicronate titanate (PZT) has been the most widely used piezoelectric material in arrays since it has a high electromechanical coupling coefficient (kt), high dielectric constant (s) and low mechanical and dielectric losses. The high dielectric constant is necessary for the small element size of the linear array. KNbC>3 has a very high coupling coefficient, high acoustic impedance and low dielectric constant that make it an excellent choice R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 for single element transducer. Since KNbC>3 has a higher coupling coefficient than LiNbCb, KNb0 3 transducers could have better device sensitivity. However, KNbC>3 can change domain under stress resulting in a reduced kt in the thickness mode (Nakamura et al., 2004) and therefore can not be easily mechanically focused without adverse effects. 1.2 Single Element Transducers Single element transducers have been developed in frequencies above 20 MHz. A comparison of transducer performance was shown by Zhao et al. (1999) and Snook et al. (2002). Transducers made with a variety of materials using a three millimeter diameter aperture and two to three f/# (focal length divided by the diameter) were compared, shown in Table 1.2 (Snook et al., 2002a). Front acoustic matching layers were used for the PZT fiber composite, LiNb0 3 , PbTiC>3 transducers to impedance match them to tissue/water. Conductive backing materials provided electrical connection to the piezoelectric elements. PZT-fiber composite transducer had a parylene matching layer and epoxy (Epotek 301, Epoxy Technology Inc., Billerica, MA) backing with conductive silver- coated microspheres. The PVDF transducer had no acoustic matching layers and an epoxy (Epotek 301) backing with sliver coated-microspheres. The LiNbC>3 transducer had two acoustic matching layers consisting of Insulcast 501 (American R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 Materials Frequency (MHz) Bandwidth (%) Insertion Loss (dB) PZT Fiber 53.6 47 34.4 Composite PVDF 48.1 118 45.6 LiN b03 44.5 74 21.3 PbTi03 45.1 47 23.7 Table 1.2. Transducer performance (From Snook et al., 2002). Safety Technologies, Roseland, NJ) epoxy with mixed silver powder (Aldrich Chem. Co., Milwaukee, WI), and a lens made from Epotek 301. The backing used was E- solder 3022 (Von Roll Isola Inc., New Haven, CT), a conductive backing with an acoustic impedance of 5.9 MRayls. The PbTi03 transducer had a parylene matching layer and an E-solder 3022 conductive backing. The PVDF transducer had the worst sensitivity of the transducers due to its low kt. However, PVDF had the largest bandwidth of all the transducers because of its low acoustic impedance and low Qm . The PZT fiber composite transducer had an average bandwidth and sensitivity. The high insertion loss was probably the result of a poor width to height aspect ratio of the ceramic pillars resulting in more lateral coupling. The LiN b03 and PbTi03 transducers had the best sensitivity with better R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 than average bandwidth which was due to their high kt and acoustic matching layers. This study shows that material properties greatly influence transducer performance and ultimately affect imaging quality. Besides material properties influencing transducer performance the relationship between frequency and wavelength dictates the resolution of the image. There is an inverse relationship between frequency and wavelength. Therefore, an increase in frequency translates into a smaller wavelength thus better axial and lateral resolution. The expense of this better resolution is a decrease in the depth of penetration. Table 1.3 shows the values of depth of penetration, axial and lateral resolution in skin for varying frequencies of 7.5 to 100 MHz (Thiboutot, 1999). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 9 Frequency (MHz) Axial Resolution (pm) Lateral Resolution (pm) Depth of Penetration (mm) 5 200 400 400 10 150 300 150 20 50-100 200-300 6-7 40 39 120 4 50 30 94 3-4 100 11 30 2 Table 1.3. Frequency, Axial Resolution, Lateral Resolution, Depth of Penetration (Thiboutot, 1999). 1.3 Review of High Frequency Linear Arrays High frequency arrays with frequencies above 20 MHz have been difficult to fabricate due to the small scale size. Ritter (2000a) described the fabrication of a 30 MHz linear array using a 2-2 composite. The pitch between elements was on the order of two wavelengths (100 microns). However, a smaller pitch equivalent to one wavelength (50 microns) would be more desirable to reduce grating lobes. The highest frequency linear array fabricated to date using a 2-2 composites and pitch of X has a center frequency of 35 MHz (Cannata, 2004). A 30 micron pitch would be R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 10 necessary for an array approaching a center frequency o f 50 MHz. This requires a kerf on the order of 10 microns, which is very difficult to dice mechanically. A method that has shown to overcome some of the limitations from traditional dice and fill techniques is interdigitial pair bonding. Interdigital pair bonding (Liu et al., 2 0 0 1 ) has been shown to be capable of producing kerf widths smaller than mechanical dicing. Smaller kerf width and ceramic post widths enable better ceramic width to height aspect ratios necessary for frequencies above 30 MHz. Fine grain ceramics or single crystals with improved dice ability would be useful to fabricate these composites. Improved mechanical strength, high dielectric constant and small ceramic width to height ratios would be desirable to achieve the small element size of high frequency (> 30 MHz) linear array transducers. 1.4 Review of High Frequency Annular Arrays Recently, there has been a renewed interest in annular arrays at frequencies above 20 MHz (Snook, 2004; Brown et ah, 2004; Ketterling et ah, 2005) because they bridge the gap between single element and linear and phased array transducers. A 45 MHz annular array ( 6 elements) made with a fine grain Lead Titanate (PbTiCL) as described by Snook, et ah used laser dicing to physically isolate elements and low temperature wire soldering for interconnect. This array suffered from an unusually high insertion loss that was hypothesized to be caused by an acoustic lens used for R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 11 transmit focusing. However, this study was successful in showing that the use of material with low planar coupling, along with laser dicing was effective in reducing inter-element crosstalk. In a far simpler design a 50 MHz (7 element) annular array was fabricated by patterning concentric annuli on un-diced PZT-5H substrate (Brown et al., 2004). Due to the lack of kerfs and high planar coupling for PZT-5H the electrical crosstalk between elements was large. However, despite this the radiation pattern of the array showed side lobes -60 dB below the main lobe. Another approach by Ketterling et al. (2005) also used an un-diced piezoelectric substrate to make a high frequency (40 MHz) array. Instead o f wire bonding or soldering, a single sided polyimide flex circuit was used for interconnect. Ohmic bonding was used to connect a poled piece of polymer PVDF to the defined circuit traces. This array performed well despite its simple construction. 1.5 Review of High Frequency Ultrasound Imaging Currently high frequency ultrasonic imaging has been primarily carried out by using ultrasound backscatter microscopy (UBM) which is a static ultrasonic imaging tool for visualizing microscopic structures. The UBM system consists of an ultrasound transducer, position control system and electronic system: pulser, receiver, amplification, data acquisition and image display system. The ultrasound transducer is mechanically scanned while being pulsed in order to admit an acoustic R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 12 beam over the specimen, the received signals are amplified and combined to from an image. The high frequency and bandwidth transducers used can provide clinical physicians with a tool to make clinical measurements and to evaluate very small structures (normal and pathologic). The current state of the art in UBM including imaging applications, system design, high frequency transducers, data acquisition and motion control were discussed by Foster et al. (2000a). UBM is a research-based tool, which has been used clinically within the last decade. A 100 MHz B-mode ultrasound backscatter microscope for imaging multicellular spheroids and intact human ocular tissue was designed by Sherar et al. (1989). This biomicroscope made possible imaging of internal structure in living specimens on a microscopic scale. Their imaging system included the following system components: transducer, motion system and scan converter with a frame rate of five frames per second. They used an f/# = 2 transducer with a lateral and axial resolution of 36 microns. Passmann and Ermert developed a UBM system in the 20 to 250 MHz frequency range (Passmann and Ermert, 1996). Being aware of the effects of tissue attenuation on spatial resolution they used tightly focused transducers with high energy density to improve lateral resolution. Two significant improvements made to the UBM system were: 1) implementation of synthetic aperture focusing and 2) B/D- scan. The B/D stands for brightness and depth, where a fixed focused single element transducer is progressively mechanically scanned closer to its target. The R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 13 B/D scan improves the resolution in the depth of field. However, the image acquisition is much more time consuming. Using synthetic aperture, different transmitter signals at each depth were used and pseudoinversely prefiltered according to the tissue transfer function improving the image quality. Pulse compression was performed with nonlinear frequency modulated chirp signals to gain signal energy for inverse filtering. Ophthalmologic applications of UBM include the normal structures of the anterior segment including the cornea, iris, Schlemm's canal, ciliary body, and sclera as well as contrasting images of pathology. Some of the pathology examined includes corneal pathology, tumors of the iris and ciliary body, and angle closure between the cornea and iris from pupilliary glaucoma. Foster et al. (1990) performed clinical evaluations using a 50 to 100 MHz ultrasound backscatter microscope. Images had a field of view of four by four mm with axial and lateral resolution ranging from 20 to 50 pm. This first study was performed on 40 patients with varied ocular pathology. A water-bath was used to couple the transducer to the eye. Current ophthalmologic applications and eye banking using UBM were described by Rosenwasser (1999). The review applies to the evaluation of pathology of ocular diseases relevant to UBM, and principles of eye banking. Disease processes discussed were conjunctival and iris melanoma, small lesions of the anterior segment, other forms of neoplasia, intraocular cysts, narrow angle glaucoma, as well as imaging intraocular foreign bodies. Rosenwasser also explains the use of R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 14 UBM with spectroscopy having the potential in determining cell type origins for a variety of tumors. Current methods in treatment of patients including Laser In Situ Keratomileusis (LASIK), a corneal surgery, which creates a partial thickness flap of tissue in the cornea by removing tissue from the base of the cornea with excimer laser ablation then replacement of the hinged flap. LASIK surgery can result in a refractive error, thinning of the cornea, and an unwanted interface within the comeal stroma. UBM is useful in assessing the performance of LASIK surgery. UBM can be used for pachymetry to for corneal pachymetric mapping. Comeal pachymetric mapping is useful in assessing optical aberration after transplantation; this can lead to improving visual performance post transplantation. Applications of high frequency skin imaging include tumor staging, boundary definition, studies of the response of tumors to therapy, investigations of inflammatory skin conditions such as psoriasis and eczema, basic studies of skin aging, sun damage and the effects of irritants. Improved resolution has enhanced calculating the perimeter of small skin lesions, and in vivo skin thickness measurements to characterize nonmalignant skin disease. UBM images of normal skin, seborrhoeic keratosis and malignant melanoma were displayed. Turnbull et al. (1995a) investigated UBM skin imaging in the 40 to 100 MHz frequency range obtaining axial resolution between 17 and 30 microns and lateral resolution between 33 and 94 microns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 15 Current dermatological applications in UBM were reviewed by Thiboutot (1999). Some o f the pathologies discussed were: diseases of the epidermis, psoriasis, diseases of the dermis, scleroderma/morphea, lipodermatosclerosis, radiation- induced fibrosis, wounding healing, mixed epidermal/dermal disease, contact dermatitis/urticaria (tissue edema), skin atrophy, subcutaneous disease, cellulites/abscess, panniculitis/lipomas, skin tumors, basal and squamous cell carcinomas, melanoma, angiomas, port-wine stains, kaposi’s sarcoma, disease of skin appendage, and hidadrenitis suppurativa. If skin pathologies such as skin tumors, basal and squamous cell carcinomas, and melanomas can be detected using high frequency ultrasound efficiently then this diagnosis method could be used in the clinic reducing the need for exploratory surgery. The study of sebaceous glands and hair follicles using high frequency ultrasound would also be useful in assessing the effects of acne medicines and hair loss. Turnbull et al. (1995b) also reported high frequency ultrasound imaging of mutant phenotyping in the mouse embryo. Further high frequency ultrasound of the mouse was conducted by Foster et al. (2002). The mouse was imaged from day 5.5 of embryogenesis to adulthood. Since mice share 90% of our genetic code, the mouse has been a choice for many researchers to study the effects of mutation. Transgenic and mutant mice with cardiac defects were studied to examine the cardiac structure and function using 40 MHz ultrasound imaging by Srinivasan et al. (1998). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 16 1.6 Thesis Aims and Goals The advantage of ultrasonic annular arrays over conventional single element transducers has been in its capabilities to transmit focus at multiple points throughout the depth of field, as well as receive dynamic focus. The capability of annular arrays to focus on both transmit and receive has enabled better resolution throughout the depth of field as compared to single element transducers. Annular array transducers have also been used because of lower costs compared to conventional linear and multidimensional array transducers due to the fewer number of channels required to form an image. In the past annular arrays have been used in diagnostic imaging systems (Aditi et al., 1982; Foster et al., 1989). Today, annular arrays have been replaced by linear arrays and multi-dimensional arrays due to improvements in electronics. Commercial imaging systems can accommodate the large number of radio frequency (RF) channels necessary for imaging with these types of arrays. However, linear and multidimensional array imaging systems are not commercially available at frequencies greater than 20 MHz. The fabrication of these types of transducers can be quite problematic considering the dimensional constraints posed by an increase in frequency. Overall array element isolation and interconnect seem to be the two greatest fabrication difficulties to overcome. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 17 This study describes the development of a high frequency (> 50 MHz) annular array transducer made by mating a poled piece of copolymer P(VDF-TrFE) film to a two sided flexible circuit. The P(VDF-TrFE) material was chosen because of its low acoustic impedance, low dielectric constant and low lateral coupling and higher longitudinal coupling coefficient with respect to PVDF. This study shows that the copolymer material can be bonded to the two sided flexible circuit and housed within a brass tube for RF shielding. The annular array was used in conjunction with a transmit beamformer and dynamic receive focusing in order to image a wire target phantom. Range resolution measurements were made with the annular array imaging system to show improved resolution throughout the depth of field as compared to single element transducers. The details of the methodology involved to design such an array are described in the following chapters. It is also the aim of this work to present preliminary ultrasound B mode images using this transducer in conjunction with the multi-channel digital beamformer (Cao et al., 2003). The use of the multi-channel digital beamformer can implement electronic receive focusing which will improve the lateral resolution throughout the depth of field. Preliminary images of the cornea of the rabbit’s eye were shown. The acoustic modeling of annular array transducers using the same fabrication method described in this work using piezoelectric materials such as PMN-PT and PZT-5H 1-3 composites will be accessed. The feasibility of fabricating a 1 -3 composite annular array transducer using the same flex circuit with R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 18 frequency of 30 MHz will be discussed. New tuning and impedance methodology may make it possible to select the desired center frequency while using the same 9 pm P(VDF-TrFE) film. In order to tune to a specific frequency within the frequency pass band of the material a specific inductor value will be used to cancel the phase at that resonant frequency. Furthermore, impedance matching using a step down transformer and coaxial cable will be incorporated. 1.7 Thesis Outline The following chapters describe in detail the modeling, fabrication, characterization and imaging of the annular arrays developed using this technology. The second chapter details the initial design and modeling that was undertaken to predict the performance of the annular array; this includes transducer electrical equivalent circuit models and ultrasound field models. The third chapter describes the fabrication and characterization of high frequency annular arrays. Transducer characterization includes pulse echo, electrical impedance, insertion loss, crosstalk, and acoustic intensity measurements. The fourth chapter describes the imaging performed with the annular transducer. Preliminary images include wire phantom and rabbit’s eye obtained in-vitro. The fifth chapter describes the future work of using different material composites for annular array transducers. Preliminary results on acoustic modeling, fabrication and testing using a fine scale PZT-5H 1-3 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 19 composite annular array transducer is reported. Also, a tunable center frequency copolymer transducer is fabricated demonstrating the frequency bandwidth of the material. Finally, Chapter 6 summarizes the work accomplished in this dissertation and gives insight on future work to better design and fabricate new devices. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 0 CHAPTER2 ANNULAR ARRAY DESIGN AND MODELING 2.0 Annular Array Design The geometry chosen for this flat aperture annular array transducer was based on the findings of Arditi et al., 1982. A fundamental design for annular arrays is based on equal area for each element. If constructed properly equal area elements will have the same electrical impedance and same phase shift across the aperture. The acoustic path length difference, Ad, d r do can be seen in Fig. 2.1. do is the acoustic path length from the middle of the center annulus to the focal point and dj is the acoustic path length from the radius of an element, rj, of the center annulus to the focal point. /A d d0 ------------------------- Fig. 2.1. Path difference Ad = di-do. permission0f thern . Pynght owner FUrth 21 The path difference is: Ad = d, -d „ (2.1) Furthermore, the path difference for a flat aperture can be rewritten as: Ad: :Vr> 2+dl (2.2) Hence di can be substituted by y r,2 +d„ , therefore: Ad = d„ - d . (2.3) Ad = d f r 2 \Y i l + f v doy (2.4) The above expression can be simplified by the binomial expansion because d0 « v\. = dt (2.5) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 2 Finally the path difference reduces to: Ad = 2 d„ (2.6) The phase shift AO and time of flight At are: At = Ad r. c 2 d 0c seconds (2.7) ^ „ Ad r,‘ A^ = co • At = 27zf— = 7t c d 0A radians (2 .8) where c is the speed o f sound, f is the frequency. Path length differences can be determined for each annulus in the array. Equation 2.6 can be expanded to include radii from the outer annuli: 2 2 d , - d . = ^ ' ^ ^ T + l I 2 dn (2.9) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 23 Since all the annuli have the same equivalent area, each annulus’s calculated phase shift will be the same. Hence if there are no kerfs the radius of each element, n, can be calculated as: If a kerf is included in the design then the radius, q, of each element can be calculated as: where i is the radius integer and kerf is the spacing between elements. The following parameters were used in the initial eight annuli transducer design to determine the resonance of the film, the area of each element, and impedance. The effective inner and outer radii for each element are shown in table 2.1. The active area for an element was chosen based upon its calculated impedance, natural focus and phase shift. An area of 0.7854 mm2 corresponds to an active aperture of 1 mm diameter for the center element. The resonance frequency, f, o f the film is related to the longitudinal velocity, c, of the material and its thickness, t, by the following equation: (2.10) (2 .11) R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. 2 4 The resonance occurs when the thickness is at odd multiples of A/4. A 9 pm P(VDF-TrFE) copolymer (Measurement Specialties Inc., Wayne, PA) piezoelectric layer was chosen to provide a A/4, 63 MHz resonance due to the electroded backing materials used (Brown et al., 2000). The actual center frequency of the array should be lower than this due to attenuation and mass loading of the front acoustic array surface in a load medium of water or tissue. P(VDF-TrFE) copolymer has desirable material properties which make it well suited for the fabrication method used in this work and contributes to the overall good performance of the device. The low acoustic impedance (Z a ) allows for better acoustic matching to water/tissue, the low relative clamped dielectric constant (es /s0 ) has better electrical impedance matching for larger area elements and the low lateral piezoelectric strain constant (dai) contributes to low acoustic coupling between adjacent elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 25 Element # Inner Radius (pm) Outer Radius (pm) Width (pm) 1 0 500 500 2 530 729 199 3 759 909 150 4 939 1064 125 5 1094 1 2 0 2 108 6 1232 1330 98 7 1360 1449 89 8 1479 1561 82 Table 2.1. Geometry for the annular array. The clamped capacitance for a piezoelectric disc in thickness mode resonance is (Kino, 1987): c ^ The clamped dielectric constant is s 33. The area is A (m ) and t (m) is the resonance thickness of the piezoelectric disc. The clamped capacitance was important in determining the approximate electrical impedance of the resonant film. The magnitude of the film impedance is (Rizzoni, 2000): R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 6 F F F (2'14) 2^/c6'o For a plane piston transducer, zj is the location of the last axial maximum. In the near field pressure oscillates but in the far field the pressure decays 1/zj. The equation for this near to far field transition is governed by (Shung, 1992): = ■ r 1 T (2.15) where X is the wavelength of the medium and r is the radius. After the transition point the beam begins to diverge by angle 0, following this relationship (Shung, 1992): 6 - sin '.6 1 A N (2.16) The predicted values of impedance magnitude (|Z|), clamped capacitance (C0 ), and natural focus (fi) are shown in Table 2.2. A center frequency of 50 MHz was chosen for the impedance and the natural focus calculation. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 7 radius area (mm2 ) |Z| (Q) Co(pF) Z | (mm) 1 mm 0.7854 588 5.4 8.33 Table 2.2. Calculated impedance, capacitance, and focus for the center element. 2.1 Equivalent Circuit Modeling A single annulus was modeled using PiezoCAD (Sonic Concepts Inc., Woodinville, WA), based on Krimholtz, Leedom and Matthaei (KLM) model (Krimholtz et al, 1970). The model included a 9 pm P(VDF-TrFE) copolymer piezoelectric layer, a one millimeter diameter aperture (1 pm copper, 0.05 pm chrome, 0.45 pm gold) electrode on a 50.8 pm polyimide film with an unloaded epoxy (Epotek 301, Epoxy Technology Inc., Billerica, MA) backing. A pulse echo response from the model shown in Fig. 2.2, displays the 65 nano-second, -20dB pulse length. The equivalent axial resolution corresponding to this -20 dB pulse length in water is 48 pm (Turnbull et al., 1995a). The Fourier transform of the pulse echo response simulation shows a center frequency o f 56 MHz with 45 % bandwidth in Fig. 2.3. At 56 MHz, the modeled impedance magnitude and phase were 479 £2 and - 74° respectively (Fig. 2.4). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 8 0.5 T3 ■ o CL - 0.5 100 120 140 160 60 80 20 40 0 Time (ns) Fig. 2.2. Simulated time domain pulse echo response (normalized). The time domain pulse echo response has a -20 dB pulse length o f 65 ns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 9 -10 05 « -2 5 -30 -35 -40 60 40 20 Frequency (MHz) Fig. 2.3. Simulated frequency spectrum of the pulse echo response has a - 6 dB bandwidth of 45 %. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3 0 1 4 0 0 -10 1200 -20 -30 1000 -40 ^ 800 -50 -60 600 -70 400 -80 -90 1 0 0 30 Frequency (MHz) Fig. 2.4. Simulated electrical impedance magnitude (solid line) and phase angle (dotted line) for an array element. 2.2 Pressure Field Modeling The pressure field modeling of the annular array was performed using Field II (Jensen et al., 1992). Field II is a set of programs for simulating ultrasound transducer fields and ultrasound imaging using linear acoustics with Matlab (Mathworks, Inc., Natick, MA). The programs are capable of calculating the emitted and pulse-echo fields for both the pulsed and continuous wave case using a variety of arbitrary apertures. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 31 Field II uses linear system theory to calculate the pressure field for either the pulse wave or continuous wave case. The field generated by any transducer is calculated by convolving the spatial impulse response with the excitation function. The spatial impulse response equals the received response for a spherical wave emitted by a point. The pulse-echo response is found by convolving the transducer excitation function, the spatial impulse response from the emitting aperture and the spatial impulse response from the receiving aperture. Using Field II, an eight element annular array was created to model the beam profile in Fig. 2.5 (Snook, 2004). Each annulus was populated with small rectangular elements approximately 30 pm in diameter. The size of the rectangular elements used to approximate the aperture was chosen due to computational restraints. The center frequency and bandwidth of the transducer modeled was 50 MHz and 50%, respectively. The sampling frequency was set to 500 MHz. The following beam profile simulations show different scenarios of transmit and receive focusing. The -6 , -12, -18, -24, -30, -36, -42, -48, -54, -60 dB contour lines are displayed on each o f the two-way intensity field plots. For the following scenarios dynamic receive and transmit focusing was performed with 10 pm (6.7 nano-second) resolution from 6 mm to 18 mm. The following scenarios simulated were: 1) no focusing (Fig. 2.6); 2) receive focusing only (Fig 2.7); 3) transmit focusing only (Fig. 2.8); 4) transmit and receive focusing (Fig. 2.9). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 01__________I__________I __________I __________I__________I__________I ---------------I ---------------1 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 x [mmJ Fig. 2.5. Annular array aperture used in Field II simulations. Each annulus in the array is subdivided into the equal aperture elements. A X kerf width separates each annulus from its nearest neighbor. When no focusing was used in the simulation (Fig. 2.6) only the near field has reasonable imaging capabilities. In the case 2 and 3, implementing either dynamic receive or transmit focusing showed the same result. The -12 dB depth of focus was extended from 6 mm to 18 mm, while - 6 dB was only extend from 6 mm to 9 mm. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 33 The side lobe levels also decreased due to dynamic focusing. However, in a real time imaging system it would be more advantageous to only use dynamic receive focusing. Finally, implementing both transmit and receive dynamic focusing extends the - 6 dB depth of field from 6 mm to 16 mm there by showing a significant improvement over either transmit or receive dynamic focusing alone. Since dynamic transmit focusing is not realistic to perform in real time imaging systems, a single transmit focus at 1 2 mm with dynamic receive focusing was simulated (Fig. 2.10). The result show a - 6 dB depth of field from 10 mm to 14 mm with decreased side lobe levels from 6 mm to 18 mm. This result shows that this is a better focusing scenario than just having either receive or transmit dynamic focusing. Foster et al. (1989) considered a two transmit foci approach with dynamic receive focusing with a 12 element annular array at 4.5 MHz. Similarly, a two transmit foci approach at 50 MHz with dynamic receive focusing was simulated. The beam profile with transmit foci at 10 mm and 12 mm with dynamic receive focusing (Fig. 2.11) shows a slightly extended depth of field as compared to the single transmit focus. Therefore, a two foci transmit system is not ideal for extending the depth of field from 6 mm to 18 mm. The number of transmit foci was increased to further explore what the minimum number of transmit foci were needed to extend the - 6 dB depth of field from 6 mm to 18mm. The following figures: Fig. 2.12,2.13,2.14, show the transmit foci for 4, 7, 13 zones. Ultimately, 13 transmit foci with dynamic receive focusing R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3 4 showed the minimum number of transmit zones need to extend the - 6 dB depth of field from 6 mm to 18 mm. Simulated phantom images confirm this result later in this chapter. 0.5 beam -1 -0.5 Distance(mm) perpendicular to beam axis to Fig. 2.6. Simulated annular array two-way beam profile with no focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 35 22 8 2 0 E 0 3 < D X ! o)15 c „o 0 3 1 1 0 I 5 22 b o -1.5 -1 -0.5 0 0.5 1 1.5 Distance(mm) perpendicular to beam axis Fig. 2.7. Simulated annular array two-way beam profile with dynamic receive focusing only. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. Distance (mm) along b ea m axis 36 20 15 10 5 0 -1.5 -1 -0.5 0 0.5 1 1.5 Distance(mm) perpendicular to beam axis Fig. 2.8. Simulated annular array two-way beam profile with dynamic transmit focusing only. .30 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 37 Distance(mm) perpendicular to beam axis Fig. 2.9. Simulated annular array two-way beam profile with both dynamic transmit and receive focusing. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. -1.5 -1 -0.5 0 0.5 1 1.5 Distance(mm) perpendicular to beam axis Fig. 2.10. Simulated annular array two-way beam profile with a 12 mm transmit focus and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. -1.5 -1 -0.5 0 0.5 1 1.5 Distance(mm) perpendicular to beam axis Fig. 2.11. Simulated annular array two-way beam profile with a 10 mm & 12 mm transmit foci and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 0 20 -60 - 48. 0.5 beam -1 -0.5 Distance(mm) perpendicular to beam axis to Fig. 2.12. Simulated annular array two-way beam profile with 4 transmit zones: 8mm, 10 mm, 12 mm, 14 mm transmit foci and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 41 8 20 ?/ -60 r*<f 0.5 -1 -0.5 Distance(mm) perpendicular to beam axis -1.5 beam to Fig. 2.13. Simulated annular array two-way beam profile with 7 transmit zones: 6 mm, 8mm, 10 mm, 12 mm, 14 mm, 16 mm, 18 mm transmit foci and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 42 c o - '" 7 ■60 0.5 beam -1 -0.5 Distance(mm) perpendicular to beam axis -1.5 to Fig. 2.14. Simulated annular array two-way beam profile with 13 transmit zones: 6 mm, 7 mm, 8mm, 9 mm, 10 mm, 11 mm, 12 mm, 13 mm, 14 mm, 15 mm, 16 mm, 17 mm, 18 mm transmit foci and dynamic receive focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 43 Simulated images of a wire phantom were acquired using the modeled annular array transducer developed in this work. The virtual phantom consisted of vertically spaced targets separated by 1.5 mm each. Twenty five scan lines were acquired per element over a distance of 1 mm corresponding to a 40 pm scan line width. The simulated RF scan lines were automatically beamformed to show the image. For the following scenarios dynamic receive and transmit focusing were performed with 6.7 nano-second resolution from 6 mm to 18 mm. The following scenarios simulated (Fig. 2.15) were: 1) no focusing; 2) one transmit focus and dynamic receive focusing; 3) transmit and receive dynamic focusing. The - 6 dB resolutions were measure from each virtual target for the above scenarios (Fig. 2.16). The - 6 dB lateral resolution at the focus (10.39 mm) was 115 pm after dynamic focusing. Both transmit and receive dynamic focusing improved the lateral resolution throughout the depth field and reduced side lobe levels. The dynamic transmit focusing plays an especially important role in further improving the - 6 dB resolution as compared to just using receive focusing alone, especially in the far field. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 4 - 0.5 0 0.5 - 0.5 0 0.5 - 0.5 0 0.5 mm Fig. 2.15. Simulated phantom images: no focusing (left), one transmit focus and dynamic receive focusing (middle), both transmit and receive dynamic focusing (right). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 45 500 Sim. w/o Dynamic Focus - 0 - Sim. w/ T&R Dynamic Focus Sim, w/ R Dynamic Focus 450 400 350 100 50. Range (mm) Fig. 2.16. Simulated lateral resolution (- 6 dB) from wire targets before and after dynamic focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 46 Another Field simulation with transmit and receive focusing at 12 mm from the transducer (Fig. 2.17) was performed to justify the number of elements chosen for this annular array. The aperture was expanded by increasing the number of equal area elements resulting in an improvement of lateral resolution by the narrowing of the - 6 dB beam width. Increasing the number of elements decreased side lobe levels. Finally, the sensitivity of the transducer can be improved by increasing the number of elements (Brown et al., 2004). 1 annuli annuli annuli annuli annuli annuli annuli annuli 15 annuli = -50 < -6 0 -1 -0.5 0 0.5 1 Lateral Distance (mm) Fig. 2.17. Simulated two-way radiation patterns for annular array with increasing number of elements at 12 mm from the transducer. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. 47 The effect of increasing aperture while keeping the same number of elements is shown in Fig. 2.18 for a two way radiation response with transmit and receive focusing at 12 mm from the transducer. A 2 mm, 3 mm, and 4.4 mm aperture was simulated with eight elements each with equal area and a 30 pm kerf. The center frequency was kept at 50 MHz with 50 % bandwidth. As what can be seen is that the smaller aperture array has lower side lobes than the larger apertures. The reason for this is because a small aperture array has a small path length difference between adjacent elements therefore a smaller phase shift. Therefore having a smaller phase shift between elements is desirable. The depth of field is also limited by increasing frequency. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 48 — 2 mm aperture — *-• 4.4 mm aperture — 9— 6 mm aperture -10 -20 1 - 4 0 ■ q. E -50 -60 -70 -80 0.5 Lateral Distance (mm) -0.5 -1.5 Lateral Fig. 2.18. Simulated two-way radiation patterns for annular array with increasing aperture size at 12 mm from the transducer. Finally the center frequency: 30 MHz, 50 MHz, 80 MHz, each with 50 % bandwidth was varied to see the effects on beam width and side lobe levels (Fig. 2.19). Transmit and receive focusing was implemented 12 mm from the transducer face. The results were as expected such that the beam width decreased with an increase in frequency. However, as a result of keeping the same aperture size and thus the same acoustic path difference when the wavelength decreased the phase shift across the aperture increased causing the side lobes to rise. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 9 30 MHz 50 MHz 80 MHz -10 -20 -40 E -50 -60 -70 -80 -1.5 -0.5 Lateral 0.5 Lateral Distance (mm) Fig. 2.19. Simulated two-way radiation patterns for annular array with increasing frequency at 12 mm from the transducer. The directivity pattern (one-way) for each individual annulus was also modeled, Fig. 2.20, and was later used to compensate the measured insertion loss for diffraction (Snook et al., 2005). The side lobe levels shown in this image increased for the outer annuli due to diffractive effects. The directivity pattern for the entire 8 element array at a 12 mm focus is shown in Fig. 2.21. The side lobe levels were found to be 37 dB lower than the main lobe. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 50 _ -1 0 TJ (D ,N 8 -20 o cd-30 ~ o < L > T 3 5 - 4 0 Q . E < -50 -60 Center Element 2nd Element 3rd Element 4th Element 5th Element 6th Element 7th Element 8th Element -80 -60 -40 -20 0 20 Angle (degrees) 40 60 80 Fig. 2.20. Field II simulated directivity pattern for each annular array element at 12 mm from the transducer. As expected, due to a decrease in width, side lobe levels were higher in amplitude for the outer annuli. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 51 -10 2-20 in -30 XJ 0 X J i - 4 0 Q . -50 -60 -80 -60 -40 -20 Angle (degrees) Fig. 2.21. Field II simulated directivity pattern for the 8 element annular array transducer at a 12 mm focus. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 52 CHAPTER 3 ANNULAR ARRAY FABRICATION AND CHARCTERIZATION 3.0 Annular Array Fabrication Issues There is a degree of difficulty to manufacture high frequency annular arrays due to the small scale size required at frequencies above 50 MHz. If a piezo-ceramic material is used a tremendous amount of care must be taken to mechanically lap it to the M2 resonance thickness. Since most piezo-ceramics have high acoustic impedances on the order of 30 MRayls, MA acoustic matching layers on the order of 10 pm need to be added to the substrate. Again, lapping these matching layers can be very time consuming and meticulous. The advantage of using a copolymer film is that it does not need to be lapped nor does it need acoustic matching layers. Interconnect methods typically used for high frequency annular array transducer manufacturing include ultrasonic wire bonding and low temperature soldering (Snook, 2004; Brown et al., 2004). These interconnect methods can be problematic when trying to interconnect to a small element on the order o f two wavelengths or less at high frequency. Therefore, in this study the copolymer substrate was ohmic bonded to a flexible circuit consisting of a two layer gold electroded polyimide surface. In choosing this fabrication method the annular array transducer can be more conducive to mass production. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 53 3.1 Annular Array Flexible Circuit Design A CAD schematic of the annular array flexible circuit interconnect is shown in Fig. 3.1. The gray traces are the active annuli and ground plane on the top side of the flex while the black traces are on the bottom side connected to the top through micro vias. The bottom traces terminate to solder pads for coaxial cable interconnect. A ground plane surrounds the entire circuit on the top side and is connected to the bottom side through vias. The total annular array aperture was 3.12 mm with each annulus area of 0.7854 mm2. The flexible circuit was fabricated at Microconnex (Snoqualmie, Wa.). The advantage of using a double sided flexible circuit for array interconnect was the ability to have the electrode pattern bonded to each element on the top layer with micro vias connecting the (+) traces to the cable interconnect pads on the bottom side of the circuit. Thus the array could be fully shielded from the outside RF noise. The top layer electrode pattern consisted of eight annuli of equal area with a 30 pm separation; the electrode thickness consisted of 1 pm copper, 0.5 pm chrome, and 0.45 pm gold. The center electrode had an aperture o f 1 mm. Each 30 pm via hole was electroplated with copper. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 54 Fig. 3.1. A CAD of the annular array flexible circuit. 3.2 Annular Array Fabrication Procedure The fabrication procedure began by cleaning the flexible circuit with acetone, alcohol, and deionized water. The flex circuit was then dried using compressed air. Short wires were soldered to the solder pads with a low temperature indium-based solder (#1E) from Indium Corporation of America (Utica, NY). The bottom of double sided tape was attached to the top of a glass plate and then the non adhesive side of a polyimide tape was placed on top of double sided tape. The annuli pattern of the flex circuit was placed face down on the top of the adhesive side of the polyimide tape. Then the flex circuit was bent upward and housed inside a custom R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 55 machined brass tube. Care was taken to ensure that the annuli were concentric with the housing. The brass tube was filled with degassed non conductive epoxy (Epotek 301). The epoxy was degassed again after filling to ensure no air bubbles were present. The parts were placed in a dry box for 24 hours allowing the epoxy to cure. The brass housing and flex circuit were removed from the adhesive tape. The parts were cleaned and dried again using the procedure above. A schematic of the flexible circuit housed inside of the brass tube is shown in Fig. 3.2. Adhesion promoter (Chemlok AP 131, Lord Corp., Erie, PA) was applied to the surface of the annuli electrodes and allowed to dry of one hour. Nine micron P(VDF-TrFE) piezoelectric material was applied to the top side of the flexible circuit interconnect using non conductive epoxy, Epotek 301. A jig was used to apply pressure to the copolymer material and secure it in place. The jig was placed inside the dry box at room temperature for 24 hours. Bonding the flexible circuit directly to the piezoelectric material has provided a reliable method for fabricating high annular array transducers. Inductors were then soldered to the short wires. The one meter length, coaxial cables, 77 Cl, 40 AWG, coax # 171-1019-XX (Precision Interconnect, Portland OR), were soldered to the inductors (1.2 pH) attached to the short wires. The inductors impedance matched each element to 50 Cl at resonance by canceling the negative phase of the transducer. The coaxial cables were characterized; see Table 3.1, at 50 MHz using the same method as described by Cannata, et al. (2003). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 56 Ground Plane u/Au Electrodes 9 pm P(VDF-TrFE Brass Housing Backing Epoxy (3 MRayls) Flex circuit Ground Plane Solder Pad (+) Coaxial cable solder pads Fig. 3.2. Annular array design cross section. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Fig. 3.3. The annular array transducer shows the copolymer film attached to the front surface of the flex circuit and housing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 58 Property PI #171-1019-XX Characteristic Impedance (Zo) 77.27 - 2.06i Q Propagation Constant (y) 0.077 + 1.25i Propagation Velocity (Vp) 2.52 x 108 m/s Resistance/Unit Length (r) 8.49 Q/m Capacitance/Unit Length (c) 51.4 pF/m Inductance/Unit Length (1) 0.306 pH Conductance/Unit Length (g) 560 pS Attenuation/Unit Length 0.67 dB/m Table 3.1. Properties of the 75 Q coaxial cable (PI #171-1019-XX, Precision Interconnect, Portland OR) characterized at 50 MHz, used to impedance match annular array elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 59 3.3 Annular Array Characterization The pulse echo response, electrical impedance, insertion loss and crosstalk were measured. Electrical impedance measurements were taken using an HP4194 (Agilent Technologies, Englewood, CO) impedance analyzer equipped with Z-probe attachment by connecting the coaxial cables of each annulus to the active and ground calibrated/compensated clips on the impedance probe. The water loaded impedance measurement for each annulus with impedance matching is shown in Fig. 3.4. The electrical impedance magnitude and phase at resonance before and after matching is shown in Table 3.2. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2( 30 40 50 60 70 80 90 100 Frequency (MHz) 100 O ) -100 100 80 40 20 Frequency (MHz) Fig. 3.4. Electrical impedance magnitude (top) and phase (bottom) for each annulus in water. Each element was impedance matched with a 1.2 pH inductor and a A/4 75 Cl coaxial cable. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 61 Element # |Z| D . without matching 0 ° without matching |Z| Q with impedance matching O ° with impedance matching 1 549 -84 52 0 2 552 -84 53 0.616 3 536 -84 56 0.0245 4 539 -84 57 0.395 5 546 -84 52 0.493 6 548 -84 60 0.22 7 549 -84 52.5 0.22 8 501 -84 60 0.515 Table 3.2. Impedance measurement of annular array elements before and after matching at 55 MHz. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 2 The pulse echo analysis consisted of measuring the received echo pulse from the reflection of a quartz target in a deionized water bath (Fig. 3.5). A Panametrics 5900 pulser/receiver (Waltham, MA) was used to excite each annulus and receive the echo waveform. The received echo waveform was digitized and displayed on a Lecroy LC534 (Chestnut Ridge, NY) oscilloscope using 50 Q coupling. The center frequency and - 6 dB bandwidth were measured for each annulus and listed in Table 3.3. The pulse echo response from the center element is shown in Fig. 3.6 and corresponding normalized frequency spectrum Fig. 3.7 for the annular array with impedance matching. Oscil oscope ---------i . C a ; -------- . 1 » ' Pulser/Reciever Transducer Water Bath Quartz Fig. 3.5. Pulse echo analysis setup. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Amplitude (normalized) 63 0.5 -0.5 0 50 100 150 200 Time (ns) Fig. 3.6. Measured time domain pulse echo response of the center array element with a -20 dB pulse length of 60 ns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Magnitude (dB) I I l 6 4 -25 -30 -35 80 30 Frequency (MHz) Fig. 3.7. Normalized frequency spectrum of the center array element with a - 6 dB bandwidth of 49%. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 65 Element # Center Frequency (MHz) -6 dB Bandwidth (%) -20 dB pulse lengths (ns) 1 55 49 60 2 55 56 56 3 55 54 57 4 55 56 59 5 55 47 63 6 56 39 69 7 60 39 86 8 52 39 91 Table 3.3. Pulse echo response of annular array elements with impedance matching. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 6 The two way insertion loss for each annulus was measured. The amplitude of the sinusoid signal from an arbitrary function generator (Sony/Tetronix model AFG2020, Beaverton, OR) set in burst mode was recorded with a 50 G coupling on the oscilloscope for various frequencies over the array’s pass-band (Fig. 3.8). An array element was then connected to the function generator with the oscilloscope set at IMG coupling. In order to have ample signal to noise ratio the return peak echo amplitude signal from the quartz target at the element’s center frequency was measured 6 mm from the transducer. The initial insertion loss is calculated using the following equation: IL = 20 log ' V ' yV, , dB (3.1) where Vo is the peak-to-peak voltage of the received echo waveform; Vi is the peak- to-peak voltage o f the initial burst. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 7 Oscilloscope I 50Q 2 1MQ Water Bath Quartz Fig. 3.8. Insertion loss measurement setup. The signal loss from the attenuation in water at 2.2 x 10'4 dB/mm-MHz2 (Lockwood et al., 1994) and transmission into the quartz target at 1.6 dB (Selfridge, 1983) was compensated in all insertion loss calculations. Loss due to diffraction was also compensated for in the final measurement by using the modeled one-way angular response for each annulus V(cp) (Snook et al., 2005). The equation used to calculate the diffractive loss from Snook et al., 2005 was: R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 68 j V v ’M s* DL = 10 log ^ -------------- dB (3.2) o where cp/ and c p 2 are the capture angles from the inner and outer radii to twice the focal distance of the initial measured insertion loss (Fig. 3.9) and V(cp) is the amplitude of the one way modeled angular response (Snook et al., 2005). The numerator represents the power over the capture angle. Reflector Interface inner d Fig. 3.9. Annular array element and its mirror image used to calculate the diffractive loss (From Snook, 2004). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 6 9 Futhermore, from geometry < p i and < p o can be calculated as: ( j) x = tan ( j) 2 - tan d rad \ u J ( p \ o uter v d j rad (3.3) (3.4) The compensated diffraction and uncompensated diffraction insertion loss values for each annulus at 55 MHz are recorded in Table 3.4. Element # Diffraction Diffraction _____________ Uncompensated IL (dB) Compensated IL (dB) 1 33.42 31.98 2 42.06 32.69 3 43.86 32.05 4 47.00 33.07 5 48.26 33.29 6 50.73 34.45 7 52.08 34.68 8 53.67 36.27 Table 3.4. Annular array insertion loss measurement compensated for attenuation in water and diffraction effects. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 0 The combined electrical and acoustical crosstalk was measured between adjacent elements (Fig. 3.10). The annular array transducer was placed in a deionized water bath opposite an absorptive rubber target for this measurement. The arbitrary function generator set in burst mode and the peak amplitude of the sinusoid signal across each element was recorded with a 1 MQ coupling on the oscilloscope for frequencies over the device pass-band. The peak signal amplitude of each adjacent element was simultaneously measured (1 MQ coupling) across the device pass-band (Fig. 3.11). The crosstalk between adjacent elements was less than - 29 dB at 55 MHz and deemed to be acceptable isolation for this study. Lecroy Oscilloscope Function Generator U W a 1 mq Coupling Ch 1. Ch 2. • • Element 1 i Element 2 Water Bath Rubber Target Fig. 3.10. Cross talk measurement setup. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 71 -20 Annuli 1&2 Annuli 2&3 -®- Annuli 3&4 Annuli 4&5 ■ - * — Annuli 5&6 Annuli 6&7 Annuli 7&8 _530 40 50 60 70 80 Frequency (MHz) Fig. 3.11. Crosstalk measurement between adjacent elements. A crosstalk model was developed by Guess et al., 1995 to further understand intuitively how current moves from one element across the kerf to the adjacent element (Fig. 3.12). Voltages Ei and E2 are from the two neighboring elements Zk and Z j. The crosstalk impedance is defined as Z cr. The currents now can be defined as ii, i2, i3 where ii is the current flowing through Zk; i2 is the current flowing through Z cr; i3 is the current flowing through Z j and R l , the load on the oscilloscope. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 72 C R Fig. 3.12. Crosstalk model from Guess et al., 1995. The following equations were deduced from the Guess et al. (1995) crosstalk model to derive a crosstalk equation which could be used after measuring the electrical impedances Z c r (electrical impedance between Z k and Z j) and Z j (electrical impedance from the adjacent element) using the HP4194 impedance analyzer: i , = - ^ P-5) z , • _ E\ ~ ^2 2 ' ZC R ( 3.6 ) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 73 Now Z j and R^are combined in parallel, therefore i3 becomes: E, i , = ------------- — 3 Z; R i 'k l (3.7) Z,+Rl Therefore the sum of the currents at the node is: h — *3=0 (3.8) E\ E2 _ E2(Z, + Rl ) _ ^ (3.9) Z(:r Z, ■ Rl = (3.10) E, ZC R (Zi + R l)+Zj R l Cross Talk = 20-log— (3.11) E, Cross Talk = 20 log 52-50 405(52 + 50)+52-50 -24.5 5dB (3.12) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 4 The crosstalk calculation at 55 MHz was made from: Z* = 52 £2 (impedance magnitude measured from the center element of the array in water), Z cr = 405 Q (impedance magnitude measured from the center element to the adjacent element in water) and R l = 50 Q (load across the impedance analyzer). This calculated result of -24.55 dB representing the electrical part of the total crosstalk was worse than the measured crosstalk in the previous experiment of -30.14 dB. There is no apparent reason why the electrical impedance crosstalk measurement should be more than the measured summed acoustic and electrical crosstalk measurement in the previous experiment. Actually, the expected result was that it would be less. Ideally, it would be beneficial to have either large electrical impedance between elements or a lower element impedance to provide the maximum isolation and lowest crosstalk. 3.4 Acoustic Output Measurements A high frequency exposimetry system was developed by Snook, 2000. This system included a water tank, 3-axis position control system (Burleigh, Inc., Fishers, NY), a calibrated 0.04 mm needle hydrophone (Precision Acoustics, Dorchester, Dorset, UK) with pre amplifier and booster amplifier (25 dB gain with 125 MHz bandwidth), and Lecroy LC534 oscilloscope. This system was used with the annular array transducer developed in this work in order to measure the intensity output of the transducer. The needle hydrophone was calibrated up to a frequency of 60 MHz R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 75 on 10/29/1999 (Reference U 1176) by the National Physics Laboratory (Teddington, UK). The calibrated end-of-cable loaded sensitivity is shown in Figure 3.13. The values in between measurements were interpolated using a cubic spline function in Matlab. The end of cable loaded sensitivity is much lower at high frequencies greater than 20 MHz. The effective spot size of the hydrophone is 100 pm (Snook, 2000) which is still less than the modeled beam width of the transducer. 0.16 0.14 20 Frequency (MHz) Fig. 3.13. End of cable loaded sensitivity for needle hydrophone. The Panametrics pulser was set to 32 pJ energy level with a PRF of 200 Hz and attached to an eight way power splitter (Mini Circuits, ZCSC-8-1, Brooklyn, R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 6 NY). An analog delay beamformer was used to create a transmit focus for the array and each delay line cable from the corresponding beamformer was attached to the cable of each annuli. The analog beamformer will be described in more detail in the following chapter. The transducer was placed inside of the water tank and aligned on axis with the hydrophone. The acoustic beam from the transducer was focused to the tip of the hydrophone. The received voltage wave form, V(t), was sampled at 1 GHz and signal averaged (31:1) on the oscilloscope. The data was saved to file for further processing (Fig. 3.14). The acoustic output measurements can be calculated from V(t). A detailed description of the definitions of all the acoustic output measurements can be found from the US Food and Drug Administration (FDA, 1985), the American Institute of Ultrasound in Medicine (AIUM, 1992), the National Electrical Manufacturers Association (AIUM/NEMA, 1992) and the International Electrotechnical Commission (IEC, 1992). The pulse intensity integral was calculated from V(t) using the following equation (Snook, 2000): k 2(o P I I = w ^ m ^ ' i p is the density of the propagation medium (kg/m ), c is the phase velocity in the medium (m/s) and M ^fc) is the loaded end-of-cable sensitivity (pV/Pa) at the center R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 7 frequency. The pulse intensity integral is shown in Fig. 3.15. It is also used to calculate the spatial peak pulse average intensity (Isppa) and the spatial peak temporal average intensity (Ispta)- The following equations were used to calculate Isppa and Ispta (Snook, 2000): I S P P A ~ m a X P II (3.14) where T is the pulse duration. Ispta- Isppa*T*PRF* 1000 (3.15) The FDA regulation policy uses derated values of intensity to account for attenuation in tissue. The derated Isppa and Ispta values were also calculated by multiplying by the derated factor (Snook, 2000): d f = g - 0 .2 3 x 0 .3 x * = ^ -0 .0 0 6 9 f L .zt ^ Table 3.5 shows the calculated intensity values for the annular array transducer. Notice that the intensity values are very below the FDA maximum acoustic output R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 8 levels as stated in the “510(k) Guide for measuring and reporting acoustic output of diagnostic ultrasound medical devices,” Food and Drug Administration, Washington, D.C., (1985). The Ispta.3 must be below 17 mW/cm2 for ophthalmic uses. However, these values again are again dependent on the PRF and the amount of energy used to drive the transducer. This device has not been approved by the FDA nor has any of the acoustic output measurements made in this work been reported to the FDA. 0.03 0.025 0.02 0.015 0.01 05 I 0.005 -0.005 - 0.01 -0.015 7.5 7.6 7.7 7.8 7.9 Time [us] Fig. 3.14. Measured received pulse wave form from the hydrophone. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 9 0.8 ,0.6 < N ,0.4 0.2 7.5 7.7 7.8 7.9 8 7.6 Time [us] Fig. 3.15. Pulse intensity integral. Center Frequency 45 MHz Ispta 0.170 mW/cm2 Pulse Duration 36.25 ns I sppa 23.40 W/cm2 PRF 200 Hz IsPTA.3 0.004 mW/cm2 Focal Depth 1.18 cm IsPPA.3 0.60 W/cm2 Table 3.5. Acoustic output measurements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 0 CHAPTER 4 ANNULAR ARRAY IMAGING 4.0 Introduction to Imaging High frequency imaging applications include intravascular ultrasound, ophthalmologic and dermatologic ultrasound imaging (Foster et al., 2000). The penetration depth of ultrasound varies according to frequency and attenuation (Foster et al., 1993). The minimum separation of targets perpendicular to the beam’s path is called lateral resolution. Since the beam width varies with depth along the beam’s path, the lateral resolution is proportional to f and A , (Foster et al., 1993). f o If a constant velocity of 1540 m/s is chosen for ct and the center frequency of the transducer is varied from 1 to 100 MHz then in Fig. 4.1 the resolution is shown for f*’s 1, 2, 3. As it can be seen the lowest f* has the best lateral resolution. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 81 F/1 F/2 F/3 V) c 2 o E Frequency (MHz) Fig. 4.1. Lateral resolution as a function of frequency for varying f/#,s. The axial resolution ( A r) is the separation of two targets along the beam’s path. The relationship between pulse duration, Pd, speed of sound, ct, in tissue, and Af the bandwidth (MHz) define axial resolution (Foster et al., 2000): A. ~ (4.2) * 2 The axial resolution improves with broader bandwidth shown in Fig. 4.2. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 2 800 700 600 500 400 o E 300 200 100 100 Bandwidth (MHz) Fig. 4.2. Axial resolution as a function of bandwidth in MHz. During the propagation of an ultrasound wave through tissue its acoustic energy is degraded as a function of distance (Shung, 1992): „ r , -«* (4'3> P = P0 e a is the attenuation coefficient (dB/cm). This dissipation of acoustic energy into thermal energy is called attenuation due to absorption and scattering from losses in the tissue and at boundaries. Attenuation effects increase at higher frequencies and depend on tissue type. The attenuation can be defined as (Ye et al., 1995): R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. A frequency dependence y, of 1.2 (Foster et al., 1993) was assumed in the following calculation along with initial attenuation, ao (ao values for eye (0.01 dB/mm), skin (0.1 dB/mm), and blood (0.04 dB/mm). The attenuation coefficients for eye, skin, and blood can be found in Ye et al., 1995; Pan et al., 1998; Lockwood et al., 1991 (Fig. 4.3). The penetration as a function of frequency for 80 dB dynamic range can be found in Fig. 4.4. The depth of penetration decreases as the frequency is increased. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 84 25 eye blood skin 20 m 15 100 Frequency (MHz) Fig. 4.3. Attenuation for eye, skin, and blood as a function of frequency. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 85 — eye — blood * — skin 2 | 1 0 1 10 o 10 o 1 ,2 10 10 10 Frequency (MHz) Fig. 4.4. Maximum penetration as a function of frequency. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 6 4.1 Anatomy of the Eye An anatomical drawing of the eye is shown in Fig. 4.5. The cornea is the clear front window of the eye. The cornea acts as the first refractive surface that bends light toward the lens. The iris is the colored part of the eye. The iris serves as an aperture stop which regulates the amount of light that enters the eye. The pupil is the opening in the middle of the iris. The pupil accommodates the amount of light that is passed through the cornea by changing size. The lens is the most refractive transparent structure inside the eye that focuses light rays onto the retina. cornea conjunctiva' sclera c h o r o id ^ f . cham ber a n g le ” retina aq u eo u s vitreous Fig. 4.5. Anatomy of the Eye (Eyesight Insight, [web page] Feb 2000). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 87 4.2 Anatomy of the Skin The anatomy of the skin can be seen in Fig. 4.6. The epidermis is the thin outer layer of the skin. The layers of the epidermis include the stratum corneum (homy layer), stratum granulosum, stratum spinosum (Malphigian layer) and stratum basale (Thiboutot, 1999). The keratinocytes differentiate from the basale layer in columns to the corneum appearing as squamous cells. The dermis is the middle layer of the skin containing blood vessels, lymph vessels, hair follicles, sweat glands and sebaceous glands. The dermis is mostly connective tissue of type one collagen fibers. Under the dermis is subcutaneous fat of adipose tissue. The dimensions of the various skin structures are in Table 4.1. ctflajsfi cells. smatgtamd Fig. 4.6. Anatomy of the Skin (University of Maryland Medicine, [web page] January 2001). R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. 88 Skin Structure Depth (pm) Width (pm) Epidermis 50-100 Dermis 120-180 Hair Follicle (t) 150-400 50-100 Hair Follicle (v) 800-1200 30-60 Sebaceous Gland 150-250 150-500 Sweat Gland 200 400 Table 4.1. Skin Structure dimensions (from Thiboutot, 1999). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 89 4.3 Wire Phantom Imaging Images o f a wire phantom were acquired using the annular array transducer developed in this work. The phantom consisted of vertically spaced 0.02 mm diameter tungsten wires (California Fine Wire Co., Grover Beach, CA) separated by 1.5 mm. For this imaging test a single transmit focus analog beamformer was made using specific length delay line coaxial cables (RG/58 C5779.18.10 General Cable, Highland Heights, KY). This coax was fully characterized at 50 MHz to calculate the propagation velocity (Table 4.2) in order to determine the proper delays for each array element. The transmit time delays and the corresponding cable lengths were calculated based on the acoustic path differences from each annulus to a 12 mm focal point (Table 4.3). Property G.C. RG/58 C5779.18.10 Characteristic Impedance (Zo) 51.15 -0.244iQ Propagation Constant (y) 0.013 + 1.40 i Propagation Velocity (Vp ) 2.246 x 108 m/s Resistance/Unit Length (r) 1.08 Q/m Capacitance/Unit Length (c) 87.03 pF/m Inductance/Unit Length (1) 0.249 pH Conductance/Unit Length (g) 118 pS Attenuation/Unit Length 0.113 dB/m Table 4.2. Properties of the 50 Q coaxial cable (General Cable RG/58 C5779.18.10) characterized at 50 MHz, used in the fixed focus transmit beamformer. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 9 0 Delays (ns) Cable Length (m) Element 61 13.64 1 53 11.88 2 45 10.04 3 36 8.14 4 28 6.17 5 19 4.16 6 9 2.10 7 0 0 8 Table 4.3. Calculated time delays and corresponding cable lengths for a single focus (12 mm) transmit beamformer. A Panametrics 5900 pulser set at 32 pj was used in conjunction with an 8 way power splitter (MiniCircuits, ZCSC-8-1, Brooklyn, NY) to deliver an equal amplitude pulse to each annulus on transmit. The carefully trimmed coaxial delay lines were placed between the power splitter and the annuli. For receive a one meter length coaxial probe connected a single annulus to the receiver portion of the Panametrics 5900. Twenty five scan lines were acquired per element over a distance of 1 mm corresponding to a 40 pm scan line width. The averaged (32:1) received echo waveform was displayed on an oscilloscope (50 £2 coupling) and downloaded to file for post processing. A total of eight RF scan lines were recorded at each position in order to collect the received waveform from each annulus. Dynamic receive focusing time delays were calculated based on every sample corresponding to a 4 ns resolution and applied to each received waveform. Once all the waveforms were delayed they were coherently summed to form one corresponding scan line. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 91 The phantom image before and after dynamic focusing is shown in Fig. 4.7. The lateral resolution was greatly improved by using the dynamic focusing technique. The - 6 dB lateral resolution at the focus was 108 pm after dynamic focusing. The lateral beam profile, before and after dynamic focusing, is shown for a wire target at 10.391 mm from the transducer in Fig. 4.8. Figure 4.9 shows a comparison between the measured and simulated lateral resolution throughout the depth of field before and after dynamic focusing. Dynamic focusing improved the lateral resolution throughout the depth field and reduced side lobes. The phantom images show that the geometric focus was actually shifted closer toward the transducer to 10.5 mm rather than 12 mm. This result was most likely due to diffraction effects such as a spherical aberration (Kossoff, 1979). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 92 7 8 9 10 11 12 13 14 15 -0.5 0 0.5 -0.5 0 0.5 Scale (mm) Fig. 4.7. Measured wire target phantom images (48 dB dynamic range): before dynamic focusing (left) and after (right) dynamic focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 93 — After DF — Before DF -10 -a -30 -40 0.5 Lateral Distance (mm) Fig. 4.8. The lateral beam profile measured from a 20 pm wire target at the focus before and after dynamic focusing (DF). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 94 600 ■ ■ ■ © ■ ■ • M easured w /o Dynam ic F o cu s M easured w / D ynam ic F ocu s Sim ulated w / Dynam ic F o cu s 500 100 R ange (m m ) Fig. 4.9. Measured and simulated lateral resolution (- 6 dB) from wire targets before and after dynamic focusing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 95 4.4 Annular Arrav Imaging System The annular array imaging system was developed by Cao et al. (2002, 2003) to be used with the high frequency annular array transducers. Cao et al. designed a six channel digital beamformer for high frequency annular array ultrasound transducers (Fig. 4.10). The beamformer allows programmable time delays for each element in the array to enable both focusing during transmit and dynamic focusing during receive. PC: Image display User interface T ransducer Position C 'ontrol Digital Back-End Digital I/O C ard (N l PC 1-6534) transm it Heam('orming I Digital Beamforming D igital Signal Kcceive Processor I # # 1 licam form ing (A D S P -2 10651.) (X ilinx I PGA) Annular Array Transducer < S :E r:> T ransceiver Gain Control T A m plification and TGC A D conversion Fig. 4.10. Annular array imaging system (From Cao et al., 2003). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 96 4.5 Preliminary Ultrasound Images Preliminary ultrasound images were acquired using the annular array ultrasound system with the transducer developed in this work. A comparison was made between a single element transducer of equal aperture to that of the annular array. Ultrasound images of a wire phantom and rabbit cornea in-vitro were compared. A linear spacing phantom contained wire targets placed on machined steps of 1.5 mm vertical spacing and 0.65 mm lateral spacing using 0.02 mm diameter tungsten wires (California Fine Wire Co., Grover Beach, CA) secured in all specified locations. The wire target image from the annular array with cable only is shown in Fig. 4.11. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 97 1 2 3 4 5 Lateral distance (mm) Fig. 4.11. Wire target image using annular array. Since the annular array has low sensitivity and the receiver electronics has a large noise floor the signal to noise ratio of the image is quite small. The ultrasound image of the rabbit’s cornea in-vitro was also acquired using the annular array transducer (Fig. 4.12). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 98 1 2 3 4 5 Lateral distance (mm) Fig. 4.12. Ultrasound image of rabbit’s cornea in-vitro using the annular array. The top layer o f the rabbit cornea is visible in the ultrasound image. The curvature of the rabbit cornea decreases the intensity o f the reflected ultrasound wave incident to the transducer. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 99 The imaging performance of the annular array was compared to a 3 mm aperture P(VDF-TrFe) single element transducer. The single element transducer had a center frequency of 45 MHz and a bandwidth of 75 %. The insertion loss measured at the center frequency was -35 dB. The wire target phantom image can be seen in Fig. 4.13. 1 2 3 4 5 Lateral distance (mm) Fig. 4.13. Wire target image using single element transducer. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 100 As seen by the phantom image the lateral resolution is degraded throughout the depth of field when using a single element transducer with a single focus point. The ultrasound image of the rabbit’s cornea in-vitro was acquired using the single element transducer (Fig. 4.14). The top and bottom layer of the rabbit cornea and the lens are visible in the ultrasound image. 1 2 3 4 5 Lateral distance (mm) Fig. 4.14. Ultrasound image of rabbit’s cornea in-vitro using the single element transducer. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 101 4.6 Annular Imaging System Limitations A major limitation of the annular array imaging system is the low signal to noise ratio. Since no signal averaging was used in this imaging experiment the image quality was quite poor and noisy. Improvements to the electronic system are necessary in order to compensate for the high insertion loss of the P(VDF-TrFE) annular array transducer. The excitation pulse used for this system was a unipolar spike with an amplitude of -60 volts and a - 6 dB pulse length of 12 ns. The 9 pm copolymer film has a max AC operating field of less than 25 V/pm (Measurement Specialties, Wayne, PA). An increase of voltage above the max operating field would cause nonlinearity in the pulse echo response. Modifications are necessary to the pulser excitation shape and amplitude in order to increase signal to noise ratio during transmit. A monocycle pulse shape would be beneficial because its center frequency response would excite the same frequency pass band as the transducer. An increase in transmit signal amplitude to 200 volts peak to peak would provide an increase in SNR. The receiver is also too noisy and its noise floor should be minimized. The number of channels can also be expanded to eight so that the entire array can be used. Currently, 8 bit analog to digital (A/D) conversion is being used which limits the number of gray levels to 256. Higher bit A/D can increase the dynamic range of the system. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 102 CHAPTER 5 DEVELOPMENT FEASABILITY OF COMPOSITE AND TUNABLE COPOLYMER ANNULAR ARRAY TRANSDUCERS 5.0 Introduction The P(VDF-TrFE) annular array transducer had a low acoustic impedance and low d3 i which resulted in a device with adequate bandwidth and low acoustic crosstalk. However, the insertion loss was still high therefore in order to improve the sensitivity it would be advantageous to use a piezoelectric material with a higher coupling coefficient. However, if a traditional ceramic were used the array would suffer from a higher crosstalk. It would therefore be beneficial to pick a material with lower acoustic impedance for better coupling to load medium thereby improving sensitivity and bandwidth, and a lower d3 i for lower element cross coupling. A good compromise between copolymer and ceramic would be a 1-3 composite annular array transducer. The feasibility of fabricating a 1-3 composite annular array transducer using the same flex circuit with a center frequency of 30 MHz has been accessed. The same characterization methods described earlier in this work was used so that the characterization measurements of the composite annular array could be compared to the copolymer annular array. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 103 Tuning and impedance methodology make it possible to select the desired center frequency while using the same 9 pm P(VDF-TrFE) film. In order to tune to a specific frequency within the frequency pass band of the material a specific inductor value will be used to cancel the phase at that resonant frequency. Furthermore, impedance matching using a step down transformer and coaxial cable was incorporated. 5.1 Piezoelectric Composites Previous work from Ritter (2000) and Cannata (2004) both show the use of a 2-2 composite for the fabrication of linear array transducers at or above 30 MHz. A 2-2 composite is a good choice for a linear array transducer because the length of the ceramic is much longer than its thickness. For an annular array, it would be better to use a 1-3 composite material because of its lower acoustic impedance, homogeneity of elements, and lower cross coupling. A 1-3 composite high frequency single element transducer was reported by Snook et al. (2002). The 1-3 composite used had a 45 % volume fraction and fiber diameters of 17 pm (Meyer, 1998). The 1-3 composite was a PZT-5A fiber composite in an Epotek 301 epoxy. Since PZT-5A has a lower ss/so than PZT-5H the composite had a low es /so of 200. A low £s/so is advantageous for single element transducers with large apertures but not for small aperture annular array transducers. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 104 Since the commercial availability of these 1 -3 fiber composites (Meyer, 1998) is limited, another method for fabricating high frequency 1-3 composite can be used. Interdigital pair bonding is a current method available to fabricate fine scale 1- 3 composites for arrays above 20 MHz (Liu et al., 2001). In this process two ceramic pieces are diced with a coarse pitch and then are interdigitated to form a composite with a small pitch necessary for high frequencies above 20 MHz. As a result of the small epoxy kerf the volume fraction of the composite is high as compared to traditional dice and fill composites. The kerf width should be less than: Where Vs is the shear velocity of the epoxy and Fc is the center frequency of the transducer (Ritter, 2000). In order to avoid lateral mode resonances the ceramic width should be (Ritter, 2000): kerf width < 4xV 2 x F c (5.1) Ceramic Width < Ceramic Height 2 (5.2) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 105 5.2.1 PZT-5H/Epoxy 1-3 Composite Acoustic Modeling The following parameters were calculated using the constituent equations from Smith and Auld, 1991 for a PZT5H/Epotek 301 1-3 composite: density (p) (Fig. 5.1), clamped dielectric constant (8 S 33/£o) (Fig. 5.2), coupling coefficient (kt) Fig. (5.3), longitudinal velocity (Vl) (Fig. 5.4) and acoustic impedance (ZA ) (Fig. 5.5). The properties from TRS600FGHD (TRS Ceramics, State College, PA) and Epotek 301 (Table 5.1) were used in the calculation. Table 5.2 shows the parameters of the 1-3 composite for the different ceramic volume fractions up to 55%. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 106 Piezoceramic TRS600FGHD1 Cbn (N/m2 ) 14 x 101 U CE 12 (N/m2 ) 7.5 x 101 0 Ce1 3 (N/m2 ) 9.0 x 101 0 Ce3 3 (N/m2 ) 12.1 x 101 0 e3i (C/m2) -5.7 e3 3 (C/m2) 25.8 es3 3 /e0 1350 p (Kg/m3) 7500 k3 3 0.79 kt 0.54 Polymer Epotek 3012 ci i (N/m2) 0.8075875 x 101 U C 12 (N/m2) 0.4596205 x 101 0 En/So 4 Vs (m/s) 1230 Vl (m/s) 2650 P (Kg/m3 ) 1150 1. www.trsceramics.com 2. Calculated after Ritter, 2000. Table 5.1. Properties of fine grain PZT-5H (TRS600FGHD) and epoxy (Epotek 301). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 107 8000 7000 6000 C O 3000 2000 1000 20 80 100 v, Volume Fraction Ceramic (%) Fig. 5.1. Variation with volume fraction of piezoelectric ceramic, v, of a composite’s density, p. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 108 1400 1000 800 600 400 -a a. 200 100 v, Volume Fraction Ceramic (%) Fig. 5.2. Variation with volume fraction of piezoelectric ceramic, v, of a composite’s relative clamped dielectric constant, sS 33/eo- R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 109 0.8 0.7 0.6 0.5 0.3 0.2 0.1 40 60 v, Volume Fraction Ceramic (%) 80 100 Fig. 5.3. Variation with volume fraction of piezoelectric ceramic, v, of a composite’s coupling coefficient, Kt. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 5000 4500 2500 80 100 40 60 v, Volume Fraction Ceramic (% Fig. 5.4. Variation with volume fraction of piezoelectric ceramic, v, of a composite’s longitudinal velocity, Vl. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Impedance Z (MRayls) 111 40 35 30 25 20 15 10 5 0. 0 20 40 60 80 100 v, Volume Fraction Ceramic (%) Fig. 5.5. Variation with volume fraction of piezoelectric ceramic, v, of a composite’s acoustic impedance, ZA . R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 112 Vol. % ceramic P (Kg /mj) es3 3 /s0 VL (m/s) ZA (MRayls) Kt 0 1150 4 2650 3.047 0 5 1467 73 3077 4.510 0.505 10 1785 142 3341 5.964 0.605 15 2102 210 3516 7.392 0.652 20 2402 280 3640 8.810 0.679 25 2737 349 3734 10.220 0.696 30 3055 417 3809 11.63 0.708 35 3372 486 3869 13.05 0.716 40 3690 555 3920 14.46 0.722 45 4007 624 3965 15.89 0.726 50 4325 693 4004 17.32 0.729 55 4642 761 4041 18.76 0.730 Table 5.2. Calculated properties of PZT-5H/Epoxy 1-3 composite. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 113 5.2.2 Fabrication of 1-3 Composite using IPB The 1 -3 composite was prepared using the interdigital pair bonding (IPB) method according to the procedure in Liu et al. (2001). 2” x 2” glass plates with uniform thickness within 4 pm are cleaned with acetone, alcohol and deionized (DI) water. Pieces of TRS600FGHD are cleaned using the same solutions as above. After the ceramic and glass pieces are dried. The two pieces of TRS600FGHD were waxed onto a glass plate using a 60° C refined parafin wax on a hot plate and pressed with a rubber jig to remove air bubbles. Using various coarse to fine grit sand paper the pieces were lapped to 500 pm and polished using 12 pm AI2O3 powder. A dicing saw (Tear 864-1, Thermocarbon Inc., Casselberry FL) was used with a saw blade thickness of 35 pm ( Tanaka Systems Inc., Mountian View, CA) to cut kerfs 120 pm deep at a 70 pm pitch resulting in a gap width of 37 pm and a ceramic width of 22 pm. This process was repeated again at a right angle to the original cut. Once the two pieces were fabricated they were impregnated with Epotek 301 and interdigitally inserted together and allowed to cure in a dry box for 24 hours. The resulting kerf width was 7.5 pm between ceramic posts. The IPB composite was lapped 400 pm on one side to remove the excess ceramic material. The ceramic backbone of the composite was then cleaned and waxed onto a glass plate for further processing. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 114 5.2.3 PZT-5H/Epoxy 1-3 Composite Equivalent Circuit Modeling Furthermore, the volume fraction can be calculated as (Liu et al., 2001): W 2 VF = 7 ----------------------------------------------------------- (5.2) (Wt + W j where Wc is the width of the ceramic post and Wk is the width of the kerf. The ceramic volume fraction was therefore 55 %. The acoustic parameters were calculated for the 1-3 composite with a 55 % volume fraction and used in the equivalent circuit model. Quarter wavelength thickness matching layers were calculated based on the findings of Desilets et al., 1978 (Table 5.3) and the 1-3 composite with a volume fraction of 55 %. Zc represents the acoustic impedance of the composite and Zl represents the acoustic impedance of the load (1.5 MRayls). The load impedance is typically water or human tissue. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 115 Impedance First Layer ( Z i ) MRayls Second Layer ( Z 2) MRayls One Matching Layer Two Matching Layers r j 1/ 3*7 2 /3 A : A 3.48 Zc4/7Zl 3/7 6.35 Zc 1/7Zl 6/7 2.15 Table 5.3. Impedance calculation of matching layers from Desilets et al., 1978. A single annulus was modeled using PiezoCAD (Sonic Concepts Inc., Woodinville, WA), based on Krimholtz, Leedom and Matthaei (KLM) model (Krimholtz et al, 1970). The model included a 40 pm 1-3 composite piezoelectric layer, a one millimeter diameter aperture (1 pm copper, 0.05 pm chrome, 0.45 pm gold) electrode on a 50.8 pm polyimide film with an unloaded epoxy (Epotek 301, Epoxy Technology Inc., Billerica, MA) backing. The 1st matching layer used was an epoxy loaded with silver particles with an acoustic impedance of 7.33 MRayls with a thickness of 13 pm. The 2n d matching layer used was Parylene with an acoustic impedance of 2.6 MRayls with a thickness of 15 pm. Inductive tuning (0.1 pH) and one meter length coaxial cable, 50 Q, was used for impedance matching. A pulse echo response from the model shown in Fig. 5.6, displays the 84 nano-second, -20 dB pulse length. The equivalent - 20 dB axial resolution corresponding to this pulse length in water is 63 pm (Turnbull et al., 1995). The Fourier transform of the pulse echo response shows a center frequency of 30 MHz R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 116 with 75 % bandwidth, Fig. 5.7. At 30 MHz, the modeled impedance magnitude and phase were 53 Q . and 35° respectively (Fig. 5.8). 0.5 "O T3 a. 100 0 200 300 400 500 Time (ns) Fig. 5.6. Simulated pulse echo response from 1-3 composite. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 117 -20 -40 T3 3 -60 -80 -100 -120 20 40 60 80 100 Frequency (MHz) Fig. 5.7. Simulated frequency response from 1-3 composite. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 10 20 30 40 50 60 70 80 90 100 Frequency (MHz) Fig. 5.8. Simulated pulse electrical impedance from 1-3 composite, magnitude (solid line), phase (dashed line). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 119 5.2.4 1-3 Composite Annular Array Fabrication Procedure The composite was cleaned with trichlorethylene, acetone, alcohol, DI water and dried. A 500 A chrome/ 1 0 0 0 A gold electrode was deposited onto the top of the wafer by sputtering. A ceramic ring was glued around the composite onto the glass plate using 5-minute epoxy. A 1% Alconox solution was used to clean the substrate followed by a DI water rinse and forced air dry. An adhesion promoter, Chemlok AP-131 (Lord Corp., Cary, NC), was wiped across the surface of the electrode and allowed to dry for 30 minutes. The first matching layer consisted of an epoxy including a mixture of 5 g Insulcast 501 and 0.65 g Insulcure 9 (American Safety Technologies, Roseland, NJ). Six grams of silver particles (Sigma-Aldrich Inc., St. Louis, MO) was mixed with 2.5 g of this epoxy and degassed. The mixture was degassed to remove all air bubbles and then poured into the ceramic ring. In order to compact the silver particles to the surface of the electrode the piece was centrifuged at 3000 rpm for 10 minutes. The epoxy was then allowed to cure in the dry box for 24 hours followed by a post cure at 40° C for 4 hours. New spacers were added to the glass plate to be same height of the epoxy ring. Using various coarse to fine grit sand papers the matching layer was is lapped to within 30 pm of the final thickness. Then a 12 pm AI2O3 powder slurry was used to polish the matching layer to a final thickness of 19 pm. The composite was then removed from the glass plate and cleaned with acetone, alcohol, and DI water. The composite matching layer down R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 120 was waxed onto a glass plate using 60° C refined paraffin was on a hot plate and pressed with a rubber jig to remove air bubbles. Using various coarse to fine grit sand papers the ceramic layer was lapped to within 50 pm of the final thickness. Then the composite layer was polished with a 12 pm AI2O3 powder slurry to within 20 pm of the final thickness. Nine micron diamond powder slurry was used to polish the composite to a final thickness of 50 pm. The composite was cleaned with acetone, alcohol, and DI water. Then using the same method described earlier in this work the unelectroded side of the composite was bonded to the flex circuit using Epotek 301 and allowed to cure in the dry box for 24 hours. The ground electrode (500 A chrome/ 1 0 0 0 A gold) was then sputtered on the surface of the transducer. All the elements of the composite were then poled at a high voltage of 2 0 0 volts corresponding to 4 V/pm for the 50 pm composite in air at room temperture. A parylene matching layer was then evaporated onto the surface of the transducer to a thickness of 24 pm. Inductive tuning (1.2 pH) and 1.66 meter length coaxial cable, 50 f2, was used for impedance matching. The 38 AWG coaxial cable (New England Electric Wire Company, Lisbon, NH) Part No. N12-38T-125-1 was characterized at 30 MHz (Table 5.4). R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 121 Property N12-38T-125-1 Characteristic 53.55 -5 .5 1 iQ Impedance (Zo) Propagation Constant (y) 0.066 + 0.977i Propagation Velocity 1.929 x 108 m/s (VP ) Resistance/Unit Length 8.934 fl/m ( 0 Capacitance/Unit Length 96.47 pF/m (c) Inductance/Unit Length 0.275 pH (1) Conductance/Unit -635.5 pS Length (g) Attenuation/Unit Length 0.57 dB/m Table 5.4. Properties of the 50 Q coaxial cable (N12-38T-125-1, New England Electric Wire Company, Lisbon, NH) characterized at 30 MHz, used to impedance match annular array elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 1 2 2 5.2.5 1-3 Composite Annular Array Fabrication Issues and Characterization The most critical fabrication issue when using a flex circuit directly bonded to the unelectroded composite is the thickness of the epoxy bond line. The epoxy bond line acts as an extra electrical capacitance in series with the transducer thereby creating a voltage divider drawing the majority of current instead o f the composite layer (Fig. 5.9). A thinner bond line (< 1 pm) is more advantageous because it makes the electrical capacitance higher thereby deceasing its electrical impedance. p '-'com posite c , epoxy Fig. 5.9. Voltage divider between composite and epoxy bond line. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 123 The reason that this happens has to do with the dielectric constants of the epoxy bond line and the composite. The epoxy has a higher electrical impedance due to its low dielectric constant as compared to that of the composite (Table 5.5). Material Thickness (pm) ss C(pf) |Z| (O) Epoxy 1 4.1 27.8 191 P(VDF-TrFe) 9 7 5.4 982 Composite 50 761 33.67 158 Table 5.5. Transducer material comparison of calculated impedance values based on thickness. This additional capacitance is detrimental to the composite transducer’s performance by raising the overall electrical impedance magnitude and making the phase more negative. As a result the sensitivity of the device decreases as well as the bandwidth due to the impedance mismatch. The copolymer does not suffer as much from the electrical capacitance of the epoxy bond line since the capacitance of the copolymer is much lower. Therefore, the copolymer film has the majority of the voltage across it as compared to the epoxy bond line. The epoxy bond line is acoustically transparent to the copolymer film and composite because of the thickness. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 124 The pulse echo response, impedance, insertion loss and crosstalk were measured for the fabricated composite array using the same methods described earlier in this work. The electrical impedance magnitude and phase at 30 MHz before and after matching is shown in Table 5.6. The water loaded impedance measurement for each annulus with impedance matching is shown in Fig. 5.10. The center frequency, - 6 dB bandwidth and -20 dB pulse lengths were measured for each annulus and listed in Table 5.7. The longer ring down effect from the decreased width to height ratio o f the outer annuli was evident in the -20 dB pulse length measurements. Element # |Z| Q . without matching 0 ° without matching |Z| Q with impedance matching O ° with impedance matching 1 262 -82 46 16 2 265 -83 46 2 2 3 257 -83 49 15 4 262 -80 46 14 5 227 -75 48 14 6 276 -84 47 25 7 307 -83 41 41 Table 5.6. Impedance measurement of annular array elements before and after matching at 30 MHz. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 4 > degrees |Z| Q 125 400 200 100 Frequency (MHz) 100 -100 50 60 Frequency (MHz) 100 Fig. 5.10. Measured electrical impedance in water for all annuli: magnitude (top) and phase (bottom). Each annulus was impedance matched with a 1.2 pH and a UA 50 coaxial cable. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 126 Element # Center Frequency (MHz) - 6 dB Bandwidth (%) -20 dB pulse lengths (ns) W/H ratios 1 27 46 1 2 0 1 0 . 0 0 2 28 44 168 3.98 3 29 35 227 3.00 4 27 38 301 2.50 5 28 32 295 2.16 6 31 27 330 1.96 7 27 18 384 1.78 Table 5.7. Pulse echo response of annular array elements with impedance matching. The pulse echo response and corresponding frequency response from the first element are shown in Fig. 5.11 and Fig. 5.12 for the annular array with impedance matching. The diffraction compensated and diffraction uncompensated insertion loss values for each annulus at 30 MHz are recorded in Table 5.8. The signal loss from the attenuation in water at 2.2 x 10-4 dB/mm-MHz2 (Lockwood et al., 1994) and transmission into the quartz target at 1.6 dB (Selfridge, 1983) was diffraction compensated in the diffraction uncompensated insertion loss calculation. These results are comparable to the diffraction compensated values of the P(VDF-TrFe) annular array transducer. Due to the electrical capacitance issues of the bonding R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 127 technique used in this work the composite’s full kt potential (0.73) was not completely fulfilled. A new fabrication technique must be implemented that will increase the connectivity of the flex circuit electrodes to the composite thereby increasing the capacitance significantly to lower the insertion loss. A possible solution could include electroding the bottom side of the composite and using photolithography to etch an annuli pattern on the composite corresponding to the electroded flex circuit. In essence having electrodes on both the composite and the flex circuit could create a short circuit through the epoxy bond line thereby eliminating the capacitance effect. This method would complicate the fabrication procedure by having to align the flex circuit electrodes to the annuli pattern on the composite. Even though this process would complicate the simple array fabrication it may be necessary in order to make the device actually perform as predicted in the equivalent circuit models. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Amplitude (normalized) 128 0.6 0.4 0.2 - 0.2 -0.4 - 0.6 0 100 200 300 400 500 600 700 800 900 Time (ns) Fig. 5.11. Measured time domain pulse echo response for the center array element. The -20 dB pulse length was 120 ns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 129 -10 -20 -40 -50 30 Frequency (MHz) Fig. 5.12. Measured normalized frequency spectrum for the center array element. The measured center frequency was 27 MHz with a - 6 dB bandwidth of 46 %. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 130 Element # Diffraction Uncompensated IL (dB) Diffraction Compensated IL (dB) 1 -35.23 -33.65 2 -39.85 -31.82 3 -42.80 -30.10 4 -46.64 -31.63 5 -47.91 -32.44 6 -49.09 -32.02 7 -54.09 -36.32 Table 5.8. Annular array insertion loss measurement compensated for attenuation in water and diffraction effects. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 131 The combined electrical and acoustical crosstalk was measured between adjacent elements. The crosstalk measurement between adjacent elements in the pass-band is shown in Fig. 5.13. The crosstalk remained below -14 dB at 30 MHz for all adjacent elements. The acoustic crosstalk can be attributed mainly to greater lateral mode acoustic coupling in the composite. Mechanical dicing may be beneficial to further decrease acoustic crosstalk between elements. There is also an increased electrical crosstalk as compared to the copolymer film. The capacitance of the composite over the separation between electrodes is greater due to the high dielectric constant o f the composite thereby causing reduced crosstalk impedance between elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 132 -12 -14 -16 # / Q) 1 Lyj ■ a fj % -2 o h 03 J, C O 7/ -22'Lj Annuli 1:2 Annuli 2:3 Annuli 3:4 Annuli 4:5 Annuli 5:6 Annuli 6:7 -24 -26 25 35 20 Frequency (MHz) Fig. 5.13. Composite annular array crosstalk measurement between adjacent elements. 5.2.6 PMN-33%PT 1-3 Composite Acoustic Modeling Single crystal PMN/PT and PZN/PT composites offer higher coupling coefficients than conventional ceramic composites. Transducers fabricated with single crystal composites have shown broader bandwidth and higher sensitivity than their ceramic composite counterparts (Ritter et al., 2000b). However, most single crystal composites have a lower clamped dielectric constant than ceramic composites. As a result the array elements are usually larger or a stacked transducer R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 133 approach needs to be taken to match for impedances. Annular array transducers would benefit from a 1 -3 single crystal composite because of the higher sensitivity and broader bandwidth. The following parameters were calculated using the constituent equations from Smith and Auld, 1991 for a PMN-33%PT/Epotek 301 1-3 composite: density (p) (Fig. 5.14), clamped dielectric constant (sS 33/8 o) (Fig. 5.15), coupling coefficient (kt) Fig. (5.16), longitudinal velocity ( V l) (Fig. 5.17) and acoustic impedance (Z a ) (Fig. 5.18). The properties from PMN-33%PT from Zipparo and Oakley, 2001(Table 5.9) and Epotek 301 were used in the calculation. Table 5.10 shows the parameters of the 1-3 composite for the different ceramic volume fractions up to 55%. Since the longitudinal velocity was lower than that of the PZT-5H/Epoxy composite at 55 %, the resonance frequency of the annular array would be lower for a composite of the same thickness. However, the PMN-PT array would have better sensitivity and bandwidth. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Single Crystal PMN-33%PT Can (N/m2) 11.1 x 101 U CEi2 (N/m2 ) 10.2 x 101 0 Ce1 3 (N/m2 ) 10.1 x 101 0 Ce3 3 (N/m2 ) 10.5 x 101 0 e3. (C/m2 ) -3.7 e3 3 (C/m2 ) 15 £S 3 3 /£0 680 p (Kg/m3 ) 8060 k3 3 0.93 kt 0.59 Table 5.9. Properties of PMN-33%PT from Zipparo et al., 2001. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 135 9000 8000 7000 C O 6000 O ) m 4000 3000 2000 1000 20 100 v, Volume Fraction Ceramic (%) Fig. 5.14. Variation with volume fraction of piezoelectric ceramic, v, of a PMN-PT composite’s density, p. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 136 700 < 600 C O C O C O 500 c C C • + — * C f l c o O .o o 0 0 b 400 300 200 T3 Q. E 100 0 5 o 100 40 60 v, Volume Fraction Ceramic (%) 80 Fig. 5.15. Variation with volume fraction of piezoelectric ceramic, v, of a PMN-PT composite’s clamped dielectric constant, s^ / sq. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 137 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 40 v, Volume Fraction Ceramic (%) 60 Ceramic 80 100 Fig. 5.16. Variation with volume fraction of piezoelectric ceramic, v, of a PMN-PT composite’s coupling coefficient, Kt. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 138 4400 4200 4000 3800 o 3600 ^ 3400 ,-i 3200 3 3000 2800 2600, 100 v, Volume Fraction Ceramic (%) Fig. 5.17. Variation with volume fraction of piezoelectric ceramic, v, of a PMN-PT composite’s longitudinal velocity, V l. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 139 30 « 25 S' a: & 20 a > o c a s T J 0 a . E 40 60 v, Volume Fraction Ceramic (%) 100 80 Fig. 5.18. Variation with volume fraction of piezoelectric ceramic, v, of a PMN-PT composite’s acoustic impedance, ZA . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 Vol. % PMN-PT P (Kg /mJ) SS 33/So VL (m/s) ZA (MRayls) Kt 0 1150 4 2650 3.047 0 5 1495 38 2716 4.063 0.473 1 0 1841 73 2786 5.130 0.603 15 2186 107 2836 6 . 2 0 0 0.671 2 0 2532 142 2874 7.270 0.714 25 2877 176 2905 8.350 0.742 30 3223 2 1 1 2931 9.440 0.762 35 3568 245 2955 10.540 0.776 40 3914 279 2978 11.650 0.786 45 4259 314 2999 12.770 0.793 50 4605 348 3022 13.910 0.797 55 4950 383 3046 15.070 0.799 Table 5.10. Calculated properties of PMN-PT/Epoxy 1-3 composite. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 141 5.3.0 Tunable Copolymer Annular Array Transducer A tunable copolymer transducer was modeled, fabricated and characterized using all the methods previously described in this work. Inductive tuning and impedance methodology made it possible to select a desired center frequency of 30 MHz while using the same 9 pm P(VDF-TrFE) film. Thereby it is possible to choose specific inductor values to tune to certain resonant frequencies. In order to further reduce the high electrical impedance at lower frequencies a step down transformer was used. The transformer used in conjunction with a coaxial cable further improved the bandwidth of the device throughout the pass band. 5.3.1 Equivalent Circuit Modeling of a Tunable Copolymer Annular Array Transducer A single annulus was modeled using PiezoCAD (Sonic Concepts Inc., Woodinville, WA), based on Krimholtz, Leedom and Matthaei (KLM) model (Krimholtz et al, 1970). The model included a 9 pm P(VDF-TrFE) copolymer piezoelectric layer, a one millimeter diameter aperture (1 pm copper, 0.5 pm chrome, 0.45 pm gold) electrode on a 50.8 pm polyimide film with an unloaded epoxy (Epotek 301, Epoxy Technology Inc., Billerica, MA) backing. A 3.9 pH inductor was used to cancel the phase at 30 MHz. Then a step down transformer with a turn R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 142 ■ y ratio of N = 1.5 was used to reduce the transducer impedance by a factor of N . With a \J4 72 Cl coaxial cable and transformer the bandwidth was improved across the pass band of the device. A pulse echo response from the model shown in Fig. 5.19, displays the 125 nano-second, -20dB pulse length. The equivalent axial resolution corresponding to this pulse length in water is 93 pm (Turnbull et al., 1995). The Fourier transform o f the pulse echo response simulation shows a center frequency of 36 MHz with 36 % bandwidth, Fig. 5.20. At 30 MHz, the modeled impedance magnitude and phase were 52 Cl and 6 6 ° respectively (Fig. 5.21). 0.6 0.4 T J ^ 0.2 = - 0.2 -0.4 - 0.6 500 300 400 200 100 0 Time (ns) Fig. 5.19. Modeled time domain pulse echo response for the center array element. The -20 dB pulse length was 125 ns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Magnitude (dB) 143 -10 -20 -40 -70 -80 -90 80 100 20 30 Frequency (MHz) Fig. 5.20. Modeled normalized frequency spectrum for the center array element. The measured center frequency was 36 MHz with a - 6 dB bandwidth of 36%. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. o Iz l 144 300 100 250 200 100 -100 40 50 60 Frequency (MHz) 80 90 100 Fig. 5.21. Modeled electrical impedance magnitude (solid line) and phase (dashed line) for center annulus in water. The element was impedance matched with a 3.9 pH inductor, step down transformer N = 1.5 and a A/4 72 Q coaxial cable. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 145 5.3.2 Fabrication and Characterization of a Tunable Copolymer Annular Array Transducer The same fabrication procedure was used as previously described in this work. Inductors (3.9 pH) were soldered to the ends of short wires connected to the flex circuit solder pads. The transformers (N = 1.5) were then soldered to short wires and then to the inductors. Then A/4 length coaxial cables, 72 f2, 40 AWG, coax # 171-1019-XX (Precision Interconnect, Portland OR), were soldered directly to the transformers. The coaxial cables were characterized; see Table 5.11, at 30 MHz using the same method as described by Cannata et al., 2003. The pulse echo response, impedance, insertion loss and crosstalk were measured using the same methods described earlier in this work. The water loaded impedance measurement for each annulus with impedance matching is shown in Fig. 5.22. The electrical impedance magnitude and phase at resonance before and after matching is shown in Table 5.12. The center frequency and - 6 dB bandwidth were measured for each annulus and listed in Table 5.13. The pulse echo response and frequency response from the first element are shown in Fig. 5.23 and Fig. 5.24 for the annular array with impedance matching. The inductor was used to cancel the phase out resonance thereby improving the sensitivity of the device at that frequency. An adverse effect of using this tuning is that it is specific to one frequency therefore a low bandwidth band pass filter is created thereby the transducer will have a lower R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 146 - 6 dB bandwidth. The addition of the step down transformer serves to decrease the transducer impedance further to match closer to 50 Q over a broad band of frequencies. Ideally it would be better to use a larger aperture to decrease the impedance thereby creating a better impedance match at lower frequencies. Property PI #171-1019-XX Characteristic 70.58 - 6.55i Q Impedance (Zo) Propagation Constant (y) 0.059+ 0.762i Propagation Velocity 2.47 x 108 m/s (VP ) Resistance/Unit Length 9.14 O/m (r) Capacitance/Unit Length 57.17 pF/m (c) Inductance/Unit Length 0.283 pH (1) Conductance/U nit -166 pS Length (g) Attenuation/Unit Length 0.51 dB/m Table 5.11. Properties of the 72 O coaxial cable (PI #171-1019-XX, Precision Interconnect, Portland OR) characterized at 30 MHz, used to impedance match annular array elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 147 300 200 100 40 50 60 7 Frequency (MHz) 100 O ) 20 30 40 50 60 70 80 90 100 Frequency (MHz) Fig. 5.22. Measured electrical impedance in water for all annuli: magnitude (top) and phase (bottom). Each annulus was impedance matched with a 3.9 pH inductor, transformer (N = 1.5) and a A 74 72 Q coaxial cable. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 148 Element # |Z| Q without matching ® ° without matching |Z| Q with impedance matching ® ° with impedance matching 1 936 -84 61 8 2 932 -85 59 8 3 933 -85 61 7 4 906 -85 61 4 5 882 -85 58 17 6 883 -85 62 1 2 7 883 -85 65 1 2 8 826 -85 6 6 - 2 Table 5.12. Impedance measurement of annular array elements before and after matching at 30 MHz. R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission. 149 Element # Center Frequency (MHz) - 6 dB Bandwidth (%) -20 dB pulse lengths (ns) 1 32 29 175 2 31 29 225 3 30 2 2 281 4 32 28 164 5 30 34 147 6 30 34 182 7 32 30 155 8 30 27 229 Table 5.13. Pulse echo response of annular array elements with impedance matching. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 150 0.6 0.4 0.2 T J 1 - 0.2 Q . -0.4 - 0.6 0 100 200 300 400 500 Time (ns) Fig. 5.23. Measured time domain pulse echo response for the center array element. The -20 dB pulse length was 175 ns. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 151 -10 = -30 -50 -60 30 40 Frequency 50 Frequency (MHz) 60 Fig. 5.24. Measured normalized frequency spectrum for the center array element. The measured center frequency was 32 MHz with a - 6 dB bandwidth of 29 %. The signal loss from the attenuation in water at 2.2 x 10' 4 dB/mm-MHz2 (Lockwood et al., 1994) and transmission into the quartz target at 1. 6 dB (Selfridge, 1983) were compensated in all insertion loss calculations. The diffraction uncompensated and diffraction compensated insertion loss values for each annulus at 30 MHz were recorded in Table 5.14. This array suffered from larger insertion loss as compared to higher frequency copolymer array because of impedance matching. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 152 Element # Diffraction Uncompensated IL (dB) Diffraction Compensated IL (dB) 1 -40.57 -38.99 2 -48.48 -40.45 3 -51.30 -38.60 4 -52.29 -37.27 5 -53.94 -38.46 6 -55.63 -38.56 7 -54.60 -36.83 8 -55.85 -37.99 Table 5.14. Annular array insertion loss measurement compensated for attenuation in water and diffraction effects. The combined electrical and acoustical crosstalk was measured between adjacent elements (Fig. 5.25). The crosstalk remained below -23 dB at 30 MHz for all adjacent elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 153 -35' -40 1 N — - Annuli 1:2 — « . . . Annuli 2:3 Annuli 3:4 ---------- Annuli 4:5 Annuli 5:6 — ► - Annuli 6:7 ----- Annuli 7:8 25 30 Frequency (MHz) Fig. 5.25. Tuned copolymer annular array crosstalk measurement between adjacent elements. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 154 CHAPTER 6 SUMMARY AND FUTURE WORK 6.0 Summary of Work This study shows that it is possible to fabricate a high frequency annular array using a single sheet of copolymer material and a double sided polyimide flex circuit. This simple yet effective approach for annular array fabrication proved to be reliable and reproducible, and is well suited for mass production of future devices at high frequency. It was initially thought that the primary performance tradeoff for this simple design would be unsatisfactory levels of inter-element crosstalk. However, the measured crosstalk of less than - 29 dB at the center frequency proved to be satisfactory for annular array beamforming. The low crosstalk levels observed can be mainly attributed to the low planar coupling coefficient of the copolymer film. The copolymer annular array transducer built for this study performed well under standard array characterization tests. The device sensitivity was improved by electrically matching each element to an impedance magnitude of 50 Q and 0° phase at resonance. Since copolymer transducers typically have high insertion loss it was important to impedance match the transducer in order to maximize signal to noise ratio for imaging. The average round trip insertion loss measured for the array and R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 155 compensated for diffraction effects was - 33.5 dB. The measured average center frequency and bandwidth of an element was 55 MHz and 47 % respectively. An adverse effect of using inductive tuning was the narrow band filtering effect on the device’s bandwidth. The high frequency and short pulse length of this array makes it well suited for high frequency medical imaging applications. A vertical wire phantom was imaged using a single focus transmit beamformer and dynamic focusing receive beamformer. The wire phantom imaged showed improved lateral resolution over a range of 9 mm after a dynamic focusing algorithm was applied on receive. This fabrication technique can be applied to annular array designs higher than the 55 MHz device reported. However, care should be taken when designing the component layer thicknesses of the flex circuit. The polyimide, chrome/gold, epoxy and copper layers can all contribute to spurious resonances that will deteriously affect the resonance of the transducer. The aperture size of the array and the number of elements must be carefully chosen. Reducing the aperture results in a smaller phase difference between elements thus reducing the side lobe levels (Brown et al., 2004). Increasing the number of elements while keeping the path difference the same has shown an increase of 12 dB in relative pressure amplitude and a decrease of side lobe levels (Brown et al., 2004). Furthermore, the feasibility of fabricating composite and tunable copolymer annular transducers was accessed. The material properties of a 1-3 PZT 5H/Epoxy R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 156 composite was modeled based on ceramic volume fraction. These results were used in circuit equivalent 1-D models to predict the performance of the annular array transducer. A high density 1-3 PZT-5H/epoxy composite with a ceramic width of 22 pm and kerf width of 7 pm was fabricated using interdigital pair bonding method. The composite was then lapped to the appropriate thickness and matching layers were added to enhance acoustic performance. The composite was then bonded to the flex circuit. The measured average center frequency and bandwidth of an element was 28 MHz and 34 % respectively. It was found that the electrical capacitance of the epoxy bond line reduced the sensitivity and bandwidth of the device significantly than what was expected. Even with electrical impedance matching using an inductor and coaxial cable the average insertion loss for the array was - 32 dB. Therefore the sensitivity and bandwidth of the device was not restored to its full potential. The reduced - 20 dB pulse length for each element can be attributed to the acoustic impedance of the composite thus decreasing the reverberations at the composite and bond line interface as compared to a ceramic. A key assumption of the kerfless array design is that the lateral resonance would be below the frequency band of the device the since there are no cut kerfs between elements thus eliminating the width to height aspect ratio (Morton and Lockwood, 2001). However, the - 20 dB ring increased as the electrode element width to thickness ratio decreased but this may have been caused by the acoustic crosstalk (> -14 dB throughout - 6 dB bandwidth) in the material itself due to the ceramic post width to height ratio of 1 :2 . R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 157 A tunable copolymer transducer was modeled based on the fact that the phase could be canceled at a particular frequency. Thus using the same 9 pm copolymer film a 30 MHz transducer was fabricated. In order to improve the device bandwidth a step down transformer and coaxial cable were used. The measured average center frequency and bandwidth of an element was 31 MHz and 29 % respectively. Insertion loss (-38 dB) and cross coupling (< -23 dB across - 6 dB bandwidth) values were greater when compared to the 50 MHz copolymer annular array. The larger crosstalk may be attributed to more lateral acoustic coupling at 30 MHz shown by the low level ring down as well as the worse electrical impedance mismatch. For a 30 MHz annular array transducer the performance would better if the aperture were redesigned to be a larger aperture incorporating a thicker copolymer film. Nonetheless, due the broad bandwidth and capacitive nature of P(VDF-TrFE) the frequency can be tuned to any desired frequency within the -20 dB frequency pass band. The tradeoffs for tuning transducers include increased insertion loss, cross coupling and lower bandwidth. The center frequency and bandwidth of the tuned copolymer array and composite array were similar. Some differences were in the measured insertion loss, crosstalk levels, and ring down. The composite array showed a 6 dB improvement in sensitivity as compared to the tuned copolymer array. Since the copolymer material has a low lateral coupling coefficient compared to composites the tuned copolymer array crosstalk level was 9 dB lower than that of the composite array. The ring down R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 158 effect of the tuned copolymer array was also less than the composite array because of larger effective electrode width to height (polymer thickness) ratios and reduced lateral crosstalk. The performance of the 50 MHz copolymer array was superior to both the composite and the tuned copolymer annular array. The reason for this is simply because the array aperture and copolymer film were designed specifically for the 50 MHz center frequency and electrical impedance. The flex circuit polyimide substrate is also acoustically well matched to the copolymer film. The other annular designs used the same flex circuit as the 50 MHz copolymer annular array therefore this flex circuit was not ideal for them. However, this flex circuit was used to see if any initial performance problems would result from using flex circuit interconnect in general. In this work it was discovered that the composite epoxy bond line caused a major deficit in device performance. 6.1 Future Work Many improvements can be made to the existing design and fabrication of the annular array transducers described in this work. The flex circuit can be further improved by implementing ground shields between elements this will electrically isolate elements from fringing effects there by reducing the overall electric crosstalk at high frequencies. Ground shields can also be implemented between traces on the bottom side of the flex circuit. Electrode thicknesses should be further reduced to R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 159 5,000 A to avoid spurious resonances the copolymer film. It would also be advantageous to reduce the polyimide thickness from 50.8 pm to several microns. The aperture of the device could be made smaller this would further reduce side lobe levels. If a smaller aperture array than presented in this work would be designed then using materials with a higher dielectric constant would be necessary for electrical impedance matching purposes. Since composites have low acoustic impedance compared to ceramics and higher electromechanical coupling coefficient then if they are incorporated into the array design then there should be an improvement in sensitivity and bandwidth. It will be necessary to develop a better array interconnect strategy since it has been shown in this work that using unelectroded composite with high dielectric constant as compared to copolymer with low dielectric constant that the epoxy bond line has an adverse effect on array performance. If an interconnect solution can be implemented for composites then arrays with high bandwidth and low insertion can be fabricated. It may be possible to overcome the electrical capacitance of the non conductive epoxy bond line by using low temperature anisotropic conductive materials to bond the composite to the flex circuit or to electrode the backside of the composite with the same electrode pattern as on the flex circuit. If a photolithography technique is implemented then it would be crucial to align the electrode pattern on the composite to that on the flex circuit. Another solution may be to electrode the entire composite bottom then bond it to the flex followed by laser R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 160 dicing the annuli pattern. However, these possible solutions would further complicate the simple fabrication process developed in this work. Some initial investigations have been performed by Demore and Lockwood (2005) on the effective width of the annulus as compared to that of the electrode width. Further study on finite element modeling could determine whether dicing the composite and filling the kerfs with epoxy would reduce cross coupling or increase it by changing the width to height ratios of the annuli. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 161 BIBLIOGRAPHY AIUM, “Acoustic Output Measurement and Labeling Standard for Diagnostic Ultrasound Equipment,” Amer. Inst. Ultra. Med., Washington, DC, (1992). AIUM/NEMA, “Standard for real-time display of thermal and mechanical acoustic output indices on diagnostic ultrasound equipment,” Nat. Elec. Manuf. Assoc., Washington, DC, (1992). Anatomy o f the Eye. Eyesight Insight, [web page] Feb 2000; http://members.aol.eom/insighteye/anatl.htm#diagram. [Accessed 16 April 2001]. Anatomy o f the Skin. University of Maryland Medicine, [web page] January 2001; http://www.umm.edu/dermatology-info/anatomy.htm. [Accessed 29 June 2001]. Arditi M., Taylor W.B., Foster F.S., Hunt, J.W., “An annular array system for high resolution breast echography,” Ultrasonic Imaging, vol. 4, pp. 1-31, 1982. Brown L.F., “Design considerations for piezoelectric polymer ultrasound transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 47, no. 6 , pp. 1377-1396, 2000. Brown J.A., Demore-Morton C.E., Lockwood G.R., “Design and Fabrication of Annular Arrays for High-Frequency Ultrasound,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 51, no. 8 , pp. 1010-1017, 2004. Cannata J.M., Design andfabrication o f high frequency single element ultrasonic transducers using lithium niobate, M.S. Thesis in Bioengineering, The Pennsylvania State University, 2000. Cannata J.M., Ritter T., Chen W.H., Silverman R.H., Shung K.K., “Design of efficient, broadband single element (20-80 MHz) ultrasonic transducers for medical imaging applications,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 50, no. 11, pp. 1548-1557, 2003. Cannata J.M., High Frequency (>20 MHz) Ultrasonic Arrays fo r Medical Imaging Applications, Ph.D. Thesis in Bioengineering, The Pennsylvania State University, 2004. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 162 Cao P.J., Shung K.K., “Design of a real time digital beamformer for a 50 MHz annular array ultrasound transducers,” Proc. 2002 IEEE Ultrasonics Symposium, pp. 1619-1622, 2002. Cao P.J., Hu C.H., Shung K.K., “Development of a real time digital high frequency annular array ultrasound imaging system,” Proc. 2003 IEEE Ultrasonics Symposium, pp. 1867-1870, 2003. Chen W.H., Development o f a high frequency ultrasound imaging system utilizing single element and annular array transducers, Ph.D. Thesis in Bioengineering, The Pennsylvania State University, 2002. Demore C.E.M., Lockwood G.R., “An investigation of the effective width of elements in kerfless annular arrays,” Proc. 2005 IEEE Ultrasonics Symposium, in publication. Desilets C.S., Fraser J.D., Kino, G.S., “The design of efficient broadband piezoelectric transducers,” IEEE Transactions on Sonics and Ultrasonics, vol. 25, no. 3, pp. 115-125, 1978. FDA, “510(k) Guide for measuring and reporting acoustic output of diagnostic ultrasound medical devices,” Food and Drug Administration, Washington, D.C., (1985). Foster F.S., Larson J.D., Mason M.K., Shoup T.S., Nelson G., Yoshida H., “Development of a 12 element annular array transducer for realtime ultrasound imaging,” Ultras. Med. Biol., vol. 15, no. 7, pp. 649-659, 1989. Foster F.S., Pavlin C.J., Starkoski B., Harasiewicz K., “Ultrasound backscatter microscopy of the eye in vivo,” Proc. 1990 IEEE Ultrasonics Symposium, pp. 1481- 1484, 1990. Foster F.S., Lockwood G.R., Ryan L.K., Harasiewicz K.A., Berube L., Rauth A.M., “Principles and applications of ultrasound backscatter microscopy,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 40, no. 5, pp. 608-617, 1993. Foster F.S., Pavlin, C.J., Harasiewicz K.A., Christopher D.A., Turnbull D.H., “Advances in ultrasound biomicroscopy,” Ultras. Med. Biol., vol. 26, no. 1, pp. 1-27, 2 0 0 0 a. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 163 Foster F. S., Harasiewicz K.A., Sherar M.D., “A history of medical and biological imaging with polyvinylidene fluoride (PVDF) transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 47, no. 6 , pp. 1363-1371, 2000b. Foster F.S., Zhang M.Y., et al., “A new ultrasound instrument for in vivo microimaging of mice,” Ultras. Med. Biol. vol. 28, no. 9, pp. 1165-1172, 2002. Guess J.F., Oakley C.G., Douglas S.J., Morgan R.D., “Cross-talk paths in array transducer,” Proc. 1995 IEEE Ultrasonics Symposium, pp.1279-1282,1995. IEC, “Requirements for the declaration of the acoustic output of medical diagnostic ultrasonic equipment,” Inter. Electrotech. Comm., (1992). Jensen J.A. and Svendsen N. B. “Calculation of pressure fields from arbitrarily shaped, apodized, and excited ultrasound transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 39, no. 2, pp. 262-267, 1992. Kari N., Design and fabrication o f high frequency single element ultrasonic transducers using potassium niobate single crystals, M.S. thesis in Bioengineering, The Pennsylvania State University, 2001. Ketterling J.A., Aristizabal O., Turnbull D.H., Lizzi F.L., “Design and fabrication of a 40 MHz annular array transducer,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 52, no. 4, pp. 672-681, 2005. Kino G.S. Acoustic Waves: Devices, Imaging, and Signal Processing. Englewood Cliffs, New Jersey: Prentice Hall, 1987. Kossoff, G., “Analysis of focusing action of spherically curved transducers,” Ultras. Med. Biol., vol. 5, no. 4, pp. 359-365, 1979. Krimholtz R., Leedom D.A. and Matthaei G.L., “New equivalent circuit for elementary piezoelectric transducers,” Electron. Letters, vol. 6 , no. 13, pp. 398-399, 1970. Ladabaum I., Jin X., Soh H., Atalar A., Khur-Yakub B.T., “Surface Micromachined Capacitive Ultrasonic Transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 45, no. 3, pp. 678-690, 1998. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 164 Liu R., Harasiewicz K.A., Foster F.S., “Interdigital pair bonding for high frequency (20-50 MHz) ultrasonic composite transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 48, no. 1, pp. 299-306, 2001. Lockwood G.R., Ryan L.K., Hunt J.W., Foster F.S., “Measurement of the ultrasonic properties of vascular tissues and blood from 35 to 65MHz,” Ultras. Med. Biol., vol. 17, no. 7, pp. 653-666, 1991. Lockwood G.R., Turnbull D.H., Foster F.S., “Fabrication of high frequency spherically shaped ceramic transducers,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 41, no. 2, pp. 231-235, 1994. Meyer R., High frequency (15-70 MHZ) 1-3 PZT fiber/polymer composites: fabrication and characterization, Ph.D. thesis in the Intercollege Program in Materials, Pennsylvania State University, 1998. Nakamura K., Chiba N., Ito S., “Conversion of 45° rotated x-cut KNBO3 plates to y- cut plates by compression,” Proc. 2004 IEEE Applications o f Ferroelectrics, pp. 98- 101, 2004. Morton, C.E., Lockwood, G.R., “Design of a 40 MHz Annular Array,” Proc. 2001 IEEE Ultrasonics Symposium, pp. 1135-1138, 2001. Pan L., Zan L., Foster F.S., “Ultrasonic and viscoelastic properties of skin under transverse mechanical stress in vitro,” Ultras. Med. Biol., vol. 24, no. 7, pp. 995- 1007, 1998. Passman C., Ermert H., “A 100-MHz ultrasound imaging system for dermatologic and ophthalmologic diagnostic,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 43, no. 4, pp. 545-552, 1996. Ritter T. Design, fabrication, and testing o f high frequency (>20 MHz) composite ultrasound imaging arrays, Ph.D. thesis in Bioengineering, The Pennsylvania State University, 2000a. Ritter T., Geng X., Shung K.K., “Single crystal PZN/PT polymer composites for ultrasound transducer applications,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 47, no. 4, pp. 792-800, 2000b. Rizzoni G. Principles and Applications of Electrical Engineering. Boston: McGraw Hill, 2000. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 165 Rosenwasser G., “Ultrasonic biomicroscopy in ophthalmology and eye banking,” Proceedings o f the SPIE - The International Society fo r Optical Engineering, vol. 3664, pp. 17-22, 1999. Selfridge A.R., The design and fabrication o f ultrasonic transducers and transducer arrays, Ph.D. thesis in Electrical Engineering, Stanford University, 1983. Sherar M.D., Starkoski B.G., Taylor W.B., Foster F.S., “A 100 MHz B-scan ultrasound backscatter microscope,” Ultrasound Imaging, vol. 11, no. 2, pp. 95-105, 1989. Shung K. Kirk, M.B. Smith, and B. Tsui. Principles of Medical Imaging. New York: Academic Press, 1992. Silverman R., “Improved System for Sonographic Imaging and Biometry of the Cornea,” J. Ultras. Med., vol. 16., pp. 117-124, 1997. Srinivasan S., Baldwin H.S., Aristizabal O., “Noninvasive in uterus imaging of mouse embryonic heart development using 40 MHz echocardiography,” Circulation vol. 98, pp. 912-918, 1998. Smith W.A., Auld B.A., “Modeling 1-3 composite piezoelectrics: thickness-mode oscillations,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 38, no. 1, pp. 40- 47, 1991. Snook K.A., Piezoelectric polymer transducer fabrication and design o f a high frequency exposimetry system, M.S. thesis in Bioengineering, The Pennsylvania State University, 2000. Snook K. A., Zhao J.Z., Alves C. H., Cannata J. M., Chen W.H., Meyer Jr. R.J., Ritter T.A., Shung K.K, “Design, fabrication, and evaluation of high frequency, single-element transducers incorporating different materials,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 49, no. 2, pp. 169-176, 2002a. Snook K., Shrout T., Shung K.K., “Design of a 50 MHz annular array using fine- grain lead titanate,” Proc. 2002 IEEE Applications o f Ferroelectrics, pp. 351-354, 2002b. Snook K. A., Design o f a High Frequency Annular Array fo r Medical Imaging, Ph.D. thesis in Bioengineering, The Pennsylvania State University, 2004. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 166 Snook K.A., Hu C.H., Shrout T.R., Shung, K.K., “High Frequency Ultrasound Annular Array Imaging, Part I: Array Design and Fabrication,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., this work has been accepted by IEEE for publication, 2005. Thiboutot D.M., “Dermatological applications of high-frequency ultrasound,” Proceedings o f the SPIE - The International Society fo r Optical Engineering, vol. 3664, pp. 7-16, 1999. Turnbull D.H., Starkoski B.G., Harasiewicz K.A., Semple J.L., From L., Gupta A.K., Sauder D.N., Foster F.S., “A 40-100 MHz B-scan ultrasound backscatter microscope for skin imaging,” Ultras. Med. Biol., vol. 21, no. 1, pp. 79-88, 1995a. Turnbull D.H., Bloomfield T.S., Foster F.S., Joyner A.L., “Ultrasound backscatter microscope analysis of early mouse embryonic brain development,” Proc. Natl. Acad. Sci., vol. 92, pp. 2239-2243, 1995b. Ye S.G., Harasiewicz K.A., Pavlin C.J. Foster F.S., “Ultrasound characterization of normal ocular tissue in the frequency range from 50MHz to 100MHz,” IEEE Trans. Ultrason., Ferroelec., Freq. Contr., vol. 42, no. 1, pp. 8-14, 1995. Zhao J.-Z., Alves,C.H., Snook K.A., Cannata J.M., Chen W.-H., Meyer, R.J. Jr., Ayyappan S., Ritter T.A., Shung K.K., “Performance of 50 MHz transducers incorporating fiber composite, PVDF, PbTi03 and LiNb03,” Proc. 1999 IEEE Ultrasonics Symposium, pp. 1185-1190, 1999. Zipparo M.J., Very high frequency (50 to 100 MHz) ultrasonic transducers for medical imaging applications, Ph.D. thesis in Bioengineering, The Pennsylvania State University, 1996. Zipparo M. and Oakley C.G., “Finite element modeling of PZN-PT and PMN-PT single crystal materials,” Proc. 2001 IEEE Ultrasonics Symposium, pp. 1017 -1022, 2001. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Array transducers for high frequency ultrasound imaging
PDF
Development of back-end processing system for high frequency ultrasound b-mode imaging
PDF
Cytoarchitecturally conformal multielectrode arrays for neuroscience and neural prosthetic applications
PDF
Acoustic microfluidic PZT transducers and temperature -compensated film bulk acoustic resonators
PDF
A high frequency array- based photoacoustic microscopy imaging system
PDF
Design and fabrication of a high frequency PMN-PT needle transducer for retinal blood flow measurement
PDF
A CMOS frequency channelized receiver for serial-links
PDF
Development of ceramic-to-metal package for BION microstimulator
PDF
High frequency ultrasonic phased array system and its applications
PDF
Improved contrast in ultrasound imaging using dual apodization with cross-correlation
PDF
Gyrator-based synthesis of active inductances and their applications in radio -frequency integrated circuits
PDF
Investigation of a switching G mechanism for MEMS applications
PDF
Development of high-frequency (~100mhz) PZT thick-film ultrasound transducers and arrays
PDF
Finite element analysis of the effects of stem geometry, surface finish and cement viscoelasticity on debonding and subsidence of total hip prosthesis
PDF
Single-cell analysis with high frequency ultrasound
PDF
Investigation of several important phenomena associated with the development of Knudsen compressors
PDF
English phoneme and word recognition by nonnative English speakers as a function of spectral resolution and English experience
PDF
A model of cardiorespiratory autoregulation in obstructive sleep apnea
PDF
Transducers and signal processing techniques for simultaneous ultrasonic imaging and therapy
PDF
A passive RLC notch filter design using spiral inductors and a broadband amplifier design for RF integrated circuits
Asset Metadata
Creator
Gottlieb, Emanuel John
(author)
Core Title
Development of high frequency annular array ultrasound transducers
School
Graduate School
Degree
Doctor of Philosophy
Degree Program
Biomedical Engineering
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
engineering, biomedical,OAI-PMH Harvest,Physics, Acoustics
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Shung, K. Kirk (
committee chair
), Cannata, Jonathan (
committee member
), Kim, Eun Sok (
committee member
), Yen, Jesse (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-608245
Unique identifier
UC11341367
Identifier
3220107.pdf (filename),usctheses-c16-608245 (legacy record id)
Legacy Identifier
3220107.pdf
Dmrecord
608245
Document Type
Dissertation
Rights
Gottlieb, Emanuel John
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
engineering, biomedical
Physics, Acoustics