The basics of ultrasonic transducer array design in the frequency range useful for medical imaging are discussed. Performance parameters of importance in transducer design are considered, including sensitivity, coupling constant, band width, frequency downshift, pulse duration, beam focusing properties, and electrical matching. 2D and 3D effects must also be taken into account. The advantages of computer modeling in 3D with finite element analysis code are highlighted. The principles of multi-element array transducers useful for 2D real-time scanning are reviewed. 2D arrays provide the opportunity of focusing the elevation beam dimension in the short axis, or make possible full 3D scanning volumes.
Computer simulation has been used widely in medical ultrasonic transducer designs. One could now construct and test an transducer entirely ion a virtual basis. Such a virtual design and testing procedure not only can save us time and money, but also provide better understanding on design failures and allow us to modify designs more efficiently and economically. This paper is intended to give a brief review on virtual transducer design procedure using a few examples and also list the necessary precautions while using such advanced design tools.
With the wide acceptance of ultrasound medical imaging as the non-invasive diagnostic modality of choice, sonography equipment must offer the tools to complete the diagnosis, including multi-frequency operation for difficult-to-image patients. The trade-off between greater depth of penetration at low frequency for large organs and the improved detail resolution at high frequency is an essential capability that necessitates wideband transducer design and matching system hardware. This paper presents a phased array transducer design with variable ceramic thickness in the elevation direction. The design offers tow major contributions: first, -6dB round trip fractional bandwidth is increased by as much as 120 percent. This is done by controlling the thickness of the crystal from the middle to the outer edge. Since each sampling point in the crystal resonates freely at half wavelength in its fundamental mode, extended bandwidth is achieved for the single element in the phased array. This method has considerable advantage over the usual methods, such as backing the transducer with a matched lossy material. The drawback to backing the transducer is that the acoustic power consumed by the backing represents a severe insertion loss, especially if optimum bandwidth is desired. The second contribution of this design is the use of software to control the elevation slice thickness with axial symmetry around the 2D imaging plane. This is done by controlling the excitation frequency on transmit, and filtering on receive, thereby controlling the transmit and receive apertures independently during imaging. Compared to the standard elevation sampled 1.5D or 2D arrays with an increased number of hardware system channels and extensive cable wires needed, the new design offers simplicity and cost effectiveness. This represents a key development, especially with the advent of second harmonic imaging, both from a point of view of bandwidth requirement and slice thickness on receive. This paper also discusses other advantages of the design, presents experimental and simulation results, and shows a Schlieren video segment of the performance versus standard uniform thickness.
Practical high power phased arrays have been developed and tested in our laboratory. The results show that they offer significant advantages over conventional single focused ultrasound transducers for noninvasive ultrasound surgery. These advantages include but are not limited to: compensation of phase distortion induced by overlying tissues, potential for inducing optimal energy delivery patterns, and generation of large focal spots for tumor coagulation. This paper will describe some of our results with the therapy phased array.
Ultrasonic imaging has been suggested for guidance of high intensity focused ultrasound therapy. This is typically implemented using two different ultrasonic transducer systems. However the need for two transducers may pose practical difficulties such as alignment and different coordinate systems. In this paper we investigate the possibility of using the same physical transducer array for performing both therapy and imaging. A spherically shaped 1D 64-element high intensity focused ultrasound transducer capable of operating in therapeutic and imaging modes was designed and fabricated. In vitro experiments were conducted to show that this transducer is capable of creating well defined lesions 30-50 mm deep into bovine muscle samples. Furthermore, an experimental pulse-echo system was designed to collect full synthetic aperture data using this transducer. Images of multiple-wire and speckle-generating phantoms are shown to illustrate the imaging capability of this transducer. Although the image quality achieved with this array is inferior to that obtained by conventional diagnostic imaging transducers, it is sufficiently high to produce image features suitable for guidance.
This report describes ultrasonic transducers that are designed to expedite thermal-necrosis treatment of tumors, particularly in the eye. The spherical-cap transducers employ pairs of parallel strip-electrodes to generate focal-zone beam patterns that are narrow in one direction and exhibit a number ofprominent lobes over a larger width in the orthogonal direction. Diffraction analysis and thermal modelling are employed to derive information for designing such asymmetric beams and producing continuous, asymmetric thermal lesions within tumors. Compared with typical ellipsoidal lesions, the "chicklet" shaped lesions produced by these beams permit a larger tissue volume to be treated as the transducer is scanned across a tumor. This fact can reduce the number ofscans required to treat entire tumors and may significantly reduce overall clinical treatment times. Keywords: ultrasonic therapy, ultrasonic transducers, thermal modelling, thermal necrosis, high-intensity focused ultrasound
Finite element modeling is being adopted in the design of ultrasonic transducers and imaging arrays. Impetus is accelerated product design cycles and the need to push the technology. Existing designs are being optimized and new concepts are being explored. This recent acceptance follows the convergence of improvements on many fronts: necessary computer resources are more accessible, lean, specialized algorithms replacing general-purpose approaches, and better material characterization The basics of the finite element method (REM) for the coupled piezoelectric-acoustic problem are reviewed. We contrast different FEM formulations and discuss the implications of each: time-domain versus frequency domain, implicit versus explicit algorithms, linear versus nonlinear. Beyond discussions of the theoretical underpinnings of numerical methods, the paper also examines other modeling ingredients such as discretization, material attenuation, boundary conditions, farfield extrapolation, and electric circuits. Particular emphasis is placed on material characterization, and this is discussed through an actual "modelbuild-test" validation sequence, undertaken recently. Some applications are also discussed. Keywords: Arrays, Attenuation, Finite Element Method, Imaging, Piezoelectric, Transducer, Ultrasound
The development of a relatively rapid, but accurate, technique for charactensing pulsed fields from array transducers is reported. The approach enables direct confirmation of the effectiveness of many single element Iarray design, construction and activation procedures. The techniques rests on the employment of a PVDF hydrophone of novel design. Whereas conventional approaches derive field characteristics from point measurements, the new hydrophone allows a direct measurement of the fields 'directivity spectrum' — which efficiently generalises the angular spectrum approach to wideband pulses. The directivity spectrum is shown to encapsulate significant features of both near and far field output characteristics, as well as tightness of focus, even though all measurements are conducted at any convenient distance from the transducer face. The new method is demonstrated in the context of measurements of the fields from typical medical ultrasound transducers. The following field and transducer characteristics are shown : directivity, acoustic axis direction, effective transducer/field coherence, tightness of focus, effective radiating area, effective apodisation, and element uniformity. The relative simplicity of the technique is not compromised when measuring angle-emission characteristics. The theoretical basis for the new field measurement technique is presented, and its advantages over the more usual angular spectrum approach with point measurements, are also discussed. Keywords: Ultrasound; Transducer characterisation; Field characterisation; Directivity spectrum; Large aperture hydrophone.
This paper reports a combined finite element and experimental study of composite and 1D array transducers. The main properties characterized are the electrical impedance and beam pattern. In order to calculate the beam pattern of a transducer immersed in water, the pressure distribution and the normal velocity at the interface of water and transducer were calculated by using ANSYS. These results were then input into the Helmholtz integral to calculate the beampattern. The impedance curve was obtained by performing harmonic analysis using ANSYS. Related experiments were also conducted to verify these calculated results, good agreement was achieved between the experiments and simulations. Because of the similarity between the structures of 2-2 composite and 1D array, the same FEA modeling procedure for 2-2 composites was extended to study a 1D array. Particularly, the cross talk level and directivity of each element in an array were studied.
An important measurement for any new material, for either science or engineering applications, is the determination of the elastic constants. Some exotic new materials, particularly in the crystalline form, may at first be available only in very small samples, and it is challenging to measure such small samples. Recently a new technique, resonant ultrasonic spectroscopy (RUS) has been developed for measuring the elastic constants of very small, as well as large, solid materials. The RUS technique involves lightly contacting a sample with two transducers and using a swept continuous wave excitation to determine a set of resonance frequencies for the sample. A computer calculation is then sued to determine all of the elastic constants from the resonance frequencies. In the past, the technique was hindered by the time required for the computer calculation, but now common desk-top computers can perform the calculation in minutes. For small or large samples, the technique has an advantage in that it is not necessary to bond transducers to the sample; the sample may simply be placed in a holder, and the elastic constants determined in a matter of minutes.
An ultrasonic transducer usually consists of a piezoelectric element, an acoustic lens, matching layers and a backing materials. The foremost important issue in an ultrasonic transducer designing is the choice of the right materials. Therefore, determination of the bulk modulus B and the shear modulus (mu) in a broad megahertz frequency range for those materials is essential.
Resonance technique for determining the physical properties of piezoelectric material can be difficult to implement for some low symmetry systems. Inconstancy may be introduced because several samples are needed and the degree of poling depends on sample geometry. The ultrasonic method on the other hand, allows the determination of a complete set of elastic, piezoelectric, dielectric constants for materials of certain symmetries. However, some of these independent constants can not be directly measured from the phase velocities of pure modes, they need to be derived by solving a complicated coupled Christoffel equation, some relatively large errors may be introduced into these derivations. We found that if an additional length-longitudinal vibrator is used to assist the measurements, the computation would be greatly simplified and the final results became more accurate. As an example, the elastic, piezoelectric, dielectric constants and electromechanical coupling factors have been determined for a PZT-5H piezoelectric ceramic by using the combined method. This method can also be extended to piezoelectric crystals with symmetry point group of 4mm and 3mm.
Despite its beginning as an imaging tool, the scanning acoustic microscope has become the tool of choice for obtaining quantitative information about the mechanical properties of a sample when microscopic spatial resolution is required. Essential points in the experimental use of the scanning acoustic microscope will be reviewed with emphasis placed on issues, materials and developments of concern to a medical transducer designer. A simple methodology will be discussed for measuring the elastic moduli of isotropic materials along with expressions to determine the error propagation of the technique.
Piezoelectric material lie at the heart of ultrasonic transducers. Recent advances in materials development include submicron piezoelectric ceramics (PZT) which lead to improvements in feature size, i.e. aspect ratio, element width, etc., for linear arrays and high frequency transducers. In contrast to fine grain ceramics, single crystal materials based on Relaxor-PT ferroelectrics offer electromechanical coupling coefficients > 90 percent with a range of dielectric permittivity allowing flexibility in transducer engineering in regard to electrical impedance matching. Using KLM modeling, very high bandwidth performance > 120 percent is projected. Specific examples of high frequency 1-3 composites and 1D linear array transducers fabricated from new piezoelectric materials, including sol-gel derived PZT fibers, are presented.
Over the last decade, several methods were utilized to develop novel piezoelectric ceramic/polymer composites for transducer applications. Solid freeform fabrication (SFF) is one of the methods that have been emphasized recently. SFF techniques have been used to fabricate polymer, metal or ceramic structures on a fixtureless platform, directly from a computer aided design file. During design verification or the product development stage, SFF techniques offer great flexibility to manufacture prototypes with various shapes, sizes and functionality. Several SFF techniques, fused deposition modeling, fused deposition of ceramics, and sanders prototyping were used to fabricate a variety of novel piezoelectric ceramic and ceramic/polymer composite transducers at Rutgers University. The composites were processed either by a direct, indirect or soft tooling route. A variety of novel composite structures, including annular ring, hexagonal pattern with octagonal rods, and oriented fibers, have been made using the flexibility provided by the above processes. Volume fraction gradients have been incorporated into some of these designs with the ceramic volume fraction decreasing form the center towards the edges, following wither a linear, exponential or gaussian distribution. Novel radial composites being pursued in our labs, including fabrication of fine scale, large area, flexible PZT composites using thin PZT fibers, and the preparation of a new PNN-PZ-PT composition for ultrasonic transducer applications. The design, fabrication and electromechanical properties of these structures are also discussed in this paper.
Electrostrictive materials, such as the ceramic PMN/PT/La, operating above Tmax with a DC bias field behave as a piezoelectric ceramic materials with CINF symmetry. The effective piezoelectric and electromechanical coupling coefficients are found to be linear as a function of the DC bias field up to about 0.5MV/m, while the elastic constant and the dielectric constant are found to have a quadratic dependence on the DC bias field. Above 0.5 MV/m the piezoelectric and the electromechanical coupling constants begin to saturate due to higher 4th order electrostriction. In essence these materials behave as tunable piezoelectric materials with the piezoelectric coefficient being directly proportional to the electrostrictive coefficient and the DC bias field up to saturation>. The properties of DC biased resonators of this material are derived from a non-linear theory based on the Taylor's series expansion of the thermodynamic potentials to 3rd and higher order terms in field and stress. The resonance equations for the DC biased length extensional resonator are presented and it is shown that DC biased resonance techniques can be used to measure the electrostrictive and other higher order coefficients at frequencies of interest to the ultrasonics community. The experimental apparatus used to measure these properties will be described and the limitations with regards to isolation of the measurement signal and the DC bias signal will be discussed. We will show that these materials, in conjunction with standard piezoelectric ceramics, offer the transducer design engineer an extra degree of freedom and the feasibility of unique transducer designs that will allow, for example, multiple beam patterns from the same circular/linear array using an adjustable DC bias profile on the array or the possible use of the field dependence of the compliance to fabricate electrically active backing materials. In conclusion we discuss how a better understanding of the macroscopic theory of piezoelectric and electrostrictive materials can benefit the transducer designer.
In ultrasonic medical imaging it is desirable to have the maximum beam sensitivity along the transmission axis. However, the presence of gratin and side lobes greatly affects the transducer performance. It is known that the grating lobes can be reduced by non-uniform spacing of elements in the composite. In spite of this knowledge, it has been found to be difficult to fabricate piezocomposites with complex designs using traditional processing routes. The ceramic element spacing can be varied easily using solid freeform fabrication (SFF) techniques. In this work SFF techniques, including Sanders Prototyping (SP) and fused deposition of ceramics were used to make many novel piezoelectric ceramic/polymer composite transducers. A variety of 2-2 PZT-5H/spurr epoxy volume fraction gradient samples have been fabricated. Many mathematical functions, including linear, gaussian and exponential gradients were designed using Pro Engineer software. Novel oriented composites have also been fabricate where the ceramic elements are at an angle to the thickness direction. The piezoelectric properties are found to change with the orientation of piezoelectric rods. The optimum properties have been observed at an orientation of 30 degrees to the vertical where the total contribution from the d33, d31 and d15 components could be the highest. These oriented composites may be used to focus the acoustic beam at a point by varying the orientation angle of the rods within the same composite. The design, fabrication and electromechanical properties of these composites are discussed in this paper.
Basic digital beamforming architectures will be reviewed. Architectures will be analyzed for spectral characterization and hardware considerations which consist of A/D converter requirements, data storage strategies, and computational complexity. The goals of parallel processing will be illustrated along with the basic forms of parallel computation.Granularity issues which measure the ratio of the amount of computation done in a parallel task to the amount of inter-processor communication will be illustrated. Software systems will be described that allow the beamforming architectures to be modeled and tested for data throughput requirements as well as computational loading with respect to application specific concurrent architectures. One specific structure, MPI, will be described and used to illustrate how parallel processing architectures may be synthesized and tested.
High-performance and efficient beamforming circuitry is very important in large channel count clinical ultrasound systems. Current state-of-the-art digital systems using multi-bit analog to digital converters (A/Ds) have matured to provide exquisite image quality with moderate levels of integration. A simplified oversampling beamforming architecture has been proposed that may a low integration of delta-sigma A/Ds onto the same chip as digital delay and processing circuitry to form a monolithic ultrasound beamformer. Such a beamformer may enable low-power handheld scanners for high-end systems with very large channel count arrays. This paper presents an oversampling beamformer architecture that generates high-quality images using very simple; digitization, delay, and summing circuits. Additional performance may be obtained with this oversampled system for narrow bandwidth excitations by mixing the RF signal down in frequency to a range where the electronic signal to nose ratio of the delta-sigma A/D is optimized. An oversampled transmit beamformer uses the same delay circuits as receive and eliminates the need for separate transmit function generators.
High frequency polymer transducers have been used in a variety of medical imaging applications since they were first introduced by Sherar and Foster in the late 1980s. The transducers are intrinsically broadband and the flexibility of the polymer material makes fabrication relatively easy. Unfortunately, piezoelectric polymer materials have a low dielectric constant. Unless a large aperture is used, the electrical impedance of the transducer will be high, and the receiver sensitivity will be poor. This problem can be avoided by placing a high impedance pre-amplifier inside the transducer housing. Placing the pre-amplifier close to the transducer is important to avoid standing waves between the high output impedance of the transducer and the high input impedance of the pre-amplifier. We have recently developed a process for fabricating high frequency spherically shaped polymer transducers in which an integrated circuit die is mounted just beneath the surface of the transducer. In this paper we describe a theoretical and experimental analysis of the noise performance of these devices. The signal-to-noise ratio at the output of the pre-amplifier is estimated by combining a simple noise model for the amplifier with a KLM model of the transducer. This analysis provides a useful way of evaluating different transducer/pre-amplifier combinations. Excellent agreement between the model predictions and experimental results proves the validity of this approach.
In the last five years we have been actively developing capacitive micromachined ultrasonic transducers (cMUT) since they have potential advantages over piezoelectric transducers, such as ease of fabrication in single elements and arrays, broad bandwidth and high efficiency. We report on research efforts in the theoretical understanding of cMUT with an improved electrical equivalent circuit model as well as its actual implementation through microfabrication. First, we present a process sequence that has allowed us to make reproducible devices with sealed membranes. The impact of electrode metalization on the impedance, bandwidth and efficiency of the devices will be discussed, and experimental results will be compared to theoretical models. Our study in the paper indicates that the best cMUT performance can be achieved with the appropriate fabrication process under certain constraints. Our most recent devices have been designed to have an input impedance with areal part of 50 Ohms at a frequency around 5 Mhz. A through transmission measurement gave a dynamic range of better than 100 dB while operating in the frequency range of 1-10 MHz. Operated without tuning, the devices are capable of operation from dc to 10s of MHz which is achieved because the devices are not resonant. In summary, we will present a novel technology capable of delivering surface micromachined ultrasonic transducers that are efficient, and easy to make single element and multiple elements array transducers. These devices can also be integrated on chip with transmitter and receiver electronics.
Single crystal Pb(Zn113Nb2i3)03 (PZN) I PbTiO3 (PT) solid solutions have been investigated as the active materials in medical diagnostic transducers. The compositionally controlled properties of this crystal system allow the application of this material to transducers which span the frequency range of modern diagnostic transducers. Phased array elements operating at 5MHz were constructed from rhombohedral PZN I 4.5% PT single crystals and compared to PZT 5-H elements. The single crystal element displayed similar sensitivity over a broader bandwidth response. Transducer construction techniques employed differ from standard ceramic arrays. Crystallographic alignment was found to be a significant contributor to piezoelectric performance. The crystal structure and phase diagram play a major role in processing steps requiring heating such as poling and electroding. Emphasizing the versatility of the PZN/PT single crystals, a 20 MHz single element was also constructed, using tetragonal PZN /12% PT. Its response was compared to theoretical pulseechoes generated via the KLM model.
The development of single crystal relaxor-PT piezoelectrics is an exciting advance in ultrasound transducer technology. The high electromechanical coupling coefficients and variable dielectric constant could be used to significantly enhance bandwidth and sensitivity of array transducers. In this study 1-3 composites of single crystal material were engineered and applied to an array design. Both predicted and actual performances are reported and compared to array designs using PZT-5H based 1-3 composite material. In order to take advantage of the performance enhancement of single crystal materials a 1-3 composite connectivity was selected. Two techniques were used to tile together small pieces of crystal into a larger composite plate suitable for array applications. A coupling coefficient of .81 and an acoustic impedance of 15.6 Mrayls were obtained using a 58% volume fraction of single crystal. A comparable PZT composite displayed a coupling coefficient of .66 and an acoustic impedance of 1 7.0 Mrayls. A sufficiently fine spatial scale resulted in lateral resonances being well above the thickness mode resonance for both materials. One-dimensional modeling using the Redwood equivalant circuit in PSpice was used to investigate matching and backing. An optimization scheme resulted in a modeled bandwidth of 134% and a -6 dB pulse length of only one cycle using a single front matching layer and a castable backing. Eight element arrays of single crystal and PZT-5H composite were constructed to verify the theoretical results.
Forty-channel phased array probes using a O.9lPb(Zn113Nb213)03-O.O9PbTiO3 (PZN-PT 91/9) single crystal have been fabricated for greater sensitivity and broader bandwidth properties. The largest crystal by the self-flux method had the dimensions of about 43 X 42 X 40 mm. The fabricated probes have a center frequency of 3.5MHz and an aperture of about 8.5 X 8.5 mm. All transducer elements were successfully cut by mean value the fabrication process, and the dispersion of echo signals was within 20% of the average. Echo amplitudes of these PZN-PT 9 1/9 single-crystal probes are as much as 5 to 8 dB higher than those of conventional PZT ceramic probes. Moreover, the bandwidth ofthe PZN-PT 91/9 probes is 20 to 25 points broader than that of conventional PZT ceramic probes. This confirms that PZN-PT 91/9 single-crystal probe provides great improvements in the sensitivity and bandwidth of phased array probes.
Ceramic ultrasonic transducers with a frequency response <50MHz using lead zirconate titanate (PZT) layers in the 5-4Otm range have been difficult to achieve by bulk or thin film techniques. Advances in the production of small sized PZT ceramic powder allow the development ofthinner ceramic layers for this application. Sol gel composite thin film technology also provides a new technique for producing ceramic coatings of a thickness that successfully bridges the gap between traditional thin film and bulk techniques. So! gel composite PZT layers of 5-7Om have been coated on substrates (aluminum, platinized silicon, stainless steel, .. .) that can withstand the thermal processing of the ceramic. The thickness mode response of a thin piezoelectric layer supported by a thick substrate has been modeled from first principles using complex material constants. The LevenbergMarquart non linear regression technique has been used to extract the thickness mode elastic stiffness, dielectric constant and piezoelectric constant of the PZT, and the elastic stiffness of the substrate from the layered structure. This non-destructive technique allows for a reliable assessment of the quality of a coating prior to the fabrication of a transducer. The elastic stiffness of the substrate is not lossy enough for the required broadband response of an imaging transducer. However, aluminum can be preferentially etched, releasing the ceramic coating. Therefore it is possible to transfer the PZT film to a more suitable backing material. A processing sequence for single element PZT transducers in the frequency range of 50-200MHz has been developed. Characterization of transducers has been performed using pulse-echo techniques and by creating real time B-scan images of agar phantoms and biological tissue. Methods for patterning the PZT composite coatings are being developed with the intent to fabricate a linear array in the 40-60 MHz frequency range. Due to the very fine patterning and high concentration of cuts required for a high aspect ratio linear array, the limits of conventional etching techniques are surpassed. Laser micromachining using a frequency doubled Nd:YAG and a KrF excimer laser have the ability to pattern the array structure. In both cases, laser cuts <1Om wide have been achieved.