This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
The vastness of the ocean offers great challenges for small and compact underwater platforms development. With various exploratory and communication related applications, the demand for high precision acoustic and non-acoustic sensors are expanding. This paper proposed two miniaturize ultrasonic transducer (MUT) designs approaches adopting micro-electro mechanical system (MEMS)-based technology. Two sensing mechanisms were studied, namely, piezoelectric (pMUT) and capacitive (cMUT). Responses of each sensing mechanism were identified using Finite Element Method (FEM) for different materials which were ZnO, PZT and quartz for pMUT and Si3N4, Teflon and PDMS for cMUT. Analyses conducted including mechanical, piezoelectric and harmonic frequency. PZT were found to be the most responsive material with 1.3Ã-10-3 Âµm/V and 2750.2 pC/V of response but carry the lowest resonant frequency at 85.9 kHz. For cMUT design, PDMS carry the highest sensitivity at 1.6 Î¼m/Pa, with respect to Si3N4 and Teflon. However, further analysis revealed that resonance for PDMS occur at 515.9 kHz.
Keywords: Micro-acoustic, piezoelectric, capacitive, underwater
Compact and small underwater platforms can be found in many demanding underwater applications such as for navigation, imaging and communications. It includes robotic-based platform such as the autonomous or remotely operated underwater vehicles (AUV/ROV), as well as stationary platforms including anchor and buoy for sensor networks. Multiple sensors have been utilized involving acoustic and non-acoustic, with miniature designs adopting MEMS and microelectronics fabrication technology forecasted to carry huge potential these coming years . Many advantages offered by miniature transducers including low power consumptions, higher sensitivity, lightweight, embedded electronics and bio-compatibility. Depth sounding, pressure measurement, obstacle avoidance, navigation, imaging and communication are several applications utilize acoustic wave, with micro-ultrasonic transducer (MUT) design. Currently, only two sensing mechanisms are available at the micro-structure sensor design level, i.e. piezoelectric, known as pMUT and capacitive or cMUT [2-5].
In principle, pMUT membrane structure contains functional and structural layers. Functional layer was formed by a layer of piezoelectric film, sandwich between two electrodes, sputtered on both side of piezoceramic. Electrostriction process occurs each time the piezoceramic layer is deformed by inbound acoustic signal as well as when voltage is applied across two electrodes. Structural layers provide support and usually are placed under the functional layer. Silicon based materials are popular choice for its superior elasticity with high yield and tensile strength, providing good support and damping for the whole vibrating structure. Various fabrication techniques have been reported from conventional layer-by-layer deposition on silicon or glass substrate  to adhesive wafer bonding using polymer adhesives . Vast selections of piezoelectric material available include zinc oxide (ZnO), lead zirconate titanate (PZT), aluminum nitride (AlN) and quartz. This paper intends to investigate the pMUT device response using different piezoelectric materials. ZnO, PZT and quartz were selected for this study, considering well documented deposition procedures and lower cost solutions.
On the other hand, cMUT operates based on capacitive principles, where the deformation of the membrane layer caused by the impinging acoustic signal creates a vibration in accordance to the received signal. CMUT has a basic structure of a top electrode fabricated on the membrane layer and separated from the bottom electrode by a specific thickness of gap. This structure thus creating a capacitance structure with one electrode is allowed to move while the other one is fixed. The vibration of the membrane will regulate the separation distance between both electrodes, thus creating a variation of capacitance value. The effects of electrodes, membrane dimension, array configuration and device structure [8-11] have been previously studied. In order to optimise the performance of this kind of micro scale transducer, the material used for the membrane element in cMUT is Silicon Nitride, Si3N4. This is because of its material properties and ease of fabrication factor until to date. However, in theory, other polymer based material such as polydimethylsiloxane(PDMS) and Teflon are also suitable as the membrane element based on their application in microfluidic environment [12-13]. In addition, the elastomic properties of such materials can increase the detection sensitivity due to large deflection per unit pressure in acoustic detection application. This study investigates this issue (suitability of PDMS and Teflon for cMUT application) and includes the Finite Element Analysis (FEA) response of PDMS and Teflon, along with existing material, Si3N4 , as the vibrating element in cMUT .
Two miniaturize transducer design were proposed in this study namely pMUT and cMUT. Both designs are projected to be utilized on compact underwater platforms for various applications. Responses of each transducer design were determined using finite element methods. Key materials selection issues for pMUT and cMUT have been investigated and comparative approach is utilized throughout this work. However, the scope of investigation is limited to a single element design instead of transducer array.
pMUT design under investigations is shown in Fig. 1. It consists of six material layers, with piezoelectric film sandwiched between 0.5 Âµm of aluminum electrodes each. Top electrode is partially sputtered on top of the piezoelectric film. Structural wafer layers consist of bottom-etched silicon-on-insulator (SOI), leaving silica and silicon layers as part of vibrating membrane after bottom silicon layer was etched. The sixth layer is a polymer adhesive, Cytop, placed between functional and structural wafer layer, bonding those two forming a diaphragm. A layer of PDMS polymer used to package the device that consist multiple elements of pMUT in an array. Finally, nickel aluminum bronze was sputtered to encapsulate pMUT element from seawater. Package pMUT was then placed inside nickel aluminum bronze housing. Element configuration in an array and housing analysis are not within the scope of this paper.
Nickel aluminum bronze
film encapsulation (1Âµm)
Piezoelectric film (40Âµm)
Figure 1: pMUT cross section with nickel aluminum bronze housing
For a single layer circular shape diaphragm with a fixed edge, magnitudes of deflection correspond to the amount of applied pressure as given in Equation 1:
Where p is applied pressure in Pascal, a is surface area, h is device thickness, E is Young modulus and m is 1/Ï…, with Ï… is Poisson ratio. Plus and minus sign indicates diaphragm deflection upward and downward respectively. The system was represented with Butterworth-VanDyke (BVD) equivalent impedance model, as shown in Fig. 2, featuring a series of motional inductance, Lm, motional capacitance, Cm and motional resistance, Rm in parallel with stack capacitance, Co; representing vibrating mass, piezo-elasticity, mechanical losses and static parallel plate capacitor respectively.
Figure 2: BVD equivalent impedance model
Stack capacitance, Co was observed from DC analysis, while frequency analyses produced series resonant frequency, fs as well as parallel resonant frequency, fp. Additionally, motional capacitance, Cm and motional inductance, Lm can be calculated as follows:
The basic structure of cMUT is depicted in Fig. 3.
Figure 3: Schematic diagram of a cMUT.
The function of each layer is summarized in Table 1.
Table 1:The structural layer of cMUT
Normally, passivation layer is added on the top structure as a preventive material and/or as an impedance matching layer. The insulation layer is added between substrate and gap to avoid an electrical shorting during its operation. The deflection of the membrane depends on several factors such as the amount of bias voltage, the acting pressure on the membrane, the structure dimension and the flexural rigidity of the membrane material. The bias voltage is required to produce an electrostatic condition needed for cMUT operation, either in transmission or reception and determine the static deflection of the membrane structure. The sensitivity increases with the bias voltage. In response to the pressure wave signal, the deflection of the membrane,w will be maximum at the centre and can be represented by Equation 4:
where po is the total pressure, including the electrostatic force resulting from the DC bias, a is the membrane radius and D is the material dependant flexural rigidity. D is given by:
E, t and Ï… are the Young modulus, Poisson ratio and membrane thickness, respectively.
All analyses and characterizations were done using Analyzer tools package within Coventorâ„¢ software. For pMUT design, it is assumed that the model to be a multilayered plate with all outer surfaces clamped at the fixed edge. For finite element meshing, all models were simplified with only the vibrating part left as shown in Fig. 4.
Figure 4: Side view of vibrating membrane with surfaces boundary condition for FEA
There are five important surfaces, defined on the vibrating membrane of pMUT. First is the outer edge surface, Sfix. Next surface is on the top of the model, Stop whom will receive the inbound acoustic signal. Another two surfaces located on top and bottom of piezoelectric layer and being in contact with top and bottom electrodes, denoted with Spzt and Spzb respectively. And lastly, the lowest surface which is the bottom part of the diaphragm, Sbot. These simplified models were then split into two separate regions with different mesh setting. Both regions however have undergone the same linear tetrahedron meshing. Piezoelectric coefficients of all material layers were set to be zero except for piezoelectric material. Stiffness matrix, piezoelectric strain coefficients and dielectric coefficients for ZnO, PZT and quartz are obtainable from Coventor reference guide. Young modulus, E and Poisson ratio, Ï… were taken as the measure of elastic coefficients of all isotropic materials as in Table 2.
TABLE 2: Mechanical properties of isotropic materials
Nickel aluminum bronze
During piezoelectric analysis, DC voltage was applied across the electrodes and magnitude of membrane deflection is observed. According to Fig. 4, positive voltage is applied when Spzt is supplied with potential with reference to Spzb and vice versa for negative voltage. Additional query revealed the amount of charge, Q generated at Spzt and Spzb thus stack capacitance, Co of the device at supplied voltage, V can be calculated based on Co=Q/V. By using mechanical solver, pressures were applied on Stop and Sbottom to mimic the reception and projection of sound wave. Deflection that occur on +Z and -Z direction of the membrane was observed. For both piezoelectric and mechanical analyses, Sfix is assumed to be neutral electrically and fix mechanically. Finally, frequency analysis was carried out to determine resonant frequency of the pMUT. Closed-circuit condition was applied where zero potential was supplied on both electrodes. Same analysis cycle were carried out, using three different piezoelectric materials with other structural parameters were kept constant.
For cMUT modeling, a structure with specification given in Table 3 was realized by using ANSYS software. The main focus was to investigate the characteristic of three different materials; Si3N4, PDMS and Teflon as membrane element. The deflection behavior of the membrane would be plotted as a comparison on the sensitivity of sensing mechanism between these materials.
Table 3: Model specifications
Size of Membrane (Âµm)
700 x 700
Thickness of Membrane (Âµm)
Gap Thickness (Âµm)
The modeling was performed in 3D and ÂµMKSV unit. The membrane shape was selected to be in square shape. SOLID 95 and SOLID226 element are selected to model the membrane and gap, respectively. For simplicity, the top electrode, where the electrical potential was to be applied, was modeled by assigning the nodes located at the outermost layer of the membrane element. Using the same approach, the nodes at the lower gap material was assigned as bottom electrode where the 0 V potential will be applied. These approaches save computational tasks especially in 3D environment, and at the mean time still preserving the original structure.
For this structural analysis, three material properties that contribute to structural deformation are Young Modulus, Poisson ratio and density. All properties for each material of interests are given in Table 4. The membrane and the gap were glued together using 'vglued' command to ensure the continuity for solving couple field problem . The meshing process was executed using tetrahedral element. By setting the zero displacement boundary condition at every edges, the processing is run to obtain the parameter of interests. In cMUT, the static deflection needs to be analyzed to investigate the initial deflection when the bias voltage is applied.
This is done by simulating the centre node displacement resulted from different value of bias voltage applied onto the top electrode. The range of bias voltage is between 10 V to 300 V. Then the mechanical response of the structure is studied by simulating the relationship between the applied pressure and the resulted centre node deflection in the presence of specific bias voltage. A range of 10 to 50 Pa is used with the bias voltage is fixed to 200 V. Then the analysis is continued by performing the modal analysis on the structure with different membrane material. This analysis is intended to extract the vibration modes of the membrane with their respective natural frequency. However, only the first modal frequency is of our interest as it produces the maximum deflection at the centre point. This modal analysis is also used to anticipate the resonance peak frequency in harmonic analysis. Harmonic analysis is then performed to investigate the harmonic response of the structure at certain range of operating frequency. Based on the performed modal analysis, two set of frequency range are used as the frequency range for harmonic analysis. For Si3N4 and Teflon, the analysis is carried out for a frequency range between 10 to 100 kHz while for PDMS, 10 to 1MHz is used.
Table 4: Mechanical properties of cMUT membrane materials
Results & Discussions
Linear responses have been observed due to applied voltage on pMUT using ZnO, PZT and Quartz piezoelectric material as illustrated in Fig. 5. Positive voltage has resulting maximum upward deflection at the center of the membrane. At the same thickness, PZT has the highest deflection with 1.3Ã-10-3 Âµm/V of response, followed by ZnO and Quartz with 3.0Ã-10-5 Âµm/V and 5.0Ã-10-7 Âµm/V of responses respectively. The analyses were extended to estimate the amount of charge on the top surface of the piezo, Spzt. The charge response curve is shown in Fig. 6 with PZT carries 2750.2 pC/V. ZnO and Quartz piezo layer carry slightly different amount of charge responses with 19.45 pC/V and 7.92 pC/V. Based on the voltage applied across the electrodes, stack capacitance Co of the pMUT according to BVD model can be calculated using Co = Q/V, with assumption that displacement is negligible. The FEA conducted only take into account the anisotropic dielectric in piezo layer and ignore fringing field through the air.
Figure 5: Responses of pMUT to applied voltage for (a) Quartz (b) ZnO and (c) PZT
Applied voltage (V)
Figure 6: Charge response of pMUT in transmitting mode
In mechanical analysis, series of mechanical pressure were applied on top and bottom surfaces of the pMUT which is Stop and Sbot. Positive deflection is observed when upward pressure was applied on Sbot while negative deflection occurs due to downward pressure on Stop. Downward and upward movements of the membrane mimicking the vibration cycle during transmitting and receiving acoustic signal. The thickness of the structural layer consisting of silica and silicon have been optimized and balanced with functional layer of sputtered piezoelectric material. Encapsulation layers consisting PDMS polymer and thin film nickel aluminum bronze provide additional mass and damping so that transmitting and receiving responses of the pMUT is equal. Fig. 7 shows that inbound and outbound pressure responses of pMUT are equal with PZT are the most responsive at 2Ã-10-5 Âµm/Pa followed by ZnO and Quartz with 4Ã-10-6 Âµm/Pa and 8Ã-10-6 Âµm/Pa of responses respectively.
(a) Inbound (b) Outbound
Figure 7: pMUT responses to mechanical pressure
With both aluminum top and bottom electrodes are grounded or supplied with zero potential, closed-circuit resonance analysis was conducted to determine resonant frequency of pMUT with Cm and Lm of the BVD equivalent circuit are in series. By shorten the electrodes and neglecting losses, Rm the piezoelectric effect is assumed to be eliminated with only structural effects are taken into account for resonant frequency derivation. Resonant frequency of the pMUT using three different piezoelectric materials is shown in Fig. 8.
Figure 8: Frequency analysis of pMUT using ZnO, PZT and Quartz.
Transducer with PZT piezoceramic was found to be in resonance mode at 85.9 kHz while pMUT with ZnO and Quartz piezo layer have 193.6 kHz and 188.5 kHz of resonant frequencies respectively. The amount of deflection during closed-circuit resonant analysis is invalid, and modal analysis should be conducted to determine the amount of deflection at the resonance. In modal analysis, the resonant frequency is computed as the eigenvalues of the undamped and homogenous equation of motion of the vibrating structure of system.
The response of the designed structure is measured from the deflection profile obtained from the finite element analysis. Fig. 9 illustrates the static deflection of three different membrane materials for cMUT application.
Figure 9 : Static deflection of Si3N4 , Teflon and PDMS
PDMS and Teflon both show a large deformation under the effect of bias voltage condition. This behaviour is due to low Young Modulus especially for PDMS, which makes the material displaying more deflection upon specific force. Si3N4 has the least static deflection, even under high bias voltage. The same phenomenon observed in the relationship between the external pressure and centre point deflection. This relationship represents the behavior of the device during its operation under external frequency varying pressure signal. The graph from Fig. 10 is used to measure the sensitivity of the device by means of how much it deflects per unit pressure.
Figure 10: Relationship between external pressure and centre point deflection
The more it is deflected, the larger the variation in the gap separation, so the effect it has on capacitance changed also becomes more observable. From Fig. 10 the sensitivities of Si3N4, Teflon and PDMS are found to be 0.034 Î¼m/Pa, 0.67 Î¼m/Pa, and 1.6 Î¼m/Pa respectively.
The modal analysis yields a result as in Table 4. The result is only confined to the first vibration mode (fundamental frequency) in order to estimate the fundamental resonance frequency of the structure during operation.
Table 4: First mode frequency of different membrane materials
at design dimension (Modal Analysis)
Fig. 11 represents the harmonic analysis of three different membrane materials. Two range of frequency range are used for clarification purpose. For comparison purpose, only the resonance frequencies of the first mode vibration are considered.
Figure 11: First mode frequency using harmonic analysis for
(a) Si3N4 and Teflon (0-100kHz) (b) PDMS (0-1MHz)
The resonance frequency between the modal and harmonic analysis yield an error of 2%, 6% and 8% for Si3N4 , Teflon and PDMS respectively. This is due to step size sampling frequency in harmonic analysis. As can be seen, at this design configuration, the resonance frequency of Si3N4 and Teflon are below 100kHz, but the PDMS vibrates with a wider bandwidth with the resonance occurs at MHz range.
pMUT design with polymer adhesive layer have been proposed specifically for wafer bonding fabrication technique. Three piezoelectric materials have been successfully studied namely PZT, ZnO and Quartz. PZT carry the highest charge response, however mechanical analysis revealed that PZT lack of stiffness resulting in major deflection when pressure is applied compare to ZnO and Quartz. By having the least charge response, quartz piezoelectric crystal also less stiff compared to ZnO, resulting in higher resonant frequency. In the middle of it, ZnO produced enough charge amounts and vibrate moderately due to applied cyclic pressure with reference to PZT and quartz. From frequency analysis, three resonant frequencies were obtainable from three different piezoelectric materials. At less than 100 kHz, many underwater applications tailing such as deep penetration depth sounding and low resolution-long range imaging. For ZnO and Quartz with resonant frequencies higher than 150 kHz, high resolution imaging and obstacle avoidance sonar seems the right applications. Furthermore, this study has featured the usage of thin nickel aluminum bronze film as encapsulation layer, protecting the transducer from corrosive sea water. Previously we have suggested the usage of nickel aluminum bronze alloy as top electrodes, covering the pMUT from sea water .
Meanwhile, the response of cMUT analysis concludes that there are advantages and disadvantages associated with each material to be implemented as a membrane. The use of rubber-like material such as PDMS and Teflon increases the sensitivity to external pressure signal, but unfortunately these materials also suffer from large static deflection that significantly change the original structure of cMUT. So at high DC bias voltage , Si3N4 is still superior albeit the sensitivity to the external pressure reduced by the order of two compared two PDMS. In order to obtain a good tradeoff between the static and dynamic deflection, other factors such as application and sensitivity tolerance also need to be considered. In frequency range perspective, high resonance frequency of PDMS membrane makes it suitable for higher frequency operation such as in imaging application. Si3N4 and Teflon (at the design geometry) are more suitable for lower frequency operation such as in acoustic communication and navigation.