University of Southern California Dissertations and Theses

Ultrasonic microelectromechanical system for microfluidics, cancer therapeutics and sensing applications

(USC Thesis Other)

Ultrasonic microelectromechanical system for microfluidics, cancer therapeutics and sensing applications

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ULTRASONIC MICROELECTROMECHANICAL SYSTEM FOR MICROFLUIDICS, CANCER THERAPEUTICS AND SENSING APPLICATIONS by Lingtao Wang A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (ENGINEERING) December 2013 Copyright 2013 Lingtao Wang ii Acknowledgements First, I want to express my sincere thanks to my advisor, Dr. Eun Sok Kim, for his warm encouragement and thoughtful guidance. He has been supportive since the first day I began Ph.D. study at USC. Ever since, Dr. Kim has supported me not only by providing a research assistantship over 5 years, but also academically and emotionally supporting me to explore infinite possibilities in MEMS field, without which it would have been impossible to carry out this work. My thesis committee guided me through these years. Thank you to Dr. Mitchell Gross and Dr. Chongwu Zhou for reading this thesis and their advice on this thesis. Dr. Gross supported and guided me through the SFAT cytolysis project. He taught me valuable biological techniques in cancer researches, helped me come up with the SFAT testing setup and guided me over almost three years of development in the project. Dozens of people have helped and taught me immensely in MEMS group and CAMM at USC. My deepest gratitude goes to Dr. Youngki Choe, Dr. Shih-jui Chen, Dr. Anderson Lin in MEMS group, who trained me and taught me valuable skills regarding design and fabrication of MEMS devices. Dr. Choe helped me a lot when I first joined this group and provided indispensable advice and support for my researches. My special thanks go to Dr. Yi-Jia Li, Dr. Chun-Peng Liao and Joanna Chen in CAMM, for your tremendous help to our biological experiments, including iii experiments planning, cell sample preparations, and results analysis. All the results about SFAT for cancer therapeutics presented in this thesis are due to a close collaboration and fruitful discussions we had together. I also want to thank my colleagues, Qian Zhang, Lukas Baumgartel, Arash Vafanejad, and Yufeng Wang for their friendship, discussions and assistance in research. I also appreciated Dr. Changyang Lee in Dr. Shung’s group for his great help in hydrophone measurements. Finally, but most importantly, I want to express my deepest thanks to my beloved beautiful wife, Moqi Liu, for supporting me for everything and encouraging me throughout these years. Your smile and kindness has turned my Ph.D. years sweet and fruitful. Also, I would like to give special thanks to my parents, my parents-in- law and my grandparents, for their sacrifices and endless love for me. iv Table of Contents Acknowledgements ...................................................................................................... ii List of Tables ............................................................................................................ viii List of Figures ............................................................................................................. ix Abstract ...................................................................................................................... xv Chapter 1 Introduction ............................................................................................. 18 1.1 Review of droplet ejections and directional ejector control methods .................. 18 1.2 Review of ultrasound for cancer therapeutics ..................................................... 21 1.3 A simple, low-cost and accurate frequency shift detection method using a phase-locked-loop circuit ........................................................................ 23 1.3.1 Detections of Doppler frequency shift in an ultrasonic velocity sensing system ............................................................................................ 24 1.3.2 Measurements of FBAR-based oscillator’s frequency shift for biochemical sensing applications ............................................................... 25 1.4 Overview of the chapters ..................................................................................... 27 Reference ................................................................................................................... 27 Chapter 2 High-frequency SFAT Ejector with an Electrical Control of Directional Droplet Ejections .................................................................................... 32 2.1 Working principle and design .............................................................................. 32 2.1.1 Phase-varied lens design for directional droplet ejections ......................... 34 2.1.2 Dual-frequency half and half lens design ................................................... 38 2.2 Device fabrication and testing setup .................................................................... 39 2.3 Experimental results ............................................................................................ 43 2.4 Summary .............................................................................................................. 49 Reference ................................................................................................................... 49 v Chapter 3 Localized Cytolysis System using Ultrasonic Self Focusing Acoustic Transducers ................................................................................................. 50 3.1 Localized cell lysis by ultrasonic transducers ..................................................... 50 3.2 Design and fabrication of SFAT for cell lysis ..................................................... 51 3.3 Modeling of SFAT’s focus and bilayer cellular membrane model ..................... 55 3.4 Cell culture and tumor spheroids formation ........................................................ 64 3.5 Testing setup and experimental procedure .......................................................... 65 3.5.1 Fluorescent microscopy .............................................................................. 66 3.5.2 Semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) ...................................................................................... 68 3.6 Experimental results ............................................................................................ 69 3.6.1 Highly focused and localized ultrasonic cytolytic effect by the SFAT ........................................................................................................... 69 3.6.2 Considerations for other effects caused by ultrasound exposures .............. 74 3.6.3 Sensitivity to acoustic irradiation varies in benign and malignant cell lines ...................................................................................................... 75 3.6.4 Sensitivity to acoustic irradiation varies due to variations in cell’s cytoskeletons ............................................................................................... 78 3.6.5 Sensitivity to acoustic irradiation varies due to variations in cell’s cytoskeletons ............................................................................................... 78 3.7 Summary .............................................................................................................. 84 Reference ................................................................................................................... 85 Chapter 4 Self Focusing Acoustic Transducers (SFATs) with 10-mm Focal Length for Cancer-specific Localized Cytolysis of 3D Cell Spheroids in 3D Matrigel ........................................................................................... 89 4.1 3D cell spheroids in 3D Matrigel environment ................................................... 89 4.2 Material and methods .......................................................................................... 92 4.2.1 Fabrication and design of SFAT with 10 mm focal length working at 17.3 MHz .................................................................................. 92 4.2.2 Fabrication and design of SFAT with 10 mm focal length working at 2.14 MHz .................................................................................. 94 4.2.3 Assays for spheroids formation and transfer .............................................. 95 vi 4.2.4 Testing setup and experiment procedure for spheroids cell lysis by SFAT working at 17.3 MHz and 2.14 MHz .......................................... 96 4.3 Results and discussions for spheroids cell lysis by SFAT working at 17.3 MHz ............................................................................................................. 97 4.3.1 Hydrophone measurement of focused ultrasound with 10 mm focal length ................................................................................................. 97 4.3.2 Localized cell lysis of spheroids in 3D Matrigel environment ................... 98 4.3.3 Localized cell lysis of spheroids in “unsolidified” Matrigel environment .............................................................................................. 100 4.3.4 Localized cell lysis of spheroids in the purified collagen gel environment .............................................................................................. 103 4.4 Results and discussions for spheroids cell lysis by SFAT working at 2.14 MHz ........................................................................................................... 104 4.4.1 Localized cell lysis of 2D monolayer cells ............................................... 104 4.4.2 Localized cell lysis of 3D spheroids ......................................................... 106 4.5 Summary ............................................................................................................ 108 Reference ................................................................................................................. 109 Chapter 5 Phase-locked-loops (PLLs) for Ultrasonic Velocity Sensing System and FBAR oscillator sensor frequency shift detection ............................... 111 5.1 Phase-locked-loop for Doppler frequency shift detection in ultrasonic velocity sensing ................................................................................................. 111 5.2 Velocity sensing system design ......................................................................... 113 5.3 Experimental results .......................................................................................... 118 5.3.1 Frequency response of the transmitter and the receiver ........................... 118 5.3.2 Dynamic range of frequency tuning of the PLL frequency shift detector ..................................................................................................... 119 5.3.3 Real time measurement of the velocity sensing system ........................... 120 5.4 A phase-locked loop for FBAR and HBAR oscillator frequency shift detections ........................................................................................................... 122 5.5 System design of FBAR oscillator sensor and phase-locked loop frequency shift detection .................................................................................... 124 5.6 Results and discussions ...................................................................................... 126 5.6.1 Testing setup ............................................................................................. 127 vii 5.6.2 Measurements of system’s dynamic range and voltage output ................ 128 5.6.3 Measurements of voltage output in real-time ........................................... 130 5.6.4 Real-time measurement of FBAR sensor for temperature sensing .......... 131 5.7 Summary ............................................................................................................ 134 Reference ................................................................................................................. 135 Chapter 6 Conclusion and Future Directions ......................................................... 137 6.1 Conclusion ......................................................................................................... 137 6.2 A combinatory array of directional ejectors with electrical control of the droplet directions ......................................................................................... 139 6.3 Combing a FBAR sensor with antibody mobilizations, a PLL frequency shift detector and SFATs for a Lab-on-a-chip cancer diagnosis and therapeutics system ..................................................................... 139 Reference ................................................................................................................. 141 Bibliography ............................................................................................................ 142 viii List of Tables Table 2-1: Ejection directions and corresponding operating frequencies for different phase-shifts (P-S) and different electrodes, i.e., round electrode (RE) and RE without center part (RE w/o center). .................................................................................................. 47 Table 2-2: Ejection directions and corresponding operating frequencies for different dual-frequency ejector designs ......................................... 48 Table 3-1: Specific values used in the simulation .................................................. 60 Table 3-2: Acoustic Intensity Thresholds .............................................................. 77 Table 3-3: Summary of Acoustic Intensity Threshold Variation of MCF- 10A ....................................................................................................... 84 Table 4-1: Summary of Acoustic Intensity Thresholds (AITs) of MCF-7, 22RV1 and MCF-10A spheroids in various 3D biological environments ....................................................................................... 103 Table 5-1: Parameter values in the low pass filter in PLLs ................................. 117 ix List of Figures Figure 2.1: (a) and (c) Cross-sectional views of the ejector with the acoustic lens and the packaged ejector, respectively; (b) Photo of ejector lens showing the regular rings with 2 π phase-shift between adjacent rings (on the left) and the phase varied rings with 2.5 π phase-shift between adjacent rings (on the right). ............................................................................................... 34 Figure 2.2: Effect of propagation path difference on wave interference with Fresnel ring pattern ....................................................................... 35 Figure 2.3: Simulation results of the particle displacement vector and ejection direction (red arrow) for the phase-varied ejector working at different operating frequencies (phase-shift of the phase-varied rings is 2.5 π) .................................................................... 37 Figure 2.4: Fabrication process of the ejectors with electrically adjustable droplet direction .................................................................................... 40 Figure 2.5: Measurement setup for ejection: the droplet ejection is stroboscopically captured with a delay time controller (to control the observation time) and an LED ............................................ 41 Figure 2.6: Photos of various phase-shift-ring lens and electrode shapes .............. 42 Figure 2.7: Pictures of dual-frequency ejectors ring pattern: (a) left half ring for 17 MHz, right half ring for 19 MHz; (b) left half ring for 18 MHz, right half ring for 19 MHz; (c) left half ring for 18 MHz, right half ring for 17 MHz; (d) left half ring for 17.5 MHz, right half ring for 18 MHz. ................................................. 43 Figure 2.8: Experimental results with the phase-varied ejector (phase- shift of the phase-varied rings is 2.5 π) .................................................. 44 Figure 2.9: Experimental results with the dual-frequency ejector (the operating frequencies of left and right half of lens are 18 MHz and 19 MHz, respectively.) ......................................................... 45 Figure 2.10: Simulation and experimental results for the phase-varied directional ejector with half regular rings and half 2.5 π phase-varied rings. ................................................................................ 45 Figure 2.11: Experimental results for the dual-frequency directional ejector with left half rings at 18 MHz and right half rings at 19 MHz. ................................................................................................ 46 Figure 2.12: Time evolution of the droplet formations and directional ejections at different operating frequencies (f w ) and observation times. ................................................................................. 47 x Figure 3.1: (a) Cross-sectional view of the cytolysis device with an acoustic Fresnel lens on a PZT transducer, a Si acoustic chamber, and a Si cell culture chamber; (b) Schematic of the observation method of cytolysis ........................................................... 53 Figure 3.2: Fabrication process of the cell lysis device with a PZT transducer with acoustic lens, an ultrasound chamber and a Si cell culture chamber .............................................................................. 54 Figure 3.3: Coordinate system for the equations leading to the acoustic pressure distribution produced by SFAT. ............................................. 57 Figure 3.4: Schematic illustrating a bilayer cell membrane under uniform acoustic pressure ................................................................................... 58 Figure 3.5: Estimation of acoustic intensity at the focal point: (a) the simulation of acoustic pressure distribution along one axis (say, x-axis) on the focal plane when the voltage applied to SFAT is 10 volts; (b) the simulation of the relative acoustic pressure distribution on the cross-sectional x-z plane, when 10 volts is applied to the SFAT; and (c) simulated area strain of cells membrane generated by focused acoustic pressure (acoustic intensity I ac =1.944W/cm 2 ). The value of area strain is set to k s =0.03, the diameter of area of the free bilayer membrane confined by proteins in the membrane is 2a=25nm. .............................................................................................. 62 Figure 3.6: Three steps of fast screening based on a SFAT and the fluorescent stained cytolysis assay ....................................................... 67 Figure 3.7: Cytolysis with electric power applied to the device at 17.3 MHz: (a) The SFAT-sonicated 22RV1 cells were labelled by SYTOX dead cell dye and examined under a fluorescent microscope. The SFAT was operated at 17.3 MHz with applied acoustic intensity, at the focal spot, of (from top to bottom) 0.117W/cm 2 , 0.150 W/cm 2 , 2.217 W/cm 2 , and 8.88 W/cm 2 . The cytolysis areas were around 100 µm. (b) SYTOX fluorescent intensity within the focal area versus acoustic intensity. (c) The RNAs released from the cells were collected before and after the ultrasound irradiation. After the reverse transcription and PCR amplification of the collected RNAs, one housekeeping gene GAPDH was separated by electrophoresis and quantified by ImageJ software, and the measured amount of change due to the sonication is shown. .............................................................................. 70 Figure 3.8: Measured size of the cytolysis area caused by various acoustic intensities ................................................................................ 72 xi Figure 3.9: Labeled lysed cells caused (a) by the device working at 17.3 MHz and 9 mW electric power applied (the labeled size is 80 × 120 µm2); (b) by the device working at 56 MHz and 0.10 W electric power applied (the labeled size is 55 × 80 µm 2 ); (c) by the device working at 56 MHz and 0.18 W electric power applied (the labeled size is 70 × 70 µm 2 ). .................................. 73 Figure 3.10: (a) Superimposed bright field and fluorescence images showing selective cytolytic effects of SFAT at 1 W/cm 2 on benign (RWPE1, MCF-10A and Detroit 551) and malignant (22RV1, MCF-7 and A375) cell lines that represent prostate, breast and skin tissues. (b) The actin filament patterns in each cell line being visualized with green color by fluorescent conjugated phalloidin (these photos were taken before the ultrasound irradiation). (c) The actin distribution intensity in each cell line: the actin distribution intensity was quantified with fluorescence intensity using ImageJ software across the lines shown in the corresponding panels in the actin staining microscopy (b). ..................................................................................... 77 Figure 3.11: Cytolytic effect of SFAT irradiation on 3D tumor spheroids from benign (RWPE1 and MCF-10A) or malignant (22RV1 and MCF-7) cell lines. The bar represents 100 µm. The lysed cells were stained by SYTOX green fluorescence dye in the 3D spheroids when the applied peak acoustic intensities were 2.4 W/cm 2 . The incident direction of the focused acoustic beam is perpendicular to the plane of the photo. .................................. 79 Figure 3.12: Simulated maximum area strain vs. area compression modulus (k s ) as a function of free lipid area (2 πa) for 2.4 W/cm 2 acoustic intensity ...................................................................... 81 Figure 3.13: The change of the acoustic intensity sensitivity due to the alterations of cytoskeleton organization by actin polymerization inhibitors (MCF-10A cells are treated with Y- 27632, blebbistatin and nocodazol prior to acoustic irradiation): (a) actin pattern visualized by fluorescent- labelled phalloidin staining and microscopy, the across lines shown in the corresponding panels on the left are indicative of actin distribution density in the case of actin localization); (b) the AIT measured by acoustic irradiation after drug treatments; (c) the correlation between actin distribution intensity and AIT represented by a bar graph (error bars show standard deviations calculated by 10 individual cells) ......................... 83 xii Figure 4.1: Photos of (a) the front view and (b) the side view of a SFAT with an acoustic chamber; (c) the front view and (d) the back view of a spheroids chamber with a transparent Parylene- Si x N y -Parylene bottom membrane ........................................................ 92 Figure 4.2: Fabrication steps of the SFAT, the acoustic chamber and the spheroids chamber. ............................................................................... 93 Figure 4.3: Photos of (a) front view a SFAT with patterned electrodes working at 2.14 MHz and (b) the SFAT with an adjustable acoustic chamber for adjusting the distance from the SFAT’s acoustic lens .......................................................................................... 94 Figure 4.4: Three steps of determining the AITs of spheroids based on a SFAT and the fluorescent stained cytolysis assays as well as pictures of (a) grown-up MCF-7 spheroids and (b) grown-up MCF-10A spheroids in the 3D matrigel environment .......................... 97 Figure 4.5: (a) Measured acoustic intensity distribution along x-axis on the focal plane and (b) measured acoustic intensity values vs. voltages applied to the SFAT. .............................................................. 98 Figure 4.6: In the normal Matrigel, (a) the AIT of MCF-7 spheroids is increased from 0.113 W/cm 2 (for spheroids in “unsolidified” Matrigel) to 11.10 W/cm 2 due to the crosslinks between spheroids and Matrigel; and (c) MCF-7 spheroids in normal Matrigel are lysed by acoustic intensity of 15.14 W/cm 2 . In pure collagen gel, AITs of (b) MCF-7 and (d) 22RV1 spheroids are increased to 15.14 W/cm 2 because of the crosslinks of collagen between the gel and the spheroids. ................. 100 Figure 4.7: In “unsolidified” matrigel, (a) MCF-7 and (b) 22RV1 spheroids are lysed with acoustic intensity of 0.113 W/cm 2 and (c) MCF-10A spheroids cannot be lysed. .................................... 102 Figure 4.8: MCF-7 Monolayer cells treated by SFAT working at 2.14 MHz with 10 mm focal length with various acoustic intensities ............................................................................................ 105 Figure 4.9: MCF-10A Monolayer cells treated by SFAT working at 2.14 MHz with 10 mm focal length with various acoustic intensities ............................................................................................ 106 Figure 4.10: 3D spheroids of MCF-7 treated by SFAT working at 2.14 MHz with various duration times ....................................................... 106 Figure 4.11: 3D spheroids of MCF-10A treated by SFAT working at 2.14 MHz with various duration times from 1 second to 1 minute ............ 107 Figure 5.1: Schematic of a phase-locked-loop ...................................................... 112 xiii Figure 5.2: Schematic diagram of velocity sensing .............................................. 115 Figure 5.3: Close-up detail of the cascade PLL module ....................................... 116 Figure 5.4: Schematic of a 3rd order passive low pass filter ................................ 117 Figure 5.5: Simulation results of the phase noise of a PLL .................................. 117 Figure 5.6: Testing setup to characterize the frequency response of (a) the transmitter and (b) the receiver. .................................................... 119 Figure 5.7: Frequency response of (a) MEMS receiver, (b) MEMS transmitter ........................................................................................... 120 Figure 5.8: Output voltage vs. received frequency. (a) Frequency sweeping from 28.8174 kHz to 28.8177 kHz, the rate is 0.1 Hz/s (b) output voltage vs. received frequency .................................. 120 Figure 5.9: Pictures of (a) a packaged MEMS transmitter, (b) MEMS receiver and preamplifier, (c) close up picture of MEMS receiver, (d) a transmitter is placed on the surface of shaker and a receiver is enclosed in a metal box fixed by a holder. .............. 121 Figure 5.10: Voltage output when the shaker is driven a by sine wave with varying frequency, 1Hz, 0.5Hz, 0.25Hz and 0.1Hz. ........................... 122 Figure 5.11: Schematic of FBAR-based Pierce oscillator circuit ........................... 123 Figure 5.12: Schematic of FBAR frequency shift detection system ....................... 124 Figure 5.13: The frequency spectrum peaks at the oscillation frequencies of a HBAR oscillator, a FBAR oscillator and a frequency synthesizer .......................................................................................... 127 Figure 5.14: Pictures of the fabricated FBAR oscillator and designed PLL frequency shift detecting system ........................................................ 128 Figure 5.15: Testing setup to measure the dynamic range of the frequency shift detector: f IF is changing from 49.99 MHz to 50.01 MHz, while R is set at 5. The voltage output is linearly changing from 0 V to 5 V according to the frequency shift. The noise level is ±0.54 V (R=5). ....................................................................... 129 Figure 5.16: Real-time measured frequency difference f IF =f LO -f r (blue line with smaller amplitude) and voltage output V out (red line with larger amplitude). The detection range is from 49.99 MHz to 50.01 MHz (R=5), and the V out tracks the f IF . ..................................... 130 Figure 5.17: Measured voltage output V out vs the frequency shift from about -10 kHz (f IF = 49.99 MHz) to about +10 kHz (f IF = 50.01 MHz). The slope of the frequency shift detector is 0.207 V/kHz. ....................................................................................... 131 xiv Figure 5.18: Real-time voltage output of the frequency shift detector measuring the FBAR oscillator frequency shift due to the temperature variations: (a) at the room temperature, the oscillation frequency is lower than the reference frequency; (b - c) as the blue ice is taken close to the oscillator, the oscillation frequency jumps up because of the dramatic decrease of the environment temperature; (d - e) the oscillation frequency gradually decreases as the ice is taken away from the oscillator; and (f) the ambient temperature changes back to the initial value. ........................................................ 133 Figure 5.19: The frequency changes of the FBAR oscillator measured by a frequency counter as the ice is approaching to and departing from the FBAR oscillator ................................................................... 134 xv Abstract This thesis presents ultrasonic Micro-Electro-Mechanical Systems (MEMS) designed for microfluidic, biomedical and physical sensing applications, including a nozzleless micro-ejector with “phase-varied” and “dual-frequency” acoustic lens for the electrical control of droplet direction; high-frequency Self-Focusing Acoustic Transducers (SFAT) for three-dimensional localized cell lysis for cancer therapeutics; and the design of sensing system with a phase-locked loop frequency shift detection for a Doppler velocity sensing system and for FBAR-based oscillator sensors. For a multi-directional acoustic ejector with capability of electrical control on the droplet ejection angle, a novel design of the acoustic lens with “phase-varied” and “dual-frequency” patterns are developed and tested. With the novel lens, the direction of the droplet ejection depends monotonically on the operating frequency of the driving signal. The newly developed ejector consistently ejects uniform droplets in diameter of 70 µm, with electrical control of the directional angle from - 30° to 35° (with respect to normal direction of liquid surface plane) as the operating frequency is varied from 16.78 MHz to 19.08 MHz. For three-dimensional localized cell lysis for cancer therapeutics, Self Focusing Acoustic Transducers (SFAT) are designed and fabricated for focusing acoustic energy within an area of 100 μm in diameter at 800 μm focal length for applications xvi requiring high spatial resolutions. Then, the SFAT with large focal area of 1 mm and long focal length of 10 mm are demonstrated for fast ultrasound treatments with large area coverage in a three-dimensional environment. In order to minimize potentially non-specific heat or cavitation effects by acoustic irradiations, operational parameters are optimized to study bio-effects of the device in the absence of tissue heating or biological effects due to cavitations. In the case of 2D monolayer cells, the Acoustic Intensity Thresholds (AIT) for cell lysis in various cell lines representative of benign and malignant prostate, breast and skin cells are investigated. And it is observed that lower AIT’s in cancer cells over non-malignant variants. Actin staining of cytoskeletal structures indicates an association between a pattern of diffuse and less organized actin filaments with decreased AIT. Moreover, the same trend of decreased actin organization and decreased AIT is observed following treatments of changing actin patterns in the MCF-10A breast epithelial cell line. These results suggest that biomechanical properties make malignant cells specifically sensitive to cytolysis caused by this form of acoustic energy. In the case of 3D cell spheroids, the acoustic intensity thresholds (AITs) of spheroids for cancer-specific cytolysis of malignant cells (breast cancer MCF-7 and prostate cancer 22RV1) are investigated. According to experiments with the spheroids in the 3D Matrigel environment, the crosslinks between Matrigel and spheroids greatly increase AITs of 22RV1 and MCF-7 cell spheroids from 0.11 W/cm 2 to 11.10 ~ 15.14 W/cm 2 . In addition, in both “unsolidified” (without crosslink) and solidified (with crosslink) Matrigel environments, spheroids of non-malignant cell line MCF-10A are not lysed even xvii with two to three times higher acoustic intensities than the AITs of MCF-7. Therefore, the general trend associating a lower AIT with a more malignant still stands for MCF-7 and MCF-10A 3D cell spheroids in the 3D Matrigel environment. For the design of ultrasonic sensing systems, a novel, highly-sensitive ultrasonic Doppler velocity sensing system, and a hand-held, low-cost but highly sensitive sensing system built with an FBAR-based oscillator are designed, fabricated and tested. The ultrasonic velocity sensing system is a compact velocity sensing system, in which MEMS ultrasonic transducers are incorporated with phase-locked-loop (PLL) circuitry for frequency detection and signal processing. The achieved voltage- velocity sensitivity is 0.22 V/(mm/s) and the minimum detectable velocity is 0.67 mm/s, corresponding to 0.11 Hz in Doppler frequency. Also, the output of the PLL is a DC voltage linearly related to the velocity, and there is no need to convert the frequency shift to analog voltage. The FBAR-based sensor with a phase-locked-loop (PLL) for ppm-level detections and signal processing is a novel potable sensing system for remote sensing applications. The achieved voltage-frequency sensitivity is 1.035 V/kHz with the minimum detectable frequency shift of 4.81ppm (5.21kHz), and the dynamic range is 134 ppm (145kHz). Also, the output of the sensing system is purposely designed to be a DC voltage that is linearly related to the frequency shift. To our knowledge, it is the first FBAR-based sensing system that converts frequency shift to DC voltage. 18 Chapter 1 Introduction This thesis presents ultrasonic Micro-Electro-Mechanical Systems (MEMS) designed for microfluidic, biomedical and physical sensing applications, including a nozzleless micro-ejector with a phase-varied acoustic lens for electrical controls of droplet directions (Chapter 2); high-frequency Self-focusing Acoustic Transducers (SFAT) for three-dimensional localized cell lysis for cancer therapeutics (Chapter 3 and Chapter 4); and the design of sensing system with a phase-locked loop frequency shift detector for a Doppler velocity sensing system and for a FBAR oscillator sensor (Chapter 4). General reviews of these areas are covered in this chapter. Design, fabrication and testing of devices and experimental results are shown in Chapter 2 – 4. 1.1 Review of droplet ejections and directional ejector control methods Micro-droplet ejectors have a great potential in applications like ink-jet printing [1, 2], drug delivery [3, 4] and DNA synthesis [5-7]. Droplet ejection mechanisms that have been developed include nozzle-based droplet ejection with piezoelectric materials [2, 8], thermal [1, 9], electrostatic actuations [10], as well as nozzleless 19 droplet ejections with a focused acoustic beam [11-14]. Those methods have been studied in order to achieve a high resolution, a fast dispensing speed, and a good repeatability at low cost. In the piezoelectric or electrostatic actuation methods, a mechanical bending caused by piezoelectric or electrostatic actuation generates a hydrostatic pressure underneath the liquid chamber, therefore forms a droplet at the nozzle. In the thermal methods, a heater produces gas bubbles, and generates liquid droplets ejected from a nozzle through the growth and collapse of bubbles. Those methods generate nozzle-defined droplets, which may cause the nozzle clogging due to the small size of nozzles. In addition, the nozzle-based ejector could only eject droplets in a direction perpendicular to nozzle’s liquid surface. Nozzle-less acoustic-wave ejectors, however, can readily be designed for directional ejection. A focused acoustic beam could generate high particle velocity and eject droplet from the liquid surface without any nozzle. Various acoustic focusing mechanisms have been developed. A spherical acoustic lens could be fabricated by mechanical grinding and polishing [11], which have a high fabrication cost and a complicated process. For planar acoustic lenses, traditional Fresnel acoustic lenses require a tight control on lens thickness, in order to obtain phase difference of π between the waves traveling through the liquid and the lens [15]. Surface acoustic waves (SAWs) generated by the inter-digital transducers (IDT) are also used for focusing a beam. However, the F-number of IDT lens is smaller than one because of the small Raleigh angle [12]. By designing the electrodes into annular rings, acoustic waves generated from self-focused acoustic transducer (SFAT) with 20 Fresnel Half-Wave-Band Sources are focused on the liquid surface to form the near- field constructive interference [3, 14]. The fabrication of SFAT is very simple, but the device has issues of the undesirable heat generation and effects of lateral electric field. In order to have more effective focus, lens employing parylene air-reflectors based on SFAT has been developed [13]. Instead of patterning electrodes, air cavities in annular rings are formed on the top of the circular electrodes to focus acoustic waves by having constructive wave interference at the focal point. This new type of lens has wide tolerance for its lens geometry, and has been shown to be very effective in focusing acoustic waves for droplet ejection. For many applications such as the DNA synthesis application, directional droplet ejectors are highly desirable to deliver different solutions to the same spot without a physical movement of the ejectors [7]. An ejector with nozzles is difficult to realize directional droplet ejection due to the fact that droplets usually exit from a nozzle in a direction perpendicular to the nozzle plane. The SFAT acoustic-wave directional ejectors used sectored electrode patterns electrodes for oblique ejections, but the ejection direction is fixed once the ejector is fabricated [16]. In this work, a novel design of a multi-directional acoustic ejector is developed and demonstrated. To obtain electrically adjustable direction of droplet ejections, a phase-varied lens and dual-frequency lens are designed to generate unbalanced body force at the liquid-air interface. In this way, the direction of the droplet ejection is controllable, and it depends on the operating frequency of the ejector. 21 1.2 Review of ultrasound for cancer therapeutics Studies relating to ultrasonic wave propagation in tissue have important diagnostic and therapeutic applications. Diagnostically, transmission and reflection of ultrasonic waves at low intensities (acoustic intensity of spatial peak temporal average I SPPA <0.1 W/cm 2 ) are used as a non-invasive imaging modality [17]. Therapeutically, high-intensity focused ultrasound (HIFU) has been used to deliver energies (I SPTA >10 3 W/cm 2 over 1-5 seconds) associated with rapid tissue heating and coagulative necrosis [18]. Very intense and brief acoustic pulses (acoustic intensity of spatial peak temporal peak I SPTP =10 5 W/cm 2 for 0.1 µs) are used for extracorporeal lithotripsy of kidney stones, as the intense acoustic pulses cause micro-bubble formation and rapid collapse (inertial cavitation) at the interface of the solid-liquid boundary [19]. Biomedical applications have utilized thermal and non-thermal effects of acoustic energy in tissue for therapeutic intent. In recent years, High Intensity Focused Ultrasound (HIFU) has been developed as a new treatment for a variety of primary localized cancers, including breast, prostate, kidney, liver, bone, and uterus cancers. The predominant biologic effect occurs as tissue heating as the low frequency (1 – 4 MHz) energy is absorbed in the tissue and converted into heat [20]. However, limitations of thermal effects produced by HIFU are a relatively large focal area and non-specific effects on both healthy and diseased tissues, as heat will 22 inevitably transfer outside of the targeted area, causing unwanted tissue damage. Even though HIFU has been available for localized prostate cancer more than a decade, it has found limited utility due, in part, to unintentional damage to normal tissues causing edema, necrosis, and bladder outlet obstruction [20, 21]. Cavitation is the most commonly considered non-thermal effect of acoustic energy with tissue. At very high acoustic intensities (> 10 3 W/cm 2 ), gas bubbles may be generated from submicron-sized cavitation nuclei and undergo steady vibration or rapid collapse. These gas bubbles can form on, and interact with, cells or tissues to produce biological effects [22]. Cavitation is the primary mechanism for the effects of extracorporeal lithotripsy of kidney stones and may contribute to a minor component of the tissue effects produced by HIFU [23]. However, the very high pressures and erratic forces produced by cavitation may result in damage to normal cells and tissues. Recently, Micro-Electro-Mechanical System (MEMS) techniques are used to design and fabricate a Self-Focusing Acoustic Transducer (SFAT) to study effects of acoustic irradiation on living cells focused in a size of sub-millimeter in diameter. And the SFAT are used to provide a focal point of acoustic intensity as high as 25 W/cm 2 [24]. At this high acoustic intensity, the acoustic waves produced cavitation effects that detached monolayer of non-malignant vascular cells grown on a glass slide from the focal area of 150 micrometers in diameter. 23 In this work, SFAT generate a focused ultrasound in a low intensity range (0.1-1 W/cm 2 ), in order to study the bio-effects of the highly focused acoustic energy without thermal or cavitation effects. That is because high intensity induced thermal or cavitation effects potentially cause non-specific acoustic irradiation on biological samples. At low intensity, the focused acoustic beam generates negligible heat, as reported in [13] and confirmed in our experiments, and delivers direct radiation force on cells or tissues without cavitation. Consequently, the interactions between the cell membrane and the acoustic beam are dependent on the mechanical properties of the cells. It has been reported that malignant cells possess different biomechanical properties due to changes in cytoskeletal architecture [25]. Hence, sensitivity to ultrasonic irradiation may differ between malignant and non-malignant cells. Our experimental results also suggest that direct radiation pressure produced by the SFAT devices causes a differential cytolytic effect on malignant and benign cells due to differences in the cytoskeletal organization. 1.3 A simple, low-cost and accurate frequency shift detection method using a phase-locked-loop circuit The phase-locked-loop (PLL) circuit is widely used in RF and wireless communication applications for timing clock recoveries, clocking synchronizations and the frequency synthesis [26]. In this thesis, we introduce a simple but accurate frequency shift detection method employing PLLs in two applications. The first 24 application is using the frequency shift detector to measure the Doppler frequency shift in an ultrasonic velocity sensing system. And second application is detecting a film bulk acoustic resonator (FBAR) based oscillator’s frequency shift for biochemical sensing applications. 1.3.1 Detections of Doppler frequency shift in an ultrasonic velocity sensing system The velocity sensing systems are widely used in speed measurement and navigation applications, such as GPS navigation system [27], automobile adaptive cruise control system [28], flow rate sensors [29] and air velocity sensors [30]. Laser velocimetry, millimeter- or micrometer-wave velocity sensor, and ultrasonic velocity sensing system are most common methods [31-35]. The Doppler effect is widely employed in velocity sensing systems. The Doppler method measures the movement of an object by detecting the Doppler-frequency-shift between the original laser light, millimeter- or micrometer waves, or ultrasound signal and received signal from the object. Laser Doppler velocimetery (LDV) has desirable features, including high sensitivity, high accuracy and non-invasive detections. But it suffers from large optical setup, high accuracy requirement for optical alignment, and high cost. Therefore, a novel self-mixing technique with small laser diodes are used in laser velocimetry to reduce the size, improve the alignment and lower the cost. However, it still requires a high-cost and complicated spectrum analysis on self-mixing interferometric signal [36]. For millimeter- or micrometer-waves or ultrasound 25 methods, traditionally, to extract the information about velocity from Doppler signal, the velocity sensing system requires high-calculation-cost Digital Signal Processing (DSP) components [34, 37] or bulky spectrum analyzer instruments [38] for velocity conversion and calculation. Therefore, the complexity of the system increases the cost and causes the difficulties in practical applications. In order to design a portable and more integrated ultrasonic velocity sensing system for real time measurement applications, we developed a low-cost ultrasonic Doppler velocity sensing system employing MEMS transducers and phase-locked loops (PLLs) circuits. This system has many advantages, including simple structure, low calculation cost for velocity measurement, high sensitivity and DC voltage output signal representing velocity information. Moreover, it facilitates detecting very low velocity (the range of sub mm/s) for a personal navigation system. 1.3.2 Measurements of FBAR-based oscillator’s frequency shift for biochemical sensing applications Having advantages of high sensitivity, tiny sizes, flexibility in shapes, real-time measurements, and low-cost in mass production, micro-machined film-bulk-acoustic resonators (FBAR) have been developed and used widely for various sensing applications, including gas sensors, chemical and biological sensors [39]. The sensors have been demonstrated for detecting various vapors [40], DNA sequencing [41], protein sensing [42], and immunodetections [43, 44]. The demonstrated 26 sensitivity of biological sensing could be down to ng/mL in liquid. In the biological sensing applications, the surface of a FBAR sensor is usually chemically modified, so specific chemicals or biological molecules could bond to the surface of FBAR. FBAR-based sensors generate high frequency acoustic waves and changes of FBAR’s resonant frequency represent that sensing chemicals or biological molecules bonded on the FBAR’s surface. There are various methods of measuring the resonant frequency shift, such as optical measurements by a laser Doppler vibrometer [43], and electrical measurements using impedance analyzer [45] and network analyzer [42, 46]. In optical measurement method, the laser focuses on the surface of the resonator and measures the deflection of the resonator. However, the optical method has a limited dynamic range, and it requires large optical setup and complicated alignment of optical lenses. Measurements with an impedance analyzer or a network analyzer have a high accuracy and show the results in real-time. But the bulky size of instruments makes it difficult to use in portable applications. In biological sensing applications, a high-throughput array usually consists tens to hundreds of sensors. It is impractical to connect all sensors to expensive and bulky network analyzers, since one analyzer has only 2 – 4 input ports at most. In this work, we developed a portable and simple FBAR-based sensing system with a PLL circuit, in which the FBAR-based oscillator is designed for sensing functions and the PLL is used to convert the FBAR’s frequency shift to variations of a DC voltage output. It provides possibilities to design an array of FBAR sensors with a PLL sensing array for high-throughput biological sensing applications. 27 1.4 Overview of the chapters In Chapter 1, review of the traditional methods used for directional droplet ejections, ultrasonic cancer therapeutics, velocity sensing and biological sensing applications, as well as the motivation of the thesis work associated with these topics are described as the introductions to the thesis. In Chapeter 2, the working principle, design, fabrication and testing of the electrical controllable directional droplet ejector are presented. In Chapter 3 and Chapter 4, the design and fabrication of SFAT for localized cytolysis for 2D monolayer cells and 3D cell spheroids, as well as the investigations of cell’s resistances to the ultrasound in various cell lines are described. In Chapter 5, PLL-based frequency shift detector working with ultrasonic MEMS transducers as a velocity sensing system and FBAR-based sensors are presented. Finally, Chapter 6 presents conclusions and future research proposals. Reference [1] R. Nayve, et al., "High-resolution long-array thermal ink jet printhead fabricated by anisotropic wet etching and deep Si RIE," Microelectromechanical Systems, Journal of, vol. 13, pp. 814-821, 2004. [2] N. Reis, et al., "Ink-jet delivery of particle suspensions by piezoelectric droplet ejectors," Journal of Applied Physics, vol. 97, pp. 094903-094903-6, 2005. [3] H. Y. Yu, et al., "Chembio extraction on a chip by nanoliter droplet ejection," Lab on a Chip, vol. 5, pp. 344-349, 2005. 28 [4] G. MacBeath and S. L. Schreiber, "Printing proteins as microarrays for high- throughput function determination," Science, vol. 289, pp. 1760-1763, Sep 8 2000. [5] M. Schena, et al., "Quantitative Monitoring of Gene-Expression Patterns with a Complementary-DNA Microarray," Science, vol. 270, pp. 467-470, Oct 20 1995. [6] D. J. Duggan, et al., "Expression profiling using cDNA microarrays," Nature Genetics, vol. 21, pp. 10-14, Jan 1999. [7] L. Chuang-Yuan, et al., "On-Demand DNA Synthesis on Solid Surface by Four Directional Ejectors on a Chip," Microelectromechanical Systems, Journal of, vol. 16, pp. 1130-1139, 2007. [8] M. J. Ahamed, et al., "A Piezoactuated Droplet-Dispensing Microfluidic Chip," Microelectromechanical Systems, Journal of, vol. 19, pp. 110-119, 2010. [9] T. Xu, et al., "Inkjet printing of viable mammalian cells," Biomaterials, vol. 26, pp. 93-99, Jan 2005. [10] L. Sukhan, et al., "An Electrostatic Drop-On-Demand Micro Droplet Ejector with a Pole Type Nozzle," in Nano/Micro Engineered and Molecular Systems, 2006. NEMS '06. 1st IEEE International Conference on, 2006, pp. 1068-1071. [11] S. A. Elrod, et al., "Nozzleless droplet formation with focused acoustic beams," Journal of Applied Physics, vol. 65, pp. 3441-3447, 1989. [12] U. Demirci, "Acoustic picoliter droplets for emerging applications in semiconductor industry and biotechnology," Microelectromechanical Systems, Journal of, vol. 15, pp. 957-966, 2006. [13] C.-Y. Lee, et al., "Acoustic Ejector with Novel Lens Employing Air- Reflectors," in Micro Electro Mechanical Systems, 2006. MEMS 2006 Istanbul. 19th IEEE International Conference on, 2006, pp. 170-173. [14] D. Huang and E. S. Kim, "Micromachined acoustic-wave liquid ejector," Journal of Microelectromechanical Systems, vol. 10, pp. 442-449, Sep 2001. [15] S. C. Chan, et al., "Finite element analysis of multilevel acoustic Fresnel lenses," Ieee Transactions on Ultrasonics Ferroelectrics and Frequency Control, vol. 43, pp. 670-677, Jul 1996. 29 [16] C.-Y. Lee, et al., "Nanoliter droplet coalescence in air by directional acoustic ejection," Applied Physics Letters, vol. 89, pp. 223902-223902-3, 2006. [17] D. L. Miller and J. Quddus, "Diagnostic ultrasound induces membrane damage in phagocytic cells loaded with contrast agent in vitro," in 2000 Ieee Ultrasonics Symposium Proceedings, Vols 1 and 2, S. C. Schneider, M. Levy, and B. R. McAvoy, Eds., ed New York: Ieee, 2000, pp. 1947-1950. [18] S. Madersbacher, et al., "Effect of high-intensity focused ultrasound on human prostate cancer in vivo," Cancer Res, vol. 55, pp. 3346-51, Aug 1 1995. [19] D. Dalecki, "Mechanical bioeffects of ultrasound," Annu Rev Biomed Eng, vol. 6, pp. 229-48, 2004. [20] C. Netsch, et al., "Development of bladder outlet obstruction after a single treatment of prostate cancer with high-intensity focused ultrasound: experience with 226 patients," J Endourol, vol. 24, pp. 1399-403, Sep 2010. [21] A. Blana, et al., "Factors predicting for formation of bladder outlet obstruction after high-intensity focused ultrasound in treatment of localized prostate cancer," Urology, vol. 71, pp. 863-7, May 2008. [22] M. W. Miller, et al., "A review of in vitro bioeffects of inertial ultrasonic cavitation from a mechanistic perspective," Ultrasound in Medicine and Biology, vol. 22, pp. 1131-1154, 1996. [23] C. G. Chaussy and S. F. Thuroff, "Robotic high-intensity focused ultrasound for prostate cancer: what have we learned in 15 years of clinical use?," Curr Urol Rep, vol. 12, pp. 180-7, Jun 2011. [24] X. Q. J. Zhu, J. Oiler, C. Yu, Z. Wang, C. Lee, T.K. Hsiai, E.S. Kim, and H. Yu, "Localized Cell Lysis Using Self Focused Acoustic Transducers," Transducers '09, IEEE International Conference on Solid-State Sensors and Actuators, Denver, CO, vol. June 21 - 25, pp. pp. 608 - 611, 2009. [25] B. Lu, et al., "A study of the autofluorescence of parylene materials for [small mu ]TAS applications," Lab on a Chip, vol. 10, pp. 1826-1834, 2010. [26] B. Razavi, Monolithic phase-locked loops and clock recovery circuits: theory and design: Wiley. com, 1996. [27] D. M. Bevly, "Global positioning system (GPS): A low-cost velocity sensor for correcting inertial sensor errors on ground vehicles," Journal of Dynamic 30 Systems Measurement and Control-Transactions of the Asme, vol. 126, pp. 255-264, Jun 2004. [28] K. Seok-Tae and N. Cam, "A millimeter-wave multifunction sensor for wireless monitoring of displacement and velocity," in Wireless Communications and Applied Computational Electromagnetics, 2005. IEEE/ACES International Conference on, pp. 451-453. [29] W. Huilin, et al., "Laser Doppler velocity measurement system," in 2010 2nd International Asia Conference on Informatics in Control, Automation and Robotics (CAR 2010), 6-7 March 2010, Piscataway, NJ, USA, 2010, pp. 288- 90. [30] A. Le Duff, et al., "Real-time acoustic velocity measurement in the air by means of laser doppler velocimetry," in 18th IEEE Instrumentation and Measurement Technology Conference. Rediscovering Measurement in the Age of Informatics, May 21, 2001 - May 23, 2001, Budapest, Hungary, 2001, pp. 234-238. [31] K. Imou, M. Ishida, T. Okamoto, Y.Kaizu, A. Sawamura, and N. Sumida, "Ultrasonic Doppler Sensor for Measuring Vehicel Speed in Forward and Reverse Motions Inlcuding Low Speed Motions," Agricultral Engineering International: the CIGR Journal of Scientific Research And Development, vol. 3, p. 007, 2001. [32] S. Hirata, et al., "Ultrasonic distance and velocity measurement by low- calculation-cost Doppler-shift compensation and high-resolution Doppler velocity estimation with wide measurement range," Acoustical Science and Technology, vol. 30, pp. 220-3, 2009. [33] Y. Zhang, et al., "Laser Doppler Velocimetry based on self-mixing effect in vertical-cavity surface-emitting lasers," in Electronic Measurement and Instruments, 2007. ICEMI '07. 8th International Conference on, 2007, pp. 1- 413-1-416. [34] K. Seoktae and N. Cam, "On the development of a multifunction millimeter- wave sensor for displacement sensing and low-velocity measurement," Microwave Theory and Techniques, IEEE Transactions on, vol. 52, pp. 2503- 2512, 2004. [35] H. X. Zheng, "A displacement and velocity measurement technique using millimeter-wave sensor," International Journal of Infrared and Millimeter Waves, vol. 26, pp. 1277-1290, Sep 2005. 31 [36] L. Scalise, et al., "Self-mixing laser diode velocimetry: application to vibration and velocity measurement," Instrumentation and Measurement, IEEE Transactions on, vol. 53, pp. 223-232, 2004. [37] O. P. Sahu and A. K. Gupta, "Measurement of Distance and Medium Velocity Using Frequency-Modulated Sound/Ultrasound," Instrumentation and Measurement, IEEE Transactions on, vol. 57, pp. 838-842, 2008. [38] J. Zhang, et al., "A low-cost doppler radar system with 24 GHz PLL for remote detection of heart beat," Microwave and Optical Technology Letters, vol. 50, pp. 3139-3142, 2008. [39] W. Pang, et al., "Piezoelectric microelectromechanical resonant sensors for chemical and biological detection," Lab on a Chip, vol. 12, pp. 29-44, 2012. [40] A. Lin and K. Eun Sok, "Selectivity and Long-Term Reliability of Resonant Explosive-Vapor-Trace Detection based on Antigen-Antibody Binding," in Micro Electro Mechanical Systems, 2009. MEMS 2009. IEEE 22nd International Conference on, 2009, pp. 316-319. [41] Z. Hao, et al., "Sequence Specific Label-Free DNA Sensing Using Film-Bulk- Acoustic-Resonators," Sensors Journal, IEEE, vol. 7, pp. 1587-1588, 2007. [42] H. Zhang, et al., "Label-free detection of protein-ligand interactions in real time using micromachined bulk acoustic resonators," Applied Physics Letters, vol. 96, pp. 123702-3, 2010. [43] K. S. Hwang, et al., "In-situ quantitative analysis of a prostate-specific antigen (PSA) using a nanomechanical PZT cantilever," Lab on a Chip, vol. 4, pp. 547- 552, 2004. [44] A. Lin, et al., "Label-Free Detection of Prostate-Specific Antigen with FBAR- Based Sensor with Oriented Antibody Immobilization," 2011. [45] J. D. Adams, et al., "Nanowatt chemical vapor detection with a self-sensing, piezoelectric microcantilever array," Applied Physics Letters, vol. 83, pp. 3428- 3430, 2003. [46] A. Lin, et al., "Real-time label-free detection of DNA synthesis by FBAR- based mass sensing," in Ultrasonics Symposium (IUS), 2010 IEEE, 2010, pp. 1286-1289. 32 Chapter 2 High-frequency SFAT Ejector with an Electrical Control of Directional Droplet Ejections This chapter describes the working principle, design, fabrication and testing of the electrically controllable directional droplet ejectors. 2.1 Working principle and design The ejector consists of a lead zirconate titanate (PZT) transducer, an air-reflector lens with annular rings pattern, a reservoir and a microfluidic channel (Figure 2.1(c)). Since the thickness of PZT transducer is 127 µm, the PZT’s resonant frequency f r in thickness mode is v/2d=17.3 MHz, where v is sound velocity of PZT (v ~ 4394 m/s) and d is the thickness of PZT. The parylene lens with air cavities on the top of PZT substrate is used to control the transmission and reflection of acoustic waves from the PZT. The transmission and reflection of acoustic waves at the interface between two materials depend on the matching of acoustic impedances. The transmission coefficient between two materials is Γ=1-[(Z 2 /cos α 2 -Z 1 /cos α 1 )/(Z 2 /cos α 2 +Z 1 /cos α 1 )] 2 , where Z 1 and Z 2 are the acoustic impedances of two adjacent materials, and α 1 and α 2 are the incident and transmitted angles of the acoustic waves. Since the acoustic impedance of air (415 Rayl at room temperature) is much smaller than the acoustic impedance of PZT (29.6 MRayl), the acoustic waves produced by the PZT are 33 mostly reflected at the transducer-air interface due to impedance mismatch. On the other hand, the acoustic impedance of parylene is between the acoustic impedance of transducer (29.6 MRayl) and the acoustic impedance of water (1.5 MRayl). Therefore, the parylene acoustic lens helps to have efficient acoustic transmission from PZT to water. In this way, the acoustic waves are transmitted into liquid through the parylene, but the waves are reflected by the air reflector as illustrated in Figure 2.1a. The ejector with parylene reflectors has no need to control the thickness of parylene layer or the air gap distance [1]. A reservoir and a microfluidic channel are fabricated using silicon wafers in order to get continuous ejection (Figure 2.1 (c)). 34 Figure 2.1: (a) and (c) Cross-sectional views of the ejector with the acoustic lens and the packaged ejector, respectively; (b) Photo of ejector lens showing the regular rings with 2 π phase-shift between adjacent rings (on the left) and the phase varied rings with 2.5 π phase-shift between adjacent rings (on the right). 2.1.1 Phase-varied lens design for directional droplet ejections The parylene lens is patterned into phase-varied rings in order to produce the electrically adjustable oblique ejections of nanoliter droplets. The phase-varied lens consists of two parts: the left half circular lens with a regular Fresnel half-wave lens pattern, and the right half circular lens with a ‘phase-varied’ lens pattern. The cross- sectional view of the ejector and a picture of a fabricated ejector lens are shown in Figure 2.1(a) and Figure 2.1(b), respectively. The left part of the lens is patterned into Fresnel half-wavelength lens, through which acoustic waves are focused at the liquid-air interface. With the Fresnel half- wavelength lens, the path-length difference between the adjacent source rings is a 35 wavelength λ l (i.e., acoustic-wave velocity in liquid divided by the resonant frequency f r of the PZT transducer) and the phase difference is 2 π. The ring dimensions are designed to meet the path-length differences from the edges according to [2]: R k o − R k i = λ l 2 = v 2 f r (2.1) R k i − R k −1 o = λ l 2 = v 2 f r (2.2) where and are the path-lengths, from the inner and outer edges of the source rings to the focal point, respectively; λ l , v, and f r are acoustic wavelength in liquid, acoustic velocity and the resonant frequency of PZT sheet, respectively. As shown in Figure 2.2, R 1 ( ) and R 3 ( ) are the path-lengths from the inner edges of the source rings to the focal point, and R2 ( ) and R4 ( ) are the path-lengths from the outer edges of the source rings to the focal point. Figure 2.2: Effect of propagation path difference on wave interference with Fresnel ring pattern 36 Based on Eq. (2.1) and Eq. (2.2), the radiuses of the ring edges in the left part of the lens are determined by: r k i = (2k − 1) λ l 2 × ( (2k − 1) λ l 2 + 2F) (2.3) r k o = k λ l × (k λ l + 2F) (2.4) where F is focus length, while and are the radiuses of the inner edge and outer edge of the kth source ring, respectively, with k=1,2,3 …. Being different from the left part of the lens, the right part of the lens is patterned into a “phase-varied” lens, through which destructive interferences of acoustic waves are introduced. In the “phase-varied” lens, the path-length difference between the adjacent source rings differs from integer multiple of wavelength λ l , and consequently, the path-length differences between the edges should satisfy: R k o − R k i = L 1 ≠ λ l 2 (2.5) R k i − R k −1 o = L 2 ≠ λ l 2 (2.6) L 1 + L 2 ≠ λ l (2.7) where L 1 is the path-length difference between the edges of the source rings, while L 2 is the difference between the edges of the air gaps. Then the radiuses of the edges of the source rings in the right part of the lens are determined by r k i = b× (b+ 2F) (2.8) 37 r k o = a× (a+ 2F) (2.9) a= k(L 1 + L 2 ) (2.10) b= a − L 1 (2.11) The constructive and destructive interferences, from left and right parts of lens respectively, produce unbalanced body force at the focal point in the lateral direction. Therefore the lens causes droplet ejections in a direction oblique to the normal of the liquid surface. Also, in the half and half ‘phase-varied’ lens design, the ejection direction changes with the operating frequency f w , since the phase shifts through the lens vary according to the frequency of the driving signal f w . Figure 2.3: Simulation results of the particle displacement vector and ejection direction (red arrow) for the phase-varied ejector working at different operating frequencies (phase-shift of the phase-varied rings is 2.5 π) 38 The phase differences between the two adjacent rings in the two parts are Δϕ 1 =ϕ l ×f w /f r and Δϕ 2 =ϕ r ×f w /f r, , where ϕ l =2π and ϕ r =2π×(L 1 +L 2 )/ λ l . With the proper design of the phase-shift ring pattern, the ejection direction changes from right to left monotonically as the operating frequency increases. We chose ϕ r as 2.5π (L 1 = λ l /2, L 2 = 3λ l /4) and 3.25 π (L 1 = 3λ l /2, L 2 = 7λ l /8) based on our simulation results (Figure 2.3). 2.1.2 Dual-frequency half and half lens design Also, the dual-frequency half and half ring pattern ejectors are designed. According to the equations of radii of rings: r n 2 +F 2 −F= n λ l 4 (2.12) where n = 1, 3, 5…. Two different frequencies and are chose to design left and right lenses. We use in left half lens design and in right half rings design to design the dimensions of the rings in left and right lenses, respectively. In this way, when the working frequency , the phase difference between two adjacent rings in left half part of ejector is , while that phase difference in right half of ejectors is . Therefore, the acoustic waves from left half part of the lens form constructive interference at the focal spot, while the acoustic waves from the right part form destructive interference at the focal spot. The unbalanced body force in the lateral direction generates directional droplet 39 ejections, which are pointing to the right side. On the other hand, when the operating frequency is , the phase differences between two adjacent rings in left and right lens are and , respectively. The unbalanced body generates directional droplet ejections pointing to the left side. Therefore, it is expected to observe opposite droplet ejection directions at frequencies and , and to have changing directions at the frequency between and . 2.2 Device fabrication and testing setup Figure 2.4 shows the brief fabrication steps for the ejector. The acoustic transducer is made of the 127-µm-thick of the PZT sheet (PSI-5A4E, Piezo Systems). Nickel layers on the both sides of the PZT sheet are firstly patterned into circular electrodes. Then, we spin-coat 3-µm-thick photoresist (AZ5214) on the electrodes, and pattern the photoresist into “phase-varied” lens or “dual-frequency” lens. The patterned photoresist is used as a sacrificial layer to form the source rings and air gaps in the following steps. A 4.5-µm-thick parylene D is deposited as the lens material, followed by RIE etching release holes on the parylene and removing the photoresist with acetone. After the gaps are dry and filled with air, additional 3.5- µm-thick payrlene D is deposited to cover the release holes and form the air cavity underneath the parylene lens. To fabricate reservoirs and micro-channels, two silicon wafers are wet etched by potassium hydroxide (KOH) solution to form liquid chambers and microfluidic channel in between. The final step is to bond the PZT 40 sheet to the two silicon wafers containing the micro-channel and the reservoir by epoxy. Figure 2.4: Fabrication process of the ejectors with electrically adjustable droplet direction To actuate the fabricated ejectors, a pulsed sinusoidal signal is applied. A pulse generator triggers a RF signal generator to create a sinusoidal pulsed signal with 60 V pp and a minimum pulse width of 22 µs. With a typical pulse repetition frequency (PRF) of 60 Hz, the frequency of sinusoidal driving signal is varied from 16 MHz to 19 MHz. The driving signal is amplified through an RF power amplifier, before being applied to an ejector. The ejection process is stroboscopically recorded with a 41 CCD camera (SONY SSC-DC54A) fixed at the end of a microscope with a flashing LED, which is controlled by a delay time controller. The testing setup is illustrated in Figure 2.5. Figure 2.5: Measurement setup for ejection: the droplet ejection is stroboscopically captured with a delay time controller (to control the observation time) and an LED Figure 2.6 shows pictures of various phase-shift ejectors’ lens patterns. In the figure, the left half of the lens in each photo is a regular ring pattern with 2 π phase- shift, while the right half of the lens is phase-varied ring pattern with different phase- shifts. Different electrode patterns are also investigated. For the same lens design, round-shape electrodes and circular electrodes without center are fabricated and 42 tested. The white color part of the ring is air cavity under the parylene lens. The rings designed with larger phase-shift have wider rings. Figure 2.7 shows pictures of fabricated ejectors with “dual-frequency” half and half lenses. In the figure, all electrodes are patterned into a circular shape without a center. The left half and the right half of the lens are designed for different working frequencies and , respectively. Figure 2.6: Photos of various phase-shift-ring lens and electrode shapes 43 Figure 2.7: Pictures of dual-frequency ejectors ring pattern: (a) left half ring for 17 MHz, right half ring for 19 MHz; (b) left half ring for 18 MHz, right half ring for 19 MHz; (c) left half ring for 18 MHz, right half ring for 17 MHz; (d) left half ring for 17.5 MHz, right half ring for 18 MHz. 2.3 Experimental results The ejection directions of the phase-varied ejectors and the dual-frequency ejectors are characterized as a function of the frequency of driving signals. To measure the direction of ejections, a droplet position at the moment when the droplet is just separated from the chamber is recorded as a reference, and a droplet position after a 1000 μs time delay is also recorded to calculate the ejection angle (Figure 2.8 and Figure 2.9). As seen in Figure 2.8, the fabricated phase-varied ejector is capable of ejecting liquid droplets at multiple angles, which depend on the applied working RF frequency. It demonstrates electrical controllability of the ejection direction over a wide range. With 60 V pp voltage level of the pulsed driving signal, pulse width of 22 µs and PRF of 60 Hz, the ejection angle is electrically adjustable from -15.1° to 6.4°, 44 as the working RF frequency is varied from 16.34 MHz to 17.24 MHz. The size of the ejected droplet is about 70 µm, which is similar to the acoustic wavelength. With a proper design of the phase-shift ring pattern, the ejection direction changes from right to left monotonically as the operating frequency is increased from 16 MHz to 18.5 MHz (for an ejector with 2.5 π phase-shift rings), according to the experimental and simulation results (Figure 2.10). Based on the simulation results, it is noted that the angle range is from -38° to 10° as the frequency increases. Up to 17.5 MHz, the angle change is small, because the left part of the lens with a regular Fresnel half- wave pattern is the main contributor to the ejection direction at the operating frequency close to the resonant frequency of the PZT sheet. Figure 2.8: Experimental results with the phase-varied ejector (phase-shift of the phase-varied rings is 2.5 π) 45 Figure 2.9: Experimental results with the dual-frequency ejector (the operating frequencies of left and right half of lens are 18 MHz and 19 MHz, respectively.) For dual-frequency ejectors, the ejection angles are not monotonically increasing as the operating frequency is above 19 MHz (Figure 2.11). According to the measurement (Table 2-1), the widest range of directional angle -23.4° ~ 7.5° is obtained with an ejector having 18 MHz – 19 MHz dual-frequency pattern. The trend of direction changes along with the operating frequency increase is not monotonic. Figure 2.10: Simulation and experimental results for the phase-varied directional ejector with half regular rings and half 2.5 π phase-varied rings. 46 Figure 2.11: Experimental results for the dual-frequency directional ejector with left half rings at 18 MHz and right half rings at 19 MHz. For “phase-varied” lens design, the time evolution of the droplet formations and directional ejections at various operating frequencies are shown in Figure 2.12. The droplet separation time is less than 200 µs, while the liquid-surface relaxation time is less than 300 µs. The clear droplet photos captured with an optical strobing at different observation times show consistent and stable directional droplet ejections at various operating frequencies. Based on the measurement results, it is noted that the speed of droplet ejection becomes less, as the ejection direction is more oblique. It is because more destructive interferences are involved in more oblique directional ejection. The initial droplet velocity at the droplet separation moment ranges from 1.3 m/s to 2.7 m/s. 47 Figure 2.12: Time evolution of the droplet formations and directional ejections at different operating frequencies (f w ) and observation times. We have fabricated and tested various ejectors with different ring patterns of the phase-varied lenses, dual-frequency lenses and electrode shapes (Figure 2.6 and Figure 2.7). According to the measurement results (Table 2-1 and Table 2-2), the widest range of directional angle (-30° ~ 35°) is obtained with an ejector having 3.25π phase-shift rings pattern. Table 2-1: Ejection directions and corresponding operating frequencies for different phase-shifts (P-S) and different electrodes, i.e., round electrode (RE) and RE without center part (RE w/o center). P-S = 2.5 π RE P-S = 2.5 π RE w/o center P-S = 3.25 π RE P-S = 3.25 π RE w/o center (MHZ) (MHZ) (MHZ) (MHZ) 16.34 -23.1 16.34 -15.1 16.78 -30.2 16.67 -30.0 17.86 -19.4 16.78 -6.1 17.73 -2.6 16.95 -21.0 18.18 3.6 17.00 4.5 17.91 10.6 17.54 4.9 17.24 6.4 19.08 35.0 17.86 16.0 48 Table 2-2: Ejection directions and corresponding operating frequencies for different dual-frequency ejector designs 17 – 19 MHz 18 – 19 MHz 17 – 18 MHz 17.5 – 18 MHz (MHZ) (MHZ) (MHZ) (MHZ) 17.30 7.3 16.78 -23.4 17.30 6.7 17.50 10.05 17.89 -9.2 17.14 -19.6 18.50 3.8 18.16 -8.3 18.69 12.8 18.52 16.1 19.76 -4.5 18.99 3.1 19.88 -26.6 19.43 7.5 20.40 0.4 20.03 1.9 The experimental results show that, for more oblique ejection, although the droplet velocity is less, the ejection stability is better. This is because, when the ejection direction is close to the perpendicular direction, both the right and left parts contribute large and similar power to the ejection direction. However, the displacement amplitudes contributed by these two parts are out of phase, resulting in more unstable ejection. At one moment within a sinusoidal cycle, the left part contributes more power than the right part; at another moment within the same cycle, the right part contributes more than the left part. Also, with the current design, it is found that the control of the ejection direction is not continuous over the wide range of the direction, and the ejections are stable at discrete frequencies. According to Table 2-1, the droplet ejections are stable at 3-4 discrete frequencies. The working frequencies mainly depend on the lens pattern design and the liquid level in the ejector. 49 2.4 Summary In the chapter, a multi-directional acoustic ejector with “phase-varied” lens and “dual-frequency” lens have been fabricated and characterized. Electrical control of the ejection direction has been demonstrated by varying the RF frequency of the sinusoidal driving signal. Stable and consistent ejections of uniform droplets (70 - 80 µm in diameter) have been obtained. With a proper design of the phase-varied ring pattern, the ejection direction changes from right to left monotonically as the operating RF frequency is increased from 16 to 18.5 MHz (for an ejector with 2.5 π phase-shift rings). The widest range of directional angle is from -30° to 35°, which is obtained with the ejector having 3.25 π phase-shift rings pattern. Reference [1] C.-Y. Lee, et al., "Acoustic Ejector with Novel Lens Employing Air- Reflectors," in Micro Electro Mechanical Systems, 2006. MEMS 2006 Istanbul. 19th IEEE International Conference on, 2006, pp. 170-173. [2] D. Huang and E. S. Kim, "Micromachined acoustic-wave liquid ejector," Journal of Microelectromechanical Systems, vol. 10, pp. 442-449, Sep 2001. 50 Chapter 3 Localized Cytolysis System using Ultrasonic Self Focusing Acoustic Transducers 3.1 Localized cell lysis by ultrasonic transducers Cell lysis involves a process of rupturing cell membrane, allowing intracellular components out of the cell, and ultimately leading to cell death. Localized cell lysis with several microns precision will provide unprecedented opportunities in cancer diagnosis and therapeutics. Current microfluidic systems for cell lysis use thermoelectric lysis [1], electrical lysis [2], chemical lysis [3]. But these cell lysis platform do not have location selectivity. Also, these methods require physical or chemical contact with cells for cell lysis, therefore may introduce contaminations during the cytolysis process. Conventional ultrasonic lysis or sonoporation uses strong ultrasonic waves, and relies on cavitating bubbles as focusing agents to concentrate energy and stresses [4- 6]. These approaches are hard to control, because cavitation bubbles will undergo a fast collapse and generate heat and stress. Self Focusing Acoustic Transducer (SFAT) is a good candidate to realize localized cell lysis with hundred micron-level spatial resolutions. It is because SFAT 51 is able to generate high-frequency acoustic waves with a good ultrasound focusing. In previous study with an SFAT [7], high acoustic intensity has been applied to generate the cavitation bubbles at the focal spot for cell lysis. With low acoustic intensity at the focal spot, violent cavitating bubbles can be avoided [8]. In this chapter, design and fabrication of SFATs working at fundamental (17.3 MHz) and 3 rd harmonic resonant frequencies (51.9 MHz) are presented. Then, the minimum acoustic intensities for prostate and breast 2D monolayer cells are investigated. Three prostate cell lines (22RV1, RWPE-1, RWPE-2) and two breast cell lines (MCF-7, MCF-10) are tested with the focused ultrasound from SFATs with various acoustic intensities, in order to characterize cell’s sensitivity to ultrasound and Acoustic Intensity Threshold (AIT) for causing localized cytolysis. 3.2 Design and fabrication of SFAT for cell lysis The SFAT-based cytolysis device consists of a Lead Zirconate Titanate (PZT) transducer with an air-reflector acoustic lens, an acoustic chamber and a cell culture chamber, as shown in Figure 3.1a. Since the thickness of the PZT is 127 µm, the PZT’s fundamental resonant frequency f r for the thickness mode is 17.3 MHz. Parylene lens with air reflectors on the PZT transducer focuses acoustic waves at the focal spot [9]. The rings’ dimensions are designed to meet the path-length differences from the edges according to [10]: 52 R k o − R k i = λ l 2 = v 2 f r (3.1) R k i − R k −1 o = λ l 2 = v 2 f r (3.2) where and are the path-lengths, from the inner and outer edges of the source rings to the focal point, respectively; λ l , v, and f r are acoustic wavelength in liquid, acoustic velocity and the resonant frequency of PZT sheet, respectively. The radiuses of the ring edges of the lens are determined by r k i = (2k − 1) λ l 2 × ( (2k − 1) λ l 2 + 2F) (3.3) r k o = k λ l × (k λ l + 2F) (3.4) where F is focus length, while and are the radiuses of the inner edge and outer edge of the kth (k=1,2,3 …. ) source ring, respectively. The height of the acoustic chamber is 800 µm, which is equal to the focal length F of the acoustic lens. The acoustic chamber is filled with ultrasonic gel (AQUASONIC® 100) for acoustic impedance matching with the PZT transducer. Benign and malignant cells were cultured on the parylene-D membrane in a form of 2D monolayer cells in a cell culture chamber. The silicon nitride membrane is a mechanical support for the parylene layer on it. The transparency of the parylene and silicon nitride (Si x N y ) membranes facilitates the following fluorescence assays of cytolysis with a fluorescent microscope. 53 For cytolysis, the PZT transducer generates acoustic waves in response to applied electric power. Fresnel acoustic lens focuses acoustic waves by making the waves arrive at the targeted focal spot in phase, therefore creates constructive acoustic waves interference for ultrasound focusing. Thus, at the focus spot, the highly concentrated acoustic-radiation exerts a high pressure to damage the cell membrane. The cytolysis caused the SFAT is a localized effect that cells located out of the focal spot are not affected by the focused ultrasound and cause no cytolysis effect. Figure 3.1: (a) Cross-sectional view of the cytolysis device with an acoustic Fresnel lens on a PZT transducer, a Si acoustic chamber, and a Si cell culture chamber; (b) Schematic of the observation method of cytolysis The fabrication procedures of SFATs and cell culturing chambers are listed as following. The transducer is built with a 127-µm-thick PZT sheet (PSI-5A4E, Piezo Systems) with patterned circular nickel electrodes on both sides. 54 Figure 3.2: Fabrication process of the cell lysis device with a PZT transducer with acoustic lens, an ultrasound chamber and a Si cell culture chamber The acoustic lens is fabricated on the patterned electrodes by: 1) spin-coating and patterning 3-µm-thick sacrificial photoresist on the PZT substrate; 2) depositing a 4.5-µm-thick parylene D membrane for the lens material, followed by patterning of the parylene for etch release holes with oxygen plasma in RIE (Reactive Ion Etching) system; 3) removing the sacrificial layer by soaking the device in acetone for 5 hours and then allowing the acetone to evaporate through the air gaps naturally; and then 4) depositing an additional 3.5-µm-thick parylene D to cover the release holes. To 55 fabricate the 800-µm-thick acoustic chamber, two 400-µm-thick silicon wafers are wet etched with potassium hydroxide (KOH) solution and bonded to the PZT sheet. Cell culture chambers are fabricated in a 400-µm-thick silicon wafer with 1-µm- thick Si x N y on both sides by: 1) patterning the Si x N y on one side of the wafer, and forming the chamber structure with Si x N y diaphragm using KOH wet etching; 2) dicing the wafer into small chambers, and depositing a 3 µm thick parylene layer on the chamber chips to form the parylene/Si x N y /parylene membrane layers. Finally, the parylene surface is treated by RIE oxygen plasma for 10 seconds to make the parylene hydrophilic for facilitating cell attachment (Figure 3.2). The dimension of the cell culture chamber is 6×14×0.4 mm 3 . For the experiment, fabricated cell culturing chambers with cultured cells are placed on the SFAT for the localized cytolysis. 3.3 Modeling of SFAT’s focus and bilayer cellular membrane model Cell lysis modeling consists of two parts: (1) the acoustic pressure distribution simulation of SFAT on the focal plane; and (2) the mechanical response of the bilayer cell membrane under the acoustic pressure generated by the SFAT. In the first part of cell lysis modeling, a MATLAB program is developed to calculate dimensions of acoustic lens according to the operating frequency and focal length based on Eq. (3.3) and Eq. (3.4), and to simulate the distribution of acoustic 56 intensity and focal spot location, in 3-D spaces. The acoustic pressure P A generated by a SFAT at an estimation point is: P A (x,y,z)= ω 2 ρ 0 φ(x,y,z) (3.5) with ω and being the angular frequency of acoustic waves and mass density of liquid medium, respectively. The potential at any point in anisotropic liquid medium is in the form of a solution of Green’s function [11]: φ(x,y,z)= − 1 2 π u 0 e − αR −jkR R s ∫ ds ' (3.6) where the integral surface s is the active area of the Fresnel lens over the electrode with u 0 being the vibration amplitude at the transducer surface; R being the distance between the sound source point on the transducer surface and the estimation point; k being the wavenumber of acoustic waves; and α being the attenuation coefficient in the transmission media (Figure 3.3). The vibration amplitude of the transducer u 0 =S 3z ×d 0 , where S 3z is the strain due to the electric field and the stress applied on the transducer’s surface, while d 0 is the thickness of the transducer. 57 Figure 3.3: Coordinate system for the equations leading to the acoustic pressure distribution produced by SFAT. Since the stress T applied on the transducer is the acoustic pressure at the plane z = 0, it is able to derive the following equations: S 3z = s 33 T +d 33 E (3.7) T = −2 π f ×S 3z ×d pzt ×Z media (3.8) where is elastic compliance; d 33 is piezoelectric coefficient of the transducer; E is the electric field applied to the transducer; f is the operating frequency of the transducer; and Z media is the acoustic impedance of the liquid media. Replacing the stress T in Eq. (3.7) with Eq. (3.8), we solve S 3z in the term of electric filed E, and thus u 0 : S 3z = d 33 E 1+2 π f ×s 33 ×d PZT ×Z Media (3.9) 58 As the second part of ultrasonic cell lysis modeling, non-thermal interactions of ultrasound pressure and biological tissue are studied and simulated using a Matlab program. The dynamic responses of the cell plasma membrane in cancer cells and non-cancerous cells to ultrasound treatment from SFATs are explored. The interactions of ultrasound and cells can be analyzed with the bilayer cellular membrane model [12]. The advantage of this model is that it can explain the direct interactions between the acoustic pressure and the cell membrane over a wide range, from low-intensity (I< 100 mW/cm 2 , non-cavitational ultrasound intensity) to high- intensity (I > 100 mW/cm 2 , non-thermal but potentially cavitational ultrasound intensity). Figure 3.4: Schematic illustrating a bilayer cell membrane under uniform acoustic pressure 59 Bilayer sonophore (BLS) model is based on a piece of circular bilayer cell membrane with two monolayer leaflets, which is bound by a circular ring of transmembrane proteins (Figure 3.4). The diameter of the enclosed membrane is from 50 nm to 100 nm, which is much smaller than the focal area of the SFAT (100 μm in diameter). Thus, it is assumed that a uniform acoustic pressure P A is applied on the membrane. Based on the Rayleigh-Plesset equation, the deflection of the cell membrane H(t) is governed by the equation [12]: d 2 H dt 2 + 3 2R ( dH dt ) 2 = 1 ρ 1 R [P in +P ar −P 0 +P A sin ωt −P s (R) − 4 R dH dt ( 3 δ 0 μ s R + μ 1 )] (3.10) where R= (a 2 +H 2 )/2H is the radius of curvature of the upper leaflet; is the density of liquid media; is the gap pressure between two leaflets, which is determined by the shape of the membrane; is attraction or repulsion pressure between two leaflets; is the atmospheric pressure in the liquid media; is the acoustic pressure; is the pressure due to the circumferential tension in the membrane; is the dynamic viscosity of the liquid media; is the dynamic viscosity of the membrane; and is the initial thickness of the membrane. The equations for , , and are as follow [12]: P in =P 0 [1+ H 6 Δ (3+ H 2 a 2 )] −1 (3.11) P ar = 2 (a 2 +H 2 ) A r [( Δ h(r)+ Δ ) m − 0 a ∫ ( Δ h(r)+ Δ ) n ]rdr (3.12) 60 P s (R)= 2k s H 3 a 2 (a 2 +H 2 ) (3.13) h(r)= R 2 −r 2 −R+H (3.14) where is an initial gap between two leaflets; A r is the pressure when the gap between two leaflets is the initial gap; m is the exponent in the repulsion term; n is the exponent in the attraction term; and k s is the area compression modulus of the membrane. Specific values used in the simulation are shown in Table 3-1. The maximum deflection H of the cell membrane due to the acoustic pressure can be calculated by solving the differential equation Eq.(3.10). The area strain and surface tension of the cell membrane are calculated by the following equations: ε A = S −S 0 S 0 = ( H a ) 2 (3.15) T =k s ε A =k s ( H a ) 2 (3.16) Rupture tension strengths (or cell lysis tension strengths) for some bilayer lipid vesicles are measured through pipette suction, and reported to range from 3 mN/m to 10 mN/m [13]. For an area modulus k s = 0.03 N/m, the lysis strain is then in the range of 0.1 to 0.33. Table 3-1: Specific values used in the simulation Parameters Symbol Unit Value Attenuation Coefficient α m -1 0.324 [10] Characteristic Acoustic Impedance Z media Pa×s/m 150×10 6 [11] 61 (water) Density of the liquid media (water) ρ 1 kg/m 3 1056 Dynamic viscosity of the liquid media (water) μ 1 Pa/s 10 -3 Dynamic viscosity of the membrane μ s Pa/s 0.05 [12] Initial gap between two leaflets Δ m 1.4×10 -9 [12] Pressure when the gap between two leaflets is the initial gap A r Pa 10 5 The exponent in the repulsion term m 5 [12] The exponent in the attraction term n 3.3 [12] Area compression modulus k s N/m 0.03 Static pressure in the water P 0 Pa 10 5 Based on the cell lysis modeling, the relative acoustic intensity distribution in the medium and the cell membrane strain caused by the acoustic pressure are calculated. As shown in Figure 3.5, the focal plane is located 800- μm away from the transducer’s top surface, and the diameter of the focal spot is 75 μm. As expected, the diameter of the focal spot is close to the wavelength of acoustic waves in the water ( λ=c water /ƒ=1484/17.3×10 6 =85.7µm). Acoustic pressure at the focal spot is about 8 times higher than the pressure outside the focal area, which is similar to the results in [14]. The simulation results are compared with measurement results of acoustic intensity more than 2 mm away from the transducer using a hydrophone (PAHPM04/1, Precision Acoustics, UK), and the absolute value of simulated acoustic intensities are calibrated based on the measurements. This way, the acoustic intensity at the focal point is estimated. According to the results, when 1 W electric 62 power is applied to the SFAT, the acoustic intensity at the focal point is 16.67 W/cm 2 (Figure 3.5). Figure 3.5: Estimation of acoustic intensity at the focal point: (a) the simulation of acoustic pressure distribution along one axis (say, x-axis) on the focal plane when the voltage applied to SFAT is 10 volts; (b) the simulation of the relative acoustic pressure distribution on the cross-sectional x-z plane, when 10 volts is applied to the SFAT; and (c) simulated area strain of cells membrane generated by focused acoustic pressure (acoustic intensity I ac =1.944W/cm 2 ). The value of area strain is set to k s =0.03, the diameter of area of the free bilayer membrane confined by proteins in the membrane is 2a=25nm. The typical size of a cell is 10 - 20 micrometers in diameter, which is about 1/8 size of the focal area. Therefore, cells located within the focal area are considered to 63 undergo uniform acoustic pressure when the SFAT device is activated. Figure 3.5a and Figure 3.5c show the simulated acoustic pressure on the focal plane (without considering the reflection of the waves from the parylene membrane). The simulation takes into account of the acoustic effect of cavity resonance and cell absorption. It has had no problem with grid convergence, since it is not a finite element modeling. The reflection from the parylene membrane is negligible because the acoustic wavelength in paryelene ( ~ 0.12 mm) is much larger than the thickness of the parylene membrane (6 μm). Within the focal area, the acoustic intensity is much higher than that outside the focal area. Figure 3.5c shows the cell membrane’s strain caused by the acoustic pressure distribution generated by SFAT (Figure 3.5a). The strain calculation is based on the bilayer sonophore (BLS) model [12]. Comparing the results shown in Figure 3.5a and Figure 3.5c, it is noted that cells located within the focal area experience much higher strains caused by the acoustic pressure than cells outside the focal area. During ultrasonic irradiation, bilayer lipids of cell membrane are compressed to and pulled away from each other accordingly to the oscillation of acoustic pressure. In this way, the acoustic pressure exerts area strains on the cell membrane, and eventually causes cell membrane to rupture, as the cell membranes tension reaches 0.003-0.01 N·m -1 and the area strain is 0.1-0.3 [13]. 64 3.4 Cell culture and tumor spheroids formation MCF-7, A375 and Detroit 551 cells are maintained in DMEM (GIBCO) supplemented with 10% (v/v) fetal bovine serum (FBS) (Irvine Scientific), 1 mM sodium pyruvate, nonessential amino acids, and 100 μg/mL of penicillin/streptomycin (Irvine Scientific) and Normocin (InvivoGen). MCF-10A cells are maintained in DMEM with 10% (v/v) FBS, penicillin/streptomycin (100 μg/ml), nonessential amino acids, EGF (20 ng/ml), Hydrocrotisone (0.5 μg/ml), Insulin (10 μg/ml) and Cholera toxin (100 ng/ml). 22RV1 cells are maintained in RPMI with 10% (v/v) FBS and penicillin/streptomycin (100 μg/ml). RWPE1 cells are maintained in Keratinocyte-SFM media from Invitrogen. All cells are incubated in a humidified chamber supplemented with 5% CO 2 at 37°C. To prepare the cell samples, cell culture chambers are placed in wells of a 24- well cell culture plate, and sterilized with UV irradiation for 30 minutes prior to cell seeding. After the sterilization, 5×10 4 cells are seeded on the parylene surface of the cell culture chamber and maintained for the cell growth in a CO 2 incubator at 37 o C to reach 80% to 90% confluence. For cytoskeleton alteration, MCF-10A cells are treated with 5µM of nocodazole, 5µM of blebbistatin, or 50µM of Roc inhibitor Y-27632 for 1 hour prior to ultrasound irradiation. 65 3.5 Testing setup and experimental procedure For an acoustic pressure measurement, a calibrated hydrophone (PAHPM04/1, Precision Acoustics, UK) is used to characterize the SFAT’s focus in a water tank. The size of the hydrophone’s sensing area is 40 μm in diameter, and the frequency range of the hydrophone is from 1 MHz to 30 MHz. In the measurement, the hydrophone is positioned in the water tank and moved by a three-axis positioning motor controlled by a computer. The acoustic pressures at multiple spots will be measured in order to profile the acoustic pressure distribution generated by the designed SFAT. Cells are seeded and grown on the parylene membrane of the cell culturing chamber as shown in Figure 3.1a. Before the cell lysis experiment, the cells are maintained in an incubator as described above. Then, the experiments are performed outside the incubator at room temperature in ambient atmosphere. The cell culture chambers are taken out of the incubator, and placed on the SFATs for cytolysis experiments. To conduct the cytolysis experiment, multiple pulses of sinusoidal signal are applied to the transducer to produce pulses of acoustic waves. A pulse generator in conjunction with a sinusoidal function generator is used to create a pulsed sinusoidal signal with 0.05 - 5 V pp , pulse width of 1 s, and a duty cycle of 50%. The pulsed signal is amplified through an RF power amplifier. The amplified voltage is ranging between 10 and 300 V, and it determines the power applied to the transducer. Cells are exposed to acoustic irradiation for 1 minute. Then, the cells are 66 treated with SYTOX fluorescent dye and fixed for 10-30 minutes following the acoustic exposure (Figure 3.6). The control groups are treated in the same procedures and timings except the step of acoustic irradiation. It is showed that the cells without the acoustic irradiation are not taking SYTOX fluorescent dye. The temperature variation of cytolysis process is measured using 2-D IR thermal imager (FLUKE 62 mini IR thermometer). The temperature of culture media in the cell culture chamber is recorded every 30 seconds. 3.5.1 Fluorescent microscopy After the acoustic irradiation, it takes 10 – 30 minutes to treat the cells with SYTOX staining, fix the cells and do the fluorescent microscopy. The bio-effects of acoustic energy are examined by SYTOX Dead Cell staining (Invitrogen) and fluorescence microscopy, as shown in step 3 in Figure 3.6. SYTOX dead cell staining is taken up by damaged or dying cells, but it is not able to cross the plasma membrane of live cells. The cells are first incubated with SYTOX solution for 10 min and then fixed in 4% formaldehyde solution for 20 min prior to fluorescence microscopy examinations. A double staining kit, such as Calcein-AM/ Propidium iodide (PI) or a combination of SYTOX and CellMask Plasma Membrane stain, would have enhanced the contrast of the bright-field images in Figure 3.7, as it produces red fluorescence for cell membrane and green fluorescence for dead cells (as shown in Figure 3.9). However, it is observed that the red color fluorescence of the double 67 staining interfered with the auto-fluorescence of the parylene D membrane. The autofluorescence of the parylene membrane is around 570 nm (red fluorescence) [15], and does not interfere with the SYTOX green fluorescence (excitation: 504 nm; emission 523 nm). Figure 3.6: Three steps of fast screening based on a SFAT and the fluorescent stained cytolysis assay For cytoskeleton (F-actin) staining, 5×10 2 of MCF-7, MCF-10A, RWPE1, 22RV1, A375 and Detorit 551 cells are seeded in an 8-well glass chamber slides (Nunc Lab-Tek II) one day before the staining. The cells are washed twice in pre- warmed PBS and then fixed in 3.7% formaldehyde solution in PBS for 10 minutes at room temperature. After the washing, the cells are permeabilized with 0.1% Triton X-100 in PBS for 10 min. The cells are subsequently blocked with 1% BSA in PBS. F-Actin staining is carried out by incubating cells with 165 nM of Alexa Fluor 488- phalloidin for 30 min at room temperature. Then, the cells are washed three times in PBS and finally mounted in ProLong Gold reagent (Invitrogen). The staining results 68 are examined by fluorescence microscopy (AXIO; Zeiss). The actin distribution intensity is measured by phalloidin labeling and subsequently quantified by image densitometer software (ImageJ). The stress fiber distribution in a cell is quantified by checking the fluorescent intensity along a line across the nuclei and over the cell. It is based on the method described in [16, 17]. The fluorescence intensity is quantified with Image J software, and then calculated and plotted by using Microsoft Excel software. A representative trace for each cell line is shown in Figure 3.10. 3.5.2 Semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) Semi-quantitative RT-PCR is also used to monitor the cytolytic effect of acoustic irradiations. Conditioned media are collected from the cell culture chamber before and after one time sonication with SFAT. RNA is extracted and reverse transcribed to complementary DNAs (cDNAs). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is amplified by polymerase chain reaction (PCR). The products are visualized by SYBR© green staining after DNA gel electrophoresis, and the DNA amount is quantified by ImageJ software. For each acoustic intensity level, a ratio of the GAPDH message obtained after the sonication over that before the sonication is presented. 69 3.6 Experimental results 3.6.1 Highly focused and localized ultrasonic cytolytic effect by the SFAT 22RV1 prostate cancer cells are grown as a monolayer in the cell culture chamber, and the cells are sonicated by the SFAT operating at 17.3MHz with 1 second pulse width and 50% duty cycle for 3-minute. The peak acoustic intensity at the focal spot is set from 0.117 and 8.880 W/cm 2 (corresponding to applied electric power of 1.3 to 530 mW). SYTOX fluorescent dye staining is performed basically according to manufacturer’s instructions. Following acoustic irradiations or a control treatment, 22RV1 cells are washed once with PBS and incubated with 1 µM SYTOX green (dissolved in culture media) for 10 minutes. Then the cell sample is fixed in 4% formaldehyde solution at room temperature for 20 minutes before microscopy examinations. Since SYTOX green is diluted in conditional cell medium (with CO 2 ), cells are not dead during the staining process. Several control experiments without acoustic irradiation are performed, and no localized death of cells or SYTOX green- positive cells are found during the staining procedure. For all experimental conditions, and on each experimental day, a negative control experiment is performed, in which cells are handled in a similar fashion outside of the incubator without activation of the device. In all cases, SYTOX uptake or increase in free GAPDH mRNA is not observed with the control treatments (Figure 3.7b and Figure 3.7c). 70 Figure 3.7: Cytolysis with electric power applied to the device at 17.3 MHz: (a) The SFAT-sonicated 22RV1 cells were labelled by SYTOX dead cell dye and examined under a fluorescent microscope. The SFAT was operated at 17.3 MHz with applied acoustic intensity, at the focal spot, of (from top to bottom) 0.117W/cm 2 , 0.150 W/cm 2 , 2.217 W/cm 2 , and 8.88 W/cm 2 . The cytolysis areas were around 100 µm. (b) SYTOX fluorescent intensity within the focal area versus acoustic intensity. (c) The RNAs released from the cells were collected before and after the ultrasound irradiation. After the reverse transcription and PCR amplification of the collected RNAs, one housekeeping gene GAPDH was separated by electrophoresis and quantified by ImageJ software, and the measured amount of change due to the sonication is shown. The effects of sonication at different acoustic intensities is demonstrated in Figure 3.7a. The damaged cells within the focal area are labeled by green colored SYTOX fluorescence dye. The left column shows effect of sonication with SYTOX green imaged with confocal fluorescent microscopy; the middle column shows the same field with bright field phase contrast imaging; and the right column shows a 71 merged image. Although the experiments show a dose-dependent increase in SYTOX green uptake at the focal spot as the acoustic intensity increases, the number of damaged cells is not linearly correlated with the acoustic intensity applied on the cells within the focal area (Figure 3.7c). The bright field of the microscopy shows the cells outside the focal area are not compromised at any power level. The area of cell lysis is in a circular shape with diameter of 100 +/- 20 µm. The affected area is close to the focal area of the SFAT predicted in the simulation, which is 85 µm in diameter. The experimental results validated the simulation results that the size of focused area mainly depends on the frequency of ultrasonic waves. Therefore, the high spatial precision of 100 µm for localized cell lysis can be achieved by using SFAT working at 18 MHz. It is also observed that cells were detached from the parylene membrane surface within the focal area as very high acoustic intensities (>10 W/cm 2 ) applied. This result is similar to our pervious study [7]. However, in the experiments with acoustic intensity from 0.1 to 9 W/cm 2 , inertia cavitation phenomenon such as bubble generations and collapses are not observed during the ultrasound irradiation. It is noted that the focused acoustic waves alone can exert enough radiation forces to damage the cell membrane located within the focal area. In addition, the cytolytic effects are confirmed by RT-PCR again, in which the results show a dose-dependent increase of the release of the GAPDH mRNA in the collected solutions after the sonication (Figure 3.7c). The results suggest that the prostate malignant cancer cells 22RV1 can be easily damaged by ultrasound with the acoustic intensity as low as 0.15 W/cm 2 . The minimum acoustic intensity for lysing 72 cells is defined as the acoustic intensity threshold (AIT). The experimental results show that the cells are lysed, when 9 mW electrical power is applied to the device. The acoustic intensity at the focal spot is calculated to be about 0.150 W/cm 2 . Therefore, the acoustic intensity threshold (AIT) of cell lysis for the prostate cell line of 22RV1 is determined to be 0.15 W/cm 2 . Figure 3.8: Measured size of the cytolysis area caused by various acoustic intensities The measured size of cell lysis area is shown in Figure 3.8. With various electric power applied, the typical lysed area is from about 2.28×10 -9 m 2 to 3.922×10 -9 m 2 , which is not increased proportionally to the applied power (Figure 3.8). The SFAT based on a transducer generating acoustic waves at 51.9 MHz (3 rd harmonic resonant frequency of 127 µm-thich PZT) are also fabricated and tested. In Figure 3.9, it is compared the sizes of the cell lysis areas produced by the devices working at the fundamental (Figure 3.9a) and 3 rd harmonic frequencies (Figure 3.9 b 73 and c). In the case of the 3 rd harmonic devices, 0.10 and 0.18 W electric powers are applied, resulting in acoustic intensities of 4.02 and 7.156 W/cm 2 at the focal spot, respectively. The size of cell lysis area caused by the 3 rd harmonic device is around 50 µm in diameter, which is around 45% of the focal spot of the fundamental device. However, the acoustic intensity used for cell lysis by the 3 rd harmonic device is much higher than the acoustic intensity by the fundamental device. The reason for the smaller cell lysis area is: at the 3 rd harmonic frequency, the wavelength of the acoustic wave is one third of the wavelength at the fundamental frequency. Figure 3.9: Labeled lysed cells caused (a) by the device working at 17.3 MHz and 9 mW electric power applied (the labeled size is 80 × 120 µm2); (b) by the device working at 56 MHz and 0.10 W electric power applied (the labeled size is 55 × 80 µm 2 ); (c) by the device working at 56 MHz and 0.18 W electric power applied (the labeled size is 70 × 70 µm 2 ). 74 The temperature variation during the cell lysis is monitored and recorded by 2-D IR thermal imager (FLUKE 62 Mini IR thermometer). The temperature is measured at various points along the vertical direction out of the plane. The measured temperature increase over a whole plane in the cell culture chamber is less than 1°C, when an acoustic intensity level of 10 W/cm 2 is applied at the focal spot. 3.6.2 Considerations for other effects caused by ultrasound exposures Other than cell lysis effects caused by ultrasound exposures, “sonoporation” is usually used to describe the transient phenomenon of increased cell permeability induced by acoustic energy. In the presence of brief pulses of acoustic energy, a cell- impermeable agent or particles can pass through cell membrane into viable cells [4]. However, conditions must be carefully chosen to balance transduction efficiency against cytotoxicity typically observed with certain agents or with certain acoustic intensities. In our experiments, sonication occurs in the presence of normal growth media. No obvious cytopathic effect is observed immediately after acoustic irradiation. The uptake of the cell impairment SYTOX dye and the quantity of GAPDH mRNA release are used as markers for observing permanent cell damage after the ultrasound is applied. In addition, we employ a cell lysis model, to support our hypothesis that the high acoustic pressure and large area strain produced by the SFAT device are sufficient to cause damages of cell membrane. Therefore, the cytopathic effects observed in these experiments are considered as “cytolysis.” 75 Although high acoustic intensity can induce cavitation bubbles, which concentrate acoustic energy and produce great force and stress for rupturing cell membrane [4, 6], well-focused acoustic waves could also exert large body force at the focal spot and rupture cell membrane [8]. In the high power condition, both energy concentration mechanisms mentioned above exist [18], and cavitation bubbles can be observed. However, in the low power condition, it is implied that the wave focusing mechanism dominates, and cell lysis could be realized without the inertial cavitation and the bubble generation [19]. No bubble and cavitation effects are observed during the experiment when only 9 mW power is applied to the SFAT. Also, the experiments are repeated using degassed PBS as a control, and the experimental results are similar to the results obtained using non-degassed PBS. Given that the degassed PBS solution contains few air bubbles therefore hardly produces cavitation in the degassed solution, cell lysis observed in degassed PBS solutions indicates SFAT could produce localized cytolysis effects without cavitation. 3.6.3 Sensitivity to acoustic irradiation varies in benign and malignant cell lines Previous studies have shown that some cancer cells are more sensitive to high levels of acoustic energy at low frequencies delivered in continuous mode than their corresponding normal cells [20]. However, the effects of high frequencies and low energies delivered in pulsed mode have not been investigated. Thus, the cytolysis 76 effects in various cell lines representing malignant and normal tissues is studied in this section. When the monolayer cells are exposed to the focused ultrasound of 1 W/cm 2 , localized cytolysis is observed in malignant cells by showing SYTOX green uptake, while no effect is observed in various non-tumorigenic cell lines (Figure 3.10). Moreover, the AITs for several cell lines are investigated and summarized in Table 3-2. To determine the AITs of various cell lines, more than 10 independent replicates in cell lysis experiments have been conducted, and the most representative acoustic intensities calculated by SYTOX uptake are picked up as AITs (> 80% cell lysis rate). Overall, the range of AITs observed for cancer cells (MCF-7, 22RV1 and A375; AITs are from 0.12 to 0.3 W/cm 2 ) is considerably lower than that observed AITs for non-malignant variants of each cell line (RWPE1, MCF-10A and Detroit 551; AITs are from 0.6 to 2.21 W/cm 2 ). This general trend associating a lower AIT with a more malignant growth phenotype has been observed across other cancer and benign cell lines. Overall, our data supports a cancer-specific cytolysis effect from the acoustic beam produced by our current SFAT device operating at low intensities ( ≤ 0.3 W/cm 2 ). 77 Figure 3.10: (a) Superimposed bright field and fluorescence images showing selective cytolytic effects of SFAT at 1 W/cm 2 on benign (RWPE1, MCF-10A and Detroit 551) and malignant (22RV1, MCF-7 and A375) cell lines that represent prostate, breast and skin tissues. (b) The actin filament patterns in each cell line being visualized with green color by fluorescent conjugated phalloidin (these photos were taken before the ultrasound irradiation). (c) The actin distribution intensity in each cell line: the actin distribution intensity was quantified with fluorescence intensity using ImageJ software across the lines shown in the corresponding panels in the actin staining microscopy (b). Table 3-2: Acoustic Intensity Thresholds Prostate cell lines Breast cell lines Skin cell lines AIT W/cm 2 RWPE1 benign 22RV1 malignant MCF-10A benign MCF-7 malignant Detroit 551 benign A375 maligna nt 0.60±0.10 0.15±0.05 2.21±0.35 0.12±0.05 2.21±0.4 0 0.30±0.1 0 78 3.6.4 Sensitivity to acoustic irradiation varies due to variations in cell’s cytoskeletons In Section 3.6.1, it is showed that the SFAT with 800 μm focal length produces localized cell lysis at the focus spot. To determine whether the SFAT’s focused acoustic beam can penetrate into tissues, the penetration of the cytolytic effect is tested in three-dimensional (3D) tumor spheroids made from benign and malignant cell lines. Spheroids are placed directly on the parylene membrane in the cell culture chamber. The spheroids are irradiated by focused ultrasound from SFAT, and they are stained by SYTOX green fluorescent dye with the similar procedure as described in Section 3.5.1. It is noted that the penetration of ultrasound causes the cell lysis in a size of around 100 µm into the the tumor spheroids for malignant (RV1 and MCF-7), but not benign (RWPE1 and MCF-10A) 3D tumor spheroids. The lysed cells are stained by SYTOX fluorescence dye in the 3D spheroids when the applied peak acoustic intensity is 2.4 W/cm 2 . This result further confirms the increased sensitivity for cancer cells over normal cells to the acoustic energy produced by SFAT. 3.6.5 Sensitivity to acoustic irradiation varies due to variations in cell’s cytoskeletons In previous studies, it has been shown that different cell lines have different sensitizations to acoustic pressure, which may depend on their cellular stiffness [12]. Thus, it is hypothesized that the acoustic intensity thresholds (AITs) are different in malignant and non-malignant cell lines due to variations in cell’s biophysical 79 properties. Specifically, variations of the sensitivity to acoustic irradiations in benign cells and cancer cells are possibly due to the reason that prostate cancer cell line 22RV1 and breast cancer cell line MCF-7 have less area stiffness than their corresponding normal cell line, due to disorganized cytoskeletal filaments [16]. Figure 3.11: Cytolytic effect of SFAT irradiation on 3D tumor spheroids from benign (RWPE1 and MCF-10A) or malignant (22RV1 and MCF-7) cell lines. The bar represents 100 µm. The lysed cells were stained by SYTOX green fluorescence dye in the 3D spheroids when the applied peak acoustic intensities were 2.4 W/cm 2 . The incident direction of the focused acoustic beam is perpendicular to the plane of the photo. In general, cell cytoskeleton is the internal framework of a cell, composed largely of actin filaments as well as microtubules and intermediate filaments, which resists 80 deformation and reorganizes in response to external forces [21-25]. The dominant influence of actin in determining the stiffness of cells is shown by signal cell stiffness measurement using atomic force microscopy (AFM) [16] and laser ablation [25]. In [24], cell’s responds to various types of mechanical stresses, the influences of cytoskeleton on mechanical properties of cells, and the potential of cytoskeletal elements detecting and integrating physical stimuli are reviewed. F-actin network’s nonlinear elastic response to mechanical stress is studied, and the response depends on the density of networks and the degree of cross-linking [22]. As cytoskeletal changes accompany malignant transformation as well as biomechanical properties [16], it is hypothesized that differences in cytoskeletal organization and cellular compliance might contribute to observed variations in sensitivity to acoustic irradiation. Therefore, actin filament patterns in various cell lines are examined in Figure 3.10. Compared to benign cell lines, the malignant cells exhibited more peripheral staining for actin and less stress fibers. Taken together, the results from actin pattern staining and acoustic intensity sensitization experiments provided the strong evidence indicating that the cytoskeleton polymerization has high correlation with acoustic intensity sensitization. In order to predict the direction of AIT change due to the cytoskeleton alteration, Bilayer Sonopore (BLS) model is employed to analyze the interactions between cells and ultrasound. The interactions between the ultrasound and the cell membrane are governed by modified Rayleigh-Plesset equation [12]. Since cell’s mechanical 81 properties are indications of the cellular actin cytoskeleton patterns, reduction of cytoskeleton density would result in decreased cell membrane stiffness. Hence, it is possible to simulate the magnitude of area strain of cell membrane induced by focused acoustic waves, and investigate the variations of the strain in accordance with changes in area stiffness of cell membrane. According to the simulation results based on the BLS model, for a given acoustic intensity, a lower area strain of the membrane is produced for the cells having higher cytoskeleton actin density and higher stiffness (that is basically area compression modulus k s in Figure 3.12). Figure 3.12: Simulated maximum area strain vs. area compression modulus (k s ) as a function of free lipid area (2 πa) for 2.4 W/cm 2 acoustic intensity 82 To experimentally investigate the correlation of cytoskeleton organization and acoustic intensity sensitization, the cytoskeleton organization in MCF-10A benign breast cells has been altered by actin polymerization inhibitor treatments (Blebbistatin and Y-27632). Then, the changes in sensitization toward acoustic intensity are measured (Figure 3.13a). The drug treatments of Blebbistatin and Y-27632 modify actin-microtubulin polymerization so as to lower cell stiffness [21, 26-29]. As expected, it is found that MCF-10A cells treated with both inhibitors are more sensitive to acoustic treatment (Figure 3.13b) than non-treated MCF-10A cells. The actin patterns in blebbistatin- or Y-27632-treated and non-treated cells are examined. It is also revealed that the cytoskeleton polymerization is correlated with acoustic intensity sensitization (Figure 3.13a). Furthermore, the cytoskeleton polymerization in MCF-10A cells is altered by nocodazole, which enhances cell stiffness via microtubules [21]. As expected, the nocodazole-treated cells were less sensitive to ultrasound treatments (Figure 3.13b). Taken together, these results suggest that changes in cytoskeletal organization and biomechanical characteristics which differ between normal and malignant cells contribute to the cytolytic effect of highly focused acoustic energy generated from our SFAT device. Table 3-3 summarizes effects of the drugs, cytoskeleton variations quantified by F-actin fluorescence imaging, and measurement results of AIT for drug treated cells. 83 Figure 3.13: The change of the acoustic intensity sensitivity due to the alterations of cytoskeleton organization by actin polymerization inhibitors (MCF-10A cells are treated with Y-27632, blebbistatin and nocodazol prior to acoustic irradiation): (a) actin pattern visualized by fluorescent-labelled phalloidin staining and microscopy; (b) the AIT measured by acoustic irradiation after drug treatments; (c) the correlation between actin distribution intensity and AIT represented by a bar graph (error bars show standard deviations calculated by 10 individual cells) 84 Table 3-3: Summary of Acoustic Intensity Threshold Variation of MCF-10A Drug Treatment Effects Cytoskeleton Variations Acoustic Intensity Thresholds (AIT) (W/cm 2 ) FTS (50 μM) Ras Inhibitor 19.3% actin density AIT ~ 2.4 shCDC42 (2 M.O.I) Regulation of the cell cycle 34.8% actin density 0.6<AIT<2.4 NSC23766 (12 μM) Rac1 Inhibitor 13.8% actin density 0.6<AIT<2.4 Y-27632 (50 μM) Rock Inhibitor 72.0% actin density AIT ~ 2.4 Blebbistatin (5 μM) Blebs inhibitor 63.0% actin density; Reduce stiffness AIT< 2.4 No Drug Treatment N/A 100.0% actin density AIT ~ 9.6 Nocodazol. (5 μM) Synchronize the cell division cycle 117.0% actin density; Increase stiffness AIT > 9.6 3.7 Summary In this chapter, the design, fabrication and testing of a novel SFAT micro-device with peak acoustic intensity of 0.1 - 10 W/cm 2 is described. And a localized cytolytic effect without any appreciable cavitation or heat caused by the acoustic beam is presented. The focal area is 100 microns in diameter and the focal length is 800 µm above the device. However, adjacent cells or tissues outside the focal area are not affected during the localized cytolysis. The AITs of various malignant and benign cell lines are investigated and compared. The comparison shows that malignant cell lines have lower AITs than benign cell lines do. The cytoskeletal drug treatments and their results in AIT changes confirm the positive correlation of the AIT of cell lysis and the cytoskeleton organizations. These experimental results show SFAT’s 85 potential for a low-power, cancer-specific, and localized cell lysis device with 100 micron resolution. The MEMS-based SFAT preserves the safety and effective tissue penetration of conventional ultrasonic technology while providing micron-level precision in focusing and control of acoustic energies. Due to the versatile fabrication technology for SFAT, the focal spot can easily be designed to be anywhere between 5 and 200 µm in diameter, significantly smaller than conventional ultrasonic transducers. For a focal spot less than 50 µm in diameter, the ultrasonic frequency will have to be higher than 30 MHz, which is too high for an ultrasonic transducer fabricated with conventional technology. Thus, the SFAT, due to its micron-level precision and capability to produce specific cytolysis effect on malignant cells and tissue, is expected to be incorporated into a hand-held surgical instrument, or other similar devices, for cancer therapeutic procedures. Reference [1] N. Privorotskaya, et al., "Rapid thermal lysis of cells using silicon-diamond microcantilever heaters," Lab on a Chip, vol. 10, pp. 1135-1141, 2010. [2] N. Bao, et al., "Single-cell electrical lysis of erythrocytes detects deficiencies in the cytoskeletal protein network," Lab on a Chip, vol. 11, pp. 3053-3056, 2011. [3] D. Di Carlo, et al., "On-chip cell lysis by local hydroxide generation," Lab on a Chip, vol. 5, pp. 171-178, 2005. [4] M. W. Miller, et al., "A review of in vitro bioeffects of inertial ultrasonic cavitation from a mechanistic perspective," Ultrasound in Medicine and Biology, vol. 22, pp. 1131-1154, 1996. 86 [5] D. L. Miller and J. Song, "Tumor growth reduction and DNA transfer by cavitation-enhanced high-intensity focused ultrasound in vivo," Ultrasound in Medicine & Biology, vol. 29, pp. 887-893, 2003. [6] P. A. Prentice, et al., "Spatially controlled sonoporation of prostate cancer cells via ultrasound activated microbubble cavitation," in Microtechnology in Medicine and Biology, 2005. 3rd IEEE/EMBS Special Topic Conference on, 2005, pp. 158-159. [7] X. Q. J. Zhu, J. Oiler, C. Yu, Z. Wang, C. Lee, T.K. Hsiai, E.S. Kim, and H. Yu, "Localized Cell Lysis Using Self Focused Acoustic Transducers," Transducers '09, IEEE International Conference on Solid-State Sensors and Actuators, Denver, CO, vol. June 21 - 25, pp. pp. 608 - 611, 2009. [8] P. Marmottant and S. Hilgenfeldt, "Controlled vesicle deformation and lysis by single oscillating bubbles," Nature, vol. 423, pp. 153-156, May 2003. [9] C. Y. Lee, et al., "Acoustic ejector with novel lens employing air-reflectors," MEMS 2006: 19th IEEE International Conference on Micro Electro Mechanical Systems, Technical Digest, pp. 170-173, 2006. [10] D. Huang and E. S. Kim, "Micromachined acoustic-wave liquid ejector," Journal of Microelectromechanical Systems, vol. 10, pp. 442-449, Sep 2001. [11] G. S. Kino, Acoustic waves: devices, imaging, and analog signal processing: Prentice-Hall, 1987. [12] B. Krasovitski, et al., "Intramembrane cavitation as a unifying mechanism for ultrasound-induced bioeffects," Proceedings of the National Academy of Sciences, vol. 108, pp. 3258-3263, February 22, 2011 2011. [13] K. Olbrich, et al., "Water Permeability and Mechanical Strength of Polyunsaturated Lipid Bilayers," Biophysical Journal, vol. 79, pp. 321-327, 2000. 87 [14] U. Demirci, "Acoustic picoliter droplets for emerging applications in semiconductor industry and biotechnology," Microelectromechanical Systems, Journal of, vol. 15, pp. 957-966, 2006. [15] B. Lu, et al., "A study of the autofluorescence of parylene materials for [small mu ]TAS applications," Lab on a Chip, vol. 10, pp. 1826-1834, 2010. [16] Q. S. Li, et al., "AFM indentation study of breast cancer cells," Biochemical and Biophysical Research Communications, vol. 374, pp. 609-613, 2008. [17] V. Swaminathan, et al., "Mechanical Stiffness Grades Metastatic Potential in Patient Tumor Cells and in Cancer Cell Lines," Cancer Res, vol. 71, pp. 5075- 5080, August 1, 2011 2011. [18] J. Zhu, et al., "Localized cell lysis by Self Focused Acoustic Transducers," in Solid-State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, 2009, pp. 608-611. [19] D. Dalecki, "MECHANICAL BIOEFFECTS OF ULTRASOUND," Annual Review of Biomedical Engineering, vol. 6, pp. 229-248, 2004. [20] F. Lejbkowicz, et al., "The response of normal and malignant cells to ultrasound in vitro," Ultrasound Med Biol, vol. 19, pp. 75-82, 1993. [21] S. Sen and S. Kumar, "Cell-Matrix De-Adhesion Dynamics Reflect Contractile Mechanics," Cell Mol Bioeng, vol. 2, pp. 218-230, Jun 2009. [22] J. Stricker, et al., "Mechanics of the F-actin cytoskeleton," J Biomech, vol. 43, pp. 9-14, Jan 5 2010. [23] F. J. Byfield, et al., "Endothelial actin and cell stiffness is modulated by substrate stiffness in 2D and 3D," J Biomech, vol. 42, pp. 1114-1119, May 29 2009. 88 [24] P. A. Janmey and C. A. McCulloch, "Cell mechanics: integrating cell responses to mechanical stimuli," Annu Rev Biomed Eng, vol. 9, pp. 1-34, 2007. [25] J. Wu, et al., "Investigation of in vivo microtubule and stress fiber mechanics with laser ablation," Integr Biol (Camb), vol. 4, pp. 471-9, May 2012. [26] K. M. Thuet, et al., "The Rho kinase inhibitor Y-27632 increases erythrocyte deformability and low oxygen tension-induced ATP release," Am J Physiol Heart Circ Physiol, vol. 301, pp. H1891-6, Nov 2011. [27] T. Mizutani, et al., "Cellular stiffness response to external deformation: tensional homeostasis in a single fibroblast," Cell Motil Cytoskeleton, vol. 59, pp. 242-8, Dec 2004. [28] H. Darenfed, et al., "Molecular characterization of the effects of Y-27632," Cell Motil Cytoskeleton, vol. 64, pp. 97-109, Feb 2007. [29] Y. Dou, et al., "Blebbistatin specifically inhibits actin-myosin interaction in mouse cardiac muscle," Am J Physiol Cell Physiol, vol. 293, pp. C1148-53, Sep 2007. 89 Chapter 4 Self Focusing Acoustic Transducers (SFATs) with 10-mm Focal Length for Cancer-specific Localized Cytolysis of 3D Cell Spheroids in 3D Matrigel In this chapter, three-dimensional (3D) localized cytolysis of cell spheroids by a SFAT with 10 mm focal length is presented. In the previous chapter, it has been demonstrated that an SFAT with 800µm focal length caused localized cytolysis on monolayer cells. Since practical ultrasonic cancer therapeutics are usually performed in a 3D biological environment, design, fabrication and testing of a SFAT having a long penetration depth and focusing acoustic energy with 10 mm focal length and a focal area of 160 µm in diameter are described in this chapter. Also, the acoustic intensity thresholds (AITs) of spheroids for cancer-specific cytolysis of malignant cells (breast cancer MCF-7 and prostate cancer 22RV1) are investigated. 4.1 3D cell spheroids in 3D Matrigel environment 3D cell spheroids assays have been developed as important in vitro assays to study normal stem cells and cancer stem cells (CSCs) found in brain, breast, prostate and other cancers [1-3]. Cell spheroids are usually cultured from single cells in Matrigel mixture. Typically, the spheroid formed in a spherical structure with a size of 100 to 200 μm in diameter. It has been demonstrated that spheroids formed from 90 stem cells and CSCs in Matrigel are able to differentiate and generate complex functional structures in one to two weeks [3, 4]. Since tumor studies using standard serial transportation methods in mice require relatively long time, the spheroids assays could provide a useful, fast and predictive in vitro model of the therapeutic response of normal tissues and tumors to possible therapeutic methods, such as chemotherapies, radiation therapies and ultrasonic therapies [1]. In Chapter 3, a SFAT with a focal length of 800 μm has been designed for locally lysing 2D monolayer cells cultured on a parylene-D membrane in a micro-fabricated silicon chamber. However, practical ultrasound treatments are performed in a complicated biological environment, where cells grow in a form of 3D tissue structure and enclosed by extracellular matrix (ECM). To examine SFAT’s ultrasound focusing function and localized cytolytic effects in the 3D biological environment, the 3D spheroids assay is a good basis for studies of solid tumors in vitro. Being different from cells cultured in monolayer, a spheroid formed from a single stem cell has a spherical structure with multiple cellular layers. Also, microenvironments of spheroids in Matrigel and crosslinks between Matrigel and spheroids are similar to real growth conditions experienced by living cells in vivo. In addition, the thickness of Matrigel with spheroids embedded could be adjusted from 1 mm to 1cm. Therefore the spheroids assay is a flexible tool for testing ultrasonic transmissions and reflections in the 3D biological environment. For experiment results observations, the fluorescent staining assays are used. The transparency of the Matrigel facilitates the staining observations under a fluorescent microscope. 91 In previous 2D monolayer cell model, it has been showed that sensitivity to acoustic irradiation varies in benign and malignant cell lines. The general trend associating a lower AIT with a more malignant growth phenotype has been observed in MCF-7 and MCF-10A breast cell lines, as well as in 22RV1 and RWPE1 prostate cell lines [5]. In this chapter, the differences of sensitivity to acoustic irradiation between benign cell spheroids and malignant cell spheroids in 3D biological environments are investigated. Also, in 3D Matrigel culture system, the biophysical properties of the cellular microenvironment and the extracellular matrix’s (ECM) response to the ultrasonic radiations forces are considered as impactful factors contributing to the AITs of cell spheroids. 92 Figure 4.1: Photos of (a) the front view and (b) the side view of a SFAT with an acoustic chamber; (c) the front view and (d) the back view of a spheroids chamber with a transparent Parylene-Si x N y -Parylene bottom membrane 4.2 Material and methods 4.2.1 Fabrication and design of SFAT with 10 mm focal length working at 17.3 MHz The SFAT working at 17.3 MHz with 10 mm focal length is fabricated on a 127- μm-thick Lead Zirconate Titanate (PZT) substrate (Figure 4.1). The size of active area of the SFAT depends on the ring number of SFAT’s acoustic lens and its focal length. As shown in Figure 4.1, a SFAT working at 17.3 MHz with 10 mm focal length and a 7-ring acoustic lens has an active area of 6.6 mm in diameter, which is 93 about 3 times larger than a SFAT having a 7-ring acoustic lens with a focal length of 800 μm. Figure 4.2: Fabrication steps of the SFAT, the acoustic chamber and the spheroids chamber. The detailed fabrication steps of a SFAT are described in Section 3.2. A laser-cut acrylic chamber with a height of 9 mm is bonded to the SFAT as an acoustic chamber (Figure 4.2). Ultrasonic gel (AQUASONIC® 100) is filled in the acoustic chamber to improve acoustic coupling and ultrasound transmissions. A spheroids chamber consists of a 400- μm-thick silicon cavity with a parylene(3 μm)-Si x N y (1 μm)-parylene(3 μm) transparent membrane and a laser-cut acrylic chamber. 94 4.2.2 Fabrication and design of SFAT with 10 mm focal length working at 2.14 MHz Since focal size of the SFAT depends on the wavelength of acoustic waves in the media, in order to increase the focal area of a SFAT, the SFAT working at 2.14 MHz is fabricated on a 1.02-mm-thick PZT substrate (Figure 4.3). Figure 4.3: Photos of (a) front view a SFAT with patterned electrodes working at 2.14 MHz and (b) the SFAT with an adjustable acoustic chamber for adjusting the distance from the SFAT’s acoustic lens The size of a SFAT’s lens working at 2.14 MHz with 6 rings is 19 mm in diameter, the focal length is designed to be 10 mm, and the estimated wavelength in media is 690 μm. Electrodes of the SFAT are patterned into Fresnel lens, and one layer of 3- μm-thick parylene is deposited on the device. Two layers of laser-cut acrylic chambers are bonded to the SFAT as an acoustic chamber. The thickness of bottom layer of the acoustic chamber is 5 mm. The height of acoustic chamber is adjustable by using different spacers with various thicknesses as the second layer of the acoustic chamber, which are from 1.5 mm to 6.5 mm. Ultrasonic gel 95 (AQUASONIC® 100) is filled in the acoustic chamber to improve acoustic coupling and ultrasound transmissions. 4.2.3 Assays for spheroids formation and transfer Cells are counted and suspended in 1:1 Matrigel (BD Bioscience)/culture medium in a total volume of 250 μL. The mixture is placed in a well of a 24-well plate, solidified at 37 °C for 15 minutes, and then cultured with 1 mL culture medium with 10% Fetal Bovine Serum (FBS) solution. The culture medium is changed every 3 days. Grown spheroids are counted at the seventh or fourteenth day after seeding. To transfer spheroids from the plate to a spheroid chamber, Matrigel are digested in 500 μL of dispase (BD Bioscience) at 37 °C for 30 minutes. Then, the digested solution with spheroids are centrifuged, washed two to three times with 1xPBS, and re-suspended in 50% Matrigel with culture medium containing 10% FBS. The Matrigel with spheroid mixture (usually 100 μL) is placed in a spheroid chamber, and solidified at 37 °C for 15 minutes. In the second experiment group, the “unsolidified” Matrigel without crosslinks is made for comparisons. To form “unsolidified” Matrigel, spheroids with dispase is re- suspended in 1:1 Matrigel/culture medium, and the mixture is placed in a spheroid chamber. However, the Matrigel mixture could not be solidified, because the dispase in the mixture cleaves the crosslinks of Matrigel. In the third experiment group, spheroids are embedded in 83% neutralized rat’s tail collagen I (BD Biosciences), after being cultured in Matrigel for two weeks. 96 Then, the mixture (usually 100 μL) is transferred to the spheroids chamber for ultrasonic radiations. 4.2.4 Testing setup and experiment procedure for spheroids cell lysis by SFAT working at 17.3 MHz and 2.14 MHz To conduct the cytolysis experiment, a pulsed sinusoidal signal is applied on the SFAT transducer for acoustic pulses. The pulsed sinusoidal driving signal has a pulse width of 57.8 μs (containing 1000 cycles of 17.3 MHz sinusoidal signal) for the SFAT working at 17.3 MHz, or 467 μs (containing 1000 cycles of 2.14 MHz sinusoidal signal) for the SFAT working at 2.14 MHz, a pulse repetition frequency (PRF) of 60 Hz, and a voltage range from 0 V to 600 V (after an RF power amplifier). A spheroids chamber filled with spheroids/Matrigel mixture is stacked on the acoustic chamber for ultrasonic irradiations. During the treatment, the PZT transducer generates acoustic waves in response to applied pulsed driving signal, and Fresnel acoustic lens focuses acoustic waves at designed focal spot of 10 mm focal length. Thus, at the focus spot, highly concentrated acoustic radiation exerts high pressure to damage the cell spheroids. After one-minute treatment, the bio-effects of acoustic energy are examined by SYTOX Green (Invitrogen) dead cell staining and fluorescence microscopy. SYTOX Green is taken up by damaged or dying cells, but does not cross the plasma membrane of living cells. Spheroids are incubated with SYTOX Green solution for 10 minutes prior to fluorescence microscopy examination. The AIT for spheroids cytolysis could be evaluated by monitoring 97 fluorescence-marked lysed cells in the focal area (Figure 4.4). Figure 4.4: Three steps of determining the AITs of spheroids based on a SFAT and the fluorescent stained cytolysis assays as well as pictures of (a) grown-up MCF-7 spheroids and (b) grown-up MCF-10A spheroids in the 3D matrigel environment 4.3 Results and discussions for spheroids cell lysis by SFAT working at 17.3 MHz Photos of grown-up MCF-7 and MCF-10A spheroids in the Matrigel are shown in Figure 4.4. The spheroids density in the spheroid chamber is 10 spheroid/ μL. After transferring spheroids to spheroids chambers, spheroids are embedded in Matrigel and distributed in a 3D Matrigel environment. 4.3.1 Hydrophone measurement of focused ultrasound with 10 mm focal length To verify the focal effects of SFAT with 10 mm focal length, the SFAT working 98 at 17.3 MHz is characterized by calibrated hydrophone (PAHPM04/1, Precision Acoustics, UK) in DI water. According to measurement results, the SFAT produces five times larger acoustic intensity within the focal area (-6 dB beam width ~160 μm) on the focal plane than outside the focal area (Figure 4.5a). In z-direction, -6dB depth of focus is 700 μm. The acoustic intensity versus applied voltage at the focus is also characterized (Figure 4.5b). The trend line in Figure 4.5b shows that the acoustic intensity at the focal point is proportional to the square of the applied electric voltage. Figure 4.5: (a) Measured acoustic intensity distribution along x-axis on the focal plane and (b) measured acoustic intensity values vs. voltages applied to the SFAT. 4.3.2 Localized cell lysis of spheroids in 3D Matrigel environment First, 22RV1, MCF-7 and MCF-10A spheroids suspended in normal Matrigel are tested using the SFAT with 10 mm focal length. Spheroids are cut from Matrigel using dispase for transferring spheroids to spheroids chambers. After three times 99 PBS-wash, all dispase is removed from spheroids suspension. Then, spheroids are suspended in 100 μL 50% Matrigel and transferred to a spheroids chamber for ultrasonic radiations. The Matrigel/spheroids mixture is solidified within 15 minutes. One-minute ultrasound treatment has been shown to cause no displacement of the spheroids from their original locations in the Matrigel. But focused acoustic waves transmit through ultrasonic gel in the acoustic chamber and Matrigel in the spheroid chamber, and result in cell lysis in the spheroids. When applied acoustic intensity is above the AIT of spheroids, localized cell lysis effects are observed by SYTOX staining assays. MCF-7 cell spheroids located in the focused region could be lysed at the acoustic intensity of 11.10 W/cm 2 . The size of localized cytolysis area is about 150 μm in diameter (Figure 4.6), which is similar to the size of -6dB beam width of the SFAT. This result demonstrates that SFAT with 10 mm focal length is able to produce focused acoustic waves and induce localized cytolytic effects by penetrating the long distance of ultrasonic gel and Matrigel. However, there is no localized cytolysis effect observed in MCF-10A cell spheroids even when applied acoustic intensity is as high as 30.85 W/cm 2 . Thus, differences of resistivity to ultrasonic radiations between MCF-7 and MCF-10A spheroids are observed in the Matrigel environment, in line with our previous observations on monolayers of MCF-7 and MCD-10A cells. It is noted that the AIT of MCF-7 spheroids is two orders of magnitude higher than the AIT of MCF-7 monolayer cells cultured on the surface of parylene membrane (0.12 W/cm 2 [6]). Therefore, in 3D Matrigel environment, cells in spheroids have 100 interactions with surrounding microenvironment and change their responses to ultrasonic radiations. Figure 4.6: In the normal Matrigel, (a) the AIT of MCF-7 spheroids is increased from 0.113 W/cm 2 (for spheroids in “unsolidified” Matrigel) to 11.10 W/cm 2 due to the crosslinks between spheroids and Matrigel; and (c) MCF-7 spheroids in normal Matrigel are lysed by acoustic intensity of 15.14 W/cm 2 . In pure collagen gel, AITs of (b) MCF-7 and (d) 22RV1 spheroids are increased to 15.14 W/cm 2 because of the crosslinks of collagen between the gel and the spheroids. 4.3.3 Localized cell lysis of spheroids in “unsolidified” Matrigel environment In the second group of experiment, in order to test the direct cytolytic effect of acoustic energy on spheroids, crosslinks of collagen in Matrigel is inhibited. Dispase in spheroids suspension is partially removed by a PBS-wash. Then, spheroids with dispase are transferred to the spheroids chamber with 1:1 Matrigel/culture medium 101 mixture. Since dispase cleaves the crosslinks of collagen between spheroids and Matrigel, the Matrigel mixture cannot be solidified completely, and is more compliant than normal Matrigel. The compliance of “unsolidified” Matrigel can be noticed by deformations of the gel when we gently shake the spheroids chamber. AITs of MCF-7, 22RV1 and MCF-10A spheroids in “unsolidified” Matrigel are also investigated. MCF-7 and 22RV1 cell spheroids positioned within the focal area can be lysed at the acoustic intensity of 0.113 W/cm 2 (Figure 4.7). It is interesting that the AITs of MCF-7 and 22RV1 cell spheroids in the “unsolidified” Matrigel are similar to the AITs of MCF-7 and 22RV1 monolayer cells. In 2D monolayer cases, the AIT of MCF-7 is 0.12 W/cm 2 and the AIT of 22RV1 is 0.15 W/cm 2 [6]. It is likely that cells in spheroids embedded in “unsolidified” Matrigel have the similar response to the ultrasonic radiations as 2D monolayer cells. And it is also possible that interactions between cell spheroids and surrounding 3D Matrigel environment are reduced and minimized by dispase’s cleavage effects. Dispase also causes compliance of “unsolidified” Matrigel, which is another possible reason to the decrease of AITs of spheroids in “unsolidified” Matrigel. AIT of MCF-7 spheroids in normal Matrigel is 98 times larger than AIT of MCF- 7 spheroids in “unsolidified” Matrigel. Given the dramatic AIT differences between normal Matrigel case and “unsolidified” Matrigel case, it is highly possible that crosslinks of collagen between spheroids and Matrigel would change the interactions between spheroids and ultrasonic waves in 3D Matrigel environment, and alter the AIT of spheroids. In other words, the crosslinks change spheroid’s response to the 102 ultrasound radiation and increase AITs of spheroids in 3D Matrigel environment. Figure 4.7: In “unsolidified” matrigel, (a) MCF-7 and (b) 22RV1 spheroids are lysed with acoustic intensity of 0.113 W/cm 2 and (c) MCF-10A spheroids cannot be lysed. MCF-10A cell spheroids in “unsolidified” Matrigel could not be lysed by focused ultrasound with an acoustic intensity as high as 0.915 W/cm 2 , which is more than 8 times higher than AITs of MCF-7 and 22RV1 spheroids in “unsolidified” Matrigel. Therefore, MCF-7 malignant cell spheroids still have lower AIT than MCF-10A benign cell spheroids in the “unsolidified” Matrigel environment. Although the AIT of MCF-10A cell spheroids, in normal or “unsolidified” Matrigel, has not been decided yet, MCF-10A spheroids are not lysed even with 2 to 3 times higher acoustic intensities than the AITs of MCF-7 spheroids. Consequently, the general trend associating a lower AIT with a more malignant still stands for MCF-7 103 and MCF-10A spheroids in Matrigel environments. 4.3.4 Localized cell lysis of spheroids in the purified collagen gel environment In the third group of experiment, spheroids are suspended in the purified collagen gel and transferred to the spheroids chamber for ultrasonic radiations. Gel with spheroids embedded is solidified within 15 minutes. In the collagen gel, crosslinks of collagen is similar to the Matrigel case. Various acoustic intensities are applied for investigating the AITs of different spheroids in the gel. AIT of MCF-7 spheroids in the collagen gel is 15.14 W/cm 2 , which is 36% higher than the AIT of the spheroids in normal Matrigel. AIT of 22RV1 spheroids is the same as the AIT in Matrigel. However, there is no observation of lysed MCF-10A spheroids with the applied acoustic intensity as high as 30.85 W/cm 2 . It is noted that the crosslinks of collagen increase the AIT of spheroids in the 3D collagen environment. Table 4-1: Summary of Acoustic Intensity Thresholds (AITs) of MCF-7, 22RV1 and MCF-10A spheroids in various 3D biological environments Cell Lines Acoustic Intensity Thresholds (W/cm 2 ) “Unsolidified” Matrigel (without crosslinks) Normal Matrigel (with crosslinks) Collagen gel (with crosslinks) MCF-7 0.113 11.10 – 15.14 15.14 22RV1 0.113 – 0.135 15.14 15.14 MCF-10A > 0.915 >30.85 >30.85 104 4.4 Results and discussions for spheroids cell lysis by SFAT working at 2.14 MHz 4.4.1 Localized cell lysis of 2D monolayer cells For the SFAT working at 2.14 MHz with 10 mm focal length, the size of localized cell lysis area is tested and measured by following the protocols described in Section 3.5. Monolayer MCF-7 cells are treated with SFAT with various acoustic intensities from 47.61 W/cm 2 to 428.50 W/cm 2 . The pulse repetition frequency is 60 Hz, the number of cycles in each pulse is 1000, and the duration time of the treatment is 1 minute. According to experimental results, MCF-7 cells are detached from the parylene membrane when the applied acoustic intensity is as low as 5.29 W/cm 2 . In Figure 4.8, more than 60% of cells are detached from the parylene-D membrane when the applied acoustic intensity is 47.61 W/cm 2 . Also, the attached cells are lysed and stained by the SYOTX green fluorescent. For samples treated with an acoustic intensity as high as 259.20 W/cm 2 , almost all cells within the focal area of SFAT are detached from the parylene membrane and cells located at the edges of the membrane are also lysed by the high intensity ultrasound. 105 Figure 4.8: MCF-7 Monolayer cells treated by SFAT working at 2.14 MHz with 10 mm focal length with various acoustic intensities For MCF-10A monolayer cells, there are few lysed cells when the applied acoustic intensity is 47.61 W/cm 2 . As shown in Figure 4.9, the localized cell lysis effects of 2D monolayer MCF-10A cells are observed when the applied acoustic intensity is higher than 160.00 W/cm 2 . The size of the localized cell lysis area is about 1 mm in diameter. When the acoustic intensity is as high as 428.50 W/cm 2 , the localized cell lysis area is extended to be 2 mm in diameter, and some cells within the focal area are detached from the cell membrane. According to the experiment results, the AIT of MCF-10A monolayer cells by SFAT working at 2.14 MHz is 160.00 W/cm 2 , and the AIT of MCF-7 monolayer cells by SFAT working at 2.14 MHz is less than 5.29 W/cm 2 . The focal area of the SFAT working at 2.14 MHz is about 1 mm in diameter. There are 30 times difference between AIT of MCF-7 cells and MCF-10A cells treated by the SFAT working at 2.14 MHz. 106 Figure 4.9: MCF-10A Monolayer cells treated by SFAT working at 2.14 MHz with 10 mm focal length with various acoustic intensities 4.4.2 Localized cell lysis of 3D spheroids Figure 4.10: 3D spheroids of MCF-7 treated by SFAT working at 2.14 MHz with various duration times 107 For a fast ultrasound treatment with a large coverage area, a short treatment time is useful and required. A 3D localized cell lysis of spheroids by SFAT working at 2.14 MHz with a focal length of 10 mm is tested with the applied acoustic intensity of 428.50 W/cm 2 and various duration times from 1s to 1 minutes. In the experiment, the frequency of the acoustic waves is 2.14 MHz, the PRF is 60 Hz, the number of cycles in each pulse is 1000, the driving voltage amplitude is 450V, and the duration times is from 1s to 1 minute. Figure 4.11: 3D spheroids of MCF-10A treated by SFAT working at 2.14 MHz with various duration times from 1 second to 1 minute For the spheroids of MCF-7 and MCF-10A, spheroids within the focal area of 2 mm in diameter are lysed within 1 second, when the acoustic intensity of 428.50 W/cm 2 is applied. As the duration time increases, more spheroids are lysed and more localized staining of the lysed spheroids are found in the fluorescent pictures (Figure 108 4.10 and Figure 4.11). For MCF-7 spheroids, the high-intensity treatment with 1 minute duration time breaks most spheroids within the focal area and creates a circular shape within the Matrigel (Figure 4.10). For MCF-10A spheroids, because the spheroids are dispersed with long treatment of 1 minute, the staining of fluorescent dye is spread in the Matrigel and the cloud-like staining pattern is formed (Figure 4.11). 4.5 Summary In this chapter, a SFAT working at 17.3 MHz with 10 mm focal length and 160- μm focus beam width (-6 dB) and a SFAT working at 2.14 MHz with 10 mm focal length are designed and fabricated. Localized cytolysis and long penetration depth of the SFAT has been demonstrated using 3D cell spheroids assays in the 3D Matrigel environment. By using SFAT working at 2.14 MHz, the size of localized cell lysis and AITs of 2D monolayer cells of MCF-7 and MCF-10A by SFAT working at 2.14 MHz are tested and measured. For the ultrasound with frequency of 2.14 MHz, the AIT of MCF-7 cells is 12.74 W/cm 2 while the AIT of MCF-10A cells are 160.00 W/cm 2 . For 3D cell lysis of cell spheroids in 3D matrigel environment, with high-intensity ultrasound of 1034 W/cm 2 , the treatment duration time of 1 second is long enough for causing localized spheroids cell lysis for MCF-7 and MCF-10A cells. 109 By using SFAT working at 17.3 MHz, AITs of cell spheroids in various 3D extracellular environments are investigated (Table 4-1). The experiment results show variations of AITs of cell spheroids between malignant MCF-7 spheroids and benign MCF-10A spheroids in the 3D Matrigel environment. This difference of AITs shows potential for cancer-specific treatment by ultrasonic radiations. It is also found that AIT of MCF-7 spheroids in normal Matrigel with crosslinks differs from the AIT in “unsolidified” Matrigel without crosslinks. It implies that crosslinks of collagen change interactions between cell spheroids and surrounding environment, and change biophysical response of spheroids to the ultrasonic radiations. The crosslinks increase AITs of 22RV1 and MCF-7 cell spheroids in Matrigel by two orders of magnitude. These preliminary results will further be studied to better understand biological effects of ultrasound on 3D cell spheroids in a 3D complex extracellular environment for developing efficient cancer-specific ultrasound therapy. Reference [1] J. E. Visvader and G. J. Lindeman, "Cancer stem cells in solid tumours: accumulating evidence and unresolved questions," Nat Rev Cancer, vol. 8, pp. 755-768, 2008. [2] D. Ponti, et al., "Isolation and In vitro Propagation of Tumorigenic Breast Cancer Cells with Stem/Progenitor Cell Properties," Cancer Res, vol. 65, pp. 5506-5511, July 1, 2005 2005. [3] C.-P. Liao, et al., "Cancer-Associated Fibroblasts Enhance the Gland-Forming Capability of Prostate Cancer Stem Cells," Cancer Res, vol. 70, pp. 7294-7303, September 15, 2010 2010. 110 [4] G. Dontu, et al., "In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells," Genes & Development, vol. 17, pp. 1253- 1270, May 15, 2003 2003. [5] W. Lingtao, et al., "A Self-Focusing Acoustic Transducer That Exploits Cytoskeletal Differences for Selective Cytolysis of Cancer Cells," Microelectromechanical Systems, Journal of, vol. 22, pp. 542-552, 2013. [6] L. Wang, et al., "A Self-Focusing Acoustic Transducer That Exploits Cytoskeletal Differences for Selective Cytolysis of Cancer Cells," Microelectromechanical Systems, Journal of, vol. PP, pp. 1-11, 2012. 111 Chapter 5 Phase-locked-loops (PLLs) for Ultrasonic Velocity Sensing System and FBAR oscillator sensor frequency shift detection 5.1 Phase-locked-loop for Doppler frequency shift detection in ultrasonic velocity sensing The ultrasonic velocity sensing system is to measure the relative velocity of a transmitter and a receiver by detecting the Doppler frequency shift. According to the Doppler Effect, relative velocity v between the transmitter and the receiver is calculated by v= c( f t − f r ) 2 f t (5.1) where c, f t , and f r are the sound velocity in air, the constant transmitted frequency, and the received frequency, respectively. In Eq. (5.1), c and f t are constant, given that the transmitter is working at the fixed frequency and the sound velocity in air is constant. Therefore, the frequency of received signal f r changes linearly proportional to the velocity of the object. As a result, the relative velocity of the transmitter and the receiver could be measured by detecting the frequency shift between f r and f t . 112 Figure 5.1: Schematic of a phase-locked-loop In this work, a phase-locked-loop (PLL) is used to detect the frequency shift by monitoring the control voltage output of a phase detector in PLL. The phase-locked- loop is a feedback loop to synthesize frequency of the output signal with frequency of the input (Figure 5.1). Initially, a voltage control oscillator (VCO) generates an unbiased output signal f out0 , which is also called the free running frequency of a VCO. Then, a phase-frequency detector compares phase and frequency differences between the input and the output signal from VCO, and generates an output voltage according to the differences. Next, the output signal from the phase detector is then filtered by a low-pass-filter and converted to a DC control voltage, in order to adjust the output frequency of VCO. To form a loop, the output signal is sent back to the phase-frequency detector again. Therefore, the frequency of VCO is continually adjusted by the control voltage until the output signal is synthesized with the input signal. In this case, a stable control voltage from the phase-frequency detector indicates the frequency differences between the input signal and the initial output signal from VCO. 113 Given that VCO’s gain is K vco (Hz/V), the control voltage is Vc, and input frequency is , the control voltage can be found by the equation: (f r − f out0 )=V c ×K vco (5.2) We set f out0 equal to the transmitting frequency f t , and combine Eq. (5.1) and Eq. (5.2) to derive the relationship between the control voltage output and the relative velocity of the transmitter and the receiver: v= c×V c ×K vco 2f t (5.3) Since sound speed c, the transmitting frequency and K vco are constant, the control voltage output is linearly proportional to the relative velocity of the transmitter and the receiver. 5.2 Velocity sensing system design A schematic view of the entire velocity sensing system is shown in Figure 5.2. In the system, a MEMS transmitter is composed of a PZT bimorph diaphragm and a silicon top cover [1]. The transmitter is built on an 8 × 8 mm 2 , 254 µm thick PZT bimorph diaphragm with top and bottom patterned electrodes. The fundamental and the second harmonic resonant frequency of the transmitter are 14.3 kHz and 28.8 kHz, respectively. The driving signal is generated by a function generator, and then amplified by a piezo-amplifier, in order to drive the fabricated MEMS transmitter. 114 A MEMS receiver consists of a ZnO thin film sandwiched by two aluminum electrodes, sitting on the top of 1 µm thick, 3×3 mm 2 low-stress silicon nitride diaphragm [2]. It is mounted and wire-bonded to a preamp Print Circuit Board (PCB), which is enclosed in a metal box for electromagnetic shielding. The operating frequency of the receiver is 28.8 kHz, which is the second harmonic resonant frequency of the transmitter. The frequency choice is based on the measured frequency response of the transmitter and the receiver. At the operating frequency of 28.8 kHz, the sensitivity of the receiver is high enough to detect the transmitted signal (Figure 5.4). When the transmitter and the receiver are moving with respect to each other, the frequency of the received ultrasound is shifted from the transmitting signal, due to the Doppler Effect. A cascade PLL module converts the frequency shift to a DC voltage output, which is proportional to the relative velocity of the transmitter and the receiver. 115 Figure 5.2: Schematic diagram of velocity sensing A cascaded PLL module consists of two phase-locked-loops, which are connected in series (Figure 5.3). In order to stabilize the received signal, the first PLL is used to synthesize its output with the received ultrasound signal. Therefore, the output frequency of the first PLL is the same as the received ultrasound signal. The second PLL is used to linearly convert the frequency shift, due to the Doppler effect, included in the received signal to an output of DC voltage. A frequency divider is added in each PLL stage, and the divided frequency of VCO is 10 MHz by a divider number of 347. Therefore, the phase-frequency detector in the second PLL compared the Doppler shifted frequency to the frequency of the transmitted signal. 116 Figure 5.3: Close-up detail of the cascade PLL module A 3rd order passive low pass filter is used in the each PLL to filter out high frequency noise. In the PLL, the channel spacing is determined by the frequency of the transmitted signal, which is 28.8 kHz. Therefore, the loop bandwidth is designed to be 3 kHz. In addition, the phase margin is set to be 45 degree and the input capacitance of the VCO is 3 pF. Then, all parameters in the low pass filter could be determined by using simulation software (ADIsimPLL 3.0 Analog Device Inc.). Table 5-1 shows all simulation values of parameters. 117 Figure 5.4: Schematic of a 3rd order passive low pass filter Table 5-1: Parameter values in the low pass filter in PLLs Component Value C1 10 nF R2 1.5k ohm C2 110 nF R3 2.8k ohm C3 3 nF Figure 5.5: Simulation results of the phase noise of a PLL To reduce the system noise floor, two voltage-controlled crystal oscillators (Crystek VCXO-083-10MHz) and phase detectors (Analog Devices ADF4002) with extremely low phase noise (-150 dBc/Hz at 10 kHz) are used to build the cascaded 10 100 1k 10k 100k 1M Frequency (Hz) -160 -150 -140 -130 -120 -110 -100 -90 -80 -70 -60 Phase Noise (dBc/Hz) Phase Noise at 10.0M Hz Total Loop Filter Chip Ref VCO 118 PLLs structure. The simulation result of total phase noise is -109 dBc/Hz at 1 kHz offset (Figure 5.5). 5.3 Experimental results 5.3.1 Frequency response of the transmitter and the receiver First, the frequency response of the fabricated MEMS transmitter and receiver are characterized by using the calibrated ultrasound microphone and micro-speaker. In the characterization of the transmitter, a calibrated microphone is used in the testing setup to measure the acoustic pressure generated by the transmitter. The acoustic pressure is measured by a M31 calibrated microphone lied 5 mm away from the transmitter. During the measurement, the transmitter is driven by a 190 V pp sinusoidal signal varying from 50 Hz to 40 kHz (Figure 5.6a). To test the frequency response of the receiver, an ultrasonic microspeaker and a calibrated microphone with a flat frequency response up to 40 kHz are used in the setup, in order to measure the electric signal output from the receiver by a normalized acoustic pressure (Figure 5.6b). The frequency response of the receiver and the transmitter are shown in Figure 5.7. At the operating frequency of 28.8 kHz, the sensitivity of MEMS receiver with a pre-amp was 18 mV/Pa, and the acoustic pressure generated by the transmitter was 17.5 Pa. 119 Figure 5.6: Testing setup to characterize the frequency response of (a) the transmitter and (b) the receiver. 5.3.2 Dynamic range of frequency tuning of the PLL frequency shift detector Secondly, the dynamic range of frequency tuning of PLLs and the response of PLL to the frequency variation are measured. In the testing, the transmitter and the receiver are placed still, but the transmitter is driven by a linearly varying frequency signal by a function generator with the center frequency of 28.81755 kHz. The frequency variation rate is 0.1 Hz/s. According to the results, the control voltage changes linearly along with the transmitting frequency until the transmitting frequency is out of the dynamic range of PLL. The frequency response of PLL (PLL’s sensitivity) is measured by the control voltage output range versus the dynamic range of PLL (Figure 5.8). According to measurement results, the frequency dynamic range is from 28.8174 kHz to 28.8177 kHz, and the sensitivity of PLL S 1 is 1.32 V/Hz. 120 Figure 5.7: Frequency response of (a) MEMS receiver, (b) MEMS transmitter Figure 5.8: Output voltage vs. received frequency. (a) Frequency sweeping from 28.8174 kHz to 28.8177 kHz, the rate is 0.1 Hz/s (b) output voltage vs. received frequency 5.3.3 Real time measurement of the velocity sensing system Finally, to test the real time response, the receiver is fixed still on a holder and the transmitter is mounted on the moving surface of a shaker (B&K mini-shaker 4810). The shaker is sinusoidally driven at a low frequency (from 0.1 Hz to 1Hz), yielding a relative velocity of v=D 0 ×ω×cos(ωt), where D 0 and ω are displacement 121 amplitude and driving angular frequency, respectively (Figure 5.9). As a result, the output voltage of the second stage PLL changes sinusoidally along with the moving velocity of the shaker. It is demonstrated that the system is able to measure the relative velocity between the transmitter and the receiver in real-time (Figure 5.10). Figure 5.9: Pictures of (a) a packaged MEMS transmitter, (b) MEMS receiver and preamplifier, (c) close up picture of MEMS receiver, (d) a transmitter is placed on the surface of shaker and a receiver is enclosed in a metal box fixed by a holder. According to the measurement result, the reference voltage is 0.8 V, which is the output voltage for zero relative velocity. The sensitivity S of the entire system can be calculated, either from the PLL’s sensitivity S 1 and the Doppler Effect, or from the results shown in Figure 5.10. Then, the sensitivity of the velocity sensing system is: S = 2S 1 f t c = V out(max) −V out(min) 2D 0 ω (5.4) 122 The measured sensitivity is 0.22 V/(mm/sec). Since the noise level of the VCXO control signal is about 0.15 V peak-to-peak , the minimum detectable velocity is 0.67 mm/sec. Figure 5.10: Voltage output when the shaker is driven a by sine wave with varying frequency, 1Hz, 0.5Hz, 0.25Hz and 0.1Hz. 5.4 A phase-locked loop for FBAR and HBAR oscillator frequency shift detections For a FBAR working in the thickness extensional mode, the resonant frequency of a FBAR is determined by the thickness of sandwiched electrode-piezo-electrode layers. As additional mass is added onto the FBAR surface, FBAR’s resonant 123 frequency decreases. The mass-frequency relationship is determined by the Sauerbrey equation [3]: Δf f r ≈− ρ m d m ρ 0 d 0 (5.5) where Δf is the resonant frequency shift, f 0 is the base resonant frequency; ρ m and ρ 0 are the densities of the added mass and the resonator, respectively; d m and d 0 are the thickness of the added mass and the resonator, respectively. The equation shows an approximately linear relationship between the resonant frequency shift and the added mass. Figure 5.11: Schematic of FBAR-based Pierce oscillator circuit HBAR-based Pierce oscillators is studied in [4]. The oscillation frequency is dependent on resonantor’s series resonant frequency f r , values of C 1 and C 2 , as well as the FBAR’s series and parallel capacitances C x and C p : 124 f o = f r × 1+ C x C p +C 1 +C 2 (5.6) According to Eq. (5.6), the frequency output of FBAR oscillator has a linear relationship with resonator’s resonant frequency. Therefore, the mass loading effect on the FBAR will cause the frequency shift on the oscillator. Figure 5.12: Schematic of FBAR frequency shift detection system 5.5 System design of FBAR oscillator sensor and phase-locked loop frequency shift detection In order to detect the frequency shift and convert the shift to a DC voltage change, a phase locked loop (PLL) frequency shift detection system shown in Figure 5.12 has been designed and tested. The frequency output of the oscillator f r is multiplexed with a reference frequency f LO for the frequency difference f IF =|f r - f LO |. Typically, the FBAR oscillator’s frequency ranges from 1.1 GHz to 1.4 GHz. 125 Therefore, the reference frequency generated by a PLL frequency synthesizer (LMX2541 Evaluation board, Texas Instrument) is set to be 50MHz or 100 MHz off from the f r . The phase noise floor of the reference frequency is -225 dBc/Hz and ±200 Hz variance at 1.0~1.4 GHz. After a band-pass filter and a frequency divider R, the divided frequency f IF /R is very close to the output frequency of VCXO (10.000000 MHz, the N divider is set to 1) in the PLL frequency shift detector. The phase noise floor of VCXO used in the PLL is -150 dBc/Hz (Crystek VCXO-083- 10MHz). Because a phase-frequency detector compares phase and frequency difference between the input and the output signal from the VCXO and generates an output voltage according to the difference, the VCXO’s frequency follows the divided signal f IF /R, and the control voltage (V out ) of VCXO changes linearly proportional to the frequency difference between f IF /R and f VCXO . The voltage output V out is related to the sensitivity of the detector S, the frequency divider R and the frequency shift Δf of the FBAR oscillator as follows: V out = Δf × S R (5.7) Thus, V out is proportional to the oscillator’s frequency shift caused by added mass on the FBAR’s surface. In FBAR’s biological sensing applications, a typical frequency shift is from 10 kHz (10 ppm) to 300 kHz (300 ppm), which requires a highly sensitive instrument for frequency shift detections. In the frequency shift detection system (Figure 5.12), a down conversion mixer and a simple PLL frequency synthesizer convert the base frequency of the FBAR oscillator from about 126 1 GHz down to tens of MHz, and keep the frequency shift information. In the PLL frequency shift detection, the VCXO working at 10 MHz has a tuning range of 400 ppm, which can cover the frequency shift over 4 kHz with extremely good sensitivity (if R is set to 1) or over 400 kHz (if R is set to 100). Therefore, the PLL frequency shift detector has an additional advantage of an adjustable detection range (through varying the R divider). 5.6 Results and discussions Both HBAR- and FBAR-based oscillators are tested with the PLL frequency shift detector. High-tone Bulk Acoustic Resonators (HBARs) are fabricated on sapphire substrate with piezoelectric ZnO films, and measured to have more than 10 times higher quality factor (Q) than FBARs [4]. The frequency synthesizer has the frequency range from 1.1 to 1.3 GHz, and the wide tuning range of the synthesizer enables a large detection range of the sensor. The frequency spectrum peaks at the oscillation frequencies of the HBAR oscillator, FBAR oscillator and frequency synthesizer, as shown in Figure 5.13. According to the spectrum analyzer measurement results, the maximum frequency fluctuations of the HBAR oscillator, FBAR oscillator and frequency synthesizer are 5 kHz (4.72 ppm), 20 kHz (15.3 ppm), and 0.4 kHz (0.34 ppm), respectively. Since the frequency synthesizer has a much narrower peak and a smaller frequency 127 fluctuation than the HBAR and FBAR oscillators, the synthesizer does not add noise to the detection system. Figure 5.13: The frequency spectrum peaks at the oscillation frequencies of a HBAR oscillator, a FBAR oscillator and a frequency synthesizer 5.6.1 Testing setup The fabricated FBAR oscillator and the designed PLL frequency shift detector are shown in Figure 5.14. The mixer output f IF and the voltage output V out are connected to a frequency counter and an oscilloscope, respectively, for testing functions of the PLL frequency shift detector. The dynamic range of the frequency shift detector and the rate of the frequency shift over the voltage change are measured by slowly changing the output frequency of the frequency synthesizer. In 128 this way, the frequency difference f IF between the FBAR’s oscillation frequency f r and the reference frequency f LO increases as the reference frequency increases. By monitoring voltage output V out versus the frequency difference f IF , the dynamic range of the frequency shift and the voltage-frequency sensitivity of the frequency shift detector could be measured. Figure 5.14: Pictures of the fabricated FBAR oscillator and designed PLL frequency shift detecting system 5.6.2 Measurements of system’s dynamic range and voltage output In a real-time measurement, the DC voltage output of the frequency shift detector increases proportionally as f IF increases (Figure 5.15). The frequency of the oscillator f r is 1.087808 GHz (with +/-5 kHz) and the frequency of the frequency synthesizer 129 f LO is originally set to be 1.137799 GHz. Therefore, the frequency shift f IF =|f r - f LO |=49.991 MHz and the voltage output of the PLL frequency detector is initially 0 V. As the frequency of the frequency synthesizer increases, the voltage output increases linearly to follow f IF until it is above the upper limit of the dynamic range of the detector. According to measurement results, as R is set to be five, the dynamic range of the frequency shift detector is 20 kHz (i.e., from 49.99 MHz to 50.01 MHz). The range of the voltage output is from 0 V to 5V. Figure 5.15: Testing setup to measure the dynamic range of the frequency shift detector: f IF is changing from 49.99 MHz to 50.01 MHz, while R is set at 5. The voltage output is linearly changing from 0 V to 5 V according to the frequency shift. The noise level is ±0.54 V (R=5). Since the VCXO’s tuning range is 24.15 kHz/5 = 4.83 kHz, which is 483 ppm of the VCXO frequency at 10 MHz, the dynamic range of the frequency shift detector can be increased by increasing the dividend of the frequency divider R, as the dynamic range of the frequency shift is R×4.83 kHz. For example, the dynamic 130 range of the system can be adjusted from 4.83 kHz (R=1) to 144.9 kHz (R=30) by changing R. Figure 5.16: Real-time measured frequency difference f IF =f LO -f r (blue line with smaller amplitude) and voltage output V out (red line with larger amplitude). The detection range is from 49.99 MHz to 50.01 MHz (R=5), and the V out tracks the f IF . 5.6.3 Measurements of voltage output in real-time To demonstrate the real-time frequency shift measurement, the DC voltage V out and the frequency difference f IF =f LO -f r are recorded (Figure 5.16). In the measurement, the f IF is varied by changing the frequency output of the synthesizer f LO . The R is set to be five and the dynamic range of the frequency shift detector is about 20 kHz. As shown in Figure 5.16, the DC voltage output V out changes linearly according to the frequency variations of f IF in real-time. 131 The noise floors of HBAR- and FBAR oscillators are 5 kHz and 20 kHz, respectively, and dominate the noise floor of the whole detection system. Therefore, the minimum detectable frequency shifts are 4.72 ppm and 15.3 ppm for HBAR and FBAR oscillator, respectively. Figure 5.17 shows the measured V out versus f IF when f IF varies from -10 to +10 kHz with respect to the reference frequency of 50 MHz. The slope is 0.207 V/kHz. Since the R is set to be five, the sensitivity of the system is S=Slope×R=1.035 V/kHz. Figure 5.17: Measured voltage output V out vs the frequency shift from about -10 kHz (f IF = 49.99 MHz) to about +10 kHz (f IF = 50.01 MHz). The slope of the frequency shift detector is 0.207 V/kHz. 5.6.4 Real-time measurement of FBAR sensor for temperature sensing The developed FBAR sensing system has been tested for temperature sensing by bringing blue ice close to FBAR oscillator for changing FBAR sensor’s oscillation 132 frequency (Figure 5.18). Initially, the oscillation frequency of the FBAR oscillator is 1.305 GHz and the frequency synthesizer is adjusted to 1.225 GHz, so that the median frequency of f IF is 80 MHz, as R is set to be 8. Since R is 8, the dynamic range of the frequency shift detector is 38.64 kHz. As the ice is approaching to the sensor, the ambient temperature decreases, and the FBAR’s oscillation frequency increases since the FBAR has a negative temperature coefficient (Figure 5.18a and b). There is a fluctuation during the movement of the ice, likely due to electromagnetic interferences from a moving object (Figure 5.18b). When the ice is taken away from the sensor, the oscillation frequency decreases and goes back to the initial value. As demonstrated in Figure 5.18, the voltage output follows the same trend as the frequency variation, and also detects the small frequency changes due to the interference from the hand movement. The frequency change of the FBAR oscillator in the experiment shown in Figure 5.18 has been recorded with a frequency counter. To compare the frequency change with the voltage output V out , the actual frequency of the FBAR oscillator is measured with a frequency counter, and is shown in Figure 5.19. The marked points, a – f, correspond to the critical points in Figure 5.18 that shows the voltage response due to the temperature change. The measurement shows that the voltage output V out of the frequency shift detector tracks the frequency shift of the FBAR oscillator in real time, and the change of the DC voltage output V out is linearly proportional to the oscillation frequency shift. 133 Figure 5.18: Real-time voltage output of the frequency shift detector measuring the FBAR oscillator frequency shift due to the temperature variations: (a) at the room temperature, the oscillation frequency is lower than the reference frequency; (b - c) as the blue ice is taken close to the oscillator, the oscillation frequency jumps up because of the dramatic decrease of the environment temperature; (d - e) the oscillation frequency gradually decreases as the ice is taken away from the oscillator; and (f) the ambient temperature changes back to the initial value. 134 Figure 5.19: The frequency changes of the FBAR oscillator measured by a frequency counter as the ice is approaching to and departing from the FBAR oscillator 5.7 Summary In the first part of this chapter, a low-cost and portable Doppler velocity sensing system has been developed. The system can be used for zero-velocity updating in an inertial navigation system for guiding human movements. MEMS piezoelectric ultrasonic transmitter and receiver are fabricated and embedded in the system. To reduce the calculation cost of the velocity estimation of the Doppler frequency shift, a cascaded PLLs module is employed. The PLL module directly converts the Doppler frequency shift to a DC voltage output, and consequently, reduces the complexity of the system. The sensing system has a minimum detectable velocity of 135 0.67 mm/sec and a sensitivity of 0.22 V/(mm/sec). Experiments show a real-time detection of velocity from 0.67 mm/s to 1.78 mm/s. In the second part of this chapter, a highly sensitive sensing system based on FBAR oscillators has been designed and fabricated. A PLL-based frequency shift detection module is employed in the system to achieve ppm-leveled frequency shift detections. Also, the system has a DC voltage output, which is easy to measure and integrate with IC circuits for post-sensing data processing. It has been demonstrated that the sensing system is able to measure the frequency shift of FBAR sensor with a minimum detectable frequency shift down to 4.81 ppm. And the voltage-frequency sensitivity of the system is 1.035 V/kHz. The adjustable dynamic range of the system (4.8 kHz to 144.9 kHz) enables it to detect various sensing events with different signal amplitudes. Reference [1] Y. Choe, et al., "High Fidelity Loud Microspeaker Based on PZT Bimorph Diaphragm," in 6th International Conference and Exhibition on Device Packaging, Scottsdale/Fountain Hills, AZ, 2010, pp. 269 – 270. [2] L. Baumgartel and E. S. Kim, "Use of compressively-stressed zinc oxide to increase microspeaker response," Proceedings of Meetings on Acoustics, vol. 9, pp. 030001-7, 2010. [3] W. Pang, et al., "Piezoelectric microelectromechanical resonant sensors for chemical and biological detection," Lab on a Chip, vol. 12, pp. 29-44, 2012. 136 [4] Y. Hongyu, et al., "HBAR-Based 3.6 GHz oscillator with low power consumption and low phase noise," Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on, vol. 56, pp. 400-403, 2009. 137 Chapter 6 Conclusion and Future Directions 6.1 Conclusion In conclusion, this thesis presents ultrasonic Micro-Electro-Mechanical Systems (MEMS) designed for microfluidic, biomedical and physical sensing applications. The nozzleless micro-ejectors with “phase-varied” and “dual-frequency” acoustic lens are fabricated and demonstrated for the electrical control of droplet direction. Stable and consistent ejections of uniform droplets (70 - 80 µm in diameter) have been obtained. With a proper design of the acoustic lens pattern, the ejection direction changes from right to left monotonically as the operating RF frequency is increased. The widest range of directional angle is from -30° to 35°. Self-Focusing Acoustic Transducers (SFATs) for three-dimensional localized cell lysis for cancer therapeutics are designed, fabricated and tested. Localized cytolysis and long penetration depth of the SFAT has been demonstrated in the biological model of 2D monolayer cells and 3D spheroids cells in 3D Matrigel environment. Acoustic Intensity Thresholds (AITs) of various 2D monolayer cells and spheroids in different 3D extracellular environments are investigated. The experiment results show variations of AITs of cell spheroids between malignant cells and benign cells either in 2D monolayer cells or in 3D Matrigel environments. This 138 difference of AITs shows potential for localized cancer-specific treatment using the SFAT. The design of sensing system including MEMS sensors and a phase-locked loop (PLL) frequency shift detector for Doppler velocity sensing and FBAR-based sensing are presented. A low-cost, portable Doppler velocity sensing system has been developed. The system can be used for zero-velocity updating in an inertial navigation system for guiding human movements. The sensing system has a minimum detectable velocity of 0.67 mm/sec and a sensitivity of 0.22 V/(mm/sec). Experiments show a real-time detection of velocity from 0.67 mm/s to 1.78 mm/s. Also, a highly sensitive sensing system based on FBAR oscillators has been designed and fabricated. A PLL-based frequency shift detection module is employed in the system to achieve ppm-leveled frequency shift detections. It has been demonstrated that the sensing system is able to measure the frequency shift of FBAR sensor with minimum detectable frequency shift down to 4.81 ppm. And the voltage-frequency sensitivity is 1.035 V/kHz. The adjustable dynamic range of the system (4.8 kHz to 144.9 kHz) enables it to detect various sensing events with different signal amplitudes. 139 6.2 A combinatory array of directional ejectors with electrical control of the droplet directions For many applications such as the DNA synthesis application, directional droplet ejectors are highly desirable to deliver different solutions to the same spot without a physical movement of the ejectors. An array of multiple ejectors could be integrated on a single device. For each ejector, the lens is designed and patterned into “phase- varied” acoustic lens for finely tuning the direction of the droplet ejections. All ejectors in the array are aligned in a circular shape and targeting at the same spot at the center of the circle on the top of the device. The operating frequencies of ejectors in the array are dynamically controlled in order to adjust the droplet ejections to the targeted spot, though different solutions in the ejectors have different density and acoustic properties. By using this ejectors array, it has potential to setup a versatile micro-fluidic platform for high-throughput biological assays with multiple chemicals involved. 6.3 Combing a FBAR sensor with antibody mobilizations, a PLL frequency shift detector and SFATs for a Lab-on-a-chip cancer diagnosis and therapeutics system It has been demonstrated that a FBAR with PSA antibody mobilizations is capable of highly sensitive detection on Prostate Specific Antigen (PSA) in solution 140 as low as several ng/mL [1]. A SFAT cell lysis device could release cell contains by breaking the cell membrane. It is possible to combine SFATs and a FBAR sensor in one system for cancer diagnosis by detecting cancer biomarkers. In the system, a SFAT releases cell contains; microfluidic channels and a droplet ejector transport the sample solution; a FBAR with antibody immobilizations works as a biochemical sensor to detect cancer biomarkers; and a PLL frequency shift detector converts the resonant frequency shift of the FBAR sensor into a DC voltage change. The SFAT device could cause localized cell lysis within an area in size of 100 μm in diameter. Released antigen proteins from cancer cells could be used as a biomarker for immunodetections by an antibody immobilized FBAR. The sampling media transportation from the SFAT to the FBAR sensing component is through microfluidic channels and a droplet ejector. The ejector will also be used for a chemical extraction process through a aqueous two phase system (ATPS) [2], in which the ejector will eject biomarker chemicals in the top layer of two phases liquid-liquid solutions. Due to specific bindings between the targeting antigen and the antibody immobilized on the FBAR, the FBAR’s resonant frequency changes according to the concentrations of antigen in the solution. A PLL frequency shift detector could convert the amount of frequency shift to a DC voltage output, which is easy to compare with a reference voltage level for detections. The designed system will be developed as a portable and miniaturized Lab-on-a-chip system for cancer diagnosis. 141 Reference [1] A. Lin, et al., "Label-Free Detection of Prostate-Specific Antigen with FBAR- Based Sensor with Oriented Antibody Immobilization," 2011. [2] H. Y. Yu, et al., "Chembio extraction on a chip by nanoliter droplet ejection," Lab on a Chip, vol. 5, pp. 344-349, 2005. 142 Bibliography A. J. D. Adams, et al., "Nanowatt chemical vapor detection with a self-sensing, piezoelectric microcantilever array," Applied Physics Letters, vol. 83, pp. 3428-3430, 2003. M. J. Ahamed, et al., "A Piezoactuated Droplet-Dispensing Microfluidic Chip," Microelectromechanical Systems, Journal of, vol. 19, pp. 110-119, 2010. B. N. Bao, et al., "Single-cell electrical lysis of erythrocytes detects deficiencies in the cytoskeletal protein network," Lab on a Chip, vol. 11, pp. 3053-3056, 2011. L. Baumgartel and E. S. Kim, "Use of compressively-stressed zinc oxide to increase microspeaker response," Proceedings of Meetings on Acoustics, vol. 9, pp. 030001-7, 2010. D. M. Bevly, "Global positioning system (GPS): A low-cost velocity sensor for correcting inertial sensor errors on ground vehicles," Journal of Dynamic Systems Measurement and Control-Transactions of the Asme, vol. 126, pp. 255-264, Jun 2004. A. Blana, et al., "Factors predicting for formation of bladder outlet obstruction after high-intensity focused ultrasound in treatment of localized prostate cancer," Urology, vol. 71, pp. 863-7, May 2008. F. J. Byfield, et al., "Endothelial actin and cell stiffness is modulated by substrate stiffness in 2D and 3D," J Biomech, vol. 42, pp. 1114-1119, May 29 2009. C. 143 S. C. Chan, et al., "Finite element analysis of multilevel acoustic Fresnel lenses," Ieee Transactions on Ultrasonics Ferroelectrics and Frequency Control, vol. 43, pp. 670-677, Jul 1996. C. G. Chaussy and S. F. Thuroff, "Robotic high-intensity focused ultrasound for prostate cancer: what have we learned in 15 years of clinical use?," Curr Urol Rep, vol. 12, pp. 180-7, Jun 2011. Y. Choe, et al., "High Fidelity Loud Microspeaker Based on PZT Bimorph Diaphragm," in 6th International Conference and Exhibition on Device Packaging, Scottsdale/Fountain Hills, AZ, 2010, pp. 269 – 270. L. Chuang-Yuan, et al., "On-Demand DNA Synthesis on Solid Surface by Four Directional Ejectors on a Chip," Microelectromechanical Systems, Journal of, vol. 16, pp. 1130-1139, 2007. D. D. Dalecki, "MECHANICAL BIOEFFECTS OF ULTRASOUND," Annual Review of Biomedical Engineering, vol. 6, pp. 229-248, 2004. D. Dalecki, "Mechanical bioeffects of ultrasound," Annu Rev Biomed Eng, vol. 6, pp. 229-48, 2004. H. Darenfed, et al., "Molecular characterization of the effects of Y-27632," Cell Motil Cytoskeleton, vol. 64, pp. 97-109, Feb 2007. U. Demirci, "Acoustic picoliter droplets for emerging applications in semiconductor industry and biotechnology," Microelectromechanical Systems, Journal of, vol. 15, pp. 957-966, 2006. D. Di Carlo, et al., "On-chip cell lysis by local hydroxide generation," Lab on a Chip, vol. 5, pp. 171-178, 2005. 144 G. Dontu, et al., "In vitro propagation and transcriptional profiling of human mammary stem/progenitor cells," Genes & Development, vol. 17, pp. 1253-1270, May 15, 2003 2003. Y. Dou, et al., "Blebbistatin specifically inhibits actin-myosin interaction in mouse cardiac muscle," Am J Physiol Cell Physiol, vol. 293, pp. C1148-53, Sep 2007. D. J. Duggan, et al., "Expression profiling using cDNA microarrays," Nature Genetics, vol. 21, pp. 10-14, Jan 1999. E. S. A. Elrod, et al., "Nozzleless droplet formation with focused acoustic beams," Journal of Applied Physics, vol. 65, pp. 3441-3447, 1989. H. Z. Hao, et al., "Sequence Specific Label-Free DNA Sensing Using Film-Bulk- Acoustic-Resonators," Sensors Journal, IEEE, vol. 7, pp. 1587-1588, 2007. S. Hirata, et al., "Ultrasonic distance and velocity measurement by low-calculation- cost Doppler-shift compensation and high-resolution Doppler velocity estimation with wide measurement range," Acoustical Science and Technology, vol. 30, pp. 220-3, 2009. Y. Hongyu, et al., "HBAR-Based 3.6 GHz oscillator with low power consumption and low phase noise," Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on, vol. 56, pp. 400-403, 2009. D. Huang and E. S. Kim, "Micromachined acoustic-wave liquid ejector," Journal of Microelectromechanical Systems, vol. 10, pp. 442-449, Sep 2001. W. Huilin, et al., "Laser Doppler velocity measurement system," in 2010 2nd International Asia Conference on Informatics in Control, Automation and Robotics (CAR 2010), 6-7 March 2010, Piscataway, NJ, USA, 2010, pp. 288-90. 145 K. S. Hwang, et al., "In-situ quantitative analysis of a prostate-specific antigen (PSA) using a nanomechanical PZT cantilever," Lab on a Chip, vol. 4, pp. 547-552, 2004. I. K. Imou, M. Ishida, T. Okamoto, Y.Kaizu, A. Sawamura, and N. Sumida, "Ultrasonic Doppler Sensor for Measuring Vehicel Speed in Forward and Reverse Motions Inlcuding Low Speed Motions," Agricultral Engineering International: the CIGR Journal of Scientific Research And Development, vol. 3, p. 007, 2001. J. X. Q. J. Zhu, J. Oiler, C. Yu, Z. Wang, C. Lee, T.K. Hsiai, E.S. Kim, and H. Yu, "Localized Cell Lysis Using Self Focused Acoustic Transducers," Transducers '09, IEEE International Conference on Solid-State Sensors and Actuators, Denver, CO, vol. June 21 - 25, pp. pp. 608 - 611, 2009. P. A. Janmey and C. A. McCulloch, "Cell mechanics: integrating cell responses to mechanical stimuli," Annu Rev Biomed Eng, vol. 9, pp. 1-34, 2007. K. G. S. Kino, Acoustic waves: devices, imaging, and analog signal processing: Prentice-Hall, 1987. B. Krasovitski, et al., "Intramembrane cavitation as a unifying mechanism for ultrasound-induced bioeffects," Proceedings of the National Academy of Sciences, vol. 108, pp. 3258-3263, February 22, 2011 2011. L. A. Le Duff, et al., "Real-time acoustic velocity measurement in the air by means of laser doppler velocimetry," in 18th IEEE Instrumentation and Measurement 146 Technology Conference. Rediscovering Measurement in the Age of Informatics, May 21, 2001 - May 23, 2001, Budapest, Hungary, 2001, pp. 234-238. C.-Y. Lee, et al., "Acoustic Ejector with Novel Lens Employing Air-Reflectors," in Micro Electro Mechanical Systems, 2006. MEMS 2006 Istanbul. 19th IEEE International Conference on, 2006, pp. 170-173. C.-Y. Lee, et al., "Nanoliter droplet coalescence in air by directional acoustic ejection," Applied Physics Letters, vol. 89, pp. 223902-223902-3, 2006. C. Y. Lee, et al., "Acoustic ejector with novel lens employing air-reflectors," MEMS 2006: 19th IEEE International Conference on Micro Electro Mechanical Systems, Technical Digest, pp. 170-173, 2006. F. Lejbkowicz, et al., "The response of normal and malignant cells to ultrasound in vitro," Ultrasound Med Biol, vol. 19, pp. 75-82, 1993. Q. S. Li, et al., "AFM indentation study of breast cancer cells," Biochemical and Biophysical Research Communications, vol. 374, pp. 609-613, 2008. C.-P. Liao, et al., "Cancer-Associated Fibroblasts Enhance the Gland-Forming Capability of Prostate Cancer Stem Cells," Cancer Res, vol. 70, pp. 7294-7303, September 15, 2010 2010. A. Lin and K. Eun Sok, "Selectivity and Long-Term Reliability of Resonant Explosive-Vapor-Trace Detection based on Antigen-Antibody Binding," in Micro Electro Mechanical Systems, 2009. MEMS 2009. IEEE 22nd International Conference on, 2009, pp. 316-319. A. Lin, et al., "Label-Free Detection of Prostate-Specific Antigen with FBAR-Based Sensor with Oriented Antibody Immobilization," 2011. A. Lin, et al., "Real-time label-free detection of DNA synthesis by FBAR-based mass sensing," in Ultrasonics Symposium (IUS), 2010 IEEE, 2010, pp. 1286-1289. 147 W. Lingtao, et al., "A Self-Focusing Acoustic Transducer That Exploits Cytoskeletal Differences for Selective Cytolysis of Cancer Cells," Microelectromechanical Systems, Journal of, vol. 22, pp. 542-552, 2013. B. Lu, et al., "A study of the autofluorescence of parylene materials for [small mu ]TAS applications," Lab on a Chip, vol. 10, pp. 1826-1834, 2010. M. G. MacBeath and S. L. Schreiber, "Printing proteins as microarrays for high- throughput function determination," Science, vol. 289, pp. 1760-1763, Sep 8 2000. S. Madersbacher, et al., "Effect of high-intensity focused ultrasound on human prostate cancer in vivo," Cancer Res, vol. 55, pp. 3346-51, Aug 1 1995. P. Marmottant and S. Hilgenfeldt, "Controlled vesicle deformation and lysis by single oscillating bubbles," Nature, vol. 423, pp. 153-156, May 2003. D. L. Miller and J. Quddus, "Diagnostic ultrasound induces membrane damage in phagocytic cells loaded with contrast agent in vitro," in 2000 Ieee Ultrasonics Symposium Proceedings, Vols 1 and 2, S. C. Schneider, M. Levy, and B. R. McAvoy, Eds., ed New York: Ieee, 2000, pp. 1947-1950. D. L. Miller and J. Song, "Tumor growth reduction and DNA transfer by cavitation- enhanced high-intensity focused ultrasound in vivo," Ultrasound in Medicine & Biology, vol. 29, pp. 887-893, 2003. M. W. Miller, et al., "A review of in vitro bioeffects of inertial ultrasonic cavitation from a mechanistic perspective," Ultrasound in Medicine and Biology, vol. 22, pp. 1131-1154, 1996. T. Mizutani, et al., "Cellular stiffness response to external deformation: tensional homeostasis in a single fibroblast," Cell Motil Cytoskeleton, vol. 59, pp. 242-8, Dec 2004. 148 N. R. Nayve, et al., "High-resolution long-array thermal ink jet printhead fabricated by anisotropic wet etching and deep Si RIE," Microelectromechanical Systems, Journal of, vol. 13, pp. 814-821, 2004. C. Netsch, et al., "Development of bladder outlet obstruction after a single treatment of prostate cancer with high-intensity focused ultrasound: experience with 226 patients," J Endourol, vol. 24, pp. 1399-403, Sep 2010. O. K. Olbrich, et al., "Water Permeability and Mechanical Strength of Polyunsaturated Lipid Bilayers," Biophysical Journal, vol. 79, pp. 321-327, 2000. P. W. Pang, et al., "Piezoelectric microelectromechanical resonant sensors for chemical and biological detection," Lab on a Chip, vol. 12, pp. 29-44, 2012. D. Ponti, et al., "Isolation and In vitro Propagation of Tumorigenic Breast Cancer Cells with Stem/Progenitor Cell Properties," Cancer Res, vol. 65, pp. 5506-5511, July 1, 2005 2005. P. A. Prentice, et al., "Spatially controlled sonoporation of prostate cancer cells via ultrasound activated microbubble cavitation," in Microtechnology in Medicine and Biology, 2005. 3rd IEEE/EMBS Special Topic Conference on, 2005, pp. 158-159. N. Privorotskaya, et al., "Rapid thermal lysis of cells using silicon-diamond microcantilever heaters," Lab on a Chip, vol. 10, pp. 1135-1141, 2010. R. B. Razavi, Monolithic phase-locked loops and clock recovery circuits: theory and design: Wiley. com, 1996. 149 N. Reis, et al., "Ink-jet delivery of particle suspensions by piezoelectric droplet ejectors," Journal of Applied Physics, vol. 97, pp. 094903-094903-6, 2005. S. O. P. Sahu and A. K. Gupta, "Measurement of Distance and Medium Velocity Using Frequency-Modulated Sound/Ultrasound," Instrumentation and Measurement, IEEE Transactions on, vol. 57, pp. 838-842, 2008. L. Scalise, et al., "Self-mixing laser diode velocimetry: application to vibration and velocity measurement," Instrumentation and Measurement, IEEE Transactions on, vol. 53, pp. 223-232, 2004. M. Schena, et al., "Quantitative Monitoring of Gene-Expression Patterns with a Complementary-DNA Microarray," Science, vol. 270, pp. 467-470, Oct 20 1995. S. Sen and S. Kumar, "Cell-Matrix De-Adhesion Dynamics Reflect Contractile Mechanics," Cell Mol Bioeng, vol. 2, pp. 218-230, Jun 2009. K. Seok-Tae and N. Cam, "A millimeter-wave multifunction sensor for wireless monitoring of displacement and velocity," in Wireless Communications and Applied Computational Electromagnetics, 2005. IEEE/ACES International Conference on, pp. 451-453. K. Seoktae and N. Cam, "On the development of a multifunction millimeter-wave sensor for displacement sensing and low-velocity measurement," Microwave Theory and Techniques, IEEE Transactions on, vol. 52, pp. 2503-2512, 2004. J. Stricker, et al., "Mechanics of the F-actin cytoskeleton," J Biomech, vol. 43, pp. 9- 14, Jan 5 2010. L. Sukhan, et al., "An Electrostatic Drop-On-Demand Micro Droplet Ejector with a Pole Type Nozzle," in Nano/Micro Engineered and Molecular Systems, 2006. NEMS '06. 1st IEEE International Conference on, 2006, pp. 1068-1071. 150 V. Swaminathan, et al., "Mechanical Stiffness Grades Metastatic Potential in Patient Tumor Cells and in Cancer Cell Lines," Cancer Res, vol. 71, pp. 5075-5080, August 1, 2011 2011. T. K. M. Thuet, et al., "The Rho kinase inhibitor Y-27632 increases erythrocyte deformability and low oxygen tension-induced ATP release," Am J Physiol Heart Circ Physiol, vol. 301, pp. H1891-6, Nov 2011. V. J. E. Visvader and G. J. Lindeman, "Cancer stem cells in solid tumours: accumulating evidence and unresolved questions," Nat Rev Cancer, vol. 8, pp. 755- 768, 2008. W. L. Wang, et al., "A Self-Focusing Acoustic Transducer That Exploits Cytoskeletal Differences for Selective Cytolysis of Cancer Cells," Microelectromechanical Systems, Journal of, vol. PP, pp. 1-11, 2012. J. Wu, et al., "Investigation of in vivo microtubule and stress fiber mechanics with laser ablation," Integr Biol (Camb), vol. 4, pp. 471-9, May 2012. X. T. Xu, et al., "Inkjet printing of viable mammalian cells," Biomaterials, vol. 26, pp. 93-99, Jan 2005. Y. H. Y. Yu, et al., "Chembio extraction on a chip by nanoliter droplet ejection," Lab on a Chip, vol. 5, pp. 344-349, 2005. 151 Z. H. Zhang, et al., "Label-free detection of protein-ligand interactions in real time using micromachined bulk acoustic resonators," Applied Physics Letters, vol. 96, pp. 123702-3, 2010. J. Zhang, et al., "A low-cost doppler radar system with 24 GHz PLL for remote detection of heart beat," Microwave and Optical Technology Letters, vol. 50, pp. 3139-3142, 2008. Y. Zhang, et al., "Laser Doppler Velocimetry based on self-mixing effect in vertical- cavity surface-emitting lasers," in Electronic Measurement and Instruments, 2007. ICEMI '07. 8th International Conference on, 2007, pp. 1-413-1-416. H. X. Zheng, "A displacement and velocity measurement technique using millimeter-wave sensor," International Journal of Infrared and Millimeter Waves, vol. 26, pp. 1277-1290, Sep 2005. J. Zhu, et al., "Localized cell lysis by Self Focused Acoustic Transducers," in Solid- State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, 2009, pp. 608-611.

Abstract (if available)

Abstract This thesis presents ultrasonic Micro-Electro-Mechanical Systems (MEMS) designed for microfluidic, biomedical and physical sensing applications, including a nozzleless micro-ejector with “phase-varied” and “dual-frequency” acoustic lens for the electrical control of droplet direction

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Piezoelectric MEMS for acoustic sensing, contactless RF switching, and resonant mass sensing

PDF

Piezoelectric ultrasonic and acoustic microelectromechanical systems (MEMS) for biomedical, manipulation, and actuation applications

PDF

Magnetic spring in electromagnetic vibration energy harvester and applications of focused ultrasonic transducer

PDF

Acoustic ejector employing lens with air-reflectors and piezoelectrically actuated tunable capacitor

PDF

Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications

PDF

Parylene-based biomems sensors for multiple physiological systems

PDF

Integrated wireless piezoelectric ultrasonic transducer system for biomedical applications

PDF

A high frequency array- based photoacoustic microscopy imaging system

PDF

Zero-power sensing and processing with piezoelectric resonators

PDF

Electromagnetic energy harvesting from vibrations

PDF

High frequency ultrasonic phased array system and its applications

PDF

Design and development of ultrasonic array transducers for specialized applications

PDF

High-frequency ultrasound array-based imaging system for biomedical applications

PDF

Development of high-frequency (~100mhz) PZT thick-film ultrasound transducers and arrays

PDF

Highly integrated 2D ultrasonic arrays and electronics for modular large apertures

PDF

MEMS piezoelectric resonant microphone arrays and their applications

PDF

High-resolution data acquisition with neural and dermal interfaces

PDF

High-frequency ultrasonic transducers for photoacoustic applications

PDF

2D ultrasonic transducer array’s design and fabrication with 3D printed interposer and applications

PDF

Development of back-end processing system for high frequency ultrasound b-mode imaging

Asset Metadata

Creator Wang, Lingtao (author)

Core Title Ultrasonic microelectromechanical system for microfluidics, cancer therapeutics and sensing applications

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Electrical Engineering

Publication Date 11/13/2013

Defense Date 10/01/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cancer therapeutics,FBAR sensors,micro-ejector,micro-electromechanical systems,OAI-PMH Harvest,ultrasound transducers,velocity sensors

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kim, Eun Sok (committee chair), Gross, Mitchell (committee member), Zhou, Chongwu (committee member)

Creator Email lingtao@usc.edu,wltlucky7@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-346909

Unique identifier UC11296058

Identifier etd-WangLingta-2150.pdf (filename),usctheses-c3-346909 (legacy record id)

Legacy Identifier etd-WangLingta-2150.pdf

Dmrecord 346909

Document Type Dissertation

Format application/pdf (imt)

Rights Wang, Lingtao

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA