DEVELOPMENT OF NOVEL ACOUSTIC WAVE BIOSENSOR PLATFORMS BASED ON MAGNETOSTRICTION AND FABRICATION OF MAGNETOSTRICTIVE NANOWIRES Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. ______________________________ Suiqiong Li Certificate of Approval: ______________________________ ______________________________ Bryan A. Chin Zhongyang Cheng, Chair Professor Associate Professor Materials Engineering Materials Engineering ______________________________ ______________________________ Barton C. Prorok Curtis Shannon Associate Professor Professor Materials Engineering Chemistry and Biochemistry ______________________________ ______________________________ Tung-shi Huang Joe F. Pittman Associate Professor Interim Dean Nutrition and Food Science Graduate School DEVELOPMENT OF NOVEL ACOUSTIC WAVE BIOSENSOR PLATFORMS BASED ON MAGNETOSTRICTION AND FABRICATION OF MAGNETOSTRICTIVE NANOWIRES Suiqiong Li A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama August 4, 2007 iii DEVELOPMENT OF NOVEL ACOUSTIC WAVE BIOSENSOR PLATFORMS BASED ON MAGNETOSTRICTION AND FABRICATION OF MAGNETOSTRICTIVE NANOWIRES Suiqiong Li Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. ______________________________ Signature of Author ______________________________ Date of Graduation iv VITA Suiqiong Li, daughter of Tiankai Li and Surong Xu, was born on June 2, 1973, in Beijing, P. R. China. She attended Beijing University of Technology (formerly known as Beijing Polytechnic University), Beijing, for five years and graduated with a Bachelor of Science degree in Material Science and Engineering in July, 1996. In August of 2001 she entered the Graduate School at Auburn University to pursue a doctoral degree in Material Science and Engineering. She has performed research for five years under the guidance of Dr. Zhongyang Cheng. v DISSERTATION ABSTRACT DEVELOPMENT OF NOVEL ACOUSTIC WAVE BIOSENSOR PLATFORMS BASED ON MAGNETOSTRICTION AND FABRICATION OF MAGNETOSTRICTIVE NANOWIRES Suiqiong Li Doctor of Philosophy, August 4, 2007 (B.S., Beijing University of Technology, 1996) 269 Typed Pages Directed by Zhongyang Cheng There is an urgent need for biosensors that are able to detect and quantify the presence of a small amount of biological threat agents in a real-time manner. Acoustic wave (AW) devices, whose performance is defined by mass sensitivity (S m ) and merit quality factor (Q value), have been extensively studied as high performance biosensor platforms. However, current AW devices face some challenges in practical applications. In this research, two types of AW devices ? magnetostrictive microcantilever (MSMC) and completely free-standing magnetostrictive particle (MSP) ? were developed. The research consists of two parts: 1) Design and the feasibility study of MSMC and MSP based sensor technology; 2) Fabrication and characterization of micro/nano MSPs vi made of amorphous Fe-B alloy. Both MSMC and MSP based sensors are wireless/remote and work well in liquid, which makes the sensors good candidates for in-situ detection. The performance of MSMC was simulated and compared with the state of art AW devices: microcantilevers. The MSMC exhibits the following advantages: 1) remote/wireless driving and sensing; 2) ease of fabrication; 3) works well in liquid; 4) exhibits a high Q value (> 500 in air); 5) well suited for sensor array development. MSMCs in milli/micro sizes were fabricated and their performance was characterized in air and liquid. The experimental results confirm the advantages of MSMC mentioned above. The in situ detection of the yeast cells and Bacillus anthracis spores in water were performed using MSMC biosensors. MSPs in the shape of strip and bar were investigated. Strip-shape MSPs in milli/micro sizes were fabricated. The resonance behaviors of MSPs at the even and odd vibration modes were analyzed. MSP exhibits a S m about 100 times greater, and a Q value about 10 times greater, than MCs. A multiple-sensor and a multiple-target approach were developed to further enhance the performance of MSP-based sensors. A unique methodology was created to detect the target species on the sensor surface at different locations by combining even and odd harmonic mode signals. As with other AW devices, a smaller size results in a higher S m . To create micro/nano sized MSMC & MSP sensors, amorphous Fe-B thin films and nanowires were fabricated using electrochemical deposition. The microstructure, morphology, composition and magnetic properties of the fabricated nanowires were determined. It is found that the films and the nanowires are excellent candidates for developing micro/nano MSPs and MSMCs. vii ACKNOWLEDGMENTS The research conducted in this dissertation would not have been possible without the support and help of many people. I would like to express my sincere thanks and gratitude to my advisor Dr. Zhongyang Cheng. He provided me with generous guidance and invaluable knowledge, both professionally and personally, to help me reach my potential during my stay at Auburn. I would like to give my appreciation and sincere thanks to Dr. Bryan A. Chin, Dr. Barton C. Prorok, Dr. Curtis Shannon, Dr. Tung-shi Huang and Dr. Yucheng Feng for serving on my committee and for their many insightful suggestions and kind help. Thanks go to many of my colleagues and friends, Dr. Zhimin Li, Milind D. Arbatti, Xiaobing Shan, Liling Fu, Lisa Orona, Kewei Zhang, Peixuan Wu, Xin Yang, Levar Odum, Dr. Liwei Wang, Jiehui Wan, Roy Howard, L. C. Mathison, Dr. Hong Yang, and others too many to mention, for their never-wavering support and kind help. I would like to acknowledge the financial support of this work from USDA, FAA, USGS and 3M. I also appreciate the financial support from DOE and the PNNL-Summer Research Institute (SRI) for the research carried out at Pacific Northwest National Lab (PNNL). I had the pleasure of working at PNNL with Dr. Chongming Wang, Dr. Scott Lea, Mark Engelhard, and Bruce Arey. They have been invaluable mentors for me. Most importantly, I would like to thank my family for their encouragement and unconditional support throughout my education. viii Style manual or journal used: IEEE Transactions on Magnetics Computer software used: Microsoft Word, Excel, OriginPro 7.5, Peakfitting TM ix TABLE OF CONTENTS LIST OF FIGURES ......................................................................................................... xiv LIST OF TABLES........................................................................................................... xxi CHAPTER 1 INTRODUCTION .............................................................................................................. 1 1.1 Conventional Bacteria Detection Techniques........................................................... 2 1.2 Review of Current Biosensor Techniques ................................................................ 5 1.2.1 Electrochemical Biosensors............................................................................... 6 1.2.2 Optical Biosensors ............................................................................................. 9 1.3 Acoustic Devices .................................................................................................... 11 1.3.1 Operating Principle .......................................................................................... 11 1.3.2 Common Acoustic Devices.............................................................................. 14 1.3.3 Microcantilevers .............................................................................................. 15 1.4 Fundamentals of Magnetostriction and Magnetostrictive Materials ...................... 18 1.4.1 Magnetoelastic Effects..................................................................................... 18 1.4.2 Joule Magnetostriction in Ferromagnetic Materials ........................................ 21 1.4.3 Response of Magnetostrictive Material under Different DC Bias................... 24 1.4.4 Piezomagnetism in Polarized Magnetostrictive Materials............................... 26 1.4.5 Magnetomechanical Coupling Coefficient ...................................................... 27 1.4.6 Magnetostrictive Materials .............................................................................. 28 1.4.7 Applications of the Magnetostrictive Materials............................................... 30 References..................................................................................................................... 33 CHAPTER 2 RESEARCH OBJECTIVES ............................................................................................. 42 x 2.1 Design and Feasibility Study of Magnetostrictive Microcantilever (MSMC) as a Biosensor Platform................................................................................................ 42 2.2 Design and Feasibility Study of Magnetostrictive Particle (MSP) as a High Performance Biosensor Platform .......................................................................... 43 2.3 Synthesis of Amorphous Fe-B Alloy and Fe-B Nanowires for Magnetostrictive Biosensor Platform Application............................................................................ 43 CHAPTER 3 CHARACTERIZATION OF RESONANCE BEHAVIOR OF MAGNETOSTRICTIVE AW DEVICES.................................................................................................................. 45 3.1 Set-up (A) Based on Helmholtz Coil and Lock-in Amplifier................................. 45 3.1.1 Configuration of the Setup............................................................................... 45 3.1.2 Resonance Frequency and Q Value Determination......................................... 52 3.2 Set-up (B) Based on Network Analyzer ................................................................. 59 3.2.1 Configuration of the Setup............................................................................... 59 3.2.2 Resonance Frequency and Q Value Determination......................................... 62 References..................................................................................................................... 67 CHAPTER 4 DESIGN AND FEASIBILITY STUDY OF BIOSENSOR BASED ON MAGNETOSTRICITVE MICROCANTILEVER........................................................... 68 4.1 Introduction............................................................................................................. 68 4.2 Design and Operation Principle of MSMC as Biosensor Platform ........................ 69 4.2.1 Configuration and Operation Principle of MSMC........................................... 69 4.2.2 Natural Resonance Behavior of Magnetostrictive Microcantilever................. 72 4.2.2.1 Flexural (Bending) Resonance Frequency of Cantilever with One End Fixed ............................................................................................................... 72 4.2.2.2 Determination of Eigenvalues for Flexural Resonance Oscillation....... 75 4.2.3 Mass Sensitivity of MC & MSMC .................................................................. 76 4.2.4 Comparison of Mass Sensitivity of MSMC and Other MCs ........................... 79 4.2.5 Comparison of MSMC with Current MCs....................................................... 83 4.3 Experimental........................................................................................................... 85 4.3.1 Fabrication of MSMCs .................................................................................... 85 4.3.2 Characterization of Resonance Behavior of MSMC ....................................... 85 xi 4.3.3 Characterization of Mass Sensitivity of MSMCs ............................................ 86 4.3.4 Demonstration of In-situ Yeast Cells Detection Using the MSMC Sensors ... 86 4.3.5 Detection of Bacillus Anthracis Spores Using MSMC Biosensor................... 87 4.4 Results and Discussions.......................................................................................... 88 4.4.1 Resonance Behavior of MSMC in air.............................................................. 88 4.4.2 Resonance Frequency Changing with the MSMC Dimensions....................... 91 4.4.3 Stability of the MSMC..................................................................................... 93 4.4.4 MSMC Performance in Liquid ........................................................................ 96 4.4.5 MSMC Array ................................................................................................... 98 4.4.6 Characterization of the Mass Sensitivity of MSMC...................................... 100 4.4.7 Yeast Detection in Water............................................................................... 103 4.4.8 Detection of Bacillus Anthracis Spore Using MSMC-based Biosensor........ 105 4.5 Conclusion ............................................................................................................ 111 References................................................................................................................... 112 CHAPTER 5 DESIGN AND FEASIBILITY STUDY OF MAGNETOSTRICTIVE PARTICLES AS BIOSENSOR PLATFORM............................................................................................ 114 5.1 Introduction........................................................................................................... 114 5.2 Design and Operation Principle of Magnetostrictive Particles (MSPs)................ 117 5.2.1 Configuration of MSPs and Their Resonance Behavior................................ 117 5.2.2 The Mass Sensitivity (S m ) of MSPs............................................................... 121 5.2.2.1 Mass Sensitivity (S m ) of MSPs with Uniform Mass Load................... 121 5.2.2.2 Mass Sensitivity (S m ) of MSP with Non-uniform Mass Load............. 126 5.2.3 Operation Principle of MSPs as Biosensor Platform..................................... 132 5.2.4 Multiple-sensor Approach Based on MSPs ................................................... 135 5.2.5 Multiple-target Approach Based on MSPs .................................................... 137 5.3 Experimental......................................................................................................... 139 5.3.1 MSP Sensor Platforms and the Material........................................................ 139 5.3.2 Characterization of Resonance Behavior of MSPs........................................ 139 5.3.3 Mass Sensitivity of the MSPs ........................................................................ 140 5.4 Results and Discussion ......................................................................................... 141 xii 5.4.1 Resonance Frequency and Q value of MSPs ................................................. 141 5.4.2 MSP Performance in Liquid .......................................................................... 144 5.4.3 Influence of DC Bias on Resonance Behavior of MSPs................................ 146 5.4.4 Influence of AC Driving Field on Resonance Behavior of MSPs................. 148 5.4.5 Resonance Frequency Change with MSP?s Lengths ..................................... 151 5.4.6 Sensitivity of MSPs with Uniform Mass Load.............................................. 152 5.4.7 Mass Sensitivity of MSPs with Non-uniformed Mass Load ......................... 155 5.4.8 Demonstration of Multiple-Sensor Approach................................................ 159 5.4.9 Demonstration of Multi-target Detection....................................................... 161 5.5 Conclusion ............................................................................................................ 163 References................................................................................................................... 164 CHAPTER 6 SYNTHESIS OF AMORPHOUS IRON-BORON ALLOY FOR MAGNETOSTRICTIVE BIOSENSOR PLATFORM APPLICATION ....................... 165 6.1 Introduction........................................................................................................... 165 6.2 Experimental......................................................................................................... 168 6.2.1 Materials ........................................................................................................ 168 6.2.2 Preparation of Deposition Solution................................................................ 169 6.2.3 Electrochemical Deposition of Amorphous Fe-B Thin Film......................... 169 6.2.4 Characterization of the Fe-B Thin Film......................................................... 172 6.2.4.1 The Microstructure Determination Using X-ray Diffraction............... 172 6.2.4.2 Composition Determination using AES and XPS ............................... 173 6.2.4.3 Magnetization Loop Measurement ...................................................... 175 6.2.4.4 Resonance Behavior Characterization of Fe-B Alloy.......................... 176 6.2.4.5 Thermal Stability of the Fe-B Thin Film ............................................. 176 6.3 Results and Discussion ......................................................................................... 177 6.3.1 Electrochemical Deposition of Amorphous Fe-B Alloy................................ 177 6.3.2 Microstructure of the Fe-B Thin Films.......................................................... 179 6.3.3 Composition of the Fe-B Film ....................................................................... 183 6.3.4 Magnetic Properties of the Fe-B films........................................................... 188 6.3.5 Resonance Behavior of deposited Fe-B Alloy............................................... 192 xiii 6.3.6 Thermal Stability of the Amorphous Fe-B Films .......................................... 196 6.3.7 Feasibility Study of Micro-Magnetostrictive Particle (MSP) Fabrication Based on Fe-B Alloy ............................................................................................... 199 6.3.7.1 ?Top Down? Fabrication Process ......................................................... 200 6.3.7.2 ?Bottom Up? Fabrication Process......................................................... 205 6.5 Conclusion ............................................................................................................ 208 References................................................................................................................... 209 CHAPTER 7 SYNTHESIS OF AMORPHOUS FE-B NANOWIRES AS MAGNETOSTRICTIVE BIOSENSOR PLATFORM............................................................................................ 211 7.1 Introduction........................................................................................................... 211 7.2 Experimental......................................................................................................... 216 7.2.1 Fabrication of Amorphous Fe-B Nanowires and Nanowire Arrays .............. 216 7.2.2 Characterization of Fe-B Nanowires and Nanowire Arrays.......................... 219 7.2.2.1 Morphology of Fe-B Nanowires and Nanowire Arrays ...................... 219 7.2.2.2 Microstructure of Fe-B Nanowires ...................................................... 219 7.2.2.3 Composition of Fe-B Nanowires ......................................................... 219 7.2.2.4 Thermal Stability of Fe-B Nanowairs.................................................. 220 7.2.2.5 Magnetic Properties of Fe-B Nanowire Arrays ................................... 220 7.3 Results and Discussions........................................................................................ 222 7.3.1 Morphology of Fe-B Nanowires.................................................................... 222 7.3.2 Microstructure of Fe-B Nanowires ................................................................ 225 7.3.3 Composition of Fe-B Nanowires ................................................................... 228 7.3.4 Thermal Stability of Fe-B Nanowire ............................................................. 231 7.3.5 Magnetic Properties of Fe-B Nanowires........................................................ 234 7.4 Conclusion ............................................................................................................ 238 References................................................................................................................... 240 CHAPTER 8 PERSPECTIVES ............................................................................................................ 246 xiv LIST OF FIGURES 1-1 Scheme of traditional visible cell counting method for microbial population detection...................................................................................................................3 1-2 DNA duplication by polymerase chain reaction (PCR)...........................................5 1-3 Schematic diagram of a typical biosensor ...............................................................6 1-4 Scheme of immuno-FET biosensor. V G and V D are Gate voltage and Drain voltage, respectively, for generating an initial current flow. The antigen-antibody interaction on the Gate surface associates with the potential change at the Gate, which can be detected and used for bio-sensing......................................................8 1-5 Operation principle of the SPR biosensor..............................................................10 1-6 Scheme of vibration amplitude vs. frequency of an AW sensor, where f 0 and f are the resonance frequency of the sensor at zero mass load and at certain mass load respectively ............................................................................................ 12 1-7 Scheme of Joule magnetostriction .........................................................................20 1-8 Magnetization hysteresis loop of a ferromagnetic material...................................23 1-9 Hysteresis loop for Joule magnetostriction............................................................23 1-10 Response of a magnetostrictive material under different DC bias ........................25 1-11 Scheme of the conventional directions for a polarized magnetostrictive material. ??........................................................................................................................26 3-1 Schematic of set-up (A) based on a Helmholtz coil and a lock-in amplifier.........46 3-2 Picture of set-up (A) based on a Helmholtz coil and a lock-in amplifier ..............48 3-3 Background signal detected by a pair of pickup coils ...........................................49 3-4 Experimental output (black curve) of a magnetostrictive cantilever measured using setup A, and its corresponding corrected results (wine curve) based on Equation (3-2) and (3-3). The size of the cantilever is 2.8 mm x 1 mm x 35 ?m...... ...............................................................................................................................51 3-5 Typical resonance spectrum of a magnetostrictive AW device: (a) the typical output of a magnetostrictive particle; (b) the typical output of a magnetostrictive microcantilever ......................................................................................................53 xv 3-6 Baselines for fitting the resonance frequency peak from the amplitude signal. The baseline is chosen through two points; one of the points is where the amplitude reaches minimum...................................................................................................54 3-7 Lorentz fitting curve for the measured amplitude resonance peak of a magnetostrictive particle. Solid black curve is the measured signal, while dashed blue curve is the fitting curve.................................................................................55 3-8 Lorentz fitting results for the detected phase signal of a magnetostrictive cantilever measured in air. The black solid line is the detected signal, while the dashed blue line is the fitting curve .......................................................................58 3-9 A stimulus/response test system ............................................................................59 3-10 Definitions of S-parameters ...................................................................................60 3-11 Set-up (B) based on a network analyzer ................................................................61 3-12 Magnitude and phase curve of S 11 parameter changing with the frequency of a magnetostrictive strip, whose size is 10 mm x 1 mm x 30 ?m..............................64 3-13 The equivalent impedance based on S 11 data shown in Fig. 3-12. The conversion is based on Equation (3-8) .....................................................................................64 3-14 Lorentz fitting results for the measured S 11 magnitude peak a magnetostrictive strip in the size of 6.4 mm x 1 mm x 30 ?m. The black solid line is the detected signal, while the dashed blue line is the fitting curve............................................65 4-1 Structure of the magnetostrictive microcantilever (MSMC), where L and W are the length and width of the cantilever, while h is the thickness of the MSMC beam, including the magnetostrictive layer and the inactive layer. (a) MSMC under zero magnetic field; (b) MSMC under an external magnetic field................................70 4-2 Operating/detecting principle of the MSMC as a biosensor platform ...................71 4-3 Rectangular cantilever beam with a length of L, width of w and thickness of h, rigidly fixed at one end and free at the other .........................................................73 4-4 Curves of function tan(?/2), -tan(?/2) and coth(?/2). The Eigenvalues for bending resonance oscillation of a cantilever beam with one end fixed can be graphically obtained from the intersections of the functions....................................................75 4-5 MSMC (a) with a uniform mass load on the surface; and (b) with a point mass load at the cantilever tip.........................................................................................78 4-6 (a) Theoretical mass sensitivity calculated based on Equation (4-18) for MCs made of Metglas TM 2826 MB, PZT4, PZT5A, and Si. The material properties used in calculation are listed in Table 4-2, while the ratio of MCs? length to width is chosen as 5. (b) Theoretical mass sensitivity calculated based on Equation (4-18) and (4-19) for MCs made of Metglas TM 2826 MB ................................................82 xvi 4-7 The phase difference between the driving magnetic field and the magnetic signal from a MSMC. The size of the MSMC is 4 mm x 1.5 mm x 30 ?m. For this MSMC, the first six harmonic vibration modes were observed ............................89 4-8 Resonance frequency of the fundamental and the first flexural modes vs 1/L 2 for the MSMCs with a thickness about 30 ?m, where L is the length of the MSMCs. The solid triangles and solid squares are the experimental results for the fundamental and the first flexural mode, while the solid lines are results of theoretical calculation ............................................................................................93 4-9 Resonance frequency of a MSMC with size of 3.05 mm x 1 mm x 30 ?m measured in air over 24 hours................................................................................95 4-10 MSMC performance in liquid. (a) Fundamental mode resonance spectra for a MSMC in different media. (b) First flexural mode resonance spectra for a MSMC in different media...................................................................................................97 4-11 Scheme of a MSMC array characterized using setup A ........................................99 4-12 Spectrum of a MSMC array consisting of three MSMCs....................................100 4-13 Mass sensitivity characterization of an MSMC (4mm x 12.mm x 30?m): (a) The fundamental harmonic mode resonance spectra for the MSMC with different mass loads; (b) The first harmonic mode resonance spectra for the MSMC with different mass loads; (c) ?f versus ?m of the MSMC for the first and second harmonic resonance ..................................................................................... 101/102 4-14 Response of fundamental harmonic mode for real-time yeast cells detection using MSMCs: (a) frequency shift with time after MSMC-A and MSMC-B immersed in yeast suspension with concentration of 1 mg/ml and 2 mg/ml, respectively; (b) the picture of yeast cells on the MSMC-B surface after the sensor was tested in 2 mg/ml yeast cell suspension. (c) the picture of yeast cells on the MSMC-A surface after the sensor was tested in 1 mg/ml yeast cell suspension ..............................104 4-15 Real-time resonance frequency shifts as the MSMC/Phage biosensor exposed to B. anthracis suspension with increasing concentration. The MSMC biosensor is in the size of 2.7 mm x 1.0 mm x 35 ?m .................................................................107 4-16 SEM images of the MSMC surface after the biosensor was exposed to the spore suspension with increasing concentration. (a) at the tip of the cantilever beam; (b) at the middle of the cantilever beam....................................................................108 4-17 The saturated resonance frequency shifts as the MSMC/phage biosensors, in the size of 2.7 mm x 1 mm x 35 ?m, were exposed to spore suspensions with different concentrations ......................................................................................................109 5-1 Magnetostrictive particle (MSP) configuration. (a) MSP in strip-shape and (b) MSP in bar-shape.................................................................................................118 xvii 5-2 Theoretical S m of MSPs and MSMCs made of a commercially available magnetostrictive alloy 2826MB TM . For strip-shape MSPs, the ratio of the length to width is set 5 (? = 5), and the ratio of the width to thickness is set 10 (? = 10); for bar-shape MSPs, the ratio of the length to diameter is set 10 (? = 10). For MCs, the ratio ? of the length to width is set 5..............................................................125 5-3 (a) Amplitude profile of displacement of MSP for the first (n = 1) harmonic mode; (b) schematic illustration for the deformation of the MSP at different times for the first (n = 1) harmonic mode resonance ................................................................127 5-4 (a) Amplitude profile of displacement of MSP for the second (n = 2) harmonic mode; (b) schematic illustration for the deformation of the MSP at different times for the second (n = 2) harmonic mode resonance................................................128 5-5 (a) Amplitude profile of displacement of MSP for the third (n = 3) harmonic mode; (b) schematic illustration for the deformation of the MSP at different times for the third (n = 3) harmonic mode resonance....................................................129 5-6 Node points of free-standing longitudinal resonance for different harmonic modes. The red dot lines represent the location of the nodes on the MSP.......................130 5-7 Operating/detecting principle of a MSP sensor ...................................................133 5-8 Scheme of multiple-sensor approach based on MSPs .........................................136 5-9 Illustration of the response of the detection using multiple-sensor approach......137 5-10 Scheme of multiple-target approach based on MSP sensors to detect different target simultaneously ...........................................................................................138 5-11 The amplitude (a) and phase (b) signal measured by setup A from a MSP. The size of the MSP is 24.9 mm x 5.0 mm x 30 ?m. For this MSP, first three harmonic vibration modes were presented...........................................................................141 5-12 The resonance spectra of a MSP measured using setup B. The size of the MSP is 10.0 mm x 2.0 mm x 30 ?m. For this MSP, first three harmonic vibration modes were presented .....................................................................................................142 5-13 Illustration of the position of MSP and pickup coil during multiple modes measurement ........................................................................................................143 5-14 MSP performance in air and water for (a) the first harmonic mode, (b) second harmonic mode, and (c) third harmonic mode.....................................................145 5-15 The resonance spectra for the first harmonic mode of a MSP, which were measured under same AC field (amplitude is 0.16 Oe MRS), but different DC bias using setup A. The size of the MSP is 25 mm x 5 mm x 30 ?m .........................146 5-16 Magnetostriction (?) of a typical magnetostrictive material changing with external magnetic field (H)................................................................................................147 5-17 (a) Spectra for the first resonance mode of a MSP, in the size of 24.3 mm x 5 mm x 30 ?m, measured under the different AC driving magnetic fields. (b) Normalized spectra for the results shown in (a) ..................................................149 xviii 5-18 Resonance frequencies for the first harmonic mode of MSPs with different lengths..................................................................................................................152 5-19 Resonance frequency shift caused by uniform mass load for the MSPs with different dimensions. The first harmonic mode was used here ...........................153 5-20 Mass sensitivities versus lengths for MSPs with different dimensions. Solid triangles are experimental data; solid lines are the results obtained by the theoretical simulation. The ratio of the length to width of MSPs shown in this picture is 10..........................................................................................................155 5-21 (a) Non-uniform mass loads on the MSP surface; (b) The MSP dimensions and the mass load sizes for characterizing the mass sensitivity with non-uniform mass loads .....................................................................................................................156 5-22 Shifts in MSP resonance frequencies caused by the mass loads at different locations, where (a) is for first harmonic mode; (b) is for second harmonic mode??.............................................................................................................158 5-23 Demonstration of multiple sensor approach ........................................................161 5-24 Demonstration of the multiple-target approach ...................................................162 6-1 Electrochemical cell for Fe-B thin film deposition..............................................171 6-2 Illustration of Fe-B film specimens used for magnetization loop measurement .175 6-3 Cyclic voltammogram of Fe-B deposition on Cu/Cr/glass working electrode....178 6-4 XRD patterns of Fe-B thin films on Cu/Cr/glass substrate deposited by plating baths listed in Table 6-2.......................................................................................180 6-5 XRD patterns of Fe-B films deposited on the Cu/Cr/glass substrate from plating baths C, D, E and F???????????????????????..181 6-6 XRD pattern (40 o ~ 50 o ) of the film deposited from bath C and the fitting results base on Lorentz function. The black solid curve is the experimental data, while the red dotted curve and green dashed curve are the fitting results.....................182 6-7 The XPS spectrum of the Fe-B film surface as deposited. The Fe-B film was deposited from plating bath D under a current density of 5 mA/cm 2 ..................183 6-8 XPS spectra of B1s for deposited Fe-B film surface and about 5 nm, 15 nm, and 35 nm from the surface. The Fe-B film was deposited from plating bath D under a current density of 5 mA/cm 2 ................................................................................185 6-9 SEM image of the cross-section of the Fe-B film deposited from bath D under a current density of 5 mA/cm 2 ................................................................................186 6-10 Compositions of Fe-B film along its thickness direction. The film was deposited from bath D under a current density of 5 mA/cm 2 ...............................................187 6-11 Hystersis loops of Fe-B films deposited from bath (a) D, (b) E and (c) F under a current density of 5 mA/cm 2 ................................................................................189 xix 6-12 Hystersis loops of Fe-B film deposited from bath D with different thicknesses? ?..........................................................................................................................191 6-13 Resonance frequency versus reciprocal of the length for the resonators made of Fe-B film on Cu/Cr substrate. The films were deposited from baths D, E, and F under a current density of 5 mA/cm 2 ...................................................................194 6-14 Q values for the strips made of 2826 MB ribbon and 2826MB ribbon with Cu/Cr bilayer ..................................................................................................................196 6-15 XRD patterns of Fe-B films deposited from bath D after annealing at different temperatures.........................................................................................................197 6-16 Flow chart of ?top down? microelectronics fabrication of MSP..........................200 6-17 Images for fabricated MSPs using ?top down? method. The designed size for MSPs shown in (a) is 300 ?m x 60 ?m x 1 ?m, while the designed size for MSPs shown in (b) is 110 ?m x 20 ?m x 1 ?m..............................................................202 6-18 Slope can be observed on the MSP edge for fabricated MSPs............................203 6-19 (a) Defects in MSPs caused by the cracks in the original deposited Fe-B film; (b) Defects on the edge of fabricated MSPs due to the impurities in the original film. ..........................................................................................................................?204 6-20 Flow chart of ?bottom up? microelectronics fabrication of MSP.........................205 6-21 (a) SU-8 pattern on the Au electrode/substrate; (b) the MSP deposited in the opening of patterned SU-8. The MSPs were deposited from bath D under current density of 5 mA/cm 2 ............................................................................................207 7-1 Template synthesis of Fe-B nanowires................................................................218 7-2 Scheme of rectangular sample with several layers nanowires in the diameter of 200 nm and the length of 4.5 ?m.........................................................................221 7-3 SEM image for multilayer sample B ...................................................................221 7-4 SEM images for Fe-B nanowires and nanowire arrays fabricated in the template with the pore size of (a) 200 nm, (b) 100 nm, (c) 50 nm, and (d) 10 nm; and focused SEM images of individual nanowire fabricated in the template with the pore size of (e) 200 nm, and (f) 50 nm ................................................................223 7-5 (a)TEM image of the Fe-B nanowires fabricated in template with pore diameter of 50 nm; (b) Electron diffraction patterns of this nanowire; (c) Focused TEM image of the Fe-B nanobars fabricated in template with pore diameter of 50 nm .........226 7-6 (a) TEM image of a Fe-B nanowire fabricated in template with pore diameter of 10 nm; (b) Selected area electron diffraction patterns of this nanowire; (c) Focused TEM image of Fe-B nanowire fabricated in template with pore diameter of 10 nm ...............................................................................................................227 xx 7-7 (a) Individual nanowire deposited in the template with pore diameter of 50 nm under a current density of 5 mA/cm 2 ; (b) Composition distribution of point 6 on this nanowire along its diameter direction; (c) Composition distribution of this nanowire along its length direction......................................................................230 7-8 Composition distribution along the length direction of a nanowire deposited in the template with pore diameter of 50 nm under a current density of 10 mA/cm 2 ....231 7-9 The TEM image and diffraction pattern of Fe-B nanowires fabricated in template with pore diameter of 50 nm and annealed at 200 o C for 1 hour.........................232 7-10 The TEM image and diffraction pattern of Fe-B nanowires fabricated in template with pore diameter of 50 nm and annealed at 300 o C for 1 hour.........................232 7-11 (a) The TEM image and diffraction pattern of the Fe-B nanowire fabricated in template with pore diameter of 50 nm and annealed at 400 o C for 1 hour. (b) The comparison of standard electron diffraction patterns of BCC Fe and FeO with the electron diffraction pattern from (a) ....................................................................233 7-12 Magnetization hystersis loops of nanowires with diameter of 200 nm (a, b), 100 nm (c, d) and 50 nm (e)........................................................................................237 xxi LIST OF TABLES 1-1 Properties of some magnetostrictive materials ......................................................30 3-1 Fitting results of amplitude data based on Lorentz function using different baselines.................................................................................................................56 3-2 Fitting results of phase data based on Lorentz function using different fitting parameters..............................................................................................................58 3-3 S 11 magnitude peak fitting results based on Lorentz function using different fitting parameters..............................................................................................................66 4-1 Graphically obtained Eigenvalues for the first ten flexural resonance modes.......76 4-2 Mechanical properties of commercially available magnetostrictive alloy, piezoelectric ceramic and silicon...........................................................................81 4-3 Comparison of different microcantilever-based sensor platforms.........................84 4-4 Theoretical/experimental resonance frequency ratios of the nth-mode resonance frequency to the fundamental resonance (n = 0) frequency for the first to third mode, and the ratios of the nth-mode resonance frequency to the first resonance (n = 1) frequency for the second and third mode .......................................................91 4-5 Fitting results of the fundamental mode resonance frequency based on Lorentz function over different frequency ranges for the MSMC used in stability study ..95 4-6 Fitting results of the first mode resonance frequency based on Lorentz function over different frequency ranges for the MSMC used in stability study.................96 4-7 MSMC performance in liquid................................................................................98 4-8 Fitting results of resonance frequencies of MSMC-A in pure water based on Lorentz function over different frequency ranges ...............................................104 4-9 Fitting results of resonance frequencies of MSMC-B in pure water based on Lorentz function over different frequency ranges ...............................................105 5-1 MSP performance in air and water ......................................................................144 5-2 Resonance behavior of a MSP (25 mm x 5 mm x 30 ?m) measured under same AC field, but different DC biases ........................................................................148 5-3 Resonance behavior of a MSP (24.3 mm x 5 mm x 30 ?m) measured under same DC bias, different AC fields ................................................................................150 xxii 5-4 Fitting results based on Lorentz function with different fitting ranges for the resonance peak shown in Fig. 5-15 measured under AC field with amplitude of 0.32 Oe.................................................................................................................151 5-5 Mass sensitivity of MSPs with different dimensions...........................................153 5-6 Mass sensitivities of the MSP for mass loads at different locations....................159 6-1 Properties of bulk Fe 80 B 20 alloy prepared by liquid-quenching...........................168 6-2 Composition of Baths for Fe-B thin film deposition ...........................................171 6-3 Compositions of Fe-B film surface, and about 5 nm, 15 nm and 35 nm from the surface. The film was deposited from plating bath D under a current density of 5 mA/cm 2 ................................................................................................................184 6-4 Compositions of Fe-B film deposited from bath D under a current density of 5 mA/cm 2 along its thickness direction ..................................................................187 6-5 Coercivity and squareness of amorphous Fe-B films ..........................................190 6-6 Coercivity and squareness of amorphous Fe-B films deposited from bath D with different thicknesses.............................................................................................192 6-7 Resonance frequency and Q value for the resonators made from Fe-B alloy deposited from bath D before and after heat treatment........................................199 7-1 Composition of amorphous Fe-B alloy deposition solution ................................217 7-2 Coercivity for amorphous Fe-B nanowires..........................................................238 7-3 Coercivity of multi-layer Fe-B nanowires along different directions..................238 1 CHAPTER 1 INTRODUCTION Micro-organisms, such as bacteria and virus, are widely found in the nature and the environment, such as food, soil, water and the intestinal tracts of humans and animals. Although most of these micro-organisms have essential functions in nature and beneficial relations with humans, certain bacteria or viruses are harmful and can cause serious and infectious diseases. It is estimated that infectious diseases cause about 40% of the total annual deaths world-wide [1]. In the United States, approximate 76 million food borne illnesses occur each year, which accounts for 325,000 hospitalizations and 5,000 deaths [2]. The United States Department of Agriculture (USDA) reports that medical costs and productivity losses due to major food pathogens result in $2.9~$6.7 billion per year [3, 4]. The recent outbreak of E. coli O157 infections related to spinach was large and deadly, which shows that better control and prevention methods are needed in the food industry [5]. Since the 9-11 terrorist attack, the requirement of monitoring threat agents for biological weapons is growing. The spread of bio-threat microorganisms can cause high fatality rate diseases throughout a huge population. Prevention of microbial diseases depends on effective and rapid detection of various pathogenic micro-organisms in food, clinic medicine and the environment. Therefore, analytical technologies for detecting and quantifying the presence of a small amount of biological threat agents in a real-time manner are urgently needed in food safety, medical diagnostics, environmental 2 monitoring, and in public safety/security areas. Effective biological-detection methods should meet a number of criteria, such as response time, sensitivity and specificity [1, 6, 7]. Rapid detection is critical for disease control. Meanwhile, extremely sensitive detection methods are required since for some pathogenic organisms, such as E. coli O157:H7, the presence of even several cells may be an infectious dose [1, 6, 8]. Additionally, since most of micro-organisms exist in liquid environments, detection methods that can be directly carried out in liquid are highly desired. At the same time, from the industry point of view, the microbiological detection methods should be inexpensive, robust and easy to operate [1, 9]. This research focuses on developing novel high performance biosensor platforms for effective biological detection. Two kinds of novel acoustic wave (AW) sensor platforms based on the magnetostrictive effect ? magnetostrictive microcantilever (MSMC) and magnetostrictive particle (MSP) ? are introduced as high performance biosensors. The results demonstrate the feasibility of developing high performance biosensors using MSP and MSMC. As the AW sensor platforms, the sensitivity of MSP and MSMC increases with reducing size. To develop an extremely sensitive platform, MSPs in micro/nano scale were fabricated. The MSMC and MSP have the potential to significantly advance the biological analysis technology and allow us to better protect the public. 1.1 Conventional Bacteria Detection Techniques A. Visible Cell Counting Technique Traditional method for enumerating bacteria is based on the ability of bacteria in a sample to grow into visible colonies on a defined nutrient medium [10-12]. As shown in Fig. 1-1, basically in this method, the pre- enriched cell sample is spread onto agar 3 plates that are subsequently incubated at a growth temperature. During incubation, bacteria cells grow into visible colonies. Theoretically, a colony is derived from a single bacteria cell. The count of cells in the original sample can be obtained by counting the colony number and multiplying it by dilution factor [12]. While this method can be very sensitive, it is time-consuming due to the long incubation period (typically 24-72 hours). Also, colony enumeration can not provide a definitive identification for unknown biological cells. Thus, complex tests are required for any confirmed identification, which requires well trained personnel [1, 12]. This method is not suitable for quick field testing. Fig. 1-1. Scheme of traditional visible cell counting method for microbial population detection [13]. Sample collection Incubation Enumeration Plate Culture Method Normally 24 hours Colony 4 B. Polymerase Chain Reaction (PCR) The polymerase chain reaction (PCR) method has been extensively used in microbial detection since it was introduced in 1985 [6, 14-18]. By duplicating the DNA (deoxyribonucleic acid) of a small quantity of target pathogens over billions of times, PCR provides a sensitive method to determine the presence of small amounts of bacteria. Fig. 1-2 illustrates the principle of PCR. The DNA duplication involved three steps: denaturation, primer annealing and primer extension. In the first step, the DNA is heated and the double strand separates to two single stranded DNA; in the second step, at lower temperature, the primers, which are short DNA strands and uniquely match to portion of the DNA to be amplified, bind to the single stranded DNA; in the third step, the DNA- primers are incubated at higher temperature with DNA polymerase and a new copy of target DNA forms [15, 19]. By repeating PCR cycles, a large amount of identical copies of target DNA can be generated and detected. PCR is highly sensitive. However, the PCR method requires rigorous sample preparation, complex reactive components of limited shelf life, precise temperature regulation, sophisticated hardware, a complex detection process, and trained personnel [1, 6, 15]. 5 Fig. 1-2. DNA duplication by polymerase chain reaction (PCR) [20]. 1.2 Review of Current Biosensor Techniques Conventional microbiological tests must be undertaken in laboratories because of complicated testing procedures, expensive equipment and the requirement for qualified operators. Therefore, in the last decade, there has been intensive research focused on the development of portable, rapid, specific and sensitive techniques to suit the purpose of real-time and in-field detections. Biosensor techniques show particular promise to meet these requirements, and thus has been intensively studied [1]. A biosensor is an analytical device incorporating a biological sensing element (probe/receptor) to a transducer (platform) system. The specific biological reaction between the sensing element and the interested species agent causes a physical/chemical change at the bio-interface. The transducer/platform converts the observed change (physical or chemical) into a measurable signal, which is sent to an output system to Denaturation Annealing Extension 6 be amplified, processed and displayed [6]. A schematic diagram of a typical biosensor is shown in Fig. 1-3 [6]. The sensing elements can be DNA/RNA, proteins, enzymes, antibodies, or phages [6, 21]. According to their transducer/platform type, biosensors can be divided into electrochemical, optical, thermal, and mass sensors [1, 9, 22, 23]. Fig. 1-3. Schematic diagram of a typical biosensor. 1.2.1 Electrochemical Biosensors At present, electrochemical technology plays a very important role in clinical diagnosis, biochemical analysis and environmental monitoring. Electrochemical sensors obtain the analytical information from the electrical signal that results from the interaction of the target analyte and the recognition elements. Based on the electrical signals that the transducer detects, the electrochemical sensors can be classified into: conductimetric/impedimetric, potentiometric and amperometric [9]. When micro-organisms metabolize uncharged and weakly charged substances, such as carbohydrates, fats or proteins, into highly charged end products, such as organic acids, fatty acids, and amino acids, both conductance and capacitance of the medium change, causing the change of the impedance of the medium [1, 9]. The amount of the Enzymes Antibodies Nucleic acids Cells Phages Electrochemical Optical Acoustic Thermal Sample Bio-receptor Transducer Output system Signal 7 charged metabolites is proportional to the metabolization rate. Conductimetric/impedimetric biosensors detect the microbial concentration of the sample by measuring the change in electrical conductivity/impedance [1, 9, 24]. This method has been used to detect specific food pathogens [25, 26]. In potentiometric biosensors, the biological active materials, such as enzymes or antibodies, are immobilized on the surface of the electrode. When the target antigens in a sample react with the corresponding active materials, the potential between the enzyme/antibody immobilized electrode and a reference electrode will change [27]. The potential difference between the electrodes is proportional to the logarithms of the concentration of the analyte species [1, 27]. Based on this principle, potentiometric biosensors, such as Immuno-Field-Effect-Transistor (FET) sensor and light addressable potentiometric sensor (LAPS), are developed for detection of pathogenic bacteria [1, 28- 31]. Fig. 1-4 shows a scheme of an Immuno-FET biosensor. A FET is a semiconductor device, in which the conductivity of the n-channel region in the p-type silicon is controlled by the potential at the oxide insulator (Gate) and is measured as a voltage (V D ) is applied between two n-type silicon (Source and Drain). The biological action between the interested species and the biological active materials immobilized on the Gate surface associates with the potential change at the Gate, which can be detected and used for bio- sensing [1, 9]. 8 Fig. 1-4. Scheme of immuno-FET biosensor. V G and V D are Gate voltage and Drain voltage, respectively, for generating an initial current flow. The antigen-antibody interaction on the Gate surface associates with the potential change at the Gate, which can be detected and used for bio-sensing. Instead of measuring the potential change, amperometric biosensors measure the current generated by the chemical/biological reaction of an electroactive species under an applied potential, which is proportional to the concentration of the electroactive species in the solution [9, 32]. Compared to the logarithmic concentration dependence in potentionmetric system, the amperometric system has a linear relationship. This makes amperometric biosensors more suitable for bacterial detection [1]. However, amperometric sensors suffer from poor selectivity, because their selectivity is governed by the redox potential of the electroactive species in the sample solution. Therefore, the current measured by the sensor can include the contributions of several chemical species [9]. n-Si (Drain) n-Si (Source) channel V D Gate antigen antibody p-Si V G 9 To obtain adequate sensitivity and selectivity, electrochemical biosensors employ specific enzymes, which are only active in a narrow temperature or pH range, for the capture and catalytic generation of the product [9]. 1.2.2 Optical Biosensors Optical biosensors are very attractive for their direct (label-free) and quick detection of bacteria [6]. The direct fluoresce technique examines the fluorescent components of the bacteria. Under ultraviolet light, certain biospecies can emit light, usually with a lower wavelength. This phenomenon is known as fluorescence [1]. Fluorescence biosensors identify bacteria by measuring the characteristic fluorescent radiation of cells under an UV source [33, 34]. This approach is simple and rapid. However, their major limitation is that they only can detect the bacteria which contain or produce fluorescent components [1]. Some enzyme-catalyzed biochemical reactions result in emitting photons, a reaction which is known as bioluminescence [1]. Luminescence biosensors employ this phenomenon to detect a wide range of microorganisms [35-38]. The luminescence biosensors have very high specificity. However the main disadvantage is the relatively long response time which needs several hours, as well as the lack of sensitivity when very low numbers of bacterial cells are to be detected [1, 22]. Currently, the most attractive optical biosensor transducer is based on surface plasmon resonance (SPR). SPR is a quantum electro-optical phenomenon. When light hits the surface of a metal, energy carried by photons of the light can excite resonant oscillation of electrons on the surface of the metal. The ?excited? free electron portion of the surface metal layer is called plasmon [6]. The SPR is sensitive to the environment 10 of the metal surface [9]. A SPR biosensor is made of two types of materials with different refractive indices; which normally are a metal (gold) and a glass. The incident light is reflected at the interface of these two materials. Utilizing total internal reflection, SPR is excited along the metal surface at a specific incident angle (? SPR ), at which the intensity of the reflected light reduces to minimum. The change of the condition of the metal surface, such as antibody-antigen binding, affects the SPR in the metal layer and results in the changes of ? SPR [9, 39]. Therefore, SPR biosensors are able to detect interested species binding to receptors immobilized on the transducer surface by measuring the shift of ? SPR . Fig. 1-5 shows the operation principle of the SPR biosensor. Fig. 1-5. Operation principle of the SPR biosensor. SPR biosensors have been widely used in detecting food borne pathogens and chemical and biological agents [40-42]. SPR biosensors can provide easy to use, real- time, and label-free detection. However, at present, SPR sensors are mainly used in the Antigen Antibody Gold layer Glass layer Prism Monochromatic incident light Reflected light ? 11 laboratory. Also, they are relatively expensive, and have difficulty measuring target species in very low concentration environments [1, 39]. 1.3 Acoustic Devices In the last decade, acoustic wave (AW) devices attracted a great deal of attention among the biosensor platforms, since, as a sensor platform, they can offer many advantages, such as a real-time detection, simplicity of use, and cost effectiveness [6, 43- 45] . 1.3.1 Operating Principle In AW devices, acoustic waves are generated causing oscillation of the device. The operating principle of all AW sensor platforms is the same. That is, when an AW device is operated as a resonator, its resonance frequency changes with the mechanical load on the device surface, as shown in Fig. 1-6. For developing a biosensor, the AW device is usually immobilized with a bio-molecular recognition layer, such as antibody and phage, on its surface to react with the target species [21, 45], which results in a change in the recognition layer, such as the mass and elastic property. Thus, the presence of target species can be detected by monitoring the shift in device?s resonance frequency. 12 Fig. 1-6. Scheme of vibration amplitude vs. frequency of an AW sensor, where f 0 and f are the resonance frequency of the sensor at zero mass load and at certain mass load respectively. The comparison of different AW devices can be found in several review articles [45- 49]. The mass sensitivity of an AW sensor is the most important parameter to characterize its performance. Both normalized mass sensitivity (S? m ) and mass-detection sensitivity (S m ) have been introduced, as shown in Equation (1-1) and (1-2) respectively, to compare the sensitivities of these devices. Normalized mass sensitivity (S ? m ) [47]: ' 0 ' 1 dm df f S m ?= (1-1) f 0 Frequency Vibration Amplitude f 13 where f 0 and f are the uninhibited resonance frequency and the resonance frequency under mass load (m?), respectively. The mass load, m?, is the mass load per unit area rather than the absolute mass on the surface of the AW device. Mass-detection sensitivity (S m ): dm df S m ?= (1-2) where m is the absolute mass of the load. Although S? m is widely used in bulk AW sensors, such as conventional TSM (thickness shear mode resonator) and SAW (surface acoustic wave resonance) for monitoring the thickness of the mass load, the S m is more reasonable and is widely used to compare AW devices for detecting a small mass, such as single bacterium/spore cell [46, 48]. Therefore, in this dissertation, S m is used to characterize the mass sensitivity of AW devices. It should be noted that neither S? m nor S m completely describes the performance of an AW biosensor. The sensitivities defined above only show the shift in resonance frequency due to a mass load. Clearly, the ability to determine the resonance frequency accurately also affects overall device mass sensitivity. Therefore, besides mass sensitivity, the performance of an AW device is determined by another parameter: quality merit factor (Q value). The Q value is defined as the ratio of the energy stored in the resonant structure to the total energy losses per oscillation cycle. The Q value represents the sharpness of the resonance peak of an AW device. A higher Q value means a sharper resonance peak, which would result in a higher precision in determining resonance frequency and a smaller minimum detectable frequency change. Therefore, for 14 developing high performance sensors, the AW devices with a higher S m and a higher Q value are desirable [50-52]. 1.3.2 Common Acoustic Devices At present, the majority of AW devices are based on piezoelectric materials, since by using the piezoelectric effect, the devices can be easily driven and sensed. The AW devices that have been widely investigated include the thickness shear mode (TSM) resonator, surface acoustic wave (SAW) device, flexural plate wave (FPW) device [43] and microcantilever (MC) [50]. Among the AW devices investigated, the thickness shear mode (TSM) resonator, also known as the quartz crystal microbalance (QCM), has met with the greatest commercial success in applications, such as plating thickness monitors [43]. TSM typically consists of a ZT-cut quartz disk with circular electrodes deposited on both sides. Application of an electric potential between the electrodes causes a shear deformation across the crystal due to piezoelectric effect. The shear mode oscillation is most sensitive to changes in mass at the crystal surface [6, 43]. By coating the TSM surface with biological active elements, TSM have been used to detect microbial contamination [53-55]. However, TSM sensors have the lowest sensitivity among AW devices [55]. The SAW sensor consists of a piezoelectric crystal with two patterned interdigitated electrodes on the same side of the crystal surface. When an alternating voltage is applied on one of the interdigitated electrodes, an acoustic wave, whose propagation is restricted to crystal surface, is excited and travels along the surface until it is received by the other interdigitated electrodes [6, 43]. This surface wave is very sensitive to the surface perturbation. Therefore, SAW sensors exhibit a much higher sensitivity than TSM, 15 as well as a lower cost and a better reliability [6, 56]. However, the surface acoustic wave becomes severely decreased in the liquid, thus SAW sensors can not work well in liquid media [6]. In a FPW device, an acoustic wave is excited in a thin plate, whose thickness is a fraction of the acoustic wavelength. Two interdigitated transducers are fabricated on the top of the plate acting as wave generator and wave receiver, respectively. During the wave propagating through the plate, the shape of the entire plate is like a flag waving in the wind [43, 57]. FPW have much higher S? m [43]. Due to the fabrication difficulty, FPW device does not gain a great progress in terms of miniature. 1.3.3 Microcantilevers In 1990?s, researchers began to convert the micromachined cantilevers used in atomic force microscopy (AMF) into a new sensor platform [58, 59]. The microcantilevers (MCs) as a type of MEMS-based devices have many advantages over other AW devices, such as compact size and easy integration with analysis circuit [50, 60]. More importantly, the S m of MCs is superior to that of traditional AW transducers. The MCs can detect a much smaller mass ? as small as the mass of a single bacterium or spores even virus, due to their small size and mass [50-52]. For example, B. Ilic has demonstrated that silicon- based microcantilevers in length about 20 ?m are capable to detect a single E.coli cell [61], and more recently, the detection of a single virus or even single NDA molecule has been demonstrated in vacuum using the nano-cantilevers with the length less than 5 ?m [62-64]. These features are very attractive for current and future technologies. Therefore, a lot of efforts have been spent to develop MCs [45, 50, 60]. A typical microcantilever consists of a long and thin micro-beam fixed at one end 16 with a support. There are two operation modes of microcantilevers working as sensor platforms [60]: static mode or dynamic mode. In static mode, MCs work as stress sensors by measuring the deflection/bending of the beam due to changes in the surface stress because of the adsorption of molecules. In dynamic mode, MCs are AW devices, whose resonance frequencies change due to the mass load attached on the beam [45, 60]. In the medium, MCs would undergo Brownian motion due to the mechanical thermal noise. Even though the amplitude of the Brownian motion is very small (0.01 ~ 0.1 nm), this slight vibration affects MCs? practice applications [52]. Operating MCs in dynamic mode can eliminate this effect since the MCs are excited at the resonance frequency. Therefore, most applications of MCs as sensor platforms employ dynamic mode. Dynamic MCs investigated so far can be simply categorized into two types [50, 60]: the single beam MCs, such as silicon MCs [62, 64], and the composite-beam MCs, such as piezoelectric MCs [65, 66]. MCs made of non-piezoelectric materials, such as silicon, are driven by mechanical force. Usually, an external piezo-material is mounted near the cantilever, and the MC is actuated by mechanically shaking as an alternating electric field is applied on the driving piezo-material [60]. The vibration of the MC is detected by optical methods, such as optical level deflection detection [60] and interferometry [67]. Although the sensitivity of optical methods is very high, the required optical system is complicated, expensive, and occupies a significant amount of space, contrary to the development of compact and cost- effective detection devices. More importantly, optical sensing methods limit the application of the devices in liquid due to the dispersion of the laser beam in the medium [60]. Current high sensitive biological detection using silicon MCs is carried out by first 17 exposing the sensing surface to the target analyte, followed by drying the microcantilever before making the measurements in either air or vacuum [61, 67]. However, for most of devices to detect biological threat agents, a test conducted in liquid is required. A great deal of effort has paid and a great progress has been made on piezoelectric- based MCs [65, 68-70]. Compared to silicon MCs, the piezoelectric-based MCs are much easier to actuate and sense since its resonance behavior can be simply characterized using the impedance spectrum (impedance versus frequency). Besides, driving and sensing by electric signal makes it possible to integrate the driving and characterizing circuits in a chip with the MC. Theoretically, two types of piezoelectric-based MCs can be used as sensor platform: unimorph and bimorph. The unimorph consists of a piezoelectric layer and a substrate layer, which can either conduct or insulation. The bimorph, which consists of two piezoelectric layers, has two types: parallel and series connections. Due to the complicate structure and the need for electrical connections for actuating and sensing, the fabrication of piezoelectric MCs is much difficult. Additionally, when the MCs have to be employed in liquid, especially conductive liquid, extra efforts are needed to insulate the electrodes. The major drawback of MCs as sensor platforms is that the amplitude of the resonance peak and the Q value reduce significantly when the cantilever is operated in a viscous environment, such as air or liquid. This may limit their practical applications. Even though a MC has a high mass sensitivity, the shift in the resonance frequency caused by the mass load might not be identified due to the low Q value. For regular rectangular MCs, the Q value of silicon MCs in air is more than 100 [50, 60], while that of piezoelectric MCs in air is less than 100 [65, 66]. When MCs are operated in liquid, 18 most of MCs exhibit a Q value about 10 [50]. Many approaches have been investigated to increase the Q value of the MCs and to improve their performance in liquid [71, 72]. For example, for a piezoelectric MC, by employing a PZT layer with improved piezoelectric properties and the design with a nonpiezoelectric extension, a Q value higher than 300 in air has been achieved [72], while the performance in liquid can be improved by submerging only the nonpiezoelectric extension tip in liquid to reduce viscous damping effects [73]. However, high-resolution detection in liquid using MCs still remains a challenge. Recently, piezoelectric diaphragms with bending mode have been reported as a high performance AW sensor platform employed in liquid since the liquid viscosity has very little effect on the Q value of diaphragms [74, 75]. 1.4 Fundamentals of Magnetostriction and Magnetostrictive Materials 1.4.1 Magnetoelastic Effects James P. Joule first discovered that the dimensions of iron changed when it was exposed under a magnetic field in 1842. Since then, the magnetoelasticity has been intensively studied and enormous technical applications based on magnetoelasticity have been developed [76]. The magnetoelasticity is a science dealing with the coupling between the magnetic and elastic properties of a magnetic material. The size and shape of a magnetic material can vary with the magnitude and direction of its magnetization: this is the direct magnetoelastic effect, known as magnetostriction. Conversely, the magnetic properties of a material depend on the applied mechanical stress: this is the inverse magnetoelastic 19 effect. These effects can be observed in most of the ferromagnetic materials [76, 77]. The main magnetoelastic effects that are concerned for technical applications are summarized as follows [76]: ? Volume magnetostriction: The volume of a magnetostrictive materials changes due to the magnetic interaction between the materials and an external magnetic field. Its inverse effect is the pressure dependence of Curie temperature. ? Joule magnetostriction is a phenomenon that a magnetostrictive material stretches or compresses along the external magnetic field direction. Inversely, the magnetization of a magnetostrictive material varies with the mechanical stress and strains. ? Wiedemann effect is a variant of Joule magnetostriction: a helical magnetic field applied to a magnetostrictive wire causes a twisting in the wire. Its inverse effect is called Matteuci effect: the change of the magnetization in a magnetostrictive wire because of torsional deformation in the wire. ? ?E effect describes the deviation of Young?s modulus (E) of a magnetostrictive material from Hook?s law. This is because that the stress applied on the material changes its magnetization, which will cause extra deformation in the material due to the magnetostriction. In this research, we developed the novel AW biosensor platforms based on Joule magnetostriction. Joule magnetostriction is an anisotropic deformation, as shown in Fig. 1-7 [78]. Under an external magnetic field, a magnetostrictive material elongates or contracts along the external magnetic field direction to minimize the sum of the magnetoelastic and 20 elastic energies. This strain is expressed as ? and is described as the relative change in length measured along the external field (H) direction (Equation (1-3)). ll /?=? (1-3) When the magnetic saturation is achieved, ? reaches a limiting value, which is called saturation magnetostriction, ? s . If the material elongates along the external field, ? s is positive; while ? s is negative if the strain is contraction. Below the saturation, the ? is very sensitive to the external magnetic field. ? is also very sensitive to the pre-stress condition in the material [76, 78]. Fig. 1-7. Scheme of Joule magnetostriction. External magnetic field H = 0 H l ? s > 0 l External magnetic field H ? 0 ? s < 0 l External magnetic field 21 1.4.2 Joule Magnetostriction in Ferromagnetic Materials An external magnetic field (H) alerts the magnetic moments inside of a magnetic material, which results in magnetic induction (B) within the material, as well as magnetization (M) of the material. Their relations are described in Equations (1-4) [77]: HB 0 ??= (1-4 a) MHB 0 ?? += 0 (1-4 b) HM )1( ?= ? (1-4 c) where ? and ? 0 is the permeability of the material and the permeability of vacuum respectively. In ferromagnetic materials, the magnetic moments align spontaneously to form small domains without the presence of an external magnetic field. This is the result of the compensation between magnetostatic energy and quantum mechanical exchange energy. The magnetostatic energy is reduced for antiparallel magnetic moments, while the quantum mechanical exchange energy is reduced when magnetic moments align to each other. To minimize the energy, individual domains spontaneously form in ferromagnetic materials with an energetically favorable domain size [77]. Due to the spontaneous magnetic domain within the ferromagnetic material, its magnetization under an external magnetic field exhibits a hysteresis loop. A typical M-H hysteresis loop of a ferromagnetic material is shown in Fig. 1-8 as solid black curve [77]. As all magnetic domains in the material aligned to the external field direction, the material reached technical saturation magnetization, M s . When H is reduced to zero, some domains remain to align to the magnetization direction, which results a remanent magnetization, M r . The remanent magnetization can be removed by applying 22 a reversed magnetic field with a magnitude of H c , called coercive field [77]. Ferromagnetic materials with small M r and H c are called magnetically soft materials [77]. For magnetically soft materials, the area inside the M-H loop is very small and the slope dM/dH is quite steep, which is favorable for minimizing the energy loss and amplifying the output. Therefore, for developing high performance AW sensor, magnetostrictive materials with soft magnetic properties are highly desired. The blue dotted line in Fig. 1-8 shows the M-H loop under the condition that there is no hysteresis (M r = 0 and H c = 0). For magnetostrictive materials, an external magnetic field results in the magnetization and generates strain in the material. The strain (x ij ) caused by magnetostriction is related to the square of the magnetization (M k M l ), which can be expressed as [79]: lkijklij MMNx = i, j, k,l = 1-3 (1-5) where N ijkl is the magnetostrictive coefficient. Therefore, only considering the strain along the external magnetic field and based on Equations (1-4) and (1-5), under saturation magnetic field, we can obtain: 2 M?? (1-6) Thus, as shown in Fig. 1-9, although the magnetization of a ferromagnetic material is an odd function of the applied magnetic field H, the magnetostriction is an even function of H [78]. Due to the hysteresis loop for magnetization, the ?-H loop of a ferromagnetic material exhibits typical butterfly-shape, which is shown in Fig. 1-9 as solid black curve [78]. The magnetic softer the material, the narrower the ?-H loop. The blue dotted curve in Fig. 1-9 shows the ?-H relation if there is no hyteresis for magnetization (blue dotted line in Fig. 1-8). 23 Fig. 1-8. Magnetization hysteresis loop of a ferromagnetic material. Fig. 1-9. Hysteresis loop for Joule magnetostriction. M s M r M H c H hysteresis no hysteresis ? H 24 1.4.3 Response of Magnetostrictive Material under Different DC Bias As shown in Fig. 1-9, the behavior of magnetostrictive materials under external magnetic field H is non-linear. Assume, for a magnetically soft material, its ?-H curve is shown in Fig. 1-10. Based on Equation (1-6), near the origin (region I), the magnetostriction is a quadratic function of the external magnetic field [76]. When there is no DC bias (H DC = 0), that is only an AC external magnetic field, tHH AC ?sin 0 = , is applied on the material, the corresponding strain in region I would be: tHH ?? 22 0 sin)( ? )2sin1( 2 2 0 t H ??? t HH ?2sin 22 2 0 2 0 ?? (1-7) Equation (1-7) indicates that the material oscillates at a double frequency of applied external field. Also, as shown in Fig. 1-10, the observed strain is very small [76]. On the other hand, if an AC magnetic field (H AC ) is applied on the material with a DC polarizing magnetic field (H DC = H 1 ? 0), and H DC is much bigger than H AC , as shown in Fig. 1-10 (region II), the resultant stain would be: ACDC H ??? +=)( (1-8) As shown in Fig. 1-10, when H AC is small, AC ? shows quasi-linear relationship with H AC at region II. That is: )()( ??? ACDC HH +? (1-9) Equation (1-9) indicates that the obtained strain is almost proportional to the external AC driving magnetic field. Therefore, the material exhibit piezomagnetic behavior in 25 region II. Also, as shown in Fig. 1-10, the enhanced strain, and thus the enhanced output signal, can be obtained at this region. Therefore, generally, magnetostrictive devices are operated by super-imposing a small AC magnetic field on an applied DC bias field in order to avoid double frequency and obtain amplified quasi-linear output [76]. Based on Fig. 1-10, under different DC bias (H 1 , H 2 ), similar linear outputs can be obtained from the magnetostrictive device, but the amplitudes of the outputs would be different. Thus, for a magnetostrictive device, high response can be obtained at optimized DC bias field. Fig. 1-10. Response of a magnetostrictive material under different DC bias. H H 1 H 2 ? Region I Region II Region III 26 1.4.4 Piezomagnetism in Polarized Magnetostrictive Materials As discussed in above section, magnetostrictive materials behave in a quasi-linear manner if a bias magnetic field H DC is applied on the material. That is, magnetostrictive materials behavior as piezomagnetic materials under the induced polarization, such as applying a bias magnetic field H DC or a mechanical pre-stress T 0 on the material [80, 81]. The direction of H DC or T 0 is the polarization direction. Based on IEEE standard, the polarization direction is conventionally defined as the 3-direction, while directions perpendicular to this axis can be chosen as the 1- or 2-direction, as shown in Fig. 1-11 [81]. Fig. 1-11. Scheme of the conventional directions for a polarized magnetostrictive material. In the polarization direction, the application of a mechanical stress, T, generates a mechanical strain, S, as well as a magnetic induction, B, in the material. Conversely, the application of a magnetic filed, H, produces not only a magnetic induction, B, but also a mechanical strain, S. Considering only the variations around the polarized state, the behaviors of the piezomagnetic material can be described as following [76, 78]: H DC 3 1 2 27 kkijkl H ijklij HdTsS += i, j, k = 1-3 (1-10) k T ikklikli HTdB ?+= i, j, k = 1-3 (1-11) where H s is compliance at constant H, d is the piezomagnetic constant and T ? is permeability at constant T, respectively. They are called the magneto-elastic coefficients. The value of piezomagnetic constant, d, is affected by the applied DC bias, which causes the different responses of the material under different DC bias, as discussed in previous section. In order to write the magneto-elastic tensors in the form of a matrix array, Viogt rules are introduced. Based on Viogt?s definition, ij and kl is replaced by p and q respectively according to [78]: 11 ? 1, 22 ? 2, 33 ? 3, 23 or 32 ? 4, 31 or 13 ? 5, and 12 or 21 ? 6 Therefore, elastomagnetic matrices can be defined by the following equations [78, 81]: kkpq H pqp HdTsS += (1-2) k T ikqiqi HTdB ?+= (1-3) where i, k = 1-3 and p, q = 1-6. 1.4.5 Magnetomechanical Coupling Coefficient The ability of a magnetostrictive material to convert magnetic energy to mechanical energy ? or vice versa - is characterized by the magnetomechanical coupling coefficient, k, which is defined as [76]: mel mel EE E k 2 2 = (1-14) 28 where mel E is the magnetoelastic energy, el E and m E is the input elastic and magnetic energy, respectively. Considering a magnetostrictive material as shown in Fig. 1-11, the actuating is normally applied along the polarization direction (3-direction). It is assumed that the transverse actuating fields are negligible (H 1 = H 2 = 0), also the radial stresses are zero (T 1 = T 2 = 0) and there is no shear effect (T 4 = T 5 = T 6 = 0) [80]. In this case, the longitudinal magnetomechanical coupling coefficient, k 33 , is defined as [78, 80]: TH s d k 3333 2 332 33 ? = (1-5) This parameter is very crucial to the magnetostrictive sensors, actuators and transducers. Therefore, magnetostrictive materials with a large coupling factor are favorable for developing high performance magnetostrictive actuators. 1.4.6 Magnetostrictive Materials Magnetostrictive materials have been intensively investigated and the magnetostrictive devices, such as magnetostrictive actuators and sensors, have been wildly used in current industry. The magnetostrictive materials used for actuators are required to have a large magnetostriction, which means they should be able to generate a large deformation under a small magnetic field. Conventionally, magnetostrictive actuators were built based on Ni or Fe-based alloys that exhibits a saturation magnetostriction about 40 x 10 -6 [76]. However, after World War II, most of magnetostrictive actuators were replaced by piezoelectric ceramics because of their much larger strains [76, 82]. Great attentions have been paid on magnetostrictive materials again due to the 29 discovery of giant magnetostriction in some rare-earth based alloys in 1970?s [82]. It was discovered that R(rare-earth)Fe 2 compounds, have the magnetostriction 100 time higher than that of conventional magnetostrictive materials at room temperature [82]. But they are highly magnetocrystalline anisotropic, which causes it is very difficult to switch the magnetization away from its easy magnetic axis. Huge magnetic field is required to saturate the material. In 1975, A. E. Clark discovered Tb 0.3 Dy 0.7 Fe 2 alloy, which have giant magnetostriction and can be saturated at low magnetic field. This alloy was named as Terfenol-D (Ter for Terbium, Fe for iron, nol for Naval Ordnance Laboratory and D for Dysprosium). The remarkable prosperities of Terfenol-D make it comparable with piezoelectric ceramics due to its large strain response and high power density in the low- frequency range. Terfenol-D is commercial available [82]. However, Terfenol-D is brittle and exhibits poor mechanical properties. Recently, the Fe 1-x Ga x alloys (Galfenol) shows the promise to overcome the limitation of Terfenol-D due to their large magnetostriction at low-field as well as strong and ductile mechanical properties [83-85]. Galfenol in bulk form has been intensively studied by several groups, while, more recently, nanowires of Galfenol have been fabricated and studied [83-89]. The best material for AW magnetostrictive sensor is not a material with giant magnetostriction but a material that can exhibit high Q value during oscillation. Magnetostrictive materials with soft magnetic properties, high Young?s modulus and high magnetomechanical coupling coefficient are favorable for AW sensor applications [82]. Excellent magnetoelastic and mechanical properties of magnetically soft metallic glasses make them good candidates for developing high performance magnetostrictive sensors. Metallic glasses are fabricated by quench the melting alloy at the cooling rate of 30 10 6 degree/s [82]. With amorphous structure, they exhibit soft and almost isotropic magnetic properties. The magnetomechanical coupling coefficient, as high as 0.97, has been achieved in an iron rich metallic glass, Metglas ? 2605 SC, by annealing the material ribbon in a magnetic field perpendicular to the ribbon length [82]. Properties of some magnetostrictive materials are listed in Table 1-1. TABLE 1-1 Properties of Some Magnetostrictive Materials [76, 82, 90, 91] Material T C (K) ? s (10 -6 ) k 33 max Young?s modulus (GPa) Nickel 631 -36 0.31 210 Fe 65 Ni 35 773 40 0.32 ~ 110 Magnetite (Fe 3 O 4 ) 853 40 0.36 / Terfenol (TbFe 2 ) 698 1750 0.35 94 Terfenol-D 653 1100 0.75 43 Metglas ? 2605SC 643 30 0.97 100 - 110 Metglas ? 2826MB 626 12 / 100 - 110 1.4.7 Applications of the Magnetostrictive Materials The main industrial applications of giant magnetostrictive materials are actuators, such as sonars, fuel injectors, loudspeakers, vibration dampers, fast relays, rapid shutters, and magnetostrictive motors [82]. Due to the high magnetomechanical coupling factor and high energy density, underwater sonars based on Terfenol-D exhibit better performance than the ones built based on piezoelectric ceramics [80, 82]. Magnetostrictive metallic glasses are widely used to making high sensitive 31 magnetostrictive sensors. The conventional applications include strain gages, force sensors, pressure sensors, torque sensors, and position sensors [82, 92]. Since last decade, Fe-based amorphous alloys have been used to build magnetostrictive AW sensors [82]. For example, magneto-surface-acoustic-wave sensors are constructed by depositing magnetostrictive thin film on the conventional SAW substrate [93]. The velocity of the SAW propagating in the magnetostrictive film is controlled by an applied magnetic field. A change of external factors, such as magnetic field, force or pressure, affects this velocity, and thus can be sensed [93, 94]. Magnetostrictive AW sensors based on metallic glasses are also widely employed as surveillance devices by retailers to prevent theft [95]. In this system, a thin amorphous alloy strip is sealed in a small plastic box. The strip can move freely in the box and acts as a resonator. Also, a thin hard magnetic strip is placed at the bottom of the box to polarize the magnetostrictive strip. This sensor is attached on the item needed to be surveilled as a label. The alarm system, installed at the store exits, consists of a transmitter and an antenna. The transmitter generates a magnetic field pulse (~ 58 kHz), which has the same frequency as the resonance frequency of the magnetostrictive strip [95, 96]. If the item has not been paid for and brought within the alarm system, the magnetostrictive strip oscillates due to the magnetic pulse. This oscillation results in an emission of a magnetic pulse, which lasts longer than the actuating pulse, and then is picked up by the antenna [95]. When the item is paid, the sensor can be deactivated by demagnetizing the hard magnetic strip. This anti-theft system can provide a fast and wide-area detection without false alarm. Also, it is very inexpensive [95]. Recently, Crime et al. started to develop chemical and biological sensors based on the 32 magnetostrictive AW platform. The sensors are made from commercial amorphous metallic glass ribbon and are fixed at the middle for support. By measuring the resonance frequency shift of the magnetostrictive sensors, they have successfully detected temperature, pH value, pressure, viscosity, and bioagents, such as E. coli and rivin [97- 106]. 33 References [1] D. Ivnitski, I. Abdel-Hamid, P. Atanasov, and E. Wilkins, "Biosensors for detection of pathogenic bacteria," Biosensors & Bioelectronics, vol. 14, pp. 599- 624, 1999. [2] P. S. Mead, L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe, "Food-related illness and death in the United States," Emerging Infectious Diseases, vol. 5, pp. 607-625, 1999. [3] E. C. Alocilja and S. M. Radke, "Market analysis of biosensors for food safety," Biosensors & Bioelectronics, vol. 18, pp. 841-846, 2003. [4] J. C. Buzby, Roberts, T., Lin, C.T.J., MacDonald, J.M., "Bacterial Foodborne Disease: Medical Costs And Productivity Losses," Washington D.C., USDA Economic Research Service Agricultural Economics Report No. 741:100, 1996. [5] "http://www.hhs.gov/asl/testify/t061115.html." [6] P. Leonard, S. Hearty, J. Brennan, L. Dunne, J. Quinn, T. Chakraborty, and R. O'Kennedy, "Advances in biosensors for detection of pathogens in food and water," Enzyme and Microbial Technology, vol. 32, pp. 3-13, 2003. [7] M. N. Velasco-Garcia and T. Mottram, "Biosensor technology addressing agricultural problems," Biosystems Engineering, vol. 84, pp. 1-12, 2003. [8] A. E. Greenberg, Trussel, R.R., Clesceri, L.S., Franson, M.A.H. (Eds.), Standard Methods for the Examination of Water and Wastewater: American Public Health Association, Washington, D.C., 1992. [9] L. D. Mello and L. T. Kubota, "Review of the use of biosensors as analytical tools in the food and drink industries," Food Chemistry, vol. 77, pp. 237-256, 2002. [10] E. de Boer and R. R. Beumer, "Methodology for detection and typing of foodborne microorganisms," International Journal of Food Microbiology, vol. 50, pp. 119-130, 1999. [11] C. W. Kaspar and C. Tartera, "Methods for Detecting Microbial Pathogens in Food and Water," Methods in Microbiology, vol. 22, pp. 497-531, 1990. [12] N. S. Hobson, I. Tothill, and A. P. F. Turner, "Microbial detection," Biosensors & Bioelectronics, vol. 11, pp. 455-477, 1996. [13] L. J. Harris, "http://ucce.ucdavis.edu/freeform/UC_GAPs/documents/ Extension_Presentations3035.pdf." 34 [14] R. K. Saiki, S. Scharf, F. Faloona, K. B. Mullis, G. T. Horn, H. A. Erlich, and N. Arnheim, "Enzymatic Amplification of Beta-Globin Genomic Sequences and Restriction Site Analysis for Diagnosis of Sickle-Cell Anemia," Science, vol. 230, pp. 1350-1354, 1985. [15] Y. Kim, T. R. Flynn, R. B. Donoff, D. W. Wong, and R. Todd, "The gene: The polymerase chain reaction and its clinical application," Journal of Oral and Maxillofacial Surgery, vol. 60, pp. 808-815, 2002. [16] O. Traore, C. Arnal, B. Mignotte, A. Maul, H. Laveran, S. Billaudel, and L. Schwartzbrod, "Reverse transcriptase PCR detection of astrovirus, hepatitis A virus, and poliovirus in experimentally contaminated mussels: Comparison of several extraction and concentration methods," Applied and Environmental Microbiology, vol. 64, pp. 3118-3122, 1998. [17] E. Giakoumaki, M. Minunni, S. Tombelli, I. E. Tothill, M. Mascini, P. Bogani, and M. Buiatti, "Combination of amplification and post-amplification strategies to improve optical DNA sensing," Biosensors & Bioelectronics, vol. 19, pp. 337-344, 2003. [18] S. Riyaz-Ul-Hassan, V. Verma, and G. N. Qazi, "Rapid detection of Salmonella by polymerase chain reaction," Molecular and Cellular Probes, vol. 18, pp. 333- 339, 2004. [19] A. van Belkum, "Molecular diagnostics in medical microbiology: yesterday, today and tomorrow," Current Opinion in Pharmacology, vol. 3, pp. 497-501, 2003. [20] "http://users.ugent.be/~avierstr/principles/pcr.html." [21] V. A. Petrenko, Vodyanoy, V.J., "Phage Display for Detection of Biological Threat Agents," Journal of Microbiological Methods, vol. 53 (20), pp. 243-252, 2003. [22] R. A. Walsh, Electromechanical design handbook: McGraw-Hill : New York, 2000. [23] R. S. Sethi, "Transducer Aspects of Biosensors," Biosensors & Bioelectronics, vol. 9, pp. 243-264, 1994. [24] P. Feng, "Commercial Assay Systems for Detecting Foodborne Salmonella - a Review," Journal of Food Protection, vol. 55, pp. 927-934, 1992. [25] B. Swaminathan and P. Feng, "Rapid Detection of Food-Borne Pathogenic Bacteria," Annual Review of Microbiology, vol. 48, pp. 401-426, 1994. 35 [26] P. Silley and S. Forsythe, "Impedance microbiology - A rapid change for microbiologists," Journal of Applied Bacteriology, vol. 80, pp. 233-243, 1996. [27] P. D. Patel, "(Bio)sensors for measurement of analytes implicated in food safety: a review," Trac-Trends in Analytical Chemistry, vol. 21, pp. 96-115, 2002. [28] S. S. Deshpande and R. M. Rocco, "Biosensors and Their Potential Use in Food Quality-Control," Food Technology, vol. 48, pp. 146-150, 1994. [29] A. G. Gehring, D. L. Patterson, and S. I. Tu, "Use of a light-addressable potentiometric sensor for the detection of Escherichia coli O157 : H7," Analytical Biochemistry, vol. 258, pp. 293-298, 1998. [30] C. Ercole, M. Del Gallo, L. Mosiello, S. Baccella, and A. Lepidi, "Escherichia coli detection in vegetable food by a potentiometric biosensor," Sensors and Actuators B-Chemical, vol. 91, pp. 163-168, 2003. [31] F. Patolsky, G. F. Zheng, and C. M. Lieber, "Nanowire-based biosensors," Analytical Chemistry, vol. 78, pp. 4260-4269, 2006. [32] D. Ivnitski, T. Wolf, B. Solomon, G. Fleminger, and J. Rishpon, "An amperometric biosensor for real-time analysis of molecular recognition," Bioelectrochemistry and Bioenergetics, vol. 45, pp. 27-32, 1998. [33] T. M. Rossi, Warner, M., Bacterial identification using fluorescence spectroscopy: VCH Pulisher, 1985. [34] S. A. Glazier and H. H. Weetall, "Autofluorescence Detection of Escherichia-Coli on Silver Membrane Filters," Journal of Microbiological Methods, vol. 20, pp. 23-27, 1994. [35] J. I. Prosser, "Molecular Marker Systems for Detection of Genetically-Engineered Microorganisms in the Environment," Microbiology-UK vol. 140, pp. 5-17, 1994. [36] J. I. Prosser, K. Killham, L. A. Glover, and E. A. S. Rattray, "Luminescence- based systems for detection of bacteria in the environment," Critical Reviews in Biotechnology, vol. 16, pp. 157-183, 1996. [37] S. Ramanathan, W. P. Shi, B. P. Rosen, and S. Daunert, "Bacteria-based chemiluminescence sensing system using beta-galactosidase under the control of the ArsR regulatory protein of the ars operon," Analytica Chimica Acta, vol. 369, pp. 189-195, 1998. [38] R. Blasco, M. J. Murphy, M. F. Sanders, and D. J. Squirrell, "Specific assays for bacteria using phage mediated release of adenylate kinase," Journal of Applied Microbiology, vol. 84, pp. 661-666, 1998. 36 [39] J. Homola, S. S. Yee, and G. Gauglitz, "Surface plasmon resonance sensors: review," Sensors and Actuators B-Chemical, vol. 54, pp. 3-15, 1999. [40] A. N. Naimushin, C. B. Spinelli, S. D. Soelberg, T. Mann, R. C. Stevens, T. Chinowsky, P. Kauffman, S. Yee, and C. E. Furlong, "Airborne analyte detection with an aircraft-adapted surface plasmon resonance sensor system," Sensors and Actuators B-Chemical, vol. 104, pp. 237-248, 2005. [41] V. Koubova, E. Brynda, L. Karasova, J. Skvor, J. Homola, J. Dostalek, P. Tobiska, and J. Rosicky, "Detection of foodborne pathogens using surface plasmon resonance biosensors," Sensors and Actuators B-Chemical, vol. 74, pp. 100-105, 2001. [42] A. V. Samoylov, V. M. Mirsky, Q. Hao, C. Swart, Y. M. Shirshov, and O. S. Wolfbeis, "Nanometer-thick SPR sensor for gaseous HCl," Sensors and Actuators B-Chemical, vol. 106, pp. 369-372, 2005. [43] D. S. Ballantine, White, R.M., Martin, S.J., Ricco, A.J., Frye, G.C., Zellers, E.T., Wohltjen, H., Acoustic wave sensors: theory, design and physico-chemical applications: Academic Press, 1997. [44] O. Tamarin, C. Dejous, D. Rebiere, J. Pistre, S. Comeau, D. Moynet, and J. Bezian, "Study of acoustic Love wave devices for real time bacteriophage detection," Sensors and Actuators B-Chemical, vol. 91, pp. 275-284, 2003. [45] R. Raiteri, M. Grattarola, H. J. Butt, and P. Skladal, "Micromechanical cantilever- based biosensors," Sensors and Actuators B-Chemical, vol. 79, pp. 115-126, 2001. [46] T. Thundat, P. I. Oden, and R. J. Warmack, "Microcantilever sensors," Microscale Thermophysical Engineering, vol. 1, pp. 185-199, 1997. [47] M. D. Ward and D. A. Buttry, "Insitu Interfacial Mass Detection with Piezoelectric Transducers," Science, vol. 249, pp. 1000-1007, 1990. [48] J. W. Grate, S. J. Martin, and R. M. White, "Acoustic-Wave Microsensors .2," Analytical Chemistry, vol. 65, pp. A987-A996, 1993. [49] J. W. Grate, S. J. Martin, and R. M. White, "Acoustic-Wave Microsensors .1," Analytical Chemistry, vol. 65, pp. A940-A948, 1993. [50] N. V. Lavrik, M. J. Sepaniak, and P. G. Datskos, "Cantilever transducers as a platform for chemical and biological sensors," Review of Scientific Instruments, vol. 75, pp. 2229-2253, 2004. [51] J. Mertz, O. Marti, and J. Mlynek, "Regulation of a Microcantilever Response by Force Feedback," Applied Physics Letters, vol. 62, pp. 2344-2346, 1993. 37 [52] A. Mehta, S. Cherian, D. Hedden, and T. Thundat, "Manipulation and controlled amplification of Brownian motion of microcantilever sensors," Applied Physics Letters, vol. 78, pp. 1637-1639, 2001. [53] S. T. Pathirana, J. Barbaree, B. A. Chin, M. G. Hartell, W. C. Neely, and V. Vodyanoy, "Rapid and sensitive biosensor for Salmonella," Biosensors & Bioelectronics, vol. 15, pp. 135-141, 2000. [54] X. D. Su, S. Low, J. Kwang, V. H. T. Chew, and S. F. Y. Li, "Piezoelectric quartz crystal based veterinary diagnosis for Salmonella enteritidis infection in chicken and egg," Sensors and Actuators B-Chemical, vol. 75, pp. 29-35, 2001. [55] B. Drafts, "Acoustic wave technology sensors," IEEE Transactions on Microwave Theory and Techniques, vol. 49, pp. 795-802, 2001. [56] D. W. Galipeau, P. R. Story, K. A. Vetelino, and R. D. Mileham, "Surface acoustic wave microsensors and applications," Smart Materials & Structures, vol. 6, pp. 658-667, 1997. [57] A. Hierlemann and H. Baltes, "CMOS-based chemical microsensors," Analyst, vol. 128, pp. 15-28, 2003. [58] T. Thundat, R. J. Warmack, G. Y. Chen, and D. P. Allison, "Thermal and Ambient-Induced Deflections of Scanning Force Microscope Cantilevers," Applied Physics Letters, vol. 64, pp. 2894-2896, 1994. [59] J. R. Barnes, R. J. Stephenson, M. E. Welland, C. Gerber, and J. K. Gimzewski, "Photothermal Spectroscopy with Femtojoule Sensitivity Using a Micromechanical Device," Nature, vol. 372, pp. 79-81, 1994. [60] C. Ziegler, "Cantilever-based biosensors," Analytical and Bioanalytical Chemistry, vol. 379, pp. 946-959, 2004. [61] B. Ilic, D. Czaplewski, H. G. Craighead, P. Neuzil, C. Campagnolo, and C. Batt, "Mechanical resonant immunospecific biological detector," Applied Physics Letters, vol. 77, pp. 450-452, 2000. [62] B. Ilic, Y. Yang, and H. G. Craighead, "Virus detection using nanoelectromechanical devices," Applied Physics Letters, vol. 85, pp. 2604-2606, 2004. [63] A. Gupta, D. Akin, and R. Bashir, "Single virus particle mass detection using microresonators with nanoscale thickness," Applied Physics Letters, vol. 84, pp. 1976-1978, 2004. [64] B. Ilic, Y. Yang, K. Aubin, R. Reichenbach, S. Krylov, and H. G. Craighead, "Enumeration of DNA molecules bound to a nanomechanical oscillator," Nano Letters, vol. 5, pp. 925-929, 2005. 38 [65] J. W. Yi, W. Y. Shih, and W. H. Shih, "Effect of length, width, and mode on the mass detection sensitivity of piezoelectric unimorph cantilevers," Journal of Applied Physics, vol. 91, pp. 1680-1686, 2002. [66] G. A. Campbell and R. Mutharasan, "Sensing of liquid level at micron resolution using self-excited millimeter-sized PZT-cantilever," Sensors and Actuators a- Physical, vol. 122, pp. 326-334, 2005. [67] B. Ilic, D. Czaplewski, M. Zalalutdinov, H. G. Craighead, P. Neuzil, C. Campagnolo, and C. Batt, "Single cell detection with micromechanical oscillators," Journal of Vacuum Science & Technology B, vol. 19, pp. 2825-2828, 2001. [68] J. W. Yi, W. Y. Shih, R. Mutharasan, and W. H. Shih, "In situ cell detection using piezoelectric lead zirconate titanate-stainless steel cantilevers," Journal of Applied Physics, vol. 93, pp. 619-625, 2003. [69] W. Y. Shih, X. P. Li, H. M. Gu, W. H. Shih, and I. A. Aksay, "Simultaneous liquid viscosity and density determination with piezoelectric unimorph cantilevers," Journal of Applied Physics, vol. 89, pp. 1497-1505, 2001. [70] H. Y. Luo, W. Y. Shih, and W. H. Shih, "Miniaturized cantilever sensor from highly piezoelectric PMN-PT thick films," Abstracts of Papers of the American Chemical Society, vol. 228, pp. U888-U888, 2004. [71] C. Lee, T. Itoh, and T. Suga, "Self-excited piezoelectric PZT microcantilevers for dynamic SFM - with inherent sensing and actuating capabilities," Sensors and Actuators A-Physical, vol. 72, pp. 179-188, 1999. [72] Z. Y. Shen, W. Y. Shih, and W. H. Shih, "Self-exciting, self-sensing PbZr0.53Ti0.47O3/SiO2 piezoelectric microcantilevers with femtogram/Hertz sensitivity," Applied Physics Letters, vol. 89, Art. No. 023506, 2006. [73] Z. Y. Shen, W. Y. Shih, and W. H. Shih, "Mass detection sensitivity of piezoelectric cantilevers with a nonpiezoelectric extension," Review of Scientific Instruments, vol. 77, Art. No. 065101, 2006. [74] Y. Xin, Z. M. Li, L. Odum, Z. Y. Cheng, and Z. Xu, "Piezoelectric diaphragm as a high performance biosensor platform," Applied Physics Letters, vol. 89, Art. No. 223508, 2006. [75] C. Ayela, L. Nicu, C. Soyer, E. Cattan, and C. Bergaud, "Determination of the d(31) piezoelectric coefficient of PbZrxTi1-xO3 thin films using multilayer buckled micromembranes," Journal of Applied Physics, vol. 100, Art. No. 054908, 2006. [76] E. Lacheisserie, Magnetostriction - Theory and Applications of Magnetoelasticity: CRC Press, Inc., 1993. 39 [77] R. E. Hummel, Electronic Properties of Materials, Third ed: Springer-Verlag New York, Inc., 2001. [78] E. T. Lacheisserie, D. Gignoux, and M. Schlenker, Magnetism - Fundamentals: First Springer Science + Business Media, Inc., 2005. [79] R. E. Newnham, Properties of Materials: Oxford University Press, 2005. [80] F. Claeyssen, N. Lhermet, R. LeLetty, and P. Bouchilloux, "Actuators, transducers and motors based on giant magnetostrictive materials," Journal of Alloys and Compounds, vol. 258, pp. 61-73, 1997. [81] IEEE, "IEEE Standard on Magnetostrictive Materials: Piezomagnetic Namenclature," IEEE Std. 319-1990, 1990. [82] E. T. Lacheisserie, D. Gignoux, and M. Schlenker, Magnetism - Materials and Applications: First Spring Science + Business Media, Inc, 2005. [83] A. E. Clark, J. B. Restorff, M. Wun-Fogle, T. A. Lograsso, and D. L. Schlagel, "Magnetostrictive properties of body-centered cubic Fe-Ga and Fe-Ga-Al alloys," IEEE Transactions on Magnetics, vol. 36, pp. 3238-3240, 2000. [84] S. Guruswamy, N. Srisukhumbowornchai, A. E. Clark, J. B. Restorff, and M. Wun-Fogle, "Strong, ductile, and low-field-magnetostrictive alloys based on Fe- Ga," Scripta Materialia, vol. 43, pp. 239-244, 2000. [85] J. R. Cullen, A. E. Clark, M. Wun-Fogle, J. B. Restorff, and T. A. Lograsso, "Magnetoelasticity of Fe-Ga and Fe-Al alloys," Journal of Magnetism and Magnetic Materials, vol. 226, pp. 948-949, 2001. [86] R. A. Kellogg, A. B. Flatau, A. E. Clark, M. Wun-Fogle, and T. A. Lograsso, "Temperature and stress dependencies of the magnetic and magnetostrictive properties of Fe0.81Ga0.19," Journal of Applied Physics, vol. 91, pp. 7821-7823, 2002. [87] A. E. Clark, K. B. Hathaway, M. Wun-Fogle, J. B. Restorff, T. A. Lograsso, V. M. Keppens, G. Petculescu, and R. A. Taylor, "Extraordinary magnetoelasticity and lattice softening in bcc Fe-Ga alloys," Journal of Applied Physics, vol. 93, pp. 8621-8623, 2003. [88] M. Wun-Fogle, J. B. Restorff, and A. E. Clark, "Magnetomechanical coupling in stress-annealed Fe-Ga (Galfenol) alloys," IEEE Transactions on Magnetics, vol. 42, pp. 3120-3122, 2006. [89] P. D. McGary and B. J. H. Stadler, "Electrochemical deposition of Fe1-xGax nanowire arrays," Journal of Applied Physics, vol. 97, Art. No. 10R503, 2005. 40 [90] S. Harada, "Thermal-Expansion Coefficient and Youngs Modulus of Hydrogenated Fcc Fe-Ni Invar-Alloys," Journal of the Physical Society of Japan, vol. 52, pp. 1306-1310, 1983. [91] Metglas, "http://www.metglas.com/PRODUCTS/page5_1_2_7_1.htm." [92] M. Pasquale, "Mechanical sensors and actuators," Sensors and Actuators a- Physical, vol. 106, pp. 142-148, 2003. [93] H. Chiriac, M. Pletea, and E. Hristoforou, "Magneto-surface-acoustic-waves microdevice using thin film technology: design and fabrication process," Sensors and Actuators a-Physical, vol. 91, pp. 107-111, 2001. [94] E. M. Simpson and W. P. Robbins, "Magnetostrictive Fe-Si Thin-Films for Tunable Saw Devices," Ieee Transactions on Magnetics, vol. 16, pp. 919-921, 1980. [95] U. Zechbauer, "To Catch a Thief," Research and Innovation: Features, Interviews, and Reports in Siemens' Science and Technology Magazine, 1999. [96] G. Herzer and G. Rauscher, "Elongated Member Serving as a Pulse Generator in an Electromagnetic Anti-Theft or Article Identification System and Method for Manufacturing Same and Method for Producing a Pronounced Pulse in The System," US Patent 5,757,272, 1998. [97] Q. Y. Cai and C. A. Grimes, "A remote query magnetoelastic pH sensor," Sensors and Actuators B-Chemical, vol. 71, pp. 112-117, 2000. [98] C. A. Grimes, K. G. Ong, K. Loiselle, P. G. Stoyanov, D. Kouzoudis, Y. Liu, C. Tong, and F. Tefiku, "Magnetoelastic sensors for remote query environmental monitoring," Smart Materials & Structures, vol. 8, pp. 639-646, 1999. [99] C. A. Grimes and D. Kouzoudis, "Remote query measurement of pressure, fluid- flow velocity, and humidity using magnetoelastic thick-film sensors," Sensors and Actuators a-Physical, vol. 84, pp. 205-212, 2000. [100] C. A. Grimes, P. G. Stoyanov, D. Kouzoudis, and K. G. Ong, "Remote query pressure measurement using magnetoelastic sensors," Review of Scientific Instruments, vol. 70, pp. 4711-4714, 1999. [101] K. G. Ong, K. F. Zeng, X. P. Yang, K. Shankar, C. M. Ruan, and C. A. Grimes, "Quantification of multiple bioagents with wireless, remote-query magnetoelastic microsensors," IEEE Sensors Journal, vol. 6, pp. 514-523, 2006. [102] S. H. Wu, Y. Zhu, Q. Y. Cai, K. F. Zeng, and C. A. Grimes, "A wireless magnetoelastic alpha-amylase sensor," Sensors and Actuators B-Chemical, vol. 121, pp. 476-481, 2007. 41 [103] K. G. Ong, X. P. Yang, N. Mukherjee, H. D. Wang, S. Surender, and C. A. Grimes, "A wireless sensor network for long-term monitoring of aquatic environments: Design and implementation," Sensor Letters, vol. 2, pp. 48-57, 2004. [104] Q. Y. Cai, K. F. Zeng, C. M. Ruan, T. A. Desai, and C. A. Grimes, "A wireless, remote query glucose biosensor based on a pH-sensitive polymer," Analytical Chemistry, vol. 76, pp. 4038-4043, 2004. [105] M. Zourob, K. G. Ong, K. F. Zeng, F. Mouffouk, and C. A. Grimes, "A wireless magnetoelastic biosensor for the direct detection of organophosphorus pesticides," Analyst, vol. 132, pp. 338-343, 2007. [106] L. G. Puckett, G. Barrett, D. Kouzoudis, C. Grimes, and L. G. Bachas, "Monitoring blood coagulation with magnetoelastic sensors," Biosensors & Bioelectronics, vol. 18, pp. 675-681, 2003. 42 CHAPTER 2 RESEARCH OBJECTIVES 2.1 Design and Feasibility Study of Magnetostrictive Microcantilever (MSMC) as a Biosensor Platform Acoustic wave (AW) devices as high performance sensor platforms have been widely investigated in order to develop highly sensitive biosensors capable of rapid detection of a small amount of target species in a small volume of sample, mostly in liquid. Among these AW sensor platforms, microcantilevers (MCs) demonstrate the highest sensitivity in the terms of minimum detectable mass. So far, investigations of MCs as biosensor platforms mainly focus on two kinds of MCs: one is the single beam MCs, such as silicon-based MCs; another one is the composite-beam MCs, such as the piezoelectric unimorph/bimorph MCs. The advantage of silicon-based MCs is that the microfabrication process of silicon is well established, but it is difficult to actuate and sense. On the other hand, piezoelectric-based MCs are easy to be actuated and sensed, but have a complicated configuration and low Q value. In this research, we introduce a novel type of MC ? magnetostrictive microcantilever (MSMC) based on magnetostriction ? as a biosensor platform. The MSMC is compared with Si-based and piezoelectric-based MCs. The advantages of the MSMC over other MCs are presented and discussed. The MSMC can be wirelessly actuated and sensed, and has a simple structure. The MSMCs are fabricated and their performance as sensor 43 platforms is characterized. To demonstrate the performance of the MSMC as a biosensor platform, yeast and Salmonella cells are detected using MSMC biosensors in a real-time manner. 2.2 Design and Feasibility Study of Magnetostrictive Particle (MSP) as a High Performance Biosensor Platform The goal of this research is to develop a novel AW sensor platform ? magnetostrictive particle (MSP) in the shape of strip or bar ? as a high performance biosensor platform. The MSP not only inherits the advantages of other AW devices, but has its own unique advantages. Based on magnetostriction rather than piezoelectriction, the MSP is a wireless and completely free-standing sensor platform. This is its principle advantage over other AW devices. In this research, the feasibility of the MSP as a biosensor platform is studied. Some important parameters of the MSP, such as the resonance behavior, mass sensitivity and Q value, are characterized. Based on the unique features of the MSP, new detection methods (multiple-sensor approach and multiple-target approach) are designed, and their feasibilities are investigated. Multiple target approach allows simultaneous detection of different types of bioagents. Multiple sensor approach will provide a solution to the challenge of detecting biological cells in a very low concentration condition. 2.3 Synthesis of Amorphous Fe-B Alloy and Fe-B Nanowires for Magnetostrictive Biosensor Platform Application The sensitivity of AW devices depends on the size of the device: the smaller the size, the higher the mass sensitivity. To obtain a high sensitivity, magnetostrictive sensors in 44 the size of micrometers or nanometers are required. In order to develop MSMC & MSP with extremely high mass sensitivity, the synthesis of magnetostrictive thin films and fabrication of micro-MSPs and magnetostrictive nanowires are investigated. This research includes two parts. The first part is to synthesize and characterize the magnetostrictive material for magnetostrictive AW device fabrication. Amorphous Fe-B alloy is synthesized and studied as the magnetostrictive material since it exhibits favorable magnetic and mechanical properties. The Fe-B thin films are synthesized by electrochemical deposition. The deposition conditions are studied and optimized. The properties and resonance behaviors of the deposited Fe-B thin films are characterized. Combining microelectronic fabrication methods and electrochemical deposition of Fe-B films, the processes of fabricating micro-size MSPs are explored. The second part of this research is to synthesize amorphous Fe-B nanowires as biosensor platforms based on the results obtained from first part of the research. The Fe- B nanowires are fabricated by template-based electrochemical deposition. The fabrication process is optimized and the fabricated nanowires are characterized. 45 CHAPTER 3 CHARACTERIZATION OF RESONANCE BEHAVIOR OF MAGNETOSTRICTIVE AW DEVICES As a sensor platform, the resonance frequency of the AW device is used as the output signal. To characterize the resonance frequency of a magnetostrictive device, two set-ups have been built/used in this research. 3.1 Set-up (A) Based on Helmholtz Coil and Lock-in Amplifier 3.1.1 Configuration of the Setup The schematic of set-up (A) is shown in Fig. 3-1. As discussed in Chapter 1, a magnetostrictive AW device is driven by a small AC magnetic field superimposing on a relatively larger DC magnetic field (bias). biasDCACdriving HHH += (3-1) In this setup, the driving magnetic field is generated by a custom designed Helmholtz coil, which consists of two pairs of coils: one pair is used to generate a DC magnetic field to polarize the magnetostrictive material, while another pair is used to generate an AC magnetic field for exciting the device. The DC Helmholtz coil is connected to a DC power supply, which applies a constant DC current on the coil. The AC Helmholtz coil is connected to a lock-in amplifier (SRS830, Standford research system, Sunnyvale, CA), 46 which provides a sine wave signal with constant RMS (root-mean-square) amplitude of voltage to the coil. Fig. 3-1. Schematic of set-up (A) based on a Helmholtz coil and a lock-in amplifier. voltage current on/off DC power supply A/I Lock-in amplifier Helmholtz coil Helmholtz coil DC AC DC AC DC bias AC driving signal pick-up coil Magnetostrictive AW device Magnetic field 47 A uniform driving magnetic field is generated in the middle of the Helmholtz coil. The AC magnetic field sweeps over a broad frequency range to drive the magnetostrictive device, which is placed in the middle of the Helmholtz coil. The shape of the magnetostrictive material changes in response to the driving field, thus the magnetostrictive AW device works as a resonator. Due to the magnetic nature of the magnetostrictive material, the oscillation of the device will result in the emission of a magnetic signal. In this setup, a pair of home made pickup coils is used to measure the magnetic signal emitted from the magnetostrictive device. The reason of employing a pair of pickup coils is to eliminate the background. Since the pickup coil is also placed in the middle of Helmholtz coils, it can pickup the driving magnetic field, which causes strong background and might even surpass the signal from the device. To eliminate this background, the pair of pickup coils consists of two same coils. However, they were wound in opposite directions and connected in series so that the output signal is zero if there is no device inside. The magnetostrictive AW device is placed in one of the pickup coils. Therefore, the outputs of the pickup coils only reflects the magnetic signal from the device due to its oscillation and are sent back to the lock-in amplifier. The lock-in amplifier has two output signals: amplitude and phase. The amplitude represents the oscillating amplitude of the magnetostrictive device; while the phase represents the phase different between the AC driving magnetic field and the magnetic signal from the device. Both the amplitude and the phase versus frequency were recorded by a computer using a data acquisition program to obtain the resonance spectrum. 48 Fig. 3-2. Picture of set-up (A) based on a Helmholtz coil and a lock-in amplifier. For the ideal case, the background signal detected by the pair of pickup coils should be zero. However, it is very difficult to make the two coils exactly identical in practice. Therefore, the background signals generated in two coils does not totally cancel each other and, in this study, a non-zero signal is observed when the magnetostrictive device is not inside the pickup coils. Fig. 3-3 shows the background signal collected by a pair of pickup coils. The results shown in Fig. 3-3 reflect the amplitude & phase change with frequency. Helmholtz coil Pickup coil Lock-in amplifier DC power supply 49 Fig. 3-3. Background signal detected by a pair of pickup coils. Therefore, in this research, the measured resonance spectrum of the magnetostrictive AW device includes two signals: the background signal and the signal from the device. The effect of the background can be minimized by correcting the experimental output using collected background signal. The correction principle is shown as below. Let, I m = Measured signal; I s = Magnetostrictive device signal; I b = Background signal. Therefore, )sin()sin()sin( bbssmmm tItItII ?????? +++=+= ()() bbmmbbmm bbbbmmmm bbmmss IItIIt tItItItI tItItI ?????? ???????? ?????? sinsincoscoscossin sincoscossinsincoscossin )sin()sin()sin( ?+?= ??+= +?+=+ Let, bbmm IIA ?? coscos ?= ; bbmm IIB ?? sinsin ?= . 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 0.00004 0.00006 0.00008 0.00010 0.00012 0.00014 78 80 82 84 86 88 Ampl i t ud e (RSM vo ltage ) (V) phase amplitude Ph ase (deg re e) Frequency (Hz) 50 ? ? ? ? ? ? ? ? + + + += +=+ t BA B t BA A BA tBtAtI ss ?? ???? cossin cossin)sin( 2222 22 Let, 22 cos BA A + =? ; 22 sin BA B + =? . () ()?? ?????? ++= ++=+ tBA ttBAtI ss sin cossinsincos)sin( 22 22 So, 22 BAI s += ; ?? = s Thus, we can get: ()() () bmbmbm bbbmbmmmbbbmbmmm bbmmbbmm s IIII IIIIIIII IIII BAI ?? ???????? ???? ??+= +?++?= ?+?= += cos2 sinsinsin2sincoscoscos2cos sinsincoscos 22 22222222 22 22 (3-2) () ? ? ? ? ? ? ? ? ? ? = ? ? ? ? ? ? = ? ? ? ? ? ? = = bbmm bbmm s II II A B ?? ?? ? ? ?? coscos sinsin arctanarctan cos sin arctan tanarctan (3-3) Fig. 3-4 shows the measured output signal and the corrected output signal of a magnetostrictive cantilever in the size of 2.8 mm x 1 mm x 35 ?m. In Fig. 3-4, the solid black curves are the measured data, and the wine curves are the corrected data. The amplitude and phase signal are corrected based on Equation (3-2) and (3-3), respectively. The corrected signal shows a stronger and sharper peak. Clearly, the correction results in a better output signal. 51 Fig. 3-4. Experimental output (black curve) of a magnetostrictive cantilever measured using setup A, and its corresponding corrected results (wine curve) based on Equation (3- 2) and (3-3). The size of the cantilever is 2.8 mm x 1 mm x 35 ?m. 4.0x10 -5 4.1x10 -5 4.2x10 -5 4.3x10 -5 4.4x10 -5 2600 2650 2700 2750 2800 1.1x10 -5 1.2x10 -5 1.3x10 -5 1.4x10 -5 1.5x10 -5 Measured Amplitude Am plit ude (V) Frequency (Hz) Corrected Amplitude 2600 2650 2700 2750 2800 70 72 74 76 78 80 82 84 86 88 Phase Corrected Phase Measured Phase Frequency (Hz) 52 For the results reported in this research, the characterization of the device was based on the measured output signals without background correction, since the background is small enough to obtain a clear resonance frequency peak. Thus, it is expected that the real performance of the device can be better than the data reported in this dissertation. 3.1.2 Resonance Frequency and Q Value Determination As discussed above, the output signals of the lock-in amplifier are used to obtain the resonance spectrum of the device. Fig. 3-5 shows a typical resonance spectrum of a magnetostrictive AW device, where (a) is the typical output for a magnetostrictive particle, and (b) is the typical output for a magnetostrictive microcantilever. Both amplitude and phase signals are shown in Fig. 3-5. As shown in Fig. 3-5, when the frequency of driving magnetic field is the same as the natural resonance frequency of the magnetostrictive AW device, the device undergoes the resonance oscillation. At this frequency, the device has the biggest vibration, thus the emitted magnetic flux and the measured amplitude reach the maximum. This frequency is the resonance frequency (f r ). Continuously increasing the frequency of driving field, the vibration and the magnetic flux emission reach the minimum. The frequency at this point is the anti-resonance frequency (f r ? ). The correspondent phase signals around the resonance frequency are also shown in Fig. 3-5. In practice, the experimental resonance frequency (f 0 ) of the device can be determined from either the amplitude or phase signal. 53 Fig. 3-5. Typical resonance spectrum of a magnetostrictive AW device: (a) the typical output of a magnetostrictive particle; (b) the typical output of a magnetostrictive microcantilever. Amplitud e Phase f r (f 0 ) f r ? Driving frequency Amplitud e Phase Driving frequency f r f r ? (a) (b) f 0 54 It should be pointed out here that the above discussion is for the ideal case: that is the device is loss free during the vibration. Practically, due to the loss, the frequencies corresponding to the maximum and minimum magnetic signals are smaller than the real natural resonance frequency and anti-resonance frequency of the device, respectively. However, since the loss for the devices studied here is very small, the difference can be neglected. The frequency f 0 determined from the amplitude or phase signal can be assumed as the resonance frequency of the device. For a magnetostrictive particle, whose typical resonance spectrum is shown as Fig. 3- 5 (a), its amplitude signal exhibits a sharp resonance frequency peak. Therefore, the frequency corresponding to this peak is used to determine the resonance frequency of the device. In order to determine the f 0 and the Q value, the amplitude signal of the device is fitted at frequencies around the resonance frequency. Based on the experimental results, it is found that, if we choose the baseline as shown in Fig. 3-6, the Lorentz function (Equation (3-4)) can be used to fit the resonance peak. Fig. 3-6. Baselines for fitting the resonance frequency peak from the amplitude signal. The baseline is chosen through two points; one of the points is where the amplitude reaches minimum. Frequency Baseline 2 Baseline 1Amplitud e 55 The Lorentz function is expressed as: () () 2 2 0 0 4 2 wff wA DfD +? += ? (3-4) where D(f) is the amplitude as the function of driving frequency, f. D 0 is the baseline, which is determined using a two-point linear function. The A, w, f 0 are constants obtained from the fitting. The f 0 is resonance frequency of the device, the w is the width of the peak at its half height, and the ratio of the A to the w is the strength of the peak, and the ratio of the f 0 to w is the Q value. Fig. 3-7 shows the Lorentz fitting curve for the amplitude signal of a magnetostrictive particle in strip shape. The results indicate that the Lorentz function can reasonably fit the resonance peak. Fig. 3-7. Lorentz fitting curve for the measured amplitude resonance peak of a magnetostrictive particle. Solid black curve is the measured signal, while dashed blue curve is the fitting curve. 90.25 90.30 90.35 90.40 90.45 90.50 90.55 Frequency (kHz) measured amplitude A m p litud e fitting results 56 When fitting the amplitude peak using Lorentz function, we need to choose the baseline (parameter D 0 ), and the frequency range. In order to determine how these factors affect the fitting results, we chose four different baselines to fit the data. The baselines are chosen as shown in Fig. 3-6, and they all pass the minimum point of the amplitude but with different ranges. The fitting results are shown in Table 3-1. The baseline and the frequency ranges have little effect on f 0 . With the frequency range increasing, the calculated Q value increases a little, with a variation of about 5%, which is a tolerable error. This gives us confidence to determine Q and f 0 by using the Lorentz function. TABLE 3-1 Fitting Results of Amplitude Data Based on Lorentz Function Using Different Baselines Baseline frequency range (kHz) f 0 (kHz) w Q(=f 0 /w) 90.15 ? 90.55 90.416 0.064 1413 90.2 ? 90.55 90.417 0.063 1435 90.25 ? 90.55 90.417 0.06 1507 90.3 ? 90.55 90.418 0.056 1614 Average 90.417 ? 0.0007 0.061 ? 0.003 1492 ? 78 For a magnetostrictive microcantilever, whose typical resonance spectrum is shown as Fig. 3-5 (b), the phase signal is used to determine the resonance frequency (f 0 ) and the Q value of the device, since, in this case, the resonance phase is a better indicator to show how strong the resonance is. The resonance frequency (f 0 ) is determined by the corresponding frequency of the phase peak. For simplicity, the Q value of the device is determined as Shih did to determine the Q value for piezoelectric-based microcantilevers [1]. That is, the Q value is defined as ?f/f 0 , where f 0 is the resonance frequency of the 57 device determined by the phase peak, and ?f is the width of the phase peak at its half height. To precisely determinate the resonance frequency, the phase data was fitted at frequencies around the resonance. As shown in Fig. 3-8, Lorentz function (Equation (3-4)) can fit the phase peak very well. In this fitting, D(f) is the phase angle as the function of driving frequency, f. D 0 is the baseline, which can be a constant or a linear function dependent on the frequency range. The A, w, f 0 are constants obtained from the fitting. The f 0 is resonance frequency of the device; the w is the width of the phase peak at its half height. Therefore, the Q value is calculated as w/ f 0 . Fig. 3-8. Lorentz fitting results for the detected phase signal of a magnetostrictive cantilever measured in air. The black solid line is the detected signal, while the dashed blue line is the fitting curve. 85.5 86.0 86.5 87.0 87.5 88.0 1800 1820 1840 1860 1880 1900 1920 1940 1960 phase (degree) detected phase signal frequency (Hz) fitting results 58 To determine the effects of fitting parameters, D 0 and frequency range, on the fitting results, we chose three different frequency ranges to fit the measured data shown in Fig. 3-8. For each range, both a constant baseline and a linear baseline were used. The fitting results are shown in Table 3-2. The baseline and the frequency ranges have very little effects on the determination of f 0 and Q value. This indicates Lorentz function works very well to fitting the measured phase signal. TABLE 3-2 Fitting Results of Phase Data Based on Lorentz Function with Different Fitting Parameters Frequency range (Hz) Baseline f 0 (Hz) w Q(=f 0 /w) Linear, D2 1885.58 10.08 187 1865 - 1907 Constant, D2 1885.63 10.02 188 Linear, D2 1885.58 9.95 189 1855 -1915 Constant, D2 1885.64 9.96 189 Linear, D2 1885.58 9.92 190 1845 - 1925 Constant, D2 1885.64 9.95 189 Linear, D2 1885.58 ? 0 9.98 ? 0.07 189 ? 1 Average Constant, D2 1885.637 ?0.003 9.98 ? 0.03 189 ? 0.6 It should be pointed out that the Q value determined here is not the intrinsic Q value of the AW device. The output we analyzed is the magnetic signal measured by the pickup coil, not the real deformation of the device. The couplings between the driving coil, device and pickup coil, as well as the measurement circuit, all contribute to the final output. The Q value calculated here reflects the Q value of the whole system, 59 also includes the fitting error. The improvement of the driving/sensing system might be able to increase the measured Q value. 3.2 Set-up (B) Based on Network Analyzer 3.2.1 Configuration of the Setup This characterization system was built based on a network analyzer (HP 8751A) with an S-Parameter adapter (HP 87511). A network analyzer is an electric analysis instrument, which provides a known stimulus to a device and measures its response [2]. As shown in Fig. 3-9, the response can be the reflected response or the transmitted response. Fig. 3-9. A stimulus/response test system. A network analyzer characterizes the device by determining the ratios of incident signal to the response signal over a certain frequency range. The output of a network analyzer is S-parameter. The definition of an S-parameter is dependent on the testing configuration. A network analyzer provides two output ports for device characterization. Therefore, four S-parameters can be defined, which is shown in Fig. 3-10 [2]. Device under test Incident signal Reflected signal Transmitted signal 60 Fig. 3-10. Definitions of S-parameters. S-parameters are expressed as the letter S with two subscripts. The first subscript represents the port number, at which the response signal is measured; while the second subscript is the port number, at which the incident signal is applied. In this set-up, we use one-port configuration to characterize the device using reflected signal, that is, the test parameter is the S 11 parameter, which is defined as follow [2]: 2 11 ? ? ? ? ? ? ? ? == incident reflected incident reflected Voltage Voltage Power Power S (3-5) S 11 parameter is a complex number including a Real and Imaginary component pair. The magnitude of S 11 parameter (normally logarithm format is used), which represents the power ratio of reflected signal to the incident signal, is calculated as follow [2]: Log magnitude 22 10 ImaginaryReallog20 += (3-6) The phase of S 11 parameter, which represents the phase difference between reflected signal and incident signal is obtained through Equation (3-7) [2]. Real Imaginary tan 1? =phase (3-7) Reflected signal Reflected signal S 11 Incident signal Transmitted signal Device under test Port 1 Port 2 Transmitted signal Incident signal S 12 S 21 S 22 61 Fig. 3-11. Set-up (B) based on a network analyzer. On/off Port 1 Port 2 coil magnetostrictive AW device test chamber magnet magnetic field 62 The schematic of this set-up is shown in Fig. 3-11. A homemade coil was wound around the test chamber, and connected to the port 1 of the network analyzer. The magnetostrictive device is placed in the middle of the coil. The network analyzer applies a RF signal sweeping over a certain frequency range to the coil and generates an AC magnetic field in the coil to drive the device. To obtain a relatively uniform magnetic field, the length of the coil is at least twice of the length of the device. S 11 parameter is measured at port 1, and its magnitude and phase are calculated based on Equation (3-6) and (3-7). A permanent magnet is placed close to the chamber as the DC magnetic bias. The applied DC magnetic field can be tuned by slightly adjusting the distance between the test chamber and the magnet. Before putting the magnetostrictive device into the test chamber, the analyzer was calibrated to zero base-line to set the intrinsic reflection signal of the coil as zero. Therefore, the recording signal would only reflect the signal from the magnetostrictive device, which represents its oscillating amplitude. The resonance spectrum of the magnetostrictive device is obtained by recording the magnitude and phase of S 11 parameter changing with the frequency. 3.2.2 Resonance Frequency and Q Value Determination Fig. 3-12 shows the S 11 parameter changing with the frequency of a magnetostrictive strip, whose size is 10 mm x 1 mm x 30 ?m. Both the magnitude in logarithm format and the phase are shown in Fig. 3-12. When the frequency of the incident signal equates to the natural resonance frequency of the device, the device undergoes resonance oscillation, and the amount of incident power converted into the elastic energy reaches maximum. At this point, the reflected power reaches minimum, and thus the magnitude of S 11 reaches minimum. 63 The measured S 11 data can be converted to the equivalent complex impedance (Z) of the device using Equation (3-8). 11 11 1 1 50 S S Z ? + = (3-8) Based on Equation (3-8), the correspondent impedance based on the S 11 data shown in Fig. 3-12 was calculated and shown in Fig. 3-13. The magnetostrictive AW device vibrates under the driving field, and then emits a magnetic signal, which generates an electric field in the coil. At resonance frequency (f r ), the magnetostrictive device has the biggest vibration and causes the biggest magnetic flux change. Therefore, at f r , the current through the coil reaches the maximum and the impedance reaches the minimum, as shown in Fig. 3-13. Continuously increasing the driving frequency, the vibration and the magnetic flux change reach the minimum, thus the magnitude of impedance reaches the maximum. The frequency at this point is the anti-resonance frequency (f r ? ). 64 Fig. 3-12. Magnitude and phase curve of S 11 parameter changing with the frequency of a magnetostrictive strip, whose size is 10 mm x 1 mm x 30 ?m. Fig. 3-13. The equivalent impedance based on S 11 data shown in Fig. 3-12. The conversion is based on Equation (3-8). 222000 224000 226000 228000 230000 232000 -20 -15 -10 -5 0 -80 -60 -40 -20 0 20 40 log ( m a gnitude of S 11 ) ( d B) Frequency (Hz) f 0 Phase o f S 11 (d eg re e ) 222000 224000 226000 228000 230000 232000 30 40 50 60 70 80 -100 -80 -60 -40 -20 0 20 40 60 80 Lo g (magn itude o f imped ance) (log ?) Frequency (Hz) Phase of impe dance (de g ree ) f r f r ' 65 The method for characterizing the magnetostrictive device is similar to that used for setup A. In this research, for simplicity, the magnitude of the S 11 signal is used to determine the resonance frequency (f 0 ) and the Q value of the device. The f 0 is the frequency at which the S 11 magnitude reaches the minimum. The Q value is defined as ?f/f 0 , where ?f is the width of the peak at its half height. The measured data are fitted to obtain precise f 0 and Q value. Also, it is found that Lorentz function (Equation (3-4)) is a good candidate to fit S 11 magnitude peak. Fig. 3-14 shows the fitting curve for the S 11 magnitude signal of a magnetostrictive strip in the size of 6.4 mm x 1 mm x 30 ?m. Fig. 3-14 Lorentz fitting results for the measured S 11 magnitude peak of a magnetostrictive strip in the size of 6.4 mm x 1 mm x 30 ?m. The black solid line is the detected signal, while the dashed blue line is the fitting curve. 345000 345500 346000 346500 347000 347500 0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.05 Ma gn itud e o f S 11 (d B) measured S 11 signal Frequency (Hz) fitting curve 66 Different frequency ranges and baselines were chosen to fit the data shown in Fig. 3- 14 to determine the effects of the fitting parameters. The fitting results are shown in Table 3-3. The baselines and the frequency ranges do not have much effect on the determination of f 0 and the Q value, which gave us confidence of using the Lorentz function to fit the data. TABLE 3-3 S 11 Magnitude Peak Fitting Results Based on Lorentz Function using Different Fitting Parameters Frequency range (Hz) Baseline f 0 (kHz) w Q(=f 0 /w) Linear, D2 346.11 0.544 636 345000 - 347200 Constant, D2 346.10 0.548 631 Linear, D2 346.11 0.544 636 345300 - 347000 Constant, D2 346.10 0.549 630 Linear, D2 346.12 0.539 642 345000 - 346800 Constant, D2 346.10 0.560 618 Linear, D2 346.11 ? 0.007 0.542 ? 0.003 638 ? 5 Average Constant, D2 346.10 ? 0 0.552 ? 0.007 626 ? 7 67 References [1] J. W. Yi, W. Y. Shih, and W. H. Shih, "Effect of length, width, and mode on the mass detection sensitivity of piezoelectric unimorph cantilevers," Journal of Applied Physics, vol. 91, pp. 1680-1686, 2002. [2] "Agilent 4395A Network/Spectrum/Impedance Analyzer Operation Manual," Agilent Technologies. 68 CHAPTER 4 DESIGN AND FEASIBILITY STUDY OF BIOSENSOR BASED ON MAGNETOSTRICITVE MICROCANTILEVER 4.1 Introduction Based on the comparison of different AW devices, it is found that MEMS-based MCs provide a much better sensitivity in terms of the minimum detectable mass [1-7]. In other words, the MCs have the capability to detect a much smaller mass attachment. Additionally, MCs have many advantages over bulk AW devices, such as compact size and easy integration with analysis circuits. Therefore, it is interesting to employ MCs as a sensor platform for developing high performance biosensors [6-8]. As discussed in Chapter 1, much research has been conducted on MC-based biosensors. However, even though these MCs exhibit good performance and many advantages, they are facing some challenges in practical biological detection. Detection using silicon-based MCs requires complicated and expensive optical equipment. Compared to silicon-based MCs, the piezoelectric MCs are easy to actuate and sense. However, the piezoelectric MCs have much more complicated structure and exhibit a smaller Q value than silicon-based MCs. More importantly, both silicon-based and piezoelectric MCs face difficulties when they are operated in liquid [4, 5]. However, most of biological analyses are carried out in liquid medium. In this chapter, a design of a novel type of MC, which is made of magnetostrictive 69 materials and referred to as magnetostrictive microcantilever (MSMC), is presented. The MSMC can be employed as a high performance biosensor platform, which is operated based on magnetostriction rather than piezoelectricity. The MSMCs inherit all the advantages of other MCs. For example, MSMC has the similar sensitivity as other MCs. Additionally, due to the magnetostrictive effect and the magnetic nature, the MSMC can be driven and sensed wirelessly. There is no physical connection between the MSMC sensor and the interrogation system. In our preliminary research, the MSMCs were fabricated and their resonance behaviors were characterized. It is experimentally found that the MSMCs exhibit high Q value and work well in either air or liquid. For MSMCs operated in air, a Q value of more than 500 was obtained. For MSMCs operated in water, the Q value reaches more than 30. The application of the MSMC as a high performance biosensor platform is demonstrated by detecting yeast and Salmonella cells in water. 4.2 Design and Operation Principle of MSMC as Biosensor Platform 4.2.1 Configuration and Operation Principle of MSMC The configuration of the MSMC is shown in Fig. 4-1 (a). Similar to piezoelectric unimorph, the MSMC consists of two layers: one is active (magnetostrictive), and the other is inactive. The length of the magnetostrictive layer changes under an external magnetic field. This dimension change is restricted by the inactive layer, which causes the bending of the cantilever beam, as illustrated in Fig. 4-1 (b). Therefore, the bending oscillation of the cantilever can be actuated by an alternating magnetic field. Because of the magnetic nature of the magnetostrictive material, the oscillation of the cantilever beam results in an emission of a magnetic signal, which can be sensed/measured 70 wirelessly using a pick-up coil. That is, MSMC can be actuated and sensed without direct physical connections between the sensor and the interrogation device. Therefore, it allows developing a sensor for in-situ biological analysis. This is the principle advantage the MSMC offered over other MCs. Additionally, compared with piezoelectric MCs, MSMC has a simple structure and does not need to have electronic circuit on board, which greatly simplifies the fabrication process. Fig. 4-1. Structure of the magnetostrictive microcantilever (MSMC), where L and W are the length and width of the cantilever respectively, while h is the thickness of the MSMC beam, including the magnetostrictive layer and the inactive layer. (a) MSMC under zero magnetic field; (b) MSMC under an external magnetic field. Magnetostrictive material (actuating layer) w h L Non-magnetostrictive materials Magnetostrictive material (actuating layer) Non-magnetostrictive materials (a) (b) 71 The operating principle of a MSMC as a biosensor is similar to other MC sensor platforms. To form a biosensor, the surface of the MSMC is immobilized with the bio- receptor or bio-molecular recognition element. When the MSMC sensor is placed in the environment to be tested, the target species will be captured by the bio- receptor/recognition element. The captured species is a mass load for the MSMC, which results in a change in the resonance frequency. As described above, the oscillation of the MSMC can be actuated by a magnetic field and its resonance frequency can be detected by a pick-up coil. Therefore, as shown in Fig. 4-2, the presence of target species can be wirelessly determined by monitoring the shift in the MSMC sensor?s resonance frequency. Fig. 4-2. Operating/detecting principle of the MSMC as a biosensor platform. Mass load (Bacteria/Spores) Driving Coil Pick-up Coil Driving Magnetic Field Resultant Magnetic Signal Oscilla tio n Amplitude f naked sensor f mass load Frequency Bio-receptor Frequency Driv ing Field Amplitude 72 4.2.2 Natural Resonance Behavior of Magnetostrictive Microcantilever 4.2.2.1 Flexural (Bending) Resonance Frequency of Cantilever with One End Fixed For a uniform cantilever beam (in the length of L, width of w and thickness of h) shown in Fig. 4-3, the wave equation of its flexural oscillation can be expressed as [9]: 0 ? 4 4 2 2 = ? ? + ? ? xA IE t ? ? ? (4-1) where ? is the flexural deformation of the beam at the point x, E ? and ? are the apparent Young?s modulus and equivalent mass density of the beam materials, respectively; A is the cross sectional area of the beam, and I is the moment inertia. For a rectangular beam of width w and height h, the A and I are given by [9]: whA = (4-2) 3 12 1 whI = (4-3) Under the assumption of a beam structure with w>>h, the apparent Young?s modulus is given by [5]: 2 1 ? ?? = E E (4-4) where ? is the Poisson?s ration and E is the Young?s modulus of the beam material. 73 Fig. 4-3. Rectangular cantilever beam with a length of L, width of w and thickness of h, rigidly fixed at one end and free at the other. For harmonic oscillations in Equation (4-1), the flexural deformation of the beam can be written as [9]: 0 ? 2 4 4 0 =? ? ? ?? ? ? xA IE (4-5) where ? 0 is the resonance frequency. The general solution of the wave Equation (4-5) is [5]: )cosh()sinh()sin()cos()( 4321 x L cx L cx L cx L cx ???? ? +++= (4-6) where, c 1 , c 2 , c 3 and c 4 are the constants, and ? is the non-dimensional parameter; while [5]: IE A L ? 2 0 4 ??? = ? ? ? ? ? ? (4-7) By considering the boundary conditions, we can obtain the resonance frequencies for x ? x ? 74 the flexural oscillations. For the cantilever beam rigidly fixed at one end and free at the other end, the boundary conditions are [5, 9]: 0)0( =? , 0 )( 0 = =x dx xd? , 0 )( 2 2 = =Lx dx xd ? , and 0 )( 3 3 = =Lx dx xd ? (4-8) The first two boundary conditions are due to the fact that one end of the beam (at x = 0) is rigidly fixed. Since there is no shear force and moment acting at the free end of the beam (at x = L), the third and forth boundary conditions exclude any bending moment and torsion [5]. Based on Equation (4-6) and the boundary conditions, the characteristic equations for which the flexural oscillation of the beam is at resonance frequencies, is given by [5, 9]: 2 tan 2 coth ?? = (4-9) 2 tan 2 coth ?? ?= (4-10) The characteristic Eigenvalues of resonance oscillation can be obtained from Equation (4-9) and (4-10). Therefore, the resonance frequencies of natural oscillation can be determined by inserting Eigenvalues into Equation (4-7): A IE L f n nn ? ? ?? ? 2 2 2 == (4-1) Combining Equation (4-2), (4-3), (4-4) and (4-11), the n th -mode flexural resonance frequency of a cantilever with one end clamped can be expressed as Equation (4-12): )1( 122 22 2 ?? ? ? ? = E L h f n n (n=0, 1, 2, ?.) (4-12) where the n ? is the dimensionless n th -mode Eigenvalue dependent on the vibration mode, 75 h and L are the thickness and length of the cantilever, while E, ? and ? are the effective Young's modulus, density and the Poisson?s ratio of the beam material. 4.2.2.2 Determination of Eigenvalues for Flexural Resonance Oscillation The characteristic Eigenvalues for bending resonance oscillation of a cantilever beam with one end fixed can be obtained from Equation (4-9) and (4-10), which can be solved graphically. Fig. 4-4. Curves of function tan(?/2), -tan(?/2) and coth(?/2). The Eigenvalues for bending resonance oscillation of a cantilever beam with one end fixed can be graphically obtained from the intersections of the functions. 012345678 -10 -5 0 5 10 15 tan(?/2) - tan(?/2) ?/2 coth(?/2) n=0 n=1 n=2 n=3 n=4 76 Fig. 4-4 shows the curves of function tan(?/2), -tan(?/2) and coth(?/2). The solutions of Equation (4-9) and (4-10) are the intersection points of function coth(?/2) with function tan(?/2) and -tan(?/2). Therefore, Equation (4-9) and (4-10) can be graphically solved from Fig. 4-4. Table 4-1 lists the graphically obtained solutions of Equation (4-9) and (4-10), as well as the Eigenvalues for the first ten flexural resonance modes. Thus, the first ten resonance frequencies can be calculated based on these Eigenvalues. TABLE 4-1 Graphically Obtained Eigenvalues for the First Ten Flexural Resonance Modes Resonance mode ? n / 2 ? n 2 n ? n = 0 0.93755 1.87510 3.5160 n = 1 2.34705 4.69410 22.03457 n = 2 3.92738 7.85476 61.69725 n = 3 5.49777 10.99554 120.90190 n = 4 7.06858 14.13716 199.85929 n = 5 8.63938 17.27876 298.55555 n = 6 10.21018 20.42036 416.99110 n = 7 11.78097 23.56194 555.16502 n = 8 13.35177 26.70354 713.07905 n = 9 14.92257 29.84514 890.732382 4.2.3 Mass Sensitivity of MC & MSMC A) Uniform Mass Load If a mass load (?m) is uniformly distributed on the surface of the microcantilever beam, as shown in Fig. 4-5 (a), the mass sensitivity of the cantilever can be expressed as [5, 10]: 77 e nn unim M f m f S 2 , ? ? ? ?= (?m << Me) (4-13) where M e is the effective mass of the cantilever beam, n f is nature resonance frequency of the cantilever, and ?f is the shift in the resonant frequency due to the mass load (?m). Therefore, for a uniform mass load, the mass sensitivity of a MC can be written as: K wL E wL S n n unim 3 2 233 2 , 1 124 )1(2 1 122 ? ? ??? ? = ? = (4-14) where )1( 23 ?? ? = E K is the material index, which is only related to the properties of the cantilever beam material. The mass load ?m added here is assumed an ideal mass load, which does not form a solid film rigidly bonded to the cantilever surface and change the bending modulus of the cantilever [10]. Otherwise, it tends to increase the bending modulus of the cantilever and thus its resonance frequency. 78 Fig. 4-5. MSMC (a) with a uniform mass load on the surface; and (b) with a point mass load at the cantilever tip. B) Point Mass at the Tip To calculate the shift in the resonance frequency due to a point mass (?m) loaded at the free end of the cantilever (as shown in Fig. 4-5 (b)), we need to consider the mass (M e ) of the cantilever beam as the equivalent effective mass ( ' e M ) at the tip of the w h L ?m w h L ?m (a) (b) 79 cantilever. For a rectangular beam, with w (width) << L (length), the effective mass at the cantilever tip is [10]: hwLMM ee ?236.0236.0 ' == (4-15) Therefore, the shift ( n f? ) in the resonance frequency due to per unit loaded mass is: e n e nn tipm M f M f m f S 472.02 ' , == ? ? ?= (4-16) Combing Equation (4-12) and (4-16), for a point mass loaded at the cantilever tip, the mass sensitivity can be written as: K wL E wL S n n tipm 3 2 233 2 , 1 12944.0 )1(472.0 1 122 ? ? ?? ? ? = ? = (4-17) Comparing Equation (4-14) and (4-17), for a given MC and mass load ?m, one can find that the S m of a cantilever for ?m evenly distributed over the entire cantilever surface is only 23.6% of that for ?m at the tip. In other words, for the same mass load ?m, the location of ?m affects the ?f. The closer the ?m to the tip, the larger the change in ?f, thus the higher the S m . 4.2.4 Comparison of Mass Sensitivity of MSMC and Other MCs Equation (4-14) and (4-17) indicate that, given the material properties of the MC beam (constant K), the mass sensitivity of a MC increases with decreasing L & w, but is independent of h. Keeping the ratio of length (L) to width (w) as constant (L = ?w), the S m of the MSMC only depends on its length. Therefore, Equations (4-14) and (4-17) become: 80 44 2 , 1 508.43 L K L S n unim ??? ?= (4-18) 44 2 , 1 268.10 L K L S n tipm ??? ?= (4-19) Therefore, if the size of the MC is reduced by a factor of ? proportionally in all dimensions ( ,LL reduce ?= ww reduce ?= ), the mass sensitivity of the MC will enhanced by a factor of ? -4 . Therefore, the smaller the size of the MC, the higher the mass sensitivity. The mass sensitivities shown in Equation (4-18) and (4-19) indicate that, given the dimensions of a MC, its mass sensitivity depends on the material index K of the cantilever beam. K is directly related to the combination of Young?s modulus, density and Poisson ratio of the material. The MC which is made of a higher Young?s modulus and a lower density material gives a higher value for K, and thus has a higher mass sensitivity. Table 4-2 lists the mechanical properties of a commercially available magnetostrictive material [8] (Metglas TM 2826 MB, Honeywell, Morristown, NJ), piezoelectric materials [11], and silicon [12]. By using the material properties shown in Table 4-2 as effective Young?s modulus, density and Poisson ratio, the mass sensitivities of MCs made of different materials can be compared. For silicon, the K value of 3.34 can be obtained. If two widely used piezoelectric ceramics (lead zirconate titanate (PZT)), PZT4 and PZT5A, are employed to make the MC, the K values of 0.41 and 0.36 are obtained, respectively. However, if the MC is made of Metglas TM 2826 MB, an amorphous alloy widely employed in magnetostrictive sensors, a K value of 0.55 is obtained. This means that for MC with the same dimensions, a MC made of Metglas TM 2826 MB exhibit higher mass sensitivity than a MC made of PZT. 81 TABLE 4-2 Mechanical Properties of a Commercially Available Magnetostrictive Alloy, Piezoelectric Ceramics and Silicon [8, 11, 12]. Metglas TM 2826 MB PZT4 PZT5A Si Young?s Modulus (GPa) 110 66 53 130 Density (Kg/m 3 ) 7.9 x 10 3 7.6 x 10 3 7.7 x 10 3 2.33 x 10 3 Poisson Ratio 0.5 0.31 0.31 0.28 K 0.55 0.41 0.36 3.34 Fig. 4-6 (a) shows the theoretical mass sensitivity calculated based on Equation (4-18) for MCs made of Metglas TM 2826 MB, PZT4, PZT5A and silicon. In Fig. 4-6 (a), the ratio of MCs? length to width is chosen as 5 (? = 5). The mass sensitivities increase with the reducing lengths of MCs. The results in Fig. 4-6 (a) indicate that MSMC has a similar mass sensitivity as other MCs. For MCs with the same dimensions, the MCs made of Si exhibit the highest S m due to the high K value of Si, while the mass sensitivity of the MSMC made from Metglas TM 2826 MB is higher than that of PZT MCs. Equation (4-18) and (4-19) indicate that the location of the mass load affects the S m of MCs. The theoretical S m of MSMC made of Metglas TM 2826 MB was calculated based on Equation (4-18) and (4-19), and is shown in Fig. 6-4 (b), where the black line is the S m for uniform mass load (Equation (4-18)), and the red line is the S m for mass load at the tip of the cantilever (Equation (4-19)). The results revealed that, for the same mass load, larger change in resonance frequency can be obtained if the mass load is closer to the tip of the cantilever beam. 82 Fig. 4-6 (a) Theoretical mass sensitivity calculated based on Equation (4-18) for MCs made of Metglas TM 2826 MB, PZT4, PZT5A, and Si. The material properties used in the calculation are listed in Table 4-2, while the ratio of MCs? length to width is chosen as 5. (b) Theoretical mass sensitivity calculated based on Equation (4-18) and (4-19) for MCs made of Metglas TM 2826 MB. 0.1 1 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12 10 13 Length (mm) MC made from Si MSMC made from Metglas TM 2826 MB Mas s Sens it iv it y ( ? f/ ? m) (Hz/ g) MC made from PZT4 MC made from PZT5A (a) 0.1 1 10 5 10 6 10 7 10 8 10 9 10 10 10 11 10 12 10 13 S m of MSMC made of 2826 MB with mass load at the tip of the cantilever beam M a s s S ens it i v it y S m ( ? f/ ? m) (Hz/g) Length (mm) S m of MSMC made of 2826 MB with uniform mass load (b) 83 4.2.5 Comparison of MSMC with Current MCs The MSMC exhibits many advantages compared with Si-based and piezoelectric- based MCs. Interrogated through magnetic field, MSMC is a wireless sensor platform. Therefore it can be used to develop sensors for in situ and in vivo detections. Additionally, since there is no physical connection required between the MSMC sensor and the interrogation system, the MSMC has a much simpler structure than piezoelectric-based MCs. More importantly, it is experimentally found that MSMCs exhibit a higher Q value and works well in liquid. This is a crucial feature for biosensor platforms since most of real-time biological analyses take place in liquid environments. The summary of the comparison of MSMC with current MCs is listed in Table 4-3. 84 TABLE 4-3 Comparison of Different Microcantilever-based Sensor Platforms Characteristics Silicon-based Piezoelectric-based MSMC Actuation Mechanical (separated bulk system) Electrical (on-broad, wire connected) Magnetically (wireless, no connection) Transduction Optical (separated bulk system) Electrical (on-broad circuit) Magnetically ( wireless, no connection) Operate in air Yes Yes Yes Operate in insulate liquid Difficult Difficult (low performance) Works well (high performance) Operate in conductive liquid Difficult Very difficult Works well (high performance) Q value High (> 100 in air) (~ 10 in liquid) Low (?100 in air) (<10 in liquid) Very high (~ 500 in air) (~ 30 in liquid) Structure Simplest Complicated Simpler Fabrication Easy Difficult Easy Overall sensitivity High Low High Array-Actuation Very difficult (individually) Difficult (each cantilever needs a driving circuit) Easy (one driving system can drive all cantilevers at the same time) Array- Transduction Very difficult (individually) Difficult (each cantilever needs analyzing circuit) Easy (one circuit can collect the signal from all cantilevers at the same time) 85 4.3 Experimental 4.3.1 Fabrication of MSMCs In this research, the unimorph type MSMCs were fabricated. The active layer is a commercial magnetostrictive material, Metglas TM 2826 MB ribbon from Honeywell (Morristown, NJ), while the inactive layer is a copper layer. The material properties of the Metglas TM ribbon are listed in Table 4-2. The Metglas TM ribbon was polished down to a 20 ?m thickness using 2000# polish paper, and then a copper (Cu) layer, 10 um in thickness, was deposited by magnetron DC sputtering. To increase the adhesion between Metglas TM and copper, prior to sputtering the copper layer, a chromium (Cr) thin film of 100 nm in thickness was deposited on the polished ribbon by magnetron RF sputtering. Strips with different lengths and widths were cut from the Metglas TM /copper film. The MSMCs, in lengths from 2 mm to 5 mm, were obtained by clamping the strip at one end using a plastic holder made of poly(methyl methacrylate). 4.3.2 Characterization of Resonance Behavior of MSMC In this research, the resonance behaviors of MSMCs were characterized using the setup A described in Chapter 3. The MSMC was put into a sample chamber that was placed inside of the Helmholtz coil. In this study, a DC magnetic field of 20 Oe was employed. The AC field, which is swept over a broad frequency range, has amplitude less than 3.2 Oe. The magnetic signal emitted from the MSMC was detected by the pair of home-made pick-up coils. The output potential signal of the pick-up coil was measured using a lock-in amplifier (SRS830, Standford research systems, Sunnyvale, CA). The lock-in amplifier has two output signals: amplitude and phase. In this research, the 86 phase signal, as used to determine the resonance frequency of piezoelectric MCs [13], was used to characterize the resonance behavior of MSMCs. To determine the resonance frequency and Q value of MSMCs, the measured phase data were fitted using the Lorentz function as described in Chapter 3. The resonance behaviors of the MSMCs with different dimensions were characterized. The performance of MSMC in liquid was studied by characterizing and comparing the resonance behaviors of the MSMCs in air, water and ethanol. 4.3.3 Characterization of Mass Sensitivity of MSMCs The mass sensitivity of MSMC was characterized by measuring the resonance spectra of the cantilever with different mass loads. In this research, a small piece of tape, in the size about 0.5 mm x 1 mm, was stuck on the free tip of the MSMC as the mass load. The shifts in the resonance frequency of the MSMC with different the mass loads were used to determine the mass sensitivity. 4.3.4 Demonstration of In-situ Yeast Cells Detection Using the MSMC Sensors In-water yeast cell detection was employed to demonstrate the feasibility of MSMCs as biosensor platforms. In the experiment, the surface of the MSMC was coated with a thin layer of poly-L-lysine. The poly-L-lysine offers a positively charged environment to which negatively charged yeast cells are attracted. Yeast cells suspensions were prepared by suspending bakers yeast (Saccharomyces cerevisae, Fleischmann?s Yeast, obtained from a local store) in de-ionized water at room temperature. Two MSMCs were employed in this study. The MSMC in size of 2.8 mm x 1.2 mm x 30 ?m was used to detect yeast cells in suspension with the concentration of 1 mg/ml, while an MSMC in 87 size of 2.6 mm x 1.1 mm x 30 ?m was employed to do the test in yeast suspension with a concentration of 2 mg/ml. To do the real-time test, the whole MSMC was immersed into the yeast suspension, and then its resonance frequency was monitored and recorded as a function of time. After the test, the MSMCs were observed under an optical microscope in a Probe station (Signatone probing division, Gilroy, CA) for viewing the cells bonded on the surface of the MSMCs. 4.3.5 Detection of Bacillus Anthracis Spores Using MSMC Biosensor MSMC biosensors were employed to detect Bacillus anthracis spores in water. In the study, the phage agent for Bacillus anthracis spores was immobilized on the MSMC surface as the biological receptor. The phage displaying the peptide EPRLSPHS on the surface was selected from the f8/8 landscape phage library and provided by the Department of Pathobiology of Auburn University. To immobilize the phage, a thin Au layer (~ 400 nm) was sputtered on both sides of the MSMC to protect against corrosion and provide a favorable binding site for phage, while a chromium layer in thickness of 100 nm was deposited between the Au layer and the MSMC for adhesion purposes. The MSMC with a freshly sputtered Au layer was immersed into 1.5 mL of phage culture with the concentration of 2.6 x 10 12 viruses/mL for 80 minutes. And then the MSMC was rinsed three times using sterile distilled water. The MSMC was now immobilized with phage on all surfaces to form a phage based biosensor and ready for spore detection. The detected agent used in this study was the spores with nonpathogenic Sterne strain of B. anthracis provided by the Department of Pathobiology of Auburn University. The MSMC/phage biosensor was put into the test chamber containing increasing concentrations (5 x 10 4 to 5 x 10 8 spores/mL) of B. anthracis suspensions. 88 And then the shifts in the MSMC?s resonance frequency were monitored and recorded as a function of time. The data was recorded at 5-minute intervals. For each concentration of spore suspension, the data were recorded for 2 hours. 4.4 Results and Discussions 4.4.1 Resonance Behavior of MSMC in air A typical resonance behavior of an MSMC in air at room temperature is shown in Fig. 4-7, where an MSMC in size of 4 mm x 1.5 mm x 30 ?m was utilized. For this MSMC, the first six resonance modes were presented. Based on the data shown in Fig. 4-7, the resonance frequencies are 1.129, 7.877, 22.178, 41.168, 50.157 and 69.466 kHz, for fundamental (n = 0) to 5 th harmonic modes, respectively. The Q values for fundamental to 5 th harmonic modes are 216, 334, 463, 516, 422 and 419 respectively. The Q values of the MSMC obtained here is much higher than that (30~100) of most currently existing regular rectangular microcantilevers [5, 14]. 89 Fig. 4-7. The phase difference between the driving magnetic field and the magnetic signal from a MSMC. The size of the MSMC is 4 mm x 1.5 mm x 30 ?m. For this MSMC, the first six harmonic vibration modes were observed. It is interesting to find that the Q value of fundamental harmonic mode (n = 0) for the MSMC studied here is lower than that of its higher harmonic modes. This is consistent with the results obtained from the ZnO nanocantilever, where the ZnO nanowire is attached to an electrochemically sharpened tungsten (W) wire using epoxy [15]. However, the results obtained from piezoelectric cantilevers, which were rigidly bonded onto a metal holder, showed that the fundamental harmonic peak has the highest Q value [10, 16]. A theoretical simulation for cantilevers with rigid clamping indicates that the Q value is lower for higher modes [17]. Therefore, considering the bonding condition, the fact that a higher Q value observed for higher modes in the MSMCs studied here and the 1.0 1.2 -15 -10 -5 0 7.6 8.0 22.0 22.4 41.2 41.6 50 51 69 70 -15 -10 -5 0 Phase ( d egree) Frequency (kHz) 90 ZnO nanocantilevers is due to the fact that the clamping condition is not strong, which is known as anchor loss [18]. The experimental data indicate that the anchor loss has a stronger influence on fundamental mode than higher modes. One can conclude that if the MSMC is rigidly clamped onto a stiff holder, the Q value of more than 500 would be expected, since the Q value of higher harmonic peaks reaches more than 500. Based on Equation (4-12) and the Eigenvalues of different flexural vibration modes listed in Table 4-1, the ratios of the nth-mode resonance frequency to the fundamental resonance (n = 0) frequency for the first to third mode, and the ratios of the nth-mode resonance frequency to the first resonance (n = 1) frequency for the second and third mode were calculated. The results were listed in Table 4-4. From the experimental data shown in Fig. 4-7, the corresponding resonance frequency ratios were also calculated, and listed in Table 4-4. Comparing the theoretical and experimental results, it is found that the ratios of the first, second and third-mode resonance frequency to fundamental resonance frequency is 6.977, 19.644, and 36.464, respectively, which are slightly higher than that obtained from Equation (4-12). The difference between the theoretical and experimental results is about 10 %. On the other hand, the ratios of the second-mode resonance frequency to the first-mode resonance frequency (which is 2.82), as well as the ratio of the third-mode resonance frequency to the first-mode resonance frequency (which is 5.23) are very close (the difference is less than 5%) to the theoretical results. These results indicate that the anchor loss has a stronger influence on the resonance frequency of the fundamental mode, which is consistent with the conclusion based on the Q value. 91 TABLE 4-4 Theoretical/experimental Resonance Frequency Ratios of the nth-mode Resonance Frequency to the Fundamental Resonance (n = 0) Frequency for the First to Third Mode, and the Ratios of the nth-mode Resonance Frequency to the First Resonance (n = 1) Frequency for the Second and Third Mode. Theoretical value Experimental value Difference f 1 /f 0 6.267 6.977 11% f 2 /f 0 17.548 19.644 12% f 3 /f 0 34.386 36.464 6% f 2 /f 1 2.80 2.816 0.6% f 3 /f 1 5.487 5.226 5% 4.4.2 Resonance Frequency Changing with the MSMC Dimensions Based on Equation (4-12), given material properties and the thickness of the MSMC, the flexural resonance frequency is expected to linearly increase with 1/L 2 . The experimentally obtained resonance frequencies of the MSMCs with different lengths and the same thickness (30 ?m) are plotted versus 1/L 2 in Fig. 4-8, where the frequencies of fundamental (n = 0) and the first (n = 1) resonance mode are presented as solid triangles and solid squares, respectively. As expected from Equation (4-12), the experimental frequencies of MSMCs are linearly dependent on the 1/ L 2 . The theoretical resonance frequencies of the MSMCs were calculated based on Equation (4-12). For the MSMCs investigated here, the E, ?, and ? should be the effective modulus, effective density and effective Poisson?s ratio of the Metglas?/copper film. The density and the mechanical properties of the Metglas? were listed in Table 4-2. 92 Considering the thickness of Metglas? (20 ?m) and the copper (10 ?m ) as well as the density of Metglas? (7.9 g/cm 3 ) and copper (8.9 g/cm 3 ), the effective density of 8.23 g/cm 3 could be used for the MSMC. Regarding the Young?s modulus: although the bulk copper has a Young?s modulus of 130 GPa, it was reported that the Young?s modulus of sputtered copper thin film with a thickness about 1 ?m is about 110 GPa [19]. Therefore, we assume the effective E of 110 GPa for the MSMCs. The Poison ratio of bulk copper and Metglas? is 0.34 and 0.5, respectively. Based on the thickness of copper and Metglas? film, the effective Poison?s ratio of 0.45 is employed. The resonance frequencies of the MSMCs with different lengths were simulated using Equation (4-12) with the effective properties and h = 30 ?m, and are shown in Fig. 4-8 as solid lines. The results indicate that the consistency between the experimental data and the theoretical calculation. 93 Fig. 4-8. Resonance frequency of the fundamental and the first flexural mode vs 1/L 2 for the MSMCs with a thickness about 30 ?m, where L is the length of the MSMCs. The solid triangles and solid squares are the experimental results for the fundamental and the first flexural mode, while the solid lines are results of theoretical calculation. 4.4.3 Stability of the MSMC In the biosensor application, the resonance frequency shift due to the mass load is the output of the MSMC-based device. Therefore, a big concern of the sensor performance is how stable the resonance frequency is. That is, whether or not the resonance frequency would drift with time. The stability for the MSMC was studied by monitoring its fundamental and first harmonic resonance frequency in air over 24 hours. An MSMC, in 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0 5 10 15 20 25 Reso nan ce F r eq uen cy o f MS MC ( k Hz) 1/L 2 (1/mm 2 ) fundamental mode first mode 94 size of 3.05 mm x 1 mm x 30 ?m, was employed in the experiment. The driving frequency was set close to its fundamental resonance frequency to keep the MSMC vibrating, while its resonance frequencies of the fundamental and first mode were recorded at different duration times. During the experiment, the temperature was controlled at 25 ? 0.2 o C, while the humidity changed from 13% to 21%. A set of the results are shown in Fig. 4-9, where the solid squares are the experimental results, and the average of the resonance frequency is given by the solid straight line. Based on the results shown in Fig. 4-9, the average of the resonance frequencies for the fundamental and first harmonic mode are 2258.41 Hz and 14206.75 Hz, respectively. Over 24 hours, the standard deviation of the resonance frequency is 0.288 Hz for the fundamental harmonic resonance and 2.51 Hz for the first harmonic resonance mode. The drift percentage is about 0.01% and 0.017% for the fundamental and first mode, respectively. To estimate the error caused by peak fitting, Table 4-5 and 4-6 shows the resonance frequencies of this MSMC obtained from fitting results using Lorentz function over different frequency ranges for the fundamental and first mode, respectively. The results indicate that about 0.15 Hz and 0.05 Hz variation in the measured fundamental and first resonance frequencies, respectively, are caused by fitting error. Also, considering the variation of the environment condition, the MSMC exhibits good stability. In real applications, any experimental error caused by a frequency drift can be minimized by resonance frequency calibration before use, while the effect of the change in environment can be compromised by employing a sensor array with a reference sensor. 95 Fig. 4-9. Resonance frequency of a MSMC with size of 3.05 mm x 1 mm x 30 ?m measured in air over 24 hours. TABLE 4-5 Fitting Results of the Fundamental Mode Resonance Frequency Based on Lorentz Function over Different Frequency Ranges for the MSMC Used in Stability Study Frequency range (Hz) Baseline f 0 (Hz) 2200 ~ 2350 Linear, D2 2258.5 2225 ~ 2300 Linear, D2 2258.4 2240 ~ 2280 Linear, D2 2258.2 Average Linear, D2 2258.4 ? 0.15 0 5 10 15 20 25 2254 2256 2258 2260 2262 2264 14190 14200 14210 14220 14230 Time (hours) fundamental mode Frequency (Hz) first mode 96 TABLE 4-6 Fitting Results of the First Mode Resonance Frequency Based on Lorentz Function over Different Frequency Ranges for the MSMC Used in Stability Study Frequency range (Hz) Baseline f 0 (Hz) 14100 ~ 14300 Linear, D2 14205.2 14125 ~ 14275 Linear, D2 14205.1 14160 ~ 14260 Linear, D2 14205.0 Average Linear, D2 14205.1 ? 0.05 4.4.4 MSMC Performance in Liquid Since most of the biological analyses are carried out in the liquid environments, the performance of biosensors in liquid is very important for real-time or in-vivo detection. To study the performance of the MSMC in liquid, the resonance behavior of an MSMC, in size of 3.3 mm x 1 mm x 30 ?m, was characterized in air, distilled water and ethanol. When measuring the resonance frequencies of the MSMC in liquid, the whole device was immersed into the distilled water or ethanol. The resonance spectra of the fundamental and first mode of this MSMC in different media are shown in Fig. 4-10 (a) and (b), respectively. In the liquid media, the oscillating cantilever beam has to push the surrounding liquid, which acts as induced mass. Also, the viscous drag of the liquid adds additional force on the cantilever beam. Therefore, the resonance frequency, signal amplitude and the Q value of the MSMC decrease due to the dampling effect caused by the surrounding liquid. Based the results shown in Fig. 4-10, the performance of this 97 MSMC is summarized in Table 4-7. For this MSMC, whose Q value is 191 and 265 for the fundamental mode and the first mode, respectively, when operated in air, it is experimentally found that the Q value is 19 and 38 for the fundamental and the first mode respectively when it is operated in water, and 9 and 27 for the fundamental and the first mode respectively when it is operated in ethanol. Considering that other cantilevers in aqueous solution rarely have a Q value above 10, the Q value of the MSMC in water is clearly higher than that of the other microcantilevers [5]. Fig. 4-10. MSMC performance in liquid. (a) Fundamental mode resonance spectra for a MSMC in different media. (b) First flexural mode resonance spectra for a MSMC in different media. 800 1000 1800 2000 -2.0 -1.5 -1.0 -0.5 0.0 Phase (degree) Frequency (Hz) in ethanol in water in air fundamental mode (a) 6000 12000 -2.0 -1.5 -1.0 -0.5 0.0 P hase (degree) Frequency (Hz) in ethanol in water in air first mode (b) 98 TABLE 4-7 MSMC Performance in Liquid fundamental harmonic first harmonic air water ethanol air water ethanol Resonance frequency (Hz) 1885 871 907 11912 6915 6182 Q value 191 16 9 265 38 27 According to the literature, the viscosity of water and ethanol is 0.89 x 10 -3 kg/m?s and 1.08 x 10 -3 kg/m?s at room temperature, respectively [20]. The densities of water and ethanol are 1 x 10 3 kg/m 3 and 0.806 x 10 3 kg/m 3 , respectively [21]. Clearly, the Q value of MSMC reduces with the increasing viscosity of the liquid. One may notice that although the Q value of the MSMC in water is higher than that in ethanol, the resonance frequency of this MSMC is lower in water. The reason behind this phenomenon may be that the density of water is higher than that of ethanol, which causes higher induced mass and bigger decrease in the resonance frequency of the MSMC. This indicates that although both the viscosity and density of the liquid affect the resonance frequency and Q value of the MSMC, the density has a stronger influence on the resonance frequency and the viscosity has a stronger effect on the Q value. 4.4.5 MSMC Array In sensor applications, an array of sensors is highly desired, since a sensor array has many advantages over individual sensors. The change in the environment, such as temperature, varies the properties of the material, and thus affects the output resonance frequency of the device. This effect can be compromised by employing a sensor array 99 that includes reference sensors. Additionally, by coating sensors with different recognition elements, a sensor array can detect multiple targets simultaneously. For current MCs, a sensor array is difficult to operate. To form a Si-based MC array, each MC requires an optical system for sensing; while for piezoelectric MC array, individual electric circuit for operating each MC is needed. As wireless sensors, arrays of MSMCs can be easily formed, and actuated/sensed by a single scan. Fig. 4-11 shows a MSMC array consisting of three MSMCs: MSMC1 (3.5 mm x1.3 mm x 30 ?m), MSMC2 (3.3 mm x 1.3 mm x 30 ?m), and MSMC3 (2.5 mm x 1.2 mm x 30 ?m). As shown in Fig. 4-11, using the same setup for measuring single MSMC, the MSMC array can be driven and sensed by a single scan. Fig. 4-11. Scheme of a MSMC array characterized using setup A. Fig. 4-12 shows the spectrum of this MSMC array measured by a single scan. Three resonance peaks were observed in the spectrum. The resonance frequencies of the three peaks are 1.35, 1.73, and 3.02 kHz, which are corresponding to the fundamental resonance frequencies of MSMC1, MSMC2 and MSMC3, respectively. The results indicate that MSMC array can be easily formed and characterized. 1 2 3 Helmholtz coil Helmholtz coil pickup coil MSMC array 100 Fig. 4-12. Spectrum of a MSMC array consisting of three MSMCs. 4.4.6 Characterization of the Mass Sensitivity of MSMC The mass sensitivity of the MSMC was characterized by measuring the resonance spectrum with different mass loads, where the mass load is a small piece of tape, in the size about 0.5 mm x 1 mm, attached at the tip of the MSMC. The resonance frequency shifts to a lower frequency with a mass load. One set of experimental results for determining the mass sensitivity of the MSMC is shown in Fig. 4-13, where an MSMC with a size of 4 mm x 1.2 mm x 30 ?m was used. The fundamental and first harmonic resonance spectra of this MSMC with different mass loads are shown in Fig. 4-13 (a) and (b), respectively. With the mass load increased up to 0.306 mg, a 295 Hz shift for fundamental harmonic resonance and an 869 Hz shift for first harmonic resonance were observed. The shift in the resonance frequency versus mass load is plotted in Fig. 4-13 (c). 1.0 1.5 2.0 2.5 3.0 3.5 83 84 85 86 87 88 resonance peak of MSMC 3 resonance peak of MSMC 2 Phase Frequency (kHz) resonance peak of MSMC 1 101 Based on the data shown in Fig. 4-13 (c), a mass sensitivity of 1003 Hz/mg and 2577 Hz/mg was obtained for fundamental and first harmonic modes respectively. Fig. 4-13. S m characterization of an MSMC (4 mm x 1.2 mm x 30?m): (a) The fundamental harmonic mode resonance spectra for the MSMC with different mass loads; (b) The first harmonic mode resonance spectra for the MSMC with different mass loads. 78 80 82 84 86 88 90 0.7 0.8 0.9 1.0 1.1 1.2 1.3 no mass (a) Phase (degree) 0.048 mg 0.098 mg Frequency (KHz) 0.23 mg 0.306 mg fundamental mode 6.25 6.50 6.75 7.00 7.25 7.50 7.75 73 74 75 76 77 78 79 80 Frequency (KHz) no mass first mode Pha s e (d eg ree ) 0.048 mg (b) 0.098 mg 0.23 mg 0.306 mg 102 Fig. 4-13. Mass sensitivity characterization of an MSMC (4 mm x 1.2 mm x 30 ?m): (c) ?f versus ?m of the MSMC for the fundamental and first harmonic resonance. Since mass loads were added on the tip of the MSMC in this research, we may use Equation (4-17) to simulate the mass sensitivity of the MSMC. Based on Equation (4-17), the theoretical mass sensitivity of a MSMC in size of 4 mm x 1.2 mm x 30 ?m is 2029 Hz/mg and 12717 Hz/mg for the fundamental and first harmonic modes, respectively. It is found that the experimental result is lower than the calculated mass sensitivity. The reason behind this is that Equation (4-17) assumed that the mass load is a point mass at the tip of the cantilever, while, in this experiment, the size of the mass load (0.5 mm x 1 mm) for sensitivity characterization is comparable to the size (4 mm x 1.2 mm) of the MSMC. 0.0 0.1 0.2 0.3 0.4 0.0 0.2 0.4 0.6 0.8 1.0 R e son ance Fr equency shif t ( k H z ) first mode fundamental mode mass load (mg) (c) 103 4.4.7 Yeast Detection in Water To demonstrate the feasibility of using the MSMC as a sensor platform, the MSMC was used to detect the yeast cells in real-time manner. A set of data is shown in Fig. 4-14. The solid triangles are the results of testing yeast cells in yeast cell suspension with 1 mg/ml concentration, where an MSMC (MSMC-A) in size of 2.8 mm x 1.2 mm x 30 ?m was utilized; and the solid squares are the results of testing yeast cells in 2 mg/ml yeast suspension, where an MSMC in size of 2.6 mm x 1.1 mm x 30 ?m (MSMC-B) was utilized. The results were obtained by monitoring the change in the resonance frequency of fundamental harmonic mode with time after the whole MSMC was immersed into the yeast cell suspension. The exponential trend lines were drawn as solid lines based on experimental results. The results show that the MSMC sensor works well in liquid and can provide real-time detection. To determine the resolution of the MSMC operating in liquid, the phase signals of MASMC-A and MSMC-B in pure water were fitted based on Lorentz function over different frequency ranges, and the obtained resonance frequencies were listed in Table 4-8 and Table 4-9, respectively. The results indicate the fitting error is ? 0.7 Hz. Therefore, the shift of 1.5 Hz in the resonance frequency can be identified. 104 Fig. 4-14. Response of fundamental harmonic mode for real-time yeast cells detection using MSMCs: (a) frequency shift with time after MSMC-A and MSMC-B immersed in yeast suspension with concentration of 1 mg/ml and 2 mg/ml, respectively; (b) the picture of yeast cells on the MSMC-B surface after the sensor was tested in 2 mg/ml yeast cell suspension. (c) the picture of yeast cells on the MSMC-A surface after the sensor was tested in 1 mg/ml yeast cell suspension. TABLE 4-8 Fitting Results of Resonance Frequencies of MSMC-A in Pure Water Based on Lorentz Function over Different Frequency Ranges Frequency range (Hz) Baseline f 0 (Hz) 1300 ~ 1600 Linear, D2 1437.1 1325 ~1550 Linear, D2 1436.3 1300 ~ 1550 Linear, D2 1435.7 Average Linear, D2 1436.37 ? 0.7 (a) 10?m 10u m 1mg/ml 2mg/ml 0 102030405060 0 10 20 30 40 50 60 70 80 90 100 Time (mins) Shi f t (Hz ) (b) (c) 105 TABLE 4-9 Fitting Results of Resonance Frequencies of MSMC-B in Pure Water Based on Lorentz Function over Different Frequency Ranges Frequency range (Hz) Baseline f 0 (Hz) 1450 ~ 1800 Linear, D2 1612.5 1500 ~ 1750 Linear, D2 1611.4 1550 ~ 1700 Linear, D2 1612.8 Average Linear, D2 1612.23 ? 0.7 It is found that the Q values of the MSMC-A and MSMC-B in the yeast cell suspension are 21 (the Q value of MSMC-A is 130 in air) and 15 (the Q value of MSMC- B is 120 in air), respectively. As discussed above, the MSMC-A and MSMC-B can easily detect a shift in resonance frequency at least 1.5 Hz. Based on the data shown in Fig. 4-12, an almost saturated shift in resonance frequency is obtained after the MSMCs were immersed in yeast suspension for 1 hour. The saturated shifts are about 35 Hz and 85 Hz for the detection in 1 mg/ml and 2 mg/ml yeast cell suspension, respectively. The optical micrograph of immobilized yeast cells on MSMC surface were taken and shown in Fig. 4-14 (b) and (c) after the MSMCs were immersed in yeast cell suspension for 1 hour. The images indicate the obtained shift in resonance frequency is due to the yeast cells attachment. 4.4.8 Detection of Bacillus Anthracis Spore Using MSMC-based Biosensor Bacillus anthracis spore is a fatal bio-threat agent. Prevention of diseases caused by B. anthracis spores depends on immediate detection and control of the pathogenic spores in 106 the environment. Although the conventional method is sensitive, it requires incubation and takes several days to complete. Combining landscape phase probes with MSMC platform provides a rapid bio-analysis tool to monitor the B. anthracis spores. MSMC/Phage biosensors, in the size of 2.7 mm x 1 mm x 35 ?m, were employed to demonstrate real-time detection of B. anthracis spores in water. Fig. 4-15 shows the real time resonance frequency shifts of a MSMC biosensor, when it was exposed to spore suspensions with increasing concentrations. The black squares, red dots and green triangles represent the resonance frequency shifts as the MSMC was exposed to 5 x 10 4 , 5 x 10 6 , and 5 x 10 8 spores/mL of spore suspensions, respectively. As the MSMC was exposed to 5 x 10 4 spores/mL of spore suspension, no obvious shift can be observed. When the MSMC was exposed to spore suspensions with the concentration of 5 x 10 6 and 5 x 10 8 spores/mL, the resonance frequency shifts quickly at first due to the binding of the spores, and then reaches the saturation after about 1 hour exposure. The higher the spore concentration, the bigger the total resonance frequency shifts. For the concentration of 5 x 10 6 spores/mL, a total frequency shift about 15 Hz was observed, while a total frequency shift about 50 Hz was observed for the concentration of 5 x 10 8 spores/mL. 107 Fig. 4-15. Real-time resonance frequency shifts as the MSMC/Phage biosensor exposed to B. anthracis suspension with increasing concentration. The MSMC biosensor is in the size of 2.7 mm x 1.0 mm x 35 ?m. After the sensor was exposed to spore suspensions, the MSMC was rinsed using distilled water and dried, and then the surface of the MSMC was observed using a SEM. The SEM images shown in Fig. 4-16 confirm the binding of the spores on the surface, and indicate that obtained shifts in resonance frequency were due to the spores? attachment. The images clearly show that the phage clone used in this study exhibits good binding capability. It also revealed that the spores? binding was not uniform: more 0 60 120 180 240 300 360 0 10 20 30 40 50 60 Fr equenc y Shift (Hz) Time (minutes) Concentration: 5*10 4 spores/mL Concentration: 5*10 6 spores/mL Concentration: 5*10 8 spores/mL 108 spores attached at end of the MSMC, while much less spores attached at the middle of the cantilever beam. According to Equation (4-14) and (4-17), mass load at free end of the beam results in a higher sensitivity and is favorable for the detection. Fig. 4-16. SEM images of the MSMC surface after the biosensor was exposed to the spore suspension with increasing concentration. (a) at the tip of the cantilever beam; (b) at the middle of the cantilever beam. (a) (b) 109 When the MSMC/phage biosensor was exposed to spore suspension, the shift in its resonance frequency eventually reached saturation if the concentration of the suspension was not changed. Fig. 4-17 shows the saturated resonance frequency shift as a function of the spore concentration, which was measured by the phage-coated MSMC biosensors in the size of 2.7 mm x 1 mm x 35 ?m. A good dose-response relationship was observed between 10 6 spores/mL to 10 8 spores/mL. From the results shown in Fig. 4-17, a detection limit of 10 5 spores/mL is obtained for the MSMCs used in this research. Fig. 4-17. The saturated resonance frequency shifts as the MSMC/phage biosensors, in the size of 2.7 mm x 1 mm x 35 ?m, were exposed to spore suspensions with different concentrations. 10 5 10 6 10 7 10 8 0 5 10 15 20 25 30 35 40 S a t u r a t e d S h i f t i n R e so na nce Fr eq uen cy (H Z ) Concentration of Bacillus Anthracis Spore ( spores/mL) 110 The detection limit is affected by the binding of spore cells on the MSMC, as well as the mass sensitivity of the MSMC. The study of spore/phage bound density on the gold- coated Metglas TM 2826MB particles revealed that, for small size particles (size comparable to the MSMC beam used in this research), the chance of spore binding was pretty low for low spore concentration (< 10 4 spores/mL) [22]. At a concentration of 5 x 10 6 spores/mL, good spore binding can be observed, and dense spore binding was observed as the particle was exposed to a 5 x 10 8 spores/mL spore suspension [22]. This is consistent with the results shown in Fig. 4-15. It is found that, for the MSMCs used in this experiment, whose resonance frequency (f 0 ) and Q value are about 2600 Hz and 200 in air, respectively, its f 0 and Q value are about 1400 and 20 when it is operated in water. Based on the discussion in previous, the resolution of MSMCs used in this experiment is about 1.5 Hz when operated in water. Based on Equation (4-14), assuming the spores uniformly binding on the surface of the MSMC, the theoretical mass sensitivity of the MSMC used in this study is about 0.6 ?g/Hz. Thus, it is difficult to obtain observable frequency shift if the attached mass is less than 1 ?g. If we assume the spores bind at the free end of the MSMC, based on Equation (4-17), theoretically, a mass load of about 0.2 ?g is needed for a detectable frequency shift. Considering that a spore cell weights about 2 pg, for the MSMC used in this research, a dense spore binding is required, thus a good dose-response relationship was observed only as the spore concentration higher than 5 x 10 6 spores/mL. Based on Equation (4-14) and (4-17), the mass sensitivity of the MSMC exponentially increases with the decreasing size of the MSMCs. The size of the MSMC used in this experiment is still relatively large. As demonstrated by other MCs, it is expected that the sensitivity and 111 the detection limit can be great improved if the MSMC with a smaller size is employed. 4.5 Conclusion The magnetostrictive microcantilever (MSMC) is introduced as a new type of MC as sensor platform for developing a high performance biosensor. The performance of the MSMCs was experimentally determined using the MSMCs made of a commercial magnetostrictive alloy?Metglas?. Compared to the piezoelectric MCs, the MSMC has a simple configuration. Based on material properties, it is expected that the MSMC sensors would exhibit a higher sensitivity than the piezoelectric MCs. More importantly, it is experimentally found that the MSMC works well in air or liquid and exhibits a much higher Q value than the other MCs. For example, the Q value of those MSMCs in air reaches more than 500 and that in water reaches 38. The experimental results also indicate that the anchor loss in MC has a stronger influence on the fundamental mode than on the higher modes. As wireless sensors, an array of MSMCs can be easily formed and characterized by a single scan. The application of the MSMC as a high performance biosensor platform is demonstrated by detecting yeast cells and B. anthracis spores in water. The MSMCs exhibit advantages over other microcantilevers in high performance biosensor development. 112 References [1] P. Leonard, S. Hearty, J. Brennan, L. Dunne, J. Quinn, T. Chakraborty, and R. O'Kennedy, "Advances in biosensors for detection of pathogens in food and water," Enzyme and Microbial Technology, vol. 32, pp. 3-13, 2003. [2] D. S. Ballantine, White, R.M., Martin, S.J., Ricco, A.J., Frye, G.C., Zellers, E.T., Wohltjen, H., Acoustic wave sensors: theory, design and physico-chemical applications: Academic Press, 1997. [3] O. Tamarin, C. Dejous, D. Rebiere, J. Pistre, S. Comeau, D. Moynet, and J. Bezian, "Study of acoustic Love wave devices for real time bacteriophage detection," Sensors and Actuators B-Chemical, vol. 91, pp. 275-284, 2003. [4] R. Raiteri, M. Grattarola, H. J. Butt, and P. Skladal, "Micromechanical cantilever- based biosensors," Sensors and Actuators B-Chemical, vol. 79, pp. 115-126, 2001. [5] C. Ziegler, "Cantilever-based biosensors," Analytical and Bioanalytical Chemistry, vol. 379, pp. 946-959, 2004. [6] B. Ilic, Y. Yang, K. Aubin, R. Reichenbach, S. Krylov, and H. G. Craighead, "Enumeration of DNA molecules bound to a nanomechanical oscillator," Nano Letters, vol. 5, pp. 925-929, 2005. [7] Z. Y. Shen, W. Y. Shih, and W. H. Shih, "Mass detection sensitivity of piezoelectric cantilevers with a nonpiezoelectric extension," Review of Scientific Instruments, vol. 77, 2006. [8] B. Ilic, Y. Yang, and H. G. Craighead, "Virus detection using nanoelectromechanical devices," Applied Physics Letters, vol. 85, pp. 2604-2606, 2004. [9] J. Merhaut, Theory of electroacoustics: New York ; London : McGraw-Hill International Book Co., 1981. [10] J. W. Yi, W. Y. Shih, and W. H. Shih, "Effect of length, width, and mode on the mass detection sensitivity of piezoelectric unimorph cantilevers," Journal of Applied Physics, vol. 91, pp. 1680-1686, 2002. [11] Guide to modern piezoelectric ceramics: Morgan Matroc, Inc. [12] S. M. Sze, Semiconductor Sensors: John Wiley & Sons, Inc., 1994. [13] W. Y. Shih, X. P. Li, H. M. Gu, W. H. Shih, and I. A. Aksay, "Simultaneous liquid viscosity and density determination with piezoelectric unimorph cantilevers," Journal of Applied Physics, vol. 89, pp. 1497-1505, 2001. 113 [14] B. Ilic, D. Czaplewski, M. Zalalutdinov, H. G. Craighead, P. Neuzil, C. Campagnolo, and C. Batt, "Single cell detection with micromechanical oscillators," Journal of Vacuum Science & Technology B, vol. 19, pp. 2825-2828, 2001. [15] K. Yum, Z. Y. Wang, A. P. Suryavanshi, and M. F. Yu, "Experimental measurement and model analysis of damping effect in nanoscale mechanical beam resonators in air," Journal of Applied Physics, vol. 96, pp. 3933-3938, 2004. [16] J. W. Yi, W. Y. Shih, R. Mutharasan, and W. H. Shih, "In situ cell detection using piezoelectric lead zirconate titanate-stainless steel cantilevers," Journal of Applied Physics, vol. 93, pp. 619-625, 2003. [17] S. Rast, C. Wattinger, U. Gysin, and E. Meyer, "Dynamics of damped cantilevers," Review of Scientific Instruments, vol. 71, pp. 2772-2775, 2000. [18] Y. H. Park and K. C. Park, "High-fidelity Modeling of MEMS resonators - Part 1: Anchor loss mechanisms through substrate," Journal of Microelectromechanical Systems, vol. 13, pp. 238-247, 2004. [19] D. T. Read, "Young's modulus of thin films by speckle interferometry," Measurement Science & Technology, vol. 9, pp. 676-685, 1998. [20] P. H. Rieger, Electrochemistry: Prentice-Hall, Inc., 1987. [21] R. A. Serway, Physics for scientists and engineers, second ed: CBS College Publishing, 1986. [22] J. Wan, H. Shu, S. Huang, B. Fiebor, I.-H. BChen, V. A. Petrenko, and B. A. Chin, "Phage-Based Magnetoelastic Wireless Biosensors for Detecting Bacillus Anthracis Spores," IEEE Sensors Journal, vol. 7, pp. 470 - 477, 2007. 114 CHAPTER 5 DESIGN AND FEASIBILITY STUDY OF MAGNETOSTRICTIVE PARTICLES AS BIOSENSOR PLATFORM 5.1 Introduction Development of biosensors based on AW devices has attracted a great deal of attention due to the fact that AW devices can provide high sensitivity and real-time detection [1-3]. However, as mentioned in Chapter 1, current AW devices face challenges in practical applications. Some AW devices, such as TSM, can work well in liquid, but have a small S m ; while some AW devices, such as MCs, exhibit a very high S m , but with a small Q value. There is an urgent need to develop high performance portable biosensors that can be used to rapidly quantify the presence of various pathogens. Also, the biosensors that are able to perform in vivo detection are highly desirable for applications such as in-body medical diagnosis, urine and excrement tests, and quality monitoring of food during processing, storage and transportation. Another challenge facing biosensors is how to bring target species onto sensors, especially when detecting pathogen cells in very low concentration environment. For example, MCs in nano-scale have achieved the capability to detect a single cell. However, based on the study in bio-reaction dynamics, the chance of the target pathogen cells binding on the sensor platform is very low, and a very long response time is required to generate measurable signals, when the sensors are 115 in very small size or when they are used to detect pathogens in very low concentration cultures [4]. Magnetic beads have been widely used in bio-detection techniques, such as biological separation, labeling and targeting, since they provide an efficient way to capture and isolate biological targets such as DNA, cell, and protein [5]. Magnetic bead consists of an inorganic core of iron oxide, such as magnetite Fe 2 O 3 , magnetite or other insoluble ferrites, coated with polymer such as dextran. Their size is from nanometers to micrometers in diameters. Bounded to a suitable antibody, they can selectively attach to the biomaterial of interest. For manipulation or purification of biological cells/molecules, magnetic beads are mixed and stirred in the solution. And then suspended beads tagged to the biomaterial of interest can be removed from the surrounding environment using a magnetic field [5]. By this way, magnetic beads can be used to separate and enrich the target cells to make low number cell detection easier. Magnetic beads have been combined with biosensor techniques to improve the biosensors? performance in very low concentration environments. For example, Brewster et al. combined antibody-coated superparamagnetic beads with the electrochemical biosensor to detect S. typhimurium [6]. Target bacteria cells were captured by antibody- coated magnetic beads, then separated and purified from the sample solution. With the aid of the magnet, the beads were localized on the surface of the disposable graphite ink electrodes. By adding an enzyme solution, the bacteria can be detected using a voltammetric system. Using this technique, a minimum of 8x10 3 cell/ml of S. typhimurium can be detected [6]. More recently, Nam et al. also reported a nanoparticle- based bio-bar-code approach to detect a protein target at low attomolar concentrations [7]. 116 With the use of the magnetic bead as separator, antibody functionalized magnetic particles can readily remove the aggregate from the assay solution to leave only the bar- code DNA, which can be quickly identified by standard DNA detection methodologies, such as gel electrophoresis, fluorophore-labeling, and scanometric approaches. This approach can improve the detection limitation about six orders compared to conventional methods [7]. However, these approaches are both labor intensive and time-consuming due to the separation and many washing steps. Also, complex instrumentation is required for the detection. In this chapter, a novel AW device ? magnetostrictive particle (MSP) in the shape of strip or bar ? is introduced as a high performance biosensor platform. A MSP is a free standing magnetostrictive sensor platform operated based on magnetostriction. A single MSP can work as an individual sensor, while a number of MSPs can also work together as multiple sensors. Interrogated through magnetic field, there are no physical connections between the sensor and interrogation system required for driving and sensing MSPs. Therefore, MSPs are wireless sensors, which enable remote in vivo detection. Furthermore, the theoretical simulation indicates that the MSPs have much better sensitivity than MCs in the term of minimum detectable mass; and the feasibility study demonstrates that the MSPs exhibit a high Q-value and work well in liquid media. The presented MSP biosensors are cost effective and disposable, since the magnetostrictive alloys are very inexpensive and environmental friendly. The MSP also exhibits unique features as a sensor platform. MSPs have the advantages to overcome the limitations of currently existing sensors in detecting target species with low concentration. Since MSPs are magnetic, similar to magnetic beads, 117 they can be stirred using a magnetic field in the sample solution, which greatly increases the chance of target species binding on the MSP sensor. Meanwhile, since the individual MSP itself is a biosensor platform, it can directly output the signal without combining to some other sensor platforms. Based on this feature, a simple and efficient approach is presented to detect target species with very low concentration. MSPs also can overcome the effect of non-uniform mass load on the sensor surface. AW mass sensors exhibit maximum sensitivity when the mass is loaded at the place where the device undergoes maximum displacement during the resonance vibration, but can not detect the mass loaded at the place where zero displacement occurs during the vibration. Since the MSP is totally free standing, multiple harmonic resonance modes can be easily actuated and sensed by a single scan. Even with non-uniform mass load, the sensitivity of MSPs can be fully used by monitoring multiple harmonic resonance frequencies. 5.2 Design and Operation Principle of Magnetostrictive Particles (MSPs) 5.2.1 Configuration of MSPs and Their Resonance Behavior The MSP sensor platform has very simple configuration: it is a free-standing particle made from a magnetostrictive alloy. In the design, the MSP can be in the shape of strip with the length of L, width of w and thickness of t, or in the shape of bar with the length of L and diameter of d. For the MSP in strip shape, its length is bigger than its width and much bigger than its thickness (L > w > t). For the MSP in bar shape, its length is much bigger than its diameter (L >> d). The configuration is shown in Fig. 5-1. 118 Fig. 5-1. Magnetostrictive particle (MSP) configuration. (a) MSP in strip-shape and (b) MSP in bar-shape. Based on the magnetostrictive effect, the MSP will be deformed under an external magnetic field. Thus, if an alternating magnetic field is applied, extensions or compressions along all directions would be generated. In the design, the resonance along the length direction, that is the longitudinal vibration mode, is used. Assuming the MSP is oriented in its basal XY plane, with the length of the MSP along the x-axis as shown in Fig. 5-1, the vibration of the MSP along its length direction (x-direction) can be described using the theoretical model that describes the longitudinal oscillation of a thin elastic (b) L d L >> d z x y o (a) y L t w L > w > t z x o 119 plate, as follows [8]: 0 ),( )1( ),( 2 2 22 2 = ? ? ? ? ? ? x txuE t txu ?? (5-1) where E, ?, and ? are the Young?s modulus, density, and Poisson ratio of the material respectively, while the u(x, t) is the displacement vector along x-axis. For harmonic oscillations, the displacement can be expressed as [9]: ti exutxu ? )(),( = (5-2) where ? is the resonance frequency. Therefore, Equation (5-1) can be written as [9]: 0)( )( )1( 2 2 2 2 =+ ? ? ? xu x xuE ? ?? (5-3) The general solution of the wave Equation (5-3) is [9]: )sin(cossin)( 0 ?+=+= kxukxBkxAxu (5-4) where E k )1( 2 ?? ? ? = is the wave number [9], and 22 0 BAu += is the displacement amplitude. When a free-standing particle shown in Fig. 5-1 is operated, the two ends of the MSP experience no restoring force and are stress free, which means that 0/)( =?? xxu at both ends of the MSP (x = 0 and x = L). Applying these boundary conditions to Equation (5-4), we can obtain: 0)cos( 0 =?ku (5-) 0)cos( 0 =+ ?kLku (5-6) Equation (5-5) and (5-6) yields: 120 2 ? ? = (5-7) ?nkL = (n = 1, 2, 3?) (5-8) From the condition of Equation (5-8), it yields: )1()1( 2 22 ?? ? ?? ?? ? = ? == E L nE kf nn (5-9) Therefore, for the magnetostrictive strip or bar shown in Fig. 5-1, its characteristic frequency of the longitudinal oscillation is given by [8]: )1(2 2 ?? ? = E L n f n (n = 1, 2, 3?) (5-10) Equation (5-10) indicates that the characteristic frequency of the MSPs with the designed structure is only dependent on their length and material properties. Considering the boundary conditions of the MSP oscillation allows the amplitude profile of the displacement at resonance to be calculated. Inserting Equation (5-7) and (5- 8) into Equation (5-4), we can obtain the displacement along the MSP length at longitudinal resonance, which is [10]. tj ex L n utxu ? ? )cos(),( 0 = (5-1) where u 0 is the displacement amplitude and ? is the angular excitation frequency (? = 2?f). In Equation (5-11), tj e ? represents the oscillation of each point of the MSP, while )cos( 0 x L n u ? represents the amplitude profile of displacement for the MSP along the length direction at resonance frequency. 121 5.2.2 The Mass Sensitivity (S m ) of MSPs 5.2.2.1 Mass Sensitivity (S m ) of MSPs with Uniform Mass Load If an ideal pure mass ?m is uniformly added on the surface of a MSP shown in Fig. 5- 1, the additional mass corresponds to an increase in the MSP density. Assuming the original mass of the MSP is M, with the mass load ?m, the density of the MSP increases by a factor of ?1 + ?m/M?. Thus Equation (5-1) becomes: 0 )1()1( 2 2 2 2 2 = ? ? ? ? + ? ? ? x u M m E t u xx ?? (5-12) Based on Equation (5-10), the characteristic frequency of a MSP with uniform mass load ?m is: mM M f M m E L n f nmass ?+ = ? ? ? ? ? ? ? ? + = )1(1 2 2 ?? (n = 1, 2, 3?) (5-13) Therefore, the change in the resonance frequency (?f) caused by the mass load (?m) can be expressed as: m M m f m ff m f n massn ? ? + ? = ? ? = ? ? 1 1 1 (5-14) When the mass load (?m) is much smaller than the mass (M) of the MSP (?m/M << 1), Equation (5-14) can be expressed as: m M m f m M m f m ff m f nn massn ? ? +? ? ? ? ?? ? ? ? = ? ? ) 2 1 1(1 11 (5-15) 122 Therefore, for a small mass load, we can obtain: M f m f n 2 ?= ? ? (?m< 10?m) deposited from different solutions were characterized by this system. For thin films, the diffraction peaks of Cu substrate can be observed. 173 The Grazing Incidence X-ray Diffractions were performed using the Philips X'Pert system. The X-ray source is a long-fine-focus, ceramic X-ray tube with Cu anode. The operating power is 45 KV and 40 mA. The incident angle was fixed at 5 o . The XRD pattern can be used to estimate the crystalline size of the materials. The diameter of the crystal size can be calculated using Sherrer?s equation [16]. ? ? cos 9.0 B t = (6-1) where t is the dimension of the crystal or the grain, ? is the wavelength of the X-ray, B is the width of the diffraction peak at its half maximum, and 2? is the peak position. 6.2.4.2 Composition Determination using AES and XPS The compositions of the surface of the deposited Fe-B films were analyzed using X- ray photoelectron spectroscopy (XPS), while the composition distribution along the Fe-B thickness was analyzed using Auger electron spectroscopy (AES). Both AES and XPS are surface analysis techniques to determine the elemental composition and the chemical state of the atoms of the materials. XPS technique employs monoenergetic X-ray to irradiate the sample. The energy of X-ray photons is absorbed by atoms at the sample surface, causing the emission of electrons due to the photoelectric effect. The kinetic energy of emitted electrons represents the binding energy of the atoms of the materials. Since each element has a unique set of binding energies, by measuring the energies of photoelectrons emitted from the atoms, XPS can be used to identify the elements and determine the composition of the materials. The elemental binding energies vary with the chemical potential of the atom and polarization of compounds. Therefore, XPS can also used to identify the chemical state of the materials by analyzing the shifts in 174 the elemental binding energies [17]. In this research, the compositions of the surface of the Fe-B films were analyzed by XPS, which was performed using a Physical Electronics Quantum 2000 Scanning ESCA Microprobe. This system uses a focused monochromatic Al K? x-ray source and a spherical section analyzer. The X-ray beam with 100 ?m diameter was employed and incident normal to the sample, while the photoelectron detector was at 45? off-normal. The scan area was 1.4 mm x 0.2 mm rectangle on the sample. The compositions of the Fe-B film surface, and at depths of approximately 5 nm, 15 nm, 35 nm were analyzed by etching the film using 2 kV Ar + ion-beam. The sputter rate for these ion gun conditions was calibrated at a rate of 4.3 nm/min for a known SiO 2 reference material. AES is widely used to determine elemental composition of materials by analyzing the Auger electrons emitted from the excited sample. The Auger electron emission is initiated by bombarding the sample surface with electrons. The incident electron can knock off the inner shell electrons of the sample atom and create an excited ion with inner shell vacancy. To relax to lower energy state, an electron from higher energy level drops to the inner shell vacancy, which results in the simultaneous emission of an Auger electron. The kinetic energy of the Auger electron equals to the energy difference of the ionized state and the relaxed final state, and is characteristic to the structure of the atom. By detecting the number of Auger electrons as a function of electron energy, AES can determine the elements and the concentration of the elements in the sample material. Except H and He, AES can detect all elements with high sensitivity [18]. To study the uniformity of the deposited Fe-B film along its growth direction, the composition profile along film thickness direction was analyzed by AES, which was 175 performed using a Physical Electronics Model 680 Nanoprobe. Fe-B films, 3 ~ 4 ?m in thickness, deposited on Cu/Cr/glass substrate, were broken. The e-beam with a diameter of ~20 nm was focused on the cross-section surface of the films, and then the compositions of selected spots along the thickness were analyzed by the AES. Before the analysis, cross-section surface was sputtered off 50 nm to eliminate the contamination and the oxidation. The sputter rate for the ion gun was calibrated at a rate of 1.2 nm/min for a known SiO 2 reference material. 6.2.4.3 Magnetization Loop Measurement The magnetization (M-H loop) of the Fe-B thin films were measured by a DSM model Vibrating Sample Magnetometer (VSM) with the H field applied parallel to the surface of the Fe-B thin films. The specimens used for the measure were the Fe-B thin films with the Cu/Cr (total 2.5 ?m) substrate, as shown in Fig. 6-2. Two groups of Fe-B thin films were studied. One group consisted of the films deposited from the solutions with different KBH 4 concentration (plating bath D, E and F). The thickness of the Fe-B films is 12 ?m. Another group consisted of the films deposited form solution D, but with different thicknesses that are 50 nm, 3 ?m, and 12 ?m. Fig. 6-2. Illustration of Fe-B film specimens used for magnetization loop measurement. Fe-B film Cu/Cr substrate H 176 6.2.4.4 Resonance Behavior Characterization of Fe-B Alloy The resonance behavior of Fe-B alloy was examined by measuring the resonance behavior of strip-like magnetostrictive resonators made from deposited Fe-B films. Three groups of resonators made of amorphous Fe-B films, deposited under current density of 5 mA/cm 2 from plating bath D, E and F, respectively, were studies in this research. The Fe-B films in the thickness of 12 ?m was deposited on the Cu(2.5 ?m)/Cr(100 nm) coated glass substrate, and then the film was peeled off from the glass slide with the Cu/Cr subsrate. The Fe-B/Cu/Cr film was cut to strips with different sizes by hand to form resonators. The resonance spectra of the resonators were measured using the set up B, which was descried in Chapter 3. The magnitude of S 11 output from the analyzer is used to characterize the resonance behavior of the Fe-B resonators. The determination of the resonance frequency and Q value of the MSP was discussed in Chapter 3. 6.2.4.5 Thermal Stability of the Fe-B Thin Film The thermal stability of the Fe-B thin films was studied by annealing the deposited films from bath D under N 2 atmosphere at 200, 300, and 400 o C for 1 hour. The heating rate is 10 o C /min, while the films were cooling to room temperature in the oven. The microstructures of the films before and after annealing were studied using The Grazing Incidence X-ray Diffractions, which were performed using the Philips X'Pert system. The X-ray source is a long-fine-focus, ceramic X-ray tube with Cu anode. All the XRD scans were performed at room temperature. 177 6.3 Results and Discussion 6.3.1 Electrochemical Deposition of Amorphous Fe-B Alloy The Fe-B thin films were deposited from the aqueous solution containing of FeSO 4 ?7H 2 O, KBH 4 , NaOH, and KNaC 4 H 4 O 6 ?4H 2 O. The reaction mechanisms for Fe-B alloy deposition have been discussed in several literatures [19-21]. The deposition process is the combination of the electrochemical process, which Fe is deposited on the working electrode surface under the external applied electric field, and the catalytic chemical process, which BH 4 - anions reduce Fe 2+ to Fe with subsequent codeposition of Fe and B on the working electrode surface. The process involves the dehydrogenation and oxidation of the BH 4 - . BH 4 - + 4OH - ? B(OH) 4 - + 2H 2 + 4e BH 4 - + 8OH - ? B(OH) 4 - + 4H 2 O+ 8e The above reactions provide electrons for the reduction of iron and boron: Fe 2 + + 2e ? Fe B(OH) 4 - + 3e ? B + 4OH - Fig. 6-3 shows the cyclic voltammogram (CV) of the deposition of Fe-B alloys on the working electrode at room temperature. The black line is the CV curve of the deposition from solution D, while the red line is the CV curve of the deposition from solution A. There is no KBH 4 in solution A. In this case, only Fe was deposited. The deposition conditions for these two solutions were same. For plating bath D, at around -1200 mV, the deposition rate increased rapidly and a peak exhibits at about -1250 mV, indicating onset of a bulk deposition. For plating bath A, the deposition rate starts to increase rapidly at about -1300mV and a peak 178 exhibits at about -1350 mV. The results indicate the chemical reduction of Fe 2+ by KBH 4 on the working electrode surface increases the deposition rate and pushes the start of the bulk deposition to a lower voltage. KBH 4 is the source of the Boron, as well as the reduce agent for the deposition process. The reaction mechanisms of Fe-B alloy deposition indicate H 2 evolution. H 2 bubbles can be seen forming and streaming off of the working electrode surface during depositions. For lower current densities, fewer bubbles evolve from the film surface; while for higher current densities, the rate of H 2 evolution is larger, thus more and larger bubbles will form. The strong H 2 evolution causes the film surface is rough and less compact. Therefore, to obtain Fe-B films with better quality, lower current densities (< 10 mA/cm 2 ) were be used in this research. Fig. 6-3. Cyclic voltammogram of Fe-B deposition on Cu/Cr/glass working electrode. 0 -200 -400 -600 -800 -1000 -1200 -1400 -1600 -10 0 10 20 30 40 Curre nt Dens it y (m A/c m 2 ) Potential (E) vs. Ag/AgCl (mV) Bath A Bath D 179 6.3.2 Microstructure of the Fe-B Thin Films The microstructure of the Fe-B film is critical for its properties. As discussed in the ?introduction? part, Fe-B alloy with amorphous structure is highly desired for fabrication of magnetostrictive AW devices. Based on literatures, the Fe-B alloys obtain amorphous microstructure when the B concentration is higher than 16 atm% [20, 21]. As the resource of Boron, the concentration KBH 4 in the deposition bath would affect the Boron composition in the Fe-B films, thus the microstructure of the films. The microstructures of the Fe-B films deposited from baths with different KBH 4 concentrations were studied by X-ray diffraction. Fig. 6-4 shows the XRD patterns of Fe-B films on Cu/Cr/glass substrate deposited from the solutions listed in Table 6-2. The films are in the thickness of 1 ~ 2 ?m, and their microstructures were studied by the Rigaku X-ray system. Since the thickness of the films is small, the diffraction peaks of Cu substrate can be observed in Fig. 6-4. The results in Fig. 6-4 show that, for the film deposited by the baths with lower concentration of KBH 4 (bath A, B), the diffraction peaks corresponding ?-Fe reflections <110> and <211> can be clearly observed, which indicates the formation of crystallized Fe structure. For the films deposited by the baths with KBH 4 ? 16g/L (bath D, E and F), only the diffraction peaks from Cu substrate can be observed, which indicates the amorphous structure of the deposited Fe-B films. The result clearly shows that the microstructure of the films depends on the concentration of potassium borohydride in bath. This is consistent with other studies of electrochemically deposited Fe-B alloy [20, 21]. With increasing KBH 4 concentration in the plating bath, the concentration of Boron in the deposited Fe-B films increases, thus the microstructure of the film tends to become amorphous. 180 Fig. 6-4. XRD patterns of Fe-B thin films on Cu/Cr/glass substrate deposited by plating baths listed in Table 6-2. To eliminate the effect of the Cu substrate on the X-ray diffraction patterns and confirm the amorphous microstructure, Fe-B films in the thickness of 12 ?m were deposited on the Cu/Cr/glass substrate from plating baths C, D, E and F, and then the microstructures were studied by the Rigaku X-ray system. With this thickness, the X-ray beam can not penetrate the whole Fe-B film, and the diffraction patterns only come from the Fe-B film. The XRD patterns of these Fe-B films were shown in Fig. 6-5 as black curves. For the film deposited from bath D, E or F, there is only a broad halo exhibits on 40 60 80 Intensity (a.u.) Bath C 2? (degree) Bath F Bath E Cu Bath D Bath B Fe (211) Bath A Cu Cu Fe (110) 181 the XRD pattern in the range of angle 2? = 40 o ~ 50 o , which is the characteristic of amorphous Fe-B. Based on Equation (6-1), the sizes of the ordering range are 11.7, 10.5 and 11 ? for the films deposited from bath D, E and F, respectively. The wavelength of 1.54 ? for Cu K? as X-ray source was used in the calculation. The results confirmed the amorphous microstructure in these films. However, for the film deposited from solution C, the diffraction peaks corresponding ?-Fe reflections <110> and <211> appears in the XRD pattern, which indicates the formation of ?-Fe crystallite because of the low B concentration. Fig. 6-5. XRD patterns of Fe-B films deposited on the Cu/Cr/glass substrate from plating baths C, D, E and F. 30 40 50 60 70 80 90 30 40 50 60 70 80 90 30 40 50 60 70 80 90 30 40 50 60 70 80 90 2 theta (degree) Bath C Intensity (a.u.) Bath D Bath E Bath F 182 Fig. 6-6. XRD pattern (40 o ~ 50 o ) of the film deposited from bath C and the fitting results based on Lorentz function. The black solid curve is the experimental data, while the red dotted curve and green dashed curve are the fitting results. To further analyze the data, the XRD pattern (40 o ~ 50 o ) of the film deposited from bath C was fitted using the Lorentz function. The fitted curves were shown in Fig. 6-6 as the red dotted curve and green dashed curve. The fitting results reveal a sharp diffraction peak superimposing on a broad halo, which indicates that both crystalline structure and amorphous microstructure forming in the film. The broad red dotted curve is correspondent to the amorphous structure, while the green dashed peak is the diffraction peak corresponding ?-Fe reflection <110>. Based on Equation (6-1), the crystal size of ?- Fe formed in the film is about 12 nm. Therefore, to obtain an amorphous microstructure, plating baths with KBH 4 concentration higher than 16 g/L are needed. 40 50 0 50 100 150 200 250 300 350 Intensi ty (a.u.) 2 theta experimental data 183 6.3.3 Composition of the Fe-B Film The surface composition and the composition distribution along the depth of amorphous Fe-B films deposited from plating bath D were studied. The compositions near the Fe-B film surface were analyzed by XPS in layers. Fig. 6- 7 shows the XPS spectrum of the Fe-B film surface as deposited. The Fe-B film was deposited from plating bath D under a current density of 5 mA/cm 2 , and in the thickness about 4 ?m. From Fig. 6-7, besides Fe and B peaks, huge O peaks and small peaks of C, Ca and Na can be observed. The composition analysis based on data in Fig. 6-7 was summarized in Table 6-3. The results indicated that the surface of the ?as formed? film is oxidized due to the contact with air. Also, absorbed carbon and contamination from the plating bath can be observed on the surface. Fig. 6-7 The XPS spectrum of the Fe-B film surface as deposited. The Fe-B film was deposited from plating bath D under a current density of 5 mA/cm 2 . 0 200 400 600 800 1000 1200 1400 0 20000 40000 60000 80000 100000 120000 140000 160000 In te n s i t y (a .u .) Binding Energy (eV) F e3p Fe3s B1 s F e3p C1s Ca2 p 3 NaK L L O1 s Fe2p3 Fe2p1 Fe LMM Fe LMM1 Fe LMM2 O K L L O K L L C K L L 184 TABLE 6-3 Compositions of Fe-B Film Surface, and About 5 nm, 15 nm and 35 nm from the Surface. The Film Was Deposited from Plating Bath D under a Current Density of 5 mA/cm 2 B atm % Fe atm % O atm % C atm % Na atm % Ca atm % N atm % Ar atm % As formed 5.64 30.26 50.88 12.05 0.45 0.26 0.46 / Etched off 5 nm 9.63 65.85 23.32 0.17 / / / 1.03 Etched off 15 nm 13.25 76.73 8.22 0.67 / / / 1.13 Etched off 35 nm 15.04 78.26 4.86 0.65 / / / 1.19 By etching the film using 2 kV Ar + ion-beam, the compositions of the Fe-B film at depths of approximately 5 nm, 15 nm, 35 nm were also analyzed using XPS, and the results were summarized in Table 6-3. The results show that the oxidation and contamination the surface can be removed by the Ar + ion-beam etching. The concentrations of O and C decrease dramatically along the depth, while the concentrations of B and Fe increase. After Ar ion-beam etching, some Ar atoms stick on the film and can be observed. XPS is able to obtain information on chemical states from the variations in element binding energies due to differences in the chemical potential and polarizability of compounds. An obvious shift in B binding energy was observed on the XPS spectra of the film for different depths. Fig. 6-8 shows the B spectra for the film surface and about 5 nm, 15 nm, and 35 nm from the surface measured by XPS. On the film surface as deposited, the detected binding energy of B1s is 192.8 eV, which corresponds to the binding energy of B in the compound B 2 O 3 [22]. After the surface was etched off 5 nm, 185 photoelectron peak for B1s split into two peaks: one corresponds to the binding energy of compound B 2 O 3 , while another one, with the binding energy of 188.1 eV, corresponds to the binding energy of pure Boron [22]. For the spectra at depths of 15 nm and 35 nm, the peak corresponding to compound B 2 O 3 becomes very weak, while only a strong peak of B1s for pure Boron can be observed. The results clearly shows that the formation of oxidation on the film surface. After the oxidation is removed, the components exist in the chemical status as pure elements. Fig. 6-8. XPS spectra of B1s for deposited Fe-B film surface and about 5 nm, 15 nm, and 35 nm from the surface. The Fe-B film was deposited from plating bath D under a current density of 5 mA/cm 2 . 180 190 200 Binding Energy (ev) as formed etched off 5 nm B1 s etched off 15 nm In te n s i t y (a .u . ) B1s etched off 35 nm 186 The composition of the deposited Fe-B film and the uniformity along its growth direction (thickness) was studied by AES. To perform the analysis, the Fe-B film deposited from plating bath D under current density of 5 mA/cm 2 was broken, and the cross section of the film was analyzed. Fig. 6-9 is the SEM image of the cross section of the Fe-B film. As shown on the image, 10 points were selected long the film thickness direction. The compositions of the 10 points were analyzed using AES, with the e-beam focused to a diameter of ~20 nm. Before the analysis, cross-section surface was sputtered off 50 nm by Ar ion beam to eliminate the contamination and the oxidation. The compositions of the 10 points were listed in Table 6-4, while the contribution of the composition along the length direction was shown in Fig 6-10. Fig. 6-9. SEM image of the cross-section of the Fe-B film deposited from bath D under current density of 5 mA/cm 2 . 1 ?m 187 TABLE 6-4 Compositions of Fe-B film Deposited from Bath D under a Current Density of 5 mA/cm 2 along Its Thickness Direction Atomic Concentration Table Fe:B ratio Points B (atm%) O (atm%) Fe (atm%) Fe (atm%) B (atm %) 1 13.74 42.94 43.32 76 24 2 20.84 13.9 65.26 76 24 3 18.75 12.1 69.14 79 21 4 18.51 12.94 68.56 79 21 5 19.55 11.75 68.69 78 22 6 20.75 12.74 66.51 76 24 7 19.93 13.12 66.95 77 23 8 20.64 13.96 65.4 76 24 9 20.17 15.96 63.87 76 24 10 6.48 40.07 53.45 89 11 average 17.936 18.948 63.115 78 22 Fig. 6-10 Compositions of Fe-B film along its thickness direction. The film was deposited from bath D under a current density of 5 mA/cm 2 . 0246810 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Com pos it ion (atm %) points B O Fe 188 The results show that Fe-B alloy, with designed composition (about 20 atm% B and 80 atm% Fe), was deposited from plating bath D. The composition distribution is uniform along the thickness direction except point 1 and point 10. From Fig. 6-9, point 1 is close to the film surface; the variation of the composition is due to the oxidization and contamination. Point 10 is very close to the Cu substrate. The oxygen should come from the oxidization of Cu electrode. Even though the composition is uniform, oxygen is present in the whole film. Since the plating solutions are alkalic and hydroxide ions participate in the electro-reduction process, the oxygen may be attributed to the hydroxide ions adhering to the film and the oxidation during the deposition. This behavior is also observed in the electrodeposition of Ni alloy [23]. Also, due to the internal stress, tiny cracks were observed in the deposited films, the oxygen might come from the oxidation occurring at the cracks? surface. 6.3.4 Magnetic Properties of the Fe-B films The magnetization hysteresis loops of Fe-B films deposited from plating baths D, E and F are shown in Fig. 6-11. All the films were deposited under a current density of 5 mA/cm 2 . The specimens used for the measurement were the Fe-B films, in the thickness of 12 ?m, with the Cu/Cr (total 2.5 ?m) substrate. Fig. 6-11 shows that all of these amorphous Fe-B alloys exhibit a soft magnetic property, as demonstrated by the slim loops. Based on the data in Fig. 6-11, the magnetic properties of these films were determined and summarized in Table 6-5. As shown in Table 6-5, the coercive field (H c ) of these films is about 10 Oe; while the squareness, which is defined as the ratio of the remanent (Mr) to saturation (Ms) magnetization, is around 0.1 ~ 0.2. The coercive field obtained here is much higher than that of bulk Fe-B alloy. The reason is that the 189 measured sample is not pure Fe-B film, but the Fe-B film with Cu/Cr substrate. Due to the substrate, strong stress/strain exists in the Fe-B films as well as in the interface between the film and the substrate, which would affect the observed coercive field of the whole sample. Also, since the thickness of the films was thin and the magnetization was measured parallel to the film surface, the coercive field of the film would be higher than that of bulk Fe-B alloy due to geometry anisotropy. Still the deposited Fe-B films exhibit relative soft magnetic properties and slim hysteresis loops, which is favorable to obtain high magneto-mechanical coefficient and small heat loss. These properties make the amorphous Fe-B a good candidate as the active material for magnetostrictive acoustic devices. Fig. 6-11. Hystersis loops of Fe-B films deposited from bath (a) D and (b) E under a current density of 5 mA/cm 2 . -1000 -500 0 500 1000 -1.0 -0.5 0.0 0.5 1.0 N o rma l i z ed Magn et iz at i o n H (Oe) Fe-B film deposited from bath D (a) -1000 -500 0 500 1000 -1.0 -0.5 0.0 0.5 1.0 No rmalized Ma gnetization H (Oe) Fe-B film deposited from bath E (b) 190 Fig. 6-11. Hystersis loops of Fe-B film deposited from (c) bath F under a current density of 5 mA/cm 2 . TABLE 6-5 Coercivity and Squareness of Amorphous Fe-B Films Film C Film D Film E Average H c (Oe) 9.37 10.81 9.72 9.97 ? 0.75 Squareness (Mr/Ms) 0.15 0.13 0.21 0.16 ? 0.04 The magnetization hysteresis loops of Fe-B films deposited from solution D with different thicknesses were shown in Fig. 6-12, while their coercive fields (H c ) and squarenesses were listed in Table 6-6. The thickness of the Fe-B film is 100 nm, 3 ?m and 12 ?m, respectively, while the thickness of the Cu/Cr substrate is about 2.5 ?m. The results show that the coercive field and the squareness of the Fe-B film increases with the decrease of the film thickness. The results indicate the strain/stress in the deposited film hardens the magnetic properties of the film. For electrochemical deposited film, due to -1000 -500 0 500 1000 -1.0 -0.5 0.0 0.5 1.0 N o rm alie d Ma gn et i z a t io n H (Oe) Fe-B film deposited from bath F (c) 191 the mismatch between the film and the substrate lattice, strong strain/stress exists in the film, which would increase the coercive field and the squareness. The smaller the thickness of the film, the stronger the strain/stress. Therefore, the magnetic properties of the film become harder as the thickness decreases. The results shown in Table 6-6 also imply the effect of the surface of the Fe-B film and the interface between the Fe-B film and the Cu/Cr substrate on the magnetic properties. Since great amount oxygen exists in the surface, the surface and interface play an important role to block the movement of the domain wall and the switch of the magnetic domain. As the thickness of the film decreases, the volume percentage of the surface and interface in the film increases, thus the effect of them becomes more and more significant. This causes the film, in the thickness of 100 nm, to exhibit a much harder magnetic property than that of the film in the thickness of 12 ?m. Fig. 6-12. Hystersis loops of Fe-B films deposited from bath D with different thicknesses. -600 -400 -200 0 200 400 600 -1.0 -0.5 0.0 0.5 1.0 Normal ized Ma gn eti z a t io n Applied Field H (Oe) Fe-B films deposited from bath D 3um 12um 100nm 192 TABLE 6-6 Coercivity and Squareness of Fe-B films Deposited from Bath D with Different Thicknesses Thickness 100 nm 3 ?m 12 ?m H c (Oe) 15.97 15.75 9.37 Squareness (Mr/Ms) 0.63 0.19 0.15 6.3.5 Resonance Behavior of deposited Fe-B Alloy The resonance behaviors of plated Fe-B amorphous alloys were studied by analyzing the resonance spectra of strip-like magnetostrictive resonators, which consist of two layers: 12 ?m Fe-B film and 2.5 ?m Cu/Cr substrate. The resonance frequencies and Q values of these resonators were determined setup B described in Chapter 3. The experimentally obtained first harmonic resonance frequencies versus reciprocal of the length for these resonators was plotted in Fig. 6-13, where the resonance frequencies of the resonators made from Fe-B film deposited from plating baths D, E and F were represented by black squares, green stars and red triangles, respectively. The results showed that, for all these three kinds of resonators, the resonance frequency linearly increases with the reciprocal of the strip length, which is as expected based on Equation (5-10). Based on material properties shown in Table 6-1 and Equation (5-10), the theoretical resonance frequencies of Fe-B resonators with different sizes were calculated, and shown in Fig. 6-13 as the blue line. The theoretical values exhibit higher resonance frequencies than experimental results. The reason is that, the resonance frequencies calculated from Equation (5-10) are for the naked strips made from 193 Fe-B alloy. However, the resonators used in the experiment consist of two layers: 12 ?m Fe-B film and 2.5 ?m Cu/Cr substrate. The Cu/Cr substrate is equivalent to the mass load for the Fe-B resonators, which will cause the resonance frequencies shift to lower values. Equation (5-17) indicates that the smaller the size of the resonator, the bigger the shift in the resonance frequency caused by the Cu/Cr substrate. Therefore, the difference between the theoretical value and the experiment results increases with the decreasing resonator size, which is consistent with the results show in Fig. 6-13. The velocity of AW propagating in the material can be calculated based on Equation (6-2): L v f 2 0 = (6-2) where f 0 is the fundamental resonance frequency of the free standing strip resonator; L is the length of the resonator; and v is the AW velocity in the material. Therefore, the slope of the resonance frequencies versus the reciprocal of lengths for the strip resonators represents the AW velocity in the material. Based on the results shown in Fig. 6-13, the experimental AW velocity in the films deposited from baths D, E and F are slightly different, which are 3736 ? 208 m/s, 3386 ? 139 m/s and 3124 ? 106 m/s, respectively. Considering the effect of Cu/Cr substrate as mass load, the AW velocities in these films would be higher than the experimental results. The films deposited from bath D exhibits higher AW velocity than the films deposited from baths E and F, thus is expected to have a higher Q value. Therefore, in this research, for fabricating micro/nano MSP, the Fe-B alloy was deposited from bath D. 194 Fig. 6-13. Resonance frequency versus reciprocal of the length for the resonators made of Fe-B film on Cu/Cr substrate. The films were deposited from baths D, E, and F under a current density of 5 mA/cm 2 . The resonance behavior of the resonators was characterized from a kilohertz range up to a megahertz range. The Q values of the resonators were calculated based on the method discussed in Chapter 3. For all three kinds of resonators, a Q value over 200 can be observed even though the resonators were operated at the several megahertz range. The highest Q value was observed is 311. It should be noticed that, 1) the Q value of the resonators was affected by the shape of the resonator; 2) the resonators were made by Fe- 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0.0 5.0x10 5 1.0x10 6 1.5x10 6 2.0x10 6 2.5x10 6 3.0x10 6 3.5x10 6 4.0x10 6 resonators made from film deposited from bath F 1/L (1/mm) theoretical values resonators made from film deposited from bath E resonators made from film deposited from bath D R e son a n c e F r e que ncy (Hz) 195 B alloy and Cu/Cr substrate. The resonators used in this study were cut by hand, by which the shapes were not accurately controlled. The imperfect shape will reduce the Q value. On the other hand, the Cu/Cr bilayer works as the mass load, which would greatly affect the Q value of the resonators. To simulate the effect of the substrate layer on the Q value of the resonators, commercial Metglas TM 2826 MB ribbon (30 ?m in thickness) was employed to make the magnetostrictive strip resonators with different sizes. The resonance behaviors of these strips were characterized using the setup B described in Chapter 3. Meanwhile, a Cu(5 ?m)/Cr(100 nm) bilayer was deposited on Metglas TM 2826 MB ribbon (30 ?m in thickness) to form an equivalent mass load. The 2826 MB/Cu/Cr film was also cut into strips with different sizes to make the resonators, whose resonance behaviors were measured using same setup under same conditions as that for measuring resonators made of just Metglas TM 2826 MB ribbon. The Q values of the resonators with and without Cu/Cr layer were compared, and the results were shown in Fig. 6-14. The solid squares are the Q values obtained for resonators made by Metglas TM ribbon, where the average Q is almost 2000; while the solid triangles are the Q values obtained for resonators made by the 2826MB/Cu/Cr, where the average is about 600. It clearly indicates that the substrate layer greatly reduces the Q value of the resonators. Without the substrate, a Q value about 1000 would be expected for the amorphous Fe-B alloy itself. This indicates that the amorphous Fe-B alloy is a good candidate for fabricating high performance micro/nano magnetostrictive AW sensor platforms. 196 Fig. 6-14. Q values for the strips made of 2826 MB ribbon and 2826MB ribbon with Cu/Cr bilayer. 6.3.6 Thermal Stability of the Amorphous Fe-B Films To study the thermal stability of the amorphous Fe-B films, Fe-B films deposited from bath D were annealed at 200, 300, 400, and 500 o C for 1 hour under N 2 atmosphere. The films were deposited under a current density of 5 mA/cm 2 and in a thickness of about 4 ?m. The Grazing Incidence XRD patterns of the Fe-B films before/after annealing were shown in Fig. 6-15. Based on the XRD patterns and Equation (6-1), it is found that, after annealing at 200 and 300 o C, the sizes of the particles with ordering structure in the films are 12 and 11 ?, respectively, which reveals that the film remained an amorphous microstructure. But after annealing at 400 o C, the film shows diffraction peaks 2000 3000 4000 123456789 0 500 1000 Q va lue average: 1987 length of the strips (mm) average:579 197 corresponding ?-Fe reflections <110> and <211>. But there is still a broad halo exhibited over the range of angle 2? = 40 o ~ 50 o , which implies only part of the material is crystallized. Fitting the XRD pattern (40 o ~ 50 o ) of the film annealed at 400 o C indicates that ?-Fe crystal, in the size of 20 nm, formed in the amorphous film. After annealing at 500 o C, the XRD pattern exhibits strong diffraction peaks corresponding ?-Fe reflections, which indicates the crystallization of the ?-Fe. The Fe-B films exhibit good thermal stability up to 300 o C. Fig. 6-15. XRD patterns of Fe-B films deposited from bath D after annealing at different temperatures. The thermal stability of the amorphous Fe-B alloy also was studied by analyzing the resonance spectra of the resonators made from film deposited from bath D under a 40 60 80 100 2 theta (degree) as deposited Intensity (a.u.) 200 o C 300 o C 400 o C 500 o C Fe (110) Fe (211) Fe (200) 198 current density of 5 mA/cm 2 . The Fe-B film (12 ?m in thickness) on Cu/Cr substrate (2.5 ?m in thickness) was cut into strips with different sizes to form resonators. The resonance behaviors of the resonators were characterized using setup B. Then the resonators were heated to 250 o C in N 2 atmosphere and cooled to room temperature. The heating rate is 10 o C/min, while the resonators were air-cooled in N 2 atmosphere. After the heat treatment, the resonance spectra of the resonators were measured again using the same setup with the same parameters as that were used before the heat treatment. The resonance frequencies and Q values for the resonators before/after heat treatment were determined. The results are shown in Table 6-7. After the treatment, the resonance frequency increased. Commercial 2826MB material was employed to evaluate the effect of the Cu/Cr substrate on the resonance behavior of the Fe-B film before/after the heat treatment. Two kinds of resonators were made: one is cut from commercial Metglas TM 2826 MB ribbon (30 ?m in thickness), another was cut from Metglas TM 2826 MB ribbon with a Cu(5 ?m)/Cr(100 nm) bilayer deposited on its surface. The same heat treatment and analysis as that done to the Fe-B/Cu/Cr resonators were performed to the resonators made of 2826 MB and 2826 MB/Cu/Cr. For both of the two kinds of resonators, the resonance frequencies increased about 1% after the heat treatment. It indicated that the increase of the resonance frequency of the resonators was not due to the heat treatment of the Cu/Cr substrate. It most likely was caused by the relaxation of internal stress of the deposited Fe-B alloy. 199 TABLE 6-7 Resonance Frequency and Q value for the Resonators Made from Fe-B Alloy Deposited from Bath D before and after Heat Treatment Resonance frequency (Hz) Q resonator Length (mm) Before annealing After annealing Before annealing After annealing 1 2.1 915590 967740 110 187 2 3.1 570580 588520 34 86 3 2 924040 951910 67 44 4 5.6 304180 341810 180 87 5 4.1 419240 466180 81 89 6.3.7 Feasibility Study of Micro-Magnetostrictive Particle (MSP) Fabrication Based on Fe-B Alloy In order to achieve extremely high sensitivity, miniaturize the size of the MSP sensor, and precisely control the desired dimensions, the feasibility of combining microelectronic fabrication methods and electrochemical deposition of Fe-B films to fabricate micro-MSP was explored. The purpose of the process design and preliminary work in this section is to provide a useful baseline for future work on micro-MSP fabrication. Two fabrication processes were explored. One is a ?top down? process, which is to deposit the Fe-B film first, and then the MSPs are obtained by etching the film after covering the film with a patterned photoresist layer. Another process is a ?bottom up? process, which is to pattern a photoresist layer with the desired opening on a conductive substrate first, and then electrochemically deposit Fe-B alloy into the openings to obtain 200 micro-MSPs. 6.3.7.1 ?Top Down? Fabrication Process The current flow chart for this process was established, and shown as Fig. 6-16. Fig. 6-16. Flow chart of ?top down? microelectronics fabrication of MSP. Pattern photoresist substrate electrode Deposit electrode for electrochemical deposition Electrochemically deposit Fe-B film Subtractive etching Magnetostrictive Fe-B film substrate electrode photoresistphotoresist Magnetostrictive Fe-B film substrate electrode photoresistphotoresist substrate Electrode MSPMSP substrate Electrode Remove photoresist 201 The optimized process and parameters are described as follows. A thin Au layer (500 nm) was sputtered on the substrate (glass was used in this research) as the electrode for electrochemical deposition. Amorphous Fe-B film was then electrochemically deposited on the electrode/substrate from bath D. Next, a layer of AZ5214-E positive photoresist in a thickness of 1 ?m was coated on the deposited film surface by spin coating the photoresist for 30 seconds at 3000 rpm. After spin coating, the sample was immediately placed on a hot plate for soft bake at 105 o C for 1 minute. Then the photoresist was exposed through the designed mask with ultraviolet light for 4 seconds. The photoresist was developed for 1 minute using the solution made of 1 part (volume) AZ400K TM and 2 part (volume) DI water, and then the sample was rinsed with DI water and dried with compressed sir. After the development, the pattern with MSP area coved with photoresist was formed on the Fe-B film surface, and the film was ready for etching. The etchant was made of distilled water and nitric acid (10 parts water (volume): 1 part acid (volume)). During the subtractive etching, the etchant was stirred by a magnetic stirrer. The patterned Fe-B film was immersed into the etchant for 2~3 seconds, and then immediately rinsed with DI water and dried with nitrogen gas. Finally, the film was rinsed with acetone to remove the remaining photoresist and then dried with nitrogen gas. The fabricated MSPs were observed by an optical microscopy. Images for fabricated MSPs are shown in Fig. 6-17, where the designed size for MSPs shown in Fig. 6-17 (a) is 300 ?m x 60 ?m x 1 ?m, while the designed size for MSPs shown in (b) is 110 ?m x 20 ?m x 1 ?m. Due to the isotropic etching rate for amorphous materials, the size of fabricated MSP is slight smaller than the design size. Also, for thick films, slope can be observed on the MSP edge, as shown in Fig. 6-18. 202 Fig. 6-17. Images for fabricated MSPs using ?top down? method. The designed size for MSPs shown in (a) is 300 ?m x 60 ?m x 1 ?m, while the designed size for MSPs shown in (b) is 110 ?m x 20 ?m x 1 ?m. 50 ?m (a) 50 ?m (b) 203 Fig. 6-18. Slope can be observed on the MSP edge for fabricated MSPs. For the ?top down? process, besides microfabrication parameters, the quality of the plated Fe-B film greatly affects the quality of the fabricated MSPs. Small cracks caused by internal stress and big impurities due to a non-fresh plating bath will cause defects in fabricated MSPs. Fig. 6-19 (a) shows the defects in MSPs caused by the cracks in the original deposited Fe-B film; Fig. 6-19 (b) shows the defects on the edge of fabricated MSPS due to the impurities in the original film. 10 ?m slope 204 Fig. 6-19. (a) Defects in MSPs caused by the cracks in the original deposited Fe-B film; (b) Defects on the edge of fabricated MSPs due to the impurities in the original film. 10 ?m (a) 10 ?m (b) 205 The MSPs were fabricated on the electrode/substrate. To obtain MSP sensors, a major issue that must be considered is how to release the fabricated MSPs from the substrate. It is not easy to release MSP from metal electrode. In later research, conductive polymers might be chosen to replace the metals to be the electrode. By this way, the MSPs can be released from the substrate by dissolving the conductive polymer after fabrication. 6.3.7.2 ?Bottom Up? Fabrication Process The current flow chart for this process was shown as Fig. 6-20. Fig. 6-20 Flow chart of ?bottom up? microelectronics fabrication of MSP. Pattern photoresist Electrochemically deposit Fe-B alloy into the openings Remove photoresist substrate electrode photoresist substrate electrode photoresist Deposited Fe-B alloy substrate electrode MSP MSP substrate electrode Deposit electrode for electrochemical deposition 206 The optimized process and parameters are described as follows. Since the plating baths for Fe-B alloy is alkalic, negative photoresist NANO TM SU-8 2007 was used in this research. First, a thin Au layer (500 nm) was deposited on the glass substrate as the electrode for electrochemical deposition. And then the SU-8 photoresist was spin coated on the Au surface at 3000 rpm for 30 seconds. The sample was then immediately put on a hot plate for soft bake. The soft baking process is to bake the sample at 65 o C for 30 minutes and then 95 o C for one and half hour. After soft baking, the photoresist was exposed through the designed mask with ultraviolet light for 8 seconds. The sample was then put on the hot plate again for hard baking. The hard baking process is to bake the sample at 65 o C for 1 hour and then 95 o C for two hour. After hard baking, the photoresist was developed by putting into SU-8 developer for 4 minutes. The pattern was then ready for deposition. The Au electrode covered with the pattern was used as the working electrode, and the Fe-B alloy was deposited into the openings using the electrodeposition cell described as previous sections to form the MSPs with desired dimensions. Fig. 6-21 (a) shows the SU-8 pattern on the Au electrode/substrate observed using optical microscopy, while Fig. 6-21 (b) shows the optical image of the MSP deposited in the opening. The MSPs were deposited from bath D under a current density of 5 mA/cm 2 . The main issue for this process also is how to release the MSPs from the substrate. Similar with the ?top down? process, a conductive polymer can be considered to be used as the electrode. Since SU-8 is very difficult to remove after the pattern is formed, a new type of photoresist might be needed to be explored. 207 Fig. 6-21. (a) SU-8 pattern on the Au electrode/substrate; (b) the MSP deposited in the opening of patterned SU-8. The MSPs were deposited from bath D under a current density of 5 mA/cm 2 . 50 ?m Opening (a) 50 ?m (b) 208 6.5 Conclusion The Fe-B thin films were synthesized using electrochemical deposition. The deposited films exhibit amorphous microstructure and uniform designed composition. The resonance behavior of amorphous Fe-B alloy as actuating materials for magnetostrictive sensor platforms were studied. The resonators made by plated amorphous Fe-B alloy exhibited high performance when the resonators were operated in a high frequency range. The plated amorphous Fe-B film also exhibited good thermal stability and soft magnetic properties. The plated amorphous Fe-B alloy is a good candidate for fabricating high performance AW sensor platforms in micro scale and might be suitable for fabricating magnetostrictive sensors in nanoscale. 209 References [1] L. G. Puckett, G. Barrett, D. Kouzoudis, C. Grimes, and L. G. Bachas, "Monitoring blood coagulation with magnetoelastic sensors," Biosensors & Bioelectronics, vol. 18, pp. 675-681, 2003. [2] C. A. Grimes, K. G. Ong, K. Loiselle, P. G. Stoyanov, D. Kouzoudis, Y. Liu, C. Tong, and F. Tefiku, "Magnetoelastic sensors for remote query environmental monitoring," Smart Materials & Structures, vol. 8, pp. 639-646, 1999. [3] K. G. Ong, K. F. Zeng, X. P. Yang, K. Shankar, C. M. Ruan, and C. A. Grimes, "Quantification of multiple bioagents with wireless, remote-query magnetoelastic microsensors," Ieee Sensors Journal, vol. 6, pp. 514-523, 2006. [4] J. Wan, H. Shu, S. Huang, B. Fiebor, I.-H. BChen, V. A. Petrenko, and B. A. Chin, "Phage-Based Magnetoelastic Wireless Biosensors for Detecting Bacillus Anthracis Spores," IEEE Sensors Journal, vol. 7, pp. 470 - 477, 2007. [5] R. Guntupalli, J. Hu, R. S. Lakshmanan, T. S. Huang, J. M. Barbaree, and B. A. Chin, "A magnetoelastic resonance biosensor immobilized with polyclonal antibody for the detection of Salmonella typhimurium," Biosensors & Bioelectronics, vol. 22, pp. 1474-1479, 2007. [6] E. T. Lacheisserie, D. Gignoux, and M. Schlenker, Magnetism - Materials and Applications: First Spring Science + Business Media, Inc, 2005. [7] E. P. Wohlfarth, Ferromagnetic materials - A hand book on the properties of magnetically ordered substances, vol. 1: North-Holland Publishing Company, 1980. [8] E. Lacheisserie, Magnetostriction - Theory and Applications of Magnetoelasticity: CRC Press, Inc., 1993. [9] R. Ray, R. Hasegawa, C. P. Chou, and L. A. Davis, "Iron-Boron Glasses - Density, Mechanical and Thermal-Behavior," Scripta Metallurgica, vol. 11, pp. 973-978, 1977. [10] M. A. Mitchell, A. E. Clark, H. T. Savage, and R. J. Abbundi, "Delta-E Effect and Magnetomechanical Coupling Factor in Fe80b20 and Fe78si10b12 Glassy Ribbons," Ieee Transactions on Magnetics, vol. 14, pp. 1169-1171, 1978. [11] C. P. Chou, L. A. Davis, and M. C. Narasimhan, "Elastic-Constants of Metallic Glasses," Scripta Metallurgica, vol. 11, pp. 417-423, 1977. [12] N. Fujita, M. Inoue, K. Arai, P. B. Lim, and T. Fujii, "Electrochemical deposition of amorphous FeB films with soft magnetic properties," Journal of Applied Physics, vol. 83, pp. 7294-7296, 1998. 210 [13] N. Fujita, A. Tanaka, E. Makino, P. T. Squire, P. B. Lim, M. Inoue, and T. Fujii, "Fabrication of amorphous iron-boron films by electroless plating," Applied Surface Science, vol. 114, pp. 61-65, 1997. [14] I. Carabias, A. Martinez, M. A. Garcia, E. Pina, J. M. Gonzalez, A. Hernando, and P. Crespo, "Magnetostrictive thin films prepared by RF sputtering," Journal of Magnetism and Magnetic Materials, vol. 290, pp. 823-825, 2005. [15] Instruction Manual for BAS Epsilon for Electrochemistry: Version 1.50.69, 2004. [16] B. D. Cullity, Elements of X-ray Diffraction, second ed: Addison-Wesley Publishing Company, Inc., 1978. [17] D. Briggs, Handbook of X-ray and Ultraviolet Photoelectron Spectroscopy Heyden, 1978. [18] C. L. Briant and R. P. Messmer, Auger Electron spectroscopy Academic Press, 1998. [19] J. Vandenmeerakker, "On the Mechanism of Electroless Plating .2. One Mechanism for Different Reductants," Journal of Applied Electrochemistry, vol. 11, pp. 395-400, 1981. [20] W. Y. Hu and B. W. Zhang, "Electroless Deposition of Iron-Boron Alloys," Transactions of the Institute of Metal Finishing, vol. 71, pp. 30-33, 1993. [21] E. V. Rakovich, T. V. Gaevskaya, and V. V. Sviridov, "Composition and properties of electrodeposited iron-boron films," Russian Journal of Applied Chemistry, vol. 73, pp. 1167-1172, 2000. [22] "Handbook of X-ray Photoelectron Spectroscopy," Physical Electronics Company. [23] S. Gadad and T. M. Harris, "Oxygen incorporation during the electrodeposition of Ni, Fe, and Ni-Fe alloys," Journal of the Electrochemical Society, vol. 145, pp. 3699-3703, 1998. 211 CHAPTER 7 SYNTHESIS OF AMORPHOUS FE-B NANOWIRES AS MAGNETOSTRICTIVE BIOSENSOR PLATFORM 7.1 Introduction As discussed in chapter 4 & 5, the sensitivities of the MSPs and MCs increase with the decrease in their dimensions. To achieve a extremely high sensitivity, these sensors in micro/nano-scale are required. One-dimensional nanostructures, such as nanowires and nanobelts, exhibit unique properties and provide new opportunities to develop revolutionary sensors. In recent years, chemical or biological sensors based on nanowires/nanobars have attracted lots of attention due to their great potentials as high performance sensors [1-4]. For example, the use of metal-oxide nanowires as gas sensors has been widely investigated. The principle of conventional metal-oxide gas sensors is that, the adsorption of gas molecules on the sensor surface causes a significant change in its electrical conductance. Since the nanowires possess very large surface-to-volume ratios which can overcome the absorption limitation of conventional solid-state sensors, gas sensors based on nanowires can obtain extremely high sensitivity [1]. Different types of metal-oxide nanowires as chemical sensor platforms have been studied. Gas sensors based on single- crystal/polycrystalline SnO 2 nanowires have been fabricated and well characterized. These sensors exhibit excellent sensitivity toward CO, ethanol, and NO 2 [5-7]. It has 212 been demonstrated that a detection limit of ~3 ppm and response/recovery times of several seconds were achieved using SnO 2 nanowire based sensors for detecting NO 2 [5]. Recently, gas sensors based on CdO, ZnO, and In 2 O 3 nanowires have been fabricated, and exhibited high sensitivity and fast response time toward NO 2 , ethanol, and NH 3 , respectively [8-10]. Gas sensor fabricated based on carbon nanotubes also exhibit excellent performance. They can fingerprint a range of gases by monitoring the electrical breakdown of distinct gases at carbon nanotube tips. Also, they are not affect by the changes in the environment, such as temperature, humidity, and gas flow [11]. Based on the MEMS fabrication and nanowires technology, nanowire field-effect transistors (FETs) have been successfully developed and used as chemical sensors as well as biological sensors [3]. In a bulk FET, a semiconductor is used to connect the metal source and drain electrodes, in which the current between the source and drain is controlled by the voltage applied to the gate electrode on the semiconductor. A nanowire FET has similar configuration as that of bulk FET, but uses a single nanowire as the conducting channel between the source and drain electrode. The binding of molecules on the nanowire surface greatly affect its conductance. This feature makes nanowire FETs ultra-sensitive sensor platforms. The fabrication of Si-nanowire FET sensors, which were used to detect pH value and a variety of biomolecules, have been reported. The detections of single viruses and single-stranded DNA have been demonstrated using the sensors based on Si- nanowire FET [3]. Very recently, a novel biosensing platform using nanowires that consist of submicrometer layers of different metals was built [12]. By laying the different metals in specific orders, the nanowires with different characteristic patterns can be formed. These nanowire patterns can be rapidly identified by the optical reflectance 213 images, while fluorescence images are used to obtain the information of the target biological species binding on the antibodies that are immobilized on the nanowire surfaces. These nanowires form a new barcode system, which have been used to detect a variety of pathogens ranging from anthrax, smallpox to botulinum [12]. Since nanomaterials exhibit unique properties and great potential applications, the fabrication of high quality one-dimensional nanostructures has been intensively studied [13-17]. Many methods for synthesizing nanostructures have been developed, such as chemical vapor deposition [18, 19], wet chemistry [13], E-beam lithographic fabrication [20, 21], laser ablation [22], and template synthesis [17, 23, 24]. Template synthesis is the most widely used approach to grow nanomaterials with desired morphology, which was first used by C.R. Martin and his group [17, 25]. This method uses a template as a model to control the shape and the size of the nanostructures. A variety of templates have been employed, such as organic gelators [26], Proteins [27, 28], DNA [29], or carbon nanotubes [30-32]. Most of the template syntheses use two kinds of porous membranes, track-etching polycarbonate membranes and anodic aluminum oxide membranes [16, 17, 23, 25]. These membranes contain a large number of cylindrical pores with uniform diameters in the nanoscale. They are inexpensive, with high pore densities, while membranes with different pore size can be obtained based on the requirement. The desired materials are grown inside the pores of the templates, and after growth, the template can be removed to obtain individual nanomaterials or a nanomaterial array. Large area ordered one-dimensional nanostructures with high density and high aspect ratio can be fabricated using these porous membranes. Porous track-etching polycarbonate membranes are made by bombarding the 214 polycarbonate film with charged particles in a nuclear reactor. When the particles pass through film, they break the chain of the polymer and leave the damaged tracks in the film. And then the film is put into a strong etchant, by which the damaged tracks are etched away to form the pores. The pore density of the membranes is controlled by the time that the polycarbonate films are exposed to the nuclear particles, while the pore size is precisely controlled by the etching time, etchant concentration and etching temperature [25]. Porous alumina membranes are obtained by electro-oxidation of high purity Al foil in an acidic solution. During the oxidation, a porous oxide layer is grown on the Al foil. Under constant potential, pores with uniform diameter can be formed and self-arrange into a hexagonal pattern. The pores propagate through the oxide layer and are perpendicular to the surface of the Al foil. The pore size and pore density are controlled by the anodizing potential [23, 25]. The synthetic approaches of growing materials into porous membranes include electrochemical deposition [23], chemical deposition [33], chemical polymerization [34], and sol-gel deposition [16, 25]. Electrochemical deposition is the most popular method. Fabrications of different kinds of nanomaterials using template-based electrochemical deposition have been reported. Most of metal nanowires with crystalline structures, such as Au [35], Ag [36, 37], Pt [38], Cu [39, 40], Pb [41], Zn [42] etc., have been fabricated using this method, since they can be easily electrochemically deposited. Many metal oxides, such as Cu 2 O [43, 44], SnO 2 [45], In 2 O 3 [46], and TiO 2 [47], as well as semiconductors, such as Bi 2 Te 3 [48, 49], CdTe [23, 50], CdSe [23, 51], etc., have been successfully deposited by this method. 215 A big part of template based electrochemical synthesis is to fabricate magnetic nanostructures. Nanostructures of many ferromagnetic materials including Fe [52-54], Ni [17, 55], Co [56], and their alloys, such as Co-Fe [56], Fe-Ni [57, 58], Ni-Cu [59], et al., have been synthesized using template-based electrochemical deposition. The characterization of these nanostructures demonstrated that they have unique properties, such as highly anisotropic magnetization, enhanced coercivity, and abnormal temperature dependence of the coercivity, which indicate promising potential applications in high density information storage and high performance sensors [17, 60]. Up to date, most of these studies are devoted to the materials with crystalline structure and relative harder magnetic properties. Just a few investigations in nanostructures of amorphous materials, such as CoFeB and Fe 0.88-x Co x P 0.12 alloy, have been reported [61, 62]. Also, even through magnetostrictive materials have wide applications in industry, there are few studies in one-dimensional nanostructures of magnetostrictive materials. Only recently, McGary et al. reported the fabrication of giant magnetostrictive Galfenol (GaFe) nanowires [63]. The purpose of this study is to fabricate magnetostrictive nanowires using template- based electrochemical deposition for the development of highly sensitive sensor platforms. In this study, Fe-B nanowires were fabricated by electrochemically depositing amorphous Fe-B alloy into ion track-etched polymer membranes. Amorphous Fe-B nanowires and nanowire arrays, in diameter ranging from 50 to 200 nm, have been fabricated. The nanowire arrays and individual nanowires were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM), as well as Auger electron spectroscopy (AES). 216 7.2 Experimental 7.2.1 Fabrication of Amorphous Fe-B Nanowires and Nanowire Arrays The magnetostrictive Fe-B nanowires and nanowire arrays were fabricated using template-based electrochemically deposition method. The amorphous Fe-B alloy was deposited from aqueous basic solution of FeSO 4 ?7H 2 O, KBH 4 , NaOH, and KNaC 4 H 4 O 6 ?4H 2 O. Based on previous optimization, the deposition solution with the composition list in Table 7-1 was used for the fabrication. The solutions were prepared using the same procedure described in chapter 6 for Fe-B thin film deposition. Commercially available polycarbonate membranes were used as the templates. The SPI-Pore? polycarbonate membranes, circular in shape (thickness 6~10 ?m, diameter 2.5 cm), were purchased from Structure Probe, Inc [64]. Templates with pore diameters of 200 nm, 100 nm, 50 nm, and 10 nm were employed for the fabrication. To form the electrical contact for the nanobar deposition, the membrane was first sputter coated with a very thin layer Au (~50 nm) on one side. This thin layer Au can not cover the pores. The membrane was then electroplated with Au using the sputtered Au layer as seed. A three- electrode electrochemical cell was employed for the deposition. That is, the standard Ag/AgCl electrode is used as the reference electrode; the platinum mesh is used as the counter electrode; and the polycarbonate with the Au side facing the counter electrode as the working electrode. The deposition was controlled using an Epsilon TM electrochemistry analysis network from Bioanalytical System, Inc., while commercially available Orotemp 24 solution was used as the deposition solution. The Au was deposited under a current density of 0.2 mA/cm 2 for 20 minutes. After the deposition, the Au layer (~ 500 nm) completely covered the pores to serve as the electrical contact. Epoxy 217 was then applied to the Au side for electrical insulation. To precisely control the deposition process, the three-electrode configuration, again, was employed for Fe-B nanowire fabrication. The polycarbonate membrane prepared as above was used as the working electrode, with the bare side facing the counter electrode during the deposition. The deposition was under constant current density: 5 mA/cm 2 and 10 mA/cm 2 . The current density was evaluated based on the template size, pore size and pore density. The temperature was kept at 22?2 o C during the deposition. After the synthesis of the Fe-B nanobars, the membrane template was dissolved by Cl 2 CH 2 . TABLE 7-1 Composition of Amorphous Fe-B Alloy Deposition Solution KNaC 4 H 4 O 6 NaOH FeSO 4 KBH 4 Composition 90 g/L 16 g/L 20 g/L 16 g/L 218 The procedure of the Fe-B nanowire fabrication is shown in Fig. 7-1. Fig. 7-1. Template synthesis of Fe-B nanowires. Porous membrane Deposit bottom Au- electrode E-C deposit Fe-B alloy Dissolve membrane nanowires Bottom Au electrode Side view I Bottom Au electrode II Bottom Au electrode IV V Top view I 3-D view V Bottom Au electrode III Electrical insulate the Au electric contact by Epoxy 219 7.2.2 Characterization of Fe-B Nanowires and Nanowire Arrays 7.2.2.1 Morphology of Fe-B Nanowires and Nanowire Arrays The morphology of the individual nanowires and nanowire arrays were observed using scanning electron microscopy (LEO 982 SEM). Before the observation, the polycarbonate template was dissolved by Cl 2 CH 2 . To observe the individual nanowires, the Au electrode was first polished off by #2000 polish paper after the fabrication. The template was then dissolved, while the individual nanowires were collected by a magnet. 7.2.2.2 Microstructure of Fe-B Nanowires High-resolution transmission electron microscopy (JEOL 2010 HRTEM) was used to characterize the morphology and microstructure of the Fe-B nanowires. The Fe-B nanowires were placed on the Cu TEM grid coated with lacy carbon for the observation. 7.2.2.3 Composition of Fe-B Nanowires The composition of individual Fe-B nanowires was studied by Auger electron spectroscopy (Physical Electronics Model 680 Nanoprobe). Individual nanowires were released from the substrate. The e-beam with a diameter of ~20 nm was focused on individual nanowires. The compositions of selected spots along the length were analyzed by the AES. The outer surface of the nanowire was removed by Ar ion-beam sputtering and its composition was determined as a function of depth (relative to SiO 2 ). By this way, the compositions of the individual nanowire along the length direction as well as along the diameter direction were studied. 220 7.2.2.4 Thermal Stability of Fe-B Nanowairs Similar as the Fe-B thin film, the thermal stability of the Fe-B nanowires was studied by annealing the nanobars at 200, 300, and 400 o C for 1 hour in a N 2 flow. The morphology and microstructure of the nanowires annealed at different temperatures were characterized by HRTEM and electron diffraction. 7.2.2.5 Magnetic Properties of Fe-B Nanowire Arrays The magnetization (M-H loop) of the Fe-B nanowire arrays were measured by a DSM model Vibrating Sample Magnetometer (VSM). The samples were the nanowire arrays on the Au substrate with the template. The measurement was made with the H field applied parallel to nanowire axis (out of substrate plane) and perpendicular to the nanowire axis (in the substrate plan). The nanowire arrays with different diameters and lengths were measured. To analyze the dipolar reaction between nanobars, two samples with multi-layer nanowires were prepared. Sample A was made of layers of Fe-B nanowires, 200 nm in diameter and of 4.5 ?m in the length, inside the membrane. 5 layers of the membranes with nanowires were put together layer by layer using tape. And then it was cut to a 3 mm x 0.5 mm rectangular shape. The scheme is shown in Fig. 7-2. Sample B was prepared from nanowires in 50 nm diameter with the length of 3 ?m. The membrane was dissolved by Cl 2 CH 2 , and then the nanowires were collected and aligned by a magnet on a small piece of wafer. A piece of tape was cover on the nanowires to fix them on the wafer. The SEM image of this sample is shown in Fig. 7-3. For both sample A and B, the M-H loop was measured along the direction 1, 2 and 3 marked on the Fig. 7-2 and 7-3. 221 Fig. 7-2. Scheme of rectangular sample with several layers of nanowires in the diameter of 200 nm and the length of 4.5 ?m. Fig. 7-3. SEM image for multilayer sample B. Direction 3 Direction 1 Direction 2: normal to the surface wafer nanowires Nanowires in template (5 layers) Side view tape Direction 3 nanowires Top View Direction 1 Direction 2 Membrane Nanowires 3mm 0.5mm Direction 3 222 7.3 Results and Discussions 7.3.1 Morphology of Fe-B Nanowires SEM images of Fe-B nanowires fabricated in templates with pore sizes of 10 nm- 200 nm are shown in Fig. 7-4. The images show Fe-B nanowires (2-5 ?m in length) with different diameters fabricated in high density. Nanowires with uniform length can be fabricated on a substrate (as shown in Fig. 7-4 (a) and (b)) so that a sensor array can be fabricated, or can be released from the substrate to get individual nanowires (as shown in Fig. 7-4 (c) and (d)). Fig. 7-4 (e) and (f) are SEM images of individual nanowires fabricated using templates with pore size of 200 nm and 50 nm, respectively. The images revealed that the nanowires have very flat end. This morphology is expected for amorphous materials due to the isotropic growth in all direction. The flat end is very favorable for obtaining high Q value when the nanowire is operated as a resonator, since uneven ends of the nanowire gives a variation in the wire length direction and causes lower Q value. The SEM images indicated that there was no morphologic difference between the nanowires with different diameter/length and formed under different current densities. Since the polymer template is relatively soft, the pore morphology is not strictly defined. Therefore, it is observed that the diameter of nanowires increases slightly along the pore axis. Thus the final diameters of the nanowires are bigger than the pore sizes of the template. The SEM images of the nanowire arrays show that the overlap between the nanowires can be observed. This is because the nanowires were synthesized in commercial ion track membranes, in which the angle between the pore axis and the surface normal can be as high as 30 o . The condition is worse for membranes with bigger 223 pores, and becomes better for membranes with smaller pore size (bigger pore distance). Fig. 7-4. SEM images for Fe-B nanowires and nanowire arrays fabricated in the template with the pore size of (a) 200 nm, (b) 100 nm, and (c) 50 nm. (a) (b) (c) 224 Fig. 7-4 SEM images for Fe-B nanowires and nanowire arrays fabricated in the template with the pore size of (a) 200 nm, (b) 100 nm, (c) 50 nm, and (d) 10 nm; and focused SEM images of individual nanowire fabricated in the template with the pore size of (e) 200 nm, and (f) 50 nm. (d) (e) (f) 225 7.3.2 Microstructure of Fe-B Nanowires Fig. 7-5 shows the TEM image and the electron diffraction patterns of a Fe-B nanowire fabricated in the template with pore size of 50 nm. To study the details of the microstructure of the nanowire, the individual nanowire was characterized with the electron diffraction focused on the selected areas (SAED) along the nanowire. The electron diffraction patterns revealed that the microstructure of the nanowires is amorphous, and the results confirmed that the Fe-B nanowire has a uniform amorphous microstructure along the length. Fig. 7-6 is the TEM image and electron diffraction pattern of a Fe-B nanowire fabricated in the template with pore size of 10 nm. The diffraction pattern also reveals that the microstructure of the nanowire is amorphous. Nanowires with different lengths and diameters were characterized, and all exhibited a uniform amorphous microstructure. The Fe-B nanowires fabricated by template-based electrochemical synthesis exhibit same amorphous microstructure as that of Fe-B films. 226 Fig. 7-5. (a)TEM image of the Fe-B nanowires fabricated in template with pore diameter of 50 nm; (b) Electron diffraction patterns of this nanowire; (c) Focused TEM image of the Fe-B nanobars fabricated in template with pore diameter of 50 nm. (b) Bottom Middle Top (a) (c) 227 Fig. 7-6. (a)TEM image of a Fe-B nanowire fabricated in template with pore diameter of 10 nm; (b) Selected area electron diffraction patterns of this nanowire; (c) Focused TEM image of Fe-B nanowire fabricated in template with pore diameter of 10 nm. (c) (a) (b) 228 7.3.3 Composition of Fe-B Nanowires The composition of individual Fe-B nanowires was determined by Auger electron spectroscopy. An individual nanowire, as shown in Fig. 7-7 (a), was placed on the Cu TEM grid coated with lacy carbon. This nanowire is deposited using the template with the pore diameter of 50 nm under a current density of 5 mA/cm 2 . The electron beam was focused on selected points along the nanowire, and then combined with the Ar ion-beam sputtering etching, the compositions of nanowire surface, and at depths of approximately 5 nm, 10 nm, 18 nm, 30 nm and 42 nm from the surface, were measured. The composition distribution of this nanowire is shown in Fig. 7-7 (b) and (c). Fig. 7-7 (b) shows the composition contribution of the middle of the nanowire (point 6) along its diameter direction (at depths from the surface). A large concentration of oxygen was observed at the surface of the nanowire, which indicates the existence of an oxide layer on the outermost surface. It is believed that this oxidation layer is related to the passivation of nanowires in air. After removal of ~10 nm of the outermost surface, the concentration of the oxygen dramatically drops and remains constant (as does Fe and B) through the interior of the nanowire. To analyze the composition of the nanowire along its length direction, the compositions of 10 points along its length were characterized, after the nanowire was sputtering etched off 30 nm to eliminate the effect of the oxidation layer. The results were shown in the Fig. 7-7 (c). The analysis results revealed a fairly uniform composition distribution along the nanobar length direction. Except for points 1 and 10, which were located at the end of the nanowire, the variation in the ratio of Fe to B composition along the nanowire length direction was less than 3%. The analysis demonstrated that the whole nanowire exhibits uniform composition distribution except a 229 very thin oxidation layer (less than 10 nm) covering the surface. The composition contribution along the length direction of a nanowire deposited under a current density of 10 mA/cm 2 is shown in Fig. 7-8. This nanowire was also fabricated in the template with pore diameter of 50 nm. The nanowire was placed on a Si wafer. After sputtering off 30 nm, the compositions of the 10 points along the length were analyzed. The results shown in Fig. 7-8 revealed a similar uniform composition: the composition shows very small variation, expect point 1 and 10. These results demonstrate that the Fe-B nanowires fabricated by template-based electrochemical deposition exhibit uniform composition. A small amount of O can be observed through the nanobars. Same as the deposition of Fe-B films, the existence of oxygen may be attributed the hydroxide ions in the alkalic plating solutions. In this study, the atomic concentration was determined using the Auger peak signal of each element. It should be aware that the results may be affected by many factors, such as the primary electron beam energy, the efficiency of the detector, sample surface roughness, and the tilt angle of the sample [67]. Therefore, errors may arise from the system and the measurements. Additionally, in the case of analyzing the sample in nano scale, the signal may not completely come from the sample itself. Even though the electron-beam with 20 nm diameter was focused on the nanowire, about 10 % of the signal can be attributed to the material outside of the selected area, even the substrate. This may further increase the error in the resultant composition. 230 Fig. 7-7. (a) Individual nanowire deposited in the template with pore diameter of 50 nm under a current density of 5 mA/cm 2 ; (b) Composition distribution of point 6 on this nanowire along its diameter direction; (c) Composition distribution of this nanowire along its length direction. 0246810 10 20 30 40 50 60 70 Composit ion (at m %) Points Along Length B O Fe (a) (b) (c) 0 10203040 0 10 20 30 40 50 60 70 80 Composit ion ( a t m %) Distance from the Surface (nm) B O Fe 231 Fig. 7-8. Composition distribution along the length direction of a nanowire deposited in the template with pore diameter of 50 nm under a current density of 10 mA/cm 2 . 7.3.4 Thermal Stability of Fe-B Nanowire The thermal stability of the Fe-B nanowires was studied by annealing the nanowires in the N 2 environment. The TEM images and electron diffraction patterns of the nanowires annealed at 200, 300, and 400 o C for 1 hour are shown in Fig. 7-9, 7-10, and 7- 11, respectively. The results showed that the nanowires maintained an amorphous microstructure after annealing at 200 o C and 300 o C. After annealing at 400 o C, the crystalline structures appear in the nanowires. Comparing the electron diffraction pattern with the standard electron diffraction patterns, it was found that BCC iron and FeO crystal formed after annealing at 400 o C. The Fe-B nanowires exhibit a good thermal stability. 0246810 0 10 20 30 40 50 60 70 80 Concentration of Fe, B and O (atm%) Points Along Length B O Fe 232 Fig. 7-9. The TEM image and diffraction pattern of Fe-B nanowires fabricated in template with pore diameter of 50 nm and annealed at 200 o C for 1 hour. Fig. 7-10. The TEM image and diffraction pattern of Fe-B nanowires fabricated in template with pore diameter of 50 nm and annealed at 300 o C for 1 hour. Diffraction pattern Image Diffraction pattern Image 233 Fig. 7-11. (a) The TEM image and diffraction pattern of the Fe-B nanowire fabricated in template with pore diameter of 50 nm and annealed at 400 o C for 1 hour. (b) The comparison of standard electron diffraction patterns of BCC Fe and FeO with the electron diffraction pattern from (a). (a) Diffraction pattern Image Pattern of BCC Fe Pattern of FeO (b) 234 7.3.5 Magnetic Properties of Fe-B Nanowires The magnetization hysteresis loops for nanowire arrays fabricated in the membranes with pore diameter of 200, 100 and 50 nm are shown in Fig. 7-12. The magnetic field was applied perpendicular to the membrane (out-of-plane), thus parallel to the long axis of nanowires, as well as parallel to the membrane (in-plane), thus perpendicular to the long axis of nanowires. The conditions of the nanowires and their magnetic properties were listed in Table 7-2. With the diameter reduced from 200 nm to 50 nm, the coercivitive field along out-plane increased from 12 Oe to 164 Oe, while the coercivity of in-plane increased from 19 Oe to 335 Oe. For the nanowires in the diameter of 200 nm and the length of 1.8 ?m, the easy axis of magnetization is along the in-plane direction; while for other nanowires, the out-of-plane is the easy axis of magnetization. All the nanowires exhibit very small squareness both in-plane and out-of-plane. Since the nanowires have an amorphous microstructure, the magnetocrystalline anisotropy can be neglected. The coercivity of the nanowire arrays increases with the decrease of the nanobar diameter. The enhancement of the coercivity is due to the tendency of single-domain formation in the nanobars with very small diameter [17, 65]. The overall magnetic behavior of the nanowires is determined by the competition of the shape anisotropy and magnetostatic interaction between the nanowires. It is observed that, for nanowires in the diameter of 200 nm and the length of 1.8 ?m, the easy axis of magnetization is along the in-plane direction. A micromagnetic simulation performed by Ciureanu et al. demonstrated that the saturation field (M S ) of a ferromagnetic array is affected by the interaction field between nanowires along the in- plane direction and out-of-plane direction [61]. When the in-plane interaction field is 235 larger than (2?/3)M S , the interaction field will cause the easy axis of magnetization along the in-plane direction [61]. For the nanowires in diameter of 200 nm, considering the small distance between the nanowires due to the high pore density (3 x 10 8 pores/cm 2 according to the manufacturer) and the overlap between the nanowires, the interaction field in the array is very large. This will cause the in-plane axis to be the easy axis of magnetization. However, for the nanowires in diameter of 200 nm and length of 4.5 ?m, the length/diameter ratio is bigger. This indicates shape anisotropy has more effect, and the easy axis of magnetization is observed along the out-of-plane direction. For the nanowire arrays with smaller diameter, the interaction field is smaller, where the shape anisotropic will dominate the magnetic behavior. The nanowires fabricated in this study exhibited large length/diameter ratio, which is more than 20 for nanowires in diameter of 100 nm and more than 60 for nanowires in diameter of 50 nm. The shape anisotropy tends to align magnetic moments along the wire direction because of the large aspect ratio. As shown in the hysteresis loop, the ?easy axis of magnetization? is along the out-of-plane direction for the nanowires in diameter of 100 nm and 50 nm. However, it should be note that, for all nanowire arrays, the coercivity along the wire direction is smaller than the coercivity perpendicular to the wire direction. This phenomenon was observed in the Co nanowires also deposited in the polycarbonate membranes [17]. The pore density in commercial ion track etched membranes is about 4 ~ 6 x 10 8 pores per centimeter according to the manufacturer, but the pores are not well- parallel. There are strong dipolar interactions between the dense but not well-paralleled nanobars formed in the commercial membranes. Research showed that the interaction between the nanowires would decrease the coercivity [66]. This effect is much stronger in 236 out-plane axis than in in-plane axis. This might strongly decrease the out-plane coercivity to a value that is smaller than in-plane coercivity. The magnetic properties of multi-layer Fe-B nanowires (sample A and sample B) were summarized in Table 7-3. The hysteresis loops of these two samples exhibit similar magnetic behavior as the single layer nanowires. The easy axis of magnetization is still along the wire direction with smaller coercivity. It indicated that dipolar interactions between nanowires have much stronger effect along wire direction than that perpendicular to wire direction. 237 Fig. 7-12. Magnetization hystersis loops of nanowires with diameter of 200 nm (a, b), 100 nm (c, d) and 50 nm (e). -10000 -5000 0 5000 10000 -1.0 -0.5 0.0 0.5 1.0 in plane No rm aliz e d M a gne tiz a tio n diameter: 100nm length: 4.5 um (d) out of plane H (Oe) -10000 -5000 0 5000 10000 -1.0 -0.5 0.0 0.5 1.0 out of plane No rm aliz ed M agnet izat ion H (Oe) in plane diameter: 200nm length: 1.8 um (a) -10000 -5000 0 5000 10000 -1.0 -0.5 0.0 0.5 1.0 out of plane Nor m al i e d Magn et iza t i o n H (Oe) in plane diameter: 200nm length: 4.5 um (b) -10000 -5000 0 5000 10000 -1.0 -0.5 0.0 0.5 1.0 out of plane Norm alized Magne tization in plane H (Oe) diameter: 100nm length: 2.6um (c) -3000 -2000 -1000 0 1000 2000 3000 -1.0 -0.5 0.0 0.5 1.0 in plane No rmalized Mmag netization H (Oe) diameter: 50nm length: 3um (e) out plane 238 TABLE 7-2 Coercivity for Amorphous Fe-B Nanowires 200 nm 100 nm 50 nm Length ( ?m) 1.8 4.5 2.6 4.5 3 In plane 45 19 28 142 335 Hc (Oe) Out of plane 32 12 17 62 164 In plane 0.06 0.02 0.03 0.2 0.3 Squareness (Mr/Ms) Out of plane 0.01 0.005 0.01 0.03 0.2 TABLE 7-3 Coercivity of Multi-layer Fe-B Nanowires along Different Directions 7.4 Conclusion To fabricate an extremely sensitive magnetostrictive sensor platform, Fe-B nanowires were synthesized using a template-based electrochemical deposition method. The composition and microstructure of the Fe-B nanowires are directly related to their performance as a biosensor platform. The Fe-B nanowire arrays and individual nanowires were characterized using scanning electron microscopy (SEM), transmission electron microscopy (TEM), as well as Auger electron spectroscopy (AES). Morphologically, nanowires have a very flat top and a smooth cylindrical surface, which are critical factors Direction 1 Direction 2 Direction 3 Sample A Hc (Oe) 24 19 13 Sample B Hc (Oe) 299 321 214 239 for obtaining high performance as sensor platforms. Structurally, electron diffraction reveals that the Fe-B nanowires are amorphous. AES analysis indicates that Fe-B nanowirers show no significant compositional variation along the length direction. It is found that the nanowires were covered by an oxidation layer of a typical thickness of ~ 10 nm. It is believed that this oxidation layer is related to the passivation of nanobars in air. High temperature annealing and subsequent structural analysis indicate that the Fe-B nanowires possess a good thermal stability. The magnetization loop of the Fe-B nanowire arrays revealed the nanowires exhibit soft magnetic properties. The study demonstrated the fabrication of Fe-B nanowires as ultra-sensitive magnetostrictive sensor platforms. 240 References [1] A. Kolmakov and M. Moskovits, "Chemical sensing and catalysis by one- dimensional metal-oxide nanostructures," Annual Review of Materials Research, vol. 34, pp. 151-180, 2004. [2] D. H. Reich, M. Tanase, A. Hultgren, L. A. Bauer, C. S. Chen, and G. J. Meyer, "Biological applications of multifunctional magnetic nanowires (invited)," Journal of Applied Physics, vol. 93, pp. 7275-7280, 2003. [3] F. Patolsky, G. F. Zheng, and C. M. Lieber, "Nanowire-based biosensors," Analytical Chemistry, vol. 78, pp. 4260-4269, 2006. [4] Z. W. Pan, Z. R. Dai, and Z. L. Wang, "Nanobelts of semiconducting oxides," Science, vol. 291, pp. 1947-1949, 2001. [5] M. Law, H. Kind, B. Messer, F. Kim, and P. D. Yang, "Photochemical sensing of NO2 with SnO2 nanoribbon nanosensors at room temperature," Angewandte Chemie-International Edition, vol. 41, pp. 2405-2408, 2002. [6] E. Comini, G. Faglia, G. Sberveglieri, Z. W. Pan, and Z. L. Wang, "Stable and highly sensitive gas sensors based on semiconducting oxide nanobelts," Applied Physics Letters, vol. 81, pp. 1869-1871, 2002. [7] A. Kolmakov, Y. X. Zhang, G. S. Cheng, and M. Moskovits, "Detection of CO and O-2 using tin oxide nanowire sensors," Advanced Materials, vol. 15, pp. 997- +, 2003. [8] X. Liu, C. Li, S. Han, J. Han, and C. W. Zhou, "Synthesis and electronic transport studies of CdO nanoneedles," Applied Physics Letters, vol. 82, pp. 1950-1952, 2003. [9] Q. Wan, Q. H. Li, Y. J. Chen, T. H. Wang, X. L. He, J. P. Li, and C. L. Lin, "Fabrication and ethanol sensing characteristics of ZnO nanowire gas sensors," Applied Physics Letters, vol. 84, pp. 3654-3656, 2004. [10] D. J. Zhang, C. Li, X. L. Liu, S. Han, T. Tang, and C. W. Zhou, "Doping dependent NH3 sensing of indium oxide nanowires," Applied Physics Letters, vol. 83, pp. 1845-1847, 2003. [11] A. Modi, N. Koratkar, E. Lass, B. Q. Wei, and P. M. Ajayan, "Miniaturized gas ionization sensors using carbon nanotubes," Nature, vol. 424, pp. 171-174, 2003. [12] J. B. H. Tok, F. Y. S. Chuang, M. C. Kao, K. A. Rose, S. S. Pannu, M. Y. Sha, G. Chakarova, S. G. Penn, and G. M. Dougherty, "Metallic striped nanowires as multiplexed immunoassay platforms for pathogen detection," Angewandte Chemie-International Edition, vol. 45, pp. 6900-6904, 2006. 241 [13] X. Wang and Y. D. Li, "Solution-based synthetic strategies for 1-D nanostructures," Inorganic Chemistry, vol. 45, pp. 7522-7534, 2006. [14] A. K. Wanekaya, W. Chen, N. V. Myung, and A. Mulchandani, "Nanowire-based electrochemical biosensors," Electroanalysis, vol. 18, pp. 533-550, 2006. [15] Q. Gu, C. D. Cheng, R. Gonela, S. Suryanarayanan, S. Anabathula, K. Dai, and D. T. Haynie, "DNA nanowire fabrication," Nanotechnology, vol. 17, pp. R14-R25, 2006. [16] P. Aranda and J. M. Garcia, "Porous membranes for the preparation of magnetic nanostructures," Journal of Magnetism and Magnetic Materials, vol. 249, pp. 214-219, 2002. [17] J. S. J. T. M. Whitney, P. C. Searson, C. L. Chien, "Fabrication and Magnetic Properties of Array of Metallic Nanowires," Science, vol. 261, pp. 1316-1319, 1993. [18] L. Zhang, R. Tu, and H. J. Dai, "Parallel core-shell metal-dielectric- semiconductor germanium nanowires for high-current surround-gate field-effect transistors," Nano Letters, vol. 6, pp. 2785-2789, 2006. [19] X. M. Cai, A. B. Djurisic, and M. H. Me, "GaN nanowires: CVD synthesis and properties," Thin Solid Films, vol. 515, pp. 984-989, 2006. [20] W. Fritzsche, K. J. Bohm, E. Unger, and J. M. Kohler, "Metallic nanowires created by biopolymer masking," Applied Physics Letters, vol. 75, pp. 2854-2856, 1999. [21] P. D. Miller, C. P. Liu, W. L. Henstrom, J. M. Gibson, Y. Huang, P. Zhang, T. I. Kamins, D. P. Basile, and R. S. Williams, "Direct measurement of strain in a Ge island on Si(001)," Applied Physics Letters, vol. 75, pp. 46-48, 1999. [22] A. M. Morales and C. M. Lieber, "A laser ablation method for the synthesis of crystalline semiconductor nanowires," Science, vol. 279, pp. 208-211, 1998. [23] Y. Chen, "Template-directed electrosynthesis of nanostructures ", vol. Ph. D.: Auburn University, 2004. [24] K. Jain and S. T. Lakshmikumar, "Porous alumina template based nanodevices," Iete Technical Review, vol. 19, pp. 293-306, 2002. [25] A. Huczko, "Template-based synthesis of nanomaterials," Applied Physics a- Materials Science & Processing, vol. 70, pp. 365-376, 2000. [26] Y. Ono, K. Nakashima, M. Sano, Y. Kanekiyo, K. Inoue, J. Hojo, and S. Shinkai, "Organic gels are useful as a template for the preparation of hollow fiber silica," Chemical Communications, pp. 1477-1478, 1998. 242 [27] R. Bhattacharya, C. R. Patra, S. F. Wang, L. C. Lu, M. J. Yaszemski, D. Mukhopadhyay, and P. Mukherjee, "Assembly of gold nanoparticles in a rod-like fashion using proteins as templates," Advanced Functional Materials, vol. 16, pp. 395-400, 2006. [28] Y. Ono, Y. Kanekiyo, K. Inoue, J. Hojo, M. Nango, and S. Shinkai, "Preparation of novel hollow fiber silica using collagen fibers as a template," Chemistry Letters, pp. 475-476, 1999. [29] W. Shenton, T. Douglas, M. Young, G. Stubbs, and S. Mann, "Inorganic-organic nanotube composites from template mineralization of tobacco mosaic virus," Advanced Materials, vol. 11, pp. 253-+, 1999. [30] Y. G. Wang, Q. H. Li, T. H. Wang, X. W. Lin, V. P. Dravid, and S. X. Zhou, "In situ growth of nanowire on the tip of a carbon nanotube under strong electric field," Applied Physics Letters, vol. 86, 2005. [31] T. Sainsbury and D. Fitzmaurice, "Carbon-nanotube-templated and pseudorotaxane-formation-driven gold nanowire self-assembly," Chemistry of Materials, vol. 16, pp. 2174-2179, 2004. [32] M. Gao, L. Dai, and G. G. Wallace, "Glucose sensors based on glucose-oxidase- containing polypyrrole/aligned carbon nanotube coaxial nanowire electrodes," Synthetic Metals, vol. 137, pp. 1393-1394, 2003. [33] C. A. Huber, T. E. Huber, M. Sadoqi, J. A. Lubin, S. Manalis, and C. B. Prater, "Nanowire Array Composites," Science, vol. 263, pp. 800-802, 1994. [34] H. Masuda and K. Fukuda, "Ordered Metal Nanohole Arrays Made by a 2-Step Replication of Honeycomb Structures of Anodic Alumina," Science, vol. 268, pp. 1466-1468, 1995. [35] W. B. Zhao, J. J. Zhu, and H. Y. Chen, "Photochemical synthesis of Au and Ag nanowires on a porous aluminum oxide template," Journal of Crystal Growth, vol. 258, pp. 176-180, 2003. [36] G. Riveros, S. Green, A. Cortes, H. Gomez, R. E. Marotti, and E. A. Dalchiele, "Silver nanowire arrays electrochemically grown into nanoporous anodic alumina templates," Nanotechnology, vol. 17, pp. 561-570, 2006. [37] X. Y. Sun, F. Q. Xu, Z. M. Li, and W. H. Zhang, "Cyclic voltammetry for the fabrication of high dense silver nanowire arrays with the assistance of AAO template," Materials Chemistry and Physics, vol. 90, pp. 69-72, 2005. [38] C. D. Merritt and B. L. Justus, "Fabrication of microelectrode arrays having high- aspect-ratio microwires," Chemistry of Materials, vol. 15, pp. 2520-2526, 2003. 243 [39] Y. Fukunaka, M. Motoyama, Y. Konishi, and R. Ishii, "Producing shape- controlled metal nanowires and nanotubes by an electrochemical method," Electrochemical and Solid State Letters, vol. 9, pp. C62-C64, 2006. [40] Y. Konishi, M. Motoyama, H. Matsushima, Y. Fukunaka, R. Ishii, and Y. Ito, "Electrodeposition of Cu nanowire arrays with a template," Journal of Electroanalytical Chemistry, vol. 559, pp. 149-153, 2003. [41] Y. T. Pang, G. W. Meng, L. D. Zhang, W. J. Shan, X. Y. Gao, A. W. Zhao, and Y. Q. Mao, "Arrays of ordered Pb nanowires with different diameters in different areas embedded in one piece of anodic alumina membrane," Journal of Physics- Condensed Matter, vol. 14, pp. 11729-11736, 2002. [42] J. G. Wang, M. L. Tian, N. Kumar, and T. E. Mallouk, "Controllable template synthesis of superconducting Zn nanowires with different microstructures by electrochemical deposition," Nano Letters, vol. 5, pp. 1247-1253, 2005. [43] J. Oh, Y. Tak, and Y. Lee, "Electrodeposition of Cu2O nanowires using nanoporous alumina template," Electrochemical and Solid State Letters, vol. 7, pp. C27-C30, 2004. [44] X. M. Liu and Y. C. Zhou, "Electrochemical synthesis and room temperature oxidation behavior of Cu nanowires," Journal of Materials Research, vol. 20, pp. 2371-2378, 2005. [45] M. J. Zheng, G. H. Li, X. Y. Zhang, S. Y. Huang, Y. Lei, and L. D. Zhang, "Fabrication and structural characterization of large-scale uniform SnO2 nanowire array embedded in anodic alumina membrane," Chemistry of Materials, vol. 13, pp. 3859-3861, 2001. [46] M. J. Zheng, L. D. Zhang, G. H. Li, X. Y. Zhang, and X. F. Wang, "Ordered indium-oxide nanowire arrays and their photoluminescence properties," Applied Physics Letters, vol. 79, pp. 839-841, 2001. [47] X. Y. Zhang, L. D. Zhang, W. Chen, G. W. Meng, M. J. Zheng, and L. X. Zhao, "Electrochemical fabrication of highly ordered semiconductor and metallic nanowire arrays," Chemistry of Materials, vol. 13, pp. 2511-2515, 2001. [48] L. Li, Y. W. Yang, X. H. Huang, G. H. Li, and L. D. Zhang, "Pulsed electrodeposition of single-crystalline Bi2Te3 nanowire arrays," Nanotechnology, vol. 17, pp. 1706-1712, 2006. [49] C. G. Jin, X. Q. Xiang, C. Jia, W. F. Liu, W. L. Cai, L. Z. Yao, and X. G. Li, "Electrochemical fabrication of large-area, ordered Bi2Te3 nanowire arrays," Journal of Physical Chemistry B, vol. 108, pp. 1844-1847, 2004. 244 [50] A. W. Zhao, G. W. Meng, L. D. Zhang, T. Gao, S. H. Sun, and Y. T. Pang, "Electrochemical synthesis of ordered CdTe nanowire arrays," Applied Physics a- Materials Science & Processing, vol. 76, pp. 537-539, 2003. [51] D. S. Xu, X. S. Shi, G. L. Guo, L. L. Gui, and Y. Q. Tang, "Electrochemical preparation of CdSe nanowire arrays," Journal of Physical Chemistry B, vol. 104, pp. 5061-5063, 2000. [52] J. B. Wang, X. Z. Zhou, Q. F. Liu, D. S. Xue, F. S. Li, B. Li, H. P. Kunkel, and G. Williams, "Magnetic texture in iron nanowire arrays," Nanotechnology, vol. 15, pp. 485-489, 2004. [53] M. Kroll, W. J. Blau, D. Grandjean, R. E. Benfield, F. Luis, P. M. Paulus, and L. J. de Jongh, "Magnetic properties of ferromagnetic nanowires embedded in nanoporous alumina membranes," Journal of Magnetism and Magnetic Materials, vol. 249, pp. 241-245, 2002. [54] Y. Peng, H. L. Zhang, S. L. Pan, and H. L. Li, "Magnetic properties and magnetization reversal of alpha-Fe nanowires deposited in alumina film," Journal of Applied Physics, vol. 87, pp. 7405-7408, 2000. [55] C. L. Xu, H. Li, G. Y. Zhao, and H. L. Li, "Electrodeposition and magnetic properties of Ni nanowire arrays on anodic aluminum oxide/Ti/Si substrate," Applied Surface Science, vol. 253, pp. 1399-1403, 2006. [56] H. R. Khan and K. Petrikowski, "Synthesis and properties of the arrays of magnetic nanowires of Co and CoFe," Materials Science & Engineering C- Biomimetic and Supramolecular Systems, vol. 19, pp. 345-348, 2002. [57] H. Chiriac, A. E. Moga, M. Urse, and T. A. Ovari, "Preparation and magnetic properties of electrodeposited magnetic nanowires," Sensors and Actuators a- Physical, vol. 106, pp. 348-351, 2003. [58] T. Ohgai, L. Gravier, X. Hoffer, and J. P. Ansermet, "CdTe semiconductor nanowires and NiFe ferro-magnetic metal nanowires electrodeposited into cylindrical nano-pores on the surface of anodized aluminum," Journal of Applied Electrochemistry, vol. 35, pp. 479-485, 2005. [59] A. Encinas, M. Demand, L. Vila, L. Piraux, and I. Huynen, "Tunable remanent state resonance frequency in arrays of magnetic nanowires," Applied Physics Letters, vol. 81, pp. 2032-2034, 2002. [60] P. D. McGary, L. W. Tan, J. Zou, B. J. H. Stadler, P. R. Downey, and A. B. Flatau, "Magnetic nanowires for acoustic sensors (invited)," Journal of Applied Physics, vol. 99, 2006. 245 [61] M. Ciureanu, F. Beron, L. Clime, P. Ciureanu, A. Yelon, T. A. Ovari, R. W. Cochrane, F. Normandin, and T. Veres, "Magnetic properties of electrodeposited CoFeB thin films and nanowire arrays," Electrochimica Acta, vol. 50, pp. 4487- 4497, 2005. [62] D. S. Xue, J. L. Fu, and H. G. Shi, "Preparation and magnetic properties of Fe- 0.88_xCoxP0.12 amorphous nanowire arrays," Journal of Magnetism and Magnetic Materials, vol. 308, pp. 1-4, 2007. [63] P. D. McGary and B. J. H. Stadler, "Electrochemical deposition of Fe1-xGax nanowire arrays," Journal of Applied Physics, vol. 97, 2005. [64] SPI Supplys, "Standard White Polycarbonate "Track Etch" Membrane Filters." [65] A. H. Morrish, The physical principles of magnetism: New York, Wiley 1965. [66] Q. F. Liu, C. X. Gao, J. J. Xiao, and D. S. Xue, "Size effects on magnetic properties in Fe0.68Ni0.32 alloy nanowire arrays," Journal of Magnetism and Magnetic Materials, vol. 260, pp. 151-155, 2003. [67] "Handbook of Auger Electron Spectroscopy", Physical Electronic Company. 246 CHAPTER 8 PERSPECTIVES (1) As discussed previously, to obtain an extremely high sensitivity, MSMCs in a very small size are required. At current stage, the MSMCs are hand-made, which makes the fabrication of micro/nano-MSMC and the control of the MSMC?s quality difficult. At next stage, the micro-fabrication of MSMC should be explored using the established microfabrication process based on amorphous Fe-B alloy. The objectives should be: 1) to fabricate the MSMCs with precisely controlled dimension, morphology, and quality; 2) to develop MSMC arrays; 3) to fabricate and characterize MSMC in micro/nano scale; 4) to perform the practical biological detection using MSMC biosensors. (2) Fe-B nanowires were fabricated as nano-scale magnetostrictive biosensor platforms in this research. Since the nanowires should have very high resonance frequencies, the resonance behavior of the nanowires has not been characterized and the performance of the Fe-B alloy at giga-hertz range remains unknown. The current system has to be extended to characterize sensors with very high resonance frequency. Both frequency-domain and time-domain technology should be explored. (3) To fabricate micro/nano MSP/MSMC with high performance, the immobilization of self-assembled monolayer (SAMs) of some chemicals on the sensor surface has 247 to be studied in order to replace the gold layer used currently, since the gold layer (~40 nm thick) would be a extremely high mass load for micro/nano MSP/MSMCs. The SAM should be dense in order to protect the sensor from corrosion. The free end of the SAM has to be active to react with antibody/phage.