We have developed an innovative, nonresonant wide bandwidth multilayer PVDF ultrasonic transducer, well suited for both conventional clinical imaging in the frequency range 3 - 15 MHz as well as for high frequency back-scatter microscopy imaging between 20 - 100 MHz. Such operation is desirable in clinical practice, eliminating the need for the separate probes used to minimize the tradeoff between achievable penetration depths and desired image resolution. The unique properties of our transducer were achieved by stacking individual PVDF layers in a parallel or anti-parallel polarization direction following a Barker coded pattern. The thickness of a single layer determines our transducer's bandwidth; its electrical properties are similar to those of conventional PZT transducers; and its overall pulse-echo sensitivity is sufficiently high for directly interfacing with a commercially available ultrasound imaging system. Using our transducer model, key parameters of the design were predicted and compared with single layer PZT and PVDF transducers in the 3 - 15 MHz and 25 - 100 MHz frequency ranges, respectively. Several prototypes of our wide bandwidth multilayer transducers were fabricated and tested in water. Agreement between experimental results and corresponding computer predictions indicate that the multilayer design outperforms the PZT transducer with respect to axial resolution and overall pulse-echo sensitivity in the frequency range 3 - 100 MHz.

We have fabricated a miniature 120-MHz transducer for imaging the internal structure of living samples, and mounted it in a 3-mm-diameter rod-shaped probe which ensures contact with a tissue to evaluate the tissue imaging capability of the transducer. The transducer consists of a thin film of 12.5-micrometer thick ZnO sandwiched between two metal electrodes, the bottom one deposited on a sapphire substrate whose other face has a polished concave-sphere acoustic lens. Both the lens diameter and the sphere radius are 0.5 mm; that is, the F number of the lens is 1. The lens of the transducer faces outwards in the probe so that the ultrasound can be transmitted and received directly by it in the radial direction of the rod without any mirrors. As the probe rotates mechanically around its axis and shifts in the direction of the axis, a cylindrical plane created by the locus of the beam focus is located inside of the tissue. Using this scanning, we form tissue images in the C-scan mode in a cylindrical plane within the target tissue. Preliminary results for imaging an in vitro bovine kidney sample into which the probe was inserted demonstrate that the fabricated probe can image microscopic structure inside tissue samples.

In order to improve the sensitivity and the bandwidth of an ultrasonic probe, the single-element probe using single crystals of the solid solution Pb(Zn_1/3Nb_2/3)O₃- PbTiO₃ (PZN-PT) has been investigated. Single crystals of PZN-PT with 9 mol% PbTiO₃ (PZN-PT 91/9) single crystals were grown by the self flux method using PbO-based flux. It was confirmed that the single crystal of good quality with the dimensions of about 25 multiplied by 15 multiplied by 5 mm which is capable of fabricating a phased array probe for echocardiography has been grown. As a first step, a plane single-element probe of 2.0 mm in diameter and 20 MHz was fabricated for gastrointestinal lesions. Mechanical strength of the new crystal material in the fabricating process was strong enough since no damage such as cracking occurred. The sensitivity of the PZN-PT probe has been improved compared with that of the conventional PZT one. Also,the bandwidth has been noticeably broadened.

It is difficult to do significant work on transducers in an academic or small-shop setting since the machinery required for the fabrication of the micro-sized elements is expensive, and the needed human expertise takes a long time to acquire. The expense of preparing for prototype production must be committed even before the design has been tested. This paper discusses a means to allow fabrication of low-frequency models of high-frequency transducers in the average academic laboratory or model shop -- without requiring exotic saws, presses, etc., for preliminary evaluation of the design. These models will have the added advantage that they can be more realistic models of actual transducers than present computer models. In addition, we show that acoustic fields can be scaled so that field models of scattering and propagation can be used to investigate these effects.

Ultrasonic transducers having curved radiating surfaces may offer a simple solution to maintaining good lateral resolution over the large depth of field required in medical imaging. In this paper the design considerations for such a transducer that consists of a cylindrical metal housing and an ultrasonic wave generating piezoceramic disc is presented. The mechanism of focusing the radiated ultrasonic wave is studied by changing the geometry of the front surface of the metal housing. The propagation of ultrasonic wave in the surrounding medium is analyzed using the impulse response approach for the near field region and Fraunhofer's approximation for the far field. In addition, modal analysis of the transducer structure is conducted using the finite element method. The results obtained show that the geometry of the transducer housing has significant effects on the radiation characteristics of the transducer.

The design, fabrication and initial characterization of a prototype fully lambda/2 sampled, 2.5 MHz, 50 by 50 element 2D array for cardiac medical imaging applications is presented. The array utilizes a novel Z-axis electrically conductive backing layer to provide both appropriate acoustic attenuation, and an anisotropic electrical interconnect for the individual acoustic elements in the 2D array. A modular, demountable pad grid array (PGA) interconnect is used to connect the backing. The PGA is capable of terminating the full 2500 element array at a spatial pitch of 300 microns. Measurements are presented on the electrical impedance, directivity and cross talk of the 2D array module, as well as the pulse echo properties of the 2D array elements excited through the pad grid array interconnect system. The single element directivity is measured to be 35 degrees, while the nearest neighbor electrical cross talk is measured at minus 42 dB. The pulse echo waveform has a fractional bandwidth of 50%.

A 1.5D transducer is capable of only elevation aperture control or performing focusing, shading, and aperture control. Elevation steering is typically not included as one of its capabilities. Different version of 1.5D transducer are considered here. Methods of optimizing the various 1.5D array parameters are discussed and examples are used to illustrate the principles involved.

High-intensity focused ultrasound is used in many therapeutic applications such as drug activation/drug delivery, hyperthermia, cancer therapy, ultrasound surgery and myocardial ablation. Various ultrasonic systems have been proposed for these therapeutic applications. While many applicators produce adequate power levels, multiple element ultrasound phased arrays adjust for phase aberrations, focus around obstructions such as bone and/or air spaces (lungs), and follow, in real time, a moving target. Since large aperture arrays with several hundred elements are required, design compromises keep the element count and fabrication cost at a reasonable level. These trade-offs, which optimize the array aperture with respect to element count, often result in a non-ideal aspect ratio (element width to thickness), leading to lateral mode vibrations which reduce the electrical to acoustical efficiency to about 10 - 20%. These vibrations are easily observed with a laser interferometer system. Piezo composite technology, which eliminates the non-ideal aspect ratio by dividing the individual array elements into long, thin rods, provides a solution to this problem. The spaces between the rods are filled with a polymer to provide structural support and allow deposition of electrode layers to interconnect individual rods and to outline array elements. Several piezo composite transducers have been tested, and initial results show a greatly improved beam pattern and increased efficiency. Power handling capability of composites has recently improved allowing outputs in excess of 10 watts/cm² with efficiencies greater than 60%. This is sufficient for many therapeutic applications.

The potential of 1.5D ultrasonic imaging is investigated, using an elevation sampling of 7 elements (4 channels). The number of elements has been chosen in accordance with the number of channels that will be available in commercial beamformers in the future. The probe technology has been developed for curved array geometry with no limitation in radius of curvature. The transducer has been designed to introduce no limitation of electroacoustic performances, allowing fair comparison with images obtained with 1D arrays. Possible approaches for the array design are discussed. Achieved performances on prototypes of a 3.5 MHz, 40 mm radius, 128 by 7 elements, 0.5 by 2.4 mm pitch probe are presented. Electroacoustic performances are in the range of state-of-the-art 1D probes. Several figures of merit are defined and analyzed. Theoretical and experimental comparisons with 1D arrays show drastic improvement for various focusing strategies. Prototypes of curved-linear 1.5D probes with high density of elements, exhibiting performances consistent with high end diagnosis equipment, and using fabrication processes suitable for industrial production, are now available. First clinical evaluations have been conducted and confirm very promising capabilities. The results will be used for the definition of an optimized 1.5D probe family.

A model of a medical ultrasound imaging system can be broken into four major subsystems: the transducer, the array, the beamformer, and display system. In this work, each sub- system is modeled separately and combined to simulate the complete system. Transducer elements are modeled as three port systems with two mechanical ports and one electrical port. The array is modeled as a product of a point source array and finite radiating elements. The beamforming system is modeled as a frequency domain beamformer with depth focusing, utilizing a parallel computing architecture. Each of these systems offers flexibility to model a one dimensional linear phased array, a 1.5D array or a two dimensional array, or an array of arbitrary shape.

This paper presents a system model and an inversion for phased array ultrasonic imaging with a moving target within a stationary background. The imaging method is based on the inversion principle that utilizes the spatial Fourier decomposition of the obtained data in the array elements. Velocity estimation can be obtained by examining the Doppler shift caused by a target moving along a scanning beam axis and Doppler spread caused by the same target moving normal to the beam. This process includes two filters in the reconstructed spatial domain and spatial frequency domain for separating a moving target's signature from the stationary background. In addition, the Doppler shift does not respond to the pulse propagation time but instead responds to the scanning time which is much slower. This slow time method makes Doppler shift and spread computable in the processing. After the velocity estimation the shifted and smeared moving target signature is able to be correctly and clearly imaged with estimated velocity vector values in the reconstructed spatial domain by a phase compensation in the spatial frequency domain.

The fast development of computer power and computational algorithms has made it possible to design complex transducers and arrays using computer simulation. Practical example are given here for transducer modeling using 3-D finite element method (FEM) instead of the traditional 1-D equivalent circuit models. The merits and deficiencies of the frequency domain and time domain FEM formulations also are analyzed.

Diagnostic medical ultrasound systems generally use only a portion of the available bandwidth of modern transducer arrays. A new method is presented that makes optimal use of all available bandwidth. This method achieves improved lateral resolution with no loss of frame rate. The performance of the method in the presence of tissue related phase aberration and attenuation is considered. The key elements of the additional electronic hardware required to support this technique are discussed. The extension of the method to use with 1.5D arrays and annular arrays is also discussed.

Interconnecting a microminiature coaxial cable to a fully populated two-dimensional transducer array for diagnostic ultrasound is a significant challenge due to high element counts and termination densities. Pad grid array (PGA) structures allow inline interconnect to array contacts, but have not been previously demonstrated to allow elements to be interconnected at spacing matching the pitch of transducer arrays typical of 2.5 to 7.5 MHz diagnostic imaging (300 to 100 mm). PGAs provide a capable platform for experimentation and to allow transducer replacement because they are demountable. This paper reports on such a structure to allow a cabled 50 by 50 contact PGA to be connected by means of an anisotropic elastomeric anisotropic material to a 50 by 50 element transducer backing structure at 300 micron pitch, the acoustic half-wavelength normally used for 2.5 MHz imaging. Multiple flexible printed circuits provide the interconnect between coaxial cabling and a precision- drilled monolithic alumina substrate. The substrate is finished to provide a planar array of contact pads, each of which is used to contact gold wires embedded in a silicone medium which in turn connect to the backing electrodes of the transducer module. Using this approach, 97% of targeted interconnects were successfully accomplished. Acceptable pulse echo performance was demonstrated, suitable for diagnostic imaging.

Devices including ultrasound and magnetic resonance imaging probes can 'overheat' and burn human skin unless they are carefully designed and tested. An empirical study was performed to determine how much thermal power the skin can absorb without raising skin temperature to the damage point. Steady-state power and temperature measurements were recorded from seven healthy adults. Small skin areas, 1.8 to 25 cm², were heated. The data indicates a 'safe' absorption level of approximately 40 mW/cm². Near the overheating point, skin temperature increases approximately 0.8 degrees Celsius for each additional 10 mW/cm² of absorbed power.

An important part of ultrasonic transducer design is the optimization of the acoustic impedance matching layer. Since there is a large acoustic impedance mismatch between piezoceramic and human tissue the majority of the acoustic energy from the transducer would be reflected without the use of an intermediate matching layer. This matching layer can improve transducer performance; particularly the transit/receive sensitivity. We report here a study of the optimum matching layer thickens of a transducer using both experimental results and a finite element model. For reliable input parameters we have made a set of 0 - 3 alumina-epoxy composites and measured the material properties. The method used to make the composite is similar to that described by Grewe.

Relaxor ferroelectric single crystals of Pb(Zn_1/3Nb_2/3)O₃ (PZN), Pb(Mg_1/3Nb_2/3)O₃ (PMN) and their solid solutions with normal ferroelectric PbTiO₃ (PT) were investigated for ultrasonic transducer applications. Crystals offer adjustable properties not only by compositional tailoring but also by domain state engineering associated with different crystallographic orientation, which is not achievable in polycrystalline materials. Longitudinal coupling coefficients (k₃₃) as high as 94% and dielectric constants (K₃^T) in the range of 3500 - 6000 were achieved with low dielectric loss (less than 1%) using <001> oriented rhombohedral crystals of (1-x)PZN-xPT and (1-y)PMN-yPT, where x less than 0.09 and y less than 0.35. Dicing direction as well as poling direction were critical for high coupling under laterally clamped condition. Dicing parallel to the (001) yields 90% of laterally clamped coupling (k_bar) out of 94% longitudinal coupling (k₃₃) for PZN-8%PT. On the other hand, samples diced parallel to (110) exhibited no dominant mode present. Thickness coupling (k_T) as high as 64% and low dielectric constant (K₃^T) less than 600 with low loss (less than 1%) could be achieved using tetragonal crystals of (1-x)PZN-xPT and (1-y)PMN-yPT, where x greater than 0.1 and y greater than 0.4. The performance gains associated with these ultra-high coupling coefficients and range of dielectric constants are evident in relation to broader bandwidth and electrical impedance matching. Specifically, rhombohedral crystals offer the possibility of extremely broad bandwidth devices for transducer arrays and tetragonal crystals for single element transducers. Transducer simulation was performed using the KLM model. The pulse/echo response simulated a 124% bandwidth subdiced array element with a center frequency of 10 MHz. An optimized array design of the same geometry constructed of PZT 5H displays an 87% bandwidth.

The properties of piezoelectric materials operating at high frequencies greatly influence the level of transducer performance which is achievable. Selection of the appropriate material can be made based on the transducer area and operating frequency. The properties of a number of piezoceramic materials have been experimentally determined by measuring the electrical impedance of air-loaded resonators whose thickness corresponds to resonance frequencies from 10 to 100 MHz. Materials measured include commercially available high dielectric lead zirconate titanate (PZT) and lower dielectric modified lead titanate (PT) ceramics, as well as materials which have been designed or modified to result in improved properties at high frequencies. Conclusions regarding the influence of the microstructure and composition on the frequency dependence of the properties are made based on the calculated properties and microstructural analysis of each material. Issues which affect transducer performance are discussed in relation to the measurements. For larger area transducers the use of a lower dielectric constant material is shown to result in a better electrical match between the transducer and standard 50 Omega terminations. For transducers whose impedance is close to that of the connecting cables and electrical terminations, KLM model simulations show improved performance without the need for electrical matching networks which can narrow the bandwidth and introduce additional losses. Measurements of actual transducers show close agreement with the simulations, verifying the material property measurements and the performance benefits of electrically matched transducers.

This paper reports on our determination of the small signal properties of Motorola 3203 HD, including the effects of losses and dispersion, and it shows the limitations imposed by assuming that the material is loss-less. A set of PZT 3203 HD unloaded resonators manufactured by Motorola was cut to specifications outlined in the IEEE Standard on Piezoelectricity Std 176-1987 to ensure the appropriate boundary conditions of each resonance mode. Impedance spectra of the thickness, thickness shear, length and length thickness extensional modes were analyzed at the fundamental and second resonances using Smits' method while the analysis of the radial mode resonators was accomplished using a method that we had developed earlier. Using the results from the above analysis we have determined the 10 independent complex constants (S₁₁^E, S₁₂^E, S₁₃^E, S₃₃^E, S₅₅^E, d₁₃, d₃₃, d₁₅, (epsilon) ₁₁^T, and (epsilon) ₃₃^T) that define the reduced matrix for a C_(infinity ) piezoelectric material at the fundamental resonance frequency of the resonator. The use of complex material constants to account for the loss of linear systems is discussed and the relationship between the mechanical Q, dielectric tan(delta) , piezoelectric loss and the complex material constants are presented. The dispersion in the martial constants, except for S₁₂^E and S₁₃^E, was studied by evaluating the constants at the second resonance. The source of the major components of the dispersion has been determined to be due to external effects such as mode coupling for high impedance resonators and electrode sheet resistance for low impedance resonators. The size of loss components and the dispersion are shown to be significant and it is suggested that ignoring these effects will reduce the accuracy and predictive capabilities of transducer models.

The Pb(Zn_1/3Nb_2/3)O₃ (PZN)/PbTiO₃ (PT) solid solution has been grown in single crystal form. The dielectric and piezoelectric properties have been determined over a wide range of compositions. Longitudinal coupling constants in this system can be maintained at near 90% for a wide range of relative permittivities, allowing a 'designer dielectric' approach to ultrasonic transducer design. The piezoelectric transducer model developed by Kimholtz, Leedom and Matthaei (KLM) was employed to first optimize transducer design points, and then to study the behavior of these materials as operational transducers. Two types of transducers were modeled and contrasted to conventional materials, a 50 MHz single element designed for ultrasound backscatter microscopy and a 5 MHz phased array element. These two transducer designs are representative of the wide range of properties available in this system by carefully choosing a composition. Extremely high piezoelectric coupling coefficients (k₃₃ greater than 94%) and a range of dielectric constants (3000 - 5000) have been observed in these systems on the rhombohedral side of the morphotropic phase boundary (MPB). Relatively low dielectric constants (approximately 1000) and high thickness mode coupling (kt greater than 63%) were observed as typical of tetragonal formulations. A prototype single element transducer at 35 MHz was fabricated from PZN/8% PT and compared, in pulse/echo mode, to a PZT-5H transducer similar design.

Random 2-2 composite transducers are studied using FEM(ANSYS^R). Admittance curves as well as beam patterns in the nearfield are calculated and used to evaluate the performance of random 2-2 composite designs. First, the pressure and the normal velocity distributions at the interface of water and transducer are calculated using ANSYS, then, these pressure and velocity data are used to calculate the beam pattern using Helmholtz integral. Different random configurations are studied and the results are discussed.

Higher frequency ultrasound is rapidly becoming an important tool for dermatologic and ophthalmologic imaging. This brings about a need for improvement in single element transducers operating in the frequency range between 40 MHz and 100 MHz. Several piezoelectric materials may yield improved performance over common lead zirconate titanate (PZT) transducers. This study investigated several different materials incorporated into single element transducers. A static ultrasonic backscatter microscope (UBM) was constructed in the laboratory. This system allowed for a comparative testing of the imaging performance of various transducers. B-mode scans made by individual transducers show differences in image resolution. Clinically, these differences may be important to allow finer detail to be observed in a structure. Not only does this work show differences between transducers constructed from various materials, but it does so in an application-based environment. Previously, only a limited number of materials were used in such a system. This study showed results from several materials that had not been demonstrated before.

The acousto-optical schlieren method for the measurement of ultrasonic transducer field characteristics has been established as a fast way to veil the beam pattern and directivity of such devices. At the Pennsylvania State University, an Optison schlieren System made by Intec Research is used to analyze transducer beam profiles in both continuous wave and pulsed modes. The schlieren system works by suspending a transducer in a tank of water and sending a pulsed beam of coherent monochromatic light transverse to the beam path. The light received through the tank is focused on a CCD camera. The pressure of the water in the beam path changes the optical index of refraction of the water which in turn results in an image caused by Raman-Nath scattering which can be seen by the camera. Minor fluctuations in the temperature of the water, small bubbles, and particles present in the water (even after filtering the water) can cause noise in the image seen by the camera. In this paper the results of using multiple frame filtering techniques and particle perturbation are analyzed for delivering improved images. This allows better dynamic range and better spatial resolution of actual beam information. Since multiple frames are used, no blurring or other false artifacts are introduce to the images or the reconstructed beam information.

In order to improve the quality of ultrasonic images, it is necessary to narrow the width of the ultrasonic beam produced by the imaging transducer. At the same time, the side lobes of the beam should be suppressed in order to prevent their interference with the main imaging lobe of the beam. Apodization can be used to narrow beam width of an ultrasonic imaging array. While this effect has been produced through element amplitude weighting in pulsed mode linear arrays, a simpler method of apodization can be achieved simply by fabricating elements into other geometries. Currently, small rectangular elements are the standard shape used in single-dimensional ultrasonic imaging arrays. By fabricating array elements into other shapes, apodization is placed on the array. Various apodization schemes have been modeled using a computer simulation. Using the model results, 6 MHz single element transducers were constructed from lead zirconate titanate (PZT) in various geometries. All of these elements had equal surface area. Each of these elements were diced using a common silicon wafer dicing saw as is the preferred industry method for producing ultrasonic arrays. Beam patterns from these transducers were compared to the modeled results.

A new method of blood-flow measurement requires construction of transducers that act like diffraction gratings, producing beams at variable angles as a function of the signals driving them. A new transducer structure that eases fabrication and operation of such transducers is presented and its operation experimentally verified. This transducer is called a higher-order diffracting-grating transducer because, like a higher-order optical diffraction grating, the periodicity of its elements is based on a multiple number of wavelengths. The theory and experimental results for this new transducer structure are shown to agree.