USF Libraries
USF Digital Collections

Design and test of lead-zirconate-titanate flexural plate wave based actuators

MISSING IMAGE

Material Information

Title:
Design and test of lead-zirconate-titanate flexural plate wave based actuators
Physical Description:
Book
Language:
English
Creator:
Akella, Sriram
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Fpw
Pzt
Fabrication process
Sol-gel deposition
Piezoelectricity
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Current MEMS development is driven by the need to develop various 'Miniaturized Total Chemical Analysis Systems (muTAS), biological and chemical sensing, drug delivery, molecular separation, microfiltration, amplification, and sequencing systems. In this work, the use of flexural plate wave devices as an actuator has been investigated.This research was done with the aim of developing a platform to build FPW devices for use in System-On-Chip applications. It is well known that acoustic forces generated by a flexural plate wave (FPW) device can cause fluid motion, by the principle of acoustic streaming. Also the proven ability of FPW devices to cause mixing, filtration and to work as a chemical-biological sensor can be used towards building a micromachined muTAS. The effects of the IDT finger width, spacing, aperture, membrane thickness, and driving conditions on the device performance was studied to understand the impact of IDT design on device performance.
Thesis:
Thesis (M.S.E.E.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Sriram Akella.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 87 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001681113
oclc - 62782659
usfldc doi - E14-SFE0001001
usfldc handle - e14.1001
System ID:
SFS0025322:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Design and Test of Lead-Zirconate-Titanate Flexural Plate Wave Based Actuators by Sriram Akella A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Shekhar Bhansali, Ph.D. Tom Weller, Ph.D. John Bumgarner, Ph.D. Scott Samson, Ph.D. Date of Approval May 29, 2005 Keywords: FPW, PZT, Fabrication process, Sol-gel deposition, Piezoelectricity Copyright 2005, Sriram Akella

PAGE 2

Dedication To my Parents

PAGE 3

Acknowledgements I express my sincere gratitude to Dr. Sh ekhar Bhansali for all the help he has extended to me, both as my advisor and frie nd. For his guidance and time, for his faith and confidence in me, I thank him. I deeply appreciate and cherish all the times we have spent talking about so many things, academic and otherwise. I thank Dr. John Bumgarner, Dr. Scott Sa mson for the time and effort they spent in guiding me through my entire masters program. I express my gratitude to Shinzo for helping in sputtering, Sean for all the nitrid e and oxide depositions, and Dave for all the SEMs. I thank my sponsors, SMDC for funding this project and showing their interest in the ongoing research at COT. I thank Larry Langebrake for giving me the opportunity to work on the MMI project. To the people at NNRC who made my time at USF a memorable one. I would like to thank Rob Tufts, for his excellent clean-r oom guidance. I would like to thank Rich, Jay, Mike and Trung for their timely help. I thank all the members of the MEMS gr oup at USF, especially Shyam, Praveen, Abdur, Rajshekar, Kiran, Puneet, Rahul, Vanda na, Dr.Senthil, Dr. Kim, Kevin and all the other group members for all the lab related hel p. I would also like to thank Balaji, Srinath and Sarvana from the RF MEMS groups for helping me in the measurements, and for all the discussions, we had about fabrication and measurement issues. I thank you all.

PAGE 4

I greatly appreciate the contributions made by the members of my graduate committee in reviewing this dissertation and participating in the thesis defense. I thank my friends Vikram, Santosh, Sudhaka r, Roja, Saraswathi, Sunil, Prashanth and Swathi without all of whom, things would have been very different. Last but not the least; I thank my parents, wife Swetha, my brother Srikanth, grandmother, and my in-laws, for their unconditional love and support, for always respecting my decisions and sh aring my failures and succe ss with equal eagerness. I would like to thank my sister-in-law Neetha and my brother-in-law Pintu for all the fun times we had and I deeply cherish all the times we spent together.

PAGE 5

i Table of Contents List of Tables .....................................................................................................................iii List of Figures ....................................................................................................................iv Abstract .............................................................................................................................vi i Chapter 1 Introduction .........................................................................................................1 1.1 Acoustic Wave Devices .............................................................................................1 1.2 Thickness Shear Mode (TSM) Resonator .................................................................3 1.3 Surface Acoustic Wave (SAW) Device ....................................................................3 1.4 Acoustic Plate Mode (APM) Devices .......................................................................4 1.5 Flexural Plate Wave (FPW) Device ..........................................................................5 1.6 Organization ..............................................................................................................6 Chapter 2 FPW Based Actuators .........................................................................................7 2.1 Active Fluid Pumps ...................................................................................................7 2.2 Flexural Plate Wave Device (FPW) ..........................................................................7 2.3 Acoustic Streaming ...................................................................................................9 2.3.1 Flexural Plate Wave Driven Streaming ............................................................11 2.4 Acoustic Wave Generation ......................................................................................12 2.4.1 Bidirectional IDT ..............................................................................................13 2.4.2 Unidirectional IDT ............................................................................................15 Chapter 3 Piezoelectricity and Lead-Zirconate-Titanate (PZT) ........................................19 3.1 Piezoelectricity ........................................................................................................19 3.2 Derivation of Piezoelectric Effect ...........................................................................19 3.3 Poling .......................................................................................................................23 3.4 Hysteresis Curve ......................................................................................................24 3.5 Lead-Zirconate-Titanate (PZT) ...............................................................................26 3.6 PZT Deposition Methods ........................................................................................28 3.6.1 Pulsed Laser Deposition (PLD)c ......................................................................29 3.6.2 Chemical Vapor Deposition (CVD) ..................................................................29 3.6.3 Sputter Deposition ............................................................................................30 3.6.4 Sol-Gel Method .................................................................................................31 3.7 Properties of PZT ....................................................................................................32 3.7.1 Dielectric Constant and Loss Factor .................................................................32 3.7.2 Piezoelectric Coupling Coefficient...................................................................32 3.8 Influence of Structure and Composition on Physical Properties .............................33 3.8.1 Piezoelectric Properties .....................................................................................33 3.8.1.1. Influence of Orientation ............................................................................33 3.8.1.2 Influence of Composition ..........................................................................33

PAGE 6

ii 3.8.1.3 Influence of Poling Field ...........................................................................34 3.8.1.4 Influence of Film Thickness ......................................................................34 3.8.2 Ferroelectric Properties ........................................................................................35 3.8.2.1 Influence of Composition ..........................................................................35 3.8.2.2 Thickness Dependence...............................................................................35 3.8.3 Dielectric Properties ..........................................................................................36 3.8.3.1 Influence of Orientation .............................................................................36 3.8.3.2 Influence of Composition ..........................................................................36 3.8.3.3 Dependence on Film Thickness .................................................................37 Chapter 4 Fabrication and Process Flow Development .....................................................38 4.1 Background .............................................................................................................38 4.2 Issues in Fabrica tion of the FPW Devices ...............................................................39 4.2.1 Poor Boron Etch Stop.......................................................................................39 4.2.2 Silicon Nitride Membranes ...............................................................................41 4.2.3 Bottom Electrode..............................................................................................41 4.2.3.1 Bottom Electrode/Adhesion Layer ............................................................41 4.2.3.2 Stress in PZT/Bottom Electrode Stack ......................................................43 4.3 Sol-gel PZT Deposition ...........................................................................................43 4.4 Poling of the PZT Films ..........................................................................................47 4.5 Modelling of FPW Device ......................................................................................49 4.6 Mask Design ............................................................................................................51 Chapter 5 Measurements and Results ................................................................................53 5.1 Measurement Set-up ................................................................................................53 5.2 Results and Discussion ............................................................................................55 5.3 Modified Mask Design ............................................................................................63 Chapter 6 Conclusions and Future Work ...........................................................................65 6.1 FPW Fabrication ......................................................................................................65 6.2 FPW Design ............................................................................................................65 6.3 Future Work ............................................................................................................66 References ..........................................................................................................................67 Appendices.........70 Appendix A: Process Flow for FPW Device Fabrication ..................................................71 Appendix B: Membrane Etching (Using EDP).................................................................73 B.1 EDP Mixture Preparation .......................................................................................73 B.2 Waste Handling ......................................................................................................74 B.2.1 Solid Waste ......................................................................................................74 B.2.2 Liquid Waste ....................................................................................................74 B.3 Clean-up ..................................................................................................................75 B.4 Required Safety......................................................................................................75

PAGE 7

iii List of Tables Table 4.1 Various parameters listed for each of the eight designs used ............................52 Table B.1 Chemical ratios for EDP solution preparation..73

PAGE 8

iv List of Figures Figure 1.1 Schematic sketches of the four types of Acoustic sensors [3] ............................2 Figure 1.2 Illustration of a uniform IDT [4] ........................................................................3 Figure 1.3 Deformation filed due to the SAW propagating to the right along the solid surface [4] ..........................................................................................................4 Figure 1.4 Schematic of the SH-APM device sh owing the shear horizontal displacement as it propagates between the i nput and output transducers [3]..........................5 Figure 1.5 Illustration of symmetric and asym metric lamb waves generated in the FPW device [4] ...........................................................................................................6 Figure 2.1 Flexural plate wave de vice with bidirectional IDTs .........................................8 Figure 2.2 Flexural plate wave de vice with unidirectional IDT's ........................................9 Figure 2.3 The forms of the acoustic, shear visc ous and force fields near the plate surface [7] ........................................................................................................11 Figure 2.4 Electrode layout for bidirectional IDT's ...........................................................13 Figure 2.5 Pictorial view of acoustic wave generation by bidirectional IDT's in a FPW device ...............................................................................................................15 Figure 2.6 Electrode layout for unidirectional IDT's .........................................................16 Figure 2.7 Pictorial view of acoustic wave generation by unidirectional IDT's in a FPW device ...............................................................................................................18 Figure 3.1 Illustration of poling proce ss of a piezoelectric material [15] ..........................24 Figure 3.2 A typical P-E hysteresi s loop in ferroelectrics [16] ..........................................26 Figure 3.3 Unit cell of PZT Perovskite [17] ......................................................................26 Figure 3.4 Phase diagram and of PZT [18] ........................................................................27

PAGE 9

v Figure 3.5 Crystal orientation, and composition dependence of piezoelectric constant d33 for PZT thin films near MPB [21, 22] ......................................................33 Figure 3.6 Dependence of remnant polarization and coercive filed on the composition of PZT thin films [23] ......................................................................................35 Figure 3.7 Dependence of remnant polarization and coercive field on film thickness [24] ...................................................................................................................36 Figure 3.8 Variation of dielectr ic constant with change in the PT content and orientation of the PZT thin films [20, 21] ..........................................................................36 Figure 3.9 Dependence of dielectric constant and dissipation factor on film thickness [24] ...................................................................................................................37 Figure 4.1 Process flow and schematic of a FPW device ..................................................39 Figure 4.2 Picture of the cracks f ound after the drive-in process......................................40 Figure 4.3 XRD analysis of the platinum wa fers annealed at different temperatures .......43 Figure 4.4 Single step pyrolysis process flow for PZT deposition ....................................44 Figure 4.5 Process flow for three step pyrolysis for PZT deposition ................................45 Figure 4.6 Setup for spinning PZT ....................................................................................46 Figure 4.7 AFM image of PZT f ilm showing grain formation ..........................................46 Figure 4.8 EDAX analysis of th e sol-gel deposited PZT film ...........................................47 Figure 4.9 Hysteresis curve for unpoled PZT samples ......................................................47 Figure 4.10 Hysteresis curve obtained after the sample is poled [28] ...............................48 Figure 4.11 SEM image of PZT film showing good crystallization .................................48 Figure 4.12 Cross section of the solid mode l of the FPW device built in CoventorWare .49 Figure 4.13 Cross-sectional vi ew of the meshed model ....................................................50 Figure 4.14 Plot of the mask set us ed to fabricate the FPW device ...................................51 Figure 5.1 Schematic of the measurement set-up ..............................................................54 Figure 5.2 FPW characterization set-up .............................................................................54

PAGE 10

vi Figure 5.3 Picture of the FPW device under test (unshielded) ..........................................54 Figure 5.4 Ideal frequency re sponse of FPW device [29] ................................................55 Figure 5.5 Plot of transmission lo ss in dBm vs. driving frequency ...................................56 Figure 5.6 Plot of transmission loss in dBm vs. driving frequency for FPW design # 1 ...57 Figure 5.7 Plot of transmission loss in dBm vs. driving frequency for FPW design # 2 ...57 Figure 5.8 Plot of transmission loss in dBm vs. driving frequency for FPW design # 3 ...58 Figure 5.9 Plot of transmission loss in dBm vs. driving frequency for FPW design # 4 ...58 Figure 5.10 Plot of transmission loss in dBm vs. driving frequency for FPW design # 5 .59 Figure 5.11 Plot of transmission loss in dBm vs. driving frequency for FPW design # 6 .59 Figure 5.12 Plot of transmission loss in dBm vs. driving frequency for FPW design # 7 .60 Figure 5.13 Plot of transmission loss in dBm vs. driving frequency for FPW design # 8 .60 Figure 5.14 Plot of transmission loss in dBm vs. driving frequency of design 2 and 4 ....61 Figure 5.15 Plot of transmission loss in dBm vs. driving frequency of design 3 and 5 ....62 Figure 5.16 Plot of the redesigned mask set isolating each device ....................................64

PAGE 11

vii Design and Test of Lead-Zirconate-Titan ate Flexural Plate Wave Based Actuators Sriram Akella ABSTRACT Current MEMS development is driven by the need to develop various Miniaturized Total Chemical Analysis Sy stems (TAS), biological and chemical sensing, drug delivery, molecular separati on, microfiltration, amplification, and sequencing systems. In this work, the use of flexural plate wave de vices as an actuator has been investigated. This research was done with the aim of developing a platform to build FPW devices for use in System-On-Chip applicatio ns. It is well known that acoustic forces generated by a flexural plat e wave (FPW) device can cause fluid motion, by the principle of acoustic streaming. Also the proven ability of FPW devices to cause mixing, filtration and to work as a chemical-biological sensor can be used towards building a micromachined TAS. The effects of the IDT finger width, sp acing, aperture, membrane thickness, and driving conditions on the device performan ce was studied to understand the impact of IDT design on device performance. For this re search bidirectional IDTs were used, and at a later stage unidirectiona l IDTs on one end and bidirect ional IDTs on the other end can be used for devices that need to function both as a pump and as a sensor.

PAGE 12

viii The devices were tested under various conditions to understand the device performance. The devices were tested with and without a ground. The device performance was also studied when opera ted under the poling condition. Also the response of the device was tested before and after etching the membra ne and a significant improvement in the output response was obser ved with the reduction in the transmission loss. The FPW device performance was studied and it is observed that uniformly spaced IDTs which have an acoustic path of at least 25 wavelengths long are required for good device performance. It is also observed that the devices need to be isolated from one another piezoelectrically fo r better device performance.

PAGE 13

1 Chapter 1 Introduction A revolution in understandi ng and utilizing micromechanical devices has started. The utility of these devices will be enormous, and with time these microdevices will fill the niches of our lives as pervasively as electronics. What form will these microdevices take? What will actuate them, and how will they interact with the environment? We cannot foresee where the developing technol ogy will take us. The number of possible things we could try is beyond possibility [1]. Micromachining has numerous applications in microfluidics, and its use in this area has become even more important as people strive to create complete fluidic systems. A. Manz, in 1993, identified one of the possi bilities as Miniaturi zed Total Chemical Analysis System (TAS) [2]. Some of th e applications include chemical analysis, biological and chemical sensing, drug deliv ery, molecular separation, microfiltration, amplification, sequencing and other applications. These sy stems are examples of the integration of flow channels, mixers, pumps, valves, sensors, etc. which from the building blocks for these TAS systems. In this context, we present the work on design and testing of Flexural Plate Wave (FPW) devices for possible inte gration into fluidic systems for biological and chemical sensing, filters and mixers. 1.1 Acoustic Wave Devices Acoustic wave devices are attractive for use in sensors because of the fact that the wave velocity and damping are sensitive to the surrounding parameters. Acoustic wave based delay line sensors have the following f eatures: high sensitivity, very small in size,

PAGE 14

have a high signal-to-noise ra tio is an oscillato r set-up. The acoustic wave devices, based on piezoelectric crystals, which allow transduction between electrical and acoustic energies, have been fabricated in a number of configurations the most commonly utilized for sensing application are: 1. Thickness Shear Mode (TSM) resonator 2. Surface Acoustic Wave (SAW) device 3. Acoustic Plate Mode (APM) device 4. Flexural Plate Wave (FPW) device. Each of these devices ar e depicted in Figure 1.1 Figure 1.1 Schematic sketches of the four types of Acoustic sensors [3] (a) TSM resonator (b)SAW device (c)APM device (d)FPW device The waves in most cases are excited using an interdigital transducer (IDT), which consists of an interdigital metal pattern on a piezoelectric substrate. An illustration of an IDT electrode is shown in Figure 1.2. 2

PAGE 15

Figure 1.2 Illustration of a uniform IDT [4] In the Figure 1.2, d, determines the period icity of the fingers and is equal to the wavelength of the excited wave. The frequency f o = / p is the resonant frequency of the device. A is the aperture a nd d/4 is the finger spacing. 1.2 Thickness Shear Mode (TSM) Resonator The TSM resonator, also called as th e quartz crystal microbalance (QCM), typically consists of a thin di sc of AT-cut quartz with circ ular electrodes patterned on both side. Due to the piezoelectric properties of the quartz crystal, the application of a voltage between the electrodes results in a shear deformation of the crystal. The TSM resonators are generally used to measure metal deposition rates. 1.3 Surface Acoustic Wave (SAW) Device Lord Rayleigh in 1887 discovered the surf ace acoustic wave m ode of propagation in which the energy is confined very near to the surface of an isotropic solid. Rayleigh waves are useful in sensor application due to the surface confin ement of the energy, making them sensitive to surface perturbations. The surface acoustic wave is most 3

PAGE 16

conveniently excites on a piezoelectric material using an IDT patter n. An application of an alternating voltage between the alternately connected electrodes causes a periodic electric field to be imposed on the crystal gene rating a periodic strain field that produces a standing surface acoustic wave. The standing wave produces a propagating wave in both the directions with the wavefronts para llel to the transducer fingers. A crosssectional view of the strain field generated by the wave pr opagating along the surface of the crystal is shown in Figure 1.3. The penetrat ion depth of these waves is just over one wavelength. Typically these de vices operate at around 100MHz. Figure 1.3 Deformation filed due to the SAW propagating to the right along the solid surface [4] 1.4 Acoustic Plate Mode (APM) Devices Acoustic plate mode devices utilize the shear-horizontal (SH) acoustic plate mode of vibration, which was developed for sens ing in liquids. SH modes have particle displacements parallel to the device surface and normal to the direction of propagation as depicted in Figure 1.4. The absence of a surface-normal component of displacement allows each SH plate mode to propagate in contact with liquids without coupling excessive amounts of acoustic energy into th e liquid. These device s utilize thin singlecrystalline quartz plates that serve as acoustic waveguides, confining the acoustic energy between the upper and lower surfaces of the plat e. Due to this the sensitivity of the SHAPM to mass loading and other perturbations depends on the thickness of the substrate 4

PAGE 17

and both the faces of the crystal can be unde rgo displacements, hence detection can take place at both the surfaces. Figure 1.4 Schematic of the SH-APM device showing the shear horizontal displacement as it propagates between the input and output transducers [3] 1.5 Flexural Plate Wave (FPW) Device In a FPW device, the acoustic wave is excite d in a thin membrane. The unique feature of the FPW device is that its phase velocity is lower than that of most liquids. When the FPW device is immersed in a liquid, a slow mode of propagation exists in which there is no radiation from the plate. The acoustic wa ves generated in a FPW are the first order symmetric and asymmetric lamb waves, illustrated in Figure 1.5 5

PAGE 18

Figure 1.5 Illustration of symmetric and asymmetr ic lamb waves generated in the FPW device [4] 1.6 Organization The following chapters describe the met hods used in the development of the ultrasonic FPW device. Chapter 2 introduces the theory behind the system studied and the operation of the FPW device. Chapter 3 deals with the PZT thin films and there processing parameters. Chapte r 4 describes the fabricati on issues and process flow development. Chapter 5 presents experimental setup and results. 6

PAGE 19

7 Chapter 2 FPW Based Actuators 2.1 Active Fluid Pumps Fluid flow in micro channels has very high flow impedance, i.e., pressure gradient per unit volume flux. This in a cylindrical chan nel varies as the inve rse fourth power of the channel radius. In micro channels with radii in the order of micro meters, the hydrodynamic loads are enormous. In discrete pumping, there is one or more discrete pumping stations, each driving the hydrodynamic load associated with a channel length of the channel. However the number of pumping stations or the strength of these pumps must increase drastically to keep up with the drastically high flow impedances in microchannels. An alternative approach is to use dist ributed pumping as opposed to discrete pumping or the so called Active pumps. The fluid is driven all along the flow circuit via a body force [5]. Since each section of the channel is active it needs to drive only its own impedance; the channel itself actin g as a pump. Such an active pump can be realized by a phenomenon known as Acous tic Streaming as shown by White et al [6]. The FPW device is an example of a device which causes fluid motion by the principle of acoustic streaming. 2.2 Flexural Plate Wave Device (FPW) A typical FPW device consists of a piezo electric film on a 1-m thick silicon nitride membrane supported by a silicon frame. The piezo electric layer is sandwiched

PAGE 20

between a thin continuous bo ttom electrode and the interdigital transducers (IDT) top electrode. In a FPW device, when a sinusoida l excitation is given to the IDTs, they generate an acoustic wave that propagates ac ross the membrane. The IDTs can either be unidirectional (generate an acoustic wave trav eling in one directio n) or bidirectional (generates an acoustic wave that travels in both the directions). Figure 2.1 and Figure 2.2 illustrate a FPW device with bidire ctional and unidirectional IDTs. Figure 2.1 Flexural plate wave device with bidirectional IDTs Figure 2.2 shows a unidirecti onal IDT configuration, this device when excited by two sinusoidal electrical drives operating in quadrature, where one signal lags the other signal by 90 o in phase, launch acoustic waves in only one direction, appropriate for FPW device as an actuator. Figure 2.1 shows a sens ing device with bidirectional IDTs that launch acoustic waves to the right and left across the membrane. Richard Moroey and Chuck Bradley showed that FPWs with uni directional IDTs are effective micropumps [5, 6]. This pumping is caused by a phenom enon called acoustic streaming. Acoustic 8

PAGE 21

streaming is a non linear effect that results in the time-aver aged velocity field having a non-zero component paralle l to the membrane. Figure 2.2 Flexural plate wave device with unidirectional IDT's A unique feature of the FPW device is that it can be designed such that its phase velocity is lower than that of most fluids. Wh en the FPW is in contact or is immersed in a fluid, a slow mode of propagati on exists in which there is no radiation from the plate. This makes the FPW device to function well in a liquid environment and can be used effectively for pumping as well as chemi cal and biological sensing in liquids. 2.3 Acoustic Streaming The various mechanisms of fluid actua tion are rotating cylinders, diaphragm motion, electrophoresis and other physical phe nomena. One mechanism that received limited attention is the phenom ena of acoustic streaming. Acous tic streaming is a steady fluid motion, which is created when high amplitude acoustic waves travel through a dissipative medium. Part of the energy lost through acoustic dissipation is imparted to the fluid as a steady momentum by means of nonlinear hydrodynamic coupling [7]. 9

PAGE 22

10 Acoustic streaming can be cla ssified into three types de pending on the scale of the fluid motion: 1. Eckart type: streaming generated in a non-uniform free sound field, whose scale is much larger than the acoustic wavelength. 2. Rayleigh type: vortex-like streaming generated outside the boundary in a standing wave field, whose scale is comparable to the wavelength 3. Schlitching type: vortex motion generate d in a viscous boundary layer on the surface of an object placed in a sound fi eld, whose scale is much smaller than the wavelength. Acoustic streaming offers two distin ct advantages for applications in microfluidics [8]. 1. Streaming helps to overcome the larg e viscous forces associated with microfluidic flow. This is because (a ) Streaming is generated through acoustic attenuation, with very viscous losses that limit most of the microfluidic devices and (b) the streaming force is a body force and hence present wherever the acoustic field is presen t in significant amplitude and thus resulting in pumping thr oughout the fluidic network. 2. The streaming forces scale favorably as the diameter of the channels decrease. This is due to the generation of st reaming in the acous tic boundary layer, resulting in very large fo rces over a small area. In a FPW device, the flexural plate wave s near the micromachined thin membrane excites an acoustic filed in the neighboring fluid and the acoustic filed in turn causes

PAGE 23

streaming. This actuation mechanism is gentle yet effective means of transporting fluids through flow channels. 2.3.1 Flexural Plate Wave Driven Streaming Determination of the sour ce boundary condition is the fi rst step towards applying the streaming theory to the FPW-driven stre aming problem. The ultrasonically vibrating plate is the source of the acoustic energy; the source motion is that of the flexural waves traversing in the plate. The waves are vi scous loaded zeorth-order antisymmetric Rayleigh-lamb mode waves. The streaming is primarily caused by the first-order acoustic field composed of acoustic and shear viscous modal fields. For fluid loaded FPWs the acoustic field is evanescent, with a penetra ting depth of ~ 20m, and the viscous filed penetration depth is about 0.3m. Figure 2.3 The forms of the acoustic, shear viscous and force fields near the plate surface [7] The force field is directed near along th e direction of propagation of the wave, with several oscillations with in ~3m of the surface and an eventual exponential decay as shown in Figure 2.3 Forces that are generated due to the ac oustic streaming can be classified into lateral forces (i.e., act in direction parallel to the device membrane) and vertical forces (i.e., act in a direction perpendicula r to the device membrane) [9, 10, 11]. The lateral forces include the Strokes drag force and the force from the radiation pressure due to the traveling a nd standing wave. The vertical forces include the lift force, 11

PAGE 24

12 gravity, van der Waals interactions, dielec trophoresis (DEP), and Newtonian added mass 2.4 Acoustic Wave Generation The generation and detection of acoustic waves in a FPW is accomplished by the use of an IDT. The comb like structure of th e IDT is depicted in Figure 3.4. The design of the IDT determines the electrical impedance of the device as well as the operating frequency, bandwidth, and sens itive area. The IDT excites an acoustic wave in the piezoelectric material (PZT) when a RF-voltage is applied to it. This applied voltage results in a synchronously varying deformation of the PZT substrate and the generation of an acoustic wave. The wavelength most effectively excited by the IDT is equal to the periodicity of the IDT pattern. The electrical impedance of the IDT depends on a variety of factors like the electromechanical coupling factor (K 2 ), the dielectric permittivity of the substrate ( r ), and the geometry of the IDT: electrode width, electrode spacing, number of finger pairs, and acoustic aperture (IDT finger overlap length) [3]. For the low-loss and low-noise operation of the device it is important that its impedance be matched as closely as possible to the external components. The impedance of the IDT is largely capacitive due to its phys ical nature and this capacitance is be tuned out by placing an appropriatel y chosen inductor in series with the non-grounded comb of the IDT. The number of finger pairs in the IDT affects the bandwidth (BW) of the device: BW f o /N, where f o is the centre frequency and N is the number o finger pairs. The generation of the wave for both bidirectional and unidirectional wave s is explained below

PAGE 25

2.4.1 Bidirectional IDT A bidirectional acoustic wave is generated by bidirectional IDTs. In the standard bidirectional FPW device, two sets of transduc er fingers are interdigitated and spaced so that the fingers are half a wa velength apart (center-to-center). A single sinusoidal drive voltage is split by a power splitter, which produces two signals, Signal I and Signal II, which are 180 o out of phase with each other. Since the Ti/ TiN/ Pt ground electrode stack is always at a voltage half-way be tween the Signal I and Signal II, and Signal I and Signal II are symmetric about 0V, the ground plane will remain at 0V at all times. This is why devices with bidirectional IDTs do not need an actual account to th e ground. Though unnecessary in principle, the ground plane connection improves stability of the device in practice. Figure 2.4 Electrode layout for bidirectional IDT's The piezoelectric nature of PZT is used to create the periodic tension and compression in the upper half of the membrane When the signal applied to an electrode is positive with respect to the ground plane, the PZT beneath it expands in a horizontal 13

PAGE 26

direction. When the signal applied is negative with resp ect to ground pl ane, the PZT beneath the electrode contracts in a horizontal direction. This behavior is described by 1313EdS (2.1) Where S 3 is the strain in th e horizontal direction, d 31 is the piezoelectric coefficient (m/V) and E 1 is the applied electric filed (V/m) Firstly, consider the rightward traveling displacements in a device operating at resonance in Figure 2.5. The compressions an d tensions under the electrode propagate outward at a velocity determined by the dime nsions, and the material properties of the membrane and any fluid in contact. The hor izontal expansion that begins at time t o under the left most Signal I electrode will pr opagate to the right so that at time t 2 it is under the neighboring Signal II electrode. When the device is at resonance, this expansion arrives just as the voltage at the Signal II electrode is induc ing its own horizontal expansion, the net result being an even la rger horizontal expansion. The same thing happens under the second Signa l I electrode at time t 4 As this horizontal expansion passes under successive electrodes, the re sultant expansion continues to grow in magnitude. For a horizontal contraction under Signal II electrode at time t o a similar process takes place resulting in a growing magn itude of the horizontal contraction as it passes under successive electrodes. The generation and growth of the leftward traveling horizontal expansions and contractions occur in a similar way. The resu lting sequence of tens ion and compression in the PZT layer (upper half of the membrane ) causes periodic flexur e, which propagates both to the right and to left. 14

PAGE 27

An anti-symmetric traveling wave is generated when the PZT layer is approximately half of the thickness of the membrane. When the PZT layer is considerably thicker, when compared to the total membrane thickness, a bulk acoustic wave propagating in the x dire ctions might be expected. Figure 2.5 Pictorial view of acoustic wave generation by bidirectional IDT's in a FPW device 2.4.2 Unidirectional IDT The working on a unidirec tional IDT for unidirecti onal wave generation is explained as follows: each transducer finger pa ir is spaced one-quarter wavelength apart. 15

PAGE 28

This transducer finger pair pattern is repe ated at wavelength intervals to give the quadrature unidirectional interdigitated tran sducers, which is illustrated in Figure 2.6. A quadrature drive produces two si nusoidal signals, Signal I and Signal II, which are 90 o out of phase with one another. For a devi ce with unidirectional IDTs a contact with the ground is necessary to hold it at 0V becau se Signal I and Signa l II are not always symmetric about 0V. The floating fingers are us ed to improve device performance, also a periodic structure improves wave propagation and the floating electr odes help in keeping the wave fronts parallel. Figure 2.6 Electrode layout for unidirectional IDT's 16 Consider the rightward-traveling displacements in a device operating at resonance, in which Signal I leads Signal II by 90 o The horizontal expansion that begins at time t o under the left-most Signal I electrode will pr opagate to the right so that it is under the neighboring Signal II electrode at time t 1 This expansion arrives just as the voltage at the Signal II electrode is inducing horizontal expansion, th e net result being a larger horizontal expansion. A similar thing will happen under the second Signal I electrode at time t 4 As this horizontal expansion passes u nder successive electrodes, the resulting

PAGE 29

17 expansion continues to grow in magnitude. A similar process takes place for a horizontal contraction under the left most Signal I electrode at time t 2 resulting in increasing the contraction magnitude as the co ntraction continues rightwards. Now consider the leftward-traveling di splacements in a device operating at resonance. The horizontal e xpansion that begins at t 1 under the left most Signal II electrode will propagate to the left so that at time t 2 it is under the neighboring Signal I electrode. This expansion arrives just as the vol tage at the Signal I electrode is inducing its own horizontal contraction, resulting in the net horizontal displacements being cancelled out. Similarly, the leftward-traveling contractions and expansions are cancelled. The result is that no traveling wave is launched in a leftward directi on. Figure 2.7 shows a snapshot of the tension and co mpression in the FPW membrane.

PAGE 30

Figure 2.7 Pictorial view of acoustic wave generation by unidirectional IDT's in a FPW device 18

PAGE 31

Chapter 3 Piezoelectricity and Lead-Zirconate-Titanate (PZT) This chapter deals with the discussion of piezoelectricity and the deposition techniques, and the properties of PZT. The va riation in the piezoelect ric and ferroelectric properties of PZT with changes in differe nt processing parameters like composition, annealing temperatures, poling, thickne ss of the samples etc., is discussed 3.1 Piezoelectricity The asymmetry in a unit cell resulting in th e generation of electr ic dipoles due to mechanical distortion is the origin of piezoel ectricity [12]. Piezoelect ricity is defined as electric polarization produced by mechanical strains in crystals belonging to certain classes, the polarization being pr oportional to the strain and changing sign with it. This is the direct piezoelectric effect. The convers e effect, a piezoelectr ic crystal becomes strained, when electrically polarized, by an amount proportional to the polarizing field. 3.2 Derivation of Piezoelectric Effect The piezoelectric effect can be derived by considering the in ternal energy of a crystalline material at equilibrium described by the values of its state variables. For a piezoelectric material, the state variables are entropy S, temper ature T, mechanical strain mechanical stress the electric polarization P, a nd the applied electric field E. Choosing the temperature, stress and the elec tric filed as independent variables, the thermodynamic potential is given by the Gibbs free energy: PE STUG (3.1) 19

PAGE 32

where the tension notation of , P, E is ignored and U being the internal energy of the system. The differential of the above equation is: PdEdSdTdG (3.2) Rearranging the above equation and holding the independent variables constant, we obtain the following relation: STGE,)/( (3.3) PEGT ,)/( (3.4) ETG,)/( (3.5) The system can also be described by expanding the energy function as a series of powers and products of the independent variables. Including only the lower order terms and neglecting the influence of temperatur e, the Gibbs free energy is given by: EEdSGE2 1 2 12 (3.6) where s is the elastic compliance at constant electric field, d is the piezoelectric constant and is the dielectric susceptibil ity at constant stress [13]. Differentiating the above equation with re spect to E and combining with Equation 4.4 yields the direct piezoelectric effect: EdPEGT )/( (3.7) Differentiating Equation 3.6 with respect to mechanical stress and combining with Equation 4.5 yields the c onverse piezoelectric effect: dEs GE E)/ ( (3.8) 20 We can easily verify that the piezoelectr ic constant d is identical for both the direct and converse piezoelec tric effects by comparing the second, mixed partial

PAGE 33

derivative of the Gibbs free en ergy G [14]. Differentiating E quation 3.7 with respect to mechanical stress gives us: dPEG )/()/(2 (3.9) Differentiating Equation 3.8 with respect to applied electric filed E gives us: dE EG )/()/(2 (3.10) The piezoelectric constant is equivalent for both direct and converse effects since the order of differentiation is unim portant, G being a state function S. The notation used in this derivation is simplified and does not consider the tensor nature of the components. The electric fi eld and polarization ar e actually first-order tensors and the stress and strain are second-or der tensors and the piezoelectric constant is a third-order tensor. The direct and converse piezoelectric affects can be expressed in tensor notation as: IiIjkijkiE dP (3.9) iijk mn jkmn E jkEd s (3.10) The 27 independent elements of the piezoelectric tensor d ijk may be written as an array in the shape of a cube. The first suffix, i, designates the layer, while the second suffix, j and third, k denote the row and column of the element respectively. For example the i th layer can be written as a 3x3 matrix: 33 32 31 23 22 21 13 12 11 iii iii iiiddd ddd ddd (3.11) where, i ranges from 1 to 3. 21

PAGE 34

It is established that the mechanical stress and strain tensors and are symmetric, i.e., jk = jk and jk = kj [14]. It becomes necessary that the piezoelectric tensor is also symmetric with respect to j and k (d ijk = d ikj ), reducing the number of independent elements from 27 to 18, enabling expression of the piezoelectric coefficients in the 3x6 matrix commonly used. In order to do this, the j and k suffixes are replaced by single suffixes in the following manner: Tensor Notation 11 22 33 23, 32 13, 31 12, 21 Matrix Notation 1 2 3 4 5 6 This representation reduces the matrix as: 134 5 4 26 5 6 12 1 2 1 2 1 2 1 2 1 2 1 ddd ddd dddi i i ii i i i (3.12) where the s are introduced to account for th e multiplicity of the coefficients. Aligning the layers in rows and their elements in columns provides us with the simple compact 3x6 piezoelectric element matrix commonly used: 3635 34 333231 2625 24232221 1615 14 131211dddddd dddddd dddddd (4.13) When the symmetry of a materials crysta l structure is considered, many of these coefficients are found to be si mply related to zero. Even t hough it appears to be a secondorder tensor, it is important to recall that, it is still a third-order te nsor and transforms as such [14]. 22

PAGE 35

23 The piezoelectric coefficients can be meas ured using the wafer flexure method; a wafer, which has the piezoelectric material deposited on it with the bottom and top electrodes defined, is placed in a special encl osure with a provision to apply pressure from one side of the wafer. When the pressure is applied, the strain produces in the film produces a potential across the film, which is measured. From the applied pressure and the generated electric potential, the piezoelectric coefficients can be calculated. In this research the piezoelectric coefficients could not be measured due to th e lack of the fixture to make the measurements 3.3 Poling A crystal exhibiting spontaneous polarization can be visualized to be composed of negative and positive ions. In a certain temperature range, where the crystal has minimum free energy, these ions are at their equilibr ium positions and the centre of the positive charge does not coincide with the centre of the negative charge This can be visualized as an electric dipole, and the s pontaneous polarization as dipole moment per unit volume, as due to an assembly of these dipoles, which point in the same direction. Crystals, which develop an electric ch arge upon uniform heating, are labeled Pyroelelctric. If the magnitude and direction of spontaneous polarization can be reversed by an external electric fiel d, then such crystals are known as ferroelectrics. Ferroelectric films possess regions with uni form polarization called ferroelectric domains. Within a domain, all the electric di poles are aligned in the same direction. There may be many domains in a film se parated by boundaries called domain walls. Adjacent domains can have their polarization vectors in anti-parallel directions or at right angles to one another. The boundaries betw een these domains are correspondingly known

PAGE 36

as 180 or 90 domain walls. A single domain can be obtained by domain wall motion made possible by the application of a suffi ciently high electric field, the process known as poling. Poling is very important for the application of polycrystalline ferroelectric films. Ferroelectric films do not possess any piezoelectric properties owing to the random orientations of the ferroelectric domains in the ceramics before poling. During poling, a DC electric field is applied on the ferroelectric ceramic sample to force the domains to be oriented or poled. While domains cannoct be perfectly aligned with the field except when the grain or crystal is coincidentally or iented with its cor a-axis in the field direction, their polarization v ectors can be aligned to maximize the component resolved in the field direction. After poling, the electric field is removed and a remnant polarization and remnant strain are maintained in the sample, and the sample exhibits piezoelectricity. A simple illustration of the poling process is shown in Figure 3.1. However, a very strong field co uld lead to the reversal of the polarization in the domain, known as domain switching Figure 3.1 Illustration of poling process of a piezoelectric material [15] 3.4 Hysteresis Curve A consequence of the resistance to doma in switching is that polarization in a ferroelectric is hysteretic, which is another important characteristic of ferroelectrics; namely, that the polarization P is a double-valu ed function of the app lied electric field E, 24

PAGE 37

25 and so is not precisely reversib le with field. A ferroelectr ic hysteresis loop is shown in Figure 3.2. When a small electr ic field is applied, only a li near relationship between P and E is observed because the field is not large enough to switch any domain and the crystal will behave as a normal dielectric material (paraelect ric). This case corresponds to the segment OA of the curves in Figure 3.2. As the electric strength increases, a number of negative domains, which have a polarizati on opposite to the direction of the field, will be switched over in the positive direction along the field, and domain orientation begins to take place. This results in a sharply rising P with increasing field E, and the polarization will increase rapidly (segment AB ) until all the domains are aligned in the positive direction (segment BC). This is a state of saturation in which the crystal is composed of just a single domain. As the fi eld strength decreases, the polarization will generally decrease but does not return back to zero (at the point D in the Figure 3.2). When the field is reduced to zero, some of the domains will remain aligned in the positive direction and the film will exhibit a re mnant polarization Pr. The extrapolation of the linear segment BC of the curve back to the polarization axis (at the point E) represents the value of the s pontaneous polarization Ps. The remnant polarization Pr in a film cannot be removed until the applied fiel d in the opposite direction reaches a certain value (at the point F). The stre ngth of the field required to reduce the polarization P to zero is called the coercive field strength Ec. Further, increasing the field in the negative direction will cause a complete alignment of the dipoles in this direction and the cycle can be completed by reversing th e field direction once again.

PAGE 38

P Figure 3.2 A typical P-E hysteresis loop in ferroelectrics [16] 3.5 Lead-Zirconate-Titanate (PZT) Lead zirconate titanate (PZT) is a solid solution of lead titanate (PbTiO 3 ) and lead zirconate (PbZrO 3 ). It has the chemical composition of Pb(Zr x Ti 1-x )O 3 PZT has a cubic perovskite crystal structure, shown in Figure 3.3(a). The piezoelectric and ferroelectric properties exhibited by PZT can be attributed to the distortion of the perovskite crystal structure. The unit cell is shown in Figure 3. 3(b); in this unit cell the lead atoms (A cations) occupy the corners of th e cube and Ti or Zr (B catio ns) occupy a position at the centre of the cube surrounded by an octahedron of oxygen anions. Ti or Zr ( B ) O Pb (A) Figure 3.3 Unit cell of PZT Perovskite [17] 26

PAGE 39

The excellent ferroelectric behavior of PZT is related to the mobility of the Ti/Zr (B cation) within the oxygen octahedron. Ther e are three main distortions of the cubic perovskite crystal structure. For Ti rich compositions, having Ti concentration above 48% a ferroelectric tetragonal pha se is preferred. The easy po ling axis for this phase lies in the <100> family of directions. For Ti concentrations less th an 48% a rhombohedral phase of the perovskite is pref erred with an easy poling axis lying in the <111> family of directions. At a composition of 48% Ti a nd 52% Zr, a temperature independent phase (morphotropic phase boundary MPB) boundary occurs where the tetragonal and rhombohedral phases co-exist. It is in th is phase that PZT exhibits the largest piezoelectric coefficients, due to the ability of the material to orient along one of the 14 easy axes of poling, 8 in the rhombohedral phase and 6 in the tetragonal phase. Increasing the zirconium concentration, an anti-ferroelectric orthorhom bic phase develops. Figure 3.4 shows the phase diagram for PZT. Figure 3.4 Phase diagram and of PZT [18] From the diagram of the PZT solid solution system, the advantages of the system for using as piezoelectric films are evident: 27

PAGE 40

28 1. Above the curie temperatures the symmet ry is cubic, and the structure is perovskite. 2. The high curie temperatures across the whol e diagram lead to stable ferroelectric states over wide, usable temperature ranges. 3. The Morphotropic phase boundary (MPB) separating rhombohedral (8 domain state) and tetragonal (6 doma in state) ferroelectric domains is first-order, so there is necessarily a two phase region near this 52/48 Zr/Ti composition. 4. In the two phase region, the poling fi eld may draw upon 14 orientation states leading to exceptional ceramic polability (Figure 3.4). 5. As the MPB is near vertical on the phase diagram, the intrinsic property enhancement in compositions chosen near to the boundary pers ists over a wide temperature range. As the result of the unique structure of PZT, both its dielect ric and piezoelectric properties show anomalous behavior near th e MPB (Figure 3.4). Ther e is a maximum in both the relative permittivity and the electromech anical coupling coefficient; the latter is an important parameter describing the piezoel ectric properties. This feature makes PZT a practically useful piezoelectric material. 3.6 PZT Deposition Methods The PZT thin films can be grown in different ways: sputteri ng deposition, sol-gel method, pulsed laser deposition, chemical vapor deposition, ion-beam sputtering, and screen-printing. The deposition methods are re vised here under. Of all these sol-gel method is the most explored and widely used method in research.

PAGE 41

29 3.6.1 Pulsed Laser Deposition (PLD) Pulsed laser deposition, also called pul sed laser ablation, is a popular deposition method among researchers. It consists of ma terial removal by bombarding the surface of the target with short energetic pulsed of focu sed laser beam (laser ablation) this process takes place in a vacuum chamber where a gas (h ere oxygen) is held at constant pressure. Each pulse generally lasts a few tens of na noseconds. The interaction at the surface of the target involves a sequence of events: surface melting, vaporization, multi photon emission, and the generation of plasma plume normal to the surface of the target. The components of the plasma plume are then coll ected on the substrate, forming the desired film. Piezoelectric multi-component materials like PZT are very difficult to grow in the form of a thin film as small variations in th e film stoichiometry results in the degradation of the piezoelectric and ferroelectric propertie s. Control over the pr ocess is required to avoid loss of volatile component s (lead oxide PbO) and forma tion of metastable phases (pyclro-structure). PLD has some advantages with respect to other techniques. The most important being epitaxial thin films can be deposited at low substrate temperatures and with high depositi on rates [19]. 3.6.2 Chemical Vapor Deposition (CVD) Chemical vapor deposition (CVD) is widely used in semiconductor manufacturing for depositing dielectric layers and metals on substrates. CVD consists of flowing gaseous precursors at elevated temperatures. Th e condition of the substrate surface is selected to promote a chemical r eaction between the substrate and the various

PAGE 42

30 precursors. The disadvantage being the use of toxic metal alkyl precursors used and the time and effort required to properly characterize and optimize the deposition process. 3.6.3 Sputter Deposition This technique is used widely to deposit piezoelectric thin films and ahs reached a very high degree of maturity over the years. The advantages are strong adhesion to the substrate, smooth surface, low deposition temp eratures etc. The major disadvantage being to the deposition of multi-oxide films like PZT is the control over the composition and deposition rate. There are several factors which affect th e properties of the deposited film. They are the power supplied, sputtering pressure, the type of gases used in the plasma, temperature of the substrate et c. Sputtering using stoichiome tric PZT composite ceramics is expensive; the targets are prone to cr acking during high power, and produced films deficient in lead, the lead deficiency caused by the preferential resputtering of lead atoms by the negatively charged oxygen ions. Another a pproach is the use of metallic targets in a DC sputtering system which hare indepe ndently powered. This method requires the reactive sputtering in oxygen environment, or multiple metal-oxide targets can also be used. While the above deposition variations are primarily concerned with the altering condition of the target and the plasma, the conditions at the substrate also have a significant influence on the deposited films. Experiments have shown that the substrate temperature and material have a large influe nce on the morphology of the deposited films. Epitaxial films are most difficult to deposit a nd require a substrate material with similar lattice parameters as PZT and substrate temperatures of above 550 o C. Polycrystalline

PAGE 43

31 films can be deposited at similar temperatures on any substrate. The use of certain under layers facilitates the growth in the preferen tial directions. A defect pyrochlore crystal structure is obtained if the films are de posited at temperatures lower than 550 o C, and with no heating amorphous films are obtained. The problem of lead loss at elevated te mperatures during depos ition is countered by depositing excess lead to compensate for th e lead loss. Another strategy is to deposit the films at lower temperatures and ann ealing the deposited amorphous or pyrochlore films to nucleate the perovskite phase. 3.6.4 Sol-Gel Method The sol-gel method is used for much of th e research incorporating PZT layers due to its advantages over other techniques, such as control of compositi on over a large area, low initial processing temperature, high purity, and low cost. While the exact fabrication methods vary between different groups, the basic principle remains the same. Organic precursors containing the desire d metals (generally lead acetate, zirconium-n-butoxide, and titanium-i-butoxide) are dissolved in specific quantities in a solvent (usually used solvents are methoxyxethanol, 2-proponol or isoproponol). These solutions are then carefully diluted to attain a specific viscosit y for the application proc ess. This solution is then filtered and used for film deposition by apcplying a spin-coated process. A few drops are deposited on an electrode deposited wafer and spun at 1000 4000 rpm. The resulting films are dried at 100 o C 150 o C before another treatment at 300 o C 450 o C in air to pyrolyze the remaining organic materials. A shrinkage process takes place, up to 50 70 % in the thickness directio n: thus, for obtaining films w ith thickness appropriate for

PAGE 44

applications, hundreds of nanometers up to microns, multistep deposition process is considered. The films are finally annealed at temperatures in the range of 600 o C 750 o C. There are several advantages and disadvantages of depositing PZT films by this method. A major disadvantage is that the pr ocess of spinning on, pyrolyzing multiple layers for each substrate requires time and effort, making it generally unsuitable for largescale production. The high-temperature annealing treatment is another major drawback of this technique. Finally, the precursors must be selected carefully to avoid toxicity and problems arise with residual stress in these films due to densification and out gassing during annealing. 3.7 Properties of PZT 3.7.1 Dielectric Constant and Loss Factor The dielectric constant (also called the relative permittivity) r is defined as the ratio of the permittivity of the material to the permittivity of free space. The dielectric loss factor is defined as the tangent of the loss angle (tan ). It is the measure of the amount of electrical energy which is lost th rough conduction when a voltage is applied across the piezoelectric element. 3.7.2 Piezoelectric Coupling Coefficient The piezoelectric coupling coefficient is defi ned as the ratio of the mechanical energy accumulated in response to an el ectrical input or vice-versa. It is also referred to the electromechanical coupling coefficient. Mechanical energy convert ed to electrical ener gy K 2 = Input mechanical energy 32

PAGE 45

3.8 Influence of Structure and Comp osition on Physical Properties 3.8.1 Piezoelectric Properties 3.8.1.1. Influence of Orientation The influence of orientation of the stru cture on the piezoelectric properties has been discussed in two fundamental papers [20, 21]. For tetragonal PZT the piezoelectric constant d 33[001] orientation is larger than the piezoelectric constant d 33[111] orientation. Figure 3.5 shows the piezoelectric constant d 33 in the [001] and [111] orientations for different compositions. Figure 3.5 Crystal orientation, and composition dependence of piezoelectric constant d33 for PZT thin films near MPB [21, 22] However, the difference between d 33[001] and d 33[111] is much larger for rhombohedral PZT. It has been noted that the directions with maximum d33 are very close to the perovskite [001] directions. Consequently, it is suggested that PZT epitaxial thin films used for sensors and actuators adopt rhombohedral compositions near the MPB with perovskite [001] orientation. 3.8.1.2 Influence of Composition A systematic investigation of the relationship between composition and the piezoelectric constant d 33 in PZT thin films has been repo rted [21, 22]. It is observed that 33

PAGE 46

34 for PZT thin films in the rhombohedral phase with varying Ti content the piezoelectric constant d 33 increase to a maximum valve near the MPB. Similarly for PZT thin films in the tetragonal phase, the piezoelectric constant d 33 decreases from a maximum valve at the MPB with increasing Ti cont ent as depicted in Figure 3.5 3.8.1.3 Influence of Poling Field In polycrystalline unoriented films the polarization under high DC electric fields is crucial because of the piezoelectric effect cannot be obtained in the absence of a polar axis. Poling of the films leads to substantial increase of their piezoel ectric properties due to the dipole orientation during poling treatme nt. Poling at higher te mperatures and for longer time also increases the piezoelectri c properties. Significan t improvements have also being seen with poling in a DC electric filed increased in steps. The valves generally obtained are less than those of bulk ceramics, and this difference is attributed to the substrate constraints and to the restricti on on the reorientation of the ferroelectric domains. 3.8.1.4 Influence of Film Thickness It is generally observed that the increase in the thickness of the film results in higher piezoelectric coefficients. This is genera lly attributed to the ch anges in the residual stress with thickness and the presence of an interfacial layer at the electrodes whose influence on the effective prope rties is higher for thin films. However cracks in the film reduce their mechanical and electrical integrity.

PAGE 47

3.8.2 Ferroelectric Properties 3.8.2.1 Influence of Composition The ferroelectric properties show a strong dependence on composition of the PZT thin films. Thin films in the rhombohedral ph ase, with Ti content increasing to near the MPB, the remnant polarization values are high and low valves for the coercive field, as compared to the tetragonal phase where w ith increasing Ti c ontent, the remnant polarization increase with decreasing coer cive field as illustrated in the Figure 3.6. Figure 3.6 Dependence of remnant polarization and coercive filed on the composition of PZT thin films [23] 3.8.2.2 Thickness Dependence Thickness dependence of ferroelectric properties in PZT thin films obtained by sol-gel method has being observed. The PZT films with higher thickness had better remnant polarization values compared to film s with lower thickness. Also the coercive field value decrease with increasing thickness. These effects of the PZT thin films have been contributed to the greater contribution from the interfacial effects as the sample thickness decreases. As illustrated in Figure 3.7, the coercive filed increases sharply at a thickness of about 1m and the remnant polariz ation decreases gradually for films with thickness between 2m and 1m [24]. 35

PAGE 48

Figure 3.7 Dependence of remnant polarization and coercive field on film thickness [24] 3.8.3 Dielectric Properties 3.8.3.1 Influence of Orientation The influence of orientatio n of the PZT films was show n in two papers [21, 22], the results depicted in Figure 3.8. For both the rhombohedral and tetragonal phases, the dielectric constant increases as the crystal tiltin g angle increasing fr om the spontaneous polarization direction. For rhom bohedral phase the dielectric constant is maximum in the [001] direction and for films in the tetra gonal phase this is mu ch higher on the [111] direction than on the polar axis direction. Figure 3.8 Variation of dielectric constant with change in the PT content and orientation of the PZT thin films [20, 21] 3.8.3.2 Influence of Composition The maximum dielectric constant for PZT thin films is observed at the MPB for all orientations. The dependence has been expe rimentally verified in PZT films deposited 36

PAGE 49

by RF-magnetron sputtering [23]. A strong dependence o composition has bceen obtained as depicted in Figure 3.8 3.8.3.3 Dependence on Film Thickness Thickness dependence of dielectric propertie s has been found in different PZT films grown by different technique s. Kurchania and Milne [ 24] studied the thickness dependence on sol-gel prepared films with f ilms of various thicknesses from 0.25m to 10m, the results plotted in Fi gure 3.9. The variation in the dielectric properties is due to the presence of growth stresses and thermal stresses, which increase with decreasing film thickness. Figure 3.9 Dependence of dielectric constant and dissipation factor on film thickness [24] 37

PAGE 50

38 Chapter 4 Fabrication and Process Flow Development 4.1 Background The process flow development for the fle xural plate wave devices presented here began with considerations of previous ly developed FPW fabrication methods. FPW Fabrication: Figure 4.1 outlines the process flow previously developed for a basic FPW device for bi-directional IDTs. Pro cessing begins with RCA clean of high resistivity wafers ( 3000 ohm-cm) on which 5000 of th ermal oxide is grown. The oxide on one side of the wafer is tripped protecting the other si de in Buffered Oxide Etch (BOE). The wafers are then cleaned and Sp in-on-Boron is spun on the wafers which are pyrolyzed in the oven at 110 o C for 20 minutes. Next, thes e wafers go through the high temperature boron-pre-deposition in process in the furnace to diffuse the boron. After this the wafers are deglazed in Hydro-Fluoric aci d (HF) for two minutes, followed by a drive in step, a wet oxidation to grow 3000 of oxide. On the boron diffused side, titanium and platinum are evaporated (using AJA e-be am evaporator) and then annealed at 450550 o C in argon atmosphere. On these wafers sol-gel PZT is spun in a multistep process (to be explained in detail in another section) and the final PZT film with a thickness of ~1m thick is annealed in oxygen atmosphere to improve the piezoel ectric properties of PZT. On top of the PZT, gold is deposited with chromium used as the adhesion layer which is patterned to form the IDT electrodes. On the back side etch windows are opened lithographically and etched in BOE to etc th e oxide and expose the silicon. The wafer is then places in a custom built etch jig and the Si etched from the back side in ethylene-

PAGE 51

diamine-pyrocathecol (EDP) at 110oC for appr oximately 3 hours to get the FPW device with 2m thick membranes. Figure 4.1 Process flow and schematic of a FPW device 4.2 Issues in Fabrication of the FPW Devices The various issues encountered and the resulting modifications in the process flow for the FPW device fabrication ar e outlined in the following sections 4.2.1 Poor Boron Etch Stop The boron etch stop obtained by the above said diffusion process gave a very low yield in the membrane etch step, most of the membranes were etch through. The reason 39

PAGE 52

for this being that the required boron concentrations of 1.0e 19 boron atoms/cm was not attained, Spin-on-boron being a limited source diffusion pro cess. For this the diffusion process was changed and solid planar diffusi on sources were used for the boron diffusion (Pre-dep) process. Also the time of the diffu sion step was considerably increased from three hours to a seven hour process. After the diffusion process the wafers were deglazed and oxidised (drive-in step). After this step the wafers were found to have cracks on the top glass surface as shown in Figure 4.2. Figure 4.2 Picture of the cracks found after the drive-in process The reason for this was found out to be the formation of boric acid when the hot boronrich glass interacts with water vapour. This reaction causes pitting on the surface as the boric acid evaporates [25]. To overcome this problem a dry oxide was grown at 700 o C before the drive-in step us ing dry oxygen followed by a degl aze step in HF for 2 min. during this step the boron rich oxide film on the top was completely oxidising and the deglaze step completely removed the oxide a nd free of any surface defects. This two step boron diffusion step using pl anar diffusion sources resu lted in good boron-etch stop layer, which resulted in a uniform smooth membranes for the FPW devices. 40

PAGE 53

41 4.2.2 Silicon Nitride Membranes Low stress silicon-nitride was deposited on bare silicon wafers instead of the diffusion process. This change was bought about after the diffusion tube was changed and the same results were not obtained from the diffu sion process as it had to be characterized for the new diffusion tube. The nitride is also a very good etch stop as the etch rate of silicon-nitride is very low in EDP. Low st ress silicon nitride (1.5m thick) was deposited followed by oxide deposition (0.5m thick) at th e clean room in Star Centre, Largo. The oxide was deposited to reduce the overall st ress in the membrane stack. The membrane stack consisting of the nitride, the Ti/ Pt bo ttom electrode and the spun PZT thin film are all tensile in nature and hence the oxide was included in the stack due to its compressive nature. Consequently nitride and oxide of equal thickness was bei ng tried out to offer better stress compensation to the membrane stack. 4.2.3 Bottom Electrode 4.2.3.1 Bottom Electrode/Adhesion Layer PZT films have been deposited on a variety of substrates with variable results. Substrate selection is critical as single crysta ls of PZT will not grow on substrates with substantially different lattice parameters. Most MEMS and electronic applications require the PZT film to be deposited on an electrode layer. PZT f ilms have been deposited on a wide variety of conductive films like RuO2, IT O (indium-tin oxide), TiN, Pt and Si. The bottom electrode is critical because PZT diffuses through the electr ode layer inhibiting the formation of the perovskite phase, also the electrode laye rs get oxidised because of the repeated heat treatment the wafer has to undergo during the PZT deposition process.

PAGE 54

42 Pt and RuO 2 electrode layers have been found to yield PZT films with the best piezoelectric properties [26]. Platinum was chosen as the bottom el ectrode layer because it does not get oxidised and remains conductive after the P ZT deposition process. Platinum films can also be deposited with [111] preferential orientation, whic h is preferred because [111] planes of Pt and PZT have only 4% lattice mismatch, thus facilitating [111] oriented phase nucleation and growth. Ti layer was used as the adhesion layer between the Pt bottom electrode and the oxide underneath because it acts as a good diffusion barrier. But the titanium was discovered to diffuse through the platinum electrode and forms a thin oxide layer. Various solutions were devised to address this problem, like oxidising the surface of the titanium layer before platinum deposition, use of TiN, TiO 2 TiN layer was selected as it effectively reduces the titani um diffusion through the platin um electrode reducing hillock formation. For this research, a titanium, titanium-nitr ide, platinum stack as been used. The titanium film is sputtered followed by reactive sputtering of titanium-nitride in the same sputtering tool (AJA s putter tool in Star Centre, Larg o) without breaking the vacuum. This was followed by sputtering of platinum. These wafers were than annealed in argon environment in the rapid thermal annealing system. Initially the films were annealed at 450 550 o C, but later changed to a multi step anneal process where they were annealed at 700 o C for 1 minute. This was done to prevent the peeling off occurring at the titanium layer during subsequent processing.

PAGE 55

4.2.3.2 Stress in PZT/Bottom Electrode Stack The stress changes within the bottom electrode stack was observed during the annealing stepc. It was found that sputte red Ti/TiN/Pt stack was initially stressed compressively due to the intrinsic stress in the deposited films. Upon annealing, the stress in the film increases. The stress increase in the system is attributed to thermal expansion differences between the platinum film and th e silicon substrate. In the region between 400 500 o C, the stress begins to decrease due to re-crystallization that removes the structural imperfections and nucleates the [200 ] oriented platinum grains in the film as evident from the XRD analysis shown in Figure 4.3. This is typically accompanied by stress release by the formation of hillocks at the su rface of the film. Figure 4.3 XRD analysis of the platinum wafers annealed at different temperatures 4.3 Sol-gel PZT Deposition 43 Sol-gel PZT is spun on the annealed platinum wafers in a multi-step process to get PZT films of varying thickness. The process flow diagram for spinning the PZT film is detailed in Figure 4.4.

PAGE 56

Figure 4.4 Single step pyrolysis process flow for PZT deposition Sol-gel PZT solution is spun on the anneal ed platinum wafers. The solution is spun in a two step process: 500rpm for 6 seconds followed by 2100rpm at 60 seconds. The films were then dried at 365 o C for two minutes on a hot plate, placing them on the hot plate very slowly. This process of spi nning results in PZT films approximately 0.1m thick. This process is repeated 3 times follo wed by an anneal step in oxygen atmosphere for 1 minute to crystallize and nucleate the rhombohedral phase in the PZT film. The above process of spinning and a nnealing is repeated four time s and results in film which is approximately 1m thick. This multi-step process results in good films rather than a 44

PAGE 57

single step process as the heat treatment results in enorm ous shrinkages in the film thickness direction. Figure 4.5 Process flow for three step pyrolysis for PZT deposition The three step pyrolysis was also perfor med for spinning the PZT films. This was based on the fact that PZT undergoes three major volume shrinkages due to the solvent evaporation, light organic evaporation and decomposition of organics [27]. This endures complete organic removal and to avoids crac king due to the above mentioned shrinkages. Figures 4.6, 4.7, and 4.8 show the set-up for spinning PZT, the AFM-image of the PZT surface and the EDAX analysis of the P ZT film deposited. The absence of platinum 45

PAGE 58

in the analysis shows that the platinum bo ttom electrode has not diffused into the PZT layer, indicating that the bottom electrode to be intact after the repeated heat treatment (annealing process), indicating that TiN can be used as an effective barrier layer. Figure 4.6 Setup for spinning PZT Figure 4.7 AFM image of PZT film showing grain formation 46

PAGE 59

Figure 4.8 EDAX analysis of the sol-gel deposited PZT film 4.4 Poling of the PZT Films The PZT films where poled to improve their ferroelectric and piezoelectric properties. Different poling cond itions were tested: poling at a constant DC bias, poling at a constant DC bias at higher temperatures, and poling in an electric field increased in steps. After which the hysteresis curves are pl otted using the ferroelectric tester, Radiant Technologies RT66A. Figure 4.9 Hysteresis curve for unpoled PZT samples 47

PAGE 60

-300 -200 -100 0 100 200 300 -300-200-1000100200300 KV/cmuC/cm2 Figure 4.10 Hysteresis curve obtained after the sample is poled [28] Figure 4.11 SEM image of PZT film, showing good crystallization Figure 4.9 shows the hysteresi s curve for unpoled PZT samples and Figure 4.10 shows the hysteresis curve for the poled PZT sa mple. The remnant polarization of about 48

PAGE 61

95C/cm is obtained from the films. We also see a significant incr ease in the remnant polarization of the samples. Figure 4.11 shows the SEM image of the PZT film. 4.5 Modelling of FPW Device The FPW device simulation was tried to understand the intera ction between the IDT and the membrane. The solid model of th e FPW was initially built in CoventorWare. Initial trials to mesh the solid model failed as the number of nodes of the meshed model was large and beyond the limitations of the system. The solid model is as shown in Figure 4.12. The model was then meshed using a different approach, using different mesh setting (each mesh element was defined de pending upon individual layer geometry) for each individual layers and the use of ties an d links to mechanically tie the layer together. The cross-sectional view of the meshed model is shown in Figure 4.13. This approach of meshing resolved the meshing issues Figure 4.12 Cross section of the solid model of the FPW device built in CoventorWare 49

PAGE 62

Figure 4.13 Cross-sectional view of the meshed model Simulations on the meshed model were tried by applying a harmonic boundary condition on the IDT metal fing ers, with harmonic analysis with frequencies of between 3.95MHz to 4.05MHz. Simulations of the first m ode of vibration of the membrane were tried out with the above boundary conditions. These efforts were not productive as the program never converged to a solution The modelling of the device was also tried out using ANSYS; however the node limitations on the ANSYS University license di d not result in a solution. Hence the mask design was developed based on the analytical model 50

PAGE 63

4.6 Mask Design Eight designs were developed and used for testing the FPW device performance. The mask design used is illustrated in Figure 4.13 and the different parameters for each of the designs considered is tabulated in Table 4.1 Figure 4.14 Plot of the mask set used to fabricate the FPW devic e 51

PAGE 64

52 Table 4.1 Various parameters listed for each of the eight designs used Design # 1 2 3 4 5 6 7 8 Finger Width (m) 20 40 40 40 40 20 30 30 Finger Spacing (m) 20 20 40 20 40 20 20 20 Membrane Width (mm) 2.631 2.650 2.675 2.649 2.659 2.632 2.622 2.640 Membrane Length (mm) 7.724 8.107 7.744 9.712 10.687 7.540 9.460 8.850 Acoustic Distance (mm) 4.224 2.646 1.088 4.250 4.031 4.013 4.057 3.430 Resonant Frequency (MHz) 5 3.33 2.5 3.33 2.5 5 4 4 Wavelength (m) 80 120 160 120 160 80 100 100

PAGE 65

53 Chapter 5 Measurements and Results The FPW devices fabricated were tested to find the transmission losses in the device. Two sets of measurements were pe rformed on these devices: without a DC bias and with a DC bias applied to them. The ef fect of using a ground electrode was studied on the performance of the devices. Eight differe nt designs were tested each with varying finger width, finger spacing, finger aperture, acoustic distance and different membrane dimensions. Each of the designs had 20 finge r pairs both on the input side and output side. 5.1 Measurement Set-up An arbitrary function generator was used to supply the power to the FPW device. The input signal from the function generator was split using a power splitter which splits the signals into two signals out of phase by 180 o The two signals from the power splitter were connected to the input IDT pair of the FPW using DC probes. The wafer was mounted on a probe station which had five DC probes. Two DC probes were used to actuate the device at the inpu t end and two probes were used to sense the output voltage at the receiving IDT. The fifth probe wa s connected to the bottom platinum ground exposed in the corner of the wafer by PZT etching. The input signal was a sinusoidal signal with a 0dBm input power level. The frequency was swept from 1MHz to 10 MHz and the power level from the receiving end was combined using a power combiner and fed to an Electronic Spectrum Analyzer

PAGE 66

(ESA). The schematic illustration is depi cted in Figure 5.1, Figure 5.2 and Figure 5.3 shows the picture of the set-up wi th various components labelled. Figure 5.1 Schematic of the measurement set-up Figure 5.2 FPW characterization set-up Figure 5.3 Picture of the FPW device under test (unshielded) 54

PAGE 67

5.2 Results and Discussion The ideal frequency response (insertion lo ss) from an FPW is as depicted in Figure 5.4 [29], showi ng the antisymmetric A o mode and the symmetric S o mode. The A o and S o modes are the only two modes that can be generated in plates having small thickness to-wavelength ratio. Figure 5.4 Ideal frequency response of FPW device [29] The initial measurements were carrie d out by connecting Arbitrary Function Generators (AFG) in a phase lock loop so th at both the signals are out of phase by 180 o The out-put signal from the out-put IDT pair was combined using a 0-180 o power combiner and fed to an ESA. Due to the l ack of automation software to the old ESA (HP8562A), the plots could not be exported to excel for analyzing th e results. The results were plotted manually and one is shown in Figure 5.5. The plot in Figure 5.5 is not sufficient to understand the out-put characteristics of the FPW device and hence the setup was slightly modified to automatically sw eep the driving frequency from 1 MHz to 15 MHz to better understand the operation of th e device. The signal from one of the AFG was fed to a power splitter and from the pow er splitter the signal was supplied to the input electrodes. The output was combined us ing a power combiner and supplied to the 55

PAGE 68

ESA (Agilent E4411B). The obtained trace re sults were exported to a computer and plotted in Excel. Figure 5.5 Plot of transmission loss in dBm vs. driving frequency The effect of ground on the devices was te sted by plotting th e frequency response of the device, which were not etched. Eff ectively it will work as a Surface Acoustic Wave (SAW) device (thickness wavelength rati o greater than 2). The effect of poling was also tested on the device performance. The devices were driven by applying a DC bias of 4.5V and the output plotted. The fre quency plots for all the eight designs are plotted in Figures 5.6 to 5.13. 56

PAGE 69

Figure 5.6 Plot of transmission loss in dBm vs. driving frequency for FPW design # 1 Figure 5.7 Plot of transmission loss in dBm vs. driving frequency for FPW design # 2 57

PAGE 70

Figure 5.8 Plot of transmission loss in dBm vs. driving frequency for FPW design # 3 C Figure 5.9 Plot of transmission loss in dBm vs. driving frequency for FPW design # 4 58

PAGE 71

Figure 5.10 Plot of transmission loss in dBm vs. driving frequency for FPW design # 5 59 Figure 5.11 Plot of transmission loss in dBm vs. driving frequency for FPW design # 6

PAGE 72

Figure 5.12 Plot of transmission loss in dBm vs. driving frequency for FPW design # 7 Figure 5.13 Plot of transmission loss in dBm vs. driving frequency for FPW design # 8 60

PAGE 73

Eight designs were considered for the FP W device. Four of them were designed to have the same resonant frequencies with va rying acoustic path, i.e ., the distance between the IDT pairs. Figure 5.14 Plot of transmission loss in dBm vs. driving frequency of design 2 and 4 The influence of varying the acoustic path on the device performance is compared in Figure 5.14. It is observed that frequency response is observed wh en the acoustic path is an integral multiple of the wavelength; design two has a better response compared to design four. In addition, it is observed that the effect of ground stabilizes the response. 61

PAGE 74

Figure 5.15 Plot of transmission loss in dBm vs. driving frequency of design 3 and 5 From Figure 5.15, it is noted that a mini mum acoustic distance of 25 wavelengths is required for the FPW device; design 3 has an acoustic path that is equal to 6.8 wavelengths compared to 25.2 wavelengths for design 5. It is also observed that the characteristic peak at the A o mode resonant frequency cannot be seen as expected. The possible reason for this is that when the FPW deice under test is excited, so me of the energy travels along th e PZT layer to other devices on the substrates and this causes multiple reflections. The result obtained is that at the resonant frequency of each device, the re flected power will be the maximum and hence the obtained frequency plot looks like a table between the frequencies 2MHz to 5MHz rather them having a single peak valv e at the device resonant frequency. 62

PAGE 75

63 5.3 Modified Mask Design From the initial measurements performed on the devices two issues were found. One was the interference from the other device s in the form of reflections and secondly the mask design did not facilitate the use of RF probes to be used for calculating the Sparameters using a network analyzer. To addr ess these two issues the masks have been redesigned with provision for the use of RF probes and also the isolation of the devices from one another by adding an etch step to et ch the PZT layer in between the devices to isolate them piezoelectrically. The issue of dicing the individual devices can also be resolved by etching deep v-groves in between the devices to facilitate easy cleavage of the wafer.

PAGE 76

Figure 5.16 Plot of the redesigned mask set isolating each device 64

PAGE 77

65 Chapter 6 Conclusions and Future Work The intent of this research was to devel op a flexural plate wave device for use as an actuator for microfluidic applications lik e pumping, mixing and filtrations. The effect of various parameters in the IDT design ha ve been studied, primarily focusing on bidirectional IDTs. PZT sol-gel deposition process was developed and working FPW devices have been fabricated. Th e obtained results are summarized. 6.1 FPW Fabrication The process flow for both depositing solgel PZT and the FPW device fabrication has been optimized to increase the yield. The process flow for the fabrication of the FPW device is outlined in appendix A. low stress silicon-nitride was used successfully for defining the membrane. The variation in the st ress of the film during the fabrication has been minimized by the use of TiN barrier laye r, a thin oxide layer to compensate the tensile stresses in the membrane and the use of different annealing conditions. 6.2 FPW Design The bidirectional IDT electrode design has b een used, to facilitat e the study of the device as a sensor and a microfluidic pump at a later stage. The device performance under both a DC bias and under no bias has be en measured. The influence of the ground electrode on the device performance was studied. The FPW device modeling and simulations were carried out in CoventorWare 2004, and ANSYS 8.0. These efforts were not productive, as the programs never

PAGE 78

66 converged to a solution no si gnificant result could be ob tained due to hardware and software limitations. 6.3 Future Work The reflections from other devices on the same substrate have to be eliminated to obtain the true characteristics of each device. This can be obtained, by etching the PZT in between the FPW devices so that they are all disjoint. This can be achieved by redesigning a mask to etch PZT before th e IDTs are defined. One more problem faced during the testing was the scratc hing of the contact pads by the probe tips. To prevent this, the contact pads have to be ma de thicker to facilitate repeat ed testing of th e devices. This has to be achieved with out increasing the IDT finger thickness because doing so will load the membrane mechanically. The device characteristics as a micropump and a sensor have to be studied. The unidirectional IDT transducer design has to be studied for use in microfluidic applications; this facilitates the use of the FPW device for filtering and mixing applications for TAS applications. The FP W device-modeling need s to be carried out using other softwares like CFD, Kappa, FE MLAB or the code can be developed in MATLAB.

PAGE 79

67 References 1. Mohamed Gad-El-Hak, The MEMS Handbook, CRC Press. 2. A.Manz, N. Graber, and H.M.Widmer, Miniaturized total chemical analysis systems: a novel concept for chemical se nsing, Sensors and Ac tuators B, vol. c1, pp. 244, 1990. 3. D. S. Ballantine, R.M.White., S. L. Ma rtin, Acoustic Wave Sensors: Theory, Design, and Physico Chemical Applicatio ns. 1997, San Diego: Academic Press. 4. Michael J. Vellekoop Acoustic wa ve sensors and their technology, Ultrasonics, Vol 36, pp 7 14, 1998. 5. C.E. Bradley, J.M. Bustillo and R. M. White, Flow measurements in a micromachined flow system with inte grated acoustic pumping, Ultrasonics Symposium, pp 505-510, 1995. 6. C.E. Bradley and R.M. White, Acoustically driven flow in flexural plate wave devices: theory and experiment, Ultrasonics Symposium, pp 593-597, 1994. 7. R. M. Moroney, R. M. White, R. T. Howe, Ultrasonically Induced Microtransport, Proceedings of the IEEE 1991 MEMS, Nara Japan, 1991, pp 277282. 8. H. Mitome, The mechanism of generati on of acoustic stream ing, Electronics and Communication is Japan, vol.81, no. 10, 1998. 9. Q. Quan and G. J. Brereton, Mechanisms of removal of micr on-sized particles by high-frequency ultrasonic waves, I EEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, vol. 42, pp. 619, 1995. 10. A. A. Doinikov, Acoustic radiation force on a particle in viscous heat-conducting fluid I. General formula, J. Acoust. Soc. Am., vol. 101, pp. 713, 1997. 11. A. A. Doinikov, Acoustic radiation force on a particle in viscous heat-conducting fluid. II. Force on a rigid sphere, J. Acoust. Soc. Am., vol. 101, pp. 722, 1997. 12. W. G. Cady, Piezoelectricity, Do ver Publications, Inc. New York.

PAGE 80

68 13. F. Jona, and G. Shirane, Ferroelectric Crystals, Pergamon Press, 1962. 14. J. F. Nye, Physical Properties of Cr ystals, Oxford University Press, 1985. 15. Xu, Y., Ferroelectric Materials and Th eir Applications. 1991: North Holland. p. 101-210. 16. Rajsekhar Popuri, Masters Thesis, Electri cal Engineering, Un iversity of South Florida, 2003. 17. Clifford Frederick Knollenbe rg, Masters Thesis Engi neering-Materials Science and Mineral Engineering, Universi ty of California, Berkley, 2001. 18. Jaffe, B., Cook Jr., W. R. and Jaffe, H ., Piezoelectric Ceramics. 1971, Academic Press: New York. 19. Auciello A., and Kingon A. I., A Critic al Review of Physical Vapor Deposition Techniques for the Synthesis of Ferroel ectric Thin Films, Proceedings of ISAF IEEE Eighth Internationa l Symposium on Applications of Ferroelectrics, Greenville, SC, US A, August30-September 2, 1992, 320-331. 20. Du X. H. Belegundu U., Uchino K., "C rystal Orientation Dependence of Piezoelectric Properties in Lead Zirconate Titanate: Theoretical Expectation for Thin Films, Japan Journal of Applied Physics, Vol 36, pp 5580, 1997. 21. Du X. H., Zheng J., Belegundu U., Uchino K., Crystal orientation dependence of piezoelectric properties of lead zircona te titanate near the morphotropic phase boundary, Applied Physics Letters, vol 72, pp 24212423, 1998. 22. Chen .H. D., Udayakumar K. R., Gaskey C. J., Cross L. E., Applied Physics Letters, Vol 67 pp 3441, 1995. 23. Sreenivas K., Sayer M., Jen C. K., and Yamanaka K., Bulk and surface acoustic wave transduction in sputte red lead zirconate titanate thin films, Proceedings Ultrasonics Symposium, 1998, pp 291-295. 24. Kurchania R., and Milne S. J., Journal of Materials Research, Vol 14, pp 1852, 1999. 25. Glenn A. Meyer, James R. Dalcin, Boron Induced Defecs in Silicon, Application notes, 1986. 26. Ryu W. H., Chung Y. C., Choi D.-K., Yoon C.S., Kim C. K., and Kim Y. H., Computer Simulation of the Resonance Characteristics and the Sensitivity of Cantilever Shaped Al/PZT/RuO2 Biosenso r, Sensors and Actuators B, vol 07, Iss 15, 2003.

PAGE 81

69 27. Reji Thomas, Shoichi Mochizuki, Toshi yuki Mihara, and Tadashi Ishida, PZT (65/3) and PLZT(8/65/35) Thin Films by Sol-Gel Process: A Comparative Study on the Structural, microstrutural and elec trical Properties, Thin Solid Films, Vol.433, 2003. 28. Praveen Kumar Sekhar, Masters Thesis, Electrical Engineering, University of South Florida, 2005. 29. Ph. Luginbuhl, S. D. Collins, G.-A. Racine, M.-A. Gritillat and N. F. de Rooij, K. G. Brooks, and N. Setter, Flexural-Plate-Wave Actuators Based on PZT Thin Film, Sensors and Actuators A: P hysical, Volume 64, Issue 1, January 1998, Pages 41-49.

PAGE 82

70 Appendices

PAGE 83

71 Appendix A: Process Flow for FPW Device Fabrication Wafer Type N (100) Si, Double sided Polis hed, Resistivity: 3000-5000 ohm cm, Thickness: 265-285m. A. Step 1. RCA cleans 1. RCA Clean (SC1 and SC2) B. LPCVD Nitride deposition 1. Deposit 2m thick low-stress LPCVD n itride. (at Star Centre, Largo) (Used as a Membrane material and etch-stop) C. CVD Oxide deposition 1. Oxide deposition of 0.5m thick (at Star Centre, Largo) D. Bottom electrode deposition (at Star Centre Largo) 1. Sputter Ti/TiN in AJA sputtering tool. Thickness 800 /100 2. Sputter Pt 2000 thick E. Pt layer annealing 1. Rapid thermal annealing of the wafer in Ar environment. 2. 50 o C/sec ramp rate to 450 o C for 1 min. 3. 50 o C/sec ramp rate to 750 o C for 1 min. F. Sol-gel PZT deposition 1. Spin PZT sol-gel o Step 1: 6 sec, 500rpm o Step 2: 60 sec, 2100rpm. 2. Pyrolysis hot plate 360 o C, 2 min. 3. Spin PZT sol-gel o Step 1: 3 sec, 500rpm o Step 2: 60 sec, 2000rpm. 4. Pyrolysis hot plate 365 o C, 2 min. 5. Spin PZT sol-gel o Step 1: 3 sec, 500rpm o Step 2: 60 sec, 2000rpm. 6. Pyrolysis hot plate 365 o C, 2 min. 7. Rapid thermal Annealing 700 o C, O 2 atmosphere o Ramp at 50 o C/sec to 450 o C for 1 min. o Ramp at 50 o C/sec to 600 o C for 1 min. o Ramp at 50 o C/sec to 700 o C for 1 min. 8. Repeat the above process for 4 runs G. Litho for back etch window (Mask 1) 1. Spin 3000PY on front side, 3000RPM, 60sec

PAGE 84

72 Appendix A (Continued) 2. Soft Bake Hot Plate 155 o C, 60secs 3. Expose 17secs (bright field mask) 4. Hard Bake Hot Plate 110 o C, 60secs 5. Develop RD-6, 25secs 6. Inspect H. Opening of the etch window. 1. Spin AZ4620 on front side, 3000RPM, 60sec, to protect th e PZT from BOE 2. Oxide etch in 10:1 BOE for 15 min. 3. RIE in Plasma therm at clean room in USF, Tampa o 4 steps of 10 min. etches I. Metallization of the ID Ts and the bond pads 1. Deposit Cr/Au 200 /1500 thick in the e-beam ev aporator (Thin films lab ENB110). J. IDT Mask (Front Side Alignm ent) Using EVG mask aligner 1. Spin 3000PY on front side, 3000RPM, 30sec 2. Soft Bake Hot Plate 155 o C, 60secs 3. Expose 17secs (Dark Field Mask) 4. Hard Bake Hot Plate 110 o C, 60secs 5. Develop RD-6, 25secs 6. Etch back Au and Cu 7. Inspect K. PZT Poling 1. Etch PZT near the wafer flat using BOE until Pt surface is seen 2. Clean the wafer with D.I. water and blow dry. 3. Poling Voltage 15V,30 mins using DC probes 4. Inspect Using RT66A for Hysterisis Curve L. Back Side Etching for Membrane 1. EDP etching at 110 o C, 3.5 hrs. Avg. etch rate ~ 75/Hr. (using etch jig) 2. Etch 2m Membrane M. Clean the wafer and Drying 1. The wafer is thoroughly cleaned in Isoproponal and D.I.Water. 2. The wafer is dried in an oven at 110 o C, as blow drying the 2 membranes is not advisable.

PAGE 85

73 Appendix B: Membrane Etching (Using EDP) B.1 EDP Mixture Preparation Table B.1 Chemical ratios for EDP solution preparation Single Double Triple DI water 48 ml 96 ml 144 ml Catechol 48 g 96 g 144 g Pyrazine 0.9 g 1.8 g 2.7 g Ethylenediamene 150 ml 300 ml 450 ml 1. Obtain an EDP beaker, a graduated cylinde r labeled ethylenediamine, the scale, a large plastic weigh boat, a small plasti c weigh boat, and a piece of aluminum foil to cover the beaker. 2. Place an EDP waste bottle in the sink. Make sure that the hotplate is covered with two layers of clean aluminum foil. 3. Measure the DI water in the graduated cylinder and add it to the EDP beaker. 4. Weigh out the catechol in the large plastic weigh boat: add it to the beaker. Discard the weigh boat in the small EDP tr ash can. Wipe off the lip and outside of the catechol bottle before re turning it to the cabinet. 5. Weigh out the pyrazine in the small plastic weigh boat: add it to the beaker. Discard the weigh boat in the small EDP trash can. Wipe down the outside of the pyrazine bottle before returning it to th e cabinet. Wipe down the scale before returning it to the shelf. 6. Place the large graduated cylinder in th e sink. Measure out the ethylenediamine and pour it in the beaker. Immediately cove r the beaker with aluminum foil. Rinse the graduated cylinder with DI water: pour the water from the first two rinses into

PAGE 86

74 Appendix B (Continued) the EDP waste bottle. Rinse the gradua ted cylinder three more times before returning it to the shelf. Wipe down th e outside of the ethylenediamine bottle before returning it to the cabinet. 7. Place the beaker on the hotplate and insert the temperature probe through the tinfoil. Make sure that the probe is imme rsed but does not touch the sides or the bottom of the beaker. Set the probe temp erature to 110C. Do not allow EDP to boil. 8. After the EDP has reached the correct temper ature, place the wafer to be etched in the EDP beaker. The etch rate is about 75m per hour. 9. Post a notice with your name, date and time, and a phone number where you can be reached on the door of the hood. B.2 Waste Handling B.2.1 Solid Waste All solid trash generated during an EDP etch should be deposited in the small EDP trash can. This includes paper towels, weigh boats, plastic scoopers, gloves etc. B.2.2 Liquid Waste EDP and/or ethylenediamene may not be pour ed down the drain. EDP must be disposed of in properly labeled hazardous waste bottles Never tightly cap a bottle containing hot EDP waste or it may explode. Loosely place the cap on the bottle and leave it at the back of the hood until it is sufficiently cool. Then tighten the cap, wipe down the outside of the bottle, and put it in the waste chemical cabinet.

PAGE 87

75 Appendix B (contd.) B.3 Clean-up 1. When etching is finished, turn off the hot plate. Remove the temperature probe, wrap it in a wet paper towel, and set it aside. 2. Carefully pour off the EDP into the wa ste bottle using the funnel. Waste water from the first two rinses s hould also be pored into the waste bottle. Then rinse the beaker at least three more times and us e a wet towel to scrub off any dried EDP residue. Scrub the inside and the outside of the beaker. Thoroughly rinse the funnel and any teflonware. 3. When rinsing your sample, as long as the water continues to turn blue-black it should be poured into the waste bottle. When it is clear it may be pored down the drain. 4. Scrub the plastic shields, including the space between the metal frame and the plastic. 5. Scrub all the surfaces of the hood with we t paper towels. Continue to clean until the paper towels no longer turn yellow. 6. When the hotplate is cool, remove and discard the tinfoil. Carefully clean the temperature probe and place them in the shelf. B.4 Required Safety The following items of apparel are required wh en doing EDP etch; long pants and shoes. The following safety is also required: la b coat, PVC gloves, acid apron, acid gloves, safety glasses, and a face shield. It is recomm ended that users obtain a respirator for this procedure.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001681113
003 fts
005 20060215071247.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 060105s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001001
035
(OCoLC)62782659
SFE0001001
040
FHM
c FHM
049
FHMM
090
TK145 (Online)
1 100
Akella, Sriram.
0 245
Design and test of lead-zirconate-titanate flexural plate wave based actuators
h [electronic resource] /
by Sriram Akella.
260
[Tampa, Fla.] :
b University of South Florida,
2005.
502
Thesis (M.S.E.E.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 87 pages.
3 520
ABSTRACT: Current MEMS development is driven by the need to develop various 'Miniaturized Total Chemical Analysis Systems ([mu]TAS), biological and chemical sensing, drug delivery, molecular separation, microfiltration, amplification, and sequencing systems. In this work, the use of flexural plate wave devices as an actuator has been investigated.This research was done with the aim of developing a platform to build FPW devices for use in System-On-Chip applications. It is well known that acoustic forces generated by a flexural plate wave (FPW) device can cause fluid motion, by the principle of acoustic streaming. Also the proven ability of FPW devices to cause mixing, filtration and to work as a chemical-biological sensor can be used towards building a micromachined [mu]TAS. The effects of the IDT finger width, spacing, aperture, membrane thickness, and driving conditions on the device performance was studied to understand the impact of IDT design on device performance.
590
Adviser: Dr. Shekhar Bhansali.
653
Fpw.
Pzt.
Fabrication process.
Sol-gel deposition.
Piezoelectricity.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1001