The amplitude of a sound wave equals the degree of pressure change from equilibrium. Larger amplitudes lead to denser compressions, which create sound with higher intensities. The wavelength (λ) of an ultrasound is the distance (in millimeter or micrometer) between compressions or rarefactions. Wavelength can also be measured between two repeating points on a sinusoidal wave of pressure amplitude (successive wave crests) (Figure 7.2). Frequency (f) is the number of times the wave oscillates during a cycle per second (sec). Frequency also refers to the number of wavelengths passing a specific point per second. The unit of measurement is in hertz (Hz), given that 1 Hz equals one oscillation per second. Sound waves with frequencies between 15 and 20 000 Hz (20 kHz) compose the audible acoustic spectrum. Sound waves with frequencies below 15 Hz are called infrasound, whereas those with frequencies above 20 kHz are called ultrasound. Medical ultrasound typically uses frequencies between 2 and 15 MHz (15 000 000 Hz); however, specialized applications can use frequencies as high as 50 MHz. The period of a sound wave represents the time duration of one wave cycle and is equal to 1/f. The speed of sound, which is the distance traveled by the wave per unit of time, is equal to the wavelength divided by the period. Because period and frequency are inversely proportional (period = 1/f), the relationship between velocity, wavelength, and frequency can be represented as

v = λf

where v is the velocity of sound (m/s), λ is the wavelength (m), and f is the frequency (cycles/s).

The velocity of sound depends on the medium in which the wave is propagating through, and this is highly variable between different materials. The speed of the wave depends on two factors—the bulk modulus (B) and the density (ρ) of the medium. The bulk modulus is a measurement of the stiffness of a medium and how well it resists compression. Relating speed of sound in a medium with bulk modulus and density can be represented as

where the SI units are as follows: v (m/s), B (kg/ms²), and ρ (kg/m³). A highly compressible medium (eg, air) has a low speed of sound, whereas a less compressible medium (eg, bone) has a higher speed of sound. The density and speed of sound for relevant materials in medical imaging are listed in Table 7.1. The varying speeds of sound modulated at the boundaries of tissues provide the fundamental reason for contrast in ultrasound imaging.

Within a given tissue medium, ultrasound frequency is independent of changes in sound speed. Hence, the wavelength is determined by frequency and the propagation medium. Wavelength depends on the compressibility of the material the sound wave is propagating through. An ultrasound transducer with a frequency of 2 MHz will result in a wavelength of 0.77 mm in soft tissue, 0.17 in air, and 1.7 mm in bone.

Given that the speed of sound in soft tissue is approximately 1540 m/s (equal to 1.54 mm/µs), the wavelength (in mm) within soft tissues can be calculated as follows:

λ = v/f = 1.54 mm/µs/f (MHz) = 1.54 mm/10^-6 s/f (10⁶/s) = 1.54 mm/f (MHz)

The wavelength changes at the boundaries of two different media due to the change in sound speed.

The wavelength and frequency determine the spatial resolution of the image and the attenuation of the ultrasound beam energy. An ultrasound beam with high frequency, and hence a small wavelength, provides superior resolution and image detail than does a low-frequency beam. However, a high-frequency beam is limited in the depth it can penetrate. Ultrasound beams with a low frequency provide lower resolution and image detail; however, they can penetrate to a much deeper depth. Consequently, the choice between using a higher or lower frequency ultrasound beam depends on the particular clinical application. Lower frequency waves (3.5-5 MHz) are used for thicker body parts (eg, abdominal imaging), whereas higher frequency waves (7.5-10 MHz) are used for smaller body parts or when the target is close to the skin (eg, thyroid, breast).

Ultrasound machines contain acoustic transmitters, which produce different independent sound beams. These unique beams can interact with one another, leading to constructive or destructive wave inference or complex interference patterns (Figure 7.3). Interactions of two ultrasound waves that have the same frequency and phase result in a beam with increased amplitude (constructive interference). Interactions of two ultrasound waves that are 180° out of phase with one another result in an output wave with a lower amplitude (destructive interference). Interactions between ultrasound waves of slightly different frequencies result in waveforms of varying amplitudes (complex interference). The two most important factors determining the degree of interference include the phase of the sound wave and the amplitude of the interacting beams.

Acoustic energy causes displacement of particles and variations in local pressure as it propagates through a medium. The variations in pressure are referred to as pressure amplitude (P). Pressure amplitude is equal to the difference between the peak maximum value or peak minimum value and the average pressure within a medium. The peak maximum (positive pressure amplitude) and the peak negative (negative pressure amplitude) values are equal when a symmetrical waveform is present. However, the compressional amplitude typically exceeds the rarefactional amplitude to a large degree in most diagnostic ultrasound applications. The SI unit of pressure is the pascal (Pa), which equals one newton per
square meter (N/m²). Most diagnostic ultrasound beams deliver peak pressure levels of approximately 1 MPa (megapascal), which exceeds the earth’s atmospheric pressure 10-fold.

Power refers to energy per unit time. Intensity (I) is the amount of power per unit area. Intensity is proportional to the pressure amplitude squared (I ∝ P²). Therefore, doubling the pressure amplitude leads to a quadruple increase in intensity. The units of intensity in medical diagnostic ultrasound are milliwatts per centimeter square. Relative intensity is defined in units of the decibel (dB) and calculated using the following equation:

Relative intensity (dB) = 10 log (I₂/I₁)

Alternatively, relative pressure (also defined in units of dB) can be calculated as follows:

Relative pressure (dB) = 20 log (P₂/P₁)

In these equations, I₁ and I₂ refer to intensity levels and P₁ and P₂ refer to pressure amplitude values. The base 10 logarithm (“log”) is used to compress the large potential variance in intensity ratios between the incident pulse and the returning echoes. Every 10 unit change in the dB scale represents a change in intensity to the order of 10 times the magnitude, which means a 20 dB change equates to a change in 100 times the magnitude. An intensity value of 60 dB is equivalent to an incident intensity that is 10⁶ (ie, 1 million) times that of the returning echo. The dB levels are positive when the incident ultrasound intensity is greater than that of the returning echo (ie, intensity ratio > 1). Alternatively, the dB levels are negative when the incident ultrasound intensity is less than that of the returning echo (ie, intensity ratio < 1). Furthermore, a decrease of 3 dB equals a 50% loss of signal intensity. The thickness of tissue required to reduce the intensity of the ultrasound by 3 dB is referred to as the half-value thickness (HVT).

Measurements of distance, area, and volume during ultrasound examinations are possible because the speed of sound in soft tissues is nearly constant (1540 ± 15 m/s). Calibration of the instrument is readily determined depending on the round-trip time of the pulse and the echo. Careful selection of the reference points is necessary to ensure measurement accuracy. Selecting points along the direction of the ultrasound beam typically yields measurements more reliable than points measured in the lateral plane given the improved spatial resolution. Furthermore, measuring the distance between the leading edges of two objects lying in the axis of the beam is reliable because these points are less affected by variations in echo signal amplitude. The circumference of a circular object can be calculated using the measured radius (r) or diameter (d) with 2πr or πd. The area of an object or region of interest is calculated using distance measurement and by assuming a specific geometric shape. The area of a region of interest in a single image plane can be extended to 3D volume measurements by estimating the slice thickness (elevational resolution).

The acoustic impedance of a material or tissue describes its stiffness and flexibility to the ultrasound beam. In essence, it describes the degree to which particles within a medium will react and change in response to mechanical vibrations. The acoustic impedance (Z) of a material is the product of the density (ρ) and the velocity of the sound wave (v) within the material:

Z = ρ × v

with the units of kg/m³ for density and m/s for velocity. The units of acoustic impedance (kg/m²s) are expressed as the SI units of rayls (1 kg/m²s = 1 rayl). A tissue medium’s ability to resist change because of mechanical stimulation increases as the density of the medium increases. Large differences in acoustic impedance between two adjacent media lead to a large reflection in energy, whereas small differences allow for continued propagation of energy with little reflection. The lung, for example, has low acoustic impedance because of pockets of air within alveoli, which scatter the sound wave. The adjacent soft tissues have a relatively large acoustic impedance. This results in near-complete reflection of the incident sound waves when propagating to the soft tissue-lung interface. Minor reflection occurs between tissues that have similar acoustic impedances. Of tissues in the body, bone has the highest acoustic impedance (7.8 × 10⁶ rayls), whereas air has the lowest (0.0004 × 10⁶ rayls). Table 7.2 lists the relative acoustic impedances of tissues in the body relevant to diagnostic ultrasound imaging. Piezoelectric crystals have a very high acoustic impedance, much greater than that of bone.

As mentioned earlier, ultrasound energy that is perpendicular to a tissue boundary results in a portion returning directly to the source (as echoes) and the other portion being transmitted directly through the medium. Sound waves experience an 180° phase shift in pressure amplitude as they propagate from a medium of lower acoustic impedance into a medium of higher acoustic impedance.

The difference in the acoustic impedances of two tissues at a boundary results in reflection of the ultrasound beam. Reflection, which is also known as backscatter, is the principle tissue interaction of sound waves to produce an ultrasound image. Approximately 1% of the intensity of an ultrasound is reflected at a typical muscle-fat interface, resulting in nearly 99% propagation of the beam through the boundary. The energy reflected back from an interface is known as an echo. Nearly 100% of the ultrasound energy is reflected at muscle-air interfaces, which essentially leads to nonvisualization of the deeper anatomy. This explains the need to apply acoustic coupling gel in the transducer-skin interface during an examination. Any air pockets between the transducer and the skin lead to degradation of the image because of reflective interactions. Values of reflected intensities for several interfaces commonly encountered in medical diagnostic ultrasound are listed in Table 7.3. The posterior shadowing caused by bowel gas often leads to nonvisualization of certain organs in abdominal imaging, in particular the pancreas and appendix. A large fraction of the incident beam energy is reflected at bone-tissue interfaces given the significant difference between acoustic impedances. It is essentially impossible to image through bone or air given the extensive shadowing, leading to a void of echoes.

Scatter is the redirection of acoustic energy in numerous directions, which results in the production of a weak signal. The interactions between sound waves and individual molecules, particularly at small interfaces or large rough surfaces, will result in scatter. There are two general types of reflector surfaces—nonspecular and specular. A nonspecular surface does not follow the typical rules of reflection and reflects ultrasound echoes in all different directions, a phenomenon known as acoustic scattering (Figure 7.4). This occurs at irregular interfaces possessing wavelengths that are equal to or smaller than the incident ultrasound beam (typically 1 mm or less). Only a small amount of the transmitted energy reaching a nonspecular reflector returns to the transducer. The amplitudes of returning scattered echoes are much lower than those of specularly reflected echoes. However, this usually does not pose an issue because modern ultrasound receivers have a dynamic range capable of utilizing information obtained from a wide range of amplitudes. Scattered echoes originate from small, poorly reflective objects that are less angle dependent (eg, blood cells). Specular reflection occurs at large, smooth surfaces that are regularly shaped (eg, valves). For specular reflection, the angle of incidence (i) is equal to the angle of reflection (r) (Figure 7.5). As the angles i and r decrease, the chance that the reflected beam will return to the transducer and be detected as an echo increases.

The dynamic range indicates how wide a spectrum of echo signals the ultrasound system can accept without distortion. Echo signals below the dynamic range are regarded as noise, whereas signals above the range are regarded as saturated and set to the maximum level. The dynamic range is an adjustable setting in real time. Smaller ranges create a steeper gradient, which provides for more contrast on the display. This is useful in improving the appearance of a low-contrast structure, such as a mass. The downside to increasing the contrast is the production of a less clear, coarse image. Alternatively, larger dynamic ranges generate less contrast, but smoother images.

Several factors influence the amplitude of the ultrasound signal returning from the insonated tissues, including the acoustic impedance differences at the scatterer interfaces, the number of scatterers present, the size of each scatterer, and the frequency of
the ultrasound. Objects described as hyperechoic have a higher scatter amplitude in comparison to the background signal, whereas hypoechoic objects demonstrate a relatively low-scatter amplitude.

Acoustic scattering from nonspecular surfaces increases proportionally with increasing ultrasound frequencies. Most organs of the body give rise to a characteristic scatter “signature” due to the inherent tissue properties (see Figure 7.4). Specular reflection occurs at smooth interfaces that have dimensions larger than the incident ultrasound wavelength. Specular reflection is not affected by changes in frequency. The echoes from specular reflector surfaces are used to create ultrasound images.

Reflection occurs when the incident beam is at an angle perpendicular to the tissue boundary. Interactions occurring at nonperpendicular angles result in refraction of the ultrasound energy. Refraction refers to the change in direction of an ultrasound beam when moving between two tissues that have different speeds of sound (acoustic velocities). The frequency of the sound wave remains constant between the two tissue media; however, the wavelength changes to accommodate the different velocity in the second tissue. The wavelength shortens if the velocity is decreased, and vice versa. The sound pulse is refracted as it enters the second medium as a result of the change in wavelength. The angle of refraction depends on the change in the velocity of sound occurring at the tissue boundary. This is determined by Snell’s law:

sinθ_i/sinθ_t = v₂/v₁

where θ_i is the angle of incidence, θ_t is the transmitted angle, v₁ is the velocity of sound in medium 1, and v₂ is the velocity of sound in medium 2. When the velocity of sound is greater in tissue 2 compared to tissue 1, the angle of transmission is greater than the angle of incidence, and vice versa (Figure 7.6). If the velocity of sound between the two tissue media is the same, then no refraction occurs. Refraction also does not occur with perpendicular incidence. The velocity of sound is relatively low in compressible tissues, such as fluid, lung, and fat, and high in less compressible issues, such as bone.

The ultrasound machine assumes that sound waves propagate in a straight line. Any refraction results in image artifacts. Examples include the misplacement of an object within an image and an artifact known as edge shadowing, which occurs beyond a curvilinear interface. Total refraction is a phenomenon in which v₂ > v₁ and the angle of incidence of the propagating sound beam at a tissue interface is greater than the critical angle. The critical angle, designated as θ_c, is the angle that creates an angle of refraction of 90° when a wave moves from a material with a slower speed of sound into a material with a faster speed of sound. By setting the angle of incidence (θ_t) to 90° and plugging it into Snell’s law, the critical angle can be calculated as

sin θ_c = v₁/v₂

Attenuation describes the energy loss (ie, change in wave amplitude) of an ultrasound beam as it propagates through a tissue medium. This occurs because of several factors, including reflection, refraction, scatter, and absorption. The amplitude decay model is calculated as

A = A₀e^-µz

where A is the reduced amplitude of the sound wave, A₀ is the amplitude of the unattenuated wave, e refers to Napier’s constant (which is rounded to 2.71828), µ is the attenuation coefficient (expressed in decibels per cm: dB/cm), and z is distance traveled. The attenuation coefficient is the loss of intensity (dB) per centimeter that a sound wave travels through a medium. The amount of attenuation depends on the frequency of the ultrasound beam and the properties of the propagating medium. Water, for example, results in the least attenuation of an ultrasound beam. Thus, it serves as a good conductor for deeper structures. A diagnostic scenario exemplifying the good conducting property of simple fluid (eg, water, urine) is the transabdominal portion of a pelvic ultrasound examination (Figure 7.7). The urinary bladder is ideally filled before scanning to allow for a “diagnostic window” to
the uterus. Ultrasound attenuation occurring in a homogeneous tissue is exponential. Tissues and fluids each possess a characteristic attenuation coefficient, and these vary widely (Table 7.4). Another way to calculate the rate of attenuation is as follows:

Attenuation (dB) = µ × f × x

where µ is the attenuation coefficient (in dB/cm at a frequency of 1 MHz), f is the ultrasound frequency (in MHz), and x is the tissue thickness (in cm). The attenuation coefficient for soft tissues (0.3-0.8) can be rounded to approximately 0.5 dB/cm per MHz of frequency (0.5 × [dB/cm]/MHz). Multiplying the ultrasound frequency for a given sound wave with (0.5 × [dB/cm]/MHz) provides the approximate attenuation coefficient in units of dB/cm. This demonstrates that attenuation increases linearly with increasing frequencies. A 2 MHz wave will experience twice the attenuation of a 1 MHz wave per unit distance (in cm), a 5 MHz will suffer 5 times the attenuation of a 1 MHz wave, and so on. It is important to note, however, that the dB scale increases logarithmically. Thus, a linear increase of attenuation is observed with increasing frequencies, whereas the ultrasound beam intensity is accentuated exponentially with distance (Figure 7.8). The HVT refers to the thickness of a particular tissue medium required to attenuate the ultrasound beam intensity by 50%, which results in a 3 dB reduction in intensity. A 3 dB intensity reduction is of special importance because this equates to an absolute reduction in intensity by two. The equation for intensity in decibels is given as follows:

dB = 10 log₁₀ (I/I_o)

where I is intensity and I_o is the reference intensity. The HVT decreases as the ultrasound beam frequency increases. Ultrasound is attenuated to a much greater degree than is audible sound given its high frequency.

Ultrasound intensity is the amount of energy (joules) flowing through a unit cross-sectional area (cm²) per a unit of time (seconds). Joules (J) are related to watts via the following equation: 1 watt (W) = 1 J/s. The intensity is typically expressed as W/cm² (or mW/cm²). The power of an ultrasound beam is a measure of the total energy passing through the whole area of the beam per unit time. In other words, total power is the product of ultrasound intensity and the beam area:

Power = intensity (mW/cm²) × area (cm²)

This equation assumes that intensity is uniform. The product of total power and the amount of time the beam is on gives the total energy.

As previously mentioned, relative sound intensity is measured on a logarithmic scale and expressed as decibels (dB). Keeping in mind the relative intensity equation of 10 log₁₀ (I/I_o), dB values will be positive if the intensity of interest (I) is greater than the relative intensity (I_o), which represents signal amplification. Negative values occur if I is less than I_o, which represents signal attenuation. The levels of intensity in ultrasound imaging are very low, which makes the logarithmic decibel relative intensity scale useful practically. A reduction in intensity to 10% corresponds to −10 dB; a reduction to 1% corresponds to −20 dB; etc.

Absorption is the conversion of acoustic energy to other forms of energy, namely, heat. This energy is transferred to the propagating medium, which “absorbs” it. Absorption of ultrasound energy by the tissue medium is one of the most important determinants of attenuation. Three main factors influence the extent of absorption—beam frequency, the viscosity of the tissue medium, and the relaxation time of the medium. Higher frequencies lead to increased absorption because the particles within the medium travel past each other faster, generating more heat. The absorption of ultrasound in soft tissues demonstrates a proportional increase with higher frequencies. Alternatively, minimal absorption occurs in fluids. Viscosity is a measure of the friction between particles traveling past one another within a tissue medium. Greater frictional forces lead to more heat generation. Thus, ultrasound absorption increases with higher viscosity. Relaxation time refers to the length of time for particles within a medium to revert to their original positions after being displaced by an ultrasound pulse. A longer relaxation time means that displaced particles have a higher probability of encountering the next ultrasound pulse before fully relaxing. The particles may be moving in an opposite direction than the new compression pulse, which results in increased dissipation of energy from the ultrasound beam. Therefore, increases in all three factors—frequency, tissue viscosity, and tissue relaxation time—lead to increases in heat generation, and hence, absorption of the ultrasound beam.

The functional component of the transducer is made of the piezoelectric material. This material exhibits the piezoelectric effect; that is, when mechanical stress is applied, the net polarization of the material changes as a result of the changes in the alignment of the molecular dipoles inside the material (a molecular dipole is a molecule whose center of positive charge is spatially separate from its center of negative charge), leading to electrical charges generated on the surface of the material. The piezoelectric effect is a reversible process; when an external electric field is applied, the material experiences mechanical stress and undergoes physical deformation. Such interesting characteristics of the piezoelectric material have made it useful in both generation and detection of ultrasound.

An imbalance in the charge distribution occurs when an external pressure is applied, which creates a potential difference (voltage) across the element. The voltage, which is measured by surface electrodes, is proportional to the incident mechanical pressure amplitude.

Ultrasound transducers use a synthetic piezoelectric ceramic, most commonly lead-zirconate-titanate (PZT). A multistep process gives the ceramic its piezoelectric characteristics, which begins with molecular synthesis and heating. Next, an external applied voltage orients the internal dipole elements. The orientation is permanently maintained via cooling. Finally, the material is cut and molded into a specific shape. No piezoelectric properties are exhibited by PZT in its natural state. The dipoles align in the material once heated beyond its “Curie temperature” (328°C-365°C) and an external voltage is applied. The dipoles retain their alignment once the material has cooled. No net charge is present on ceramic surfaces at equilibrium. A voltage is produced between the surfaces when compressed. Mechanical deformation occurs when a voltage is applied between electrodes attached by both surfaces (see Figure 7.10).

Resonance transducers used for pulse-echo ultrasound operate in “resonance” mode, in which a voltage of very short duration (˜1 µs) is applied. The piezoelectric material contracts and then subsequently vibrates at a natural resonance frequency (f₀), which is determined by the thickness cut. Ultrasound waves with a wavelength twice the thickness of the piezoelectric material are preferentially emitted. The operating frequency depends on the thickness and speed of sound within the piezoelectric material. The speed of sound within PZT is approximately 4000 m/s. Using this speed of sound, the wavelength can be calculated for a given frequency. As an example, a 5-MHz transducer will have a wavelength of

λ = c/f = 4000 m/s/5 × 10⁶/s

λ = 8 × 10^-4 m = 0.80 mm

The damping block, which is typically composed of tungsten or rubber in an epoxy resin, absorbs the ultrasound energy that is transmitted to the back of the crystal in addition to stray ultrasound signals from the housing unit. The damping material is located behind the piezoelectric element of the transducer. It also serves to dampen the transducer vibration to create an ultrasound pulse that has a short spatial pulse length (SPL). This is necessary to maintain the detail along the beam axis, the axial resolution. The damping block must be composed of a material that has the same acoustic impedance as the crystal. This prevents the unwanted production of echoes from a damping block-crystal interface, which would lead to reverberation artifact when the energy returns to the crystal. The damping block also serves as a mechanical pulse damper by limiting the SPL. The effect of dampening the transducer vibration, which is also known as “ring-down,” is 2-fold. Firstly, the purity of the resonance frequency is lessened. Secondly, a broadband frequency spectrum is introduced. Frequencies higher and lower than the resonance frequency are introduced with ring-down, which increases the bandwidth (range of frequencies). The Q factor represents the bandwidth of the sound propagating from a transducer, where f₀ is the center (resonance) frequency and the bandwidth is the width of the frequency distribution:

Q = f₀/Bandwidth

A transducer that has a narrow bandwidth and a corresponding long SPL is known as a high Q transducer. Those that have a wide bandwidth and a short SPL are known as low Q transducers. Medical imaging applications require a transducer
with a broad bandwidth to achieve high spatial resolution along the direction of the beam. Figure 7.11 shows the examples of high and low Q transducers and their respective relationships with SPL.

Differences in the acoustic impedance of the transducer crystal and the surface of the patient’s skin can prohibit adequate transmission of echoes into the patient. A layer of material (matching layer) placed on the front surface of the transducer increases the efficiency of energy transmission into and out of the patient. A matching layer has an intermediate impedance value, which is between that of the transducer and soft tissue. This serves to minimize the acoustic difference between the transducer and the patient. The ideal thickness of a matching layer is one-fourth of the wavelength of sound, which is determined from the center operating frequency of the transducer and the speed of sound within the matching layer. This is known as quarter-wave matching. The matching layer, in combination with acoustic coupling gel, improves the quality of the images obtained.

Most ultrasound transducers contain numerous piezoelectric elements (usually between 128 and 512), which are arranged in linear or curvilinear arrays. The length of each individual element is typically 2 to 3 mm. The multielement arrays used to produce a beam are normally either linear (sequential) or phased (Figure 7.12). Linear arrays contain between 128 and 512 elements. Linear array transducers activate a single group of approximately 20 adjacent elements to produce one line of sight. Each line of sight provides information for one line of an image. A large number of lines of sight are sequentially generated and combined to form the entire image. After the production of a line of sight, the grouped elements then listen for returning echoes. Grouped elements are used to increase the near field rather than a single element. Another line of sight (A-line) is generated by firing another group of elements that are displaced by one or two elements. With a linear array, a full rectangular field of view is produced by firing groups of elements along the entire length of the array. Curvilinear arrays produce diverging images, which result in a wider field of view than do linear arrays. Clinical applications of linear arrays include peripheral vascular ultrasound and the imaging of small body parts (eg, thyroid, scrotum). Curvilinear array transducers are used in abdominal ultrasound.

Phased arrays usually have between 64 and 128 elements, which are contained in a tighter package than those of linear arrays transducers. All of the elements are utilized during the production of an ultrasound beam. The elements are electrically activated at slightly different times, and this allows the ultrasound beam to be focused and steered through an arc without moving the transducer. Returning echoes are detected by each individual transducer element, which allows for the generation of images. Phased-array transducers are used in cardiac ultrasound because the smaller size allows for imaging between ribs.

Capacitive micromachined ultrasonic transducers (CMUTs) are a relatively new method in medical imaging to produce high-frequency ultrasound energy. Rather than conventional piezoelectricity, which utilizes PZT complexes, CMUTs are silicon based and transduce energy via changes in capacitance. CMUTs have been heavily investigated since the 1990s and have proved to be useful within medical imaging and therapy.

CMUTs are formed with silicon using micromachining techniques. The capacitor cell of a CMUT consists of a fixed electrode, known as the backplate (or bottom electrode), and a free electrode, referred to as the membrane (also top electrode or top plate), which are separated by a vacuum gap. An alternating current applied between these two elements results in membrane vibration, which generates ultrasound energy. The top plate is attracted to the conductive substrate by electrostatic force. Therefore, this concept of operation is known as electrostatic transduction. The attraction is resisted by a mechanical restoring force due to the stiffness of the plate. The CMUT can also serve as a receiver of ultrasonic waves. If an incident ultrasound wave reaches the membrane, a change in capacitance is detected as a current or voltage signal. A direct current bias voltage needs to be applied for signal detection and for transmission.

There are several different ways in which micromachined transducers are fabricated, including sacrificial release surface methods, wafer bonding, and top-down processing. Sacrificial release surface micromachining involves the formation of a cavity beneath a thin plate after a sacrificial layer is deposited or grown on a carrier substrate. The sacrificial layer is then carefully removed with an etchant. Different combinations of the sacrificial layer, plate, and substrate material can be tried to fabricate CMUTs. A process known as deep reactive ion etching provides electrical connectivity to either end of the substrate. A limitation of the sacrificial release process is the relative lack of control of important factors of the deposited layers, such as the uniformity, absolute thickness, and mechanical properties.

Wafer bonding is a method of fabricating CMUTs, which improves upon the limitations of the sacrificial release process. There are several variations of the bonding process, including simple wafer bonding, local oxidation of silicon (LOCOS), and a thick-buried-oxide process. Simple wafer bonding utilizes two wafers—a prime quality silicon wafer and a silicon-on-insulator wafer. The advantage of this process is the improved control over the thickness, uniformity, and mechanical properties of the plate, which is due largely in part to the use of a single-crystal silicon device layer. Limitations include a reduced breakdown voltage and increased parasitic capacitance between the cells, in addition to the relative tendency of electrical short circuits due to contamination during packaging. LOCOS provides for an extended insulation layer structure in the post area, which improves the aforementioned limitations of the simple wafer bonding process. The thick-buried-oxide process is a newer method, which confines and isolates the CMUT bottom electrode to the specific area under the gap where a high electric field is desired. This minimizes the chance of dielectric breakdown and parasitic capacitance in the post region.

CMUTs have several advantages over conventional piezoelectric methods of transduction, including better acoustic
matching with the propagation medium, which allows for greater bandwidth capabilities, potentially lower costs owing to improved fabrication methods, and the ability to have integrated circuits on the same “wafer.” Because a wider pulse-echo bandwidth can be employed (typically >100%), the images obtained from CMUTs offer improved axial resolution, allowing for small targets to be resolved, and improved lateral resolution. State-of-the-art CMUT probes produce images that are comparable or even superior to those of conventional PZT probes. There is much promise surrounding 2D arrays in eventually leading to improvements in efficiency, speed, volumetric imaging, and multibandwidth operation. The 2D arrays of CMUTs offer the potential for real-time 3D imaging in harmonic imaging applications and in high-frequency applications, such as intravascular imaging.

Although CMUTs offer a better axial resolution, further improvements are needed in the sensitivity and resolution to match that of piezoelectric arrays, particularly in deeper depths of penetration. CMUTs are relatively new in the field of ultrasound. Although the potential for CMUTs in diagnostic imaging is wide, the therapeutic capabilities have not yet been fully explored.

As previously discussed, different ultrasound transducers are employed for different purposes. For example, an ultrasound unit will typically come equipped with a linear array transducer of low frequency that provides good penetration. Such a transducer will optimize ultrasound imaging of the abdomen in which deep penetration of sound waves is needed. A separate phased transducer employing a high frequency is used for “small parts.” Such a transducer will have low penetration, although it will provide high spatial resolution for objects of relatively low depth below the skin surface. Furthermore, there are a large number of special purpose transducers for more specific applications. Intracavitary transducers, for example, include transvaginal, transurethral, transrectal, and transesophageal devices. These allow for high-resolution, high-frequency imaging of structures that would otherwise be difficult to image using a standard transcutaneous approach given the depth and/or proximity to cavity wall. Intravascular transducers provide information on the morphology of a vessel wall, estimate the degree of stenosis, and assess for the efficacy of vascular intervention. The differences in the appearance of curvilinear arrays, phased arrays, and intracavitary ultrasound probes are shown in Figure 7.13.