DESIGN AND IMPLEMENTATION OF VO2 -BASED TUNABLE WINDOWS AND MICROELECTROMECHANICAL OPTICAL SHUTTERS FOR EMISSIVITY MODULATION By José Luis Figueroa-Soto A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Electrical Engineering – Doctor of Philosophy 2021 ABSTRACT DESIGN AND IMPLEMENTATION OF VO2 -BASED TUNABLE WINDOWS AND MICROELECTROMECHANICAL OPTICAL SHUTTERS FOR EMISSIVITY MODULATION By José Luis Figueroa-Soto This dissertation presents the programmability of emissivity states in a monolithically in- tegrated micro window based on vanadium dioxide (VO2 ) thin films. The 400 µm window features a VO2 thin film with integrated electrodes for actuation and sensing. The phase transition was induced by resistive heating, while the electrical resistivity and optical trans- mittance (for near IR wavelength of 1550 nm) of the VO2 thin film were monitored simul- taneously. Abrupt drops in electrical resistance and optical transmittance confirmed the quality of the VO2 thin films. Electronic pulses were used to program emissivity states in the VO2 window. The emissivity programmed state was shown for a specific DC current over imposed with the programming pulse; but any emissivity state that belongs to the minor hysteretic curves can be obtained by choosing different electronic inputs. The fully mono- lithically integrated device presented here can be used for IR cloaking applications, where different emissivity values can be programmed with electronic pulses. Micrometer-sized VO2 -based devices with integrated resistive heaters of different configu- rations were fabricated. Quality of the VO2 films was confirmed by measuring the character- istic drop in transmittance and negative differential emissivity for these films. A two-interface model for optical transmittance, reflectance, and absorbance is presented. This method and analytic model presents an advantage over most typically used approaches in that it does not require direct measurements of the material’s optical constants to estimate transmit- tance. By combining the substrate and the VO2 film into one layer with a reduced optical admittance, the two-interface model was reduced to a single-layer model. Moreover, the present work demonstrates the implementation of the developed VO2 -based devices in adap- tive camouflage and shape-converting applications. Electrical pulses are used to program different emissivity states to convert geometric shapes inside a fully integrated VO2 -based electro-optical window. This resulted in the reconfiguring of thermal images to either create new shapes, or shift from one to another. A transient heat deflection model is developed which can be applied for any bimorph structure actuated electro-thermally and for which the dimensions and material properties are known. Micrometer size VO2 based cantilevers with integrated resistive heaters are fabricated and deflections due to an input of current are recorded using a high-speed camera. From the high frame rate videos, deflection measurements and dynamic response is extracted and compared to numerical data obtained from the model. The transient heat as caused by the current input is studied, and a model is developed. VO2 -based MEMS tunable optical attenuators are demonstrated. The design consists of a VO2 -based cantilever attached to a VO2 -based optical window with integrated resistive heaters for individual mechanical actuation of the cantilever structure, tuning of the optical properties of the window, or both. Optical transmittance measurements as a function of current for both heaters demonstrates that the developed devices can be used as analog optical attenuators, where the intensity of a light beam can be tuned to any value within the range of VO2 phase transition. A transmittance drop off 30% is shown for the optical window, with tuning capabilities greater than 30% upon actuation of the cantilever. Unlike typical mechanical attenuators, these devices are not restricted to binary optical states. Optical modulation of the optical window is demonstrated with an oscillating electrical input. This produces a transmittance signal that oscillates around an average value within the range off VO2 ’s phase transition. For an input current signal with fixed amplitude (fel = 0.28 Hz), tuned to be at the onset of the phase transition, a transmittance modulation of 14% is shown. Similarly, by modulating the DC-offset, a transmittance modulation of VO2 along the hysteresis is obtained. To my family and friends... and for those who enjoy a little bit of light reading. iv ACKNOWLEDGMENTS I would like to thank my advisor and friend Dr. Nelson Sepúlveda for taking a chance on me and giving me the opportunity to be part of his Applied Materials Group. I am forever grateful for this opportunity. Dr. Sepúlveda has been a great mentor, influence and friend during my academic and professional development. To the past and present members of the Applied Materials Group, thank you. Dr. David Torres, Dr. Tongyu Wang, Dr. Yunqi Cao and Dr. Wei Li, thank you for accepting me in your group and helping during that first year. To the current members of the group, Juan, Henry and Ian, thank you all for the countless advice, help and especially the good times we had. I will surely miss all the laughs and jokes we shared, to you all I am grateful. I would also like to express my gratitude to the members of my graduate committee; Dr. John Albrecht, Dr. Qi Hua Fan and Dr. Chong-Yu Ruan. I would also like to thank the people over at Wright Patterson Air Force Research Labo- ratories, especially Dr. David Torres, Dr. Harris Hall and Dr. LaVern Starman. Thank you for the guidance and help during the design and fabrication of the devices. Finally, to my family members. My mother Virginia Soto, my father José Figueroa and my sister Cristina Figueroa. Thanks for supporting my endeavors, teaching me how to be a responsible person, taking pride in my accomplishments and above all, supporting me along the way. This work is for you. v TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix CHAPTER 1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Problem Description and Motivation . . . . . . . . . . . . . . . . . . . . . . 3 1.2 Thesis Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3 Research Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4 Dissertation Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 CHAPTER 2 BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1 Vanadium Dioxide (VO2 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.1 VO2 Crystal Structure . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.2 Optical and Electrical Transition of VO2 . . . . . . . . . . . . . . . . 9 2.2 Deposition Techniques for VO2 . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.2.1 Pulse Laser Deposition (PLD) . . . . . . . . . . . . . . . . . . . . . . 12 2.3 Emissivity: A Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.4 Smart Window Designs and Emissivity Modulation . . . . . . . . . . . . . . 15 2.5 VO2 -based Actuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.6 Optical Shutters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 CHAPTER 3 DESIGN AND DEVICE FABRICATION OF VO2 -BASED WIN- DOWS, CANTILEVERS, AND MEMS TUNABLE OPTICAL SHUT- TERS AND EXPERIMENTAL SETUP . . . . . . . . . . . . . . . . 29 3.1 Design and Fabrication of Monolithically Integrated VO2 -based Micro Win- dows and Experimental Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.1.1 Fabrication Flow Process of Monolithically Integrated VO2 -based Micro Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.1.2 Experimental Setup for Monolithically Integrated VO2 -based Micro Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2 Design and Fabrication of VO2 -based Shape Converting Devices . . . . . . . 34 3.2.1 Fabrication Flow Process of VO2 -based Shape Converting Devices . . 34 3.2.2 Experimental Setup for VO2 -based Shape Converting Devices . . . . 36 3.3 Design and Fabrication of VO2 -based Cantilever Devices . . . . . . . . . . . 37 3.3.1 Fabrication Flow Process f VO2 -based Cantilever Devices . . . . . . . 37 3.3.2 Experimental Setup for VO2 -based Cantilever Devices . . . . . . . . 39 3.4 Design and Fabrication of VO2 -based Tunable Optical Shutters and Exper- imental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.4.1 Fabrication Flow Process of VO2 -based Tunable Optical Shutters . . 41 3.4.2 Experimental Setup for VO2 -based Tunable Optical Shutters . . . . 43 vi CHAPTER 4 PROGRAMMING EMISSIVITY ON FULLY INTEGRATED VO2 WINDOWS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1.1 Electro-Optical States . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1.2 Programmability of Emissivity States in VO2 . . . . . . . . . . . . . 49 4.1.3 Thermal Images of VO2 Window’s . . . . . . . . . . . . . . . . . . . 52 4.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 CHAPTER 5 A SIMPLIFIED APPROACH FOR OBTAINING OPTICAL PROP- ERTIES OF VO2 THIN FILMS, AND DEMONSTRATION OF IN- FRARED SHAPE-SHIFTING DEVICES . . . . . . . . . . . . . . . 54 5.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 5.1.1 Optical States and Emissivity of VO2 . . . . . . . . . . . . . . . . . . 54 5.1.2 Two-Surface Reflectance and Transmittance Model for VO2 films. . . 59 5.1.2.1 Two-Surface Incoherent Reflectance and Transmittance Model for VO2 films. . . . . . . . . . . . . . . . . . . . . . . . . . . 63 5.1.3 Implementation of VO2 Windows for Shape-Shifting Devices . . . . . 65 5.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 CHAPTER 6 DEFLECTION MODEL FOR ELECTRO-THERMALLY ACTU- ATED VO2 -BASED CANTILEVERS . . . . . . . . . . . . . . . . . 72 6.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.1.1 Effective Length of Cantilever . . . . . . . . . . . . . . . . . . . . . . 73 6.1.2 Heat-Transient Deflection Model . . . . . . . . . . . . . . . . . . . . 74 6.2 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 CHAPTER 7 VO2 -BASED MICRO-ELECTRO-MECHANICAL TUNABLE OP- TICAL SHUTTER AND MODULATOR . . . . . . . . . . . . . . . 81 7.1 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 7.1.1 Heat Distribution Simulation Results and Thermal Imaging Results . 81 7.1.1.1 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . 81 7.1.1.2 Thermal Images . . . . . . . . . . . . . . . . . . . . . . . . 82 7.1.2 Performance of VO2 -based tunable optical shutters . . . . . . . . . . 84 7.1.3 VO2 -based electro-optical tunable modulator . . . . . . . . . . . . . . 88 7.2 Curvature and Deflection Analysis for a Trimorph Cantilever . . . . . . . . . 92 7.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 CHAPTER 8 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 8.1 Summary of Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 8.1.1 Problems Solved in this Thesis . . . . . . . . . . . . . . . . . . . . . . 98 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 vii LIST OF TABLES Table 5.1: VO2 parameters for insulating phase used for FEM simulation. . . . . . . 66 Table 6.1: VO2 and SiO2 parameters used for the transient-heat-deflection model . . 76 Table 7.1: VO2 and SiO2 parameters used for the trimorph curvature calculation. . . 94 viii LIST OF FIGURES Figure 2.1: (a) Monoclinic structure (M1) of VO2 . (b) Rutile structure of VO2 . Red denotes vanadium atoms and blue denotes oxygen atoms. [93] . . . . 8 Figure 2.2: Unit cell of VO2 crystal. Dashed lines represent the monoclinic cell while solid lines represent the tetragonal cell.[4] . . . . . . . . . . . . . . 9 Figure 2.3: Band structure for VO2 near the Fermi level for insulating and metallic phases. [97] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 2.4: Wavelength dispersions of (a) refractive index and (b) extinction coeffi- cient at various temperatures between 25 and 120o C. Solid and broken curves indicate the data obtained during the rise and drop in tempera- ture, respectively. The insets show the temperature dependences of the optical constants at wavelengths of 500, 1000, and 1500 nm. Closed and open symbols indicate the data obtained during the rise and drop in temperature. Copyright (2007) The Japan Society of Applied Physics [45] 11 Figure 2.5: Schematic for PLD system used for VO2 deposition. [116] . . . . . . . . 13 Figure 2.6: (a) Emitted infrared energy for a blackbody as a function of wavelength. (b) Diagram depicting the incident, reflected, absorbed and transmitted energy on an object. Copyright 2018 Optotherm, Inc.[117, 118] . . . . . 14 Figure 2.7: Negative thermal emittance of the VGC film. (a) The schematic setup for measurement of thermal emissivity of the VGC film. (b-i) Thermal images recorded by the thermal camera during the temperature cycling. (j) The difference of IR temperature between the VGC film (TVIR GC ) IR and the background (TB ). The emissivity in the camera is set to 0.95 in panels (b-j). (k) The temperature-dependent emissivity of the VGC film on black tape and the background coating. The arrows in panels (j) and (k) indicate the temperature cycling loop [92]. . . . . . . . . . . . 17 Figure 2.8: (a) Emitted power of the VO2 -sapphire sample integrated over the 8- to14-µm atmospheric transmission window for heating (solid line) and cooling (dashed line), compared to the emitted power from black soot. (b) The integrated emissivity of the VO2 -sapphire sample over the 8- to14-µm wavelength range. (c) Infrared camera images of the sample (diameter = 1 cm) for increasing temperatures [122]. . . . . . . . . . . . 18 Figure 2.9: Multilayered structure for smart window[136]. . . . . . . . . . . . . . . . 18 ix Figure 2.10: Sketch of the VO2 /Ag and VO2 /Cu multilayers structures. The number of layers is here, for example, N=7[137]. . . . . . . . . . . . . . . . . . . 19 Figure 2.11: Tunable optical cavity. (a)Schematic of the plasmonic device with hexag- onal array of period = 4 µm and diameter = 1.64 µm. FTIR scan gen- erated images of the plasmonic surface acquired for (b) semiconducting (T = 295 K), (c) phase-separated (T = 320 K), and (d) metallic (T = 360 K) states of VO2 . Albert Einstein image is 1.3 x 1.7 mm2 [139]. . . . 19 Figure 2.12: Infrared emissivity vs temperature from front (a) and rear (b) surface of the 190 nm VO2 film on Si substrate. Red lines are calculated emissivity during heating cycle, blue lines are calculated emissivity during cooling cycle. Continuous lines are theoretical fit by using Maxwell Garnett effective medium theory[140]. . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 2.13: (Left) Experimental optical system used to program an image onto a VO2 thin film. (Right)Programmed and projected NIR images. (a) 55o C before laser scanning. (b) During writing laser scanning. (c)Right after writing laser scan finished. (d) Programmed image 5 minutes after scanning[128]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 2.14: VO2 -based cantilever with integrated heaters for actuation and elec- trodes to measure the resistance change of VO2 . Top plot shows the input voltage and deflection. [40]. . . . . . . . . . . . . . . . . . . . . . . 22 Figure 2.15: (a) Fabrication process of SWNT/VO2 -based cantilever actuator. (b) Top view microscope images of SWNT/VO2 and bare VO2 based can- tilevers. Zoom imaged of SEM cross-section of the SWNT/VO2 -based cantilever. [65] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 2.16: VO2 -based MEMS mirror device. (Top) SEM image with colored bi- morph legs. (Bottom) Top view of VO2 -based MEMS device[18]. . . . . . 23 Figure 2.17: Planar nanomechanical actuator with chevron-type geometry based on VO2 phase transition. [42] . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 2.18: GeTe phase change shutter on a glass substrate. Grey areas represent GeTe and gold slits represent the heater. [150] . . . . . . . . . . . . . . . 25 Figure 2.19: Design for the PCM shutter. (a) Blue represents the Ge2 Sb2 Te5 PCM, yellow represents the silica substrate. (b) Tiled filter pattern for the device. [156] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 x Figure 2.20: LC shutter (a) Real photo of sample S1 at P state (b) Polarization optical microscopy image of P state. (c) Real photo of sample S1 at FC state. (d) POM image of sample S1 at FC state. [160] . . . . . . . . . . 26 Figure 2.21: Schematic for the IR shutter structures. (a) Three layer structure based on ITO/VO2 /ITO with Pt pads that act as electrical contacts. (b) Two layer structure with the ITO bottom layer, top VO2 layer and Pt pads. [88] 27 Figure 2.22: (a) Schematic of micro-optical switch device based on W-doped VO2 . (b) Transmittance modulation (λ= 1550 nm) for the W-doped VO2 - based micro-optical switch device. [161] . . . . . . . . . . . . . . . . . . 27 Figure 2.23: (a) Cross-section view of the VO2 -based light shutter. (b) Deposited Au circuit on top of VO2 film. [87] . . . . . . . . . . . . . . . . . . . . . 28 Figure 3.1: Fabrication process for the VO2 based windows. (a) SiO2 substrate. (b) metallization of heater and electrodes. (c) SiO2 insulating layer. (d) opening of the electrodes. (e) VO2 deposition and window patterning. (f) opening of contact pads for electrical connections. . . . . . . . . . . 30 Figure 3.2: (a) SEM images for the VO2 -based windows. (b) SEM image for the 400 µm2 window. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Figure 3.3: Electro-Optical setup used to measure VO2 ’s resistance and transmit- tance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 3.4: OptoTherm, Infrasight MI320 infrared camera setup to measure VO2 ’s thermal distribution and emissivity. . . . . . . . . . . . . . . . . . . . . . 33 Figure 3.5: (A) Fabrication process for the VO2 based window. (a) 500 µm thick SiO2 substrate, (b)Deposited VO2 layer. (c) Patterning of VO2 win- dows, top windows are 1100 µm x 500 µm, lower windows are 650 µm2 . (d) Insulating layer of SiO2 , (e) Au/Ti metal deposition. (f) Pattering of metal traces. B) Optical microscope image of die containing VO2 based devices. (C) SEM images for: (1) 1100 µm x 500 µm, (2) 650 µm2 with circle and square heater configuration, (3) 650 µm2 with tri- angle heater configuration.(4) Cross-section for the 1100 µm x 500 µm device with respective thickness. (D) Cross-section for the 1100 µm x 500 µm window showing the VO2 thickness, SiO2 top layer thickness and gold thickness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 xi Figure 3.6: (A) Fabrication flow process for the VO2 -based cantilevers. (a) A 1 µm layer of SiO2 is deposited over a 500 µm thick Si wafer via PECVD. (b) Deposition and etching of a 150 nm/50 nm layer of Pt/Ti that will act as the heaters for the cantilevers. (c) A 1 µm layer of SiO2 is deposited over the metal structure.(d) Plasma etching is performed to pattern the SiO2 . (e) The structure is released via XeF2 etching. (f) A layer of VO2 is deposited via PLD over the release structure. Dashed red line shown in (f) represents where the sample was cut to image the cross-section in the SEM. (B) SEM picture of the cross-section of the final device showing each of the different layers that made up the device. A 100 nm layer of VO2 is observed followed by the 1 µm layer of SiO2 and 200 nm layer of Pt/Ti. (C) Shows a top/side view of the 550 µm x 50 µm and 450 µm x 50 µm cantilevers. . . . . . . . . . . . . . . . . . . . . . . . . . 39 Figure 3.7: (a) Optical setup with high-speed camera to record deflection. (b) Re- sistance vs Temperature curve for the 450 µm x 50 µm cantilever. (c) 450 µm x 50 µm cantilever in equilibrium state before actuation. (d) 450 µm x 50 µm after actuation with an input current of 4.8 mA . . . . 40 Figure 3.8: (a) Fabrication flow process for the VO2 -based optical shutter. Cross- section views correspond to the dotted blue line shown in figure 3.8-d. (1) 500 µm Si wafer. (2) Deposition of a 1 µm layer of SiO2 by PECVD. (3) Deposition of VO2 layer by PLD. (4) Patterning and etching of VO2 layer. (5) Deposition of a 300 nm layer of SiO2 by PECVD. (6) Pattern- ing and etching of second SiO2 layer. (7) Evaporation and metal lift-off of Au/Cr layer. (8) Deposition, patterning, and etching of a 200 nm layer of SiO2 . (9) Structure release by XeF2 etching. (b) Cross-section of the VO2 based cantilever showing the top and bottom SiO2 layer thickness and VO2 layer thickness. (c) Side angle view for a released VO2 based shutter (d) Top view SEM image for an un-released shutter. Dotted lines represent cross-section cuts. . . . . . . . . . . . . . . . . . . 43 Figure 3.9: (a) VO2 -based optical shutter experimental setup. Blue, solid line repre- sents the actuation current for the window. Red, dashed line represents the actuation current for the cantilever (tilt control). Black arrow rep- resents direction of actuation upon Joule heating (figure is not to scale) . (b) Top close up view for the cantilever/window structure. Squares represent release holes for ease of etch. Orange, red and yellow spots represents where the beam is focused on the window as seen on (a). . . 45 Figure 4.1: Simultaneous measurements for the electrical (a) and optical (b) tran- sition in the 400 µm VO2 window. . . . . . . . . . . . . . . . . . . . . . 46 xii Figure 4.2: Electrical (a) and optical transition (b) minor loops in the 400 µm VO2 window. Programming pulse is over-imposed. Inset shows the voltage input used to obtain the minor loops. . . . . . . . . . . . . . . . . . . . . 48 Figure 4.3: Time constant measurement for the 400 µm VO2 window. . . . . . . . . 49 Figure 4.4: Emissivity as a function of temperature (a) and current (b) for the 400 µm window. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Figure 4.5: Curves for mapping method used to obtain the emissivity as a function of current.(a) Temperature of VO2 window as a function of temperature of tape = temperature of heater. (b) Temperature of heater as a function of actuation current. (c) Temperature of tape = temperature of heater as a function of actuation current. . . . . . . . . . . . . . . . . . . . . . 51 Figure 4.6: Thermal image for the 400 µm VO2 window. (a) before programming pulse. (b) after programming pulse. . . . . . . . . . . . . . . . . . . . . . 52 Figure 5.1: Optical transmittance measurements for the 650 µm2 VO2 window. (b) shows the minor hysteretic loops for the same window. Measurement was taken by actuating one of the single triangle heaters and measuring inside the area of the heater. Inset on (b) shows the voltage sequence used to obtain the minor loops. . . . . . . . . . . . . . . . . . . . . . . . 56 Figure 5.2: Emissivity for VO2 windows: (a),(b): emissivity of 650 µm2 VO2 win- dow with etched SiO2 (exposed VO2 ) as a function of temperature and current, respectively. (c): emissivity for the 650 µm2 VO2 window with top layer of SiO2 over VO2 . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Figure 5.3: Curves for mapping method used to obtain the emissivity as a function of current for the 650 µm2 VO2 window.(a) Temperature of VO2 win- dow as a function of temperature of tape = temperature of heater. (b) Temperature of heater as a function of actuation current. (c) Tempera- ture of tape = temperature of heater as a function of actuation current. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Figure 5.4: (a) Transmittance as a function of wavelength. (b) Transmittance as a function of wavelength below and above transition temperature for several references. (c) Reflectance as a function of wavelength. (d) Absorbance as a function of wavelength. . . . . . . . . . . . . . . . . . . 62 Figure 5.5: Schematic for the assembly showing incident reflection and transmit- tance for both sides of the stack. . . . . . . . . . . . . . . . . . . . . . . 64 xiii Figure 5.6: (a) Transmittance as a function of wavelength. (b) Reflectance as a function of wavelength. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Figure 5.7: Temperature distribution for the 1100 µm x 500 µm VO2 window as simulated using the electric currents and Joule heating modules with an input current of 35 mA. . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Figure 5.8: (A) Thermal images for the 650 µm2 window with square and circle heater configuration for several programming pulses.(1) Corresponds to the pre-heated state using the square heater. (2) A second pulse is sent to achieve a lower emissivity state. (3) The second circular heater is activated, this creates a new circular state. (4) The current is increased on the circular heater to obtain a new square state. (5) The second heater is turned off with only the first heater maintained at the second state (2). A state with a lower emissivity but same temperature as in (2) is achieved.(6) The first heater is brought back to the pre-heated level. (B) Emissivity curve for a heating and cooling cycle. Labeled emissivities in the plot correspond to the values for each of the thermal images. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Figure 5.9: (A) Thermal images for the 650 µm2 window with triangle and triangle heater configuration for several programming pulses.(1) Corresponds to the pre-heated state using the triangle heater. (2) A second pulse is sent to achieve a lower emissivity state. (3) The second triangle heater is activated, this creates a new square state. (4) The current is increased on the triangle heater to obtained a new square state. (5) The second heater is turned off with only the first heater maintained at the second state (2). A state with a lower emissivity but same temperature as in (2) is achieved. (6) The first heater is brought back to the pre-heated level. (B) Emissivity vs current showing the corresponding states. . . . . 69 Figure 5.10: Thermal images for the 1100 µm x 500 µm VO2 window. (a) shows the pre-heated current necessary to form a number one figure. (b) a second pulse is applied to reach a lower emissivity state. (c) Image after second pulse is brought down to the pre-heated level. (d) shows the pre-heated current necessary to form a number seven figure. (e) second pulse is applied to reach a lower emissivity state. (f) Image after second pulse is brought down to the pre-heated level, a more defined seven can be seen on the lower irradiance area. . . . . . . . . . . . . . . . . . . . . . . . . 70 Figure 6.1: Deflection as a function of time: (a) is obtained directly from the high- speed video. (b) as calculated from equation (6.7). . . . . . . . . . . . . 77 xiv Figure 7.1: Temperature distribution for the VO2 -based optical shutter due to a current input. (a) Cantilever actuation current of 6 mA and window actuation current of 0 mA.(b) Cantilever actuation current of 0 mA and window actuation current of 16 mA. (c) Cantilever actuation current of 6 mA and window actuation current of 10.5 mA. da) Cantilever actuation current of 6 mA and window actuation current of 16 mA. . . . . . . . . 83 Figure 7.2: Temperature distribution for the VO2 -based optical shutter. (a) Can- tilever actuation current of 6 mA and window actuation current of 0 mA.(b) Cantilever actuation current of 0 mA and window actuation current of 16 mA. (c) Cantilever actuation current of 6 mA and window actuation current of 10.5 mA. d) Cantilever actuation current of 6 mA and window actuation current of 16 mA. . . . . . . . . . . . . . . . . . . 84 Figure 7.3: Transmittance Measurements upon actuation of window and/or can- tilever: (a) Steady-state transmittance results for separate actuation; i.e. actuation of resistive heaters for window or cantilever (tilt control). (b) Results for window actuation, followed by cantilever actuation with a current step of 6 mA. Insert shows the current actuation step supplied to the cantilever traces. (c) Transmittance minor loops due to actua- tion of the window heater. Insert shows the voltage input used to trace the minor loops. (d) Results for cantilever actuation (with a current step of 6 mA) after partial actuation of the window (x-axis). Dashed lines represent the transition points where the cantilever was actuated. Solid lines represent overall transmittance change. Legend shows the voltage/current point during the window hysteresis where the current step was applied. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Figure 7.4: (a) Transmittance minor loops due to actuation of the window heater. Red lines represents limits of voltage values used for modulation. Green line represents the current value used at the onset of the phase tran- sition. Insert shows the input current for modulation. A preheated current of 10.5 mA (4.2 V) is hold on the device (onset of phase transi- tion), then a sine wave with an amplitude of 1 V and frequency of 0.28 Hz is sent. (b) Modulated transmittance of the optical window due to the current input. (c) FFT for input current and output power for the VO2 optical window. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Figure 7.5: (a) Input current with modulated DC offset and constant voltage am- plitude with an input frequency of 1 Hz. Modulation is done for both heating and cooling cycles with an increment on the DC offset of 0.2 V (0.6 mA). (b) Transmittance with modulated DC offset and constant voltage amplitude (0.7 V) with an input frequency of 1 Hz. Transmit- tance is modulated for both heating and cooling cycles and it starts at the onset of the PT at 4.5 V or 10.5 mA. . . . . . . . . . . . . . . . . . . 92 xv Figure 7.6: Schematic of VO2 -based trimorph cantilever. . . . . . . . . . . . . . . . 92 xvi CHAPTER 1 INTRODUCTION Vanadium Dioxide (VO2 ) is a phase change material whose properties present many oppor- tunities for applications in the fields of material science and condensed matter physics [1, 2]. As first observed by Morin in 1959 [3], VO2 undergoes a particular phase change transfor- mation in its structural, electronic and optical properties above its transition temperature of TP T = 340.15 K (68 o C) [4]. This phase change can be induced via thermal [5], optical [6], electrical [7] and applied stresses and pressure [8, 9] . Upon this phase change, VO2 ’s crystal structure rapidly changes from monoclinic to a tetragonal phase. Properties such as thermal conductivity, electrical and magnetic susceptibility, Young’s modulus and index of refraction change [10, 11]. The dynamics of this phase change have been the subject of several studies [12, 13, 14, 15, 16], but a complete understanding of the physical nature has yet to be made. Due to these changes of the material properties along the PT, VO2 is a suit- able candidate for different applications such as frequency resonators [17], MEMS devices [18, 19], VO2 -based actuators [20, 21], switches and shutters [22, 23] and smart windows and coatings [24, 25, 26] among others. Even though there are other phase change materi- als, such as inorganic-organic hybrids like [(CH3 )4 P][FeCl4 ] and [(CH3 )4 P][FeBr4 ] [27], and materials like GeTe and Ge2 Sb2 Te5 [28]. Their transition temperature is still higher than that of VO2 (closest to room temperature 68 o C), with the inorganic-organic hybrids having a phase transition temperature occurring at 353 K (79.85 o C) and 359 K (85.85 o C), and GeTe transition temperature being in the range of 493.15 K to 528.15 K (220 o C to 255 o C) [29]. Although research is still being conducted on these phase change materials [30, 31], VO2 ’s low transition temperature makes it one of the most suitable options due to its low power consumption, PT close to room temperature and dramatic change in the materials properties [32]. This work presents the integration of VO2 thin films into smart optical windows for 1 emissivity modulation and shape-converting devices [33, 34], cantilever devices where VO2 is integrated in the structure for actuation purposes to achieve deflections and tunable optical shutters for transmittance modulation. Previous research has successfully incorporated VO2 thin films into smart window designs [35, 22, 36, 37, 38, 39, 25] and mechanical structures [40, 41, 42]. Of important notice is the fact that due to its intrinsic phase transition, VO2 ’s change in its optical properties is far more noticeable in the infrared region [43, 44, 45]. Since its transmittance decreases as the temperature is increased, this would equate to a change in the materials emissivity through the phase transition. This would imply that upon observing the material under an IR camera, the material will look colder as the temperature is increased. The fact that this phase transition occurs within a small range of 10 o C which would imply a low power consumption, makes VO2 a suitable candidate for smart windows, cloaking device integration and optical shutters for sensor protection applications. Besides smart windows, Vanadium Dioxide has also been integrated into micrometer sized devices, which can be easily integrated on small scale chips for a variety of applications. Previous research has shown the integration of VO2 into MEMS mirror devices [18, 46, 47, 48, 49] and micromanipulators [50, 51]. For these devices, the main component for actuation is a cantilever-type structure that is composed of one or more materials with different coefficients of thermal expansion. This will cause a strain difference upon actuation, thus a bending moment will be generated. In order to achieve deflections on these devices, an actuation mechanism is required. This can be achieved via electrostatic [52, 53, 54, 55], electrothermal [56, 57, 58], electromagnetic [59, 60], piezoelectric [61, 62, 63, 64] and photo-thermal actuation [21, 65, 66, 67]. For this work, we present the integration of VO2 into bimorph cantilevers that are actuated via integrated resistive heaters and a transient-temperature deflection model is developed. This model can be applied to any bimorph structure with integrated resistive heaters for electro-thermal actuation. Over the last decades, devices such as optical shutters, variable optical attenuators and optical modulator technologies have seen an expansion with regards to optical communica- 2 tions networks, fiber optics and integrated circuit fabrication technologies [68, 69, 70, 71, 72, 73]. Several technologies and designs for optical shutters range from micro shutter/cantilever arrays or matrices that block incoming light [74, 75, 76], thermally switched reflective mirror plates [77], to liquid crystal or dye based shutters [78, 79]. Most of these optical shutter de- signs can be adapted to different systems such as spectrometers [80], reconnaissance cameras [81], high precision experimental instruments [82], sensor protection [83, 84] and systems where optical fiber protection is required [85, 86]. Vanadium Dioxide based smart window designs have been previously implemented to fabricate mechanically static optical tunable shutters [87, 88, 23, 89]. Many of these devices consist of a single layer of VO2 over a substrate with a metallic trace on top for actuation via Joule heating. Upon actuation, modulation of the optical transmittance signal has been shown to be above 25 % for IR wavelengths ( λ = 1550 nm) [88]. In this work we present the fabrication of mechanical-optical VO2 -based tunable shutters. The device consist of the combination of the aforementioned VO2 smart windows and mechanical cantilevers to create a shutter that works as a regular mechanical shutter and also as a tunable optical shutter. 1.1 Problem Description and Motivation For this work VO2 has successfully been integrated into smart windows designs that are actuated via electro-thermal actuation for transmittance and emissivity modulation. A step-by-step guide on the design and fabrication of these devices is given. Characterization of these VO2 devices is done and demonstration of its ability to be integrated for cloak- ing and shape-converting devices is shown. Similarly the fabrication of VO2 -based bimorph cantilevers is shown and an temperature transient model for deflections is developed. Fi- nally, by combining both previous designs, MEMS tunable optical shutters where designed and successfully fabricated. The fabricated devices fully demonstrate their intrinsic tuning capabilities due to the VO2 implementation. The problems addressed in this work are as follows: 3 • Integration of VO2 on smart window designs with integrated resistive heaters that would allow for modulation of transmissivity and emissivity. • Characterization of both electrical resistance and optical transmittance of VO2 -based windows for a laser with wavelength of λ=1550 nm. • Demonstration of VO2 ’s emissivity modulation via computer controlled electronic pulses to achieve IR cloaking using a thermal camera. • Development of a two-surface transmittance, reflectance and absorbance model is shown that takes into account the effects of the film and substrate thickness on the optical properties of the device without the need to measure the materials optical constants. • Demonstration of VO2 -based smart windows ability to reconfigure thermal images into different geometrical shapes or numbers. • High-speed recordings of cantilever deflections are attained. Analysis for deflection is obtained from high-speed videos. • Development of a transient-temperature deflection model for VO2 -based cantilever bimorphs. Deflections are compared to experimental data. • Development and integration of VO2 -based smart windows and VO2 -based cantilevers into a MEMS tunable optical shutter design. • Modulation of the transmitted power is demonstrated like a regular binary shutter by actuation of the cantilever structure. • Actuation of the optical window allows for modulation of the transmitted power within the boundaries of the hysteresis of VO2 . This creates a tunable mechanical-optical shutter, no longer being a binary shutter. 4 In order to address all these problems, designs, fabrications, simulations, experiments and analysis where carried out to present solutions to these problems. By following the same fabrication guidelines provided in the following chapters, similar devices can be fabricated, and characterization and mathematical models can be implemented for future device designs that could implement other phase change materials. 1.2 Thesis Statement The major contributions of this work are the development of VO2 -based smart windows for optical modulation and the introduction of MEMS tunable optical shutters. Studies have previously reported on the implementation of VO2 for smart window integration and optical modulation [22, 36, 90, 35, 91, 92]. This study further expands the previous research by adding integrated resistive heaters for actuation via Joule heating in order to demonstrate the device optical modulation capabilities. Furthermore, this allows to demonstrate the ability to camouflage and shape-shift objects in the near infrared by exploiting VO2 ’s intrinsic emissivity change. By arranging the integrated heaters into different geometrical shapes, shape-converting becomes feasible due to the intrinsic properties of VO2 . Another major contribution is the design and fabrication of the first VO2 -based MEMS tunable optical shutters. Regular shutters either have to forms of operation, by either blocking the incoming light beam or not being on the free path of the beam. By incorporating a VO2 -based mechanical shutter with a VO2 -based optical window, a tunable mechanical-optical shutter is created capable of optical modulation within the boundaries of VO2 ’s phase transition. Thesis Statement: Vanadium dioxide (VO2 ) is integrated in the design of smart win- dows and tunable mechanical-optical shutters for infrared cloaking, shape-converting and op- tical modulation. 5 1.3 Research Contributions Step by step guide on the fabrication of VO2 smart window designs is provided for future references. By understanding the steps involved in the fabrication of the devices, any geo- metrical shape can be designed and implemented into future designs for implementation into shape-converting devices. By knowing the materials used for the design and fabrication of smart windows, the two surface optical model can be applied to obtain the optical properties of the device for characterization. Furthermore, the design and fabrication of the VO2 -based optical shutters is provided as a foundation for fabricating and further expanding on the design. 1.4 Dissertation Outline The structure of this dissertation is as follows: Chapter 2 goes into an in-depth dis- cussion on VO2 ’s phase transition, optical and electrical transition. This is followed by a description and review of emissivity, smart window design and emissivity modulation. Fi- nally a review on VO2 -based actuators and optical shutters is given. Chapter 3 presents a thorough description of the design and fabrications for the VO2 -based windows for emissiv- ity modulation. The experimental setup used to simultaneously measure the transmittance and resistance of the VO2 -based windows and thermal scope for emissivity modulation is described. The design and fabrication for the VO2 -based shape shifting devices is then pre- sented followed by the experimental setup used for measurements. A description on the VO2 -based cantilevers design and fabrication is shown, and further explanation for the high- speed camera setup to monitor deflection is presented. Finally, an in-depth description on the design and fabrication process for the MEMS tunable optical shutters is given. This is followed by presenting the experimental setup used for measurements of the optical trans- missivity for the optical shutters. Chapter 4 presents the results and analysis for the micro variable optical windows based on VO2 thin films. Actuation via resistive heating allows for the programmability of both transmittance and emissivity states. Using an infrared camera, 6 thermal images for the device are shown, which allow for the creation of a cloaking device based on emissivity modulation. In Chapter 5 further developments for VO2 -based optical windows allow for the demonstration of shape shifting devices. Several demonstrations are shown for different shape-converting capabilities exploiting the intrinsic phase transition of VO2 via electronically programmed pulses. A two surface reflectance, transmittance and absorbance model is developed. This model does not require direct measurements for the optical properties of the material and still provides a valid model that can be used to char- acterize thin films. Chapter 6 presents VO2 /SiO2 -based cantilevers that are actuated via Joule heating. Using a high-speed camera, deflections that occur within a time range of 100 µs can be monitored. This experimental data can be compared to data provided by a transient temperature-deflection bimporh model that takes into account the changes in VO2 properties along the phase transition. Chapter 7 presents the development and fabrication of VO2 -based MEMS tunable optical shutters. Finally, Chapter 8 shows a summary of the contributions presented on this work. 7 CHAPTER 2 BACKGROUND 2.1 Vanadium Dioxide (VO2 ) 2.1.1 VO2 Crystal Structure As previously mentioned in Chapter 1, VO2 possesses a monoclinic structure at room tem- perature and undergoes a phase transition above T=340 K (68 o C) which causes its crystal structure to change to a tetragonal (rutile) structure. Figure 2.1-a shows the monoclinic M1 structure of VO2 with crystallographic space group P 21 /c (C52h No. 14) [93]. From Longo et al. [94] the lattice constants and reported values for angles are aM1 =5.7517 Å, bM1 =4.5378 Å, cM1 =5.3825 Åand βM1 =122.646o . Figure 2.1-b shows the rutile structure of metallic VO2 with a tetragonal lattice of space group P 42 /mnm (D14 4h No. 136) [93]. Figure 2.1 : (a) Monoclinic structure (M1) of VO2 . (b) Rutile structure of VO2 . Red denotes vanadium atoms and blue denotes oxygen atoms. [93] The lattice parameters and the internal oxygen parameters are given by at = bt =4.58 Å, 8 ct =2.8528 Å, u=0.30000 with all axial angles measuring 90o as presented by Wentzcovitch et al. [95]. Figure 2.2 shows the crystallographic planes for monoclinic VO2 . Once the PT is induced, one has to take into account the small twist of the unit cell, the at and the ct directions from the figure coincide with bm and am . As can be noted from the figure, the am , cm , bt and ct axes are located on the same plane (010). Figure 2.2 : Unit cell of VO2 crystal. Dashed lines represent the monoclinic cell while solid lines represent the tetragonal cell.[4] 2.1.2 Optical and Electrical Transition of VO2 Vanadium Dioxide (VO2 ) temperature driven phase transition is one of the most studied MIT’s, mainly for its technological applications. One major part of this is due to the fact that it occurs close to room temperature (68 o C) [32]. At room temperature, VO2 behaves 9 as a semiconductor and possess an optical bandgap of ∼ 0.7 eV [96]. Above 68 o C, VO2 undergoes a phase transition into a metallic state. As a consequence of this the band gap collapses and its resistivity drops by up to five orders of magnitude. Besides the temperature driven transition, VO2 undergoes a structural phase transition with the metallic phase having a rutile structure and the insulating a monoclinic structure as seen in Figure 2.1. Figure 2.3 : Band structure for VO2 near the Fermi level for insulating and metallic phases. [97] Figure 2.3 shows the accepted picture of VO2 ’s band gap during its insulating phase (opening) and metallic (closing) during the MIT phase transition [97]. For the low tem- perature monoclinic phase, vanadium atoms move in a zig-zag configuration, this in turn increases the energy of the π ∗ band. Another occurrence during the monoclinic phase is the creation of V-V dimers which are tilted along the c-axis. As a result of this the d|| band is split as seen in Figure 2.3-a [93]. For the rutile phase, the vanadium (V) atoms are surrounded by an octahedron configuration of oxygen (O) atoms. During VO2 ’s optical and electrical transition, two major mechanism have been proposed: The Peierls mechanism (electron phonon mechanism) and the Mott mechanism (electron correlation mechanism) [98, 99, 95, 100, 101]. Further studies have suggested that an intermediate phase coexists between the insulator and metal phases (68 o C to 75 o C) [98] and this first order MIT phase is not driven by the structural phase transition [102]. It is important to note that it has 10 been shown that the optical phase transition of VO2 is an extremely fast process occurring in the order of femtoseconds [103]. During the MIT, VO2 ’s optical constants (n, k) abruptly change (Figure 2.4) and VO2 ’s transmittance decreases and increases as the temperature rises and drops as shown by Kakiuchida et al. [45]. Figure 2.4 : Wavelength dispersions of (a) refractive index and (b) extinction coefficient at various temperatures between 25 and 120o C. Solid and broken curves indicate the data obtained during the rise and drop in temperature, respectively. The insets show the tem- perature dependences of the optical constants at wavelengths of 500, 1000, and 1500 nm. Closed and open symbols indicate the data obtained during the rise and drop in temperature. Copyright (2007) The Japan Society of Applied Physics [45] 11 2.2 Deposition Techniques for VO2 Deposition of VO2 can be achieved in a variety of different ways. Atomic layer deposition (ALD) [104], physical vapor deposition (PVD) [105], chemical vapor deposition (CVD) [96], sputtering [106, 107] and pulse laser deposition (PLD) [108, 4, 109, 110] processes are just a few ways of depositing VO2 . Since VO2 undergoes an insulator to metal phase transi- tion, with a drop on its electrical resistance by an order of 103 or 104 above its transition temperature. Then the crystal structure and stoichiometry of VO2 thin films plays a major role on the quality of the film [110]. There has been substantial research on trying to tune VO2 ’s transition temperature by introducing dopants with high valence metal ions into the lattice of VO2 [111, 112], or by introducing epitaxial strain by using RuO2 layers [113, 114]. While these additions manage to lower the transition temperature of VO2 , it comes with a degradation on its electrical properties. 2.2.1 Pulse Laser Deposition (PLD) PLD is a growth technique commonly used for the fabrication of thin films since the late eighties. A typical PLD chamber is shown in Figure 2.5. An intense laser beam pulse is passed through an optical window of a vacuum chamber and is focused into a target, either solid or liquid, where it is partially absorbed and reflected. Typically the ablation of the target is performed with a KrF excimer laser of wavelength λ =248 nm. Once the target is illuminated at a specific power density, a significant removal of the material occurs in the form of an ejected luminous plume. This ejection of such plume depends on the material being used, its morphology and the laser pulse and duration. Material ejected from the plume is deposited on a substrate, where film growth occurs. This film growth is usually accompanied by a passive or reactive gas or ion source which may affect the ablation plume species in the gas phase or the surface reaction [115]. 12 Figure 2.5 : Schematic for PLD system used for VO2 deposition. [116] 2.3 Emissivity: A Description All objects absorb, reflect, transmit and emit energy. An object that emits energy will have an associated intensity of radiation that will be a function of its temperature, wavelength and emissivity. For a blackbody or material that is known as a perfect emitter, the material will radiate 100% of the electromagnetic energy that it its absorbing [117]. The intensity of radiation for a blackbody as a function of temperature and wavelength is given by Planck’s blackbody equation. For a blackbody, as temperature increases, the intensity of radiation will increase. This will also cause the peak intensity of radiation to shift to shorter wavelengths as seen in Figure 2.6-a. A piece of metal at room temperature will not emit any visible light to the human eye. As the object’s temperature increases, then the object will start to glow red. Upon increasing the temperature (1500 o C), the piece of metal will glow white, consequently this case the metal will emit energy at all visible wavelengths [117]. The blackbody term is used to describe a perfect absorber, or a material that absorbs all visible electromagnetic energy, and therefore looks black. 13 Figure 2.6 : (a) Emitted infrared energy for a blackbody as a function of wavelength. (b) Diagram depicting the incident, reflected, absorbed and transmitted energy on an object. Copyright 2018 Optotherm, Inc.[117, 118] . The term emissivity describes the efficiency at which a material radiates infrared energy. A perfect emitter or a blackbody as an emissivity value of 1.00, which means that no other material can irradiate more thermal energy at a given temperature. Emissivity values range from 0.0 to 1.0, with 0 being an object that will not emit infrared energy [119]. Figure 2.6-b shows an image of an object that has incident infrared energy. When the energy is upon the material, be it is solid, liquid or gas, the material will exhibit the properties of absorption, reflection and transmittance. From conservation of energy, it follows that the amount of incident energy must be equal to the sum of the absorbed, transmitted and reflected energy such that: EIncident = EAbsorbed + ET ransmitted + ERef lected (2.1) Now for an object in vacuum at constant temperature, there will be no other sources of energy input or output from the object. Therefore, the absorbed energy will increase the object’s thermal energy, such that the transmitted and reflected energies will not. Since the object remains at constant temperature, then it must follow that [118]: 14 EEmitted = EAbsorbed (2.2) Plugging equation (2.2) into equation (2.1) and setting the incident energy equal to 100% we have: 100% = Emissivity + T ransmissivity + Ref lectivity (2.3) Where the emitted energy term has been changed to emissivity, since emissivity equals the efficiency at which a material emits energy, and similarly with transmitted and reflected energy. From equation (2.3) it can be noted that there is a balance between the terms. Increasing either parameter will require a decrease in the sum of the other two parameters. For most solids, transmission of infrared energy is very low. The majority of the incident energy is either absorbed or reflected. Therefore by setting transmissivity equal to zero, equation (2.3) can be rewritten as follows [118]: 100% = Emissivity + Ref lectivity (2.4) From equation (2.4), a simple relation can be found. For a given material, if its emissivity increases, then its reflectivity must decrease. The same for the opposite case. For most materials their emissive and reflective behavior is similar for the visible and infrared regions of the spectrum. For example, ceramic materials are good emitters, but have low reflectance. Likewise, metals are highly reflective but have a low emissivity associated with them [118]. 2.4 Smart Window Designs and Emissivity Modulation Adaptive camouflage, shape-converting technologies and emissivity modulation have been the subject of research in the recent years [92, 120, 22, 121, 37, 36, 122, 123, 124, 125, 25, 126, 39, 127, 128]. To this day, many materials and methods are still being researched, ranging from clothing improvements with fabrics and special dies [129, 130, 131] to materials with reversible color changes for camouflage [132]. Of special importance is infrared camouflage, 15 which can be achieved by emissivity modulation instead of direct control of the material’s temperature [92]. An active cloaking device based on VO2 has been previously reported by Xiao et al. [92]. A VO2 layer on graphene/carbon nanotube (CNT) thin film with negative differential emissivity is developed. As shown in Figure 2.7-k, via Joule heating the emissivity of the device can be changed and thus a rapid thermal camouflage can be achieved (Figure 2.7-b- f). Of most importance is the flexible nature of the films, which can be adapted to various surfaces enabling emissivity tuning with low power consumption. The dynamics of this emittance change during the transition have been further researched by Kats et al. [122]. During the phase transition it is found that the VO2 is in an intermediate state of insulator- metal where the VO2 layer comprises nanoscale islands of both phases. Upon heating, a large negative thermal emissivity is found (Figure 2.8-b), thus allowing for a direct measurement and extraction of both wavelength and temperature dependent emissivity, and consequently allowing for its use in applications for IR camouflage and thermal regulation as seen in Figure 2.8. Other studies have successfully developed VO2 nanoparticles encased with a ZnS shell [121] for emissivity modulation. By encasing the VO2 nanoparticles with a ZnS shell, the shell allows to modify the color of the VO2 core which adapts to visible camouflage with the background. Multi-layer smart window designs with added VO2 is another form of smart window design [133, 134, 135]. Figure 2.9 shows such design where the structure has 5 layers and is designed to sensibly adjust infrared transmittance to maintain temperature [136]. By designing VO2 multilayers (Figure 2.10) thermal emissivity in the IR can be tuned. This design allows to use the window as a low emissivity filter for high temperatures that can be applied for military purposes [137]. Zheng et al. [138] developed a TiO2 (R)/VO2 /TiO2 (A) multilayer that allows for the design of a smart window for thermal regulation. Such a device posses an antireflection layer with the TiO2 (R), the VO2 acts as a thermal regulator, and the top TiO2 (A) acts as a photocatalyst, thus allowing for a regulation efficiency of 10.2%. 16 Figure 2.7 : Negative thermal emittance of the VGC film. (a) The schematic setup for measurement of thermal emissivity of the VGC film. (b-i) Thermal images recorded by the thermal camera during the temperature cycling. (j) The difference of IR temperature between the VGC film (TVIR IR GC ) and the background (TB ). The emissivity in the camera is set to 0.95 in panels (b-j). (k) The temperature-dependent emissivity of the VGC film on black tape and the background coating. The arrows in panels (j) and (k) indicate the temperature cycling loop [92]. Wavelength selective IR camouflage systems capable of operating at specific IR bands have also been developed [139]. By triggering VO2 ’s phase transition, the reflection spectra is modified and an adaptive camouflage is attained. Figure 2.11 shows the device active thermal camouflage of multispectral infrared which could be used for adaptive camouflage and infrared tagging technologies. Leahu et al. has previously reported on the asymmetric tunability of emissivity for VO2 - based structures [140]. Their design consists of a 190 nm thin film of VO2 deposited over 480 µm Si wafer. By testing the emissivity from both sides of the film (VO2 side and Si side), it was found that the emissivity drop from the Si side is almost double (Er =0.47) than that of the VO2 interface. By applying a Maxwell Garnett effective medium approximation 17 Figure 2.8 : (a) Emitted power of the VO2 -sapphire sample integrated over the 8-to14-µm atmospheric transmission window for heating (solid line) and cooling (dashed line), compared to the emitted power from black soot. (b) The integrated emissivity of the VO2 -sapphire sample over the 8-to14-µm wavelength range. (c) Infrared camera images of the sample (diameter = 1 cm) for increasing temperatures [122]. Figure 2.9 : Multilayered structure for smart window[136]. theory, numerical simulations are obtained to compare to the experimental results as seen in Figure 2.12. Dávila et al. [128] has successfully reported on the use of VO2 ’s programmability of 18 Figure 2.10 : Sketch of the VO2 /Ag and VO2 /Cu multilayers structures. The number of layers is here, for example, N=7[137]. Figure 2.11 : Tunable optical cavity. (a)Schematic of the plasmonic device with hexagonal array of period = 4 µm and diameter = 1.64 µm. FTIR scan generated images of the plasmonic surface acquired for (b) semiconducting (T = 295 K), (c) phase-separated (T = 320 K), and (d) metallic (T = 360 K) states of VO2 . Albert Einstein image is 1.3 x 1.7 mm2 [139]. optical states to project near infrared images. By a using a combination of lasers (λ =1550 nm and λ=650 nm), optical states can be programmed into the VO2 surface by photo- thermal actuation as seen in Figure 2.13. The principle of optical memory works due the fact that it is being performed within the hysteretic region of VO2 . By using the writing laser to induce the PT on VO2 and causing a drop on transmittance, then an IR laser is transmitted and measured by a laser beam profiler on the back of the sample thus creating 19 Figure 2.12 : Infrared emissivity vs temperature from front (a) and rear (b) surface of the 190 nm VO2 film on Si substrate. Red lines are calculated emissivity during heating cycle, blue lines are calculated emissivity during cooling cycle. Continuous lines are theoretical fit by using Maxwell Garnett effective medium theory[140]. an image and successfully creating a programmable NIR scene projector. Figure 2.13 : (Left) Experimental optical system used to program an image onto a VO2 thin film. (Right)Programmed and projected NIR images. (a) 55o C before laser scanning. (b) During writing laser scanning. (c)Right after writing laser scan finished. (d) Programmed image 5 minutes after scanning[128]. 20 2.5 VO2 -based Actuators Due to its intrinsic phase transition, VO2 makes a suitable candidate for MEMS de- vices actuators [141, 40, 67, 142, 51, 143, 144, 145]. Previous reports have successfully demonstrated deflections due to the change in coefficient of thermal expansion between two materials. Deflections of more than 100 µm have been previously been reported via thermal actuation on a VO2 -based coated silicon micro-cantilever with a length of 130 µm [141]. Taking advantage of VO2 ’s intrinsic memory, Cabrera et al. [40] presents a way to pro- gram mechanical displacement on a VO2 -based cantilever. Actuation is achieved through integrated resistive heaters as shown in Figure 2.14, and via electrodes, resistance can be measured to monitored the phase transition of VO2 . Recent studies have tried to optimize these types of MEMS actuators. Wang et. al. [65, 21] showed the improved capabilities in light absorption of the VO2 MEMS cantilever by adding a layer of single wall carbon nanotubes (SWNT) (see Figure 2.15). The addition of SWNT has shown that it enhances the actuator performance in terms of speed and responsivity. Moreover, SWNT could be applied to any other type of VO2 -based actuators to increase its photo-thermal efficiency. By exploiting the mechanical memory capabilities, mechanical states can be programmed in order to be used for applications such as MEMS mirrors. 21 Figure 2.14 : VO2 -based cantilever with integrated heaters for actuation and electrodes to measure the resistance change of VO2 . Top plot shows the input voltage and deflection. [40]. Figure 2.15 : (a) Fabrication process of SWNT/VO2 -based cantilever actuator. (b) Top view microscope images of SWNT/VO2 and bare VO2 based cantilevers. Zoom imaged of SEM cross-section of the SWNT/VO2 -based cantilever. [65] From this same concept, MEMS mirror devices with VO2 - based bimporh structures where reported by Torres et. al. [18]. Figure 2.16 shows the VO2 -based MEMS structure. A squared mirror platform is supported by 4 separated and individually actuated bimporh legs. Upon actuation, the phase transition generates stress levels up to 1 GPa and strain 22 energy densities near to 106 Jm−3 , thus resulting in a total maximum vertical displacement of 75 µm with a low power consumption. Manca et. al. [42] successfully developed a planar nanomechanical actuator based on VO2 . Figure 2.17 shows the chevron type geometry which further amplifies the large lattice expansion of VO2 during the phase change. Figure 2.16 : VO2 -based MEMS mirror device. (Top) SEM image with colored bimorph legs. (Bottom) Top view of VO2 -based MEMS device[18]. 23 Figure 2.17 : Planar nanomechanical actuator with chevron-type geometry based on VO2 phase transition. [42] 2.6 Optical Shutters Different shutters are used for different applications and technologies. Mechanical shut- ters are commonly integrated in laser systems to control aperture or display systems [146, 147]. They can also be adapted into bigger designs for temperature control purposes and thermal energy storage [148, 149]. Electro-optical shutters can be substantially faster than their mechanical counterpart, with switching speeds in the range of µs’s [82, 150]. Typically these shutters are integrated into systems for optical modulation of transmittance and re- flectance [151, 152], and frequency modulation [153]. Different materials, specifically phase change materials can be incorporated into the design of electro-optical shutters to further en- hance the modulation capabilities of the device. Phase change materials such as germanium telluride (GeTe) have shown a transmittance modulation of around 70 % when integrated into an array of sub-wavelength gold slits filled with GeTe (Figure 2.18) for a wavelength of 1550 nm [150, 154, 155]. 24 Figure 2.18 : GeTe phase change shutter on a glass substrate. Grey areas represent GeTe and gold slits represent the heater. [150] Chalcogenide phase change materials (GeSbTe) with large changes in their optical con- stants have also successfully been shown capable of adaptive thermal control in the infrared region [156]. Figure 2.19 demonstrates such a design for optical shutters to be used for adaptive coded aperture control in the infrared. By using the optical constants of the ma- terial (n, k ) an optimization algorithm was used to guarantee the device performance goals. One drawback for these types of devices is the fact that they are mechanically static shut- ters and require a high power consumption to trigger the phase change on the material. For example a temperature range of 150 o C to 230 o C is required for materials like GeTe [157]. Naturally, phase change materials with a low transition temperature, ease of fabrication and low power consumption are desirable for smart material fabrication. Figure 2.19 : Design for the PCM shutter. (a) Blue represents the Ge2 Sb2 Te5 PCM, yellow represents the silica substrate. (b) Tiled filter pattern for the device. [156] Light device shutters based on liquid crystals have been previously reported. Several methods for these types of shutters exist, these range from polymer-dispersed liquid crystals 25 (PDLCs) to polymer network liquid crystals (PNLC) [158, 159]. A great advantage to liquid crystal shutters is the low energy consumption. Du et. al. [160] has shown an electrically switchable light shutter with a hybrid structure of cholesteric liquid crystals and chiral polymer film. By sending a voltage signal, the light shutter can be switched from a planar transparent state to a focal conic state which is highly reflective for near infrared light. By exploiting this effect, a 30 % transmittance drop was observed for IR light by using a square wave with voltage of 250 V and a frequency of 1 Hz. Figure 2.20 shows a real picture of the liquid crystal shutter while actuated showing both states (a shows planar state and c shows focal conic state). Figure 2.20 : LC shutter (a) Real photo of sample S1 at P state (b) Polarization optical microscopy image of P state. (c) Real photo of sample S1 at FC state. (d) POM image of sample S1 at FC state. [160] Heinilehto et. al. [88] demonstrated an infrared optical shutter by incorporating VO2 in a thin film stack with ITO. By exploiting the metal-insulator phase change via Joule heating, a 26.5 % change on the transmittance was reported. Figure 2.21 shows a schematic of the two designs for the VO2 -based optical shutter. By applying a voltage through the platinum pads, the phase transition is triggered and the transmittance was monitored by using a spectrometer. 26 Figure 2.21 : Schematic for the IR shutter structures. (a) Three layer structure based on ITO/VO2 /ITO with Pt pads that act as electrical contacts. (b) Two layer structure with the ITO bottom layer, top VO2 layer and Pt pads. [88] Similarly, Soltani et. al. [161] presented the design and fabrication of a planar micro- optical switch device based on VO2 ’s phase transition. By doping with tungsten (W), the transition temperature is successfully lowered to 36 o C. A reversible transmittance switching of 45% is shown for the device for a wavelength of 1550 nm. The device is triggered by applying a voltage (up to 28 V), thus successfully triggering the phase transition of W doped VO2 for a lower phase transition temperature. Figure 2.22 shows the design schematic for the VO2 -based micro-optical switch (a) and transmittance modulation (b). Figure 2.22 : (a) Schematic of micro-optical switch device based on W-doped VO2 . (b) Transmittance modulation (λ= 1550 nm) for the W-doped VO2 -based micro-optical switch device. [161] 27 By spin coating VO2 unto a Al2 O3 sapphire substrate, Johansson et. al. developed an active light shutter for infrared light (3-5 µm) [87]. For actuation purposes, a resistive gold heater was deposited. Upon actuation a transmittance drop of 55 % and shutter actuation switching time of 15 ms was observed. Figure 2.23 shows a cross section view of the finalized device along with a top view of the optical shutter with integrated resistive heaters. Figure 2.23 : (a) Cross-section view of the VO2 -based light shutter. (b) Deposited Au circuit on top of VO2 film. [87] 28 CHAPTER 3 DESIGN AND DEVICE FABRICATION OF VO2 -BASED WINDOWS, CANTILEVERS, AND MEMS TUNABLE OPTICAL SHUTTERS AND EXPERIMENTAL SETUP 3.1 Design and Fabrication of Monolithically Integrated VO2 -based Micro Windows and Experimental Setup. 3.1.1 Fabrication Flow Process of Monolithically Integrated VO2 -based Micro Windows The design and fabrication of the VO2 -based windows consisted of a 4-mask lithography process as shown in Figure 3.1, with a minimum feature size of 10 µm. The fabrication process starts with a double sided polished SiO2 wafer (2-inch diameter, 500 µm thick, SOF50D05C2, MTI) as the substrate. The metal layers (titanium/platinum (Ti/Pt) with thicknesses of (400 Å/1500Å, respectively) used for the heater and electrodes were deposited by evaporation and patterned using lift-off technique. The Ti layer was used only for adhesion purposes. The width of the heater loop was 10 µm and the electrodes were 15 µm. The gap between the heater loop and the electrode was 10 m as well. An insulating layer of SiO2 of approximately 400 nm thick was deposited by plasma enhanced chemical deposition (PECVD). This was done in 3 steps of approximately 130 nm each to avoid possible voids through the SiO2 . After deposition, the SiO2 layer was etched by reactive ion etching (RIE) to open vias to the Pt electrodes to the VO2 thin film that is to be deposited next. A layer of VO2 , approximately 170 nm thick was deposited by pulsed laser deposition (PLD). A KrF laser operated at 10 Hz with a laser fluence of 2 J/cm2 was used, with a deposition time of 25 minutes. The substrate was maintained at 595 oC in an oxygen environment at 15 mTorr pressure. After deposition, an annealing process under the same pressure and temperature conditions was performed for 30 minutes. To avoid any undesired material residual accumulation that would have affected the results for 29 transmission experiments, the backside of the wafer was covered during the deposition with a Si wafer. This was followed by the patterning of the VO2 windows through photolithography process and RIE using reference [18]. Finally, another SiO2 etching step was performed to open contact pads to the heater and electrodes. The 2-inch wafer was then diced into individual dies, each measuring 4 mm2 . The final design consists of 4 VO2 -based windows 100 µm2 , 200 µm2 , 300 µm2 and 400 µm2 . Figure 3.2 shows SEM pictures for the VO2 -based windows. For the purpose of this experiment, the 400 µm2 was used. Figure 3.1 : Fabrication process for the VO2 based windows. (a) SiO2 substrate. (b) met- allization of heater and electrodes. (c) SiO2 insulating layer. (d) opening of the electrodes. (e) VO2 deposition and window patterning. (f) opening of contact pads for electrical con- nections. 30 Figure 3.2 : (a) SEM images for the VO2 -based windows. (b) SEM image for the 400 µm2 window. 3.1.2 Experimental Setup for Monolithically Integrated VO2 -based Micro Win- dows Figure 3.3 shows the electro-optical setup used to test the VO2 window. The die containing the four micro VO2 windows was mounted and wire-bonded into a circular package which has a hole in the middle to facilitate the transmittance measurements. The package was then mounted to a custom built printed circuit board (PCB) with a centered hole and electrical connections for both the heater and electrodes. Once wired and mounted, the PCB was placed on an X-Y-Z translational stage to align the IR laser beam with incidence normal to the window. A Thorlabs NIR laser diode (λ=1550 nm, ML925B45F) was operated below its stable power of 5 mW for measurement taking. To make sure that the beam spot was properly aligned with the window, a Thorlabs NIR laser diode (λ= 980 nm, L980P010) was used. Both laser diodes were passed through a 50:50 NIR beam splitter (Thorlabs, BS015- 50:50, 1100 nm-1600 nm), then coupled into a single mode optical fiber and focused with a lens of 15 mm focal length. The lens was mounted on a micro positioner rail to control the diameter of the beam. A laser beam profiler (LBP, Newport, Model number LBP-4-PCI) was used to assist in the alignment of the focused laser beam and to obtain an approximate 31 value of the beam diameter. After passing through the sample, the laser beam was focused into an optical sensor (S144C, Thorlabs) connected to a power meter (PM100D, Thorlabs), which communicates with a LabVIEW computer interface to facilitate data gathering. Figure 3.3 : Electro-Optical setup used to measure VO2 ’s resistance and transmittance. For measurements, the electrical contacts of the PCB were connected to a National Instruments data acquisition (NI USB-6001) control for data acquisition. To actuate the window, a heater current IH was used while the voltage across the VO2 (VV O 2 ) was mea- sured. Using a voltage divider and (VV O 2 ), the resistance of VO2 was calculated (see insert equation in Figure 3.3). For the voltage divider, a series resistor of RS = 6.67 kΩ and a 32 Figure 3.4 : OptoTherm, Infrasight MI320 infrared camera setup to measure VO2 ’s thermal distribution and emissivity. supply voltage of VC = 10 V were used to measure the VO2 resistance. Thus, the setup allows to simultaneously measure and drive the device. The electro-thermal actuation, temperature distribution and emissivity were investigated by IR thermal imaging (OptoTherm, Infrasight MI320) as shown in Figure 3.4. Emissivity measurements were taken as a function of both temperature (i.e. conductive heating) and current (i.e. resistive, or Joule heating). For the temperature measurements, the die con- taining the window was attached to a Peltier heater. In order to obtain the value of the 33 thermal emissivity, it is necessary to have a material with a well-defined emissivity (i.e. a benchmark). To this end, a piece of masking tape with a known value of emissivity (E= 0.95) [162] was placed near the window of VO2 . Then the emissivity of the VO2 was modified in the thermal camera until the temperature measured by the IR thermal imaging system in the VO2 region was equal to the temperature measured in a selected region inside the masking tape. 3.2 Design and Fabrication of VO2 -based Shape Converting De- vices 3.2.1 Fabrication Flow Process of VO2 -based Shape Converting Devices The fabrication of the VO2 -based shape-controlled devices consisted of a two-mask lithog- raphy process with a minimum feature size of 10 µm. Figure 3.5-A shows the fabrication process. A double sided polished SiO2 wafer (SOF50D05C2, MTI Corp.) with a diameter of 2 inches and a thickness of 500 µm was used as the substrate. The first step was the deposition of a layer of VO2 via pulsed laser deposition (PLD). A silicon wafer was used as a backside cover for the SiO2 wafer to avoid any material accumulation that could affect any transmittance measurement. The VO2 deposition was performed using a KrF laser operating at 10 Hz with a laser fluence of 2 J/cm2 and a deposition time of 25 minutes. The substrate was maintained at a temperature of 595 o C, and the chamber was evacuated to an oxygen environment with a pressure of 15 mTorr. A 30-minute annealing process was performed after the deposition was completed. Following the deposition, the VO2 layer was patterned and then etched by reactive ion etching (RIE). A 300 nm-thick insulating layer of SiO2 was then deposited by plasma enhanced chemical deposition. This was performed in three steps of 100 nm at a temperature of 250 o C to avoid any voids across the SiO2 , and to make sure that there is no contact (short-circuit) between the VO2 layer and metal traces. Metal layers (titanium/gold (Ti/Au) with thickness of 20 nm and 580 nm) used for the heaters were deposited by evaporation and patterned using lift off technique. This was followed by 34 another etching step to remove the SiO2 in order to expose the VO2 . This last step was performed only on some of the samples, in order to compare results between samples with exposed VO2 and covered VO2 . The final wafer was then diced to form individual dies. Each die contains: 1. two 1100 µm x 500 µm windows with 8 heaters (with a width of 10 µm), properly allocated to form a number eight shape; 2. a 650 µm2 window with a double heater configuration (square and circle); and 3. a 650 µm2 window with two triangle heaters form- ing a square and a middle heater in the center completes the device. Figure 3.5-B shows an optical microscope image of the final die with the corresponding VO2 windows. Once the wafer is diced into individual chips, it is mounted on a circular IC package before being wire bonded and mounted on a custom built PCB for both electro-optical measurements. Figure 3.5-C-D show scanning electron microscope images (SEM, JEOL 6610LV) of the VO2 samples where the VO2 layer was exposed. The cross-section SEM images are taken after cutting a single die along the center of the 1100 µm x 500 µm VO2 window (i.e. along the lines shown in the top-view images, Figure 3.5-B-C). From the cross-section images, the measured thickness for the Au/Ti, SiO2 , and VO2 layers (top-to-bottom order) are 465, 300, and 200 nm, respectively. Before the SEM images were taken, a thin coating of osmium (10 nm) was deposited over the samples using a NEOC-AT osmium CVD (chemical vapor deposition) coater (Meiwafosis Co., Ltd.). 35 Figure 3.5 : (A) Fabrication process for the VO2 based window. (a) 500 µm thick SiO2 substrate, (b)Deposited VO2 layer. (c) Patterning of VO2 windows, top windows are 1100 µm x 500 µm, lower windows are 650 µm2 . (d) Insulating layer of SiO2 , (e) Au/Ti metal deposition. (f) Pattering of metal traces. B) Optical microscope image of die containing VO2 based devices. (C) SEM images for: (1) 1100 µm x 500 µm, (2) 650 µm2 with circle and square heater configuration, (3) 650 µm2 with triangle heater configuration.(4) Cross- section for the 1100 µm x 500 µm device with respective thickness. (D) Cross-section for the 1100 µm x 500 µm window showing the VO2 thickness, SiO2 top layer thickness and gold thickness. 3.2.2 Experimental Setup for VO2 -based Shape Converting Devices The optical setup used was similar to the one used for the experiments performed on section 3.1.2, except for the implementation of a voltage divider for the resistance measurements. For the transmittance measurements, a Thorlabs NIR laser diode (λ=1550 nm, ML925B45F) operated at a current of 34.9 mA was used on the optical setup. The device did not include 36 metal electrodes to monitor the resistance of the material, in order to minimize the number of metal lines that will interfere with the transmission, reflection and emission due to the VO2 film. The sample was actuated through Joule heating by passing a current IH through the heaters, which was computer-controlled and monitored via a virtual instrument and data acquisition (NI USB-6001) system. To avoid any overheating of the sample, a limiting resistance (RL =46Ω) was connected in series with the heaters on the devices. For emissivity measurements and generation of thermal images showing the desired pro- grammed states (and shapes) on the device, an IR thermal microscope (OptoTherm, In- frasight MI320) was used. The die containing all the devices was attached to a Peltier heater. A material with a known emissivity as a benchmark (masking tape, E=0.95) [162], was placed close to the VO2 window, and the thermal microscope was focused on both sur- faces. It should be noted that the emissivity of the benchmark does not change for the range of temperatures used in this experiment (i.e. from room temperature to 100 o C). This was followed by a calibration process, where the temperature of the Peltier heater was varied, and emissivity of VO2 (as read by the microscope) was varied until the measured temper- ature in the VO2 region matched that of the used benchmark. This process results in the characterization of the emissivity of VO2 as a function of temperature (V O2 (T )). It also gives a relationship that is used to obtain the emissivity as a function of applied current i (V O2 (i)). The temperature that corresponds to an applied current i can be obtained by using V O2 (T ) to find the matching emissivity of VO2 and corresponding temperature for each value of i. 3.3 Design and Fabrication of VO2 -based Cantilever Devices 3.3.1 Fabrication Flow Process f VO2 -based Cantilever Devices The fabrication of the VO2 -based cantilevers is shown in Figure 3.6-A and follows the same steps as reported by [40]. A 1 µm thick layer of SiO2 was deposited via PECVD over a 500 µm Si wafer. This was followed by deposition of the metal that will act as the heaters for 37 actuation. They consist of a 50 nm layer of Ti(for adhesion purposes) and a 150 nm layer of Pt. The metal is then encapsulated by another 1 µm thick layer of SiO2 . The device was patterned via plasma etch and then the structure is released using XeF2 dry etching. This is followed by cutting the wafer into individual dies. After the device is properly released and diced, a VO2 layer is deposited over the dies which are mounted on a Si wafer. This VO2 layer is deposited via pulsed laser deposition (PLD) using a KrF laser which operated at 10 Hz and an energy of 450 mJ for 25 minutes. The substrate is kept at a temperature of 595 o C during the deposition and then the device is annealed for 30 minutes to remove any remaining impurities. Once the VO2 layer is deposited, the devices are mounted and wire bonded on a rectangular IC package for electrical measurements. The final device consists of 4 cantilevers with varying dimensions of 550 µm x 50 µm, 450 µm x 50 µm, 350 µm x 50 µm and 250 µm x 50 µm. For the purpose of this experiment, the 450 µm x 50 µm cantilever was used. 38 Figure 3.6 : (A) Fabrication flow process for the VO2 -based cantilevers. (a) A 1 µm layer of SiO2 is deposited over a 500 µm thick Si wafer via PECVD. (b) Deposition and etching of a 150 nm/50 nm layer of Pt/Ti that will act as the heaters for the cantilevers. (c) A 1 µm layer of SiO2 is deposited over the metal structure.(d) Plasma etching is performed to pattern the SiO2 . (e) The structure is released via XeF2 etching. (f) A layer of VO2 is deposited via PLD over the release structure. Dashed red line shown in (f) represents where the sample was cut to image the cross-section in the SEM. (B) SEM picture of the cross-section of the final device showing each of the different layers that made up the device. A 100 nm layer of VO2 is observed followed by the 1 µm layer of SiO2 and 200 nm layer of Pt/Ti. (C) Shows a top/side view of the 550 µm x 50 µm and 450 µm x 50 µm cantilevers. 3.3.2 Experimental Setup for VO2 -based Cantilever Devices To measure the resonant frequency of the cantilevers, an interferometer setup was used [163]. This setup relies on monitoring the resonant frequencies by measuring the deflection of a laser beam due to the cantilevers and a piezoelectric transducer. The reflected laser beam was focused into a detector that sends the output into a spectrum analyzer that displays the voltage as a function of frequency. For the work shown here, the resonant frequency of un-coated cantilevers (no VO2 ) was measured to find a value for the effective length of the cantilevers. Resistance measurements as a function of temperature where taken by 39 measuring the resistance drop across the VO2 while increasing the temperature of the film using conductive heating. This was done by using a Peltier heater that was driven by a temperature controller (Thorlabs TED 4015) and the resistance change was monitored via a computer controlled program. For the purpose of taking high-speed measurements of the deflection due to a step of current, a high speed camera (Olympus i-speed 2) is used. Figure 3.7-a shows the setup used for recording of deflection videos. After being wire bonded, the sample is placed on a circuit board in order to actuate the cantilever through Joule heating by passing a current across the metal traces of the device. This is all done via a computer controlled system with the help of a virtual system and data acquisition (NI DAQ-USB 6001). The high speed camera is placed in front of the device at a fixed distance and a series of lenses are used to create and focus the image. As the device is actuated, the camera captures a video with a set frame rate of 10,000 frames per second (fps) for 5 seconds. The video file is then analyzed to obtain the data points for deflection as a function of time using Tracker Video Analysis and Figure 3.7 : (a) Optical setup with high-speed camera to record deflection. (b) Resistance vs Temperature curve for the 450 µm x 50 µm cantilever. (c) 450 µm x 50 µm cantilever in equilibrium state before actuation. (d) 450 µm x 50 µm after actuation with an input current of 4.8 mA 40 Modeling software (Version 5.1.2, Douglas Brown, physlet.org/tracker). Figure 3.7-c and d shows images for the 450 µm x 50 µm cantilever before and after actuation as obtained from the video. 3.4 Design and Fabrication of VO2 -based Tunable Optical Shut- ters and Experimental Setup 3.4.1 Fabrication Flow Process of VO2 -based Tunable Optical Shutters The electro-mechanical-optical shutter device design consists of a combination of VO2 optical windows [33, 34] and VO2 -based MEMS actuators [164], where the window was added to the end of the cantilever structure. Separate resistive heaters allowed for individual actuation of the cantilever structure and the optical window; i.e., a resistive heater along the cantilever was used for controlling the mechanical position of the shutter device (Figure 3.9-a, red, dashed line); while a separate heater around the window was used for controlling the optical properties of the window (Figure 3.9-a, blue, dashed line). The fabrication process of the MEMS VO2 -based optical shutter devices is shown in Figure 3.8-a, and discussed next. The cross-section 2D diagrams corresponds to the dotted, blue line shown on Figure 3.8-d. A 1 µm layer of SiO2 was deposited via PECVD at a temperature of 250 o C on a 500 µm thick Si wafer. This was followed by approximately 160 nm of VO2 (confirmed by SEM pictures) by PLD (Pulsed Laser Deposition) using a KrF laser which operated at 10 Hz and a fluence of ∼ 2 J/cm2 for 25 minutes and a O2 pressure of 20 sccm. During deposition, the substrate was heated by a ceramic heater located behind the substrate, kept at a temperature of 595 o C during the 25 minute deposition, and a 30-minute post annealing step at the same deposition conditions. Patterning and etching of the VO2 layer was performed and followed by a 300 nm layer deposition of SiO2 (done in 3 steps of 133 nm at 225 o C, each to reduce the amount of voids in the layer). After patterning and etching the SiO2 layer, metallization and lift off of a Au/Cr film (180 nm/ 20 nm) was completed. A final encapsulating layer of SiO2 (200 nm) was deposited at a temperature of 225 o C. This was followed by dicing of the wafer into 41 individual dies that contained two identical optical shutter devices. After dicing, the devices are released via isotropic etching of the silicon substrate using XeF2 gas. The final device consisted of cantilevers with length and widths of 750 µm x 80 µm, respectively; and square (650 µm2 ) optical windows. Figure 3.8-b-c and d shows SEM images (JEOL 7500F) of the VO2 -based optical shutter (before and after release), coated with a thin layer of osmium (10 nm) using a NEOC-AT osmium CVD coater (Meiwafosis Co., Ltd.), this conductive coating is necessary to avoid charging and to improve the secondary electron signal for the SEM furthering improving the image [165]. Figure 3.9-b shows a magnification of the cantilever structure, where two pairs of traces are identified: the wider pair runs to the window and is used to heat the window square; while the other one corresponds to the cantilever heater, used to heat and actuate the cantilever structure. A VO2 gap between the cantilever structure and window was also implemented on the design. This is done to minimize the heat distribution between the cantilever and window upon actuation of the metal traces. The SEM cross- section image is taken after cutting a single unreleased die perpendicular to the cantilever (dotted red line on Figure 3.8-d). From the cross-section, the measured thickness for the bottom SiO2 , middle VO2 and top SiO2 are 1.33 µm, 160 nm and 200 nm, respectively. 42 Figure 3.8 : (a) Fabrication flow process for the VO2 -based optical shutter. Cross-section views correspond to the dotted blue line shown in figure 3.8-d. (1) 500 µm Si wafer. (2) Deposition of a 1 µm layer of SiO2 by PECVD. (3) Deposition of VO2 layer by PLD. (4) Patterning and etching of VO2 layer. (5) Deposition of a 300 nm layer of SiO2 by PECVD. (6) Patterning and etching of second SiO2 layer. (7) Evaporation and metal lift-off of Au/Cr layer. (8) Deposition, patterning, and etching of a 200 nm layer of SiO2 . (9) Structure release by XeF2 etching. (b) Cross-section of the VO2 based cantilever showing the top and bottom SiO2 layer thickness and VO2 layer thickness. (c) Side angle view for a released VO2 based shutter (d) Top view SEM image for an un-released shutter. Dotted lines represent cross-section cuts. 3.4.2 Experimental Setup for VO2 -based Tunable Optical Shutters The electro-optical setup used to test the VO2 -based optical shutter is similar to the one used for experimentation in [33, 34] and is shown in Figure 3.9-a. A die containing the device is wire-bonded to a IC package which is then mounted on a solderless breadboard 43 with the required electrical connections for actuation of both heaters. A 980 nm laser diode (Thorlabs, L980P010) with a focused diameter of approximately 80 µm, an operating current of 30 mA and coupled into an optical fiber is used to align the beam spot in the area of interest in the VO2 window and photo-detector (S144C, Thorlabs) which is connected to a power meter (PM100D, Thorlabs) with a sampling rate of 100 Hz. In order to align the beam spot, a CCD camera (Sony CCD-IRIS Hyepr HAD B&W CCTV) is used. Once the beam spot is aligned, a 1550 nm laser diode (Thorlabs, ML925B4SF) operated with a current of 15 mA and coupled into the optical fiber is used for measurements. Actuation of the device is done via Joule heating by passing a current through the heaters (represented by dashed red and solid blue lines in Figure 3.9-a), which was computer-controlled and -monitored via a virtual instrument and data acquisition (NI USB-6001) tool. A limiting resistance (RL = 325 Ω) connected in series with the heaters of the device is used to prevent overheating on the thin metal resistive heaters, due to voltage spikes in the applied signal. For experiments where both heaters need to be actuated at the same time, the window heater current was computer controlled (NI USB-6001, data acquisition tool) and the cantilever heater was manually controlled with a step input provided by a power supply. For window actuation only, the voltage input signal will go from 0 to 7 volts (0 mA to 15.7 mA) in steps of 0.1 V with a holding time of 500 ms. For the cantilever transmittance measurements, the voltage input signal was programmed to go from 0 to 3.5 volts in steps of 0.1 V and a holding time of 500 ms. For measurements where both the window and cantilever are actuated, the voltage input signal was programmed from 0 to 7.5 volts (0 mA to 16.5 mA) in steps of 0.1 V and a holding time of 500 ms. The extra voltage steps where used to accommodate the actuation of the cantilever with the supplied current step (6 mA). This is followed by the voltage returning to 0 volts following the same rate for all cases. While the input voltage is applied, the computer software will convert the input to current and each current point will be associated with a power value that comes from the photo-detector. In order to calculate the transmittance value, the power before the window is measured and is inputted into the 44 Figure 3.9 : (a) VO2 -based optical shutter experimental setup. Blue, solid line represents the actuation current for the window. Red, dashed line represents the actuation current for the cantilever (tilt control). Black arrow represents direction of actuation upon Joule heating (figure is not to scale) . (b) Top close up view for the cantilever/window structure. Squares represent release holes for ease of etch. Orange, red and yellow spots represents where the beam is focused on the window as seen on (a). computer software. The calculated transmittance is then plotted as a function current. 45 CHAPTER 4 PROGRAMMING EMISSIVITY ON FULLY INTEGRATED VO2 WINDOWS This chapter presents a micro variable optical window based on VO2 thin film. The window can be actuated via resistive heaters, which allows to program states in both transmittance and emissivity via electronic modulation. Via this change in transmittance and emissivity we can developed a smart VO2 window that could act as a cloaking device. The intrinsic hysteresis of VO2 allows for multiple optical states for a single temperature within the transition region. This is exploited to demonstrate electronically programmable emissivity states in a monolithically integrated VO2 window. 4.1 Results and Discussion 4.1.1 Electro-Optical States Characterization of both optical and electrical transitions were performed simultaneously for the 400 µm sized VO2 window as shown in Figure 4.1. In order to obtain the major Figure 4.1 : Simultaneous measurements for the electrical (a) and optical (b) transition in the 400 µm VO2 window. 46 hysteretic loops, current steps of increasing amplitude (0.1 V, or 0.63 mA) were applied to the heater electrodes until the phase transition was complete. This resulted in the major heating hysteretic loop. Then, current steps of decreasing amplitude were applied until reaching 0 mA, which resulted in the major cooling hysteretic loop. Each current step lasted 1 s, and the measurement was taken after waiting 900 ms from the beginning of the step. During this input to the heater, the VO2 ’s resistance and transmittance were being monitored. A drop of approximately 3 orders in magnitude is visible for the resistance of VO2 across its phase transition. This drop in resistance of VO2 is to be expected and confirms the overall good quality of the sample. The average resistance drop for the sample was from Ri = 631 kΩ to Rf = 676 Ω, having an average Ri /Rf ratio of 935 . For the optical transmittance, the power of the IR laser before the sample (1.22 mW) was used to normalize the transmitted power through the sample. A transmittance drop from Ti = 0.36 to Tf = 0.14 was observed for the 400 µm window, giving a ratio of Ti /Tf = .57 for the infrared region (λ= 1550 nm). To demonstrate the programmability stages of VO2 , several minor loops were measured for both electrical and optical transitions. Figure 4.2 shows the plots for resistance and transmittance. The input used for obtaining these minor loops is shown in the insert of Figure 4.2. Obtaining the minor loop plots allows for a more reliable way to program the desired values for transmittance and therefore emissivity. For this case, the same current-increase input steps used to obtain the major heating hysteretic loop were used, until reaching 15 mA. This was used as the pre-heating current, from which the first electrical/optical state was measured. Programming of a second electrical/optical state was achieved by applying an electrical pulse (also following the same step input) up to 16.97 mA (see Figure 4.1). Since this value of current is not enough to complete the phase transition of VO2 , once the pulse is over, the resistance and transmittance comes back to the pre-heating current, but following one of the minor hysteretic loops. The resulting pre-heated and programmed states in electrical resistance and optical transmittance are shown in Figure 4.1, and the corre- 47 sponding minor loop is identified in Figure 4.2. It should be noted that there is a DC shift in transmittance between the minor loops and the programming pulse, which is most likely due to a small difference in the background light when measurements were taken. Although this shows only one programmed state, essentially any electrical resistance/transmittance value that belongs to the minor loops can be programmed by simply using a different pulse magnitude or pre-heating current. In order to know the required minimum sampling rate for the 400 µm window, the device’s thermal time constant was measured. This was done by measuring the voltage in VO2 (VV O 2 ) resulting from a single current step input IH . Figure 4.3 shows the thermal time constant (τof f = 21.68 ms) for the device for the case when the step was released. This measured time constant of approximately 20 ms is much faster than the 100 ms used for the current step input pulses, which indicates that the pulses will be enough to reach steady-state. Figure 4.2 : Electrical (a) and optical transition (b) minor loops in the 400 µm VO2 window. Programming pulse is over-imposed. Inset shows the voltage input used to obtain the minor loops. 48 4.1.2 Programmability of Emissivity States in VO2 The change in VO2 ’s optical properties across its phase transition is larger for wavelengths in the infra-red (IR) region [45]. The material’s ability to radiate thermal energy (i.e. emis- sivity) also changes abruptly during the material’s phase change, which allows for selective thermal emission. Figure 4.4 shows the VO2 emissivity as a function of both temperature and current. For both plots, VO2 shows negative differential emissivity at the onset of the phase transition, which occurs around TP T ≈ 68 o C and IP T ≈ 19.9 mA. The VO2 window shows a large thermal emissivity change from 0.76 below the transition point to 0.54 above the transition point. To measure emissivity as a function of current, the temperature of the VO2 window was first measured as a function of current (Figure 4.5-b). Given that the Figure 4.3 : Time constant measurement for the 400 µm VO2 window. 49 Figure 4.4 : Emissivity as a function of temperature (a) and current (b) for the 400 µm window. emissivity of VO2 changes with temperature, a benchmark is required for calibration. Using the masking tape mentioned earlier for conductive experiments from Peltier heater would not work in this case, since it would require increasing the temperature of the tape by Joule heating through the same resistive heater used for the window. Therefore, the platinum heater is used as a temperature sensor, by monitoring its temperature as the current was applied. The heater’s temperature (which is associated to the temperature of the bench- mark tape) as a function of current (Figure 4.5-c) was used to obtain the VO2 window’s temperature (Figure 4.5-a). This allowed for mapping VO2 ’s emissivity to obtain Figure 4.4-b. On comparing the transition point for the plots on Figure 4.2 to the plot in Figure 4.4-b, there is a difference in current of IP T ≈ 4.5 mA. This is most likely due to the method for measuring the emissivity as a function of current as previously mentioned. It should be noted that the hysteresis curves have similar shapes, which suggests that the difference is most likely a DC offset, which would be corrected by an additional integrated device that can be used as a benchmark for calibration. 50 Figure 4.5 : Curves for mapping method used to obtain the emissivity as a function of current.(a) Temperature of VO2 window as a function of temperature of tape = temperature of heater. (b) Temperature of heater as a function of actuation current. (c) Temperature of tape = temperature of heater as a function of actuation current. 51 Figure 4.6 : Thermal image for the 400 µm VO2 window. (a) before programming pulse. (b) after programming pulse. 4.1.3 Thermal Images of VO2 Window’s Figure 4.6 shows a thermal image for a 400 µm window before and after actuation using an electric pulse (1 mA), supplied in the form of short current steps as depicted for the electrical and transmissivity experiments. Although both states correspond to the same temperature since after the pulse the current returns to the pre-heated value. The thermal image after the pulse clearly shows a lower irradiance, which is mapped to a lower emissivity. This is due to the lower emissivity of the VO2 after the programming pulse. 52 4.2 Summary We have developed a VO2 based window of 400 µm2 that can be used as a smart window or thermal camouflage system. The electrical and optical transition in VO2 were investigated, and hysteretic curves for minor loops where obtained. The minor loops inside the hysteresis of the VO2 allows to program any state for transmissivity inside the window. Since the optical properties of VO2 change greatly in the IR- region, then the emissivity will change during the transition. The emissivity for the VO2 film as a function of both temperature and current was determined, and confirmed that the film shows a negative differential emissivity with the phase change. Emissivity states were programmed by electronic pulses to change the window’s thermal radiation, which can be used for real-time thermal cloaking. The rapid tune-ability of both transmissivity and emissivity in VO2 suggests that the film could be incorporated in the use of adaptive thermal camouflage devices. 53 CHAPTER 5 A SIMPLIFIED APPROACH FOR OBTAINING OPTICAL PROPERTIES OF VO2 THIN FILMS, AND DEMONSTRATION OF INFRARED SHAPE-SHIFTING DEVICES In this chapter the development of VO2 -based shape-controlled micrometer-sized devices, which are triggered through integrated resistive heaters is shown. This in turn allows us to demonstrate the shape-reconfiguring capabilities of VO2 in the IR region via electronically programmed pulses. A two surface reflectance and transmittance model is shown to simulate the effects of the film and substrate thickness on the optical reflectance, absorbance, and transmittance of the developed micrometer-sized electro-optical windows. Multiple studies and approaches have been used to model VO2 -based films and electro-optical devices [166, 167, 168], which require direct measurements of the material’s optical properties. The model presented here is simpler than most of the methodologies used; and yet, still presents a valid model that can be used to characterize these devices for applications such as the use of electrical pulses to reconfigure thermal shapes. Examples for the shape-converting capabilities of VO2 based devices are shown. Thermal images show how electrical pulses are used to convert square shapes to circles and then back to squares. Similarly, triangle shapes are shown to change to square shapes and back. Finally, a VO2 -based window with several resistive heaters forming an 8-bit digital display is used to demonstrate the same shape converting capabilities, where electrical pulses are used to display thermal images of different numbers. 5.1 Results and Discussion 5.1.1 Optical States and Emissivity of VO2 To verify the quality of the devices, optical transmittance measurements for the VO2 based windows were measured as a function of current. Figure 5.1 shows both a single hysteretic 54 loop and minor loops for the 650 µm2 window (device # 4 in Figure 3.5-B). The power of the laser was measured before the sample (1.23 mW) and used for normalizing the transmitted power through the window. The device shows a transmittance drop of about 30%, which is to be expected for this infrared region (λ=1550 nm) [45]. Hysteresis curves were plotted through computer-controlled heating and cooling cycles with decreasing amplitude in order to obtain minor loops inside the major hysteretic loop. For programming optical states, a DC-bias current of 82 mA was applied to reach a temperature at the onset of the phase transition (i.e. pre-heating temperature). The optical properties at the pre-heating temperature represent the first optical state. This is followed by programming of another electrical pulse of 85 mA which is not large enough to complete the full PT of the VO2 film; thus returning to the pre-heated level through one of the minor hysteretic loops. At this point, after the pulse, the sample is again at the pre-heating temperature, but with different optical properties. The minor hysteretic loops cover a transmittance change from 32% to 2%, which means that, any transmittance between these values can be programmed on the electro-optical window by using electrical pulses. Figure 5.2-a-b shows the VO2 emissivity as a function of both current and temperature for the 650 µm2 windows (device #4 in Figure 3.5-B), where the window consists of VO2 film exposed to air (e.g. the SiO2 layer on the VO2 was removed). At a pre-heating temperature of about TP T = 67 o C (which is applied by conductive heating using a Peltier heater below the sample), the measured average value for V O2 (T )= 0.83. When the temperature is increased to a value greater than 77 o C; (across the PT), V O2 (T ) drops to 0.4. Figure 5.2-b shows the V O2 (i), where a similar drop in emissivity from 0.83 to 0.40 is observed across the phase transition. A similar emissivity drop of 0.44 was measured on VO2 /carbon composites [169], which is in agreement to the observed drop in the present devices. In order to see the effects of an added layer of SiO2 above the VO2 , emissivity measurements were performed for a similar device where the SiO2 over the VO2 window was not removed. These results are shown in Figure 5.2-c, where no significant difference on the emissivity 55 Figure 5.1 : Optical transmittance measurements for the 650 µm2 VO2 window. (b) shows the minor hysteretic loops for the same window. Measurement was taken by actuating one of the single triangle heaters and measuring inside the area of the heater. Inset on (b) shows the voltage sequence used to obtain the minor loops. 56 Figure 5.2 : Emissivity for VO2 windows: (a),(b): emissivity of 650 µm2 VO2 window with etched SiO2 (exposed VO2 ) as a function of temperature and current, respectively. (c): emissivity for the 650 µm2 VO2 window with top layer of SiO2 over VO2 . 57 Figure 5.3 : Curves for mapping method used to obtain the emissivity as a function of current for the 650 µm2 VO2 window.(a) Temperature of VO2 window as a function of temperature of tape = temperature of heater. (b) Temperature of heater as a function of actuation current. (c) Temperature of tape = temperature of heater as a function of actuation current. 58 values before and after the transition can be observed, since SiO2 ’s doesn’t show any negative differential emissivity (E=0.75) [162]. 5.1.2 Two-Surface Reflectance and Transmittance Model for VO2 films. Using a two-surface reflectance and transmittance model at normal incidence, the trans- mittance and reflectance for VO2 can be modeled at different temperatures. The following calculations will take into consideration two interfaces: VO2 / SiO2 , and VO2 / incident medium (air). The metal traces are not taken into account since the beam spot size is cen- tered in the window, which only consist of the VO2 thin film on SiO2 substrate. To obtain the real part of the index of refraction (n) for fused silica (nSiO2 ) we use Sellmeier’s equa- tion [170] for a range from 400 nm to 1800 nm. For our desired range of wavelengths, the extinction coefficient (k) for SiO2 is set at kSiO2 =0 (i.e. no absorption), which is reasonable for wavelengths in the 400-1800 nm range [171]. Values for the real and imaginary parts of the index of refraction for VO2 (nV O2 and kV O2 , respectively) for a wavelength ranging from 400 nm to 1800 nm were obtained from previous works for temperatures ranging from 25o C to 120o C [45, 172, 173, 174]. The index of refraction for VO2 (N1 ) and SiO2 (N2 ) will then be defined as: N1 = n1 − ik1 (5.1) N2 = n2 − ik2 , where (n1 ,k1 ) correspond to the optical constants of VO2 and (n2 ,k2 ) correspond to the optical constants for SiO2 . Equation (5.1) can be incorporated in the definition of optical admittance (tilted optical admittance), which takes into consideration the case of oblique incidence for both parallel and perpendicular polarization and are given by; 59 η1S = N1 cos(θ1 ) (5.2) η1P = N1 cos(θ1 ) η2S = N2 cos(θ2 ) η2P = N2 cos(θ2 ) From equation (5.2) both cos(θ1 )=cos(θ2 )=1, since we are treating the case for normal incidence. Therefore η1S =η1P =η1 and η2S =η2P =η2 . The phase shift for light going from air to VO2 (δ1 ) is then given by: 2πN1 d1 cos(θ1 ) δ1 = . (5.3) λ The electric and magnetic fields at the air/VO2 interface ((Ea ) and (Ha )), normalized with respect to the electric field at the VO2 /SiO2 interface (Eb ) become B = E a Eb and C=H a Eb , respectively. A characteristic matrix is now defined, which combines the VO2 thin film and SiO2 substrate into one interface [175]:      isin(δ1 ) B   cos(δ1 ) η1   1   =   (5.4) C iη1 (sin(δ1 ) cos(δ1 ) η2 From equation(5.4) we can define a reduced optical admittance given by; C Y = (5.5) B Here, we have successfully reduced a two-interface optical model into a single bound- ary one. Using equation (5.5) we can then calculate both the reflection coefficient ρ and transmission coefficient τ as: η −Y ρ= 0 (5.6) η0 + Y 60 2η0 τ= (5.7) η0 B + C where η0 is the optical admittance of the first medium (in this case air). The transmit- tance (T ) and reflectance (R) can then be obtained from: R = ρρ∗ (5.8) 4η0 Re(η2 ) T = (5.9) (η0 B + C)(η0 B + C)∗ where ρ∗ is the complex conjugate of the reflectance coefficient. Similarly, the absorbance on a multilayer is related to R and T (A=1 − R − T ), which can be combined with equation 5.8 and 5.9 to obtain: 4η0 Re(BC ∗ − ηm ) A= (5.10) (η0 B + C)(η0 B + C)∗ Figure 5.4 shows the calculated T , R, and A values as a function of wavelength. In the calculations, the measured VO2 thickness (SEM) of d1 =205 nm was used. From Figure 5.4-a, it is noted that the transmittance for room temperature increases with wavelength until approximately 1.2 µm. After this wavelength, the transmittance follows a constant trend for (1.2 µm-1.6µm ), and then it increases for wavelengths larger than 1.7 µm. The transmittance is reduced with increasing temperature (from 25 o C to 120 o C) to almost 0%. It can be noted that the measured value for transmittance (Figure 5.1) was around 0.31 at λ=1550 nm, in comparison to the 0.34 at λ = 1550nm for the calculated results, which is 0.03 difference from the calculated results shown in Figure 5.4-a. This could be attributed to the fact that the n and k obtained from [45] take into consideration a thin 10 nm layer of TiO2 with a VO2 layer that was around 100 nm thick. Of greater importance is that the optical constants of VO2 will vary and are largely dependent on their fabrication process [176]. Figure 5.4-b shows the calculated transmittance as a function of wavelength 61 Figure 5.4 : (a) Transmittance as a function of wavelength. (b) Transmittance as a function of wavelength below and above transition temperature for several references. (c) Reflectance as a function of wavelength. (d) Absorbance as a function of wavelength. for both regions: below and above the transition temperature, using different n and k values obtained from the literature. Below the transition temperature (for temperatures between 25-35 o C), the calculated transmittance at λ=1550 nm using the optical constants from all the references included in Figure 5.4-b is in the vicinity of the measured transmittance of 0.31 shown in Figure 5.1-a. Similarly, the calculated transmittance at λ=1550 nm drops to 0 above the transition temperature (85-100 o C). On the other hand, reflectance values increase as the temperature is increased, (Figure 5.4-c), which is indicative of the intrinsic phase transition of VO2 going from insulator (low 62 reflective material) to a metal (highly reflective material). For wavelengths between 400- 850 nm, the reflectance peak value decreases and shifts to lower wavelength values as the temperature is increased. A similar trend was measured by [174, 177]. The calculated peak reflectance value at room temperature at about 850 nm in Figure 5.4-c, drops rapidly as the wavelength is increased, reaching a minimum reflectance value at a wavelength close to 1500 nm. This was also observed in previous optical measurements on VO2 films [167, 177]. The magnitude of this calculated drop is reduced for temperatures above room temperature. Furthermore, at a wavelength of 1550 nm, the calculated reflectance increases with tempera- ture, and for 120 o C the value increases for wavelengths above 700 nm. A similar trend was measured by [167, 174]. The large variations in the reflectance of VO2 at room temperature is due to the larger film’s roughness before the phase transition. In the insulating state, VO2 has a greater surface roughness and thus as a result light its scattered [178]. Above the phase transition, the film becomes metallic and surface roughness decreases, resulting in more uniform optical properties. Figure 5.4-d shows the absorbance of the VO2 film for different temperatures. The larger variations in the reflectance curves mentioned earlier cause similar large variations in absorbance at room temperature as well. These variations follow the same trend in magnitude variations and shift in wavelength. 5.1.2.1 Two-Surface Incoherent Reflectance and Transmittance Model for VO2 films. This section will give an update on the previous optical model shown in Section- 5.1.2. The previous model treated the substrates as a one-sided slab of material of infinite depth. This assumption is made for ease of calculations but it is not real and the substrate will have a finite depth. Such finite depth would mean that the rear surfaces of the layers of the film will reflect some of the incident energy [175]. For this case the depth of the SiO2 substrate is 500 µm, which is greater than the incident wavelength of 1550 nm. Figure 5.5 illustrates the reflected and transmitted waves at the front, inside and rear surfaces for the model with 63 the assembly consisting of the combination of the VO2 layer (205 nm) and SiO2 substrate (500 µm). Figure 5.5 : Schematic for the assembly showing incident reflection and transmittance for both sides of the stack. Considering that waves are reflected successively at the front and rear surface, then the reflectance and transmittance can be written as: Ra+ + Rb− (Ta2 − Ra− Ra+ ) R= (5.11) (1 − Ra− Rb+ ) Ta+ Tb+ T = (5.12) (1 − Ra− Rb+ ) where R a and T a correspond to the reflectance and transmittance between the air and assembly, and R b and T b are the reflectance and transmittance between the rear SiO2 and air. Equations 5.11 and 5.12 are for the case where absorption is present (A 6= 0). For the case of no absorption (A = 0), then; 64 Ra + Rb − 2Ra Rb R= (5.13) 1 − Ra Rb 1 1 T =( + − 1)−1 (5.14) Ta Tb with R + − a = R a = R a . Similarly the reflectance in terms of transmittance is R a = 1-T a and R b = 1-T b when A = 0. Figure 5.6 shows the calculated transmittance and reflectance for the VO2 /SiO2 assem- bly considering incoherent interactions of light. Figure 5.6 : (a) Transmittance as a function of wavelength. (b) Reflectance as a function of wavelength. 5.1.3 Implementation of VO2 Windows for Shape-Shifting Devices Since VO2 possesses a negative differential emissivity with temperature, and selected regions in the material can be driven to different optical states. It can be integrated in larger systems for demonstrating shape-shifting devices. The following experiments are demonstrations of possible applications for the developed devices, which include reconfiguration of geometric shapes (circles, triangles and squares) and numbers shown in an 8-bit display; as seen by an IR camera. Finite element method (FEM) simulations were performed using heat transfer 65 Table 5.1 : VO2 parameters for insulating phase used for FEM simulation. Properties VO2 Density [kg/m3 ] 4760 Relative Permittivity 1 Thermal Conductivity [W/mK] 5 Coefficient of Thermal Expansion [1/K] 5x10−6 and electric currents modules in Comsol Multiphysics software, where the temperature dis- tribution inside the VO2 windows was confirmed for different currents applied to the resistive heaters. Parameters for the insulating state such as density (ρ), relative permittivity (r ) and thermal conductivity (κ) and coefficient of thermal expansion (α) shown in table (5.1) were used for the insulating state of VO2 . A 1100 µm x 500 µm rectangle of VO2 with heaters of 450 µm in length and a width of 10 µm were designed for the simulation, the same as device #1 in Figure 3.5-B. Figure 5.7 shows the simulated temperature distribution for four cases, where the heaters are selected to create the desired figures (in this case, numbers). For this case a current of 35 mA was selected for each heater, which in turn resulted in a temperature close to the center of the VO2 window being greater than 70 o C. Figure 5.8-A shows several thermal images captured for the VO2 window with the circle and squared heater configuration (device 4 in Figure 3.5-B and C). It should be noted that the resistive heaters have individual and separate electrical connections (i.e. each heater has a separate pair of electrodes for wire bonding). The preheated stage was first achieved using a preheated current (Ip h = 98 mA) on the square heater shown in Figure 5.8-1. This value of current corresponds to a known value of emissivity (E1 = 0.74), which is labeled in the emissivity plot shown in Figure 5.8-B. This was followed by increasing the current (IH = 101 mA) on the square heater to achieve a lower emissivity for the film. The temperature increase provided by the traces induces a change that follows the shape of the heater, thus giving the appearance of a square in the thermal image. A current of 50 mA was then sent 66 Figure 5.7 : Temperature distribution for the 1100 µm x 500 µm VO2 window as simulated using the electric currents and Joule heating modules with an input current of 35 mA. to the circular heater (Figure 5.8-3), further increasing the temperature and inducing the phase transition inside the region surrounded by the heaters. This will in turn change the shape of the image to a circular shape with a lower value of emissivity (E4 = 0.40). The current on the circular heater was then increased to 100 mA, transitioning the entire VO2 film inside the window to the metallic phase. This results in a larger area for the thermal image of the square shape, but with different orientation (Figure 5.8-4). With both heaters on the window activated, a total current of 201 mA was supplied, giving an emissivity value of 0.40. At this stage, the circular heater was turned off, leaving the square heater at the second stage (IH = 101 mA). This would correspond to the same applied current as the stage shown in Figure 5.8-2. However, due to the hysteretic behavior of VO2 , this current 67 Figure 5.8 : (A) Thermal images for the 650 µm2 window with square and circle heater configuration for several programming pulses.(1) Corresponds to the pre-heated state using the square heater. (2) A second pulse is sent to achieve a lower emissivity state. (3) The second circular heater is activated, this creates a new circular state. (4) The current is increased on the circular heater to obtain a new square state. (5) The second heater is turned off with only the first heater maintained at the second state (2). A state with a lower emissivity but same temperature as in (2) is achieved.(6) The first heater is brought back to the pre-heated level. (B) Emissivity curve for a heating and cooling cycle. Labeled emissivities in the plot correspond to the values for each of the thermal images. will induce an emissivity in the VO2 film of only E5 = 0.43 (see emissivity curve in Figure 5.8-B). Thus, a lower irradiance is observed. The heater was then brought down to the pre- heated current (Ip h = 98 mA), where a square with a lower emissivity value (E6 = 0.45) can be observed (Figure 5.8-6). Several shape shifting stages can be successfully programmed using combinations of heaters incorporated into the device. A different shape-conversion for VO2 windows is demonstrated using a triangle heater shape (see Figure 5.9-A). From Figure 5.9-1, a pre-heated current of 101 mA was first applied to the first triangle heater, this resulted on an emissivity value of E1 = 0.69. This was followed by a pulse increase to 107 mA, which corresponded to an emissivity of (E2 = 0.51). A clear triangle figure can be seen inside the heater region (Figure 5.9-2). By sending another pulse of 101 mA to the second heater (Figure 5.9-3), a square can be achieved with a value of emissivity (E3 = 0.4) past the full transition of the VO2 –note the emissivity curve 68 Figure 5.9 : (A) Thermal images for the 650 µm2 window with triangle and triangle heater configuration for several programming pulses.(1) Corresponds to the pre-heated state using the triangle heater. (2) A second pulse is sent to achieve a lower emissivity state. (3) The second triangle heater is activated, this creates a new square state. (4) The current is increased on the triangle heater to obtained a new square state. (5) The second heater is turned off with only the first heater maintained at the second state (2). A state with a lower emissivity but same temperature as in (2) is achieved. (6) The first heater is brought back to the pre-heated level. (B) Emissivity vs current showing the corresponding states. in Figure 5.9-B, where all the emissivity states in Figure 5.9-A are labeled. As done on the first heater, the current was increased on the second heater to 106 mA (Figure 5.9- 4), causing the center of the square to irradiate more even though it corresponds to the same value of emissivity (E4 = 0.40). This was followed by shutting down the second heater (Figure 5.9-5), going back to the second state current (I= 107 mA) with a lower value of emissivity (E5 = 0.42). The heater was then brought down to the pre-heated current (Ip h = 101 mA), where a triangle with an emissivity of E6 = 0.44 is obtained (Figure 5.9-6). Figure 5.9-B shows the several states for emissivity and current for the device. Figure 5.10 shows several thermal images for the 1100 µm x 500 µm VO2 window, where the resistive heaters were actuated to form the numbers “7” and “1”. Following a similar approach to the one used in the previous geometric shapes, the VO2 film was locally heated to program the shapes of these two numbers. 69 Figure 5.10 : Thermal images for the 1100 µm x 500 µm VO2 window. (a) shows the pre- heated current necessary to form a number one figure. (b) a second pulse is applied to reach a lower emissivity state. (c) Image after second pulse is brought down to the pre-heated level. (d) shows the pre-heated current necessary to form a number seven figure. (e) second pulse is applied to reach a lower emissivity state. (f) Image after second pulse is brought down to the pre-heated level, a more defined seven can be seen on the lower irradiance area. 5.2 Summary A two interface model was implemented to calculate the transmittance, reflectance and absorbance for the device at normal incidence. The method did not require the measurement of optical constants. Instead, using previously measured parameters from the literature, and a theoretical approach for reducing a two-interface problem into a single layer model, the transmittance for VO2 -based devices was obtained within difference of 0.03 from the mea- sured transmittance. The implemented model was also used to calculate reflectance and absorbance, and the obtained values and parameter behavior with wavelength and tempera- 70 ture resembled that of previously measured optical properties in VO2 . This method can be taken as a universal approach for any multi-layer films, considering that the optical constants (n,k) are provided. Furthermore, VO2 based micrometer-sized windows were demonstrated with integrated resistive heaters that can be actuated independently to drive the VO2 film into different regions across its phase transition. Both major and minor loops for the optical transmittance were obtained across the phase transition. Emissivity curves were obtained as a function of temperature and applied current to the resistive heater. Due to the film’s hysteretic behavior, electronic pulses were used to program different emissivities states along the film’s surface. This was used to reconfigure thermal images into different geometrical shapes or numbers. Finite element method simulations were used to determine the heat distribution along the VO2 surface. 71 CHAPTER 6 DEFLECTION MODEL FOR ELECTRO-THERMALLY ACTUATED VO2 -BASED CANTILEVERS In this chapter we present VO2 /SiO2 based cantilevers which are actuated electro-thermally (i.e. via Joule heating). Deflections are achieved due to the difference in coefficients of thermal expansions (CTE) for both materials and VO2 ’s phase transition. Upon heating, VO2 ’s monoclinic (011)M crystal plane which is parallel to the SiO2 surface, reversibly changes to a rutile (110)R plane [141]. This phase transition will induce a change in the CTE of the material which will be a contributing factor towards a bending moment. This in turn changes the curvature of the structure. This structural change along with several other varying properties such as Young’s modulus (E ), contributes to the overall achieved deflection in the structure. Using a high-speed camera, we monitored the deflections that occur within a time range of 100 µs, and obtain information regarding the dynamics of the deflection. Furthermore, expanding on previously shown models, a transient temperature- deflection bimorph model that takes into consideration the changes in VO2 properties across the phase transition is developed. Using known geometric parameters and the properties of both materials, transient deflections are estimated from the model and compared to the empirical value. It should be mentioned that the present model is not based on curve fits to empirical results, as it has been done in previous studies [164, 51]. Instead, the present work uses the thermal and mechanical properties of the materials involved to obtain a model that uses the electrical current as input, its conversion to thermal domain, and its final output in the form of mechanical deflection of a cantilever beam. This transient temperature deflection model can be applied to any bimorph cantilever structure with integrated resistive heaters. An analysis for the transient temperature due to a current input is shown. For a known input of current, it is found that upon thermal equilibrium the cantilever structure will reach a steady-state deflection. This is due to the fact that conductive and convective losses balance 72 out the increase of temperature due to the input of current. 6.1 Results and Discussion 6.1.1 Effective Length of Cantilever Figure 3.6-B and C shows scanning electron microscope images (SEM, JEOL 7500F) for a cross-section of a device after cutting along the center of the 550 µm and 450 µm cantilevers, following the cut specified in Figure 3.6-A-f. The cantilevers were designed to have a length of 450 µm. However, due to over-etching, caused by the XeF2 gas used for isotropic etching of the substrate and structure release, the effective length of the cantilever will be much longer. In order to find the effective length of the cantilever after release, the resonant frequency of an un-coated cantilever (no VO2 deposited on top) was measured and the following relation was used [179]: 1.016 E2 a22 1 fu = ( )2 (6.1) 2πL2 ρ1 where E2 = 74 GPa corresponds to the Young’s modulus of SiO2 , ρ1 = 2200 g/m3 as shown by [180], and a2 = 2 µm is the thickness of the SiO2 layer as seen from the SEM (Figure 3.6-B) and fu = 4400 Hz as measured. Solving for L, gives us an effective length for the 450 µm which turned out to be Le f f = 652.29 µm, which will be later used for the proposed model shown. To verify the quality of the VO2 layer that was deposited over the cantilever structure, a measurement of resistance as a function of temperature was carried. Figure 3.7-b shows the hysteretic behavior for the resistance of VO2 . A drop of almost three orders of magnitude can be seen from the graph, which is indicative of the good quality of VO2 . VO2 ’s phase transition typically occurs for temperatures above 68o C. From the figure it can be shown that the drop in resistance starts above 71o C. This off-set can be attributed to the method of obtaining the measurement. The Peltier heater has to heat up the IC package in which the device is bonded too, which in turn causes a temperature off-set. 73 6.1.2 Heat-Transient Deflection Model Combining a steady-state deflection model for a bi-morph and a transient time heat equation, a transient deflection model for a bi-morph is developed. This model will allow for the calculation of the deflection of any bimorph structure, as long as the basic properties of the material such as Young’s modulus (E ), density (ρ), coefficient of thermal expansion (α) and thickness are known. The embedded heater inside the bi-metal structure will not be taken into consideration due the fact that the heater is thin enough to neglect its mechanical contribution. For this model, the deflection caused by Joule heating (resistive heating along the metal traces due to a current I) is also shown and will only take into consideration the SiO2 layer of 2 µm and the VO2 layer of 100 nm. Starting from a bi-morph model as proposed by [181], the curvature of a bi-metal structure can be expressed as: 1 [6(αV O 2 − αS iO 2 )∆T (1 + m)2 ] = , (6.2) ρ h[3(1 + m)2 + (1 + mn)(m2 + 1/mn)] where ρ is the curvature of the bi-metal structure, αV O 2 is the coefficient of thermal expansion of the top layer of the structure (VO2 for this case), and αS iO 2 is the coefficient of thermal expansion of our bottom layer (SiO2 ). The quantity m is the ratio of thicknesses a E m = a1 , and n = E1 is the ratio of Young’s modulus for both layers (VO2 and SiO2 ), 2 2 h=a1 +a2 is the height of the cantilever or the sum of the thicknesses and ∆T is the difference in temperature. From the geometry of our structure, we can derive a relation between the cantilever’s curvature and deflection, such that: 1 2δ = 2, (6.3) ρ L where δ is the deflection of the cantilever and L is the length of the cantilever (Le f f for this case). Taking equation (6.2) and substituting it into equation (6.3), we get an equation for deflection as follows: 74 3L2 (αV O 2 − αS iO 2 )∆T (1 + m)2 δ= (6.4) h[3(1 + m)2 + (1 + mn)(m2 + 1/mn)] where equation (6.4) describes the steady state deflection for a bi-metal structure. Now if we consider our cantilever structure with mass ms , volume V, density ρ and initial temper- ature Ti , and the material is placed into a medium with a temperature T∞ and coefficient of heat transfer h, then the heat transfer into the body for an interval dt must be such that the increase in energy of the body during the interval dt are the same. Thus: hCHT A(T∞ − T )dt = ms Cp dT (6.5) where Cp is the specific heat of the material (VO2 ) at constant pressure. Changing ms =ρV and integrating from T = 0 to T = Ti we get: T (t) = (Ti − T∞ )e−bt + T∞ (6.6) h A where b = CHT ρV Cp . Here A and V correspond to the superficial area and volume of the cantilever. Then, τ = 1b is the time constant for the system. Taking equations (6.4) and (6.6) and knowing that ∆T = T (t) − Ti ; an expression for a transient time deflection can be derived: 3L2 (αV O 2 (t) − αS iO 2 )(1 + m)2 δ(t) = (Ti (e−bt − 1) + T∞ (1 − e−bt )) (6.7) h[3(1 + m)2 + (1 + mn)(m2 + 1/mn)] where αV O 2 (t) depends on temperature (T(t)) and therefore time dependent. In order to find the values for the coefficient of thermal expansion (CTE) for VO2 during the transition, a linearization for the temperature range during the transition was performed using the values obtained from [182]. Before the phase transition (T < 341.15K) the average value for the CTE is αav = 5.70 × 10−6 K −1 , above it (T > 358.15)K) the average value is αav = 13.35 × 10−6 K −1 . For SiO2 the value [180] used for calculations was αSiO2 = 0.55 × 10−6 K −1 . To measure the temperature that the cantilever would reach at maximum 75 Table 6.1 : VO2 and SiO2 parameters used for the transient-heat-deflection model Properties VO2 (Insulating) VO2 (Metallic) SiO2 Density [kg/m3 ] 4760 4760 2200 Young’s Modulus [GPa] 140 [184] 155 [184] 74 Thickness of Layer [m] 1x10− 7 1x10− 7 2x10− 6 Coefficient of Thermal Expansion [1/K] 5.7x10−6 [182] 13.35x10−6 [182] 0.55x10−6 [180] Convective Heat Transfer [W/m2 K] - - 350 [183] deflection (T∞ ), a device was placed on a Peltier heater and the deflection was monitored while the temperature on the Peltier heater was increased through a temperature controller. Once the maximum deflection was reached, the temperature value was found to be T ∞ =89 0C (362.15 K). In order to obtain a value for the parameter b on equation (6.6), a reference value for the convective heat transfer coefficient (h CHT ) describing the flow of heat from the SiO2 layer to the VO2 layer was needed. A value for h CHT = 350 W/m2 K as reported by Baghban et al. [183] was chosen which would be the best fit with our reported time constant obtained from experiments (table (6.1)). For this model the dissipating heat flow from the VO2 and air interface will not be taken into consideration and only the heat flow from the SiO2 and VO2 will be taken into account. This is due to the fact that the heat flow from the VO2 /air interface is much slower than that of the SiO2 /VO2 interface during the phase transition. If this was not the case, then the cantilever structure would not be able to bend because the heat flow from the VO2 /air interface would be dissipating much faster than the heat flow provided by the SiO2 /VO2 interface. Figure 6.1-a shows the deflection as a function of time for the 450 µm x 50 µm cantilever actuated via Joule heating. This curve is obtained after analyzing the high-speed video that was taken simultaneously as the cantilever was actuated. The time constant as obtained from the video was found to be τ = 5.7 ms after subtracting the last frame of actuation with the first frame of actuation and dividing it by the frame rate (10,000 fps). For a full deflection of 55 µm a time constant can be obtained by finding the value of deflection at 67% (δ = 34.6 76 Figure 6.1 : Deflection as a function of time: (a) is obtained directly from the high-speed video. (b) as calculated from equation (6.7). 77 µm). A time constant of τ = 1.5 ms is obtained, in comparison with the calculated value of τ = 1.3 ms, for a percentage difference of 14.2%. Figure 6.1-b shows the calculated deflection using equation (6.7). Setting the initial temperature of the cantilever as room temperature (T i =298.15 K), and using a value for the Young’s modulus before phase transition of E 1 = 140 GPa, and after transition of E 1 = 155 GPa [184] for completion of the calculations. As for specific heat of VO2 a value [185] of C p = 730 J/kgK was used. A maximum deflection of 39 µm was obtained theoretically in comparison with the 55 µm measured. To obtain such deflections, a current input is required along the metal traces to start the process of Joule heating. It was found that the maximum deflection attained by the cantilever was at a current value of IM ax = 4.7mA. For a known resistor, the power dissipated through it is given by: P = I 2R (6.8) where R is the value of resistance. This equation is known as the Joule heating equation. The heat produced by this resistor can be described by: H = mCp ∆T (6.9) where m is the mass for the metal trace. A relation can be done by combining the energy produced by the Joule heating process and the heat produced, such that: I 2 Rt = ms Cp ∆T (6.10) From equation (6.10), a relation for the temperature of the cantilever due to the increase in the current is obtained with the form: I(t)2 Rt T (t) = Ti + (6.11) ms Cp where equation (6.11) describes the transient increase of the temperature inside the bi- morph structure due to the current. Equation (6.11) tells us that the temperature would keep increasing with time. Taking into consideration that there will be heat dissipating from 78 the cantilever in form of convection, conduction and radiation then equation (6.11) can be rewritten in the form; I(t)2 Rt T (t) = (Ti + ) − LF, (6.12) ms Cp where LF will be known as a heat loss factor that will include radiation, conductive and convective losses. From Stefan Boltzmann’s law, it is given that the radiated power is P=AσT4 . Since our device area is small, radiation losses can be neglected. As for convective and conductive losses, they are both proportional to the difference in temperature (∆T ). When the cantilever has reached thermal equilibrium (convective and conductive losses bal- ance out and help maintain the temperature), there will no longer be a change in deflection. Thus, a constant current will lead to a constant temperature which will lead to a constant deflection. This was observed by heating up the cantilever with half the current needed for total deflection, holding the current for one hour, and observing no change in the deflec- tion. The rate of change of deflection depends upon the transfer of heat within the bimorph structure. This is governed by two factors; Joule heating and the heat transient. In an adiabatic process (no convective and radiation losses), the rate of change of the temperature 2 due to Joule heating is given by ImC RT which is faster than the heat transient distribution. p The assumption of adiabatic process is valid due to the fact that heat has to travel through the cantilever before it is dissipated to the surrounding. This in turn means that heat loss governs the final deflection and not the rate of deflection. 6.2 Summary A transient heat deflection model was developed for a bimorph structure. By knowing the material properties and dimensions, deflection calculations can be obtained on par with experimental results. This transient heat model can be applied to any bimorph structure with integrated resistive heaters provided that the properties of the materials forming the structure are known. An analysis for the current input and its temperature transient distribution is 79 also provided. It is found that upon thermal equilibrium, the cantilever structure reaches a constant deflection due to the convective and conductive losses balancing out the temperature increase due to the constant current. A high-speed camera was used to obtained high frame rate (10,000 fps) videos for the deflection of a 450 µm x 50 µm VO2 based cantilever which is actuated via Joule heating. 80 CHAPTER 7 VO2 -BASED MICRO-ELECTRO-MECHANICAL TUNABLE OPTICAL SHUTTER AND MODULATOR In this chapter VO2 -based MEMS tunable optical shutters are demonstrated. The design consists of a VO2 -based cantilever attached to a VO2 -based optical window with integrated resistive heaters for individual mechanical actuation of the cantilever structure, tuning of the optical properties of the window, or both. Optical transmittance measurements as a function of current for both heaters demonstrates that the developed devices can be used as analog optical shutters, where the intensity of a light beam can be tuned to any value within the range of VO2 phase transition. A transmittance drop off 30% is shown for the optical window, with tuning capabilities greater than 30% upon actuation of the cantilever. Unlike typical mechanical shutters, these devices are not restricted to binary optical states. Optical modulation of the optical window is demonstrated with an oscillating electrical input. This produces a transmittance signal that oscillates around an average value within the range off VO2 ’s phase transition. For an input current signal with fixed amplitude (fel = 0.28 Hz), tuned to be at the onset of the phase transition, a transmittance modulation of 14% is shown. Similarly, by modulating the DC-offset, a transmittance modulation of VO2 along the hysteresis is obtained. 7.1 Results and Discussion 7.1.1 Heat Distribution Simulation Results and Thermal Imaging Results 7.1.1.1 Simulation Results Finite element method (FEM) simulations were performed using heat transfer and electric currents modules in Comsol Multiphysics software, where the temperature distribution for the VO2 -based tunable optical shutter was confirmed for different currents applied to the 81 independent heaters and simultaneously. Parameters for the insulating state such as density (ρ= 4760 kg/m3 ), relative permittivity (r = 1) and thermal conductivity (κ= 5 W/m K) were used for the insulating state of VO2 . Figure 7.1 shows the simulated temperature distribution for four cases. Figure 7.1-a shows the case where only the cantilever is actuated with a current of iCantilever = 6 mA with a maximum temperature of 72.7 0 C attained in the cantilever structure. By actuating the cantilever, heat will propagate to the lower portions of the window, with the highest temperature area ( ≈ 55o C) located after the VO2 gap that connects the cantilever and window structure. Figure 7.1-b demonstrates the case where only the window is actuated with a current of 16 mA. Since the window traces run along the length of the cantilever, actuation of the cantilever structure is will occur. With a current of 16 mA, a heat distribution in any of the quadrants of the window is above the phase transition temperature of VO2 which is enough to cause a transmittance change. Figure 7.1-c and d shows two cases where both heaters are actuated with Figure 7.1-c being the case where the window is actuated at the current value (iph = 10.5mA) necessary to place the transmittance at the onset of the phase transition. Similarly, Figure 7.1-d shows the heat distribution with the maximum window and cantilever actuation current. With a current of 16 mA, a maximum temperature inside the window greater than 100 0 C is obtained. 7.1.1.2 Thermal Images Thermal images of the VO2 -based optical shutter where taken using a an IR thermal mi- croscope (OptoTherm, Infrasight MI320). Current was sent to both metal traces using a power supply, and the thermal image was recorded using thermography computer controlled software (Thermalize-Thermal Image Analysis Software). Figure 7.2 shows the thermal images for both the window and cantilever structure with several current inputs. Figure 7.2-a shows the thermal image for the cantilever structure with a current input of 6 mA. Figure 7.2-b shows the thermal image for the optical window with a current input of 16 mA and a cantilever input of 0 mA. A temperature range from approximately 65 o C to 90 o C 82 Figure 7.1 : Temperature distribution for the VO2 -based optical shutter due to a current input. (a) Cantilever actuation current of 6 mA and window actuation current of 0 mA.(b) Cantilever actuation current of 0 mA and window actuation current of 16 mA. (c) Cantilever actuation current of 6 mA and window actuation current of 10.5 mA. da) Cantilever actuation current of 6 mA and window actuation current of 16 mA. is seen. Figure 7.2-c and d shows the thermal image for the case where both the cantilever and window are actuated. For Figure 7.2-c, a current of 6 mA is sent to the cantilever and a current of 10.5 mA is sent to the window (the temperature required to be at the onset of the phase transition). A temperature greater than 50 o C can be seen inside the optical window. For Figure 7.2-d, a current of 6 mA is sent to the cantilever and a current of 16 mA is sent to the window. When both resistive heaters are actuated with these values for current, the average temperature inside the structure is above 75 o C. S 83 Figure 7.2 : Temperature distribution for the VO2 -based optical shutter. (a) Cantilever actuation current of 6 mA and window actuation current of 0 mA.(b) Cantilever actuation current of 0 mA and window actuation current of 16 mA. (c) Cantilever actuation current of 6 mA and window actuation current of 10.5 mA. d) Cantilever actuation current of 6 mA and window actuation current of 16 mA. 7.1.2 Performance of VO2 -based tunable optical shutters Figure 7.3-a shows the transmittance of the VO2 window as a function of the actuation current of the window (bottom x-axis) and cantilever current (top x-axis). Actuating the resistive heater trace around the window results in a single hysteretic loop with a transmit- tance drop of 23% across the phase transition. This is achieved by focusing a beam spot with a diameter of ∼ 80 µm around one of the quadrants of the window, between two release squares. The purpose of this is to avoid the beam spot shining into one of the release squares of the window (see orange spot in Figure 3.9-b), which will cause the beam (or part of it) to travel through free-space and not provide a measurement of the transmittance through the window. A transmittance drop of 23 % ( or greater) for VO2 at a wavelength of 1550 nm is on par with previous reported results as seen in [33, 34, 174]. Actuation of the metal traces of the window will result in some curving of the window structure, due to structural 84 changes in VO2 films during its phase transition [164, 40]. As a result of this, the beam spot was placed either on the left or right side of the center of the window (shown as an orange spot in Figure 3.9-b) so that the position of the beam stays inside the window during actuation. Throughout the window-heating experiments, the beam spot moved along the window surface, avoiding the release holes. This was validated through preliminary exper- iments, and was done to obtain reliable measurements of the transmittance changes across the window only (and not through free-space outside the window or release holes). If the beam spot is located close to the edge of the window, then during the experiments, the only action that would remove the window from the beam path is the actuation of the cantilever structure. When the actuation current signal is sent to the cantilever metal trace, the struc- ture bends, creating an actuation mechanism that resembles that of a mechanical shutter (tilt control mechanism). Upon enough cantilever actuation, the window is driven away from the beam path, causing the transmittance to increase, since the beam travels through free-space at this point. A current below 6.7 mA for the cantilever was selected to avoid any damage to the metal traces and window structure. For an actuation current of 6.7 mA, an increase of 21 % was observed (see Figure 7.3-a). During the cantilever actuation, the increase on transmittance is purely from the change in position of the window with respect to the beam spot and is not affected by any effects that the input current might have on the window structure. Since the heat distribution from the cantilever is not enough to reach the top edge of the window, this cantilever actuation will not induce a change of phase in the window (refer to thermal images in supplementary material). It should be noted that the measured transmittance that corresponds to free-space is about 92 %. This 8% loss is attributed to a small portion of the beam profile still being blocked by the window, even after full cantilever actuation. For the cantilever actuation experiments, the beam spot is placed around the top edge of the window (red spot in Figure 3.9-b), in order to maximize the free-space beam path upon cantilever actuation. As the current input signal is sent, the cantilever structure moves downward. While this actuation is active, the beam spot will 85 start to move upward with some part of the beam being in free space and the other part being blocked by the structure. However, even after full cantilever actuation, the edges of the beam spot could still be in contact with the edge of the window, resulting in some power loss as seen from the detector (refer to supplementary material video-2). The power measured by the photo-detector in free-space without the focusing lens was around 0.232 mW, and the power measured just before the shutter was 0.163 mW. The transmittance (T ) is calculated as the ratio of the power measured before the window (P0 ) and after passing through the window (P ) (i.e. T =P/P0 ). For measurements where both the window and cantilever are actuated, the beam spot is lower in order to accommodate space for the beam spot to stay over the window’s surface during the actuation of the window traces. This in turn resulted in a lower initial value for the transmittance. Figure 7.3-b shows the transmittance drop of the window due to the actuation of the window traces with a transmittance drop of 28 %. Once the hysteresis reaches around 0.32 of transmittance, the cantilever heater is actuated with a current step of 6 mA. This actuation causes a deflection of the device, which in turn takes the beam spot out of the window. A transmittance increase of 37 % is followed by a similar drop once the actuation current step is no longer supplied to the cantilever traces. The transmittance then follows the cooling hysteretic loop and comes back to the initial 0.61 value (refer to supplementary material video-3 for actuation video). Since the hysteresis of the window is being measured, the beam spot must be placed around the sides of the middle release square of the window or higher (yellow spot in Figure 3.9-b). The beam spot will be driven to the edge of the window by the actuation of the window traces. Once the heating curve for the window is recorded (reaches a value of 31%), the cantilever is actuated (tilt control) with the current step (i= 6 mA), resulting in an increase of transmittance. Since the window moves downward upon actuation of the window traces, and the beam spot moves upward, the beam might experience some blockage by the window’s top edge when the cantilever is actuated. This results in an increase of transmittance, but not one as high in comparison to the one where only the cantilever traces are actuated. Another factor that 86 Figure 7.3 : Transmittance Measurements upon actuation of window and/or cantilever: (a) Steady-state transmittance results for separate actuation; i.e. actuation of resistive heaters for window or cantilever (tilt control). (b) Results for window actuation, followed by cantilever actuation with a current step of 6 mA. Insert shows the current actuation step supplied to the cantilever traces. (c) Transmittance minor loops due to actuation of the window heater. Insert shows the voltage input used to trace the minor loops. (d) Results for cantilever actuation (with a current step of 6 mA) after partial actuation of the window (x-axis). Dashed lines represent the transition points where the cantilever was actuated. Solid lines represent overall transmittance change. Legend shows the voltage/current point during the window hysteresis where the current step was applied. could lead to some variance on the initial values of transmittance is due to the curvature of the window, due to the intrinsic stress of the structure, the VO2 window is not a flat surface and possess some curvature. Other factors include the variance of the VO2 material across the window structure. By taking advantage of VO2 ’s intrinsic phase transition, programming of transmittance states can be achieved. Figure 7.3-c shows the hysteretic minor loops for the VO2 while 87 the window heater is actuated. A current pulse (not the same pulse used for the modulation section.) is programmed with not enough amplitude to completely transition across the hysteresis loop. This will cause the phase transition to return through one of the minor loops inside the hysteresis. The minor hysteretic loops cover a transmittance range from 51% to 29%, thus allowing for the programming of any transmittance state as long as it is within the bounds of hysteresis. Figure 7.3-d shows transmittance of the VO2 window while both heaters are independently actuated. First, the window heater is actuated up to a certain point during the transition (e.g. 7.6 mA, red curve on Figure 7.3-d, marked by red dashed line), which marks the trigger point for the actuation of the cantilever (tilt control) with a current step of 6 mA (same one shown in Figure 7.3-b insert). Once the cantilever current step is applied at the trigger point, a sudden increase in transmittance is measured (13% for the case of the 7.6 mA –see Figure 7.3-d). This is done for several points across the phase transition (actuation step for tilt control is marked by dashed lines on Figure 7.3-d and overall transmittance change is marked by full lines), with transmittance changes due to the actuation of the cantilever ranging from 13% to 45%. This successfully demonstrates the tunable capabilities of the optical shutter. 7.1.3 VO2 -based electro-optical tunable modulator Since VO2 intrinsic phase transition allows to program optical states within the hysteresis, then modulation of the optical signal is viable within the boundaries of the phase transition. Optical modulation usually involves a light beam that can be modulated by varying the current that is used to driving it, or a material that will act as a light modulator to either manipulate the phase, amplitude or frequency of the input beam [186]. Applications for such modulators range from power modulation for high speed communications to pulse laser generation [187]. Figure 7.4-a insert shows the input current (sinusoidal wave) required for optical modulation of transmittance for the VO2 window (no current input was sent to the cantilever traces). A pre-heating current of i ph = 10.5 mA is sent to the window heater. 88 Figure 7.4 : (a) Transmittance minor loops due to actuation of the window heater. Red lines represents limits of voltage values used for modulation. Green line represents the current value used at the onset of the phase transition. Insert shows the input current for modulation. A preheated current of 10.5 mA (4.2 V) is hold on the device (onset of phase transition), then a sine wave with an amplitude of 1 V and frequency of 0.28 Hz is sent. (b) Modulated transmittance of the optical window due to the current input. (c) FFT for input current and output power for the VO2 optical window. 89 This value of current is specifically chosen to be at the onset of the phase transition. This in turn will allow a better monitoring of the transmittance drop (Tph = 0.52) when the pulse is sent. The pulse used for modulation has an amplitude of VAmp = 1 V (i =2.7 mA) and a frequency of f el = 0.28 Hz. These values where chosen to facilitate data acquisition and to avoid any damage that higher frequencies could do to the window/cantilever structure. Figure 7.4-b shows the modulated transmittance for the VO2 window as a consequence of the sinusoidal electrical input sent via resistive heaters. Once the input is sent, an initial drop of 22% is seen, followed by the transmittance being modulated between 0.43 and 0.31. It can be noted that due to the sinusoidal input, the transmittance will follow a behavior that can be described as T M odulated = TAmp sin(ωt), with a transmittance modulation of 14 %. Figure 7.4-a shows the minor hysteretic loops for the VO2 window with the marked values for preheating current and limits of modulation. Once the initial pulse is sent, the transmittance will follow through the main heating loop and return through the cooling main loop. While the modulation continues, the transmittance will follow one of the minor heating loops from 0.43, and return through one of the cooling loops to 0.31 respectively. Since VO2 is a phase transition material and possess an intrinsic hysteresis with a change in its optical and electrical properties [45], the frequency response of the system will follow a non-linear response. By taking the average output power from the optical window due to the input current, and taking a Fast Fourier Transform, a more in-depth analysis of the output frequency can be accomplished. Figure 7.4-c shows the FFT for the input current (Figure 7.4-a insert) and the FFT for the output power from the optical window as measured by the optical detector within the range of 0 Hz to 2.5 Hz. An initial peak at f T r =0.28 Hz can be observed on both plots, which corresponds to the input frequency used for modulation (f el = 0.28 Hz). Second peak and third transmittance frequency peaks are found at two and three times the initial frequency, respectively – i.e. 2 × f T r = 0.56 Hz and 3 × f T r = 0.84 Hz for the output power. In order to verify the effects of VO2 ’s non-linearity on the input signal, the total harmonic distortion (THD) was obtained from the amplitude FFT for both 90 the input and output signal (refer to supplementary material for FFT plot). A total THD of 4.07% (-27.7 dB) was calculated for the input current signal and a THD of 14.10% (-17.01 dB) was obtained for the output signal. A total 10 % distortion is attributed to VO2 ’s intrinsic phase change (refer to supplementary material for the calculation of THD). A similar modulation along the phase transition is shown for the VO2 based optical window (iCantilever = 0 mA). Figure 7.5-a shows the current input while the amplitude and frequency are kept fixed and the DC offset is modulated. For the purpose of this experiment, the DC offset is started at the onset of the phase transition (4.2 V, 10.5 mA) and its modulated by steps of 0.2 V (0.6 mA) until reaching the end of the transition at 5.6V (14.7 mA). The DC offset is then swept along the minor loops (Figure 7.3-c) driving it across the phase transition. Once the maximum DC offset is reached, it is swept back to the initial value and it follows the cooling hysteretic minor loops. Figure 7.5-b shows such response due to the modulation of the DC offset effect on the transmittance of the device. As the DC offset is increased, a greater change in the modulation of the optical signal is observed, A change that can be noted by the slope and width of the minor loop hysteresis curve in Figure 7.3-c. 91 Figure 7.5 : (a) Input current with modulated DC offset and constant voltage amplitude with an input frequency of 1 Hz. Modulation is done for both heating and cooling cycles with an increment on the DC offset of 0.2 V (0.6 mA). (b) Transmittance with modulated DC offset and constant voltage amplitude (0.7 V) with an input frequency of 1 Hz. Transmittance is modulated for both heating and cooling cycles and it starts at the onset of the PT at 4.5 V or 10.5 mA. 7.2 Curvature and Deflection Analysis for a Trimorph Cantilever Figure 7.6 : Schematic of VO2 -based trimorph cantilever. A similar relation for the curvature of a composite beam of n layers can be further developed from the foundation from chapter 6. For a regular 2 layer bimorph, the curvature 92 is given by [188]: 0+z z +z 00 E2 (α2 − αef f )t2 ( 2 1 − z0 ) + E1 (α1 − αef f )t1 ( 1 2 2 − z0 ) w (x) = 3 ∆T, (7.1) E2 ((z1 − z0 )3 − (0 − z0 )3 ) + E1 ((z2 − z0 )3 − (z1 − z0 )3 ) where αef f is the coefficient of thermal expansion term that takes into consideration both layers of the bimorph and is given by: E1 t1 α1 + E2 t2 α2 αef f = . (7.2) E1 t1 + E2 t2 For equations 7.1 and 7.2, E 1 and E 2 are the Young’s modulus of layer 1 and layer 2. Similarly, α1 , α2 , t 1 and t 2 are the coefficient of thermal expansion and thicknesses for both layers. Finally, z, represents the plane height for each of the layers. For the case where the second layer is significantly thinner than the bottom layer (t 2 < < t 1 ), the curvature can be written as: 00 E ε t w (x) = 6 2 22 2 , (7.3) E1 t1 where ε is the strain of the layer. From the curvature the strain and residual stress can be obtained as follows: E1 t21 00 ε2 = w (x) (7.4) 6E2 t2 E1 t21 00 T2 = w (x). (7.5) 6t2 Equations 7.4 and 7.5 are known as Stoney’s equation and relates the bending of a cantilever beam to the stress or strain for the thin film [188]. For a 3 layer cantilever structure (trimorph), the curvature of the structure will be defined as: 00 N w (x) = , (7.6) D 93 Table 7.1 : VO2 and SiO2 parameters used for the trimorph curvature calculation. Material CTE [1/K] Layer Thickness [µm] Height [µm] VO2 (Insulating) 5.7x10− 6 0.16 1.16 VO2 (Metallic) 13.35x10−6 0.16 1.16 SiO2 (Bottom Layer) 0.55x10−6 1.0 1.0 SiO2 (Top Layer) 0.55x10−6 0.5 1.66 where N and D are the numerator and denominator, and are given by:  1 1 N = [∆T ] E1 (α1 − αef f )t1 [ (0 + z1 ) − z0 ] + E2 (α2 − αef f )t2 [ (z1 + z2 ) − z0 ]+ 2 2 1 E3 (α3 − αef f )t3 [ (z1 + z2 ) − z0 ] (7.7) 2 1 1 D = [(z1 − z0 )3 − (0 − z0 )3 ] + [(z2 − z0 )3 − (z1 − z0 )3 ]+ 3 3 [(z3 − z0 )3 − (z2 − z0 )3 ] (7.8) and the effective coefficient of thermal expansion is given by: E1 t1 α1 + E2 t2 α2 + E3 t3 α3 αef f = . (7.9) E1 t1 + E2 t2 + +E3 t3 For the case of the VO2 -based cantilever, the first layer will correspond to the 1 µm SiO2 layer, the second one corresponds to the 0.16 µm VO2 layer. Finally, the top layer corresponds to the top 0.5 µm SiO2 layer. Figure 7.6 shows the schematic for the trimorph layers with thickness values and table 7.1 shows the parameters used for calculation. In order to calculate the effects due to the phase change on the curvature of the cantilever trimorph, two effective coefficients of thermal expansions where calculated. For the insulating phase (αef f M =1.41x10−6 K−1 ) and for the metallic phase of VO2 (αef f R =2.89x10−6 K−1 ). 94 With both effective CTE’s, then an approximation for the curvature of the trimorph structure can be made from equation 7.6, such that: 00 w (x)M = ∆T (−1.83x10−6 )[K −1 µm−1 ] (7.10) 00 w (x)M = ∆T (−1.51x10−4 )[K −1 µm−1 ] (7.11) From the results obtained for the curvature, an analysis can be made for how the curvature of the structure will change as the phase change of VO2 is induced. By going from a monoclinic phase to a rutile phase, the curvature of the trimorph will evolve towards a less negative value of curvature for the rutile phase. This will result in an overall reduction of the curvature of the trimorph structure as the phase change is induced. By knowing the evolution of the curvature after the phase change, a relation between the trimorph curvature and deflection can be given by: L2 00 δ= w (x) (7.12) 2 From equation 7.12, two deflections can be approximated for both monoclinic and rutile phases. Deflection values where approximated to δM = -0.514 ∆T[K−1 ] µm and δR = - 14994.74 ∆T[K−1 ] µm respectively. Taking into consideration the residual stress presented on the trimorph structure upon release (which explains the overall curvature and initial deflection of the structure), then once the structure is actuated and the phase transition is induced with the change in temperature, then an overall downwards (negative) deflection is obtained. A similar relation for the deflection of the trimorph structure can be made in terms of Stoney’s equation 7.4 and 7.5 (in terms of stress and strain). Since the overall change on the trimorph curvature is negative (δR = -14994.74 ∆T[K−1 ] µm) once the phase 00 change is induced, then a direct relation can be made since w (x) ∝ T and ε, which would translate to a negative (compressive) stress and strain on the trimorph structure. 95 7.3 Summary VO2 -based MEMS optical shutters with continuous (or analog) transmittance tunability have been demonstrated. The optical shutters are based on the design of VO2 cantilevers and VO2 based optical windows with integrated resistive heaters for independent actuation. Using a wavelength of λ= 1550 nm, a transmittance drop of 23% along the hysteresis is shown while actuation of the optical window occurs. Furthermore, by actuating the VO2 based cantilever, a shuttering effect is achieved with a transmittance increase of 32%. While actuating both the window and cantilever traces, the transmittance change is no longer limited to the constraints of the hysteresis of the optical window and can be further tuned with the incorporation of the cantilever actuation. This modulated shuttering effect makes it tunable and no longer a binary shutter (0 or 1 outcome). A modulation of around 37% of transmittance is achieved for a wavelength of 1550 nm. Previous shutter technologies such as cellulose-based nano and microfiber thin films and cholesteric liquid crystal films [189, 190], have successfully shown and on/off state of transmittance. Since no phase-change material is used, then this type of optical shutter is limited to either a 1 or 0 state. In comparison to a VO2 -based shutter, that introduces an intrinsic phase change thus providing tuning capabil- ities at a transition temperature (TP T = 68 o C) close to room temperature. Transmittance modulation of the optical window was demonstrated by using a sinusoidal current input with a frequency of 0.28 Hz. By using a preheating current to drive the transmittance until the onset of the phase transition, a modulation of 14% is achieved. Likewise, a modulation along the whole of the phase transition hysteresis loop of the VO2 window is shown with a modulated input with fixed amplitude and frequency. By modulating the DC-offset in steps of 0.2 V (0.3 mA), the transmittance was modulated for both heating and cooling cycles. As the DC offset was increased, an increase on the change of the transmittance modulation was obtained with an overall change in transmittance of 21%. 96 CHAPTER 8 SUMMARY 8.1 Summary of Contributions This dissertation presents the development of vanadium dioxide (VO2 )-based smart win- dows, shape converters and tunable optical shutters devices. The devices have integrated resistve heaters for actuation which allows for modulation of transmittance and emissivity. Design and fabrication process is presented for the VO2 -based windows, shape converters, cantilevers and tunable optical shutters. Programming of transmittance and emissivity states is presented for the optical windows and thermal images for the device are presented for a programming current with the corresponding emissivity value and irradiance pattern. By designing and arranging different configurations of heaters, VO2 -based shape converters were designed and fabricated. Actuation of the independent heaters allows for demonstration of shape-converting for infrared light as seen by a thermal camera. A two interface optical model that combines the substrate and VO2 -layer with a reduced optical admittance is pre- sented. The model does not required direct measurements of the materials optical constants to estimate the assembly’s optical transmittance, reflectance and absorbance. The design and fabrication of VO2 -based cantilevers is presented along with the development of a tran- sient heat deflection model that could be applied to any electro-thermally actuated bimorph structure. Using a high-speed camera, deflections due to Joule heating are recorded and deflection measurements are extracted and compared to the numerical data from the model. Finally, VO2 -based MEMS tunable optical shutters were designed and fabricated. The com- bination of VO2 -based cantilevers and smart windows allows for the mechanical control that a regular mechanical shutter would give. By adding the smart window part to the shutter, the device is no longer restricted to a binary outcome and allows for the tunability of the optical conductivity within the boundaries of hysteresis given by VO2 . Optical modulation 97 of the optical window by a sinusoidal input is shown and control of the transmittance and programming of transmittance states is presented. 8.1.1 Problems Solved in this Thesis This work addresses the following: • VO2 thin films where incorporated into the design and fabrication of smart windows and shape converters. • By combining the design of VO2 -based cantilevers and smart windows, MEMS tunable optical shutters were designed and fabricated. • Characterization of the optical transmittance of the VO2 -based smart windows and programming of transmittance states. • Characterization of the emissivity of the VO2 -based smart windows and emissivity control for infrared light. • Development of a two-interface optical thin film model used to extract the optical transmittance, reflectance and absorbance for the VO2 thin film stack without direct need of the materials optical constants. • Demonstration of the shape-converting capabilities of VO2 -based smart windows with geometrically shaped embedded resisitive heaters for independent actuation. • Development of a transient heat deflection model for bimporh structures that takes into consideration the changes in VO2 ’s properties across the phase transition. • Demonstration of the optical tuning capabilites of the VO2 -based tunable optical shut- ter. • Development of steady-state curvature and deflection analysis for a trimorph cantilever structure. 98 BIBLIOGRAPHY 99 BIBLIOGRAPHY [1] M. Imada, A. Fujimori, and Y. Tokura, “Metal-insulator transitions,” Reviews of mod- ern physics, vol. 70, no. 4, p. 1039, 1998. [2] C. Ahn, J.-M. Triscone, and J. Mannhart, “Electric field effect in correlated oxide systems,” Nature, vol. 424, no. 6952, p. 1015, 2003. [3] F. Morin, “Oxides which show a metal-to-insulator transition at the neel temperature,” Physical review letters, vol. 3, no. 1, p. 34, 1959. [4] S. Lysenko, V. Vikhnin, F. Fernandez, A. Rua, and H. Liu, “Photoinduced insulator-to- metal phase transition in v o 2 crystalline films and model of dielectric susceptibility,” Physical Review B, vol. 75, no. 7, p. 075109, 2007. [5] M. Maaza, O. Nemraoui, C. Sella, A. Beye, and B. Baruch-Barak, “Thermal induced tunability of surface plasmon resonance in au–vo2 nano-photonics,” Optics Communi- cations, vol. 254, no. 1-3, pp. 188–195, 2005. [6] S. Lysenko, A. Rua, V. Vikhnin, J. Jimenez, F. Fernandez, and H. Liu, “Light-induced ultrafast phase transitions in vo2 thin film,” Applied Surface Science, vol. 252, no. 15, pp. 5512–5515, 2006. [7] P. Markov, R. E. Marvel, H. J. Conley, K. J. Miller, R. F. Haglund Jr, and S. M. Weiss, “Optically monitored electrical switching in vo2,” Acs Photonics, vol. 2, no. 8, pp. 1175–1182, 2015. [8] C. Marini, M. Bendele, B. Joseph, I. Kantor, M. Mitrano, O. Mathon, M. Baldini, L. Malavasi, S. Pascarelli, and P. Postorino, “Probing the electronic and local struc- tural changes across the pressure-induced insulator-to-metal transition in vo2,” EPL (Europhysics Letters), vol. 108, no. 3, p. 36003, 2014. [9] L. Bai, Q. Li, S. A. Corr, Y. Meng, C. Park, S. V. Sinogeikin, C. Ko, J. Wu, and G. Shen, “Pressure-induced phase transitions and metallization in vo 2,” Physical Review B, vol. 91, no. 10, p. 104110, 2015. [10] S. Lysenko, A. Rúa, V. Vikhnin, F. Fernández, and H. Liu, “Insulator-to-metal phase transition and recovery processes in v o 2 thin films after femtosecond laser excitation,” Physical Review B, vol. 76, no. 3, p. 035104, 2007. 100 [11] V. Vikhnin, S. Lysenko, A. Rua, F. Fernandez, and H. Liu, “Charge transfer model of metal–insulator phase transition and ultrafast optical response in vo2,” optical mate- rials, vol. 29, no. 11, pp. 1385–1389, 2007. [12] Z. Tao, T.-R. T. Han, S. D. Mahanti, P. M. Duxbury, F. Yuan, C.-Y. Ruan, K. Wang, and J. Wu, “Decoupling of structural and electronic phase transitions in vo 2,” Physical review letters, vol. 109, no. 16, p. 166406, 2012. [13] M. Gatti, F. Bruneval, V. Olevano, and L. Reining, “Understanding correlations in vanadium dioxide from first principles,” Physical review letters, vol. 99, no. 26, p. 266402, 2007. [14] S. Kumar, J. P. Strachan, A. D. Kilcoyne, T. Tyliszczak, M. D. Pickett, C. Santori, G. Gibson, and R. S. Williams, “The phase transition in vo2 probed using x-ray, visible and infrared radiations,” Applied Physics Letters, vol. 108, no. 7, p. 073102, 2016. [15] T. Yao, X. Zhang, Z. Sun, S. Liu, Y. Huang, Y. Xie, C. Wu, X. Yuan, W. Zhang, Z. Wu, et al., “Understanding the nature of the kinetic process in a vo 2 metal-insulator transition,” Physical review letters, vol. 105, no. 22, p. 226405, 2010. [16] A. Cavalleri, C. Tóth, C. W. Siders, J. Squier, F. Ráksi, P. Forget, and J. Kieffer, “Fem- tosecond structural dynamics in vo 2 during an ultrafast solid-solid phase transition,” Physical review letters, vol. 87, no. 23, p. 237401, 2001. [17] Y. Cao, W. Li, J. Figueroa, T. Wang, D. Torres, C. Wang, Z. L. Wang, and N. Sepul- veda, “Impact-activated programming of electro-mechanical resonators through ferro- electret nanogenerator (feng) and vanadium dioxide,” Nano Energy, vol. 43, pp. 278– 284, 2018. [18] D. Torres, T. Wang, J. Zhang, X. Zhang, S. Dooley, X. Tan, H. Xie, and N. Sepúlveda, “Vo 2-based mems mirrors,” Journal of Microelectromechanical Systems, vol. 25, no. 4, pp. 780–787, 2016. [19] R. Cabrera, E. Merced, and N. Sepúlveda, “Performance of electro-thermally driven vo2-based mems actuators,” vacuum, vol. 2013, p. 0042, 2014. [20] T. Wang, D. Torres, F. E. Fernández, A. J. Green, C. Wang, and N. Sepúlveda, “In- creasing efficiency, speed, and responsivity of vanadium dioxide based photothermally driven actuators using single-wall carbon nanotube thin-films,” ACS Nano, vol. 9, no. 4, pp. 4371–4378, 2015. PMID: 25853931. 101 [21] T. Wang, D. Torres, F. E. Fernández, C. Wang, and N. Sepúlveda, “Maximizing the performance of photothermal actuators by combining smart materials with supplemen- tary advantages,” Science Advances, vol. 3, no. 4, p. e1602697, 2017. [22] Z. Chen, Y. Gao, L. Kang, J. Du, Z. Zhang, H. Luo, H. Miao, and G. Tan, “Vo2-based double-layered films for smart windows: optical design, all-solution preparation and improved properties,” Solar energy materials and solar cells, vol. 95, no. 9, pp. 2677– 2684, 2011. [23] M. Soltani, M. Chaker, E. Haddad, R. Kruzelecky, J. Margot, P. Laou, and S. Paradis, “Fabrication of stationary micro-optical shutter based on semiconductor-to-metallic phase transition of w-doped vo 2 active layer driven by an external voltage,” Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, vol. 26, no. 4, pp. 763–767, 2008. [24] A. Hendaoui, N. Émond, S. Dorval, M. Chaker, and E. Haddad, “Vo2-based smart coat- ings with improved emittance-switching properties for an energy-efficient near room- temperature thermal control of spacecrafts,” Solar Energy Materials and Solar Cells, vol. 117, pp. 494–498, 2013. [25] S. Wang, M. Liu, L. Kong, Y. Long, X. Jiang, and A. Yu, “Recent progress in vo2 smart coatings: Strategies to improve the thermochromic properties,” Progress in Materials Science, vol. 81, pp. 1–54, 2016. [26] G. Xu, P. Jin, M. Tazawa, and K. Yoshimura, “Optimization of antireflection coating for vo2-based energy efficient window,” Solar Energy Materials and Solar Cells, vol. 83, no. 1, pp. 29–37, 2004. [27] P.-P. Shi, Q. Ye, Q. Li, H.-T. Wang, D.-W. Fu, Y. Zhang, and R.-G. Xiong, “Novel phase-transition materials coupled with switchable dielectric, magnetic, and optical properties:[(ch3) 4p][fecl4] and [(ch3) 4p][febr4],” Chemistry of Materials, vol. 26, no. 20, pp. 6042–6049, 2014. [28] J. Akola and R. Jones, “Structural phase transitions on the nanoscale: The crucial pattern in the phase-change materials ge 2 sb 2 te 5 and gete,” Physical Review B, vol. 76, no. 23, p. 235201, 2007. [29] X. Sun, E. Thelander, J. W. Gerlach, U. Decker, and B. Rauschenbach, “Crystallization kinetics of gete phase-change thin films grown by pulsed laser deposition,” Journal of Physics D: Applied Physics, vol. 48, no. 29, p. 295304, 2015. 102 [30] L. Perniola, V. Sousa, A. Fantini, E. Arbaoui, A. Bastard, M. Armand, A. Fargeix, C. Jahan, J.-F. Nodin, A. Persico, et al., “Electrical behavior of phase-change memory cells based on gete,” IEEE Electron Device Letters, vol. 31, no. 5, pp. 488–490, 2010. [31] C. Peng, L. Wu, F. Rao, Z. Song, X. Zhou, M. Zhu, B. Liu, D. Yao, S. Feng, P. Yang, et al., “Nitrogen incorporated gete phase change thin film for high-temperature data retention and low-power application,” Scripta Materialia, vol. 65, no. 4, pp. 327–330, 2011. [32] S. D. Ha, Y. Zhou, A. E. Duwel, D. W. White, and S. Ramanathan, “Quick switch: Strongly correlated electronic phase transition systems for cutting-edge microwave de- vices,” IEEE Microwave Magazine, vol. 15, no. 6, pp. 32–44, 2014. [33] J. Figueroa, Y. Cao, T. Wang, D. Torres, and N. Sepúlveda, “Programming emissivity on fully integrated vo2 windows,” Advance Material Letters, vol. 9, no. 6, pp. 406–410, 2018. [34] J. Figueroa, Y. Cao, H. Dsouza, J. Pastrana, and N. Sepúlveda, “A simplified approach for obtaining optical properties of vo2 thin films, and demonstration of infrared shape- shifting devices,” Advanced Materials Technologies, vol. 4, no. 4, p. 1800599, 2019. [35] K. Kato, P. K. Song, H. Odaka, and Y. Shigesato, “Study on thermochromic vo2 films grown on zno-coated glass substrates for “smart windows”,” Japanese journal of applied physics, vol. 42, no. 10R, p. 6523, 2003. [36] Y. Cui, Y. Ke, C. Liu, Z. Chen, N. Wang, L. Zhang, Y. Zhou, S. Wang, Y. Gao, and Y. Long, “Thermochromic vo2 for energy-efficient smart windows,” Joule, vol. 2, no. 9, pp. 1707–1746, 2018. [37] Z. Zhang, Y. Gao, H. Luo, L. Kang, Z. Chen, J. Du, M. Kanehira, Y. Zhang, and Z. L. Wang, “Solution-based fabrication of vanadium dioxide on f: Sno 2 substrates with largely enhanced thermochromism and low-emissivity for energy-saving applications,” Energy & Environmental Science, vol. 4, no. 10, pp. 4290–4297, 2011. [38] S. Babulanam, T. Eriksson, G. Niklasson, and C. Granqvist, “Thermochromic vo2 films for energy-efficient windows,” Solar energy materials, vol. 16, no. 5, pp. 347–363, 1987. [39] K.-K. Du, Q. Li, Y.-B. Lyu, J.-C. Ding, Y. Lu, Z.-Y. Cheng, and M. Qiu, “Control over emissivity of zero-static-power thermal emitters based on phase-changing material gst,” Light: Science & Applications, vol. 6, no. 1, p. e16194, 2017. 103 [40] R. Cabrera, E. Merced, and N. Sepúlveda, “A micro-electro-mechanical memory based on the structural phase transition of vo2,” physica status solidi (a), vol. 210, no. 9, pp. 1704–1711, 2013. [41] D. Torres, T. Wang, J. Zhang, S. Dooley, X. Tan, and N. Sepúlveda, “Experimental characterization of the dynamics of vo2-based mems mirrors,” in Smart Materials, Adaptive Structures and Intelligent Systems, vol. 50480, p. V001T02A005, American Society of Mechanical Engineers, 2016. [42] N. Manca, T. Kanki, F. Endo, D. Marré, and L. Pellegrino, “Planar nanoactuators based on vo2 phase transition,” Nano Letters, vol. 20, no. 10, pp. 7251–7256, 2020. [43] R. Powell, C. Berglund, and W. Spicer, “Photoemission from v o 2,” Reviews of Modern Physics, vol. 40, no. 4, p. 737, 1968. [44] A. Barker Jr, H. Verleur, and H. Guggenheim, “Infrared optical properties of vanadium dioxide above and below the transition temperature,” Physical Review Letters, vol. 17, no. 26, p. 1286, 1966. [45] H. Kakiuchida, P. Jin, S. Nakao, and M. Tazawa, “Optical properties of vanadium dioxide film during semiconductive–metallic phase transition,” Japanese Journal of Applied Physics, vol. 46, no. 2L, p. L113, 2007. [46] D. Torres, J. Zhang, S. Dooley, X. Tan, and N. Sepúlveda, “Modeling of mems mirrors actuated by phase-change mechanism,” Micromachines, vol. 8, no. 5, p. 138, 2017. [47] D. Torres, J. Zhang, S. Dooley, X. Tan, and N. Sepúlveda, “Hysteresis-based mechan- ical state programming of mems mirrors,” Journal of Microelectromechanical Systems, vol. 27, no. 2, pp. 344–354, 2018. [48] M. Lara-Castro, A. Herrera-Amaya, M. Escarola-Rosas, M. Vázquez-Toledo, F. López- Huerta, L. Aguilera-Cortés, and A. Herrera-May, “Design and modeling of polysilicon electrothermal actuators for a mems mirror with low power consumption,” Microma- chines, vol. 8, no. 7, p. 203, 2017. [49] J. Morrison, M. Imboden, T. D. Little, and D. Bishop, “Electrothermally actuated tip- tilt-piston micromirror with integrated varifocal capability,” Optics express, vol. 23, no. 7, pp. 9555–9566, 2015. [50] M. Boudaoud, Y. Haddab, Y. Le Gorrec, and P. Lutz, “Noise characterization in millimeter sized micromanipulation systems,” Mechatronics, vol. 21, no. 6, pp. 1087– 1097, 2011. 104 [51] E. Merced, J. Zhang, X. Tan, and N. Sepúlveda, “Robust control of vo 2-coated mi- croactuators based on self-sensing feedback,” in 2013 IEEE/ASME International Con- ference on Advanced Intelligent Mechatronics, pp. 656–661, IEEE, 2013. [52] S. Saghir and M. I. Younis, “An investigation of the static and dynamic behavior of electrically actuated rectangular microplates,” International Journal of Non-Linear Mechanics, vol. 85, pp. 81–93, 2016. [53] M. A. Rosa, D. De Bruyker, A. R. Völkel, E. Peeters, and J. Dunec, “A novel external electrode configuration for the electrostatic actuation of mems based devices,” Journal of Micromechanics and Microengineering, vol. 14, no. 4, p. 446, 2004. [54] J.-E. Wong, J. H. Lang, and M. A. Schmidt, “An electrostatically-actuated mems switch for power applications,” in Proceedings IEEE Thirteenth Annual International Conference on Micro Electro Mechanical Systems (Cat. No. 00CH36308), pp. 633–638, IEEE, 2000. [55] A. Kärkkäinen, A. Oja, J. Kyynäräinen, H. Kuisma, and H. Seppä, “Stability of elec- trostatic actuation of mems,” Physica Scripta, vol. 2004, no. T114, p. 193, 2004. [56] B. Svilicic, E. Mastropaolo, B. Flynn, and R. Cheung, “Electrothermally actuated and piezoelectrically sensed silicon carbide tunable mems resonator,” IEEE electron device letters, vol. 33, no. 2, pp. 278–280, 2012. [57] T. Moulton and G. Ananthasuresh, “Micromechanical devices with embedded electro- thermal-compliant actuation,” Sensors and Actuators A: Physical, vol. 90, no. 1-2, pp. 38–48, 2001. [58] D. Yan, A. Khajepour, and R. Mansour, “Design and modeling of a mems bidirectional vertical thermal actuator,” Journal of Micromechanics and Microengineering, vol. 14, no. 7, p. 841, 2004. [59] C.-Y. Lee, H.-T. Chang, and C.-Y. Wen, “A mems-based valveless impedance pump utilizing electromagnetic actuation,” Journal of Micromechanics and Microengineer- ing, vol. 18, no. 3, p. 035044, 2008. [60] I.-J. Cho, T. Song, S.-H. Baek, and E. Yoon, “A low-voltage and low-power rf mems series and shunt switches actuated by combination of electromagnetic and electro- static forces,” IEEE Transactions on Microwave Theory and Techniques, vol. 53, no. 7, pp. 2450–2457, 2005. 105 [61] H.-C. Lee, J.-Y. Park, and J.-U. Bu, “Piezoelectrically actuated rf mems dc contact switches with low voltage operation,” IEEE microwave and wireless components letters, vol. 15, no. 4, pp. 202–204, 2005. [62] C.-B. Eom and S. Trolier-McKinstry, “Thin-film piezoelectric mems,” Mrs Bulletin, vol. 37, no. 11, pp. 1007–1017, 2012. [63] J. Y. Park, Y. J. Yee, H. J. Nam, and J. U. Bu, “Micromachined rf mems tunable ca- pacitors using piezoelectric actuators,” in 2001 IEEE MTT-S International Microwave Sympsoium Digest (Cat. No. 01CH37157), vol. 3, pp. 2111–2114, IEEE, 2001. [64] M. C. Huang, K. B. Cheng, Y. Zhou, B. Pesala, C. J. Chang-Hasnain, and A. P. Pisano, “Demonstration of piezoelectric actuated gaas-based mems tunable vcsel,” IEEE photonics technology letters, vol. 18, no. 10, pp. 1197–1199, 2006. [65] T. Wang, D. Torres, F. E. Fernández, A. J. Green, C. Wang, and N. Sepúlveda, “In- creasing efficiency, speed, and responsivity of vanadium dioxide based photothermally driven actuators using single-wall carbon nanotube thin-films,” ACS nano, vol. 9, no. 4, pp. 4371–4378, 2015. [66] E. Merced, N. Dávila, D. Torres, R. Cabrera, F. E. Fernández, and N. Sepúlveda, “Pho- tothermal actuation of v o2: Cr-coated microcantilevers in air and aqueous media,” Smart Materials and Structures, vol. 21, no. 10, p. 105009, 2012. [67] R. Cabrera, E. Merced, N. Sepúlveda, and F. E. Fernández, “Dynamics of photother- mally driven vo2-coated microcantilevers,” Journal of Applied Physics, vol. 110, no. 9, p. 094510, 2011. [68] J. B. Sampsell, “Micro-mechanical optical shutter,” Oct. 17 1995. US Patent 5,459,602. [69] D. L. Stephenson and J. P. Towhey, “Pivotable optical shutter for blocking emission from a lightguide adapter# 5,” Nov. 11 1997. US Patent 5,687,268. [70] A. A. Ahmed, “Oscillating photovoltaic optical shutter for reflective display,” Oct. 6 1992. US Patent 5,153,760. [71] T. Nguyen, C. H. Hong, N. T. Le, and Y. M. Jang, “High-speed asynchronous optical camera communication using led and rolling shutter camera,” in 2015 Seventh Inter- national Conference on Ubiquitous and Future Networks, pp. 214–219, IEEE, 2015. [72] S. A. Carlson, “Optical shutter,” Feb. 25 2003. US Patent 6,525,859. 106 [73] M. J. Sieben, J. Conradi, and D. E. Dodds, “Optical modulation system,” Mar. 9 1999. US Patent 5,880,870. [74] W. N. Carr and X.-Q. Sun, “Optical microshutter array,” July 14 1998. US Patent 5,781,331. [75] L. Lipton, “High dynamic range electro-optical shutter for steroscopic and other ap- plications,” May 26 1992. US Patent 5,117,302. [76] C. H. Grove, “Optical shutter switching matrix,” Mar. 20 1990. US Patent 4,910,396. [77] R. M. Powers and W. McCarthy, “Thermally switched reflective optical shutter,” July 13 2010. US Patent 7,755,829. [78] G. W. Kim, Y. C. Kim, I. J. Ko, J. H. Park, H. W. Bae, R. Lampande, and J. H. Kwon, “High-performance electrochromic optical shutter based on fluoran dye for visibility enhancement of augmented reality display,” Advanced Optical Materials, vol. 6, no. 11, p. 1701382, 2018. [79] D. M. Allen, “Lcd optic shutter,” June 2 2009. US Patent 7,540,644. [80] M. R. Hammer, M. K. Masters, S. R. Campbell, and P. G. Layton, “Optical shutter for spectroscopy instrument,” June 22 2004. US Patent 6,753,959. [81] B. A. Mathews and B. H. Coon, “Electro-optical reconnaissance system with forward motion compensation,” Apr. 16 2002. US Patent 6,373,522. [82] P.-W. Huang, B. Tang, Z.-Y. Xiong, J.-Q. Zhong, J. Wang, and M.-S. Zhan, “Note: A compact low-vibration high-performance optical shutter for precision measurement experiments,” Review of Scientific Instruments, vol. 89, no. 9, p. 096111, 2018. [83] C. H. Lee, B. Bihari, R. Filler, and B. K. Mandal, “New azobenzene non-linear optical materials for eye and sensor protection,” Optical Materials, vol. 32, no. 1, pp. 147–153, 2009. [84] W. Moreshead, J.-L. Nogues, and D. McBranch, “Optical limiting windows for eye and sensor protection from laser radiation.,” tech. rep., GELTECH INC ALACHUA FL, 1997. [85] I. Samuel and S. O’Riorden, “Fiber optic protective shutter,” Jan. 1 2008. US Patent 7,315,682. 107 [86] D. J. Zellak, “Internal shutter for optical adapters,” July 22 2003. US Patent 6,595,696. [87] D. Johansson, “Vo2 films as active infrared shutters,” 2006. [88] S. T. Heinilehto, J. H. Lappalainen, H. M. Jantunen, and V. Lantto, “Ir-wavelength optical shutter based on ito/vo2 /ito thin film stack,” Journal of electroceramics, vol. 27, no. 1, pp. 7–12, 2011. [89] M. Soltani, M. Chaker, E. Haddad, R. Kruzelecky, and J. Margot, “Station- ary hadamard shutter based on semiconductor-to-metallic phase transition of ther- mochromic w-doped vo2 arrays,” [90] S. Chen, Z. Wang, H. Ren, Y. Chen, W. Yan, C. Wang, B. Li, J. Jiang, and C. Zou, “Gate-controlled vo2 phase transition for high-performance smart windows,” Science advances, vol. 5, no. 3, p. eaav6815, 2019. [91] S. J. Lee, D. S. Choi, S. H. Kang, W. S. Yang, S. Nahm, S. H. Han, and T. Kim, “Vo2/wo3-based hybrid smart windows with thermochromic and electrochromic prop- erties,” ACS Sustainable Chemistry & Engineering, vol. 7, no. 7, pp. 7111–7117, 2019. [92] L. Xiao, H. Ma, J. Liu, W. Zhao, Y. Jia, Q. Zhao, K. Liu, Y. Wu, Y. Wei, S. Fan, et al., “Fast adaptive thermal camouflage based on flexible vo2/graphene/cnt thin films,” Nano letters, vol. 15, no. 12, pp. 8365–8370, 2015. [93] V. Eyert, “The metal-insulator transitions of vo2: A band theoretical approach,” An- nalen der Physik, vol. 11, no. 9, pp. 650–704, 2002. [94] J. M. Longo, P. Kierkegaard, et al., “A refinement of the structure of vo2,” Acta Chem. Scand, vol. 24, no. 2, pp. 420–426, 1970. [95] R. M. Wentzcovitch, W. W. Schulz, and P. B. Allen, “Vo 2: Peierls or mott-hubbard? a view from band theory,” Physical review letters, vol. 72, no. 21, p. 3389, 1994. [96] H. Kim, H. You, R. Chiarello, H. Chang, T. Zhang, and D. Lam, “Finite-size effect on the first-order metal-insulator transition in vo 2 films grown by metal-organic chemical- vapor deposition,” Physical Review B, vol. 47, no. 19, p. 12900, 1993. [97] D. Ruzmetov, S. D. Senanayake, and S. Ramanathan, “X-ray absorption spectroscopy of vanadium dioxide thin films across the phase-transition boundary,” Physical Review B, vol. 75, no. 19, p. 195102, 2007. 108 [98] M. Kang, S. W. Kim, J.-W. Ryu, and T. Noh, “Optical properties for the mott tran- sition in vo2,” AIP Advances, vol. 2, no. 1, p. 012168, 2012. [99] T. Rice, H. Launois, and J. Pouget, “Comment on” v o 2: Peierls or mott-hubbard? a view from band theory”,” Physical review letters, vol. 73, no. 22, p. 3042, 1994. [100] H.-T. Kim, B.-G. Chae, D.-H. Youn, S.-L. Maeng, G. Kim, K.-Y. Kang, and Y.-S. Lim, “Mechanism and observation of mott transition in vo2-based two-and three-terminal devices,” New Journal of Physics, vol. 6, no. 1, p. 52, 2004. [101] M. M. Qazilbash, M. Brehm, B.-G. Chae, P.-C. Ho, G. O. Andreev, B.-J. Kim, S. J. Yun, A. Balatsky, M. Maple, F. Keilmann, et al., “Mott transition in vo2 revealed by infrared spectroscopy and nano-imaging,” Science, vol. 318, no. 5857, pp. 1750–1753, 2007. [102] B.-J. Kim, Y. W. Lee, S. Choi, J.-W. Lim, S. J. Yun, H.-T. Kim, T.-J. Shin, and H.-S. Yun, “Micrometer x-ray diffraction study of vo 2 films: Separation between metal- insulator transition and structural phase transition,” Physical Review B, vol. 77, no. 23, p. 235401, 2008. [103] A. Cavalleri, “All at once,” Science, vol. 318, no. 5851, pp. 755–756, 2007. [104] A. P. Peter, K. Martens, G. Rampelberg, M. Toeller, J. M. Ablett, J. Meersschaut, D. Cuypers, A. Franquet, C. Detavernier, J.-P. Rueff, et al., “Metal-insulator transi- tion in ald vo2 ultrathin films and nanoparticles: Morphological control,” Advanced Functional Materials, vol. 25, no. 5, pp. 679–686, 2015. [105] F. C. Case, “Improved vo 2 thin films for infrared switching,” Applied optics, vol. 30, no. 28, pp. 4119–4123, 1991. [106] K. Rogers, J. Coath, and M. Lovell, “Characterization of epitaxially grown films of vanadium oxides,” Journal of applied physics, vol. 70, no. 3, pp. 1412–1415, 1991. [107] A. Rúa, R. D. Dı́az, N. Kumar, S. Lysenko, and F. E. Fernández, “Metal-insulator transition and nonlinear optical responseof sputter-deposited v3o5 thin films,” Journal of Applied Physics, vol. 121, no. 23, p. 235302, 2017. [108] M. Borek, F. Qian, V. Nagabushnam, and R. Singh, “Pulsed laser deposition of oriented vo2 thin films on r-cut sapphire substrates,” Applied physics letters, vol. 63, no. 24, pp. 3288–3290, 1993. 109 [109] S. Lysenko, A. Rúa, J. Figueroa, L. Chevres, and F. E. Fernández, “Ultrafast light scat- tering by transient inhomogeneities in vanadium dioxide,” in International Conference on Ultrafast Phenomena, pp. UW4A–15, Optical Society of America, 2016. [110] D. Kim and H. Kwok, “Pulsed laser deposition of vo2 thin films,” Applied physics letters, vol. 65, no. 25, pp. 3188–3190, 1994. [111] X. Tan, T. Yao, R. Long, Z. Sun, Y. Feng, H. Cheng, X. Yuan, W. Zhang, Q. Liu, C. Wu, et al., “Unraveling metal-insulator transition mechanism of vo 2 triggered by tungsten doping,” Scientific reports, vol. 2, p. 466, 2012. [112] K. L. Holman, T. McQueen, A. Williams, T. Klimczuk, P. Stephens, H. W. Zandber- gen, Q. Xu, F. Ronning, and R. J. Cava, “Insulator to correlated metal transition in v 1- x mo x o 2,” Physical Review B, vol. 79, no. 24, p. 245114, 2009. [113] H. Kim, N. Charipar, J. Figueroa, N. Bingham, and A. Piqué, “Control of metal- insulator transition temperature in vo2 thin films grown on ruo2/tio2 templates by strain modification,” AIP Advances, vol. 9, no. 1, p. 015302, 2019. [114] S. Fischer, J.-O. Krisponeit, M. Foerster, L. Aballe, J. Falta, and J. I. Flege, “Massively strained vo2 thin film growth on ruo2,” Crystal Growth & Design, vol. 20, no. 4, pp. 2734–2741, 2020. [115] P. Willmott and J. Huber, “Pulsed laser vaporization and deposition,” Reviews of Modern Physics, vol. 72, no. 1, p. 315, 2000. [116] T. Wang, Integration of Carbon Nanotube and Vanadium Dioxide Thin Films in Photo- Thermal Actuators. Michigan State University, 2017. [117] I. Optotherm, “Planck’s law,” 2018. [118] I. Optotherm, “Physics of emissivity,” 2018. [119] I. Optotherm, “What is emissivity,” 2018. [120] L. Kang, Y. Gao, H. Luo, J. Wang, B. Zhu, Z. Zhang, J. Du, M. Kanehira, and Y. Zhang, “Thermochromic properties and low emissivity of zno: Al/vo2 double- layered films with a lowered phase transition temperature,” Solar Energy Materials and Solar Cells, vol. 95, no. 12, pp. 3189–3194, 2011. 110 [121] H. Ji, D. Liu, C. Zhang, and H. Cheng, “Vo2/zns core-shell nanoparticle for the adap- tive infrared camouflage application with modified color and enhanced oxidation resis- tance,” Solar Energy Materials and Solar Cells, vol. 176, pp. 1–8, 2018. [122] M. A. Kats, R. Blanchard, S. Zhang, P. Genevet, C. Ko, S. Ramanathan, and F. Ca- passo, “Vanadium dioxide as a natural disordered metamaterial: perfect thermal emis- sion and large broadband negative differential thermal emittance,” Physical Review X, vol. 3, no. 4, p. 041004, 2013. [123] S. Vassant, I. Moldovan Doyen, F. Marquier, F. Pardo, U. Gennser, A. Cavanna, J.- L. Pelouard, and J.-J. Greffet, “Electrical modulation of emissivity,” Applied Physics Letters, vol. 102, no. 8, p. 081125, 2013. [124] H. Ji, D. Liu, H. Cheng, C. Zhang, L. Yang, and D. Ren, “Infrared thermochromic properties of monoclinic vo 2 nanopowders using a malic acid-assisted hydrothermal method for adaptive camouflage,” RSC Advances, vol. 7, no. 9, pp. 5189–5194, 2017. [125] Z. J. Coppens and J. G. Valentine, “Spatial and temporal modulation of thermal emission,” Advanced Materials, vol. 29, no. 39, p. 1701275, 2017. [126] Z. Zhang, Y. Gao, Z. Chen, J. Du, C. Cao, L. Kang, and H. Luo, “Thermochromic vo2 thin films: solution-based processing, improved optical properties, and lowered phase transformation temperature,” Langmuir, vol. 26, no. 13, pp. 10738–10744, 2010. [127] N. Davila, E. Merced, and N. Sepulveda, “Electronically variable optical attenuator enabled by self-sensing in vanadium dioxide,” IEEE Photonics Technology Letters, vol. 26, no. 10, pp. 1011–1014, 2014. [128] N. Dávila, R. Cabrera, and N. Sepúlveda, “Programming and projection of near ir images using vo2 films,” IEEE Photonics Technology Letters, vol. 24, no. 20, pp. 1830– 1833, 2012. [129] N. Puzikova, E. Uvarova, I. Filyaev, and L. Yarovaya, “Principles of an approach for coloring military camouflage,” Fibre Chemistry, vol. 40, no. 2, pp. 155–159, 2008. [130] S. Kovacevic, I. Gudlin Schwarz, and V. Durasevic, “Analysis of printed fabrics for military camouflage clothing,” Fibres & Textiles in Eastern Europe, 2012. [131] I. C. Cuthill, M. Stevens, J. Sheppard, T. Maddocks, C. A. Párraga, and T. S. Tros- cianko, “Disruptive coloration and background pattern matching,” Nature, vol. 434, no. 7029, p. 72, 2005. 111 [132] J. H. Ko, S. Yeo, J. H. Park, J. Choi, C. Noh, and S. U. Son, “Graphene-based elec- trochromic systems: the case of prussian blue nanoparticles on transparent graphene film,” Chemical Communications, vol. 48, no. 32, pp. 3884–3886, 2012. [133] Z. Huang, S. Chen, C. Lv, Y. Huang, and J. Lai, “Infrared characteristics of vo2 thin films for smart window and laser protection applications,” Applied Physics Letters, vol. 101, no. 19, p. 191905, 2012. [134] Y. Gao, H. Luo, Z. Zhang, L. Kang, Z. Chen, J. Du, M. Kanehira, and C. Cao, “Nanoce- ramic vo2 thermochromic smart glass: A review on progress in solution processing,” Nano Energy, vol. 1, no. 2, pp. 221–246, 2012. [135] S. Chen, H. Ma, X. Yi, T. Xiong, H. Wang, and C. Ke, “Smart vo2 thin film for protection of sensitive infrared detectors from strong laser radiation,” Sensors and Actuators A: Physical, vol. 115, no. 1, pp. 28–31, 2004. [136] X. Chen, Q. Lv, and X. Yi, “Smart window coating based on nanostructured vo2 thin film,” Optik, vol. 123, no. 13, pp. 1187–1189, 2012. [137] R. L. Voti, M. Larciprete, G. Leahu, C. Sibilia, and M. Bertolotti, “Optimization of thermochromic vo2 based structures with tunable thermal emissivity,” Journal of Applied Physics, vol. 112, no. 3, p. 034305, 2012. [138] J. Zheng, S. Bao, and P. Jin, “Tio2 (r)/vo2 (m)/tio2 (a) multilayer film as smart window: Combination of energy-saving, antifogging and self-cleaning functions,” Nano Energy, vol. 11, pp. 136–145, 2015. [139] S. Chandra, D. Franklin, J. Cozart, A. Safaei, and D. Chanda, “Adaptive multispectral infrared camouflage,” ACS Photonics, vol. 5, no. 11, pp. 4513–4519, 2018. [140] G. Leahu, R. Li Voti, C. Sibilia, and M. Bertolotti, “Anomalous optical switching and thermal hysteresis during semiconductor-metal phase transition of vo2 films on si substrate,” Applied Physics Letters, vol. 103, no. 23, p. 231114, 2013. [141] A. Rúa, F. E. Fernández, and N. Sepúlveda, “Bending in vo 2-coated microcantilevers suitable for thermally activated actuators,” Journal of Applied Physics, vol. 107, no. 7, p. 074506, 2010. [142] E. Merced, N. Dávila, D. Torres, F. E. Fernández, and N. Sepúlveda, “Dynamics of vo 2: Cr coated microcantilevers in aqueous media,” in 2012 IEEE Nanotechnology Materials and Devices Conference (NMDC2012), pp. 100–104, IEEE, 2012. 112 [143] J. Zhang, D. Torres, E. Merced, N. Sepúlveda, and X. Tan, “A hysteresis-compensated self-sensing scheme for vanadium dioxide-coated microactuators,” in ASME 2014 Dy- namic Systems and Control Conference, American Society of Mechanical Engineers Digital Collection, 2014. [144] A. Rúa, R. Cabrera, H. Coy, E. Merced, N. Sepúlveda, and F. E. Fernández, “Phase transition behavior in microcantilevers coated with m1-phase vo2 and m2-phase vo2: Cr thin films,” Journal of Applied Physics, vol. 111, no. 10, p. 104502, 2012. [145] R. Cabrera, E. Merced, N. Dávila, F. E. Fernández, and N. Sepúlveda, “A multiple- state micro-mechanical programmable memory,” Microelectronic Engineering, vol. 88, no. 11, pp. 3231–3234, 2011. [146] K. Singer, S. Jochim, M. Mudrich, A. Mosk, and M. Weidemüller, “Low-cost me- chanical shutter for light beams,” Review of scientific instruments, vol. 73, no. 12, pp. 4402–4404, 2002. [147] G. K. Starkweather and M. J. Sinclair, “Microelectrical mechanical structure (mems) optical modulator and optical display system,” Nov. 22 2005. US Patent 6,967,761. [148] T. Silva, R. Vicente, F. Rodrigues, A. Samagaio, and C. Cardoso, “Performance of a window shutter with phase change material under summer mediterranean climate conditions,” Applied Thermal Engineering, vol. 84, pp. 246–256, 2015. [149] R. Sharma, P. Ganesan, V. Tyagi, H. Metselaar, and S. Sandaran, “Developments in organic solid–liquid phase change materials and their applications in thermal energy storage,” Energy Conversion and Management, vol. 95, pp. 193–228, 2015. [150] M. Jafari and M. Rais-Zadeh, “A 1550 nm phase change electro-optical shutter,” in 2016 IEEE 29th International Conference on Micro Electro Mechanical Systems (MEMS), pp. 655–658, IEEE, 2016. [151] E. R. Kocher and R. P. Novak, “Fast electro-optical shutter with programmable trans- mission control,” Journal of Physics E: Scientific Instruments, vol. 13, no. 5, p. 542, 1980. [152] Y.-H. Park, Y.-C. Cho, J.-W. You, C.-Y. Park, H.-S. Yoon, S.-H. Lee, J.-O. Kwon, S.-W. Lee, B. H. Na, G. W. Ju, et al., “Three-dimensional imaging using fast micro- machined electro-absorptive shutter,” Journal of Micro/Nanolithography, MEMS, and MOEMS, vol. 12, no. 2, p. 023011, 2013. 113 [153] W. Schwenger and J. M. Higbie, “High-speed acousto-optic shutter with no optical frequency shift,” Review of Scientific Instruments, vol. 83, no. 8, p. 083110, 2012. [154] M. Jafari and M. Rais-Zadeh, “Zero-static-power phase-change optical modulator,” Optics letters, vol. 41, no. 6, pp. 1177–1180, 2016. [155] M. Jafari, L. J. Guo, and M. Rais-Zadeh, “An ultra-fast optical shutter exploiting total light absorption in a phase change material,” in Optical components and materials XIV, vol. 10100, p. 101000I, International Society for Optics and Photonics, 2017. [156] D. H. Werner, T. S. Mayer, C. Rivero-Baleine, N. Podraza, K. Richardson, J. Turpin, A. Pogrebnyakov, J. D. Musgraves, J. A. Bossard, H. J. Shin, et al., “Adaptive phase change metamaterials for infrared aperture control,” in Unconventional Imaging, Wavefront Sensing, and Adaptive Coded Aperture Imaging and Non-Imaging Sensor Systems, vol. 8165, p. 81651H, International Society for Optics and Photonics, 2011. [157] J. Pries, Y. Yu, P. Kerres, M. Häser, S. Steinberg, F. Gladisch, S. Wei, P. Lucas, and M. Wuttig, “Approaching the glass transition temperature of gete by crystallizing ge15te85,” physica status solidi (RRL)–Rapid Research Letters, p. 2000478, 2020. [158] J. Doane, N. Vaz, B.-G. Wu, and S. Žumer, “Field controlled light scattering from nematic microdroplets,” Applied Physics Letters, vol. 48, no. 4, pp. 269–271, 1986. [159] H. Guillard and P. Sixou, “Polymer network cholesteric liquid crystal operating in the near infrared,” Molecular Crystals and Liquid Crystals Science and Technology. Section A. Molecular Crystals and Liquid Crystals, vol. 364, no. 1, pp. 647–663, 2001. [160] X. Du, Y. Li, Y. Liu, F. Wang, and D. Luo, “Electrically switchable bistable dual frequency liquid crystal light shutter with hyper-reflection in near infrared,” Liquid Crystals, vol. 46, no. 11, pp. 1727–1733, 2019. [161] M. Soltani, M. Chaker, E. Haddad, R. Kruzelecky, and J. Margot, “Micro-optical switch device based on semiconductor-to-metallic phase transition characteristics of w-doped vo 2 smart coatings,” Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, vol. 25, no. 4, pp. 971–975, 2007. [162] I. Optotherm, “Emissivity in the infrared,” 2018. [163] Y. Cao, D. Torres, T. Wang, X. Tan, and N. Sepúlveda, “Enabling tunable microme- chanical bandpass filters through phase-change materials,” Smart Materials and Struc- tures, vol. 26, no. 8, p. 085032, 2017. 114 [164] R. Cabrera, E. Merced, and N. Sepúlveda, “Performance of electro-thermally driven vo2 -based mems actuators,” Journal of Microelectromechanical Systems, vol. 23, no. 1, pp. 243–251, 2013. [165] L. M. Gisela Hoflinger, “Brief introduction to coating technology for electron mi- croscopy,” 2013. [166] R. Basu, P. Magudapathy, M. Sardar, R. Pandian, and S. Dhara, “Vo2 microcrystals as an advanced smart window material at semiconductor to metal transition,” Journal of Physics D: Applied Physics, vol. 50, no. 46, p. 465602, 2017. [167] M. Currie, M. A. Mastro, and V. D. Wheeler, “Characterizing the tunable refractive index of vanadium dioxide,” Optical Materials Express, vol. 7, no. 5, pp. 1697–1707, 2017. [168] J.-H. Yu, S.-H. Nam, J. Lee, and J.-H. Boo, “Enhanced visible transmittance of ther- mochromic vo2 thin films by sio2 passivation layer and their optical characterization,” Materials, vol. 9, no. 7, p. 556, 2016. [169] S. Wang, G. Liu, P. Hu, Y. Zhou, Y. Ke, C. Li, J. Chen, T. Cao, and Y. Long, “Largely lowered transition temperature of a vo2/carbon hybrid phase change material with high thermal emissivity switching ability and near infrared regulations,” Advanced Materials Interfaces, vol. 5, no. 21, p. 1801063, 2018. [170] I. H. Malitson, “Interspecimen comparison of the refractive index of fused silica∗,†,” J. Opt. Soc. Am., vol. 55, pp. 1205–1209, Oct 1965. [171] L. Gao, F. Lemarchand, and M. Lequime, “Refractive index determination of sio2 layer in the uv/vis/nir range: spectrophotometric reverse engineering on single and bi-layer designs,” Journal of the European Optical Society-Rapid publications, vol. 8, 2013. [172] J. K. Kana, J. Ndjaka, G. Vignaud, A. Gibaud, and M. Maaza, “Thermally tunable op- tical constants of vanadium dioxide thin films measured by spectroscopic ellipsometry,” Optics Communications, vol. 284, no. 3, pp. 807–812, 2011. [173] A. Joushaghani, Micro-and nano-scale optoelectronic devices using vanadium dioxide. PhD thesis, 2014. [174] N. Mlyuka, G. A. Niklasson, and C.-G. Granqvist, “Thermochromic multilayer films of vo2 and tio2 with enhanced transmittance,” Solar Energy Materials and Solar Cells, vol. 93, no. 9, pp. 1685–1687, 2009. 115 [175] H. MacLeod and H. Macleod, Thin-Film Optical Filters, Fourth Edition. Series in Optics and Optoelectronics, CRC Press, 2010. [176] O. Konovalova, A. Sidorov, and I. Shaganov, “Interference systems of controllable mirrors based on vanadium dioxide for the spectral range 0.6-10.6 micrometer,” Journal of Optical Technology, vol. 66, no. 5, p. 391, 1999. [177] S. Lu, L. Hou, and F. Gan, “Preparation and optical properties of phase-change vo2 thin films,” Journal of materials science, vol. 28, no. 8, pp. 2169–2177, 1993. [178] S. Lysenko, F. Fernández, A. Rúa, J. Aparicio, N. Sepúlveda, J. Figueroa, K. Var- gas, and J. Cordero, “Light scattering by epitaxial vo2 films near the metal-insulator transition point,” Journal of Applied Physics, vol. 117, no. 18, p. 184304, 2015. [179] A. Rúa, F. E. Fernández, R. Cabrera, and N. Sepúlveda, “Young’s modulus of pulsed- laser deposited v 6 o 13 thin films,” Journal of Applied Physics, vol. 105, no. 11, p. 113504, 2009. [180] A. Materials, “Silica- silicon dioxide (sio2 ),” 2019. [181] S. Timoshenko, “Analysis of bi-metal thermostats,” Josa, vol. 11, no. 3, pp. 233–255, 1925. [182] D. Kucharczyk and T. Niklewski, “Accurate x-ray determination of the lattice param- eters and the thermal expansion coefficients of vo2 near the transition temperature,” Journal of Applied Crystallography, vol. 12, no. 4, pp. 370–373, 1979. [183] A. Baghban, J. Sasanipour, F. Pourfayaz, M. H. Ahmadi, A. Kasaeian, A. J. Chamkha, H. F. Oztop, and K.-w. Chau, “Towards experimental and modeling study of heat transfer performance of water-sio2 nanofluid in quadrangular cross-section channels,” Engineering Applications of Computational Fluid Mechanics, vol. 13, no. 1, pp. 453– 469, 2019. [184] W. Fan, S. Huang, J. Cao, E. Ertekin, C. Barrett, D. Khanal, J. Grossman, and J. Wu, “Superelastic metal-insulator phase transition in single-crystal vo 2 nanobeams,” Phys- ical Review B, vol. 80, no. 24, p. 241105, 2009. [185] E. Merced, X. Tan, and N. Sepúlveda, “Closed-loop tracking of large displacements in electro-thermally actuated vo2 -based mems,” Journal of Microelectromechanical Sys- tems, vol. 23, no. 5, pp. 1073–1083, 2014. 116 [186] L. N. Binh, “Optical modulation: Advanced techniques and applications in transmis- sion systems and networks,” 2017. [187] R. Paschotta, “Field guide to lasers (vol. 12),” Bellingham, WA: SPIE, pp. 33–44, 2008. [188] M. Bao, “Chapter 2 - mechanics of beam and diaphragm structures,” in Analysis and Design Principles of MEMS Devices (M. Bao, ed.), pp. 33–114, Amsterdam: Elsevier Science, 2005. [189] P. Almeida, S. Kundu, J. Borges, M. Godinho, and J. Figueirinhas, “Electro-optical light scattering shutter using electrospun cellulose-based nano-and microfibers,” Ap- plied Physics Letters, vol. 95, no. 4, p. 043501, 2009. [190] R. Kumar and K. Raina, “Electrically modulated fluorescence in optically active polymer stabilised cholesteric liquid crystal shutter,” Liquid Crystals, vol. 41, no. 2, pp. 228–233, 2014. 117