ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january

Transcription

1 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january Interdigital Pair Bonding for High Frequency (20 50 MHz) Ultrasonic Composite Transducers Ruibin Liu, Kasia A. Harasiewicz, and F. Stuart Foster, Senior Member, IEEE Abstract Interdigital pair bonding is a novel methodology that enables the fabrication of high frequency piezoelectric composites with high volume fractions of the ceramic phase. This enhancement in ceramic volume fraction significantly reduces the dimensional scale of the epoxy phase and increases the related effective physical parameters of the composite,such as dielectric constant and the longitudinal sound velocity,which are major concerns in the development of high frequency piezoelectric composites. In this paper,a method called interdigital pair bonding (IPB) is used to prepare 1-3 piezoelectric composite with a pitch of 40 m,a kerf of 4 m,and a ceramic volume fraction of 81%. The composites prepared in this fashion exhibited a very pure thickness-mode resonance up to a frequency of 50 MHz. Unlike the 2-2 piezoelectric composites with the same ceramic and epoxy scales developed earlier,the anticipated lateral modes between 50 to 100 MHz were not observed in the current 1-3 composites. The mechanisms for the elimination of the lateral modes at high frequency are discussed. The effective electromechanical coupling coefficient of the composite was 0.72 at a frequency of 50 MHz. The composites showed a high longitudinal sound velocity of 4300 m/s and a high clamped dielectric constant of 1111"0,which will benefit the development of high frequency ultrasonic transducers and especially high frequency transducer arrays for medical imaging. I. Introduction Piezoelectric ceramic or crystal/polymer composites have been one ofthe simplest, most efficient and important techniques for tailoring the properties of ferroelectric materials used in the field ofelectromechanical transducers since the early 1980s [1]. The connectivity pattern ofthe diphasic composites was considered a key feature in the development ofthe composite properties. Thus, 10 connectivity patterns were established by Newnham et al. after the appearance of the composite concept [2]. This work provided a clear means ofclassifying and analyzing composite materials. It is now common for us to use Newnham sclassificationtorefertoacompositebyapairof digits m-n, where the m refers to the self-connectivity of the ceramic phase and the n refers to the self-connectivity Manuscript received January 31, 2000; accepted June 22, This work was supported by the National Cancer Institute of Canada, the Medical Research Council of Canada, and the Sir Jules Thorne Charitable Trust. The authors are with Imaging and Bioengineering, Sunnybrook & Women s College Health Sciences Centre, University of Toronto, Toronto, Ontario M4N 3M5, Canada ( stuart.foster@swchsc.on.ca). ofthe polymer phase. The power ofcomposite materials was rapidly demonstrated in the field ofhydrophone development, where a 400 enhancement ofhydrostatic figure ofmerit over the conventional PZT ceramics was achieved using a 1-3 composite [3], [4]. The development ofpiezoelectric composites for ultrasonic transducers in medical imaging and nondestructive evaluation (NDE) resulted in significant improvements in these applications as well. These improvements resulted in enhanced electromechanical coupling coefficients and decreased acoustic impedance over conventional piezoelectric ceramics [5], [6]. Traditional PZT s have a rod-mode electromechanical coupling coefficient k 33 ofapproximately 0.75 [7], and newly developed PZN-PT relaxor single crystals have exhibited a k 33 as high as 0.94 [8]. Unfortunately, the thickness-mode electromechanical coupling coefficient k t for these two materials, by which most ultrasonic transducers are judged, is only at the level of0.5 because ofthe effects oflateral clamping and spurious lateral modes. The observed improvement in k t for composites appears to be due to a reduction ofthe lateral dimension ofthe ceramic phase with concomitant reliefofthe lateral clamping. Modeling has been used extensively in the optimization ofcomposite geometry [9] [11]. The conclusions from the modeling studies indicate that the ratio ofwidth to height ofceramic phase in a composite should be below 0.6 to realize a pure thickness mode. The effective dielectric constant has a linear relationship with the ceramic volume fraction. Moreover, disturbances from Lamb wave resonances in the epoxy medium will cause a severe reduction in the composite performance ifthe scale ofthe epoxy kerfis too large. As a result, the fabrication technology for fine scale composites has always been challenging, particularly for high frequency applications. A wide range of techniques has been developed from coarse to fine scale [12] such that fabrication of efficient, inexpensive composites working in the frequency of 1 to 10 MHz is now routine. However, the frequency limit of composites made using these approaches appears to be approximately 20 MHz. New clinical applications ofultrasound in areas such as ophthalmology, dermatology, cartilage imaging, and vascular imaging require resolution in the range of50 to 100 µm, which in turn requires the development oftransducers working in the frequency range of 20 to 60 MHz [13]. Previous studies ofpiezoelectric ceramic properties at high frequency carried out by Foster et al. [14] and Zipparo et /$10.00 c 2001 IEEE

2 300 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january 2001 al. [15] have shown that PZT ceramics, and in particular fine-grained PZT ceramics, maintain reasonable piezoelectric properties up to approximately 100 MHz. However, current approaches appear impractical for the fabrication ofcomposites at frequencies greater than 20 MHz. Only a few attempts to fabricate high frequency composites have been reported. One approach utilized fiber processing [16], [17], and a second approach employed tape lamination [18]. However, the composites made by the former approach have a rather low ceramic volume fraction and high cost, and the latter approach has a relatively complex manufacturing procedure. Recently, by combining a process called interdigital pair bonding (IPB) with conventional dice-and-fill techniques, 2-2 piezoelectric composites were fabricated with a ceramic width of36 µm and an average epoxy kerfwidth of4 µm [19]. Center frequencies as high as 26 MHz were achieved. A weak lateral mode from the ceramic transverse vibration was found at a frequency of 50 MHz. Lateral modes from the epoxy phase were very small and located in the frequency above 70 MHz. Further study showed this composite could keep the pure thickness mode with a high effective k t of0.66 up to a frequency ofapproximately 40 MHz [20]. In this paper, we extend the IPB method to the second dimension for the fabrication of 1-3 composites and show that it is possible to produce 1-3 composites with good performance up to a center frequency of 50 MHz. II. Basic Principle of IPB:Processing Procedure for 1-3 Composite A. Basic Concept of IPB A wide variety oftechniques has previously been developed for the fabrication of piezoelectric composites [12]. 2-2 or 1-3 composite production by means ofmechanical machining is often referred to as the dice-and-fill method. This method originated from Penn State [21], and has become the most widespread and standard method for fabrication of1-3 piezoelectric composites because ofits simplicity and high efficiency [5], [12]. Unfortunately, all ofthe composite fabrication techniques have limitations and difficulties in fabricating composite with ceramic post width and epoxy kerfwidth below approximately 50 and 10 µm, respectively. Recently, tape lamination has shown promise as a means ofproviding a reproducible method offabricating 2-2 composites [18]. Using this approach, a 2-2 composite is made by stacking numerous 25-µm thick ceramic sheets with an epoxy width of6 µm separating each layer. Excellent preliminary results using tape lamination 2-2 composites have been reported at frequencies as high as 30 MHz [18]. However, for higher frequencies, this approach may be limited by its complexity and the difficulties ofreliably machining PZT ceramic plates to thicknesses less than 50 µm. The surface configuration of a typical 1-3 piezoelectric composite made by the dice-and-fill method is illustrated Fig. 1. Illustration of surface configuration of a 1-3 composite made by conventional dice-and-fill. in Fig. 1. The darker grey areas represent ceramic phase with a width of w c, and the lighter grey areas represent the epoxy phase. The width ofthe epoxy phase between the ceramic posts is w p, and the diagonal separation between the ceramic posts is w d. The latter two dimensions are related to the Lamb wave resonant frequencies of f t1 and f t2, respectively, which will be discussed in Section IV. Recognizing that the thickness ofthe ceramic phase is equal to λ/2, where λ is the wavelength ofacoustic wave in the resonator, the width to thickness ratio, G, for a pure thickness-mode vibration must satisfy the relationship G = 2w c 0.6. (1) λ Eq. (1) leads to the following constraint on the lateral dimension ofceramic phase: w c 0.3λ. (2) Assuming a uniform distribution of the ceramic phase in the epoxy phase, the volume fraction of the ceramic phase is given by VF = wc 2 (w c + w p ) 2 = (0.3λ) 2 (0.3λ + w p ) 2. (3) For example, using PZT-5H with a longitudinal sound velocity of4560 m/s as the ceramic phase, plots ofvolume fraction vs frequency for various values of w p are given in

3 liu et al.:interdigital pair bonding 301 B. Processing Procedure of 1-3 Piezoelectric Composite byipb Fig. 2. Theoretical calculation of ceramic volume fraction of 1-3 composite as a function of operation frequency for three epoxy kerf widths. Fig. 2. Note that the volume fraction of the ceramic phase drops rapidly with increasing frequency. At a frequency of30 MHz, the desired epoxy kerfwidth is below 10 µm [18]. In a conventional dice-and-fill composite, the width ofthe epoxy is governed by the width ofdicing saw cuts, which are, in turn, dictated by the dicing saw blade thickness and alignment. A typical lower limit width ofcurrent dicing saw cuts is approximately 20 µm. From Fig. 2, we can see that the composites are subject to a low volume fraction at high frequencies even if we use this minimum epoxy kerfwidth. A low ceramic volume fraction in the composite will cause a low effective dielectric constant and low effective sound velocity [9] [11], which are undesired for high frequency transducers because of the difficulties in electrical matching and fabrication of thin resonators. In fact, the curves in Fig. 2 represent an ideal case that is not possible to reach because the effective sound velocity ofthe composite is also a function ofceramic volume fraction and is always lower than the velocity of the ceramic we used for calculation. Thus, based on these considerations, it is clear that new approaches are needed to maximize ceramic volume fraction for frequencies above 25 MHz. The IPB method has the capability to increase the ceramic volume fraction and decrease the kerf width of the epoxy phase for high frequency composites. This idea first appeared in 1993 [22]. The basic principle is to impregnate epoxy and interdigitally insert a pair ofidentical diced ceramic sheets. This method has been shown to be very suitable for fabrication of high frequency 2-2 composites [19], [20]. In IPB, we abandon the conventional ratio of composites in which w c w p and deliberately dice the ceramic with w c <w p. Then, by interdigitally inserting two similarly diced samples with impregnating epoxy as shown in Fig. 3, it is possible to increase the volume fraction of ceramic phase greatly while maintaining a narrower kerf. The details ofthe procedure ofthe fabrication of1-3 composite by IPB are illustrated in Fig. 3. The first step [Fig. 3(a)] involves dicing identical parallel grooves with w c <w p. Two pieces ofceramic sheets processed in Fig. 3(a) are then impregnated with epoxy and interdigitally inserted [Fig. 3(b)]. The top surface of this sandwich is lapped away to reveal the 2-2 composite as shown in Fig. 3(c). At this point, the ceramic volume fraction has been doubled, and the kerfwidth ofthe epoxy is equal to (w p w c )/ composite ofthe type shown in Fig. 3(c) is subsequently processed into 1-3 composite by repeating the processes offig. 3(a c) at right angles to the original 2-2 cutting direction. These steps are illustrated in Fig. 3(d f). III. Experimental Procedures High frequency 1-3 composites were prepared according to the IPB procedure offig. 3. Motorola 3203 HD piezoelectric ceramics and commercial PZT-5H ceramics (Materials System Inc., Littleton, MA) were used as the ceramic phase in the composites. The epoxy used for bonding and backfilling was West System R epoxy resin 105 and hardener 206 (Gougeon Brothers Inc., Bay City, MI) and EPO-TEK (Epoxy Technology, Billerica, MA). A Disco 2H/6TM (Salem, NH) automatic dicing saw was used for cutting grooves and Cr/Au electrodes were evaporated on to both sides ofsamples using a commercial evaporator made by CHA Industries (Fremont, CA). The impedance spectrum ofthe 1-3 composite was measured by an HP4191A (Palo Alto, CA) impedance analyzer. The effective dielectric constant (free and clamped) and electromechanical coupling coefficient were derived from the impedance values. A dicing saw blade with a thickness of40 µm wasused for the fabrication. The index distance perpendicular to the cutting grooves was set at 80 µm. The resultant width ofcutting grooves was 44 µm, and the width ofthe ceramic strips was 36 µm. Therefore, the average epoxy kerf width was 4 µm. The ceramic volume fraction for the composite was 81%. Two types of1-3 composites were made. Composite A was made using the PZT-5H ceramics and the EPO-TEK epoxy, and composite B was made by the Motorola 3203 HD ceramics and the West System R epoxy. The thickness for composite A was 44.4 µm and for composite B was 40.6 µm. The electrode area for composite A was 4.9 mm 2, and that for composite B was 1.21 mm 2. A photograph ofa type composite A was taken under microscope by transmission light and is shown in Fig. 4. The effective electromechanical coupling coefficient ofthe thickness mode was calculated by the following equation [23]: k 2 t = π 2 f s tan π f p f s (4) f p 2 f p

4 302 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january 2001 Fig. 3. Processing procedure for 1-3 composite by interdigital pair bonding. See text for details. where the f s and f p are series and parallel resonant frequencies, respectively. The dielectric constants offree and clamped conditions were calculated from impedance values in the frequency range suggested by Foster [14]. IV. Results and Discussions A. Electrical and Electromechanical Properties of the Composites The air-loaded impedance responses ofboth composites A and B were measured in the frequency range of 10 to 100 MHz as shown in Fig. 5. The effective free and clamped dielectric constant, ε T 33 and ε S 33, respectively, electromechanical coupling coefficient of the thickness mode k t, as well as the longitudinal sound velocity under open circuit conditions v l were obtained from the impedance spectrum. The free dielectric constants of the composites were derived from the impedance values at a frequency of 10 MHz. The clamped dielectric constants were derived from impedance values at 90 MHz. Table I gives the measured properties ofcomposites A and B. It should be pointed out that the G ratios ofwidth to thickness ofthe ceramic phase in the composites were 0.8 and 0.88 for composites A and B, respectively. However, both composites still demonstrated very pure thickness resonant modes as demonstrated in the impedance spectra offig. 5. The derived k t values (around 0.72) for both composites also confirm that the vibration is a relatively pure equivalent rod-mode vibration. In contrast to the theoretical predications for lower frequency resonators [9] [11], [24], transverse modes had little influence on the thickness-mode resonance ofthe high frequency composites reported here. The possible mechanism for this phenomenon will be discussed in the next subsection. The electrical and acoustic properties for both composites are very close, even though different types ofceramic and epoxy materials were utilized in each. This indicates that the processing procedures and subtle choice ofceramic and epoxy materials are not critical determinants ofperformance. The high dielectric constant and sound velocities for the composites are consistent with the high ceramic vol-

5 liu et al.:interdigital pair bonding 303 TABLE I Measured Properties of Composites A and B and Some Related Materials. Ceramic w c w p v 1 ε T 33 ε S 33 volume Properties (µm) (µm) k t (m/s) ε 0 ε 0 (%) 1-3 Composite A 36 ± 2 4± Composite B 36 ± 2 4± Composite 36 ± 2 4± Ceramic 3203 HD Fig. 4. Picture of 1-3 composite A made by the IPB. The shadows are 1-mm apart. ume fraction associated with IPB. The longitudinal sound velocity ofcomposite B is slightly higher than composite A, likely because the ceramic phase in composite B has a higher elastic modulus than many conventional ceramics [15]. Both composites A and B achieved a high value of k t =0.72, which is very close to the value expected for a pure rod mode (k 33 =0.75) [7], [15]. B. Influence of Lateral Modes and EpoxyType Fig. 5. Air-loaded impedance spectra for composites A (top) and B (bottom). The corresponding lateral mode resonance occurs at Similar to the 2-2 composite, the 1-3 composites are also subject to two kinds oflateral modes. One is from the transverse mode ofthe ceramic phase, which is determined by the transverse sound velocity v t and the lateral dimension w c ofthe ceramic phase. The transverse sound velocity under short circuit condition for typical PZT-5H type ceramic is given by [9] v t = c E 13 ρ ceramics = N/m kg/m 3 = 3600 m/s. (5) f l = v t 2w c = 50 MHz for wc =36µm. (6) Another major contribution to the lateral modes are from the resonance ofbloch reflection by the periodic lattice structure formed by the ceramic phase when standing Lamb waves run along the epoxy medium [25], [26]. The resonant frequencies f t1 and f t2 ofthese lateral modes are proportional to the shear wave speed v s ofthe polymer material and inversely proportional to the scales ofthe epoxy phase of w p and w d, respectively, as shown in Fig. 1.

6 304 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january 2001 Fig. 6. Amplified impedance spectra for composites A and B in the frequency range from 70 to 100 MHz. Therefore, the expected resonant frequencies are [10]: f t1 = v s 2w p = 91 MHz (7) v s f t2 = = 65 MHz (8) 2w d ( ) c where v s = = 9 N/m 2 = ρ 1100 kg/m epoxy m/s and w p =6µm, w d =8.5 µm. Here we used the wider kerfvalues because ofthe unsymmetrical location ofceramic posts in the cutting grooves. Both ofthese lateral modes were demonstrated in the 2-2 composites studied earlier [19], [20]. However, in the current 1-3 composites, we do not observe the lateral mode from the ceramic phase at 50 MHz. Two reasons could be responsible for this phenomenon. One possibility is that the lateral mode is too close to the fundamental mode to be distinguished. Another reason could be that this lateral mode was highly damped because ofthe increased number ofboundaries in 1-3 composite compared with 2-2 composite. Also in 1-3 composites, we do not see significant contribution oflateral modes from the epoxy phase. Although there are two fluctuations at the frequency of 78 and 90 MHz as shown in Fig. 6 for composites A and B, the strength is too small to be significant. The reason for the disappearance of these modes may be due to the high attenuation ofthe wave propagation at high frequency. It is well known that attenuation caused by viscous losses and thermal conduction is proportional to the square of the frequency. In addition, further attenuation will be caused by the scattering offinite-size grains or by dislocations in the matrix [27]. Finally, the randomness ofceramic element spacing caused by the IPB process itself(see Fig. 4) serves to further diminish the strength ofany resonances. This fact is in agreement with the work ofnegreira et al. [28] who tested the effects ofrandomness to improve composite performance. The asymmetrical pattern ofceramic element distribution will have little effect on the macroscopic properties ofthe composite. However, this nonuniformity in the microstructure could be an important factor in the disappearance of the resonance modes from the Lamb waves. Although experiments show that the k t ofthe composites is not strongly affected by the properties ofthe polymer phase materials [29], the choice ofepoxy with essential features such as low density, low acoustic impedance, good adhesion, and low cure contraction rate are preferred to avoid air bubbles and damage ofceramic posts in the composite. Low viscosity is usually preferred for good adhesion. In our experiments, the EPO-TEK epoxy had a very low viscosity (cure time is 48 h), and the West System R epoxy had higher viscosity with shorter cure time. It was found that the high viscosity epoxy, which offers a short cure time, worked well in our fabrication. The reason for this is due to the nature ofthe IPB processing. The ceramic strips slide inside grooves filled with epoxy, leading to the destruction ofbubbles and promoting adhesion. V. Conclusions It is simple and efficient to fabricate 1-3 composites for high frequency applications by introducing two IPB processing procedures into the conventional dice-and-fill method. 1-3 composites made by this method show very promising properties for application in ultrasonic transducers and transducer arrays up to frequencies of 50 MHz. In these experiments, it was found that a pure thickness mode and high electromechanical coupling coefficient can be obtained for a high ceramic width-to-thickness ratio, which differs from earlier theoretical predictions. The effective electromechanical coupling coefficient of the composite was 0.72 at a frequency of 50 MHz. The composites showed a high longitudinal sound velocity of4300 m/s and a high clamped dielectric constant of1111ε 0.Theanticipated lateral modes from both ceramic and epoxy phases between 50 to 100 MHz were not observed. The likely cause ofthis phenomenon is related to both increased losses in the matrix and to randomness introduced by the IPB process. The suppression ofthese modes may allow the fabrication ofcomposite up to frequencies ofapproximately 100 MHz with little additional reduction in the lateral scale ofthe ceramic phase.

7 liu et al.:interdigital pair bonding 305 References [1] R. E. Newnham, Transducers, sensors, and actuators, Jpn. J. Appl. Phys., vol. 25, suppl. 25-1, pp. 9 4, [2] R. E. Newnham, D. P. Skinner, and L. E. Cross, Connectivity and piezoelectric-pyroelectric composites, Mat. Res. Bull., vol. 13, pp , [3] M. Haun and R. Newnham, Experimental and theoretical study of 1-3 and piezoelectric PZT-polymer composite for hydrophone applications, Ferroelectrics, vol. 68, pp , [4] L. E. Cross, Ferroelectric materials for electromechanical transducer applications, Jpn. J. Appl. Phys., vol. 34, pt. 1, pp , [5] W. A. Smith, The role of piezocomposites in ultrasonic transducers, in Proc IEEE Ultrason. Symp., Montreal, pp [6] K. K. Shung and M. Zipparo, Ultrasonic transducer and arrays, IEEE Eng. Med. Biol., vol. 15, pp , [7] S. E. Park and T. R. Shrout, Characteristics of relaxor-based piezoelectric single crystals for ultrasonic transducers, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 44, pp , [8] X. C. Geng, T. A. Ritter, and S. E. Park, Characterization of electromechanical properties of relaxor-pt piezoelectric single crystals, in Proc IEEE Ultrason. Symp., Sendai, Japan, pp [9] W. A. Smith and B. A. Auld, Modeling 1-3 composite piezoelectrics: Thickness-mode oscillations, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 38, pp , [10] X. Geng and Q. M. Zhang, Resonance modes and losses in 1-3 piezocomposites for ultrasonic transducer applications, J. Appl. Phys., vol. 85, pp , [11] A. H. Nayfeh, J. J. Dong, and W. Faidi, Approximate model for wave propagation in piezoelectric materials. II. Fibrous composite, J. Appl. Phys., vol. 85, pp , [12] V. F. Janas and A. Safari, Overview of fine-scale piezoelectric ceramic/polymer composite processing, J. Amer. Ceram. Soc., vol. 78, pp , [13] F. S. Foster, C. J. Pavlin, G. R. Lockwood, L. K. Ryan, K. A. Harasiewicz, L. Berube, and A. M. Rauth, Principles and applications of ultrasound backscatter microscopy, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 40, pp , [14] F. S. Foster, L. K. Ryan, and D. H. Turnbull, Characterization of lead zirconate titanate ceramics for use in miniature highfrequency (20 80 MHz) transducers, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 38, pp , [15] M. J. Zipparo, K. K. Shung, and T. R. Shrout, Piezoelectrics for high-frequency ( MHz) single-element imaging transducers, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 44, pp , [16] T. F. McNulty, V. F. Janas, and A. Safari, Novel processing of 1-3 piezoelectric ceramic/polymer composite for transducer application, J. Amer. Ceram. Soc., vol. 78, pp , [17] R. J. Meyer, Jr., P. Lopath, S. Yoshikawa, and T. R. Shrout, High frequency 1-3 composite transducer fabricated from alkoxide-derived PZT fibers, in Proc IEEE Ultrason. Symp., Toronto, pp [18] T. Ritter, K. K. Shung, X. C. Geng, P. Lopath, R. Tutwiler, and T. R. Shrout, Composite ultrasound transducer arrays for operation above 20 MHz, in Proc. SPIE, vol. 3664, San Diego, 1999, pp [19] R. Liu, K. A. Harasiewicz, D. Knapik, N. A. Freeman, and F. S. Foster, 2-2 piezoelectric composites with high density and fine scale fabricated by interdigital pair bonding, Appl. Phys. Lett., vol. 75, pp , [20] R. Liu, D. Knapik, K. A. Harasiewicz, and F. S. Foster, Fabrication of 2-2 piezoelectric composite by interdigital pair bonding, in Proc IEEE Ultrason. Symp., pp [21] H. P. Savakus, K. A. Klicker, and R. E. Newnham, PZT- Epoxy piezoelectric transducer: A simplified fabrication procedure, Mat. Res. Bull., vol. 16, pp , [22] J. W. Sliwa, Jr., S. Ayter, and J. P. Mohr, III, Method for making piezoelectric composite, U.S. Patent , [23] IEEE Standard on Piezoelectricity, ANSI/IEEE Standard , [24] A. R. Selfridge, The fabrication of ultrasonic transducers and transducer arrays, Ph.D. dissertation, Electrical Engineering, Stanford University, Palo Alto, CA, [25] B. A. Auld and Y. Wang, Acoustic wave vibration in periodic composite plates, in Proc. IEEE 1984 Ultrason. Symp., B. R. McAvoy, Ed. Dallas, TX, pp [26] D. Certon, O. Casula, F. Patat, and D. Royer, Theoretical and experimental investigation of lateral modes in 1-3 piezocomposites, IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 44, pp , [27] G. S. Kino, Acoustic Waves: Device, Imaging andanalogue Signal Processing. Englewood Cliffs, NJ: Prentice-Hall, Inc., 1987, pp [28] C. Negreira, H. Gomez, N. Perez, I. Nunez, and J. Eiras, Lateral modes and diffracted field behaviour in non-periodical 1-3 piezocomposite transducers, in Proc IEEE Ultrason. Symp., Sendai, Japan, pp [29] K. Han and Y. Roh, The performance of a 1-3 mode piezocomposite ultrasonic transducer in relation to the properties of its polymer matrix, Sens. Actuators, vol. 75, pp , Ruibin Liu was born in Kunming City, Yunnan Province, China on July 18, He received his B.Sc. degree in electroceramics from Electrical Engineering Department of The South China University of Science and Technology, Guangzhou, China in 1984 and Ph.D. degree in materials engineering from The Shanghai Institute of Ceramics, Chinese Academia of Sciences, Shanghai, China in From 1991 to 1996, he was an engineer in Shanghai Institute of Ceramics. His research interest included fabrication of pyroelectric ceramics for the application in IR detection and imaging. From 1996 to 1999, he worked as a postdoctoral fellow in The Materials Research Laboratory of the Pennsylvania State University. His research interests there included piezoelectric actuators, single crystal thin film and electrostrictive polymers. He is presently a postdoctoral fellow of Sunnybrook & Women s College Health Sciences Centre, University of Toronto. His current research interests include piezoelectric composites and ultrasonic transducers for high frequency medical imaging. Kasia A. Harasiewicz was born in Lublin, Poland, on September 16, She received the M.Sc. degree in solid-state electronics from the Technical University of Warsaw, Poland, in From 1982 to 1991, she was a Senior Technician with the Ontario Cancer Institute. She is presently the Chief Ultrasound Research Engineer with the Sunnybrook & Women s College Health Sciences Centre. Her interests include the development of new ultrasound systems for breast imaging and backscatter microscopy. Ms Harasiewicz received the P. Eng. Title from the Association of professional Engineers of Ontario in F. Stuart Foster (M 90 SM 95) received the B.A.Sc. degree in Engineering Physics from the University of British Columbia, Vancouver, BC, Canada, in 1974, and M.Sc. and Ph.D. degrees in Medical Biophysics from the University of Toronto in 1977 and 1980, respectively. From 1980 to 1990 he was a Senior Scientist with the Ontario Cancer Institute in Toronto, Canada. He is presently a Senior Scientist with the Sunnybrook Health Science Centre, professor, and Associate Chairman of Medical Biophysics at the University

8 306 ieee transactions on ultrasonics, ferroelectrics, and frequency control, vol. 48, no. 1, january 2001 of Toronto. Dr. Foster is the recipient of a Terry Fox Cancer Research Scientist Award from the National Cancer Institute of Canada and has been involved with the development of new ultrasonic imaging systems since He has made important contributions to the development systems for the detection and evaluation of prostate, breast, and ocular cancers. His current research centres on the development of high frequency imaging systems, tissue characterization, array technology and intravascular imaging. Dr. Foster has published over 100 papers in the field of medical ultrasound imaging, and has recently authored a book on high frequency imaging. He has twice won the Ultrasound in Medicine and Biology Prize. Dr. Foster was the Distinguished Lecturer for the Ultrasonics Ferroelectrics and Frequency Control Society. In 1997 he won the Thomas Eadie Medal for major contributions to Engineering and Applied Science in Canada from the Royal Society of Canada. He is a fellow of the American Institute of Ultrasound in Medicine and serves on the editorial boards of Ultrasonic Imaging and Ultrasound in Medicine and Biology. He is currently the Chairman of the Ultrasonics Ferroelectrics and Frequency Control Society Nominations Committee.