Horn-Type Piezoelectric Ultrasonic Transducer: Modelling and Applications

The piezoelectric transducers can be categorized into sonic, ultrasonic and megasonic transducers based on the operating frequency. In each category, the design and function of the transducers vary significantly. The sonic transducers work at an audible frequency range, typically less than 20 kHz. In this frequency range, the transducers could be designed in the bending mode. Examples of the sonic transducers are the bimorph cantilever and buzzer (siren) unimorph transducer (APC International, Ltd., 2002; Yeo et al., 2008). The buzzer transducer is made of a piezoelectric disk attached to a metallic substrate. It is able to produce a sound level more than 100 dB at resonance in the frequency range of kHz (Bai et al., 2007; Miclea et al., 2008). At off-resonance conditions, the buzzer transducer can also be used as an actuator for piezoelectric diaphragm pump (Accoto et al., 2000; Woias, 2005). The ultrasonic transducers usually work at a frequency range from ~20 kHz to ~200 kHz (Abramov, 1998; Atar et al., 1999; Chen & Wu, 2002; Chu et al., 2002; Hongoh et al., 2004; Lin, 1995, 2004a; Mattiat, 1971; Medis & Henderson, 2005; Prokic, 2004; Radmanovic & Mancic, 2004; Sherrit et al., 1999a, 1999b; Tsuda et al., 1983; Wiksell et al., 2000). Transducers in this range could also be designed in the bending mode. Some examples are the tube transducer for cylindrical ultrasonic motor and disk/ring transducer for travelling wave motor (Li et al., 2007a; Uchino, 2003). But in this frequency range, a more common transducer design is the longitudinal mode. Langevin (sandwich, converter, Tonpiltz, etc.) transducer represents a typical structure (Abramov, 1998; Atar et al., 1999; Chen & Wu, 2002; Chu et al., 2002; Hongoh et al., 2004; Lin, 1995, 2004a, 2004b; Mattiat, 1971; Medis & Henderson, 2005; Prokic, 2004; Radmanovic & Mancic, 2004; Sherrit et al., 1999a, 1999b; Tsuda et al., 1983; Wiksell et al., 2000). It is sometimes connected with a horn (wave guide, booster, sonotrode, etc.) to focus or transmit energy. This type of transducer provides a wide range of applications, including welding, machining, sonochemistry, cleaning, underwater communication, ultrasonic surgery, etc. The megasonic transducers work in the frequency range of MHz. The most widely used vibration mode is the thickness mode for this range. One of the typical applications of the transducers in this category is the megasonic cleaning (Kapila et al., 2006). It provides advantages such as gentle and controllable cavitation, which will incur less damage on the cleaned parts compared with its traditional ultrasonic counterpart. Hence it is more suitable for precision cleaning.


Introduction
The piezoelectric transducers can be categorized into sonic, ultrasonic and megasonic transducers based on the operating frequency. In each category, the design and function of the transducers vary significantly. The sonic transducers work at an audible frequency range, typically less than 20 kHz. In this frequency range, the transducers could be designed in the bending mode. Examples of the sonic transducers are the bimorph cantilever and buzzer (siren) unimorph transducer (APC International, Ltd., 2002;Yeo et al., 2008). The buzzer transducer is made of a piezoelectric disk attached to a metallic substrate. It is able to produce a sound level more than 100 dB at resonance in the frequency range of kHz (Bai et al., 2007;Miclea et al., 2008). At off-resonance conditions, the buzzer transducer can also be used as an actuator for piezoelectric diaphragm pump (Accoto et al., 2000;Woias, 2005). The ultrasonic transducers usually work at a frequency range from ~20 kHz to ~200 kHz (Abramov, 1998;Atar et al., 1999;Chen & Wu, 2002;Chu et al., 2002;Hongoh et al., 2004;Lin, 1995Lin, , 2004aMattiat, 1971;Medis & Henderson, 2005;Prokic, 2004;Radmanovic & Mancic, 2004;Sherrit et al., 1999aSherrit et al., , 1999bTsuda et al., 1983;Wiksell et al., 2000). Transducers in this range could also be designed in the bending mode. Some examples are the tube transducer for cylindrical ultrasonic motor and disk/ring transducer for travelling wave motor (Li et al., 2007a;Uchino, 2003). But in this frequency range, a more common transducer design is the longitudinal mode. Langevin (sandwich, converter, Tonpiltz, etc.) transducer represents a typical structure (Abramov, 1998;Atar et al., 1999;Chen & Wu, 2002;Chu et al., 2002;Hongoh et al., 2004;Lin, 1995Lin, , 2004aLin, , 2004bMattiat, 1971;Medis & Henderson, 2005;Prokic, 2004;Radmanovic & Mancic, 2004;Sherrit et al., 1999aSherrit et al., , 1999bTsuda et al., 1983;Wiksell et al., 2000). It is sometimes connected with a horn (wave guide, booster, sonotrode, etc.) to focus or transmit energy. This type of transducer provides a wide range of applications, including welding, machining, sonochemistry, cleaning, underwater communication, ultrasonic surgery, etc. The megasonic transducers work in the frequency range of MHz. The most widely used vibration mode is the thickness mode for this range. One of the typical applications of the transducers in this category is the megasonic cleaning (Kapila et al., 2006). It provides advantages such as gentle and controllable cavitation, which will incur less damage on the cleaned parts compared with its traditional ultrasonic counterpart. Hence it is more suitable for precision cleaning.
The capability of the transducer is closely associated with the frequency, vibration mode and wavelength. As a result, different transducers will achieve different performances in a wide range of applications. In comparison, the ultrasonic transducer with longitudinal vibration mode is the only one that is able to achieve both large vibration amplitude and high power (density) simultaneously. The vibration velocity of the transducer is limited by the mechanical, thermal and electrical properties of the materials. Currently, the PZT material is only able to achieve a vibration velocity less than 1 m/s as compared to 20 m/s for the titanium material in the same ultrasonic frequency range (Li et al., 2007b;Mason & Wehr, 1970;Muhlen, 1990). Thus, by using the titanium horn, the transducer can achieve larger vibration amplitude. In addition, the transducer at the axial direction has a relative long wave length compared to the thickness mode. A large volume of the PZT material could be used to increase the power capacity of the transducer (Mason, 1964). Furthermore, the power can be focused to a small area by the horn, which further improves the power density of the transducer. Due to the superiorities of high amplitude and high power (density), the horn type transducer has been found to have many unique applications.
The shape of the horn-type transducer varies significantly depending on the applications. Conventionally, the modelling of the transducer can be based on electrical circuit theory and FEA theory. Typically, the electrical circuit theory is used to design the transducer while FEA theory is used to analyze the performance of the transducer. It is convenient to use the electrical theory to design the dimension. However, for analysis of parameters, such as mode shape, stress distribution, power consumption, etc, the FEA theory is more advantageous. In the present work, we proposed an improved electrical circuit theory, called Finite Electrical Circuit Element Modelling (FECEM), which combines functions of both design and analysis. In this method, the transducer can be represented using a network, which are the connections of electrical circuit elements. The great advantage of the proposed method is the high efficiency in the development of piezoelectric transducers. We also proposed a method to estimate the mechanical quality factor Q m which is important to estimate parameters like velocity and power consumption. Conventionally, Q m parameter was either assumed or measured. In the present work, the FECEM was also compared with ANSYS FEA 2D analysis. Consistent results have been found.
The ANSYS 2D fluid-structure acoustic modelling was also introduced in the present work. In this modelling, the system response can be simulated when the transducer is loaded in the acoustic field. This model is useful to analyze the loading effect on the transducer. It is also essential to study the distributions and magnitude of sound pressure under different boundary conditions. Lastly, in this chapter two applications of the horn-type transducer are introduced. They are thrombolysis transducer and acoustic pump. Both can be designed and analyzed using the above mentioned FECEM and ANSYS 2D theories. The thrombolysis transducer is a horntype transducer connected with a long and thin transmission wire. It can be used to emulsify the thrombus in the blood vessel. The transducer was modelled and characterized in the present work. Efficient and effective emulsification results have been observed which proves the design and analysis were successful. The acoustic pump is another application of the horn-type transducer. The pump possesses a simple structure with just a transducer and a casing. Due to the vibration of the transducer, an acoustic field will be established inside the casing. As a result, the liquid will be forced to flow due to the nonlinear acoustic effects, such as streaming and cavitation. The unique features of the pump are high pressure head, continuous flow, no body oscillation and insensitive to the tubing length. The potential applications of the pump are sonochemistry and microfluidics. Fig. 1 shows the typical structure of the horn-type piezoelectric transducer. Generally, the transducer comprises four parts, i.e., the back mass, the piezoelectric stack, the front mass and the horn. The piezoelectric stack is clamped between the front and back mass. The horn is coupled to the front mass through the method such as bolt. In this configuration, the function of the piezo stack is to generate vibration, which is then amplified by the horn. The horn together with the front and back mass also has function of impedance match to maximize the energy transfer from the piezoelectric stack to the loadings at the tip of the transducer . The horn can have various designs based on the applications, such as conical profiled, step profiled, exponential profiled, etc (Hongoh et al., 2004;Horita, 1967;Khmelev et al., 2005;Minchenko, 1969;Muhlen, 1990). To maximize the vibration amplitude, the horn can be designed with multiple sections. Each section has the same or different profiles. The total length of the transducer is usually multiples of the half wavelength.

FECEM theory
Equivalent circuit network theory is a well-accepted method in piezoelectric material study and transducer design (Li et al., , 2009(Li et al., , 2010Mason,1964;Prokic, 2004;Radmanovic & Mancic, 2004). In this method, the transducer is divided into different sections. Each section will be represented using a two-or three-port electrical network. The networks are then connected to represent the whole transducer. For example, in the case of Fig. 1, four networks can be used to represent the back mass, piezo stack, front mass and the horn. The behaviour of the transducer can then be analyzed from the obtained electrical model. Conventionally, the method is efficient in the transducer design, such as calculating the dimension based on the given frequency. But in the case of performance simulation or analysis, such as calculation of stress distribution, predication of power consumption, estimation of quality factor, etc, the application of this theory is limited. In the present work, we proposed an improved equivalent circuit theory, called Finite Electric Circuit Element Modelling (FECEM), which are convenient and efficient in both transducer design and analysis. In this method, the transducer will also be divided into multiple sections. But the section will be further divided into a number of subsections, called element. An element is represented using an electric circuit network. The connections of the networks constitute the electrical circuit model of the transducer. The number of elements can be flexibly selected based on the complexity of the structure and parameters to be solved. The advantages of this method are parameters such as mode shape, stress distribution and mechanical quality factor can be easily solved. In the following sections, the FECEM will be introduced using Langevin transducer as an example. For simplicity, the structure of horn is ignored. The transducer is considered to contain front mass, piezoelectric stack and back mass. The front and back masses are supposed to be made of the same material of aluminium. And the piezoelectric material is PZT.

Electrical circuit element
Naturally, the Langevin transducer mentioned above can be divided into front mass section, piezoelectric stack section and back mass section. The front and back mass sections are uniform cylinders. Therefore, they can be represented using the same element as shown in Fig. 2 (Lin, 1987).

Fig. 2. Front and back mass elements
This is a two-port network, in which U 1 , U 2 and I 1 , I 2 are the voltages and currents at the input and output terminals. Physically, they mean the forces and velocities at the two ends of the vibrating element. Z 1 and Z 2 are impedances in the network. They can be expressed as where  is density, c is sound speed, S is cross sectional area of the element, k is wave number and l is the length of the element. Based on Fig. 2, the two-port voltages can be expressed in terms of the two-port currents using the "Z" matrix (Balabanian & Bickart, 1981;Hayt & Kemmerly, 1993) 11 1 1 2 1 22 1 2 2 2 This is actually the constitutive relations between the forces and velocities for the elements. This relation can also be represented using the following "chain" or "ABCD" matrix (Balabanian & Bickart, 1981;Hayt & Kemmerly, 1993) 12 The Z matrix and ABCD matrix can be converted to each other based on the following Eqs. 7 and 8. This provides much convenience for different parameter calculations (Balabanian & Bickart, 1981;Hayt & Kemmerly, 1993 The piezoelectric stack comprises two parts, PZT and bolt. It can be represented using a three-port network element as shown in Fig is the static capacitance,  is the transformer ratio, Z 1bolt , Z 2bolt , Z 1pzt and Z 2pzt are the impedances from the bolt and PZT, respectively. Z 1bolt and Z 2bolt can also be expressed using Eqs. 1 and 2, respectively. The expressions for the rest parameters are shown below (Lin, 2004b;Sherrit et al., 1999aSherrit et al., , 1999b.

Finite electrical circuit element model
The vibrating elements are mechanically connected in series. Therefore the electrical model can be built as shown in Fig. 4 (Balabanian & Bickart, 1981;Hayt & Kemmerly, 1993).

Fig. 4. Finite electrical circuit element model
The output of one element is the input of the adjacent one. Therefore, the following relation exists For the front mass, assuming there are n elements, the constitutive relation will be 1 1 front front n front front n AB UU For the piezoelectric stack, assuming there are m elements, the following relation can be obtained Similarly, for the back mass, supposing the number of elements is q, so there is

Modal analysis
Modal analysis is to find the resonant frequencies and the corresponding mode shapes of the transducer. The resonant frequencies can be found using the system Z matrix. Assuming the short circuit condition, i.e.,V=0, transform the ABCD matrix T 1 into the Z matrix Z 1 based on Eq. 7, and then set the determinant to zero The solutions of the equation are the resonant frequencies. The mode shape can be obtained based on the relation in Eq. 22. Assuming I 1 =1, U 1 =0, i.e., unit velocity and zero force at the free end, the velocity at each point can hence be solved. In another word, the mode shape can be found.

Harmonic analysis
The harmonic analysis is to find the performance of the transducer as a function of frequency. The parameters interested might include magnitude of velocity, magnitude of stress, power consumption, impedance characteristics, etc. This can be done based on the system relation Eq. 27. Assuming one mechanical port is loaded with a resistor R, the other mechanical port is short circuited, i.e., no force, and the electrical port is applied with 1 V, Eq. 27 will turn to Solving the above equation, the velocity at the two ends (I 1 and I n+m+q ) can hence be obtained. The stress can then be estimated based on the mode shape, velocity obtained and the following relation (Silva, 1999) where E is Young's modus, v is velocity and x is position. The stress is an important factor that affects heat generation and vibration amplitude (Li et al., 2007b(Li et al., , 2007c).
The input current to the transducer is solved using the following relation (Lin, 2004b)  

2
IVj CI I  where I 1 and I 2 are velocities at the surface of the PZT. Solving the current, the electrical properties such as impedance and power consumption can hence be estimated accordingly.

Resistive load and mechanical quality factor
As seen in Section 3.2.2, before the harmonic analysis, an equivalent load R has to be assumed. Without it, the vibration magnitude of the transducer will go to infinity. In practice, the load can be the results of the external loadings or the internal losses. In the case of free vibration, the load is due to the internal mechanical loss, which is the function of the mechanical quality factor Q m (Chen & Wu, 2002) where M is the equivalent mass. To find M and Q m , the following method is applied.

(a) Mechanical quality factor
The mechanical loss per unit volume is estimated using the following equation (Li et al., 2007b(Li et al., , 2007c in which s is the compliance coefficient, T is the stress, tan 1 mm Q   is the mechanical loss tangent, and 2 12 ws T  is the elastic energy density. In Eq. 35, s and Q m is solely material dependant. But the stress T not only depends on the materials, but also relies on the position x. Therefore, to calculate the total loss for a single material i, the following integration should be applied 00 11 1 where A is the cross sectional area of the material, W i is the elastic energy for material i.
This equation means the equivalent quality factor Q m can be estimated using the mechanical elastic energy and the individual material quality factor.

(b) Equivalent mass M
Because the piezoelectric transducer can be modelled using both lumped parameters and distributed parameters, the following relation exists (Li, 2004) 22 0 11 22 where v x is the velocity at position x,  is density, A is cross sectional area and V t is the velocity at the tip. In Eq. 40, the left part is the kinetic energy of the transducer expressed in lumped parameters and right part in distributed parameters. From the derived mode shape mentioned earlier, v x in Eq. 40 can be easily known. Therefore the equivalent mass can be conveniently solved.

Concluding remarks
The FECEM theory is introduced in this section. This theory is useful for both transducer design and performance analysis. Based on this theory, the transducer can be divided into finite numbers of electrical cricuit elements, which are connected to form the electrical www.intechopen.com Accordingly, the modal analysis and harmonic analysis can be achieved by solving the network equations. The mechanical quality factor and equivalent mass were derived, too. The FECEM was also compared with ANSYS 2D modelling. The results show reasonable accuracy.

ANSYS 2D modelling
The piezoelectric transducer can also be analyzed using ANSYS 2D simulation (ANSYS, Inc., 2010; Li et al., 2008Li et al., , 2009Li et al., , 2010. One of the advantages of this method is that the acoustic field with various boundary conditions can be coupled to the piezoelectric transducer. Therefore, the model is useful to analyze the response of the transducer under different loading conditions. It is also essential for the study of the sound pressure distributions at different boundary conditions. Fig. 6 shows the transducer-fluid coupled model, in which the 2D model revolving about the axial of rotation will produce the 3D structure of the system. The liquid medium is water which directly contacts the mechanical structure of the piezoelectric transducer at the structure-liquid interface. For the piezoelectric transducers, the axisymmetric structure element Plane42 for metallic parts and couple field element Plane13 for PZT materials were applied. The Plane42 element has degrees of freedom UX and UY, and Plane13 element has degrees of freedom VOLT, UX and UY. For liquid medium, acoustic element Fluid29 was applied. In the acoustic domain, it has degrees of freedom PRES. And at the boundary and interface, it possesses degrees of freedom UX, UY and PRES. The boundary condition can be controlled by the absorption coefficient MU. In this case, the horizontal boundary was set "hard" and curved boundary is set "absorbing", corresponding to MU=0 and MU=1, respectively. Fig. 6 (b) shows the meshed model of the system. After meshing, an electrical load of 1 V was applied to the PZT stack. To be noted, the transducer analysis and the acoustic analysis can be modelled separately, which might be more efficient for certain problems.  Fig. 7 shows the harmonic analysis results of the transducer displacement at the resonant frequency. The colour represents the magnitude of the displacement in the unit of meter. ANSYS provides the harmonic displacement with the real and imaginary part. Because at resonance, the electrical excitation and displacement response has 90 deg phase difference, the imaginary part dominates. It is also noticed that the resonant frequency is 64510 Hz in the figure, which is different from 65845 Hz in Table 1. This is because the liquid loading changes the resonant frequency of the transducer. www.intechopen.com curved absorbing boundary. The maximum pressure is located at the centre of the transducer surface. It decreases as the distance from the transducer surface increases. The magnitude and distribution of the pressure wave are sensitive to the boundary conditions. Therefore the boundary conditions should be optimized accordingly based on the applications.

Thrombolysis transducer
Thrombolysis transducer is a device used to deal with the vascular thrombotic occlusive disease, which is a major cause of morbidity and mortality in the developed world. The development of a blood clot or thrombus in a blood vessel compromises distal blood flow and is the usual cause of a heart attack or stroke. Established treatment is the urgent removal or dissipation of the occluding thrombus. This is achieved with the use of a simple aspiration catheter, mechanical thrombectomy, or pharmacological agents such as thrombolytic drugs (Atar et al., 1999;Brosh et al., 1998;Janis et al., 2000;Siegel & Luo, 2008). Ultrasonic emulsification of the blood clot is another technique for thrombolysis. This is achieved by acoustic cavitation and mechanical fragmentation (Bond & Cimino, 1996;Chan et al., 1986). Compared with conventional mechanical thrombectomy techniques ultrasonic thrombolysis exhibits the advantage of inherent tissue selectivity (Rosenschein et al., 1994;Tschepe et al., 1994). This is because thrombus is highly susceptible to ultrasonic cavitational emulsification, while the arterial walls, which are lined with cavitation-resistant matrix of collagen and elastin, are not. Ultrasound energy has also been shown to improve myocardial reperfusion in the presence of coronary occlusion (Siegel et al., 2004). The system of ultrasonic thrombolysis generally contains a generator, a transducer and a catheter. The tip of the transducer will go through the catheter lumen to the blood clot. Then the generator will provide control signal and power to the transducer. The vibration of the tip will emulsify the blood clot. In the following sections, the thrombolysis transducer developed in NTU (Nanyang Technological University) will be introduced. Fig. 9 shows the structure of the piezoelectric thrombolysis transducer designed in the present work. The transducer consists of five parts. The first part is the end cap. It serves to prestress the adjacent PZT stack and also adjusts the mechanical impedance applied to the stack. The second part is the PZT stack, which is clamped between the end cap and the horn. It is the most crucial part of the transducer, where the vibration is generated. The third part www.intechopen.com of the transducer is the horn, which functions to magnify the displacement produced by the PZT stack. The fourth part is a long and thin transmission wire, which should be flexible but sufficiently stiff for energy transmission. The last part is a distal vibration tip that consists of a ball or a short cylinder with an enlarged diameter (~ 1.5 mm) compared to the connecting transmission wire (~ 0.5 mm). The enlarged diameter increases acoustic power emission to the surrounding liquid and blood clot. The vibration produced by the PZT stack is transmitted through the horn and then the transmission wire to the distal vibration tip. The acoustic energy emitted from the tip is finally used to emulsify the clot.

Fig. 9. Schematic illustration of the piezoelectric thrombolysis device
The transducer operates at ~26.7 kHz longitudinal vibration mode. This is a low ultrasonic frequency aiming at reducing the heat generation (Francis, 2001;Li et al., 2007b;Siegel et al., 2000). The diameter of the PZT stack is 10 mm. Maximum input power is 20 W. The length of transmission wire is 1 m, which is made of a high strength material, Ti-6Al-4V, for achieving a high vibration velocity (Li et al., 2007b;Mason & Wehr, 1970;Muhlen, 1990). For practical operations, due to the long length of the transmission wire, the bending mode could be excited, which would increase loss and decrease the efficiency of the device. To avoid this, the transmission wire should be coaxial with the horn.

FECEM modeling
Because of the high aspect ratio of the device (length/diameter), it is very advantageous to model the device using the FECEM theory. Fig. 10 shows the mode shape of the transducer. It can be seen that along the axial direction, the displacement generated by the PZT stack is first amplified by the horn, then further amplified by the transmission wire. As a result, the tip has a much larger displacement than the PZT stack. Also considering the smaller area of the vibration tip to the PZT stack, it can be said that the energy produced by the PZT stack has been focused to the tip through this design. Therefore, the tip will work effectively for the blood clot emulsification.

ANSYS 2D acoustic modeling
During practical operation, the vibration tip will be surrounded by liquid and produces an acoustic field. Fig. 11 shows the simulation of the acoustic field generated by an OD 1.5 mm and length 3 mm vibration tip. The tip is connected to an OD 0.5 mm transmission wire, and the vibration frequency of the tip is 30 kHz. The simulation was carried out using ANSYS 2D acoustic analysis. Fig. 11 shows that the maximum ultrasonic pressure is located at the top and bottom surfaces of the vibration tip, which is normal to the displacement direction. This indicates that emulsification should be most effective at these locations. The radiation area multiplied by the normal surface velocity of the tip is known as the source strength (International standard, IEC 1998;Li et al., 2009). The ultrasonic pressure amplitude is  IEC 1998;Li et al., 2009). Therefore, in the present work, a horn is applied to amplify the vibration velocity from the PZT crystal. And a ball or a short cylinder tip is attached at the distal end of the transmission wire to enhance the radiation area. Fig. 11. Acoustic pressure pattern generated by the tip in the liquid

Acoustic characterization
The acoustic properties of the transducer were characterized qualitatively. As the acoustic pressure becomes larger and larger, two phenomena might be observed around the tip of the transducer, i.e., cavitation and streaming (Abramov, 1994(Abramov, , 1998Young, 1999). Cavitation is the generation and burst of bubbles in the liquid due to the high amplitude of the acoustic www.intechopen.com pressure. Along with the burst of the bubbles is the high intensity shock wave and impinging of the liquid, which is normally very strong and even sufficient to break a hard surface. Fig. 12 (a) shows the cavitation bubble clusters generated at the vibration tip in the silicon oil. The cluster is usually generated at the center of the surface and then flows away along the acoustic axis. When the distance becomes larger from the surface, the bubbles might be agglomerating and floating upwards due to the buoyancy force. The cavitation threshold of the water is larger than silicon oil and is also frequency, temperature and static pressure dependant (Abramov, 1994(Abramov, , 1998Young, 1999). Generally in water, only when the acoustic field is very strong, visible bubble clusters can be observed. The generation of the bubble cluster at the tip surface also proves that the highest intensity of acoustic energy is around the tip. This is consistent with the earlier theoretical analysis.
The second important phenomenon is acoustic streaming, which is the flow of the liquid as the result of high and nonlinear acoustic pressure field and the generation of bubble clusters (Abramov, 1994). The flow pattern is usually that the flow is along the acoustic axis and outwards, the surrounding liquid will flow to the tip and compensate the outflows. The streaming effects have both positive and negative effect in application of ultrasonic thrombolysis. For the positive aspect, the streaming will expedite the ablation of the blood clot, especially when the fibrinolytic agents are present (Atar et al., 1999). The negative effect is that the streaming is inclined to push away the blood clot, which increases the difficulty to control, especially, the floating clot. Fig. 12 (b) is the demonstration of acoustic streaming, which pushes the water away from the tip surface. Because the demonstration is near the water surface, the droplets jumped into the air, resulting in the phenomenon so called atomization (NII et al., 2006).

Blood clot emulsification
The transducer was tested to emulsify the blood clot in an anechoic tank. The tank was filled with water and lined with sound absorption materials both at the walls and the bottom. The dimensions of the tank are 0.6 × 0.6 × 1.3 m 3 . A holder made of natural latex of 30 µm thick was used to contain the blood clot. The clot was prepared by naturally coagulating fresh rabbit blood overnight at 6 o C. During operation, the tip of the transducer was pointed at the clot surface as shown in Fig. 13. The blood clot was immediately emulsified when the power was provided. The whole procedure documents rapid clot lysis (~ 750 mg/min) and confirm the effectiveness of the transducer in thrombolysis.

Concluding remarks
The piezoelectric thrombolysis transducer consists of an end cap, a PZT stack, a horn, a transmission wire and a vibration tip. The transducers vibrate longitudinally and generate maximum acoustic pressure at the tip. The acoustic pressure induces effects such as cavitation and streaming at the tip. Blood clot can be effectively emulsified using the designed transducer.

Acoustic pump
The piezoelectric pump has various designs (Laser & Santiago, 2004;Luong & Nguyen, 2010;Yeo et al., 2008). According to frequency, the pumping mechanism will be different. At the sonic low frequency range, the diaphragm pump is a popular design, which utilizes the reciprocating movement of piezo transducer to displace the liquid. However, as frequency increases, the diaphragm mechanism becomes more and more difficult to achieve. One of the important reasons is the response mismatch between the diaphragm and valves (Hu et al., 2004;Zhang et al., 2003). However, there are other pumping mechanisms. At the ultrasonic and megasonic frequency range, the acoustic effect could be used (Chen & Lal, 2006;Frampton et al., 2004;Hasegawa et al., 2008;Koyama, et al., 2010;Li et al., 2010;Zhang et al., 2010). It has been well known that the intense and nonlinear ultrasound could induce effect such as cavitation and streaming, which are able to generate the flow of the liquid (Abramov, 1994;Li et al., 2010;Raton et al., 2007). This work will report a piezoelectric acoustic pump based on these acoustic effects.

Structure of the pump
Fig. 14 shows the structure of the piezoelectric pump designed in NTU. The pump comprises a piezoelectric transducer, a reflecting condenser and a casing. The structure of the transducer is similar as the thrombolysis transducer, but without transmission wire. There is a through hole inside the transducer working as the flow path. The transducer works approximately at 30 kHz. The reflecting condenser has a "" shape, which focused the acoustic energy at its corner. The casing has a cylindrical shape with inlet on the wall. The casing, condenser and horn surface form the boundaries of the acoustic field. Because the pump mechanism is based on the acoustic effect, the pressure magnitude and pattern inside the field will affect the performance of the pump significantly. The distribution of sound pressure level inside the casing is therefore simulated in the next section.
www.intechopen.com Fig. 15 is the simulation of the pressure pattern inside the pump casing. The casing was simplified as a cylinder in this 2D axisymmetric simulation. The length of the cylinder is one wavelength. The graph shows the right half piece of the model. The horn tip generates 1 µm displacement at 30 kHz. The colour represents the pressure magnitude in the unit of dB. Two boundary conditions are compared. Fig. 15 (a) is the absorbing condition. As mentioned earlier, a travelling wave will be generated at the tip. It propagates inside the casing and finally is absorbed at the casing boundary. The maximum pressure is located at the horn tip. The magnitude is approximately 187 dB. Fig. 15 (b) shows the hard or reflecting boundary condition. In this case, a standing wave will be established inside the casing. The maximum pressure is still at the horn tip. But the magnitude is increased to 255 dB, which is much larger than that of the absorbing boundary condition. This indicates that the reflecting boundary will be more effective than the absorbing boundary to focus the energy or increase the pressure magnitude. In practice, the reflecting boundary could be achieved by using materials with high acoustic characteristic impedance such as metals (Feng, 1999). www.intechopen.com

Performance test
Fig . 16 shows the performance of the pump. From Fig. 16 (a), it can be seen that the maximum flow rate is about 10 mL/min. And maximum pressure head is about 250 kPa. The maximum pressure head is voltage dependant. As voltage increases, the pressure head can be further increased as shown in Fig. 16 (b). Compared with the regular piezoelectric diaphragm pump whose pressure head is mostly in the magnitude of ~50 kPa (Laser & Santiago, 2004;Luong & Nguyen, 2010), the advantage of the acoustic pump is the high pressure head. The reason for this is that the acoustic cavitational effect contains large energy, which is able to generate a high force to overcome the back pressure.

Concluding remarks
The acoustic pump is based on the nonlinear acoustic effect such as cavitation and streaming. The cavitation effect contributes significantly on the high pressure feature of the pump. The advantage of the piezoelectric acoustic pump designed in NTU is simple structure, less oscillation and high pressure head.

Conclusions
The general structure of the horn-type transducer is the Langevin transducer with a horn. The advantage of such device is that both large amplitude and high power (density) can be obtained simultaneously. FECEM theory can be used for the development of the piezoelectric transducer. This method is based on two-port network theory. Parameters like dimension, frequency, stress, velocity, quality factor, impedance, etc, can all be solved using this method. It also shows consistent results with ANSYS 2D analysis. The ANSYS 2D analysis is effective for the piezoelectric structure, acoustic field, or structure-acoustic coupled field analysis. It is also important in the development of piezoelectric devices. Two applications of horn-type transducer were also introduced in this work. The thrombolysis transducer is used for blood clot emulsification. Due to high aspect ratio of the structure, it is also a good example to be analyzed using the FECEM theory. The acoustic pump, on the other hand, possesses a simple structure for pumping applications. The working principle is nonlinear acoustic effects of streaming and cavitation. The major advantage of the pump is the high pressure head. It might be used for sonochemistry or microfluidics. The piezoelectric transducer converts electric signals into mechanical vibrations or vice versa by utilizing the morphological change of a crystal which occurs on voltage application, or conversely by monitoring the voltage generated by a pressure applied on a crystal. This book reports on the state of the art research and development findings on this very broad matter through original and innovative research studies exhibiting various investigation directions. The present book is a result of contributions of experts from international scientific community working in different aspects of piezoelectric transducers. The text is addressed not only to researchers, but also to professional engineers, students and other experts in a variety of disciplines, both academic and industrial seeking to gain a better understanding of what has been done in the field recently, and what kind of open problems are in this area.