University of Southern California Dissertations and Theses

Development of fabrication technologies for robust Parylene medical implants

(USC Thesis Other)

Development of fabrication technologies for robust Parylene medical implants

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i Development of Fabrication Technologies for Robust Parylene Medical Implants Dissertation by Jessica Ortigoza-Diaz In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (BIOMEDICAL ENGINEERING) University of Southern California Los Angeles, California 2019 PRESENTED TO THE FACULTY OF THE USC GRADUATE SCHOOL ON MAY 2019 ii © 2019 Jessica Ortigoza-Diaz All Rights Reserved iii To my parents iv ACKNOWLEDGMENTS It has been a long way since I started this journey where I was truly lucky of meeting such wonderful people that help me in one way or another during this unique experience. I would like to express my gratitude to those people here. First and foremost, I thank my parents, my inspiration, for their support, advice, and unconditional love. Without them I would not be what I am now. Thank you for always believing in me during my worst times. They have sacrificed so much so that I would follow my dreams. They taught me that everything is possible if you work hard and never give up. I am grateful with live for having allowed me to be their daughter, I could not have better parents and friends. Also, I would like to thank my brother for being an important part of my life and for encouraging me to be an example of life for him. Lastly, I thank Fausto Salazar for his support, for being so understanding, and for his love. I am very grateful to Dr. Ellis Meng for her guidance and support during my Ph.D. as my graduate advisor and mentor. I thank her for the opportunity to prove myself in this experience of a lifetime. I admire her dedication, work ethic, knowledge, and creativity. I will always be grateful to her for all her advice and mentorship. I owe a special thank you to Dr. Francisco Valero-Cuevas who changed my life by providing me an opportunity to compete for a summer research internship at USC. I thank him for all his advice and guidance. The Biomedical Microsystems Laboratory, my second family, thank you for all their support and friendship over these years. I could not have done it without them. Dr. Roya Sheybani, thank you for introducing me to the world of research and guiding me during my first project in lab. Dr. Brian Kim, thank you always being willing to help and for your enthusiasm, which was always good to have whenever I was having a bad day in the clean room. Dr. Seth Hara, thank you for all the conversations full of guidance and advice at all times. Dr. Curtis Lee, thank you for your technical support. Dr. Lawrence Yu, thank you for sharing your DRIE knowledge and for your helpful conversations. Dr. Angelica Cobo, thank you for always pushing me to be out my v comfort zone and for always organizing our lab events. Dr. Alex Baldwin, thank you for your patience and your expertise in electrical engineering. Dr. Ahuva Weltman, thank you for always sharing our fabrication problems and for always looking for ways to fix them. Dr. Kee Scholten, thank you for all your invaluable guidance, advice, and knowledge whenever I needed. To Christopher Larson, James Yoo, Trevor Hudson, Eugene Yoon, and Xuechun Wang thank you for your friendship and support, I know you are going to be great Doctors in the near future. Thank you mengsters, for all the experiences we have lived together and from which I have learned a lot. No matter what path life takes us to, I will never forget you and I will always remember you with great affection. Finally, I would like to thank the USC Viterbi School of Engineering and the National Council for Science and Technology (CONACyT) for the fellowship that supported me through my graduate career. My research was also made possible through grant from the National Science Foundation (NSF). vi TABLE OF CONTENTS Acknowledgments ................................................................................................................................... iv Table of Contents ..................................................................................................................................... vi List of Tables...........................................................................................................................................viii List of Figures ........................................................................................................................................... ix Introduction to Thin-Film Parylene C Medical Implants ................................... 1 Implantable Medical Devices ........................................................................................ 2 History and Types of Parylene ..................................................................................... 4 Thin-film Parylene Device Microfabrication ............................................................... 5 State-of-the-art of Parylene-based Devices .................................................................. 7 Objectives ....................................................................................................................... 8 References ..................................................................................................................... 10 Considerations in the Fabrication of Parylene Implantable Devices ............. 14 Challenges .................................................................................................................... 15 Micromachining of Parylene Films............................................................................. 22 Discussion..................................................................................................................... 28 Conclusions .................................................................................................................. 30 References ..................................................................................................................... 30 Optimized DRIE Etching of Parylene C ............................................................... 33 Introduction.................................................................................................................. 34 Approach and goals ..................................................................................................... 38 Materials and Methods ................................................................................................ 39 Results ........................................................................................................................... 47 Discussion..................................................................................................................... 63 Conclusions .................................................................................................................. 64 References ..................................................................................................................... 65 Adhesion of Parylene to Parylene and Platinum ................................................ 67 Background .................................................................................................................. 68 Approach and goals ..................................................................................................... 71 Sample preparation and experimental methods ....................................................... 73 Results ........................................................................................................................... 78 Discussion..................................................................................................................... 86 Conclusions .................................................................................................................. 89 References ..................................................................................................................... 90 Evaluation of Parylene Insulation ......................................................................... 94 Background .................................................................................................................. 95 Approach and Goals .................................................................................................... 95 Materials and experimental methods ......................................................................... 96 Results ......................................................................................................................... 104 vii Discussion................................................................................................................... 115 Conclusions ................................................................................................................ 117 References ................................................................................................................... 118 APPENDIX A Ion Transport through Parylene Films ........................................................... 119 APPENDIX B A-174 Silane Treatment...................................................................................... 121 APPENDIX C DRIE Coupons Process Flow ............................................................................ 122 APPENDIX D DRIE Profiles ....................................................................................................... 123 APPENDIX E Dry Adhesion Forces for Parylene Interfaces ................................................ 127 APPENDIX F Parylene T-Peel Test Structures Process Flow............................................... 131 APPENDIX G Metal T-Peel Test Structures Process Flow .................................................... 133 APPENDIX H IDEs Process Flow ............................................................................................... 135 viii LIST OF TABLES Table 2-1. WVTR measurements for different un- and annealed Parylene film thickness. .......... 18 Table 2-2. Ion transport measurements for different un- and annealed Parylene film thickness. ........................................................................................................................................................ 19 Table 3-1. Passivation step of current DRIE recipe ............................................................................ 43 Table 3-2. Etch step of current DRIE recipe ......................................................................................... 43 Table 3-3. Values of process parameters tested. ................................................................................. 44 Table 3-4. DRIE recipe with ICP power optimized. ........................................................................... 51 Table 3-5. DRIE recipe with RF bias power optimized. ..................................................................... 54 Table 3-6. DRIE recipes with different Ar and SF6 flow rates ........................................................... 57 Table 3-7. DRIE recipe with O2 flow rate optimized. ......................................................................... 59 Table 3-8. Optimized DRIE recipe. ........................................................................................................ 62 Table 3-9. DRIE Parylene etch rates reported in literature................................................................ 64 Table 4-1. Comparison of wet adhesion forces reported for Parylene interfaces. .......................... 70 Table 4-2. Material combinations. .......................................................................................................... 73 Table 4-3. Properties of Parylene (n=1) at different anneal times. .................................................... 79 Table 4-4. Time to interface failure: with and without adhesion promoter. ................................... 85 Table 4-5. Atomic composition of the metal combinations. ............................................................... 86 Table 5-1. IDE experimental groups and electrode parameters. ....................................................... 96 ix LIST OF FIGURES Figure 1-1. (a) Stents [2]; and (b) pacemaker [3]. .................................................................................................... 2 Figure 1-2. (a) The CardioMEMS TM HF System [12]; and (b) the Utah array [13]. ............................................ 3 Figure 1-3. (a) Electrochemical PDMS sensor (Reprinted from Reference [14] with permission from Elsevier); (b) polyimide-silicone cuff electrode (©2001 IEEE. Reprinted, with permission, from Reference [15]); and (c) neural probe arrays. ............................................................................................ 4 Figure 1-4. Chemical structure of types Parylene. ................................................................................................. 5 Figure 1-5. Cross section of a typical device showing insulated and exposed metal features, such as, traces and electrodes, respectively. ....................................................................................................................... 6 Figure 1-6. (a) Wireless coils (Reprinted from Reference [27] with permission from Elsevier); (b) flow, pressure and patency sensors to monitor hydrocephalus treatment (©2016 IEEE. Reprinted, with permission, from Reference [28]); (c) hippocampal neural probe array (©2017 IEEE. Reprinted, with permission, from Reference [29]); and (d) retinal prosthesis that matches the curvature of the eye (Reprinted from Reference [30] with permission from Elsevier). Parylene is transparent and the opaque features are metal. .................................................................................................................... 7 Figure 2-1. Images of Parylene damage due to high heat or radiation exposure. (a) Cracked Parylene after development and soft baking at 115 °C for 3 min; and (b) air bubbles appeared under the film, during hard bake at 90 °C for 15 min, in regions previously exposed to UV light during lithography. ................................................................................................................................................. 15 Figure 2-2. Comparison of thermal annealed devices in the presence of oxygen (left and center) and under intact vacuum (right). Devices were heated at 200 °C for 48 hours. ................................................... 16 Figure 2-3. Borosilicate electrochemical cell.......................................................................................................... 19 Figure 2-4. Parylene devices where delamination was noticeable on the macro scale. (a) Parylene layers visibly split from each other; (b) detachment of the top Parylene layer from the metal and bottom Parylene layers; and (c) electrodes and metal traces began to move or slide around between the Parylene layers. ........................................................................................................................................... 20 Figure 2-5. (a) Deposited Parylene layer incorporating spherules which results in a cloudy in appearance when observed by eye; and (b) magnified photograph revealing the presence of small clusters of spherules. ..................................................................................................................................................... 23 Figure 2-6. (a) Stress-induced cracking of deposited platinum likely due to excess heat generated during the process; and (b) cracked metal traces after lift-off. ......................................................................... 25 Figure 2-7. (a) Severe wrinkling/rippling seen in sputtered deposited platinum film results from compressive stresses of a higher magnitude; and (b) wrinkled areas around the device metal features. ........................................................................................................................................................ 25 x Figure 2-8. Out-gassing of photoresist solvent results in bubbling observed in sacrificial photoresist structures sandwiched in Parylene following O2 DRIE. ....................................................................... 27 Figure 3-1. Main processes of dry etching. ............................................................................................................ 34 Figure 3-2. Directionality of etching process. ....................................................................................................... 35 Figure 3-3. Bosch process. ........................................................................................................................................ 36 Figure 3-4. Schematic of an ICP system. ................................................................................................................ 37 Figure 3-5. Schematics of (a) a patterned test coupon (2 cm × 2 cm); and (b) mask patterns and the corresponding dimensions. ....................................................................................................................... 39 Figure 3-6. Overview of fabrication process for the DRIE coupon; (a) Silicon wafer treated with A-174; (b) Parylene deposition; (c) Photoresist was spun on and patterned; (d) DRIE plasma etching; (e) Etched patterns after DRIE process; (f) Parylene patterns after removal of the masking material; and (g) Schematic of a patterned test coupon......................................................................................... 40 Figure 3-7. (a) KMPR 1025; and (b) AZ P4620 mask patterns after lithography. ............................................ 40 Figure 3-8. Overview of the replica mold process (a) etched Parylene trenches on silicon; (b) pouring and curing of silicone rubber; (c) peeling off the silicone rubber mold; (d) replica mold; and (e) SEM images of a replica mold, the circle indicates the replicated trench. ................................................... 41 Figure 3-9. Schematic of the Oxford Plasmalab 100 ICP tool. ............................................................................ 42 Figure 3-10. Cross-section view of a Parylene etched profile by DRIE. ............................................................ 44 Figure 3-11. Cross-section view of a Parylene etched profile before DRIE, after DRIE and without masking material (from left to right). ...................................................................................................... 45 Figure 3-12. Diagram of the PDMS replica mold of an etched Parylene trench showing the sidewall angle measurement. .............................................................................................................................................. 46 Figure 3-13. (a) KMPR 1025 residues left on Parylene coupon; and (b) Parylene bubbles after soaking coupons in hot Remover PG...................................................................................................................... 47 Figure 3-14. (a) Pattern definition before mounting coupon to the carrier wafer; and (b) pattern reflow after mounting coupon using AZ 4400 resist. ........................................................................................ 48 Figure 3-15. Optical micrograph of effect on photolithographically produced pattern in AZ 4620 after mounting using (a) AZ 4400 resist; and (b) double sided tape. ........................................................... 48 Figure 3-16. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate; and (b) selectivity. ICP power process parameter was varied (500, 600, 700, 800, 900, and 1000 W). ............................. 49 xi Figure 3-17. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. ICP power process parameter was varied (500, 600, 700, 800, 900, and 1000 W).................................................................................... 50 Figure 3-18. Vertical etch rate of Parylene and selectivity of the masking material, KMPR 1025, when the RF bias power was varied.......................................................................................................................... 52 Figure 3-19. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Process RF bias power (10, 20, 40, 60, 80 and 100 W) was varied. ............................................................................................................. 53 Figure 3-20. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Ar and SF6 flow rates (0, 10 and 40 sccm) were varied for the etch step. ........................................................................ 55 Figure 3-21. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Ar, for passivation and etch steps, and SF6, for etch step, flow rates (0, 10 and 40 sccm) were varied. .......................... 56 Figure 3-22. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. O2 flow rate (0, 20, 35, 40, 60, 80 and 100 sccm) was varied for the etch step. ..................................................................... 58 Figure 3-23. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. C4F8 deposition time (0, 1, 2, 3, 4, 5, 7 and 10 sec) was varied for the passivation step. ............................................................................. 60 Figure 3-24. Etch data for Parylene sample processed by DRIE: (a) vertical etch rate and selectivity, (b) lateral etch rate, (c) anisotropy; and (d) sidewall angle. C4F8 flow rate (0, 20, 40, 60, 80 and 100 sccm) was varied for the dep step. ........................................................................................................... 61 Figure 3-25. SEM images of the (a) surface electrode, (b) contact pads, (c) locking mechanism, and (d) zoom in on the surface electrode. Images without masking material, only Parylene. ..................... 63 Figure 4-1. (a) Sample schematic of T-peel test structure; (b) test structure in T-peel pulling apparatus; and (c) top and (d) lateral views of a sample under testing. ................................................................ 73 Figure 4-2. Cross-sectional view of the fabrication process for PP (right) and PMP (left) T-peel tests samples. (a) Deposition of substrate layer; (b) sacrificial layer; (c) deposition of insulation layer; and (d) etching. ........................................................................................................................................... 74 Figure 4-3. Representative raw T-peel test data. .................................................................................................. 77 Figure 4-4. Immersed samples in 1× PBS. .............................................................................................................. 78 Figure 4-5. XRD 2θ scans of un-annealed and annealed Parylene films. ......................................................... 79 xii Figure 4-6. Failure modes presented at day 0 during testing for (a) typical PMAdP peeled apart sample with no tearing; (b) PMAdP sample torn at the interface; and (c) 48h-annealed PAdMAdP sample torn at the clamping hole. .......................................................................................................................... 80 Figure 4-7. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene layers (mean ± SE, n=4): (a) peeling strength as a function of thermal annealing time for dry samples; and (b) peeling strength as a function of time soaked in warm saline. ....... 81 Figure 4-8. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene (PP) and Parylene-platinum-Parylene (PMP) layers (mean ± SE, n=4): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. ............................................................................................................................................................ 82 Figure 4-9. Delaminated (a) 48h-annealed; and (b) un-annealed PMP samples after a 3-week and 4-day soak in PBS, respectively. .......................................................................................................................... 82 Figure 4-10. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene (PP), Parylene-Ethylene glycol diacrylate-Parylene (PEGDAP), and Parylene- Diamond-like carbon-Parylene (PDLCP) layers (mean ± SE, n=4): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. .............................. 83 Figure 4-11. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-platinum-Parylene layers with and without AdPro Plus® adhesion promoter (mean ± SE, n=4). Samples featured AdPro Plus® between the base Parylene and metal layer (PAdMP), between the metal and top Parylene layer (PMAdP) and on both sides of the metal layer (PAdMAdP): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. ............................................................................................................................... 84 Figure 5-1. (a) Sample schematic of IDE structures; and (b) zoom in to observe the different finger width and pitch considered for electrode design. ............................................................................................. 96 Figure 5-2. Fabricated IDEs. .................................................................................................................................... 97 Figure 5-3. Cross-sectional view of the fabrication process for IDE. (a) Bare Si carrier wafer; (b) deposition of substrate layer; (c) platinum e-beam evaporation deposition; (d) lift-off of IDEs and contact pads; (e) deposition of adhesion layer if needed; (f) deposition of insulation layer; (g) DRIE etching; and (h) released devices. ............................................................................................................. 98 Figure 5-4. Electrical packaging of an IDE, which is attached to a ZIF connector. ......................................... 99 Figure 5-5. IDE device secured in vial caps with epoxy. ..................................................................................... 99 Figure 5-6. Immersed samples in 1× PBS. ............................................................................................................ 100 Figure 5-7. IDE devices were epoxied into vial caps for soaking study. ........................................................ 101 xiii Figure 5-8. (a) Circuit model of a Parylene-platinum-Parylene device with encapsulated electrodes under chronic soaking conditions; and (b) circuit model of the same device when water reaches the platinum electrode, creating an electrode-electrolyte interface. ........................................................ 102 Figure 5-9. Representative EIS raw data of a perfectly insulated IDE. ........................................................... 103 Figure 5-10. (a) Impedance; and (b) phase as a function of frequency. Impedance decreased and phase increased as a function of soaking time, indicating of water absorption. ........................................ 103 Figure 5-11. Optical micrographs of control IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. ..................................................................................................................................................................... 104 Figure 5-12. Optical micrographs of control IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. ........................................................... 105 Figure 5-13. Optical micrographs of annealed IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. ........................................................... 106 Figure 5-14. Optical micrographs of annealed IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. ........................................................... 107 Figure 5-15. Optical micrographs of AdPro Plus® IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. .......................................................................................................................................... 108 Figure 5-16. Optical micrographs of AdPro Plus® IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. ........................................... 109 Figure 5-17. Optical micrographs of EGDA IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. ..................................................................................................................................................................... 110 Figure 5-18. Optical micrographs of EGDA IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. ........................................................... 111 Figure 5-19. Measured impedance and phase angle at 1 Hz as function of soaking time for control devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. .................................................................................................................................. 112 xiv Figure 5-20. Measured impedance and phase angle at 1 Hz as function of soaking time for annealed devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. .................................................................................................................................. 113 Figure 5-21. Measured impedance and phase angle at 1 Hz as function of soaking time for AdPro Plus® devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. .................................................................................................................................. 114 Figure 5-22. Measured impedance and phase angle at 1 Hz as function of soaking time for EGDA devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. ................................................................................................................................................ 115 1 he development of poly-(chloro-p-xylylene) (Parylene C) as a material for microelectromechanical systems (MEMS) has come a long way since its discovery. For years, Parylene C was employed as an encapsulation material for medical implants, such as stents and pacemakers, due to its strong barrier properties and biocompatibility. However, these devices have faced challenges in materials, battery power, and functionality. Advances in MEMS technology have allowed the development and improvement of many medical implants, and the use of Parylene C as a structural and substrate material for thin-film medical implants. The choice of material for creating a medical implant is a key factor that will dictate its reliability depending on the reaction of the body to its presence. This chapter introduces implantable medical devices, the application of MEMS technology in the development of medical implants and Parylene C as material choice to fabricate complex medical implants. INTRODUCTION TO THIN-FILM PARYLENE C MEDICAL IMPLANTS T 2 Implantable Medical Devices Medical implants are devices that are place inside the body attempting to restore or enhance functionality of the body [1]. Implantable medical devices have come a long way from typical passive implants, such as stents (Figure 1-1. (a) Stents [2]; and (b) pacemaker [3].Figure 1-1a), to more active devices such as pacemakers (Figure 1-1b) due to developments in engineering, microelectronics, and materials. But they still face familiar challenges of size, power consumption, biocompatibility and efficacy. Figure 1-1. (a) Stents [2]; and (b) pacemaker [3]. Microelectromechanical systems (MEMS) technologies can provide miniaturization, wireless power operation, and integration of complex functions to implantable devices providing new opportunities to create novel implants or improve existing ones. MEMS are tiny devices produced with microfabrication techniques used to create integrated circuits [4]. The rapid growth of MEMS technology is due to the adaptation of silicon-based microlithography techniques to the topic of electro-mechanical systems, which can improve quality of life by revolutionizing medical diagnostics and treatment modalities [5]. MEMS have become popular in recent years due to their ability to sense pressure, detect motion, measure forces, pump and control fluids, and perform other important actions for medical and biological fields [6]. A different number of implantable MEMS have been developed due to the high demand for medical devices, such as, neural prostheses, cochlear prostheses, drug delivery devices, retinal prostheses, and biosensors. The large majority of these devices are machined from rigid materials, such as glass, ceramics and metals, but the overwhelming 3 majority are still rendered from silicon [7,8]. However, as a material silicon is poorly suited for medical implants due to its opacity, brittleness, and poor biocompatibility. Early examples include the CardioMEMS TM HF System and the Utah array (Figure 1-2). The CardioMEMS TM HF System (Figure 1-2a) is the first Food and Drug administration (FDA) approved MEMS-based implantable device. It uses a glass transduction mechanism to measure pulmonary artery pressure and heart rate, and wirelessly communicates data to prevent heart failure. The Utah array (Figure 1-2b) is the first MEMS based, penetrating neural prostheses, fabricated from chemically etched silicon shanks [9]. The choice of materials is driven by their amenability to existing microfabrication technologies, and in particular, the widespread familiarity with silicon’s mechanical, thermal and chemical properties. However, rigid devices have frequently demonstrated poor suitability to biological implantation. As an example, rigid neural probes are mechanically hard compared to soft brain tissue, and can induce tissue damage and subsequent chronic immune inflammatory response [10]. Silicon in particular is susceptible to corrosion, cracking, and electrical shorting, due to the electrically conductive physiological fluid encountered in most tissue. These and other material shortcomings greatly reducing implantable MEMS longevity in vivo. Biocompatibility remains perhaps the most daunting obstacle limiting device longevity and functionality [11]. Figure 1-2. (a) The CardioMEMS TM HF System [12]; and (b) the Utah array [13]. Fortunately, these limitations can be addressed using soft materials such as Parylene C, polydimethylsiloxane (PDMS) and polyimide that have been used as insulating coatings and then became a choice of substrate material to fabricate thin-film implantable devices. Free-film devices offer a combination of high biocompatibility, high flexibility, optical transparency, electrical insulation, and amenability to standard MEMS processing methods that motivated the 4 development of numerous soft thin-film devices, which to date include flexible neural implants and sensors (Figure 1-3). Figure 1-3. (a) Electrochemical PDMS sensor (Reprinted from Reference [14] with permission from Elsevier); (b) polyimide-silicone cuff electrode (©2001 IEEE. Reprinted, with permission, from Reference [15]); and (c) neural probe arrays. History and Types of Parylene Parylene is the trade name for poly-(para-xylylene), a class of semicrystalline, hydrophobic polymers which can be deposited as thin, conformal, pinhole-free films using chemical vapor deposition (CVD). Parylene was discovered by Michael Mojzesz Szwarc in the late 1940s, but was not commercialized until 1965 following the development of the Gorham deposition process at Union Carbide [16,17]. Types of Parylene that are commonly available include Parylene N, Parylene C, Parylene D, and Parylene HT (Figure 1-4). A comparison of the properties of these films can be found in [18]. Parylene N is composed of an aromatic ring with attached methylene groups [19]; it has been used as a dielectric and has the slowest deposition rate of the Parylenes [20]. Parylene C features a single chlorine atom on its benzene ring and is recognized for its chemical inertness, electrical resistivity, low moisture permeability, and proven biocompatibility. Parylene D has two chlorine atoms, resulting in similar properties to Parylene C with slightly higher temperature resistance. Parylene HT is a fluorinated variant of the Parylene N polymer and is marked by high temperature stability [21]. The two most common Parylene types in commercial applications are Parylene N and C, whereas Parylene C is the most common material encountered in MEMS devices. 5 Figure 1-4. Chemical structure of types Parylene. Parylene C, referred from here on as Parylene, is a US Pharmacopeia class VI material [22], having the highest biocompatibility certification for a plastic material. For several decades, thin Parylene coatings were used to create waterproof insulation for electronics intended for use in harsh environments, a category that increasingly includes biomedical implants. Parylene exhibits low intrinsic stress in addition to optical transparency, mechanical flexibility, and compatibility with several standard micromachining processes [23-26]. While Parylene can be used in combination with rigid substrates such as silicon and glass, it has been increasingly utilized as a flexible structural material in the growing field of thin-film Parylene medical implants. As a dynamic polymer material, Parylene presents unique challenges during microfabrication and use which are further explored in this dissertation. Thin-film Parylene Device Microfabrication Parylene-based devices are prepared through a combination of microfabrication processes and commonly consist of a simple sandwich design, with a base layer of Parylene, a thin layer of patterned metal defining traces and other components, and a top layer of Parylene. Figure 1-5 shows a flexible device in which Parylene serves as both the structural material and insulator. However, devices may also be supported on standard rigid substrates such as glass and silicon when flexibility is not required. Regardless of the final format, microfabrication requires that Parylene be supported by rigid substrate during fabrication, with optional release of a free film device towards the end of the process. Devices are realized using a combination of three process categories: additive processes (deposition), subtractive processes (etching), and patterning (e.g. 6 photolithography). Description of the processes used to fabricate the samples in this work will be presented in the following paragraphs. Figure 1-5. Cross section of a typical device showing insulated and exposed metal features, such as, traces and electrodes, respectively. Additive processes consist of depositing layers of structural and sacrificial materials. Examples of these processes are spin coating, thermal oxidation, physical vapor deposition (PVD) and chemical vapor deposition (CVD). Parylene is deposited exclusively through a highly conformal CVD process at room-temperature, with typical deposition rates on the order of ~2 µm/hour. Other deposition methods are critical for MEMS processing; spin coating is used to deposit photo-patternable polymer layers (photoresist) for use as an etch-mask or shadow-mask, or as a sacrificial layer, while PVD methods including evaporation and sputtering are used for metal deposition. Etching is a subtractive process used to selectively remove material to machine structures or define geometries. Etchants may be ‘wet’, consisting of liquid chemical solvents and corrosives, or ‘dry’, consisting of gas, vapor or plasma. Etching is typically controlled through use of a masking layer, which protects the substrate such that only areas exposed by the mask are removed. Different materials can be used as a mask, including, oxides, resist and metal, depending on the choice of etchant. Etching processes for Parylene are limited to dry methods, due to Parylene’s high chemical inertness. Reactive ion etching and deep reactive ion etching are oxygen plasma-based techniques commonly used to etch Parylene, and are expounded upon in detail in Chapter 2. 7 Photolithography is a patterning processes to transfer relief patterns to the substrate with high resolution and fidelity. Typically a thin, photo-sensitive material, photoresist, is deposited on a substrate by way of spin coating and exposed to UV light through a shadow-mask. The resulting patterned layer can serve as a temporary, sacrificial structure, or as a mask for subsequent etching/deposition processes. State-of-the-art of Parylene-based Devices Parylene is used extensively to coat printed circuit boards, wires, MEMS devices and biomedical implants, and less commonly as a structural or substrate material for electronic and MEMS devices. Still, over the last several years, many devices for a large variety of applications have been described in the literature featuring Parylene as either the structural or substrate material (Figure 1-6). Figure 1-6. (a) Wireless coils (Reprinted from Reference [27] with permission from Elsevier); (b) flow, pressure and patency sensors to monitor hydrocephalus treatment (©2016 IEEE. Reprinted, with permission, from Reference [28]); (c) hippocampal neural probe array (©2017 IEEE. Reprinted, with permission, from Reference [29]); and (d) retinal prosthesis that matches the curvature of the eye (Reprinted from Reference [30] with permission from Elsevier). Parylene is transparent and the opaque features are metal. 8 Parylene has been used as flexible coiled cable to connect medical implants to external circuitry [31-33]. Radio frequency powered coils were fabricated with Parylene for wireless transmission (Figure 1-6a) [27,34], and more recently, a new method to wirelessly transduce electrochemical impedance was developed using Parylene coils [35]. Parylene microsensors (Figure 1-6b) were developed to measure intraocular and intracranial pressures [36-38] and Parylene thermal sensors have been created to detect small flows [28,39-41]. Other devices include neurocages for in vitro neural network study [42-44], bellows for drug delivery [45-47], an electrochemical patency sensor [48], microfluidic devices [22], and electrothermal valves [49]. In the field of neural prostheses, Parylene was used to create both penetrating (Figure 1-6c) [29,50- 54] and non-penetrating microelectrode arrays [55-57]. Microfluidic channels were integrated into Parylene neural probes to inject drugs [58-60]. Parylene-based cuff electrodes, which record neural activity within peripheral nerves, were also developed [61-63]. Parylene spinal cord stimulators have also been realized with high-density electrode sites [30,64]. Parylene is also a common material for sensory neural prosthetics, such as retinal implants. High-density electrode arrays were fabricated on Parylene that was subsequently thermoformed to match the curvature of the eye (Figure 1-6d) [23,30,65]. An origami like device was fabricated to match the curvature of the eye without the use of heat [66]. Cochlear electrode arrays, which must conform to the complex anatomy of the inner ear, have been developed by exploiting Parylene’s properties as a thin and flexible substrate [67]. Objectives Parylene provides many advantages as a material for thin-film implantable devices, however, there is still much work to be done to develop reliable microfabrication protocols for Parylene that result in robust devices. With all of this in mind, it is the aim of the present work to accomplish the following research aim: 9 To advance the development of technologies for Parylene-based devices by providing an optimum process technique to selectively remove Parylene, and the improvement of Parylene- Parylene and Parylene-metal-Parylene adhesions for dry and wet environments. This work focuses on improvement of microfabrication processes to overcome the main drawbacks of Parylene. Two specific challenges are examined in the detail: the need to selectively remove Parylene to create more complex structures with a high-fidelity transferred pattern, and the need to improve dry and wet adhesion of Parylene to itself and metallic layers to improve long-term performance. These problems are significant obstacles for the implementation of Parylene-based medical implants. Chapter 2 describes some of the challenges of Parylene to be implemented as structural material, such as, thermal budget, water diffusion, adhesion, and handling. Also, the considerations in the fabrication of thin-film Parylene implants are discussed in detail as well as the solutions encountered to these problems. Chapter 3 presents an optimized DRIE process to selectively remove Parylene. The effects of several operating parameters including inductively coupled plasma and radio frequency powers, gas pressure, flow rate of reactants and passivation layer on the vertical and lateral etch rates, selectivity and sidewall angles were investigated in order to obtain high aspect ratio structures. Finally, the optimized recipe was adopted for the use in Parylene micromachining of a cuff electrode. Chapter 4 presents an extensive investigation of different methods to improve dry and wet adhesions of Parylene-Parylene and Parylene-metal-Parylene interfaces. Several strategies were used: thermal annealing, plasma treatment, adhesion promoter, and interposer layers. The adhesion strength was quantify using a 180° T-peel test as a function of soaking duration in saline solution to simulate physiological environment. Changes in film morphology and surface composition were analyzed using X-ray diffraction and X-ray photoelectron spectroscopy techniques. Chapter 5 presents a comprehensive study to evaluate and compare various parameters 10 that may improve adhesion of Parylene insulation in chronically implanted devices. Based on results obtained in Chapter 4, thermal annealing, AdPro Plus® and ethylene glycol diacrylate (EGDA) were chosen to improve reliability and hopefully extend lifetime of Parylene-based devices. Simple Parylene-based devices, interdigitated electrodes, were designed to test adhesion reliability. Devices were soaked in saline solution to simulate in vivo environment. The integrity of the insulation was evaluated using electrochemical impedance spectroscopy by measuring the impedance to observe and quantify the signal loss during the soaking test. References 1. Administration, U.S.F.D. Implants and prosthetics. https://www.fda.gov/medicaldevices/productsandmedicalprocedures/implantsandprosthetics/ 2. Boardman, H. Which stent is best for heart attacks? www.evidentlycochrane.net/which-stent-heart- attacks 3. K&J Magnetics, I. Pacemaker safety. https://www.kjmagnetics.com/blog.asp?p=pacemaker-safety 4. Grayson, A.C.R.; Shawgo, R.S.; Johnson, A.M.; Flynn, N.T.; Li, Y.; Cima, M.J.; Langer, R. A biomems review: Mems technology for physiologically integrated devices. Proceedings of the IEEE 2004, 92, 6-21. 5. Kotzar, G.; Freas, M.; Abel, P.; Fleischman, A.; Roy, S.; Zorman, C.; Moran, J.M.; Melzak, J. Evaluation of mems materials of construction for implantable medical devices. Biomaterials 2002, 23, 2737-2750. 6. Gilleo, K.; ET-Trends, L.; Warwick, R. Mems in medicine. Circuits Assembly 2005, 16, 32. 7. Robinson, D.A. The electrical properties of metal microelectrodes. Proceedings of the IEEE 1968, 56, 1065-1071. 8. Scholten, K.; Meng, E. Materials for microfabricated implantable devices: A review. Lab on a Chip 2015, 15, 4256-4272. 9. Campbell, P.K.; Jones, K.E.; Huber, R.J.; Horch, K.W.; Normann, R.A. A silicon-based, three- dimensional neural interface: Manufacturing processes for an intracortical electrode array. IEEE Transactions on Biomedical Engineering 1991, 38, 758-768. 10. Takeuchi, S.; Suzuki, T.; Mabuchi, K.; Fujita, H. 3d flexible multichannel neural probe array. Journal of micromechanics and microengineering 2003, 14, 104. 11. Onuki, Y.; Bhardwaj, U.; Papadimitrakopoulos, F.; Burgess, D.J. A review of the biocompatibility of implantable devices: Current challenges to overcome foreign body response. SAGE Publications: 2008. 12. Cardiology, D.a.I. New cardiomems data shows effectiveness in reducing heart failure readmissions. https://www.dicardiology.com/content/new-cardiomems-data-shows-effectiveness-reducing-heart- failure-readmissions 13. Jones, B. Bionic implants. https://prometheus.med.utah.edu/~bwjones/2009/08/bionic-implants/ 14. Wu, W.-Y.; Zhong, X.; Wang, W.; Miao, Q.; Zhu, J.-J. Flexible pdms-based three-electrode sensor. Electrochemistry Communications 2010, 12, 1600-1604. 15. Schuettler, M.; Stieglitz, T.; Gross, M.; Altpeter, D.; Staiger, A.; Doerge, T.; Katzenberg, F. In Reducing stiffness and electrical losses of high channel hybrid nerve cuff electrodes, 2001 Conference Proceedings of the 23rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2001; IEEE: pp 769-772. 11 16. Gorham, W.F. A new, general synthetic method for the preparation of linear poly‐p‐xylylenes. Journal of Polymer Science Part A: Polymer Chemistry 1966, 4, 3027-3039. 17. Szwarc, M. Some remarks on the ch 2 [graphic omitted] ch 2 molecule. Discussions of the Faraday Society 1947, 2, 46-49. 18. Kim, B.J.; Meng, E. Micromachining of parylene c for biomems. Polymers for Advanced Technologies 2016, 27, 564-576. 19. Fortin, J.B.; Lu, T.-M. Deposition kinetics for polymerization via the gorham route. In Chemical vapor deposition polymerization, Springer: 2004; pp 41-55. 20. Yang, G.-R.; Ganguli, S.; Karcz, J.; Gill, W.; Lu, T.-M. High deposition rate parylene films. Journal of Crystal growth 1998, 183, 385-390. 21. Fortin, J.B.; Lu, T.-M. Chemical vapor deposition polymerization: The growth and properties of parylene thin films. Springer Science & Business Media: 2003. 22. Tai, Y.C. In Parylene biomems, Robotics and Biomimetics (ROBIO). 2005 IEEE International Conference on, 2005; IEEE: pp xxxvii-xxxvii. 23. Rodger, D.C.; Li, W.; Ameri, H.; Ray, A.; Weiland, J.D.; Humayun, M.S.; Tai, Y.-C. In Flexible parylene- based microelectrode technology for intraocular retinal prostheses, Nano/Micro Engineered and Molecular Systems, 2006. NEMS'06. 1st IEEE International Conference on, 2006; IEEE: pp 743-746. 24. Senkevich, J.J.; Desu, S.B. Morphology of poly (chloro-p-xylylene) cvd thin films. Polymer 1999, 40, 5751-5759. 25. Noh, H.-S.; Huang, Y.; Hesketh, P.J. Parylene micromolding, a rapid and low-cost fabrication method for parylene microchannel. Sensors and Actuators B: Chemical 2004, 102, 78-85. 26. Kim, B.J.; Meng, E. Review of polymer mems micromachining. Journal of Micromechanics and Microengineering 2015, 26, 013001. 27. Li, W.; Rodger, D.C.; Pinto, A.; Meng, E.; Weiland, J.D.; Humayun, M.S.; Tai, Y.-C. Parylene-based integrated wireless single-channel neurostimulator. Sensors and Actuators A: Physical 2011, 166, 193-200. 28. Baldwin, A.; Yu, L.; Meng, E. An electrochemical impedance-based thermal flow sensor for physiological fluids. Journal of Microelectromechanical Systems 2016, 25, 1015-1024. 29. Hirschberg, A.W.; Xu, H.; Scholten, K.; Berger, T.W.; Song, D.; Meng, E. In Development of an anatomically conformal parylene neural probe array for multi-region hippocampal recordings, Micro Electro Mechanical Systems (MEMS), 2017 IEEE 30th International Conference on, 2017; IEEE: pp 129-132. 30. Rodger, D.C.; Fong, A.J.; Li, W.; Ameri, H.; Ahuja, A.K.; Gutierrez, C.; Lavrov, I.; Zhong, H.; Menon, P.R.; Meng, E. Flexible parylene-based multielectrode array technology for high-density neural stimulation and recording. Sensors and Actuators B: chemical 2008, 132, 449-460. 31. Huang, R.; Tai, Y.-C. In Flexible parylene-based 3-d coiled cable, Nano/Micro Engineered and Molecular Systems (NEMS), 2010 5th IEEE International Conference on, 2010; IEEE: pp 317-320. 32. Yu, H.; Wang, S.; Wang, W.; Li, Z. In Fabrication of electroplated nickel multielectrode microprobes with flexible parylene cable, Micro Electro Mechanical Systems (MEMS), 2012 IEEE 25th International Conference on, 2012; IEEE: pp 239-242. 33. Bhandari, R.; Negi, S.; Van Wagenen, R.; Solzbacher, F. In Parylene c-based flexible interface for neuroprosthetic applications, Electronic Components and Technology Conference (ECTC), 2010 Proceedings 60th, 2010; IEEE: pp 1010-1014. 34. Sun, X.; Zheng, Y.; Li, Z.; Yu, M.; Yuan, Q.; Li, X.; Zhang, H. In Design and fabrication of flexible parylene- based inductors with electroplated nife magnetic core for wireless power transmission system, Nano/Micro Engineered and Molecular Systems (NEMS), 2012 7th IEEE International Conference on, 2012; IEEE: pp 238-242. 35. Baldwin, A.; Yu, L.; Pratt, M.; Scholten, K.; Meng, E. Passive, wireless transduction of electrochemical impedance across thin-film microfabricated coils using reflected impedance. Biomedical microdevices 2017, 19, 87. 12 36. Crum, B.; Li, W. In Parylene-based fold-and-bond wireless pressure sensor, Nano/Micro Engineered and Molecular Systems (NEMS), 2013 8th IEEE International Conference on, 2013; IEEE: pp 1155-1158. 37. Yu, L.; Gutierrez, C.A.; Meng, E. An electrochemical microbubble-based mems pressure sensor. Journal of Microelectromechanical Systems 2016, 25, 144-152. 38. Meng, E.; Chen, P.-J.; Rodger, D.; Tai, Y.-C.; Humayun, M. In Implantable parylene mems for glaucoma therapy, Microtechnology in Medicine and Biology, 2005. 3rd IEEE/EMBS Special Topic Conference on, 2005; IEEE: pp 116-119. 39. Meng, E.; Tai, Y.-C. In A parylene mems flow sensing array, TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th International Conference on, 2003, 2003; IEEE: pp 686-689. 40. Li, C.; Wu, P.-M.; Hartings, J.A.; Wu, Z.; Ahn, C.H.; Narayan, R.K. In Cerebral blood flow sensor with in situ temperature and thermal conductivity compensation, Micro Electro Mechanical Systems (MEMS), 2012 IEEE 25th International Conference on, 2012; IEEE: pp 1021-1024. 41. Yu, H.; Ai, L.; Rouhanizadeh, M.; Patel, D.; Kim, E.S.; Hsiai, T.K. Flexible polymer sensors for in vivo intravascular shear stress analysis. Journal of microelectromechanical systems 2008, 17, 1178-1186. 42. Park, J.; Pine, J.; Tai, Y.-C. In Flexible neurocage array for live neural network study, Micro Electro Mechanical Systems (MEMS), 2013 IEEE 26th International Conference on, 2013; IEEE: pp 295-298. 43. He, Q.; Meng, E.; Tai, Y.-C.; Rutherglen, C.M.; Erickson, J.; Pine, J. In Parylene neuro-cages for live neural networks study, TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th International Conference on, 2003, 2003; IEEE: pp 995-998. 44. Tooker, A.; Meng, E.; Erickson, J.; Tai, Y.-C.; Pine, J. In Development of biocompatible parylene neurocages, Engineering in Medicine and Biology Society, 2004. IEMBS'04. 26th Annual International Conference of the IEEE, 2004; IEEE: pp 2542-2545. 45. Sheybani, R.; Gensler, H.; Meng, E. A mems electrochemical bellows actuator for fluid metering applications. Biomedical microdevices 2013, 15, 37-48. 46. Li, P.-Y.; Sheybani, R.; Gutierrez, C.A.; Kuo, J.T.; Meng, E. A parylene bellows electrochemical actuator. Journal of Microelectromechanical Systems 2010, 19, 215-228. 47. Gensler, H.; Sheybani, R.; Meng, E. In Rapid non-lithography based fabrication process and characterization of parylene c bellows for applications in mems electrochemical actuators, Solid-State Sensors, Actuators and Microsystems Conference (TRANSDUCERS), 2011 16th International, 2011; IEEE: pp 2347-2350. 48. Kim, B.J.; Jin, W.; Baldwin, A.; Yu, L.; Christian, E.; Krieger, M.D.; McComb, J.G.; Meng, E. Parylene mems patency sensor for assessment of hydrocephalus shunt obstruction. Biomedical microdevices 2016, 18, 87. 49. Li, P.-Y.; Givrad, T.K.; Holschneider, D.P.; Maarek, J.-M.I.; Meng, E. A parylene mems electrothermal valve. Journal of Microelectromechanical Systems 2009, 18, 1184-1197. 50. Kuo, J.T.; Kim, B.J.; Hara, S.A.; Lee, C.D.; Gutierrez, C.A.; Hoang, T.Q.; Meng, E. Novel flexible parylene neural probe with 3d sheath structure for enhancing tissue integration. Lab on a Chip 2013, 13, 554-561. 51. Kim, B.J.; Kuo, J.T.; Hara, S.A.; Lee, C.D.; Yu, L.; Gutierrez, C.; Hoang, T.; Pikov, V.; Meng, E. 3d parylene sheath neural probe for chronic recordings. Journal of neural engineering 2013, 10, 045002. 52. Hara, S.A.; Kim, B.J.; Kuo, J.T.; Lee, C.D.; Meng, E.; Pikov, V. Long-term stability of intracortical recordings using perforated and arrayed parylene sheath electrodes. Journal of neural engineering 2016, 13, 066020. 53. Seymour, J.P.; Langhals, N.B.; Anderson, D.J.; Kipke, D.R. Novel multi-sided, microelectrode arrays for implantable neural applications. Biomedical microdevices 2011, 13, 441-451. 54. Kato, Y.; Nishino, M.; Saito, I.; Suzuki, T.; Mabuchi, K. In Flexible intracortical neural probe with biodegradable polymer for delivering bioactive components, Microtechnologies in Medicine and Biology, 2006 International Conference on, 2006; IEEE: pp 143-146. 13 55. Ledochowitsch, P.; Olivero, E.; Blanche, T.; Maharbiz, M.M. In A transparent μecog array for simultaneous recording and optogenetic stimulation, Engineering in Medicine and Biology Society, EMBC, 2011 Annual International Conference of the IEEE, 2011; IEEE: pp 2937-2940. 56. Ledochowitsch, P.; Félus, R.; Gibboni, R.; Miyakawa, A.; Bao, S.; Maharbiz, M. In Fabrication and testing of a large area, high density, parylene mems µecog array, Micro Electro Mechanical Systems (MEMS), 2011 IEEE 24th International Conference on, 2011; IEEE: pp 1031-1034. 57. Chou, L.-C.; Tsai, S.-W.; Chang, W.-L.; Chiou, J.-C.; Chiu, T.-W. In A parylene-c based 16 channels flexible bio-electrode for ecog recording, SENSORS, 2014 IEEE, 2014; IEEE: pp 877-880. 58. Takeuchi, S.; Ziegler, D.; Yoshida, Y.; Mabuchi, K.; Suzuki, T. Parylene flexible neural probes integrated with microfluidic channels. Lab on a Chip 2005, 5, 519-523. 59. Suzuki, T.; Kotake, N.; Mabuchi, K.; Takeuchi, S. In Flexible regeneration-type nerve electrode with integrated microfluidic channels, Microtechnologies in Medicine and Biology, 2006 International Conference on, 2006; IEEE: pp 303-305. 60. Ziegler, D.; Suzuki, T.; Takeuchi, S. Fabrication of flexible neural probes with built-in microfluidic channels by thermal bonding of parylene. Journal of Microelectromechanical Systems 2006, 15, 1477-1482. 61. Yu, H.; Xiong, W.; Zhang, H.; Wang, W.; Li, Z. A parylene self-locking cuff electrode for peripheral nerve stimulation and recording. Journal of Microelectromechanical Systems 2014, 23, 1025-1035. 62. Kang, X.; Liu, J.-Q.; Tian, H.; Yang, B.; Nuli, Y.; Yang, C. Self-closed parylene cuff electrode for peripheral nerve recording. Journal of Microelectromechanical Systems 2015, 24, 319-332. 63. Cobo, A.M.; Boyajian, B.; Larson, C.; Scholten, K.; Pikov, V.; Meng, E. In A parylene cuff electrode for peripheral nerve recording and drug delivery, Micro Electro Mechanical Systems (MEMS), 2017 IEEE 30th International Conference on, 2017; IEEE: pp 506-509. 64. Nandra, M.S.; Lavrov, I.A.; Edgerton, V.R.; Tai, Y.-C. In A parylene-based microelectrode array implant for spinal cord stimulation in rats, Micro Electro Mechanical Systems (MEMS), 2011 IEEE 24th International Conference on, 2011; IEEE: pp 1007-1010. 65. Wang, R.; Huang, X.; Liu, G.; Wang, W.; Dong, F.; Li, Z. Fabrication and characterization of a parylene- based three-dimensional microelectrode array for use in retinal prosthesis. Journal of Microelectromechanical Systems 2010, 19, 367-374. 66. Liu, Y.; Park, J.; Lang, R.J.; Emami-Neyestanak, A.; Pellegrino, S.; Humayun, M.S.; Tai, Y.-C. In Parylene origami structure for intraocular implantation, Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS & EUROSENSORS XXVII), 2013 Transducers & Eurosensors XXVII: The 17th International Conference on, 2013; IEEE: pp 1549-1552. 67. Johnson, A.C.; Wise, K.D. In A self-curling monolithically-backed active high-density cochlear electrode array, Micro Electro Mechanical Systems (MEMS), 2012 IEEE 25th International Conference on, 2012; IEEE: pp 914-917. 14 arylene is a promising material for constructing flexible, biocompatible, and corrosion resistant MEMS devices. Historically, Parylene has been employed as an encapsulation material for medical implants, such as stents and pacemakers, due to its strong barrier properties and biocompatibility. In the past few decades, the adaptation of planar microfabrication processes to thin film Parylene has encouraged its use as an insulator, structural, and substrate material for MEMS and other microelectronic devices. However, Parylene presents unique challenges during microfabrication and during use with liquids, especially for flexible, thin film electronic devices. In particular, the flexibility and low thermal budget of Parylene require modification of the fabrication techniques inherited from silicon MEMS, and poor adhesion at Parylene-Parylene and Parylene-metal interfaces causes device failure under prolonged use in wet environments. Here, I discuss in detail the promises and challenges inherent to Parylene and present my laboratory experience in developing thin-film Parylene devices. CONSIDERATIONS IN THE FABRICATION OF PARYLENE IMPLANTABLE DEVICES P 15 Challenges Despite growing interest, many researchers report difficulties when processing thin-film Parylene devices, as well as various modes of material and device failure. These problems largely stem from a lack of well-defined protocols for the machining, use and handling of Parylene-based devices. Common processing techniques developed for semiconductor materials and glass are often incompatible with organic polymers, owing to Parylene’s limited thermal budget, gas permeability, low mechanical strength and unique chemical properties. While many of these obstacles are surmountable, solutions are rarely discussed in the literature. Here a compilation of common challenges encountered during the construction of Parylene devices and a description of current best practices to avoid these issues are presented. 2.1.1 Thermal Budget One of the most persistent challenges of Parylene is its limited thermal budget, a consequence of its thermoplastic nature; in atmosphere Parylene is subject to oxidation at temperatures greater than 100 °C, glass transition temperature around 90 °C, and a melting temperature at 290 °C [1]. These temperatures are commonly encountered and even surpassed during standard silicon micromachining and electrical packaging. Figure 2-1. Images of Parylene damage due to high heat or radiation exposure. (a) Cracked Parylene after development and soft baking at 115 °C for 3 min; and (b) air bubbles appeared under the film, during hard bake at 90 °C for 15 min, in regions previously exposed to UV light during lithography. 16 For example, during photolithography, photoresist must be soft-baked at elevated temperatures (100-120 °C) to remove residual solvent, and exothermic reactions during the UV activation of the resist can generate temperatures of up to 200 °C [2]. Soldering, plasma etching, PVD, and other methods common to micromachining require temperatures which can cause Parylene to burn, bubble, or crack (Figure 2-1). Thermal annealing is a common treatment for improving adhesion between Parylene-Parylene interfaces [3,4]. This process requires high temperature (> 200 °C), compared to Parylene’s glass transition temperature, and must be performed under vacuum to avoid the effects of oxidation (browning, wrinkling, and becoming brittle) as shown in Figure 2-2. Figure 2-2. Comparison of thermal annealed devices in the presence of oxygen (left and center) and under intact vacuum (right). Devices were heated at 200 °C for 48 hours. 2.1.2 Water Diffusion Parylene is employed as encapsulation for medical implants, such as stents, pacemakers, and neural probes [5,6], due to its strong water barrier properties. However, like all polymers, Parylene is in fact permeable to moisture and gases [7]. Both water and ions in solution (i.e. salt) can diffuse through thin layers of Parylene in less time than the intended use duration. For a Parylene-based medical device or microdevice with insulation layers only a few microns thick, limiting permeation is critical to prevent electrical shorts, corrosion of encapsulated components, and catastrophic failure. The most effective method at preventing water permeation is the use of thermal annealing to increase the crystallinity, though this only serves to slow permeation, not prevent it outright [4,8]. Unfortunately, there is scarce data quantifying the scale of the problem, 17 particularly for the very thin (~ µm) films used in Parylene-based microdevices. Below there is a description of some observations and measurements that have been collected on the topic. 2.1.2.1 Water Permeability Published values for water vapor transmission rate (WVTR) in Parylene can vary quite substantially between sources. Specialty Coating Systems (SCS) [9] and Para Tech [10], two prominent vendors for Parylene coatings, publish WVTR values of 0.0830 and 0.0550 g-mm/m 2 - day respectively at 37 °C. Menon et al. [11] measured water vapor permeation in thin, small-area Parylene structures for both annealed and un-annealed films; WVTR values calculated from reported permeation measurements are listed in Table 1. WVTR values may differ based on deposition parameters and may be non-linear with film thickness for very thin films, as transmission through defects or large pores in thin-films may dominate over diffusion-driven transmission. Table 2-1 provides WVTR measurements for annealed and un-annealed films of varying thicknesses. Films were deposited on planar silicon wafers using a PDS Labcoter 2010 (SCS, Indianapolis, IN); annealed films were heated for 48 hours at 200 °C under vacuum. The measurements were based on the beaker method for determination of water vapor permeability [12]. Glass beakers with ~40 mL of deionized water were sealed with Parylene (1.7 cm diameter circular aperture) using marine epoxy. The beakers, kept at room temperature and approximately 33% relative humidity, were weighed weekly for 2 months (mean values recorded in Table 2-1). Notably, my values differ from Menon et al. despite similar temperatures and humidity. In agreement with Menon et al., a significant decrease in WVTR values is observed after annealing films; a 25% decrease in WVTR was observed for 10 and 15 µm thick films, in general agreement with reports and anecdotes from researchers noting a decrease in water permeation and an increase in effective lifetime for annealed Parylene/Parylene-coated devices. There was no significant decrease in WVTR for the annealed 5 µm film, which indicate that films at this thickness have permeation driven by defects rather than diffusion through the bulk. 18 Table 2-1. WVTR measurements for different un- and annealed Parylene film thickness. 2.1.2.2 Ion Permeability Parylene is known for having excellent ionic barrier properties, particularly when compared to other polymers. Under salt-fog tests (ASTM B117-(03)), SCS reported no evidence of corrosion or salt deposits on Parylene-coated PCB boards after 144 hours of exposure, [9] and Mordelt et al. reported 25 µm thick layers of Parylene withstood 0.9% saline solution for up to 30 days before breakdown [13]. Very thin layers of Parylene (< 10 µm), however, are still susceptible to ion intrusion over relatively short time periods. In many Parylene-based microdevices, ionic intrusion into a Parylene insulation layer can change its dielectric properties, thereby increasing parasitic coupling between lines or decreasing shunt impedance [14,15]. In this manner, ion permeability may affect device performance even if ions never breach the Parylene barrier. Electrochemical impedance spectroscopy (EIS) was used to determine the time course of ion transport through freestanding Parylene films. Different thickness of un-annelaed and annelaed, 48 hours at 200 °C under vacuum, Parylene films were tested. The electrochemical cell (Adams & Chitteden Scientific Glass, CA, USA) allowed for the placement of a Parylene film between two chambers and reference electrodes thourgh the ports (Figure 2-3). An area of 5.3 cm 2 was exposed to the solutions. One chamber was filled with Millipore water and the other with 1x PBS. This is a three-electrode setup: each chamber contains two platinum wires (1.5 cm long) to Reference Thickness (µm) Mean WVTR (g-mm/m 2 -day) Standard Error Temperature and Relative Humidity (mean ± SE) This work Un-annealed 5 0.0194 0.0012 20.6 ± 0.14 °C, 34 ± 3% 10 0.0176 0.0010 20.7 ± 0.13 °C, 32 ± 3% 15 0.0185 0.0025 20.8 ± 0.12 °C, 33 ± 3% Annealed 5 0.0185 0.0014 20.6 ± 0.14 °C, 33 ± 3% 10 0.0128 0.0010 20.7 ± 0.14 °C, 34 ± 3% 15 0.0139 0.0024 20.7 ± 0.13 °C, 35 ± 3% Specialty Coating Systems - 0.0830 - 37 °C, 90% ASTM F1249 Para Tech - 0.0550 - 37 °C, 90% ASTM F1249 Menon et al., 2009 Un-annealed 9 0.0547 - 20 °C, 30% ASTM D1653 Annealed 9 0.0276 - 19 serve as working and counter electrodes and one Ag/AgCl reference electrode. EIS measurments were taken every 15 minutes for 3 hours daily until the solution resistance reached 10% of its original value. The solution resistance was measured at the appropiate frequency determined by minimizing the impedance phase. See testing protocol in Appendix A. Figure 2-3. Borosilicate electrochemical cell. Table 2-2 provides the ion transport measurements for annealed and un-annealed films of varying thicknesses. Notably, ions took longer to pass across the Parylene films as thickness increased and after annealing films; an increment in time was observed for 5 and 10 µm thick films after annealing, whereas a decrement in time was observed for ~15 µm thick films after annealing. This effect is due to the difference of thickness between films, un-annealed films were ~16 µm while annealed films were ~15 µm. Table 2-2. Ion transport measurements for different un- and annealed Parylene film thickness. Treatment Thickness (µm) Time (hours) Standard Error 5 1.5 0.22 Un-annealed 10 47 6.03 16 58 7.05 5 3 0.5 Annealed 10 52.8 3.25 15 22.9 1.37 20 2.1.3 Adhesion Parylene suffers from poor adhesion to itself and noble metals, such as gold and platinum, a considerable drawback in the implementation of thin-film Parylene implants. The adhesion of Parylene devices is compromised when devices are soaked in wet environments [6,8,14,16,17]. Weak adhesion can accelerate catastrophic failure of Parylene devices; as Parylene films lift off a substrate, voids form in which water vapor can condense, creating continuous paths of solution that create electrical shorts and drive further delamination (Figure 2-4). Several strategies were investigated to improve adhesion of Parylene to substrates such as silicon and glass, including melting, anchoring, surface roughening, thermal annealing, surface plasma treatment and the inclusion of chemical layers, such as silane A-174 and plasma polymerized adhesion layers [18- 20], but many of these techniques are not applicable or calibrated for adhesion of Parylene to Parylene or thin-film metals used in polymer MEMS and flexible electronics. Figure 2-4. Parylene devices where delamination was noticeable on the macro scale. (a) Parylene layers visibly split from each other; (b) detachment of the top Parylene layer from the metal and bottom Parylene layers; and (c) electrodes and metal traces began to move or slide around between the Parylene layers. Adhesion between two surfaces is achieved through chemical and physical interventions. In the case of Parylene-Parylene interfaces, adhesion is dominated by physical adsorption, whereas for Parylene-metal interfaces, adhesion is typically via a combination of hydrogen bonding and Van der Waals forces [4]. Poor adhesion between Parylene and other materials may result from differences in surface energy at the interface [21] between hydrophobic Parylene and hydrophilic metals, a phenomenon exacerbated by surface contamination. The presence of internal stress between layers [8] appears to further aggravate adhesion failure and hasten delamination. Delamination under wet conditions may be induced by water vapor condensing 21 within voids created by surface particulates present during CVD [22], suggesting that surface cleanliness is critical in maximizing adhesion. Annealing of Parylene, 48 hours at 200 °C under vacuum, has been reported to improve adhesion. Reports suggest that annealing increases the entanglement of the polymer chains while reducing stress by recrystallization of Parylene- Parylene interfaces [5,23]. Interposer layers have also been investigated to either modify the surface energy of a coated material, thereby improving chemical adhesion, or to create a barrier against water vapor, minimizing water intrusion and subsequent failure [24-26]. Plasma- enhanced Parylene is a method to improve adhesion of Parylene to other surfaces by cleaning and modifying the film within the process chamber [27]. 2.1.4 Handling Finally, due to the thin and flexible nature of Parylene devices, they can be potentially damaged by rough handling. Electrical traces incorporated into Parylene devices or ribbon cables are actually surprisingly robust, able to survive down to a bend radius of 100 µm and under fatigue testing of up to 100,000 bends [24]. However, the wrinkling or crumpling of Parylene devices, which can happen inadvertently, can cause destructive creases in thin metal connections. Parylene devices are also very light, and can easily be blown away by the nitrogen streams commonly used for cleaning microdevices, or by the venting of a vacuum chamber, or even by exhalation. They are also subject to strong static forces, owing to Parylene’s properties as an electrical insulator. Parylene devices should be gingerly handle, with tweezers, held by the edge of the device or even by a purpose built Parylene tab designed into the structure. Weakly adhesive double-sided tape can be used to secure Parylene devices temporarily, during post-processing, packaging or imaging. Despite all the challenges, Parylene is a promising material for medical applications. However, the development of low temperature processes and improvement of adhesion are necessary to guarantee the reliably long-term performance of thin-film Parylene implants. 22 Micromachining of Parylene Films Even though processes for silicon were successfully adapted to micromachine Parylene, there is a lack of well-defined protocols and standards when working with this polymer. Parylene-based MEMS and microdevices are typically constructed using a combination of bulk and surface micromachining and photolithography. In a typical process flow, a foundational Parylene layer is deposited on a support substrate, almost always a silicon wafer with its native oxide layer intact, using CVD. Subsequent layers of metal or polymer are then deposited and patterned using photolithography. Metal layers commonly serve as conductive traces or electrodes and are deposited using evaporation or sputtering and patterned by lift-off. Additional polymer layers include additional Parylene films, which may be patterned on-top of sacrificial photoresist patterns to create three-dimensional structures. Structures may be created using O2 plasma etching through a photoresist mask, to create MEMS components using bulk- micromachining, or to expose metal electrodes covered by polymer insulation. Finally, the complete device is removed from the support substrate and packaged. Below we describe challenges and solutions encountered during each step of fabrication for Parylene-based microdevices. 2.2.1 Deposition Parylene is deposited by CVD, producing a highly uniform and conformal coating [28]. Typically, the film is transparent and homogenous; however, in some instances Parylene coatings may be marred by odd “spherule” inclusions. The macroscopic appearance is hazy and white, sometimes described as “cloudy” Parylene. The microscopic appearance is presented in Figure 2-5. The spherules may be unreacted Parylene monomers that bond to each other in the gas phase prior to deposition, a result of insufficient molecular collisions before deposition caused by a high volume-to-surface-area of the deposition chamber [29]. Increasing the surface area of the chamber, by including structures with large surface-area such as a mesh, may prevent the formation of these spherules. SCS, a manufacturer of Parylene coating tools, describes the cause as high deposition rates and chamber pressure. 23 This phenomenon was repeatedly observed when coating Parylene-based devices that were previously subjected to some form of mechanical action, such as ultrasonic treatment or scrubbing during a metal-liftoff process. These devices, when insulated with CVD Parylene, frequently present with this cloudy appearance, with the densest appearances of spherules along the edges of thin-film metal structures or at locations where the mechanical action was most severe. It is unclear whether these phenomena interfere with adhesion between the Parylene and other layers and whether they impact the integrity of the Parylene coating. Figure 2-5. (a) Deposited Parylene layer incorporating spherules which results in a cloudy in appearance when observed by eye; and (b) magnified photograph revealing the presence of small clusters of spherules. 2.2.2 Lithographic Processes Microfabrication commonly entails photolithographic patterning of photoresist to serve as an etch mask, lift-off mask, or patterned sacrificial layer. Photolithography involves a series of sub-processes, such as coating, pre/post baking, UV light exposure and development. These processes involve the use of heat and UV radiation and all can compromise the integrity of Parylene films. For example, the combination of Parylene gas permeability and the off-gassing of photoresist during UV exposure [2] can lead to the formation of bubbles in Parylene film. The phenomenon tends to be more pronounced with thicker resists and higher exposure dosage of UV radiation. Photodegradation of Parylene has been reported in literature through a two-step process involving direct photolytic processes resulting in the formation of UV and IR absorbing structures, 24 followed by photo-induced oxidation of the methylene groups and benzene ring [30]. It is also well known that UV radiation can deteriorate the thermal and electrical properties of Parylene films if the doses are large (>12 J/cm2) [31]. Although the small doses of UV exposure during lithography are unlikely to reach the threshold for full photodegradation, it is hypothesized that some combination of minor oxidation and mechanical/thermal stress may be responsible for the observed phenomenon when thick resist films are used. For example, bubbles between the Parylene and substrate may appear following the UV exposure step for photolithography. Piercing the Parylene bubbles in select non-critical areas and the use of a vacuum (after piercing) was found to aid in removal of the bubbles. 2.2.3 Metal Deposition Metal deposition is a key process in the fabrication of polymer MEMS devices to create conductive elements such as traces and electrodes. The deposition of metal on Parylene, however, is rife with complications. Both sputtering and evaporation can induce significant intrinsic and extrinsic stress which can induce curvature in the resulting free-film Parylene devices. In addition, the deposition of high melting point metals, such as platinum, through high temperature processes, such as evaporation, can lead to cracking of the metal film or the underlying Parylene due to thermal stress, a result of mismatch in the thermal coefficients of expansion and film stress. Figure 2-6 shows images of platinum (2000 Å) deposited on a Parylene coated silicon wafer (10 µm thick Parylene film), with stress induced cracks. The platinum was evaporated with a Temescal BJD-1800 e-beam evaporator using an uncooled stage; during evaporation wafer temperature surpassed 110 °C as measured with temperature monitor stickers (Omega Engineering, Norwalk, CT, USA) placed on the back of the wafer. Changes in the deposition rate and tool power were insufficient to prevent cracking. Improving the thermal contact between the wafer and the stage is difficult, in part because the Parylene which coats the back of the silicon wafer serves as an insulator. Breaking the deposition into a series of four steps with 15 min pauses between each deposition can prevent cracking with metals such as titanium, evaporated at a lower temperature but is insufficient for platinum. Ultimately, the most reliable method to avoid cracking was to use a tool (CHA Industries MARK 40, Fremont, CA) with a larger throw distance 25 between the metal target and the wafer stage (22″ compared with 8″ for the Temescal) and depositing the platinum in four 500 Å steps. Heat was measured with temperature monitor stickers as previously described and was maintained below 77 °C. Figure 2-6. (a) Stress-induced cracking of deposited platinum likely due to excess heat generated during the process; and (b) cracked metal traces after lift-off. Depositing metal through sputtering can avoid thermal stress and cracking but can impart severe film stress, warping the Parylene film and forcing released devices to curve. Figure 2-7a is an image of a Parylene substrate patterned with a negative profile photoresist for metal lift-off (AZ 5214E-IR) and sputter coated with 2000 Å of platinum. The rippled appearance is typical of Parylene films coated by sputtering and highlights the severity of film stress (Figure 2-7b). Figure 2-7. (a) Severe wrinkling/rippling seen in sputtered deposited platinum film results from compressive stresses of a higher magnitude; and (b) wrinkled areas around the device metal features. 26 2.2.4 Etching Due to the inertness of Parylene it is not practical to etch it chemically. Although there are reports of wet etching Parylene using chloronapthelene or benzoyl benzonate at 150 °C [32], this is an extreme temperature for Parylene and can affect the bulk properties of unetched sections. Thus, mechanisms to selectively etch Parylene are limited to dry techniques. Parylene can be removed with oxygen plasma etching [33-36], reactive ion beam etching [37], oxygen reactive ion etching (RIE) and a deep reactive ion etching (DRIE) like method involving alternating cycles of oxygen plasma etching and fluorocarbon passivation layer deposition [33,37-41]. The latter method is referred to as O2 DRIE. Photoresist etch masks are most commonly used to lithographically pattern Parylene, despite the low selectivity (approximately 1:1). Other materials including some metals can be used as masking materials if a high etch rate and good selectivity are required [40] but the coating and etching of metal masks may induce many of the thermal mismatch problems described above. Additionally, re-deposition of metal during the etch process is common and as such photoresists remain the most common choice [41,42]. Across a broad range of processes, AZ P4620, a positive- tone photoresist, was used as a Parylene etch mask for both O2 RIE and an O2 DRIE process (presented in [38]) to etch a variety of structures into Parylene film, ranging from 1 to 15 µm deep. Selectivity rates typically vary between 0.9 and 1, for vertical etch rates varying between 0.55 and 0.80 µm/min. The most common problem observed during plasma etching of Parylene is the formation of gas bubbles, either within photoresist layers or between the support substrate and Parylene coating. The apparent cause is the off-gassing of volatile components of photoresist or remnants of the solvent used to distribute the photoresist. Minor off-gassing from photoresists may be inconsequential, or even unnoticeable, when processing on rigid substrates, however Parylene is gas-permeable and even minor volatile products can produce bubbling when under vacuum. Bubbles may deform photoresist masks, creating holes which lead to etch damage, or deform the Parylene itself making future processing difficult. Figure 2-8 shows a representative example photoresist bubbling observed following a O2 DRIE using 80 W RF and 900 W ICP power. This 27 phenomenon occurs more readily when using high power, high density, inductively coupled plasmas. Very thick photoresist layers (>10 µm) and, in particular, very thick edge beads are more susceptible, likely due to the residual solvent not driven off during soft-baking steps. The appearance of gas bubbles was also observed when processing multi-layer Parylene devices featuring three-dimensional Parylene structures defined using sacrificial layers. This is a common technique where photoresist is patterned and then coated in Parylene to create Parylene-based microfluidic or mechanical structures. These sacrificial layers are particularly susceptible to forming bubbles during subsequent etch steps, as gas expands rapidly under the low pressure required for plasma etching, faster than the rate of diffusion through the Parylene film. Figure 2-8. Out-gassing of photoresist solvent results in bubbling observed in sacrificial photoresist structures sandwiched in Parylene following O2 DRIE. The formation of these gas bubbles appears particularly sensitive to the hard-baking step following photoresist development. In a simple experiment, Parylene coated 4″ silicon wafers were masked with 10 µm thick AZ P4620 and, following development, hard-baked on a hot plate at 90 °C for either 0, 2, or 12 h. Following a brief O2 DRIE process (approximately 2 µm etch), we observed catastrophic bubbling for the sample baked for 2 h, with minimal or no bubbling for unbaked samples and samples baked for 12 h. The hypothesis is that after 2 h of hard baking, volatiles diffuse through the Parylene layer and become trapped between the Parylene and silicon wafer substrate once the wafer cools. These volatiles are responsible for later bubbling in the DRIE. By contrast, 12 h of hard baking may completely drive off photoresist volatiles from the wafer, while omitting the hard bake entirely prevents any diffusion of volatiles under the Parylene. Avoiding a hard-bake step is strongly recommended when etching Parylene through photoresist 28 masks. If a hard-bake is required, or if resist is intended for use as a sacrificial structure, a lengthy hard-bake is suggested at low (<100 °C) temperature under an inert atmosphere to minimize thermal stress and oxidation. 2.2.5 Release Releasing the device from the support or carrier wafer is the final step in Parylene MEMS processing. Most commonly, the support is a silicon wafer with a thin native oxide layer and the Parylene adhesion to this material is quite weak. In such cases, Parylene devices can be simply peeled off using tweezers, or the wafer can be submerged in water and the devices allowed to float to the surface. In some cases, however, the adhesion between Parylene and the support can be very strong and releasing Parylene without destroying the thin-film devices can be nearly impossible. In many cases this adhesion is deliberate; sacrificial adhesion layers such as aluminum [43,44] or chrome [45] have been used for complex Parylene processing and devices are then released using KOH or anodic NaCl [43,44] etching for Al, or chrome etchant for Cr [45]. The release of Parylene devices from a support substrate of polyethylene terephthalate (PET) was also demonstrated by boiling the ensemble in water for several minutes. In other scenarios, the adhesion may be unintentional or undesirable. Parylene devices annealed on wafer at 200 °C for 48 h under vacuum became virtually impossible to remove. In such situations, the only remaining option may be to etch away the underlying silicon or other carrier substrate. Discussion Applying complex, high-resolution micromachining processes to Parylene substrates is possible but will require modification of nearly every protocol developed for silicon or other rigid substrates. No single definitive set of guidelines exists but a useful compilation of best practices can be described. Efforts should be taken to minimize exposure to heat during every step of the fabrication process. Avoiding the glass transition temperature (90 °C) of Parylene is typically unrealistic but avoiding or minimizing the use of high temperature (>100 °C) processes can prevent oxidization, 29 thermal stress, the formation of gas bubbles, or irreversible changes to Parylene morphology. High temperature anneals (200 °C for 48 h in vacuum) are useful to improve moisture barrier properties or to shape Parylene by exploiting thermoplasticity, but must be done under vacuum or inert atmosphere and Parylene must be allowed to cool slowly to room temperatures over several hours. Rapid heating above 120 °C on a hotplate or during PVD can introduce destructive stress and must be avoided. Parylene is broadly amenable to most photolithography techniques, including electron- beam lithography but users should be aware of the risk of oxidation due to radiation exposure and complications from gas transport in Parylene films arising from off-gassing of thick photoresist films. Avoiding the use of a hard-bake following development, removing thick edge beads and avoiding high intensity or long duration UV exposure, are all recommended practices. Patterned photoresist masks have been used with success to transfer high-aspect ratio structures to Parylene substrates using O2 RIE and DRIE but this method requires very thick photoresist films due to low selectivity during O2 etching. Metal deposition onto Parylene films can prove incredibly challenging. Evaporative PVD can introduce thermal stress and cracking of either the Parylene or metal structures, while sputtering can introduce film-stress which can warp or wrinkle the Parylene film. These challenges are exacerbated by high-melting point metals and thicker metal films. The use of cooled stages or heat-sinking may help during evaporative PVD but the insulative nature of Parylene coatings makes such thermal control difficult. The best results were obtained by using an e-beam deposition system with a throw distance (distance between target and metal crucible) greater than 20″ and splitting depositions of thick films into multiple steps punctuated with 15- minute pauses. Finally, for those Parylene-based bioMEMS intended for chronic use under wet-saline environment, the risk of water permeation or delamination remains one of the largest obstacles. WVTR decreased significantly, for films thicker than 5 µm, following a 48-hour thermal anneal and this is consistent with various literature and anecdotal reports about the advantages of annealing Parylene devices and the difficulties using very thin Parylene as a moisture barrier. 30 The large variation in WVTR measurements between different sources prompts a call for more research into how deposition parameters change barrier properties. New methods to improve Parylene adhesion to thin-film metal would similarly help increase Parylene-device lifetime in vivo. Conclusions In conclusion, while applying micromachining to Parylene substrates can be challenging, with attention to the limited thermal budget and polymeric properties of Parylene, most processes can be successfully transferred to this thin, flexible material. The observations and recommendations listed here can serve as the basis for a series of best-practices regarding Parylene microfabrication. Continuing research is needed, particularly on the adhesion and barrier properties in the construction of multi-layer devices. By solving these and other challenges Parylene stands to become a key material in a new generation of flexible electronic microdevices and polymer MEMS. References 1. Noh, H.-S.; Huang, Y.; Hesketh, P.J. Parylene micromolding, a rapid and low-cost fabrication method for parylene microchannel. Sensors and Actuators B: Chemical 2004, 102, 78-85. 2. Reiser, A.; Huang, J.; He, X.; Yeh, T.; Jha, S.; Shih, H.; Kim, M.; Han, Y.; Yan, K. The molecular mechanism of novolak–diazonaphthoquinone resists. European Polymer Journal 2002, 38, 619-629. 3. Kim, B.J.; Washabaugh, E.P.; Meng, E. In Annealing effects on flexible multi-layered parylene-based sensors, Micro Electro Mechanical Systems (MEMS), 2014 IEEE 27th International Conference on, 2014; IEEE: pp 825-828. 4. Ortigoza-Diaz, J.; Scholten, K.; Meng, E. Characterization and modification of adhesion in dry and wet environments in thin-film parylene systems. Journal of Microelectromechanical Systems 2018, 1-12. 5. Hassler, C.; von Metzen, R.P.; Ruther, P.; Stieglitz, T. Characterization of parylene c as an encapsulation material for implanted neural prostheses. Journal of Biomedical Materials Research Part B: Applied Biomaterials 2010, 93, 266-274. 6. Hsu, J.-M.; Rieth, L.; Normann, R.A.; Tathireddy, P.; Solzbacher, F. Encapsulation of an integrated neural interface device with parylene c. IEEE Transactions on Biomedical Engineering 2009, 56, 23-29. 7. Spivack, M.; Ferrante, G. Determination of the water vapor permeability and continuity of ultrathin parylene membranes. Journal of The Electrochemical Society 1969, 116, 1592-1594. 8. Von Metzen, R.P.; Stieglitz, T. The effects of annealing on mechanical, chemical, and physical properties and structural stability of parylene c. Biomedical Microdevices 2013, 15, 727-735. 9. Systems, S.C. Scs parylene properties; 2007. 31 10. Coating, P.T. Parylene properties. 11. Menon, P.; Li, W.; Tooker, A.; Tai, Y. In Characterization of water vapor permeation through thin film parylene c, Solid-State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, 2009; IEEE: pp 1892-1895. 12. International, A. Standard test methods for water vapor transmission of organic coating films. 2003; Vol. ASTM D1653-03. 13. Mordelt, G.; Heim, P. Parylene: High-tech-beschichtung der zukunft. Metalloberfläche 1998, 52, 368-371. 14. Seymour, J.P.; Elkasabi, Y.M.; Chen, H.-Y.; Lahann, J.; Kipke, D.R. The insulation performance of reactive parylene films in implantable electronic devices. Biomaterials 2009, 30, 6158-6167. 15. Xie, X.; Rieth, L.; Tathireddy, P.; Solzbacher, F. Long-term in-vivo investigation of parylene-c as encapsulation material for neural interfaces. Procedia Engineering 2011, 25, 483-486. 16. Li, W.; Rodger, D.; Menon, P.; Tai, Y.-C. Corrosion behavior of parylene-metal-parylene thin films in saline. Ecs Transactions 2008, 11, 1-6. 17. Li, W.; Rodger, D.C.; Meng, E.; Weiland, J.D.; Humayun, M.S.; Tai, Y.-C. In Flexible parylene packaged intraocular coil for retinal prostheses, Microtechnologies in Medicine and Biology, 2006 International Conference on, 2006; IEEE: pp 105-108. 18. Huang, R.; Tai, Y. In Parylene to silicon adhesion enhancement, Solid-State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, 2009; IEEE: pp 1027-1030. 19. Charmet, J.; Bitterli, J.; Sereda, O.; Liley, M.; Renaud, P.; Keppner, H. Optimizing parylene c adhesion for mems processes: Potassium hydroxide wet etching. Journal of Microelectromechanical Systems 2013, 22, 855-864. 20. Yamagishi, F.G. Investigations of plasma-polymerized films as primers for parylene-c coatings on neural prosthesis materials. Thin Solid Films 1991, 202, 39-50. 21. Hwang, K.S.; Park, J.H.; Lee, J.H.; Yoon, D.S.; Kim, T.S.; Han, I.; Noh, J.H. Effect of atmospheric-plasma treatments for enhancing adhesion of au on parylene-c-coated protein chips. Journal of the Korean Physical Society 2004, 44, 1168-1172. 22. Schmidt, E.; McIntosh, J.; Bak, M. Long-term implants of parylene-c coated microelectrodes. Medical and Biological Engineering and Computing 1988, 26, 96-101. 23. Davis, E.M.; Benetatos, N.M.; Regnault, W.F.; Winey, K.I.; Elabd, Y.A. The influence of thermal history on structure and water transport in parylene c coatings. Polymer 2011, 52, 5378-5386. 24. Lee, C.D.; Meng, E. Mechanical properties of thin-film parylene–metal–parylene devices. Frontiers in Mechanical Engineering 2015, 1, 10. 25. Radun, V.; von Metzen, R.; Stieglitz, T.; Bucher, V.; Stett, A. Evaluation of adhesion promoters for parylene c on gold metallization. Current Directions in Biomedical Engineering 2015, 1, 493-497. 26. Xie, Y.; Pei, W.; Guo, D.; Zhang, L.; Zhang, H.; Guo, X.; Xing, X.; Yang, X.; Wang, F.; Gui, Q. Improving adhesion strength between layers of an implantable parylene-c electrode. Sensors and Actuators A: Physical 2017, 260, 117-123. 27. Bienkiewicz, J. Plasma-enhanced parylene coating. Medical device technology 2006, 17, 10. 28. Loeb, G.E.; Bak, M.; Salcman, M.; Schmidt, E. Parylene as a chronically stable, reproducible microelectrode insulator. IEEE Transactions on Biomedical Engineering 1977, 121-128. 29. Loeb, G.; Peck, R.; Martyniuk, J. Toward the ultimate metal microelectrode. Journal of neuroscience methods 1995, 63, 175-183. 30. Bera, M.; Rivaton, A.; Gandon, C.; Gardette, J. Comparison of the photodegradation of parylene c and parylene n. European Polymer Journal 2000, 36, 1765-1777. 31. Fortin, J.; Lu, T.-M. Ultraviolet radiation induced degradation of poly-para-xylylene (parylene) thin films. Thin Solid Films 2001, 397, 223-228. 32. Carbide, U. Solvent resistance of the parylenes. Parylene Products. 32 33. Meng, E.; Tai, Y.-C. In Parylene etching techniques for microfluidics and biomems, Micro Electro Mechanical Systems, 2005. MEMS 2005. 18th IEEE International Conference on, 2005; IEEE: pp 568-571. 34. Nowlin, T.; Smith, D.F. Surface characterization of plasma‐treated poly‐p‐xylylene films. Journal of Applied Polymer Science 1980, 25, 1619-1632. 35. Wang, X.-Q.; Desai, A.; Tai, Y.-C.; Licklider, L.; Lee, T.D. In Polymer-based electrospray chips for mass spectrometry, Micro Electro Mechanical Systems, 1999. MEMS'99. Twelfth IEEE International Conference on, 1999; IEEE: pp 523-528. 36. Levy, B.P.; Campbell, S.L.; Rose, T.L. Definition of the geometric area of a microelectrode tip by plasma etching of parylene. IEEE transactions on biomedical engineering 1986, 1046-1049. 37. Ratier, B.; Jeong, Y.S.; Moliton, A.; Audebert, P. Vapor deposition polymerization and reactive ion beam etching of poly (p-xylylene) films for waveguide applications. Optical Materials 1999, 12, 229-233. 38. Meng, E.; Li, P.-Y.; Tai, Y.-C. Plasma removal of parylene c. Journal of Micromechanics and Microengineering 2008, 18, 045004. 39. Majid, N.; Dabral, S.; McDonald, J. The parylene-aluminum multilayer interconnection system for wafer scale integration and wafer scale hybrid packaging. Journal of electronic materials 1989, 18, 301- 311. 40. Yeh, J.; Grebe, K. Patterning of poly‐para‐xylylenes by reactive ion etching. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1983, 1, 604-608. 41. Lecomte, A.; Lecestre, A.; Bourrier, D.; Blatché, M.-C.; Jalabert, L.; Descamps, E.; Bergaud, C. Deep plasma etching of parylene c patterns for biomedical applications. Microelectronic Engineering 2017, 177, 70-73. 42. Selvarasah, S.; Chao, S.; Chen, C.-L.; Sridhar, S.; Busnaina, A.; Khademhosseini, A.; Dokmeci, M. A reusable high aspect ratio parylene-c shadow mask technology for diverse micropatterning applications. Sensors and Actuators A: Physical 2008, 145, 306-315. 43. Ziegler, D.; Suzuki, T.; Takeuchi, S. Fabrication of flexible neural probes with built-in microfluidic channels by thermal bonding of parylene. Journal of Microelectromechanical Systems 2006, 15, 1477-1482. 44. Mercanzini, A.; Cheung, K.; Buhl, D.L.; Boers, M.; Maillard, A.; Colin, P.; Bensadoun, J.-C.; Bertsch, A.; Renaud, P. Demonstration of cortical recording using novel flexible polymer neural probes. Sensors and Actuators A: Physical 2008, 143, 90-96. 45. Wu, F.; Tien, L.; Chen, F.; Kaplan, D.; Berke, J.; Yoon, E. In A multi-shank silk-backed parylene neural probe for reliable chronic recording, Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS & EUROSENSORS XXVII), 2013 Transducers & Eurosensors XXVII: The 17th International Conference on, 2013; IEEE: pp 888-891. 33 arylene is used as an insulator and coating due to its conformality and biocompatibility. It has unique properties that make it an attractive material for medical applications. Unfortunately, mechanisms to selectively etch Parylene are limited to dry methods. Deep reactive ion etching (DRIE), an anisotropic plasma etch, can be applied to Parylene for the creation of high aspect ratio MEMS structures by controlling the profile of the etched features. In order to improve the speed and anisotropy of plasma etching of Parylene, here, the optimization of a DRIE process was investigated by exploring changes to several operating parameters: inductively coupled plasma (ICP) power, radio frequency (RF) power, reactant’s flow rates and passivation layer. The effects of these parameters on the vertical etch rate, lateral etch rate, degree of anisotropy and sidewall angle were characterized. The characterization was enabled by the application of replica casting to obtain cross sections of etched structures in a non-destructive manner. OPTIMIZED DRIE ETCHING OF PARYLENE C P 34 Introduction There is a lot of potential for thin film polymers due to their biocompatibility and amenability to micromachining. Pattern transfer of masks into Parylene films is a critical step to fabricate multi-layer Parylene-based devices. However, Parylene is neither directly photopatternable nor amenable to lift-off patterning, due to the high conformality of Parylene deposition and the chemical inertness of the material. It is not practical to chemically etch Parylene, although there are reports of the wet etching of Parylene using chloronapthelene or benzoyl benzonate at 150 °C [1]. Melting of Parylene has been also used to expose the tip of microelectrodes [2,3]. These temperatures exceed the glass transition temperature for Parylene and could change the bulk properties of the polymer. High-voltage arcing method has been also used to expose tips of microlectrodes, however, it is a difficult method to control the exposure size [3]. In order to overcome this challenge, photoablation of Parylene with an UV laser was used to expose microelectrodes with precise ablations [4]. But this technique produces debris during the laser ablation representing a challenge [5]. Thus, dry etching techniques with minimal temperature increase during the process are the most effective and practical technique to etch Parylene. Oxygen plasma etching is used for Parylene micromachining [6]. Figure 3-1. Main processes of dry etching. Dry etching is the deliberate removal of material without employing liquid chemicals or etchants. Figure 3-1 shows the processes of dry etching: (1) reactive species are generated in the plasma; (2) species are transported to the etch target and adsorb on the surface; (3) chemical reactions take place at the surface; and (4) etch by-products desorb from the surface of the etch 35 target film [7]. The most popular plasma-based etching techniques in micromachining processes are reactive ion etching (RIE) and deep reactive ion etching (DRIE). RIE is an ion-enhanced, energy-driven etching process, in which a chemical process is accompanied by ionic bombardment. DRIE is an ion-enhanced inhibitor etching process that also employs both physical and chemical etching mechanisms. Multilayer-complex structures can be produced with Parylene if there is a high level of fidelity to the mask patterns, however, pattern transfer of masks into Parylene films is difficult to achieve. Thin layers of organic materials, such as, photoresist, can be used as masking material to protect regions of the substrate material in order to transfer the features. The thin layer is patterned by photolithography techniques. After etching, the masking material is dissolved with acetone. Other materials, such as metals, can be used as masking materials if a high etch rate and good selectivity are required [6]. Selectivity is the ratio of the etch rate of the masking material and the vertical etch rate of material to be etched, thus when a metal mask is used to etch Parylene the selectivity is good since little or no mask erosion occurs during O2 plasma. However, metal can redeposit during the etch process and photoresists are preferable even though the selectivity is very low [8]. An anisotropic etching replicates the mask patterns, thus, the ideal etch profile is completely anisotropic (Figure 3-2). Isotropic etching by chemical reactions produces undercut under the mask and makes fine patterning difficult, whereas physical removal produces anisotropic profiles, where the vertical etch rate is faster than the lateral etch. The etched profile can be “controlled” by changing the plasma parameters, such as, temperature, power, pressure, and gas flow rates, to mention some. Figure 3-2. Directionality of etching process. 36 3.1.1 DRIE Theory DRIE, known as Bosch process, is a highly anisotropic technique developed at Robert Bosch GmbH (Stuttgart, Germany) in 1994 typically for etching silicon [9,10]. It consists of an alternation between a deposition of an inhibition or passivation film and the etching of the material. The Bosch process includes several physical and chemical processes: gas ionization, ion bombardment, etc. Our modified process for etching Parylene starts with an isotropic etch step of the exposed Parylene with oxygen (O2), followed by a passivation step. During the passivation step, octafluorocyclobutane (C4F8) is deposited onto the wafer. At the beginning of the next etch step, the ions that bombard the substrate remove the C4F8 passivation layer at the bottom of the trench while preventing the etch of the sidewalls (Figure 3-3). The Parylene surface is exposed to reactive oxygen-based species, after the removal of C4F8 at the bottom of the trench, which chemically etch Parylene during this period while the sidewalls remained protected. Thus, high aspect ratio structures can be achieved by repeating the etch and passivation steps. Figure 3-3. Bosch process. DRIE etchers consists of two radio frequency (RF) generators (Figure 3-4). The first one, generates plasma by means of inductively coupling RF power in the source, and the second one, independently controls the ion energy by biasing the wafer electrode operating at high and low frequencies. During the process, the wafer is cooled down by helium backside flow. The wafer is clamped to the chuck either mechanically or electrostatically to ensure thermal contact to the electrode which is cooled down with water flow. If thermal contact is insufficient, the temperature of the wafer can rise too high affecting passivation deposition. 37 Figure 3-4. Schematic of an ICP system. Even though, DRIE has facilitated the microfabrication of high aspect ratio silicon structures, there are some challenges to overcome due to the requirement for etching patterns of different size for Parylene etching [11]. 3.1.2 Effects of Operating Parameters in DRIE The etch rate, profile and selectivity depend on different operating parameters as well as the mask and etched material. Thus, operating parameters must be investigated to optimize the DRIE of Parylene. Inductively coupled plasma (ICP) power, RF bias power, chamber pressure, flow rate of reactive species, and flow rate of passivation layer are some typical tunable operating parameters. 3.1.2.1 ICP Power An effective ICP power allows high density plasma to be maintained. Thus, with a high ICP power, more etchant species are accelerated towards the wafers causing the etch rate to increase as the ICP power increases. This will also improve the selectivity. 3.1.2.2 RF Bias Power The electron energy is increased by the RF bias power. Thus, the etch rate increases with a higher RF bias power. The directionality of the reactive species of the plasma increases as the RF bias power increases by accelerating the ions towards the wafer and increasing ion bombardment. As a result, this will increase the etching at the bottom of the trench rather than at 38 the sidewalls improving the degree of anisotropy. Increasing the RF bias power will rise the temperature of the wafer, which could worsen the selectivity. 3.1.2.3 Chamber Pressure The chamber pressure affects the mean free path of the ions. The directionality of the etching is affected with high pressures, which lead to more ion collisions. Thus, the etch rate and degree of anisotropy are improved with low pressures. Unfortunately, the plasma does not strike or sustain at very low pressures. 3.1.2.4 Flow Rate of Reactive Species High flow rate of reactive species introduces more ions in the process chamber and improves the etch rate. 3.1.2.5 Flow Rate of Passivation Layer High flow rate of passivation layer means more passivation on the sidewalls and at the bottom of feature to be etched. Thus, the etch step must be longer to completely remove the passivation at the bottom of the features. The passivation also depends on the temperature of the wafer, which is kept at room temperature by cooling the chuck with water and the wafer with helium backflow at the back of the wafer. Approach and goals The goal of these experiments is to optimize a DRIE protocol for Parylene substrates, with a focus on (1) maximizing vertical etch rate, (2) improving anisotropy as measured through sidewall profile, and (3) maximizing selectivity against the photoresist etch masks common in Parylene dry etching. The ability to etch deep features while controlling the profile of the etched features will expand the size and scope of Parylene MEMS devices. The foundation of the protocol is a Bosch-etch process using an O2 ICP plasma etchant and a C4F8 passivation material. Optimization entails varying the major operating parameters, ICP power, RF bias power, flow rate of reactants and passivation gas, while measuring resulting changes in vertical etch rate, 39 sidewall profile, and selectivity. The resulting protocol will be adopted for the use in Parylene micromachining for high aspect ratio MEMS implants. Materials and Methods 3.3.1 Experimental Design Calibration patterns consisted of rectangular lines and squares with features ranging from 15 to 150 µm in width, as shown in Figure 3-5. One half of the coupon corresponds to trenches (holes) and the other half to lines (relief) depending on the masking material used. A profilometer was used to measure depth of the features prior and post etching with the masking material, and without the masking material to characterize the etching performance. Four measurements were taken from the 50 to 150 µm trenches and lines at the different time points of process mentioned before, however, measurements from the 15 and 25 µm trenches and lines were not taken due to the thickness of the profilometer scanning tip, which was too large to fit within the trench/line. Figure 3-5. Schematics of (a) a patterned test coupon (2 cm × 2 cm); and (b) mask patterns and the corresponding dimensions. 3.3.2 Sample Preparation All experiments were performed on identical test samples (2 cm × 2 cm) consisting of Parylene substrates, supported by silicon dies, masked with photoresist. Preparation of the samples is depicted in Figure 3-6. 40 Figure 3-6. Overview of fabrication process for the DRIE coupon; (a) Silicon wafer treated with A-174; (b) Parylene deposition; (c) Photoresist was spun on and patterned; (d) DRIE plasma etching; (e) Etched patterns after DRIE process; (f) Parylene patterns after removal of the masking material; and (g) Schematic of a patterned test coupon. 4” silicon wafers were treated with A-174 silane adhesion promoter (Specialty Coating Systems, IN), then coated with 12 µm of Parylene C using a commercial vapor deposition system (PDS 2010 Labcoter, Specialty Coating Systems, IN). Photoresist etch masks (15 µm thick) were spin coated followed by a baking depending on the mask material (See Appendix C for complete recipes). Two different photoresists were used as mask materials: AZ P4620 (AZ Electronic Materials, Branchburg, NJ) and KMPR 1025 (MicroChem, Westborough, MA) (Figure 3-7). The former is the standard photoresist used in microfabrication to achieve thickness of 15 µm in a single spin or up to 30 µm in a double spin. However, a material with high aspect ratio is needed to create high aspect ratio structures. Thus, KMPR 1025 was selected as masking material because thickness of 120 µm can be achieved in a single spin. Figure 3-7. (a) KMPR 1025; and (b) AZ P4620 mask patterns after lithography. 41 Wafers were diced into individual sample coupons. In total, 110 test coupons were processed (40 dies for KMPR 1025 and 70 dies for AZ P4620, 1 die per process parameter). In order to accommodate the handling mechanism of the Oxford etcher system, coupons were mounted to the center of a 4’’ silicon carrier wafer, initially using a thin layer of photoresist (AZ 4400; AZ Electronic Materials, Branchburg, NJ), and later with double sided polyimide tape to avoid reflow of the etch mask. Following DRIE, the coupon was removed from the carrier wafer by dissolving the photoresist with acetone or by carefully pushing the coupon with tweezers to detach it from the tape. The etch masks were removed with acetone and warm (50 to 70 °C) Remover PG for AZ P4620 and KMPR 1025, respectively, and finally the coupons were cleaned with isopropanol and water prior to characterization. 3.3.3 Sample Preparation for Microscopy Replica molding (Figure 3-8), a non-destructive technique, was used to analyze the sidewall angles of the cross-section of the coupons under scanning electron microscopy (SEM) after etching. Molds prepared from silicone rubber (Sylgard 184, Dow Corning, Midland, MI) with a 10:1 base-to-curing-agent ratio. Coupons were placed in aluminum foil containers (2.5 cm × 2.5 cm) and coated with silicone. The uncured silicone was degassed for 2 hours under vacuum. The silicone rubber was cured at 50 °C for 4 hours, and the molds peeled off the coupons. The molds were cut into small sections and sputtered-coated with platinum for SEM imaging. See appendix D for DRIE profiles [12]. Figure 3-8. Overview of the replica mold process (a) etched Parylene trenches on silicon; (b) pouring and curing of silicone rubber; (c) peeling off the silicone rubber mold; (d) replica mold; and (e) SEM images of a replica mold, the circle indicates the replicated trench. 42 3.3.4 Etching System and Recipe An Oxford Plasmalab 100 ICP etcher system (Figure 3-9) was used to optimize the DRIE process for Parylene. The system consists of a load lock unit, process chamber, RIE lower electrode, RF generator, ICP source, vacuum system and gas system. Figure 3-9. Schematic of the Oxford Plasmalab 100 ICP tool. Wafers are clamped mechanically by a clamp plate to the lower electrode. Helium flows through the 3-pin wafer support to have a good thermal conductance between the sample and the lower electrode to keep the temperature controlled. Process gases are feed in through the top, and plasma is inductively coupled at 13.56 MHz, which is the frequency band selected for Industrial, Scientific and Medical (ISM) applications. Once the process is done, the system automatically unloads the wafer from the process chamber and transfers it to the load lock. The current DRIE recipe is a modified version of the switched-chemistry Bosch process presented in [13], where the passivation and etching mechanisms are separated in time and continuously switched. The passivation step consists of the deposition of a Teflon-like layer, C4F8, on the sidewalls and at the bottom of the features to be etched for 3 seconds. At the beginning of the next etch step, the passivation layer is removed from the horizontal surfaces of the substrate by ion bombardment. Ar was added to improve sidewall angles. The process parameters involve in the passivation step are shown in Table 3-1. The passivation step strongly depends on the temperature of the substrate, which is kept constant by flowing helium at the back side of the 43 wafer. By increasing the RF bias power the temperature increases, which reduces the passivation layer rate resulting in a decrease in the selectivity of the masking material. Table 3-1. Passivation step of current DRIE recipe Parameter Passivation ICP power (W) 700 RF power (W) 10 C4F8 (sccm) 35 Ar (sccm) 40 Pressure (mTorr) 23 Time (sec) 3 In the etch step, the passivation layer is removed from the bottom of the etched features by ion bombardment and Parylene is etched by O2 plasma for 10 seconds. Table 3-2 shows the process parameters involve in the etch step as well as the vertical etch rate. The ICP power is responsible for the plasma density resulting in a large number of reactive species in the chamber that are accelerated towards the wafer, thus increasing the ICP power must have a strong effect on the etching. On the other hand, the RF bias power is used to accelerate ions through the wafer by determining the ion energy. An increase of RF bias power increases the vertical etch rate due to the improvement of the directionality of the plasma reactive species. This will increase the removing of the passivation layer at the bottom of the etched features. Argon was introduced to the etch step as an additional reactive species. Table 3-2. Etch step of current DRIE recipe Parameter Etch ICP power (W) 700 RF power (W) 20 O2 (sccm) 60 Ar (sccm) 40 Pressure (mTorr) 23 Time (sec) 10 Etch rate (µm/min) 0.54 44 ICP and RF powers, reactant flow rates (Ar and SF6), oxygen flow rate, passivation flow rate and passivation time were varied. Table 3-3 summarizes the range of values of the process parameters tested to optimize the DRIE protocol according to Keck Photonics clean room and the specifications of system. The recipe was calibrated by optimizing each parameter at a time, on the basis of fastest etch rate and/or largest sidewall angle, and then modulating the next parameter, and so on for all variables described in Table 3-3. Table 3-3. Values of process parameters tested. Parameter Range ICP power (W) 500-1000 RF power (W) 10-300 O2 (sccm) 20-100 C4F8 (sccm) 0-100 Ar (sccm) 0-40 SF6 (sccm) 0-40 Passivation time (sec) 0-10 3.3.5 Etch Metrics 3.3.5.1 Vertical and Lateral Parylene Etch Rates Figure 3-10 shows a characteristic DRIE profile with important features for calculating etch rate and anisotropy labelled. Figure 3-10. Cross-section view of a Parylene etched profile by DRIE. 45 In order to calculate the vertical etch rate of Parylene, the height of the etched Parylene feature, hPxC, was measured using a profilometer and divided by the etching time, t. R vertical = h PxC t (2-1) The lateral etch rate of Parylene, Rlateral, was defined as R lateral = (a − b)/2 t (2-2) where a and b are the width of the top and bottom of the etched trench, respectively. Image J software was used to measure the widths from SEM images of replica mold cross sections. The degree of anisotropy, A, is defined by A = 1 − R lateral R vertical (2-3) A perfectly anisotropic profile is defined as A = 1, where the lateral etch rate is zero, and a perfectly isotropic profile is defined as A = 0, where Rlateral = Rvertical. 3.3.5.2 Mask Etch Rate and Selectivity The selectivity is a measure of how fast the masking material is removed compared to the vertical etch rate of Parylene. Figure 3-11 shows a diagram of the parameters measured to quantify selectivity. Figure 3-11. Cross-section view of a Parylene etched profile before DRIE, after DRIE and without masking material (from left to right). First, the etched masking material is calculated to know the etch rate of the mask. The etched masking material, hmask, was defined as 46 h mask = h PR − (h PR+PxC − h PxC ) (2-4) where hPR is the initial height of the masking material, hPR+PxC is the height of the etched masking material plus the etched Parylene, and hPxC is the height of the etched Parylene after removal of the mask. Thus, the etch rate of the masking material is given by R mask = h mask t (2-5) where the etched masking material, hmask, is divided by the etching time, t. The selectivity, S, between the masking material and Parylene was calculated as the ratio of the etch rate of the masking material and the vertical etch rate of Parylene and is given by S = R vertical R mask (2-6) In order to achieve good selectivity, the etch rate of Parylene should be equal or higher than the etch rate of the masking material. 3.3.5.3 Sidewall Angle The sidewall angle, θ, was obtained from SEM images of the cross-sections of the replica molds taken from etched features and analyzed using NIH ImageJ software (Figure 3-12). Figure 3-12. Diagram of the PDMS replica mold of an etched Parylene trench showing the sidewall angle measurement. 47 Results 3.4.1 Sample Preparation 3.4.1.1 Stripping of KMPR 1025 Resist Hot Remover PG was used to strip off the KMPR 1025 resist at 70 °C for ~5 min or until the resist was completely removed after DRIE process. However, the resist is transparent and it was difficult to observe when it was completely removed (Figure 3-13a), thus some Parylene bubbles were observed after a couple of minutes in the hot solution (Figure 3-13b). The temperature of the solution was lowered to 50 °C to reduce the occurrence of Parylene bubbles but the resist removal required more time. Unlike AZ P4620 which is dissolved away, KMPR 1025 resist lifts off (Figure 3-13b). For processes which require the formation of fluidic channels, it is possible that the resist can get trapped in the channels. Thus, as an alternative to lift off, we explored the use of Piranha solution to remove the photoresist in place of Remover PG, but resist was also lifted off. The temperature of a freshly prepared Piranha solution is ~90 °C due to an exothermic reaction. To avoid bubbles associated with high temperature, resist was removed using Piranha at 60 °C. Figure 3-13. (a) KMPR 1025 residues left on Parylene coupon; and (b) Parylene bubbles after soaking coupons in hot Remover PG. 3.4.1.2 Mounting Technique Pattern deformation, or reflow, was observed after hard baking the AZ P4620 resist (Figure 3-14a) and became more severe when mounting the coupon to the carrier wafer due to the baking of the mounting resist (Figure 3-14b). 48 Figure 3-14. (a) Pattern definition before mounting coupon to the carrier wafer; and (b) pattern reflow after mounting coupon using AZ 4400 resist. Another drawback of the mounting technique was the reflow of feature edges of the calibration patterns (Figure 3-15a). The effect was eliminated by increasing the temperature of the soft bake to remove all the solvent in the resist to eliminate the need for hard baking. In subsequent experiments, double-sided polyimide tape replaced the resist as the adhesive; this approach required no additional heat, and did not induce reflow (Figure 3-15b). Polyimide was specifically chosen as it survived the high heat generated by the etching tool. Reflow is desired in some cases to avoid concentration of stress at corners of microfluidic channels. But, in this case, it is necessary to have vertical resist sidewalls to characterize the DRIE process and to avoid mask erosion. Figure 3-15. Optical micrograph of effect on photolithographically produced pattern in AZ 4620 after mounting using (a) AZ 4400 resist; and (b) double sided tape. 49 3.4.2 Characterization of the Process Parameters 3.4.2.1 Effect of ICP Power The first process parameter to be investigated was the ICP power, which was changed in both the passivation and etch steps, only in the etch step and only in the passivation step. The number of DRIE cycles was fixed at 100 to etch ~9 µm of Parylene. The vertical etch rate of Parylene and the selectivity of the etch mask, KMPR 1025 resist, are presented in Figure 3-16 as a function of ICP power. The vertical etch rate varied from 0.5 to 0.7 µm/min and the selectivity decreased as the ICP power increased only for the etch and passivation steps. Changing the ICP power for both steps, etch and passivation, improved the selectivity and the vertical etch rate at lower powers (< 700 ICP W). Due to the problems encountered when stripping the KMPR 1025 resist, the lateral etch rate, degree of anisotropy, and sidewall angles were not calculated. Figure 3-16. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate; and (b) selectivity. ICP power process parameter was varied (500, 600, 700, 800, 900, and 1000 W). We also tested the ICP power condition with an alternative masking material, using the same ICP power range for passivation and etch steps but with AZ 4620 as the masking material (Figure 3-17). The DRIE cycles were reduced to 60 to etch ~5.5 µm of Parylene due to a thinner mask, ~11.5 µm. The vertical etch rate increased as the ICP power increased, whereas the selectivity was ‘stable’ throughout the tests, ~0.96. The lateral etch rate remained relatively 50 constant but increased at 1000 W. However, the degree of anisotropy improved between 700 and 900 W. Sidewall angles increased as the ICP power increased except for 1000 W. Figure 3-17. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. ICP power process parameter was varied (500, 600, 700, 800, 900, and 1000 W). The best vertical etch rate and sidewall angle (71.40°) results were obtained when the ICP power was 900 W for the passivation and etch steps. Table 3-4 shows the DRIE recipe for an optimum ICP power. 51 Table 3-4. DRIE recipe with ICP power optimized. Parameter Passivation Etch ICP power (W) 900 900 RF bias power (W) 10 20 O2 (sccm) 0 60 C4F8 (sccm) 35 0 Ar (sccm) 40 40 SF6 (sccm) 0 0 Time (sec) 3 10 Etch rate (µm/min) 0.77 Selectivity 0.97 Anisotropy 0.62 3.4.2.2 Effect of RF Bias Power An increase in RF bias power is expected to increase the vertical etch rate due to the improvement of the directionality of the plasma reactive species, as mentioned before, thus the etch cycles were decreased to 50 to etch ~6 µm of Parylene. The RF bias power was changed for the etch and passivation steps. Figure 3-18 shows the vertical etch rate and selectivity of the masking material, KMPR 1025. The vertical etch rate linearly increased for low RF powers (≤100 W), and then plateaued for high powers (> 100 W). The the selectivity of the masking material also increased for low powers. However, at high RF bias powers (> 100 W), the interaction between the plasma and photoresist likely induced crosslinking, making the exposed photoresist difficult to remove. Different approaches were tested in order to remove the crosslinked photoresist, which include piranha solution and hot Remover PG (50 °C). Thus, the lateral etch rate, degree of anisotropy and sidewall angles were not calculated. Due to the problems encountered with removal of KMPR 1025 following etching, AZ P4620 was used as masking material from here on. 52 Figure 3-18. Vertical etch rate of Parylene and selectivity of the masking material, KMPR 1025, when the RF bias power was varied. The RF bias power was varied for passivation and etch steps using AZ P4620 as the masking material (Figure 3-19). The DRIE cycles were reduced to 40 to etch ~5 µm of Parylene. Only low powers were tested to avoid issues presented with KMPR 1025 resist. The vertical etch rate increased as the RF bias power increased (more than double over the range tested), whereas the selectivity remained relatively constant and near ~1. The lateral etch rate also increased slightly with RF bias power which in turn lead to an increase in anisotropy. Likewise, the sidewall angle slightly increased as the RF bias power increased. 53 Figure 3-19. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Process RF bias power (10, 20, 40, 60, 80 and 100 W) was varied. The best vertical etch rate (1.40 µm/min) and sidewall angle (62.86°) were obtained when the RF power was 100 W, however, an RF power of 80 W was selected as the optimum value due to issues with photoresist removal at 100 W. Table 3-5 shows the DRIE recipe for the optimum RF bias power. 54 Table 3-5. DRIE recipe with RF bias power optimized. Parameter Passivation Etch ICP power (W) 900 900 RF bias power (W) 80 80 O2 (sccm) 0 60 C4F8 (sccm) 35 0 Ar (sccm) 40 40 SF6 (sccm) 0 0 Time (sec) 3 10 Etch rate (µm/min) 1.30 Selectivity 1.02 Anisotropy 0.45 3.4.2.3 Effect of Flow Rate of Gases The effects of Ar and SF6 flow rates, standard gases available in the ICP system, were investigated for the etch step alone (Figure 3-20) and for both etch and passivation steps (Figure 3-21). For each trace, Ar flow rate was kept constant while SF6 flow rate was increased. The etch cycles were set to 40 to etch ~9 µm of Parylene. Increasing the flow rate of SF6 resulted in a decrease in the vertical etch rate. On the other hand, the selectivity was ~1 but it decreased when Ar and SF6 flow were at 40 sccm. Increasing Ar and SF6 flow decreased the lateral etch rate, except when Ar flow was kept at 40 sccm where it did not exhibit a trend. The degree of anisotropy improved as SF6 flow rate increased when Ar flow rates were 0 and 10 sccm, and anisotropy decreased as SF6 increased when Ar was 40 sccm. 55 Figure 3-20. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Ar and SF6 flow rates (0, 10 and 40 sccm) were varied for the etch step. Ar flow rate was varied for passivation and etch steps while SF6 flow rate was varied for the etch step (Figure 3-21). The methodology from the previous experiment was followed where Ar flow rate was kept constant and SF6 flow rate was increased for each trace. The etch cycles were kept at 40 to etch ~9 µm of Parylene. Varying the Ar flow rate for both steps did not improve the vertical etch rate, however, the selectivity was improved to ~1 throughout all experiments. In 56 general, better lateral etch rates were obtained when Ar flow rate was varied for both steps. Thus, the degree of anisotropy was also improved as well as the sidewall angles. Figure 3-21. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. Ar, for passivation and etch steps, and SF6, for etch step, flow rates (0, 10 and 40 sccm) were varied. Table 3-6 shows three DRIE recipes that were tested considering the best sidewall angles obtained when the flow rate of Ar was varied for both steps. 57 Table 3-6. DRIE recipes with different Ar and SF6 flow rates Recipe 1 Recipe 2 Recipe 3 Parameter Passivation Etch Passivation Etch Passivation Etch ICP power (W) 900 900 900 900 900 900 RF bias power (W) 80 80 80 80 80 80 O2 (sccm) 0 60 0 60 0 60 C4F8 (sccm) 35 0 35 0 35 0 Ar (sccm) 0 0 40 40 40 40 SF6 (sccm) 0 10 0 10 0 40 Time (sec) 3 10 3 10 3 10 Etch rate (µm/min) 1.16 1.10 0.85 Selectivity 1.05 1.05 1.04 Anisotropy 0.61 0.52 0.60 3.4.2.4 Effect of Oxygen Flow Rate The flow rate of O2 was investigated for the etch step of the three DRIE recipes shown in Table 3-6. The etch cycles were set to 25, for recipe 1 and recipe 2, and to 35 for recipe 3 to etch ~5 µm of Parylene. As expected, the vertical etch rate increased as the flow rate of O2 increased, whereas the selectivity was ‘stable’, ~1 (Figure 3-22). The lateral etch rate and the degree of anisotropy are inconclusive, but better sidewall angles were obtained. 58 Figure 3-22. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity (unfilled); (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. O2 flow rate (0, 20, 35, 40, 60, 80 and 100 sccm) was varied for the etch step. The best sidewall angle was obtained when the flow rate of O2 was 60 sccm with no Ar (Table 3-6, Recipe 1). Thus, it was selected as the optimum value. Table 3-7 shows the DRIE recipe for the optimum O2 flow rate. 59 Table 3-7. DRIE recipe with O2 flow rate optimized. Parameter Passivation Etch ICP power (W) 900 900 RF bias power (W) 80 80 O2 (sccm) 0 60 C4F8 (sccm) 35 0 Ar (sccm) 40 40 SF6 (sccm) 0 0 Time (sec) 3 10 Etch rate (µm/min) 1.15 Selectivity 1.04 Anisotropy 0.71 3.4.2.5 Effect of the Passivation Layer The deposition time of C4F8 was investigated (Figure 3-23). The etch cycles were fixed to 40 to etch ~8 µm of Parylene. As expected, the vertical etch rate decreased as the deposition time increased, whereas the selectivity was ‘stable’ ~1 throughout the experiments. The lateral etch rate decreased as the passivation time increased except for 1 and 2 seconds. The degree of anisotropy is inconclusive, whereas the sidewall angles worsen compared to O2 flow rate tests. 60 Figure 3-23. Etch data for Parylene coupon processed by DRIE: (a) vertical etch rate and selectivity; (b) lateral etch rate; (c) degree of anisotropy; and (d) sidewall angle. C4F8 deposition time (0, 1, 2, 3, 4, 5, 7 and 10 sec) was varied for the passivation step. The effects of C4F8 flow rate were also investigated (Figure 3-24). The etch cycles were fixed to 30 to etch ~6 µm of Parylene and the passivation time was 3 seconds. The addition of C4F8 had no effect on the vertical etch rate and selectivity. However, the lateral etch rate increased as the C4F8 flow rate increased from 0 to 40 sccm and it decreased when the C4F8 flow rate was 80 and 100 sccm. Thus, the degree of anisotropy fluctuated as well but better sidewall angles were obtained compared to the results obtained when the passivation time was varied. 61 Figure 3-24. Etch data for Parylene sample processed by DRIE: (a) vertical etch rate and selectivity, (b) lateral etch rate, (c) anisotropy; and (d) sidewall angle. C4F8 flow rate (0, 20, 40, 60, 80 and 100 sccm) was varied for the dep step. 3.4.3 Optimized Recipe The calibrated recipe (Table 3-8) was chosen based on the best sidewall angle (73.4°) and vertical etch rate (1.15 µm/min). The outline and contact pads of an interdigitated electrode were etched out to demonstrate that vertical sidewalls can be produced with the calibrated recipe. A 4’’ Si wafer coated with 20 µm of Parylene was patterned lithographically with the features mentioned above. The etching was divided in two steps, the first to etch 10 µm of the outline and the second to etch the contact pads and the other 10 µm of the outline. The masking material, AZ P4620 resist, was ~15 µm thick. According to the vertical etch rate, 52 loops were needed to etch 10 µm of Parylene. 62 Table 3-8. Optimized DRIE recipe. Parameter Passivation Etch ICP power (W) 900 900 RF bias power (W) 80 80 O2 (sccm) 0 60 C4F8 (sccm) 35 0 Ar (sccm) 40 40 SF6 (sccm) 0 0 Time (sec) 3 10 Etch rate (µm/min) 1.15 Selectivity 1.04 Anisotropy 0.71 The vertical etch rate decreased to 0.94 µm/min due to a loading effect; the area of a silicon wafer is 12.57 in 2 , whereas the area of the coupon is only 0.62 in 2 . Loading effect refers to a dependence of the etch rate on the etched area; the greater the area to be etched, the slower the etch rate [14]. For wide features, the loading effect can result in convex profiles at the bottom of the structure, cause by the depletion of reactant species [15]. On the other hand, the selectivity was not affected by the loading effect, and was 1. SEM pictures of the etched features are shown in Figure 3-25. 63 Figure 3-25. SEM images of the (a) surface electrode, (b) contact pads, (c) locking mechanism, and (d) zoom in on the surface electrode. Images without masking material, only Parylene. Discussion The results of this optimization confirm that DRIE could be adopted to improve the vertical etch rate of Parylene and the profile of the etched features. However, some masking issues arose during the experiments involving the RF power. The masking materials, both KMPR 1025 and AZ P4620, could not be removed after Parylene coupons were processed by DRIE due to the cross-linking of the material when the RF power was 100 W or higher. Immersion in different solutions for long periods of time were used, such as, warmed Remover PG at 50 °C, warmed acetone at 60 °C, and Piranha; however, even these were unable to fully remove the photoresists. When assisting removal using an ultrasonic bath with warmed Remover PG, the Parylene was lifted of the silicon die with the photoresist. As expected, when the ICP power was increased, the vertical etch rate increased due to the availability of etchant species at the substrate. The same effect was observed when the RF bias power was increased; vertical etch rate was increased by improving the directionality. Although the process chamber pressure was adjustable, inconsistent results were obtained due to the malfunctioning of the tool and therefore not presented. Increasing SF6 flow rate tended to 64 improve the vertical etch rate and the sidewall angle. On the other hand, when Ar flow was increased, the sidewall angles worsened and the vertical etch rate decreased in some cases. These other gases are competing with O2, and that could cause a lower vertical etch rate. Increasing O2 flow rate, decreasing and increasing the passivation time and the C4F8 flow rate had no effect on the sidewall angle. Even though completely vertical sidewalls were not obtained, the vertical etch rate and profile of the etched feature were improved compared to the unoptimized recipe. The comparison of DRIE Parylene etch rates and recipes, found in literature, are shown in Table 3-9. It can be seen that the fastest reported etch rate is 1.7 µm/min in O2 plasma when the RF bias power is high (250 W) and the pressure is low (5 mTorr). In our case, the etch rate was improved by increasing the RF (80 W) and ICP (900 W) powers. Table 3-9. DRIE Parylene etch rates reported in literature Etch Rate (µm/min) Masking material Gas Recipe Ref. 0.77 O2 800 W ICP, 23 mTorr, 60 sccm [16] 1.70 Aluminum O2 250 W RF, 400 W ICP, 5 mTorr, 20 sccm [8] 0.60 AZ P4620 O2 O2+Ar 20 W RF, 800 W ICP, 23 mTorr, 20 sccm 20 W RF, 800 W ICP, 23 mTorr, 60 sccm O2, 50 sccm Ar [13] 0.54 NA O2 100 W RF, 300 W ICP, 4 mTorr, 60 sccm [17] 0.34 NA N2O 100 W RF, 300 W ICP, 4 mTorr, 60 sccm [18] 0.34 NA O2/CF4 (20%) 30 W RF, 300 W ICP, 4 mTorr, 60 sccm [19] 0.32 NA O2/CHF3 (18%) 30 W RF, 300 W ICP, 4 mTorr, 60 sccm [20] 0.87 BNP O2+CH4/C4F8+O2 40 W RF, 2800 W ICP, 20 mTorr, 200 sccm [21] 1.15 AZ P4620 O2+SF6 80 W RF, 900 W ICP, 23 mTorr, 60 sccm O2, 10 sccm SF6 This work Conclusions The optimization of DRIE process for the removal of Parylene was done. In this work, oxygen was used as the primary reactive species, and SF6 and Ar were used to improve the 65 sidewall angles/etch profile. Process parameters were investigated to obtain vertical sidewalls. Replica molding technique was used to create cross sections of the etch structures for SEM imagining, which were used to calculate vertical etch rates, lateral etch rates, degree of anisotropy and sidewall angles. These results provide a starting point for developing new recipes. The optimized etch recipe was used to demonstrate the fabrication of complex Parylene microstructures, such as, peripheral cuff electrodes. Selective Parylene removal will further facilitate its use as a structural material in MEMS and in other biomedical applications. References 1. System, S.C. Solvent resistance of the parylene. 2016. 2. Salcman, M.; Bak, M.J. A new chronic recording intracortical microelectrode. Medical and biological engineering 1976, 14, 42-50. 3. Loeb, G.E.; Bak, M.; Salcman, M.; Schmidt, E. Parylene as a chronically stable, reproducible microelectrode insulator. IEEE Transactions on Biomedical Engineering 1977, 121-128. 4. Schmidt, E.M.; Bak, M.J.; Christensen, P. Laser exposure of parylene-c insulated microelectrodes. Journal of neuroscience methods 1995, 62, 89-92. 5. Schmiedel, C.; Schmiedel, A.; Viöl, W. In Combined plasma laser removal of parylene coatings, ISPC Conference, 2009. 6. Yeh, J.; Grebe, K. Patterning of poly‐para‐xylylenes by reactive ion etching. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 1983, 1, 604-608. 7. Nojiri, K. Mechanism of dry etching. In Dry etching technology for semiconductors, Springer: 2015; pp 11- 30. 8. Selvarasah, S.; Chao, S.; Chen, C.-L.; Sridhar, S.; Busnaina, A.; Khademhosseini, A.; Dokmeci, M. A reusable high aspect ratio parylene-c shadow mask technology for diverse micropatterning applications. Sensors and Actuators A: Physical 2008, 145, 306-315. 9. Ayon, A.; Chen, K.-S.; Lohner, K.; Spearing, S.; Sawin, H.; Schmidt, M. In Deep reactive ion etching of silicon, MRS Proceedings, 1998; Cambridge Univ Press: p 51. 10. Lärmer, F.; Urban, A. Challenges, developments and applications of silicon deep reactive ion etching. Microelectronic Engineering 2003, 67, 349-355. 11. Ohara, J.; Takeuchi, Y.; Sato, K. Development of si drie process allowing simultaneous etching from narrow and wide mask openings. IEEJ Transactions on Sensors and Micromachines 2008, 128, 442-448. 12. Xia, Y.; McClelland, J.J.; Gupta, R.; Qin, D.; Zhao, X.M.; Sohn, L.L.; Celotta, R.J.; Whitesides, G.M. Replica molding using polymeric materials: A practical step toward nanomanufacturing. Advanced Materials 1997, 9, 147-149. 13. Meng, E.; Li, P.-Y.; Tai, Y.-C. Plasma removal of parylene c. Journal of Micromechanics and Microengineering 2008, 18, 045004. 14. Yeom, J.; Wu, Y.; Shannon, M.A. In Critical aspect ratio dependence in deep reactive ion etching of silicon, TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th International Conference on, 2003, 2003; IEEE: pp 1631-1634. 66 15. Mogab, C. The loading effect in plasma etching. Journal of the Electrochemical Society 1977, 124, 1262- 1268. 16. Meng, E.; Tai, Y.-C. In Parylene etching techniques for microfluidics and biomems, Micro Electro Mechanical Systems, 2005. MEMS 2005. 18th IEEE International Conference on, 2005; IEEE: pp 568-571. 17. Shutov, D.; Kim, S.-I.; Kwon, K.-H. On the etching mechanism of parylene-c in inductively coupled o 2 plasma. Transactions on Electrical and Electronic Materials 2008, 9, 156-162. 18. Shutov, D.; Kang, S.-Y.; Baek, K.-H.; Suh, K.-S.; Kwon, K.-H. Inductively-coupled nitrous-oxide plasma etching of parylene-c films. Journal of the Korean Physical Society 2009, 55, 1836-1840. 19. Ham, Y.-H.; Shutov, D.A.; Baek, K.-H.; Do, L.-M.; Kim, K.; Lee, C.-W.; Kwon, K.-H. Surface characteristics of parylene-c films in an inductively coupled o2/cf4 gas plasma. Thin Solid Films 2010, 518, 6378-6381. 20. Ham, Y.-H.; Shutov, D.A.; Kwon, K.-H. Surface characteristics of etched parylene-c films for low- damaged patterning process using inductively-coupled o2/chf3 gas plasma. Applied Surface Science 2013, 273, 287-292. 21. Lecomte, A.; Lecestre, A.; Bourrier, D.; Blatché, M.-C.; Jalabert, L.; Descamps, E.; Bergaud, C. Deep plasma etching of parylene c patterns for biomedical applications. Microelectronic Engineering 2017, 177, 70-73. 67 he adhesion of Parylene, particularly when exposed to wet, in vivo environments, is a critical determinant of device lifetime for polymer-based implants. This chapter explores several strategies for improving the adhesion of multi-layer Parylene structures, including thermal annealing and the use of several chemical interposer layers. Interfacial adhesion of Parylene- Parylene and Parylene-platinum-Parylene films were examined using a standard T-peel test to quantify adhesion and measure film integrity under chronic exposure to saline up to 2 years. Improved adhesion and barrier properties in Parylene-Parylene films resulted from the inclusion of diamond-like-carbon and ethylene-glycol-diacrylate layers. Thermal annealing improved Parylene film integrity in wet environments, but was insufficient for improving integrity of Parylene-platinum interfaces. A 100-fold increase in adhesive strength at such interfaces was achieved using a commercially available adhesion promoter, and corresponding improvements in resistance to moisture driven delamination were observed. ADHESION OF PARYLENE TO PARYLENE AND PLATINUM T 68 Background For several decades, thin Parylene coatings have been used as waterproof insulation for electronics intended for use in harsh environments, a category that now increasingly includes biomedical implants. Parylene has been used extensively to coat silicon, metal, and glass surfaces intended for extended in vivo use, including neural recording probes [1], coronary stents [2], implantable electronics [3,4] and dental implants [5]. In addition, Parylene is used as a substrate material from which to microfabricate biomedical implants because of its low Young’s modulus and flexibility. Despite well touted barrier properties, Parylene coatings and thin film Parylene devices exposed to wet environments for extended durations can suffer from moisture permeation and delamination; this limitation has motivated a significant body of work seeking to improve the lifetime and reliability of these coatings by increasing the strength of Parylene adhesion to surfaces of interest. Many strategies have been investigated to improve Parylene adhesion, including physical modifications (e.g. melting, anchoring, surface roughening), thermal modifications (e.g. annealing), chemical modification (e.g. surface plasma treatment) and inclusion of chemical interposer layers (e.g. silane A-174 and plasma polymerized adhesion layers) [6-8]. An exhaustive search of literature reported techniques, with mechanical tensile and peels strength serving as a proxy for coating adhesion, is summarized in Appendix E. Due to its widespread use in biomedical implants Parylene has emerged as a key material in the growing field of polymer-based biomedical microdevices comprising microelectronic, microfluidic and micromechanical systems wherein the bulk structure is composed of the thin- film, flexible polymer. Examples include polymer-based neural probes [9,10], cochlear implants [11,12], retinal electrodes [13,14], and pressure sensors [15,16]. Devices frequently feature a simple symmetric design: a base layer of flexible polymer, a thin layer of patterned metal (frequently platinum), and a top, insulating layer of polymer. This approach is motivated by evidence that reductions in size and mechanical rigidity of an implant can mitigate the physiological foreign body response, and enable chronic in vivo performance. Thin and flexible polymer-based devices offer an enticing alternative to rigid and sharp implants of silicon and metal, and Parylene, owing 69 to its extensive history in biomedical applications and compatibility with standard micromachining processes, is a prime choice among available materials [14,17-19]. As with examples of Parylene coated rigid structures, thin-film Parylene devices are also subject to delamination after extended exposure to wet, in vivo conditions. While a few reports describe Parylene-based thin-film devices functioning in vivo for longer than a year [9], these examples are uncommon. Numerous reports describe moisture intrusion and subsequent delamination of Parylene films over periods of weeks and months. These failure modes reflect poor adhesion of CVD Parylene to pre-deposited Parylene films, or at Parylene-metal-Parylene interfaces [20-22]. Adhesion between Parylene layers is dominated by physical adsorption and, to an extent, chemical reaction with free radicals, whereas adhesion at Parylene-metal interfaces is typically mediated by a combination of hydrogen bonding and Van der Waals forces [18,23,24]. In either case, the formation of chemical bonds through use of an intermediary linker is possible. For example, Driesche et al. have reported the use methacrylate-based coupling agents to support Parylene-gold adhesion [25]. Without the use of such an adhesion promoter bonds are easily broken by water molecules, creating a serious problem under in vivo conditions where the interfaces encounter bodily fluids [23]. Delamination under wet conditions may be induced by water vapor condensing within voids created by surface particulates present during CVD [4]. Stringent attention to surface cleanliness may reduce the risk of this phenomenon, but even when deposited under cleanroom conditions, Parylene coatings still exhibit voids in which water vapor can nucleate and condense. In addition, there are reports that coating failure can be hastened by application of electric current across insulated metal features [20,22], the action of compounds generated during inflammatory responses [26], and the presence of internal stress between layers [27]. Parylene structure and adhesion can likewise be effected by the chemical and physical demands of typical sterilization processes, required for clinical use in vivo [28]. While there are many proposed methods for improving adhesion between Parylene and non-polymeric substrates, there are relatively few reports describing methods to improve Parylene-Parylene adhesion [29-37] or Parylene-metal-Parylene adhesion [8,31,32,35-43] in thin- film devices, and robust chronic adhesion remains elusive. Table 4-1 shows the techniques 70 reported in the literature for improving Parylene adhesion to different materials under ‘wet’ conditions, referring to chronic exposure to water and/or saline. Table 4-1. Comparison of wet adhesion forces reported for Parylene interfaces. GDMP = glow discharge polymerized methane PC = Parylene PBS = phosphate buffer solution PN = Parylene N TMS = trimethylsilane X = CH2NH2 or CHO, aldehyde and aminomethyl side group By far the most common approach is the use of thermal annealing [10,13,27,44-48]. The application of temperature and pressure during annealing can induce polymer-polymer bonding through entanglement of polymer chains [29,30], and the high temperature can alter Parylene crystal structure, limiting the rate of moisture diffusion [27,49]. Less common are reports of chemical linkers or adhesion promoting layers appropriate for Parylene multilayer structures. While silane A-174 (3-trimethoxysilylpropyl methacrylate) is commonly used to improve Parylene adhesion to glass and silicon, the adhesion mechanism relies on the presence of Test Interface Adhesion Treatment Force Environment Ref. Tensile Pt/PN None GDMP ~0.15 Pa ~3.3 Pa 3 hrs in boiling saline 2 hrs in boiling saline [42] Si/PC&PC/Si 230 °C, 30 min, 800 N > 2.98 MPa > 3.45 MPa > 1.63 MPa > 1.15 MPa > 2.51 MPa 1 week in acetone 1 week in IPA 1 week in BHF 1 week in AZ400k 1 week in MF319 [29] Peel 90° Cr/PC None 140 °C, 3 h X X, 140 °C, 3h 0 0 20 mN/mm 150 mN/mm 30 min in PBS at 37 °C [32] Au/PC Ti DLC Ti/SiOx A-174 TMS 35.3 mN/mm 6.3 mN/mm 5.1 mN/mm 157.5 mN/mm Delaminated 1 hr in boiling saline [39] Peel 180° PC/PC None 140 °C, 3 h X X, 140 °C, 3h 10 mN/mm > 100 mN/mm 2 mN/mm 3 mN/mm 30 min in PBS at 37 °C [32] PC/PC None 210 mN/mm 24 hrs in PBS at 37 °C [37] 71 hydroxyl groups absent from Parylene surfaces, and as such the compound does not promote strong adhesion between Parylene layers [31]. Also, A-174 cannot be used when sacrificial photoresist layers are present [50]. Other chemical adhesion layers have been presented to aid Parylene coating of metals such as gold through a thiol-linker group [25]. Currently, there is no commercially available chemical linker demonstrated to improve Parylene-Parylene bonding. An alternative approach entails the use of non-linking interposer layers; under this approach a thin- film is deposited prior to CVD Parylene, which improves adhesion and/or operational lifetime by modifying surface energy, providing improved barrier properties (see work with Al2O3 [37]), or supporting adhesion to an intermediate layer (e.g. thin-film metal). Plasma-polymerized films have been used for years to improve the properties of Parylene coated bulk metal structures [8,23,51], though once again there are scant references to this approach for thin-film polymer- based devices. There is no commercially available compound, or state-of-the-art protocol, established that adequately improves Parylene-Parylene thin-film adhesion, beyond the broad recommendation to thermally anneal Parylene films. Several ideas presented or proposed in the literature, including plasma treatments and interposer layers [8,31,36-39], have not been rigorously compared and most have not been tested under chronic wet conditions. This represents a critical deficit in current research, as thin-film polymer-based microdevices are increasingly proposed for biomedical implantation because they are soft and flexible, yet moisture intrusion and delamination remain frequently reported failure modes. Approach and goals The purpose of this study consists of two major objectives. (1) Investigate and characterize strategies to improve the adhesion of Parylene-Parylene films and (2) to extend the operational lifetime of thin-film Parylene devices under chronic exposure to moisture, specifically in a simulated physiological saline environment. The purpose is to advance realization of a reliable, generalizable method. This work has specific application to the creation of polymer-based microdevices intended for chronic in vivo implantation. We compared a 72 combination of thermal annealing strategies and the use of several interposer films. Samples of Parylene-Parylene bilayers were prepared with different adhesion strategies; the adhesive strength of those interfaces were measured quantitatively under ‘dry’ conditions, and as a function of soaking duration in warm saline (simulated in vivo conditions) for up to two years. We measured adhesive strength with 180° T-peel test, and recorded the number of days of exposure in a ’wet’ environment before each strategy failed. The efficiency of thermal annealing was examined as function of annealing time, and we report x-ray diffraction data as an examination of the corresponding morphological change in the Parylene films. Three chemical interposer layers were selected: amorphous diamond-like carbon (DLC), ethylene glycol diacrylate (EGDA), and a proprietary commercial compound sold under the brand name AdPro Plus®. DLC and EGDA have been previously reported to improve the barrier properties of flexible polymer films, though there is no quantitative report of their effect on Parylene-Parylene adhesion. These materials were included as they hypothetically offer improved adhesion and/or moisture barrier properties [52-55]. AdPro Plus® is a Parylene-metal adhesion promoter designed to improve adhesion of noble metals to Parylene C and is the only promoter designed specifically for this function to our knowledge. We compared the adhesive strength and operational lifetime of Parylene-Parylene films with interposer layers of DLC and EGDA, and interposer layers of platinum supported by AdPro Plus®. Platinum was chosen owing to its widespread use in polymer-based biomedical microdevices, a result of its high biocompatibility and corrosion resistance. For purposes of control we also examined the adhesion strength and operational lifetime of Parylene-platinum-Parylene films without AdPro Plus®, and with and without thermal annealing treatments. Devices were prepared with large area coverage of platinum films, an unfavorable scenario compared with functional thin-film devices, which feature very limited exposed metal. While the formulation of AdPro Plus® is proprietary and not disclosed, we provide an empirical examination of its performance compared with experimental controls: thermal annealing treatments common in current research, and EGDA and DLC interposer layers, strategies motivated by recent academic research. This work represents the most rigorous examination of strategies to improve adhesion in multi-layer Parylene-Parylene devices. 73 Sample preparation and experimental methods This study examined wet and dry adhesion in Parylene-Parylene and Parylene-platinum- Parylene film systems. Table 4-2 lists the seven material combinations prepared and tested along with corresponding abbreviations used in this article. Table 4-2. Material combinations. Interface Abbreviation Parylene-Parylene PP Parylene-platinum-Parylene PMP Parylene-AdPro Plus®-platinum-Parylene PAdMP Parylene-platinum-AdPro Plus®-Parylene PMAdP Parylene- AdPro Plus®-platinum-AdPro Plus®-Parylene PAdMAdP Parylene-ethylene glycol diacrylate-Parylene PEGDAP Parylene-diamond-like carbon-Parylene PDLCP 4.3.1 Sample fabrication Test structures for T-peel testing were fabricated using common polymer micromachining techniques. Structures consisted of a well-defined bonded area (4.3 × 3 mm) connected to two Parylene flaps (12 μm thick) each perforated with a 3.3 mm clamping hole (Figure 4-1). Figure 4-1. (a) Sample schematic of T-peel test structure; (b) test structure in T-peel pulling apparatus; and (c) top and (d) lateral views of a sample under testing. 74 Figure 4-2 illustrates the layer-by-layer process by which samples were prepared. First, Si carrier wafers were dehydrated at 110 C then coated in a base layer of Parylene (12 µm) by CVD (PDS 2010, Specialty Coating Systems (SCS), Indianapolis, IN) (Figure 4-2a). PMP samples were prepared by sputter depositing platinum (2000 Å, LGA Thin Films, Santa Clara, CA) through a photoresist mask (2 μm, AZ 5214 E-IR; Integrated Micro Materials, Argyle, TX) followed by lift- off to define the bonded area. A 4 μm thick film of photoresist (AZ 4400; AZ Electronic Materials, Branchburg, NJ) was patterned lithographically to create a sacrificial spacer (Figure 4-2b), prior to depositing the top Parylene layer (12 µm) by CVD (Figure 4-2c); the sacrificial layer assisted separation of the two Parylene layers containing the clamping regions. The sample outline and clamping holes were etched using O2 reactive ion etching (100 W:100 mTorr:5 min cycles; RIE-80 Plasma Etching System, Oxford Plasma Technology, UK) through a thick photoresist mask (30 μm, AZ 4620; AZ Electronic Materials, Branchburg, NJ) (Figure 4-2d). The sacrificial photoresist was removed with acetone, and then the released samples were rinsed with isopropanol and water. Figure 4-2. Cross-sectional view of the fabrication process for PP (right) and PMP (left) T-peel tests samples. (a) Deposition of substrate layer; (b) sacrificial layer; (c) deposition of insulation layer; and (d) etching. 75 4.3.2 Adhesion layers 4.3.2.1 Ethylene glycol diacrylate Ethylene glycol diacrylate (EGDA) is a cross-linked anchoring layer that can be grafted to a Parylene surface by initiated chemical vapor deposition (iCVD). This technique allows for thin, uniform films to be deposited conformally in a single step and without use of solvents [56]. EGDA was selected for investigation as a Parylene-Parylene adhesion layer, owing to recent reports demonstrating EDGA films as durable and moisture resistant when grafted to Parylene, despite a 30-day soak in 1× phosphate buffered saline (PBS) [52]. PEGDAP samples were prepared using an iCVD process described in detail elsewhere [52,56,57] in which a benzophenone photoinitiator is used to attach a thin (10-20 nm) EGDA film to the base Parylene layer. Successful EGDA deposition was confirmed by a decrease in water-droplet contact angle, indicating a more hydrophilic surface (55 ± 2 , compared to 90 typical for native Parylene) [52]. Samples were then processed to completion as described in section 1.3.1, with care taken to avoid any cleaning or surface treatment after EGDA deposition. 4.3.2.2 Diamond-like carbon Diamond-like carbon (DLC) is a conformal, hard, chemically inert, and low friction coating consisting of an amorphous form of carbon with diamond bonds [53,58]. Because of its potential as a barrier layer, this film was also evaluated as a potential adhesion promoter in Parylene-Parylene films. The 0.23 μm thick film was deposited by Morgan Advanced Materials using ion-beam plasma-enhanced CVD with a hydrogen concentration ranging from 30% to 40%. After receiving the coated wafers, the DLC film was cleaned by O2 plasma, and samples were then processed to completion as described in section 4.3.1. 4.3.2.3 AdPro Plus® AdPro Plus® is a biocompatible, proprietary adhesion promoter available from Specialty Coating Services, designed to improve adhesion between Parylene and metals, such as, platinum, titanium, gold, etc. [21]. Wafers were sent to SCS for treatment with AdPro Plus® and deposition of the bottom and/or top Parylene layers depending on the interface, then returned for final 76 processing. Samples were prepared with AdPro Plus® at each Parylene-platinum interface, and samples were prepared with the adhesion promoter present at both interfaces. No cleaning or surface treatment was performed after the promoter deposition to avoid removal or damage of the AdPro Plus® layer. 4.3.3 Thermal annealing A subset of test structures was thermally annealed in an attempt to improve adhesion between layers and reduce moisture permeation, by way of an induced change in crystallinity and pore size. Several annealing durations were evaluated to determine the effects of annealing on the adhesion of PP and PMP samples. As reported by Charmet et al. [7], heating Parylene above its melting point results in an amorphous morphology which is susceptible to moisture intrusion, thus samples were annealed at 200 ºC for 24, 48 and 72 hours under vacuum, then cooled to room temperature overnight. The use of a vacuum was necessary to prevent oxidation of Parylene, which occurs at temperatures higher than 125 °C in the presence of oxygen [59]. During the annealing process, flaps of the T-peel structures were separated using a Teflon film to avoid thermal bonding of the layers. 4.3.4 T-peel tests Adhesive strength was measured using a T-peel test based on ASTM standard D1876-08 [60]. Test structures were peeled apart at 180° using a custom motorized stage while force was measured using a 50 g load cell (Omega, Stamford, CT, USA). Samples were clamped onto the stage with one end fixed to a stationary post and the other to a movable post driven by a stepper motor. The motorized stage was driven at 2 mm/sec to slowly peel apart the interface at the bonded region (Figure 4-1). A characteristic T-peel measurement is shown in Figure 4-3. The force registered by the load cell increases as the bonded area begins to peel (first peak), then stabilizes to a relatively constant force during peeling, and drops to zero as the interface is fully separated. 77 Figure 4-3. Representative raw T-peel test data. The peeling strength (mN/mm) is defined as the mean force recorded on the load-cell, averaged over the period between local maximums which denote the start and end of the peeling (marked by dashed red-lines in Figure 4-3), divided by the width of the bonded area. Four samples from each experimental group were tested without exposure to solution (exempting PDLCP, for which only two samples were measured), and four samples were tested at each time point in the soaking study (as described in section 4.3.5). 4.3.5 Long-term soaking study Approximately 60 samples of every combination of experimental parameters (14 experimental groups in total) were immersed in 1× concentration PBS at 37 °C to mimic physiological conditions (Figure 4-4). The degradation of the bond was measured by T-peel test at time points 1, 4 and 7 days during the first week. Then, samples were tested weekly for a month, and then monthly for 2 months. After completion of a 3-month soak, samples were tested every 3 months until the samples delaminated or until 1 year had elapsed. Then, samples were tested every 6 months until delamination or until 2 years had elapsed. 78 Figure 4-4. Immersed samples in 1× PBS. 4.3.6 XPS After T-peel tests Parylene interfaces were cleaned and rinsed with isopropanol and deionized water, then analyzed using X-ray photoelectron spectroscopy (XPS), to determine fractional atomic composition. The XPS measurements were performed with a Kratos Axis Ultra DLD instrument (Kratos Analytical, UK) with a monochromatic Al Kα x-ray source, and probed the top 5 nm of each surface. 4.3.7 XRD The crystallinity of Parylene films, prior and post annealing, was characterized by X-ray diffraction (XRD), using an Ultima IV Powder Diffractometer (Rigaku, USA). Glass slides were coated with 12 µm of Parylene and were cleaned with isopropanol and deionized water prior to measurements. The scans were measured from 10° to 18° (incident angle) in order to measure the Parylene diffraction peaks, which are known to appear around 14° [61]. Results 4.4.1 Crystallinity/XRD analysis The XRD spectra of un-annealed and annealed Parylene films for different anneal times is shown in Figure 4-5. The diffraction peaks appeared at 2θ ≈ 14° for all tested samples. 79 Figure 4-5. XRD 2θ scans of un-annealed and annealed Parylene films. The crystallite size of the Parylene films was calculated using Scherrer’s equation, which is given by: Crystallite size= 0.9𝜆 FWHM cos 𝜃 (1) Where 𝜆 is the wavelength of Cu Kα X-ray source, FWHM (full width half maximum) is derived from the measured peak, and θ is the Bragg angle (degree of the diffraction peak). The FWHM values were calculated by fitting a Gaussian function to the peaks using OriginPro software (Northampton, MA) and are shown in Table 4-3. Table 4-3. Properties of Parylene (n=1) at different anneal times. Anneal Time (hours) FWHM (°) Crystallite Size (nm) 0 2.07±0.02 3.94±0.08 24 0.74±0.02 11.11±0.60 48 0.75±0.02 10.88±0.58 72 0.74±0.02 11.08±0.60 Error bars reflect instrument precision. 80 The FWHM decreased as anneal time increased, whereas the crystallite size of un- annealed Parylene increased after a 24-hour anneal and then was unchanged for 48 and 72 hour anneal. The percentage of crystallinity is related to the intensity of the Bragg peak as the peak intensity increases the crystallite concentration in the polymer increases as well [61]. The percentage is calculated as the ratio of the area of the crystalline peak to the whole area, amorphous plus crystallized area [62,63]. Thus, the film that was annealed for 48 hours has the greatest percentage of crystallinity. 4.4.2 T-peel tests Thermal annealing resulted in dramatic increases in the peeling force for the majority of the experimental groups, and increased the stiffness of the Parylene layers. As such, while most samples peeled apart at the bonded interface during T-peel testing (Figure 4-6a), we also observed two modes of failure: tearing at the interface (Figure 4-6b) and at one of the clamping holes (Figure 4-6c). This was observed for several PP, PMP and PEGDAP samples, and therefore the reported values must be strictly considered a lower bound of peeling strength. Figure 4-6. Failure modes presented at day 0 during testing for (a) typical PMAdP peeled apart sample with no tearing; (b) PMAdP sample torn at the interface; and (c) 48h-annealed PAdMAdP sample torn at the clamping hole. 4.4.2.1 Parylene-Parylene samples The adhesive strength of PP samples as a function of annealing time is presented in Figure 4-7a for samples annealed for a duration of 24, 48, and 72 hours at 200° C. The mean force required to peel apart PP samples increased from ~38 to ~74 mN/mm following a 24-hour anneal, increased further to ~155 mN/mm following a 48 hour anneal, and diminished for a 72 hour anneal. These samples were not subject to long-term soaking and thus these experiments are referred to as dry testing. 81 Figure 4-7b shows how adhesive strength decreased as a function of total time immersed in saline. Un-annealed samples suffered a catastrophic loss in adhesion following just a single day of soaking, and delaminated completely after 4 weeks, whereas annealed samples exhibited greater moisture resistance and longer lifetimes. The 48-hour annealed samples notably maintained minimal to no loss in adhesion over 2 years. Owing to the results of these experiments, a 48-hour anneal was used in all subsequent testing where annealed samples were prepared. Figure 4-7. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene layers (mean ± SE, n=4): (a) peeling strength as a function of thermal annealing time for dry samples; and (b) peeling strength as a function of time soaked in warm saline. 4.4.2.2 Parylene-platinum-Parylene samples The inclusion of metal (specifically platinum) between Parylene layers resulted in a dramatic decrease in adhesion; un-annealed PMP samples were peeled apart with just ~3 mN/mm prior to soaking (Figure 4-8a), while soaked samples delaminated after just 4 days in saline (Figure 4-8b). Annealing significantly increased adhesion (to a peeling force of ~22 mN/mm), however, even annealed PMP samples exhibited weaker adhesion than un-annealed Parylene- Parylene samples. Annealed PMP samples exhibited gradual adhesion loss during soaking and delaminated after 3-weeks. 82 Figure 4-8. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene (PP) and Parylene-platinum-Parylene (PMP) layers (mean ± SE, n=4): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. Figure 4-9 shows the different mechanisms of adhesion failures for annealed (Figure 4-9a) and un-annealed (Figure 4-9b) PMP samples, after 3 weeks and 4 days, respectively. In annealed samples, the metal film delaminated completely from both base and top layers of Parylene. Wrinkling of the metal film demonstrated that significant stress was introduced by the annealing process. In un-annealed samples, the top (second) layer of Parylene consistently separated from the metal film, while the adhesion between the metal film and base (first) layer of Parylene remained intact. Figure 4-9. Delaminated (a) 48h-annealed; and (b) un-annealed PMP samples after a 3-week and 4-day soak in PBS, respectively. 4.4.2.3 Parylene samples with adhesion layers Un-annealed samples with EGDA and DLC adhesion layers exhibited improved adhesion, compared to un-annealed Parylene-Parylene samples. The mean peeling force for un-annealed 83 PEGDAP and PDLCP samples was ~99 and ~58 mN/mm respectively (Figure 4-10a). Annealing improved the adhesion of PDLCP samples marginally but not for PEGDAP samples; average peeling force for annealed PDLCP samples increased to ~68 mN/mm, while the peeling force decreased to ~83 mN/mm for annealed PEGDAP samples (Figure 4-10a). Figure 4-10. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-Parylene (PP), Parylene-Ethylene glycol diacrylate-Parylene (PEGDAP), and Parylene-Diamond-like carbon-Parylene (PDLCP) layers (mean ± SE, n=4): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. Un-annealed PEGDAP samples exhibited minimal to no loss in adhesion over a period exceeding 54 weeks soaking in PBS, whereas annealed PEGDAP samples exhibited a loss in adhesion strength after 12 weeks and completely delaminated after a year in PBS. Un-annealed PEGDAP samples still retained integrity up to at least 82 weeks of simulated in vivo environment. The peeling force for un-annealed and annealed PDLCP samples dropped in the first 4 days and then it stabilized after a 4-week soak in PBS up to a month and year for un-annealed and annealed samples, respectively. Following an 8-weeks soak in PBS, un-annealed PDLCP samples presented a loss of adhesion and were completely delaminated after a year. Following T-peel tests, we observed that delamination consistently occurred at the interface between the top (second) Parylene layer and DLC layer for un-annealed and annealed PDLCP samples. PDLCP samples were curled after release due to the stress of the film. 84 4.4.2.4 Platinum samples with adhesion promoter Figure 4-11 displays the results of T-peel measurements on ‘dry’ samples of Parylene- platinum-Parylene films with AdPro Plus® deposited below, above, and on both sides of the metal film. Samples with the adhesion layer deposited prior to sputtering the platinum film (PAdMP) exhibited very weak adhesion, similar to untreated PMP samples (~3 mN/mm). In contrast, samples with the adhesion layer deposited after sputtering the platinum film (PMAdP), and samples with the adhesion layer deposited both before and after the platinum film (PAdMAdP), exhibited excellent adhesion (mean T-peel measurements of 336 and 143 mN/mm respectively). The T-peel measurement of un-annealed PMAdP samples was the highest recorded among all sample combinations. Thermal annealing of these samples yielded unexpected results, adhesion strength of PAdMP samples increased to produce T-peel measurements of ~52 mN/mm, while adhesion strength of PMAdP and PAdMAdP decreased to 94 and 87 mN/mm respectively. Most samples peeled apart during testing and the top (second) Parylene layer detached from the metal, however, all un-annealed PMAdP and annealed PMAdP and PAdMAdP samples tore at the interface and clamping holes, respectively. Figure 4-11. Average force per unit length required to peel apart annealed and un-annealed samples of Parylene-platinum-Parylene layers with and without AdPro Plus® adhesion promoter (mean ± SE, n=4). Samples featured AdPro Plus® between the base Parylene and metal layer (PAdMP), between the metal and top Parylene layer (PMAdP) and on both sides of the metal layer (PAdMAdP): (a) peeling strength of dry samples; and (b) peeling strength as a function of time soaked in warm saline. 85 Results from soak-testing reflected these same trends, and are compiled in Table 4-4. Un- annealed PAdMP samples failed after just 4 days of immersion in saline (similar to untreated PMP samples), but, if annealed, withstood 3 weeks before suffering delamination. Un-annealed PMAdP samples exhibited excellent resistance to moisture, but failed after just 2 weeks immersion in saline if annealed. In annealed samples, the metal film delaminated completely from both base (first) and top (second) layers of Parylene while in un-annealed samples, the top (second) layer of Parylene separated from the metal film. Table 4-4. Time to interface failure: with and without adhesion promoter. Interface Soaking time (weeks) PMP Annealed PMP 4 days 2 PAdMP Annealed PAdMP 4 days 3 PMAdP Annealed PMAdP 24 2 PAdMAdP Annealed PAdMAdP 3 2 4.4.3 XPS analysis Table 4-5 shows the results of XPS, detailing the atomic composition at the failing surface of the platinum interfaces after peel tests. Samples are presented corresponding to different time points in the soaking study. The presence of oxygen in platinum containing samples is likely the result of the O2 plasma cleaning process used after metal deposition and prior to the deposition of AdPro Plus®. The most notable feature of this dataset is the presence and strength of the platinum signal. In no sample was platinum detected on the top (second) layer of Parylene following t-peel testing. Instead, platinum was detected only on the bottom (first) layer of Parylene, with the exception of the annealed PAdMP sample for which no platinum was detected; this is consistent with visual observation that the stressed platinum film delaminated entirely from both Parylene layers. This dataset was useful for examining the relative strength of platinum-Parylene bonds. Notably, the platinum film consistently and preferentially remained on the bottom layer. Beyond this, 86 variation between results proved too severe to draw precise conclusions about the surface composition. Table 4-5. Atomic composition of the metal combinations. Interface Layer %C %O %Cl %Pt Parylene reference 90.00 0.00 10.00 0.00 PAdM/P (dry) Top 89.21 1.30 9.49 0.00 Bottom 63.06 20.25 8.37 8.32 48h PAdM/P (dry) Top 89.95 1.05 9.00 0.00 Bottom 89.98 0.00 10.02 0.00 PM/AdP (1 day) Top 62.57 25.05 3.46 0.42 Bottom 56.33 26.45 0.00 17.22 48h PM/AdP (2 weeks) Top 60.22 27.63 2.36 0.00 Bottom 41.02 33.42 0.00 25.56 PAdM/AdP (dry) Top 90.09 0.00 9.91 0.00 Bottom 85.31 3.16 10.50 1.02 48h PAdM/AdP (2 weeks) Top 66.09 24.76 3.25 0.00 Bottom 72.03 18.44 0.00 9.54 Discussion Our results confirm both the relatively weak adhesion between untreated Parylene layers, and the susceptibility of Parylene coatings to adhesive failure and delamination under wet conditions. This phenomenon has been noted extensively, both in literature and anecdotally, yet Parylene remains an important insulator for biocompatible implants. For free-film devices, such as those tested here, the susceptibility is exacerbated as both sides of the film offer a conduit for moisture penetration. Experimental testing of annealed samples show that the thermal treatment dramatically improves both Parylene-Parylene adhesion and barrier properties. This is in agreement with prior reports, and the mechanism can be understood as thermally induced entanglement of the polymer chains and an increase in crystallinity which reduces moisture permeation. The use of a 48-hour anneal has been commonly reported, again both in the literature and anecdotally by other researchers, and the comparison of the 24-hour, 48-hour and 72-hour datasets shows this protocol is well supported by the data. While improvements in adhesion and barrier properties are evident 87 after 24 hours, continued annealing more than doubles the adhesive strength, while the lack of improvement at the 72-hour mark suggests whatever crystallization or morphology change is occurring is accomplished at the end of 48 hours. This is further supported by the XRD data (Figure 4-5); the sharpest diffraction peaks were observed for samples annealed for 48-hours, with less crystallinity observed for 24- and 72-hour annealed samples. We measured incredibly weak adhesion of Parylene to platinum films, and the thermal annealing method improved the adhesive strength and barrier properties only slightly. Our experiments examined only platinum, and specifically sputter-deposited platinum, so we are hesitant to generalize to all metals or even platinum films deposited by other methods, however, the likely cause of this weak adherence is the mismatch in surface energy between the hydrophobic Parylene layer and the hydrophilic platinum layer (in agreement with Hwang et al. [64]), and such a mechanism would likely apply to other metals as well. Despite improvements following annealing, the lifetime of annealed PMP samples requires further improvement to achieve long term Parylene-based implants. This result is of serious concern; the majority of both polymer bioMEMS and Parylene coated medical devices incorporate conductive metal layers, and platinum specifically is a common choice. The use of EGDA and DLC interposer layers increased the lifetime of Parylene-Parylene devices under soaking conditions. Annealing produced minimal improvement, and neither the inclusion of DLC or EGDA was comparable to annealed PP in adhesive strength or lifetime (Figure 4-10). We will note that there was a considerable period of time (6 months) between the DLC deposition and the final Parylene insulation, and this risked oxidation or contamination of the DLC surface. DLC was not tested with a platinum layer as this would require a titanium adhesion layer between interposer layer and platinum, which would introduce a confounding factor. EDGA was not tested with a platinum layer as the solvents required for metal-liftoff were expected to attack the adhesion layer. AdPro Plus® was chosen for testing because it is advertised exclusively for improving adhesion at Parylene-metal interfaces. Results of both soak and T-peel tests show a dramatic improvement in barrier properties and adhesion for Parylene-platinum-Parylene samples, but 88 also revealed a critical sensitivity to temperature that may complicate processing. Un-annealed PMAdP samples exhibited the strongest adhesive force of any sample tested in this study (Figure 4-11), and a greatly extended lifetime (> 24 weeks) under saline soaking conditions. However, annealed PMAdP samples incurred a significant decrease in both adhesion and lifetime, as did annealed PAdMAdP samples. As the only difference between the un-annealed and annealed samples was the heat treatment itself, we are confident that this loss in adhesion can be ascribed to the action of the temperature on the AdPro Plus®. Our hypothesis is that the adhesion promoter denatures under high temperature, and this is supported by the failure of the PAdMP samples, which delaminated almost immediately. The deposition of platinum and associated lithography can drive samples to temperatures above 100 °C, and it would appear this is sufficient to disable the adhesion promotor. These results are further supported by the XPS data; the bottom layer of the un-annealed PAdMP exhibits a small platinum signal that vanishes following annealing, suggesting the platinum film lost adhesion after heat treatment. Conversely, the bottom layer of the un-annealed PMAdP samples exhibit a strong platinum signal, that increases following annealing, suggesting the platinum film lost adhesion to the top layer after heat treatment. In all cases, we see that the platinum film is retained on the bottom layer of Parylene, suggesting stronger adhesion of metal deposited on Parylene than Parylene deposited on metal. We note that prior work by Charmet et al. [7] examined a combination of thermal treatment (350 °C) and the use of silane A-174 adhesion promotor at the interface between silicon and CVD Parylene. Their results indicated no decrease in adhesion or insulation integrity with the addition of the heat treatment. The heat sensitivity of AdPro Plus® appears to be intrinsic to this specific formulation, and is not a general result of combining thermal treatment with chemical adhesion layers. In almost all experiments adhesion failure occurred between the top (insulating) layer of Parylene and the bottom layer regardless of material (e.g. base layer Parylene, platinum, EGDA, DLC). This agrees with previous findings that, absent the formation of a chemical bond, Parylene typically adheres poorly as deposited, particularly to smooth surfaces or materials with significant differences in surface energy [65,66]. We also observed very strong adhesion between 89 deposited platinum and the bottom (base) layer of Parylene. This may be a result of the O2 plasma descum performed prior to sputtering, which can roughen surface morphology and induce hydrophilicity in Parylene. Finally, we also note that the samples tested featured large-area platinum films, relative to smaller patterned structures commonly used in devices. Therefore, the presence of PP interfaces between metal features, such as, traces and electrodes, may improve overall adhesion of the structure. Conclusions The adhesion of thin (~10 µm) CVD Parylene films to pre-deposited layers of Parylene and metal is insufficient for robust insulation intended for chronic exposure to wet environments. Thermal annealing of Parylene under vacuum significantly improves both the barrier properties of the bulk medium and the adhesion of Parylene-Parylene interfaces, enabling samples to survive for up to 2 years in saline, and the benefits of the annealing process are exhausted after 48-hours at 200 °C. Annealing, however, is insufficient for improving the adhesion at Parylene- platinum interfaces. Interposing materials, including diamond-like carbon and ethylene glycol diacrylate, can provide improvements in both the moisture resistance and adhesion at Parylene- Parylene interfaces without reliance on heat treatments. The inclusion of AdPro Plus® between platinum films and insulating Parylene layers improved adhesion and moisture resistance by an order of magnitude beyond that of either annealed or un-annealed Parylene-platinum-Parylene films, however the adhesion proved very sensitive to elevated temperatures, and may not be compatible with additional processing. Broadly, we note that despite its reported barrier properties, very thin Parylene layers are susceptible to moisture intrusion and subsequent insulation failure/delamination. Methods which reduce water vapor permeation, including the use of moisture resistant interposer layers and crystallization under thermal annealing, can significantly reduce this effect and dramatically extend the lifetime of Parylene-based or Parylene-insulated devices exposed chronically to water or saline. Reducing moisture intrusion alone is insufficient if there is significant mismatch in surface energy between Parylene and the target, which in this study included sputtered platinum. 90 Methods that roughen the Parylene surface, or induce hydrophilicity, such as O2 plasma exposure, appear to significantly improve adhesion. Improving dry and wet adhesions of Parylene to metal is critical for long-term stability of material interfaces in Parylene-based medical implants. References 1. Hsu, J.-M.; Kammer, S.; Jung, E.; Rieth, L.; Normann, R.; Solzbacher, F. In Characterization of parylene-c film as an encapsulation material for neural interface devices, Conference on Multi-Material Micro Manufacture, 2007. 2. Wolgemuth, L. A look at parylene coatings in drug-eluting technologies. Reprinted from Medical Device & Diagnostic Industry 2005. 3. Loeb, G.E.; Bak, M.; Salcman, M.; Schmidt, E. Parylene as a chronically stable, reproducible microelectrode insulator. IEEE Trans. Biomed. Eng 1977, 24, 121-128. 4. Schmidt, E.; McIntosh, J.; Bak, M. Long-term implants of parylene-c coated microelectrodes. Medical and Biological Engineering and Computing 1988, 26, 96-101. 5. Kuppusami, S.; Oskouei, R.H. Parylene coatings in medical devices and implants: A review. Univ. J. Biomed. Eng 2015, 3, 9-14. 6. Huang, R.; Tai, Y. In Parylene to silicon adhesion enhancement, Solid-State Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009. International, 2009; IEEE: pp 1027-1030. 7. Charmet, J.; Bitterli, J.; Sereda, O.; Liley, M.; Renaud, P.; Keppner, H. Optimizing parylene c adhesion for mems processes: Potassium hydroxide wet etching. Journal of Microelectromechanical Systems 2013, 22, 855-864. 8. Yamagishi, F.G. Investigations of plasma-polymerized films as primers for parylene-c coatings on neural prosthesis materials. Thin Solid Films 1991, 202, 39-50. 9. Hara, S.A.; Kim, B.J.; Kuo, J.T.; Lee, C.D.; Meng, E.; Pikov, V. Long-term stability of intracortical recordings using perforated and arrayed parylene sheath electrodes. Journal of neural engineering 2016, 13, 066020. 10. Xu, H.; Hirschberg, A.W.; Scholten, K.; Berger, T.W.; Song, D.; Meng, E. Acute in vivo testing of a conformal polymer microelectrode array for multi-region hippocampal recordings. Journal of neural engineering 2018, 15, 016017. 11. Johnson, A.C.; Wise, K.D. In A robust batch-fabricated high-density cochlear electrode array, Micro Electro Mechanical Systems (MEMS), 2010 IEEE 23rd International Conference on, 2010; IEEE: pp 1007-1010. 12. Johnson, A.C.; Wise, K.D. In A self-curling monolithically-backed active high-density cochlear electrode array, Micro Electro Mechanical Systems (MEMS), 2012 IEEE 25th International Conference on, 2012; IEEE: pp 914-917. 13. Rodger, D.C.; Fong, A.J.; Li, W.; Ameri, H.; Ahuja, A.K.; Gutierrez, C.; Lavrov, I.; Zhong, H.; Menon, P.R.; Meng, E. Flexible parylene-based multielectrode array technology for high-density neural stimulation and recording. Sensors and Actuators B: chemical 2008, 132, 449-460. 14. Rodger, D.C.; Li, W.; Ameri, H.; Ray, A.; Weiland, J.D.; Humayun, M.S.; Tai, Y.-C. Flexible parylene- based microelectrode technology for intraocular retinal prostheses. 2006. 15. Chen, P.-J.; Rodger, D.C.; Agrawal, R.; Saati, S.; Meng, E.; Varma, R.; Humayun, M.S.; Tai, Y.-C. Implantable micromechanical parylene-based pressure sensors for unpowered intraocular pressure sensing. Journal of Micromechanics and Microengineering 2007, 17, 1931. 91 16. Gutierrez, C.A.; Meng, E. In A subnanowatt microbubble pressure sensor based on electrochemical impedance transduction in a flexible all-parylene package, 2011 IEEE 24th International Conference on Micro Electro Mechanical Systems, 2011; IEEE: pp 549-552. 17. Senkevich, J.J.; Desu, S.B. Morphology of poly (chloro-p-xylylene) cvd thin films. Polymer 1999, 40, 5751-5759. 18. Kim, B.J.; Meng, E. Review of polymer mems micromachining. Journal of Micromechanics and Microengineering 2015, 26, 013001. 19. Noh, H.-S.; Huang, Y.; Hesketh, P.J. Parylene micromolding, a rapid and low-cost fabrication method for parylene microchannel. Sensors and Actuators B: Chemical 2004, 102, 78-85. 20. Li, W.; Rodger, D.C.; Meng, E.; Weiland, J.D.; Humayun, M.S.; Tai, Y.-C. In Flexible parylene packaged intraocular coil for retinal prostheses, Microtechnologies in Medicine and Biology, 2006 International Conference on, 2006; IEEE: pp 105-108. 21. Kumar, R. In Advances in adhesion solutions for medical applications, Proceedings of the SMTA Medical Electronics Symposium, 2008. 22. Li, W.; Rodger, D.; Menon, P.; Tai, Y.-C. Corrosion behavior of parylene-metal-parylene thin films in saline. Ecs Transactions 2008, 11, 1-6. 23. Yasuda, H. Adhesion of plasma polymerized films (a model study on water sensitivity of adhesion). In Adhesion aspects of polymeric coatings, Springer: 1983; pp 193-203. 24. Zhao, Y.-P.; Wang, L.; Yu, T. Mechanics of adhesion in mems—a review. Journal of Adhesion Science and Technology 2003, 17, 519-546. 25. Driesche, S.v.d.; Habben, C.; Bödecker, A.; Lang, W.; Vellekoop, M.J. In A simple method to allow parylene-c coatings on gold substrates, Multidisciplinary Digital Publishing Institute Proceedings, 2017; p 299. 26. Cieślik, M.; Engvall, K.; Pan, J.; Kotarba, A. Silane–parylene coating for improving corrosion resistance of stainless steel 316l implant material. Corrosion Science 2011, 53, 296-301. 27. Von Metzen, R.P.; Stieglitz, T. The effects of annealing on mechanical, chemical, and physical properties and structural stability of parylene c. Biomedical Microdevices 2013, 15, 727-735. 28. Beshchasna, N.; Adolphi, B.; Granovsky, S.; Braunschweig, M.; Schneider, W.; Uhlemann, J.; Wolter, K.-J. In Influence of artificial body fluids and medical sterilization procedures on chemical stability of parylene c, Electronic Components and Technology Conference (ECTC), 2010 Proceedings 60th, 2010; IEEE: pp 1846-1852. 29. Kim, H.; Najafi, K. Characterization of low-temperature wafer bonding using thin-film parylene. Journal of microelectromechanical systems 2005, 14, 1347-1355. 30. Shu, Q.; Huang, X.; Wang, Y.; Chen, J. In Wafer bonding with intermediate parylene layer, Solid-State and Integrated-Circuit Technology, 2008. ICSICT 2008. 9th International Conference on, 2008; IEEE: pp 2428-2431. 31. Hassler, C.; von Metzen, R.P.; Ruther, P.; Stieglitz, T. Characterization of parylene c as an encapsulation material for implanted neural prostheses. Journal of Biomedical Materials Research Part B: Applied Biomaterials: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials 2010, 93, 266-274. 32. Seymour, J.P.; Elkasabi, Y.M.; Chen, H.-Y.; Lahann, J.; Kipke, D.R. The insulation performance of reactive parylene films in implantable electronic devices. Biomaterials 2009, 30, 6158-6167. 33. Chang, J.H.-C.; Lu, B.; Tai, Y.-C. In Adhesion-enhancing surface treatments for parylene deposition, Solid- State Sensors, Actuators and Microsystems Conference (TRANSDUCERS), 2011 16th International, 2011; IEEE: pp 390-393. 34. Noh, H.-s.; Moon, K.-s.; Cannon, A.; Hesketh, P.J.; Wong, C. In Wafer bonding using microwave heating of parylene for mems packaging, Electronic Components and Technology Conference, 2004. Proceedings. 54th, 2004; IEEE: pp 924-930. 92 35. Mueller, M.; Ulloa, M.; Schuettler, M.; Stieglitz, T. In Development of a single-sided parylene c based intrafascicular multichannel electrode for peripheral nerves, Neural Engineering (NER), 2015 7th International IEEE/EMBS Conference on, 2015; IEEE: pp 537-540. 36. Xie, Y.; Pei, W.; Guo, D.; Zhang, L.; Zhang, H.; Guo, X.; Xing, X.; Yang, X.; Wang, F.; Gui, Q. Improving adhesion strength between layers of an implantable parylene-c electrode. Sensors and Actuators A: Physical 2017, 260, 117-123. 37. Lee, C.D.; Meng, E. Mechanical properties of thin-film parylene–metal–parylene devices. Frontiers in Mechanical Engineering 2015, 1, 10. 38. Yang, G.-R.; Shen, H.; Li, C.; Lu, T.-M. Study of metal-polymer adhesion—a new technology: Cu plasma pib. Journal of electronic materials 1997, 26, 78-82. 39. Radun, V.; von Metzen, R.; Stieglitz, T.; Bucher, V.; Stett, A. Evaluation of adhesion promoters for parylene c on gold metallization. Current Directions in Biomedical Engineering 2015, 1, 493-497. 40. Zeniieh, D.; Bajwa, A.; Ledernez, L.; Urban, G. Effect of plasma treatments and plasma‐p olymerized films on the adhesion of parylene‐c to substrates. Plasma Processes and Polymers 2013, 10, 1081-1089. 41. Zeniieh, D.; Bajwa, A.; Olcaytug, F.; Urban, G. In Parylene-c thin film for biocompatible encapsulations with very strong adhesion and superior barrier properties, 21st International Symposium on Plasma Chemistry (ISPC 21), 2013. 42. Nichols, M.F.; Halm, A.W.; James, W.J.; Sharma, A.K.; Yasuda, H. Cyclic voltammetry for the study of polymer film adhesion to platinum neurological electrodes. Biomaterials 1981, 2, 161-165. 43. Dabral, S.; Zhang, X.; Wang, B.; Yang, G.-R.; Lu, T.-M.; McDonald, J. Metal-parylene interconnection systems. MRS Online Proceedings Library Archive 1995, 381. 44. Kim, B.J.; Washabaugh, E.P.; Meng, E. In Annealing effects on flexible multi-layered parylene-based sensors, Micro Electro Mechanical Systems (MEMS), 2014 IEEE 27th International Conference on, 2014; IEEE: pp 825-828. 45. Kim, H.S.; Najafi, K. In Wafer bonding using parylene and wafer-level transfer of free-standing parylene membranes, TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th International Conference on, 2003, 2003; IEEE: pp 790-793. 46. Ziegler, D.; Suzuki, T.; Takeuchi, S. Fabrication of flexible neural probes with built-in microfluidic channels by thermal bonding of parylene. IEEE/ASME Journal of Microelectromechanical Systems 2006, 15, 1477-1482. 47. Noh, H.-s.; Choi, Y.; Wu, C.-f.; Hesketh, P.J.; Allen, M.G. In Rapid, low-cost fabrication of parylene microchannels for microfluidic applications, TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th International Conference on, 2003, 2003; IEEE: pp 798-801. 48. Noh, H.-s.; Hesketh, P.J.; Frye-Mason, G.C. Parylene gas chromatographic column for rapid thermal cycling. Journal of Microelectromechanical Systems 2002, 11, 718-725. 49. Davis, E.M.; Benetatos, N.M.; Regnault, W.F.; Winey, K.I.; Elabd, Y.A. The influence of thermal history on structure and water transport in parylene c coatings. Polymer 2011, 52, 5378-5386. 50. Kim, B.J.; Meng, E. Micromachining of parylene c for biomems. Polymers for Advanced Technologies 2016, 27, 564-576. 51. Yasuda, H.K. Some important aspects of plasma polymerization. Plasma Processes and Polymers 2005, 2, 293-304. 52. De Luna, M.M.; Chen, B.; Bradley, L.C.; Bhandia, R.; Gupta, M. Solventless grafting of functional polymer coatings onto parylene c. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 2016, 34, 041403. 53. Tiainen, V.-M. Amorphous carbon as a bio-mechanical coating—mechanical properties and biological applications. Diamond and Related Materials 2001, 10, 153-160. 54. Mori, H.; Tachikawa, H. Increased adhesion of diamond-like carbon–si coatings and its tribological properties. Surface and Coatings Technology 2002, 149, 224-229. 93 55. Tsubone, D.; Kodama, H.; Hasebe, T.; Hotta, A. Gas barrier properties and periodically fractured surface of thin dlc films coated on flexible polymer substrates. Surface and Coatings Technology 2007, 201, 6431-6436. 56. Martin, T.P.; Lau, K.K.; Chan, K.; Mao, Y.; Gupta, M.; O'Shaughnessy, W.S.; Gleason, K.K. Initiated chemical vapor deposition (icvd) of polymeric nanocoatings. Surface and Coatings Technology 2007, 201, 9400-9405. 57. Ozaydin-Ince, G.; Gleason, K.K. Transition between kinetic and mass transfer regimes in the initiated chemical vapor deposition from ethylene glycol diacrylate. Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 2009, 27, 1135-1143. 58. Grill, A. Diamond-like carbon coatings as biocompatible materials—an overview. Diamond and related materials 2003, 12, 166-170. 59. Bera, M.; Rivaton, A.; Gandon, C.; Gardette, J. Photooxidation of poly (para-xylylene). European Polymer Journal 2000, 36, 1753-1764. 60. D-08, A. Standard test method for peel resistance of adhesives (t-peel test). 2008. 61. Hsu, J.M.; Rieth, L.; Kammer, S.; Orthner, M.; Solzbacher, F. Effect of thermal and deposition processes on surface morphology, crystallinity, and adhesion of parylene-c. Sensors and Materials 2008, 20, 87- 102. 62. Kim, H.T.; Kim, C.-D.; Lee, S.-Y.; Sohn, Y.-S. Effects of annealing temperature on parylene-c films formed by chemical vapor condensation method. Molecular Crystals and Liquid Crystals 2015, 618, 139- 145. 63. Jackson, N.; Stam, F.; O'Brien, J.; Kailas, L.; Mathewson, A.; O'Murchu, C. Crystallinity and mechanical effects from annealing parylene thin films. Thin Solid Films 2016, 603, 371-376. 64. Hwang, K.S.; Park, J.H.; Lee, J.H.; Yoon, D.S.; Kim, T.S.; Han, I.; Noh, J.H. Effect of atmospheric-plasma treatments for enhancing adhesion of au on parylene-c-coated protein chips. Journal of the Korean Physical Society 2004, 44, 1168-1172. 65. Lee, J.H.; Hwang, K.S.; Kim, T.S.; Seong, J.W.; Yoon, K.H.; Ahn, S.Y. Effect of oxygen plasma treatment on adhesion improvement of au deposited on pa-c substrates. Journal of the Korean Physical Society 2004, 44, 1177-1181. 66. Yu, Q.; Deffeyes, J.; Yasuda, H. Engineering the surface and interface of parylene c coatings by low- temperature plasmas. Progress in Organic Coatings 2001, 41, 247-253. 94 he insulation performance of Parylene films was investigated optically and electrically over time when exposed to wet, in vivo environments. This chapter explores several strategies for improving insulation properties by enhancing adhesion of functional Parylene devices. Optical testing was performed by taking optical images and electrical testing by electrochemical impedance spectroscopy (EIS). For this, interdigitated electrodes were designed and fabricated with different finger width and pitch. Bubbles and delamination were observed along platinum traces after a week in warm saline during optical testing, whereas an impedance drop was observed for electrical testing within a week in warm saline. EVALUATION OF PARYLENE INSULATION T 95 Background When considering reliability of Parylene-based devices, integrity of the Parylene film and adhesion of the Parylene film is critical. As mentioned previously, there are two interfaces of concern, the Parylene-Parylene interface and the Parylene-metal interface. These interfaces can fail under chronic exposure to saline due to solution permeation through the bulk material. Results presented in Chapter 3 showed that adhesion of Parylene-platinum-Parylene interfaces can be improved using a simple thermal annealing method, interposer layers and/or adhesion promoters. This data, however, was derived from mechanical testing, and no electrical testing has yet been presented with patterned metal features and Parylene in this work. Two methods commonly used to test insulation performance of thin film coatings are electrochemical impedance spectroscopy (EIS) [1-13] and direct current (DC) leakage current [2,6,9,10,13]. By far, EIS is the most common method to evaluate the integrity of device insulation under chronic exposure to moisture. In a common test, simple interdigitated electrode structures are used to test encapsulation; however, most studies only examine polymer films as a coating material rather than a substrate material. Thus, a comprehensive study was designed to evaluate and compare various electrode design parameters using the best methods to improve adhesion investigated in Chapter 3 to extend lifetime of Parylene-metal-Parylene devices. Approach and Goals The purpose of this study consists of two major objectives. (1) Use different materials, AdPro Plus and EGDA, and thermal annealing treatment to improve electrochemical performance of Parylene-based devices under chronic exposure to moisture and (2) to investigate the insulation performance of these devices over time using optical and electrical soaking testing. The purpose is to compare various electrode design parameters and adhesion layers to improve Parylene insulation in chronically implanted devices. 96 Materials and experimental methods 5.3.1 Sample Design Test structures were designed, consisting of a symmetric Parylene-platinum-Parylene bilayer thin-film. Platinum structures consisted of simple interdigitated comb-electrodes attached to larger contact pads for external access (Figure 5-1). IDEs were designed with varied trace pitch and width to vary the ratio of Parylene-Parylene and Parylene-metal interfaces, as the adhesion and adhesion failure mechanisms of these interfaces are known to be different. Figure 5-1. (a) Sample schematic of IDE structures; and (b) zoom in to observe the different finger width and pitch considered for electrode design. Four sample groups were prepared with different treatments: 48-hour thermal annealing, AdPro Plus® adhesion promotor, EGDA interposer layer, and an untreated control. For each experimental treatment devices were fabricated with two different trace widths, 10 and 100 µm. For each trace width, devices were prepared with varied pitch. Table 5-1 shows the different materials and electrode parameters used in this study. Table 5-1. IDE experimental groups and electrode parameters. Experimental Groups Trace width Pitch Control 48h Annealed AdPro Plus® EGDA 10 and 100 µm 5, 10, 25, 50 and 100 µm 97 5.3.2 Sample Fabrication Test structures consisted of simple platinum traces fully encapsulated between two 10 μm-thick Parylene layers and contact pads for electrical connection (Figure 5-2). Annealing, an adhesion promotor, and an interposer layer, as well as different electrode designs, were investigated for their impact on insulation integrity through improvements in adhesion. Figure 5-2. Fabricated IDEs. Figure 5-3 illustrates the fabrication process by which samples were prepared. First, Si carrier wafers were dehydrated at 110 C then coated in a layer of Parylene (10 µm) by CVD (PDS 2010, Specialty Coating Systems (SCS), Indianapolis, IN) (Figure 5-3a). Then, the electrodes and contact pads patterns were defined through lithography using a 1.5 μm mask (AZ 5214 E-IR; Integrated Micro Materials, Argyle, TX), followed by platinum deposition with e-beam evaporation (2000 Å, UCLA Nanoelectronics Research Facility, Los Angeles, CA). Lift-off defined the IDEs and contact pads (Figure 5-3b), then wafers were cleaned using O2 plasma prior deposition of EGDA, AdPro Plus® and the top Parylene layer (10 µm) (Figure 5-3c) depending on the material combination. It is critical to note that any type of cleaning process must be avoided after EGDA and AdPro Plus® depositions to prevent removal of the material layers. The sample cutout and contact pads were etched using O2 DRIE (Oxford Plasmalab 100, Oxford Plasma Technology, UK) through two photoresist masks (15 μm, AZ 4620; AZ Electronic Materials, Branchburg, NJ) (Figure 5-3d). The first mask etched through 10 µm of the cutout and the second mask through the other 10 µm of the cutout and exposed the contact pads. The etch mask was 98 stripped in acetone following an isopropanol rinse, and finally the samples were released by gently peeling them in water using tweezers. A subset of samples was thermally annealed at 200 °C for 48 hours under vacuum, to prevent oxidation of Parylene, then cooled to room temperature overnight. Figure 5-3. Cross-sectional view of the fabrication process for IDE. (a) Bare Si carrier wafer; (b) deposition of substrate layer; (c) platinum e-beam evaporation deposition; (d) lift-off of IDEs and contact pads; (e) deposition of adhesion layer if needed; (f) deposition of insulation layer; (g) DRIE etching; and (h) released devices. 5.3.3 Electrical packaging 8-channel zero insertion force (ZIF) connectors (Molex Inc., Lisle, IL) were used for electrical connection of the IDEs with a flat flexible cable (FFC; Molex Inc., Lisle, IL), which was inserted to a second ZIF connector mounted on a printed circuit board (PCB) (Figure 5-4). Contact pads required the use of a poly-ether-ether-ketone (PEEK) backing (McMaster-Carr Supply Company, Elmhurst, IL) support (250 µm thick) to provide the necessary thickness to establish a good contact between the ZIF connector and the IDEs contact pads. 99 Figure 5-4. Electrical packaging of an IDE, which is attached to a ZIF connector. Biocompatible Epo-Tek® 353ND-T epoxy (Epoxy Technology Inc., Billerica, MA) was used to seal the contact pads and ZIF connector to prevent any solution ingress that may cause electrical failure. Epoxy was also used to secure a set of IDEs to the caps of vials, which were used for the long-term soaking study (Figure 5-5). Figure 5-5. IDE device secured in vial caps with epoxy. 5.3.4 Testing Protocol/Soaking Study 5.3.4.1 Optical Testing Samples were tested under passive soak conditions with no electrical current applied. Approximately 40 samples of every material combination were immersed in 1× concentration PBS at 37 °C to mimic physiological conditions (Figure 5-6). The degradation of the encapsulation was visually inspected under the microscope every week for a month and then every two weeks after 100 that. Samples were taken out from the vials, dry with a piece of cloth, optical pictures were taken, and then samples were returned to the vials. Figure 5-6. Immersed samples in 1× PBS. 5.3.4.2 Electrical Testing The integrity of the device insulation was evaluated using EIS by monitoring the capacitance and resistance of the encapsulation during exposure to a corrosive saline environment. Approximately 40 samples of every material combination were immersed in 1× concentration PBS at 37 °C (Figure 5-7); the solution was replaced every week to maintain a constant conductivity. The integrity of the insulation was evaluated daily during the first week and then weekly for a month. EIS measurements were recorded in 1× PBS at 37 °C with an AC signal of 25 mV (rms) in the frequency range from 1 to 10 5 Hz with a Reference 600 potentiostat (Gamry Instruments, Warminster, PA). A 2-electrode setup was used where one IDE was connected to the working electrode and the other IDE was connected to the reference and counter electrodes. 101 Figure 5-7. IDE devices were epoxied into vial caps for soaking study. Figure 5-8a shows the circuit model for two perfectly insulated electrodes. The circuit consists of a solution resistance Rs, a coating resistance Rc, that is typically in the order of GΩ, and a coating capacitance Cc, which depends on the coating thickness. This equivalent circuit can be used to model the coating when water diffuses into the coating. Water diffusion into structures of the coating increases Cc as the permeability of water is much higher than that of Parylene while Rc decreases with water diffusion as ions in electrolyte increase the conductivity of the coating. An electrode-electrolyte interface is formed between the metal and solution when water reaches the electrode surface and condenses underneath the coating. The electrode-electrolyte interface is modeled as a double-layer capacitance Cdel with a charge transfer resistance Rdel connected in parallel (Figure 5-8b). 102 Figure 5-8. (a) Circuit model of a Parylene-platinum-Parylene device with encapsulated electrodes under chronic soaking conditions; and (b) circuit model of the same device when water reaches the platinum electrode, creating an electrode-electrolyte interface. A representative EIS plot, prior soaking in warm PBS (day 0), of a control Parylene coated IDE with trace width of 10 µm and pitch of 5 µm is presented in Figure 5-9, which has impedance and phase angle plotted as a function of frequency. The phase angle of approximately −90° indicates a pure capacitor behavior, suggesting that there are no defects or pinholes in the insulation film. 103 Figure 5-9. Representative EIS raw data of a perfectly insulated IDE. As soaking time increases, the impedance decreases in the low frequency range and phase increases (Figure 5-10). These EIS changes in measured data suggest degradation or water absorption in the Parylene film. Thus, impedance was recorded at 1 Hz to observe the encapsulation behavior. Figure 5-10. (a) Impedance; and (b) phase as a function of frequency. Impedance decreased and phase increased as a function of soaking time, indicating of water absorption. 104 Results 5.4.1 Optical Testing 5.4.1.1 Control Devices Optical microscopy images of control devices during the soaking study are shown in Figure 5-11. Microscopy revealed bubble formation along the metal traces for all the IDEs after a week in warm PBS. Over the course of the study, more bubbles appeared no matter the pitch. For IDEs with 5 µm pitch, bubbles are difficult to see due to dry salt on the surface of the devices. Figure 5-11. Optical micrographs of control IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. 105 Interestingly, no signs of bubble formation or metal delamination were observed for control devices with 100 µm finger width. Figure 5-12 shows the optical microscopy images up to 2 months in warm saline. Figure 5-12. Optical micrographs of control IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. 106 5.4.1.2 Annealed Parylene Devices Optical microscopy images of annealed devices (Figure 5-13) did not show bubble formation or metal delamination, after 2 months in warm saline, for all pitch combinations of devices with 10 µm finger width. These results suggest that annealing improves adhesion of IDE devices no matter the Parylene-to-platinum ratio. Figure 5-13. Optical micrographs of annealed IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. 107 Annealing had the same effect on IDE devices with 100 µm finger width where no bubbles or metal delaminated were observed. Figure 5-14 shows the optical microscopy images of annealed devices up to 2 months in warm saline. Figure 5-14. Optical micrographs of annealed IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. 108 5.4.1.3 AdPro Plus® Devices Optical microscopy images of IDE devices with AdPro Plus® are shown in Figure 5-15. These images reveled bubble formation after a month in warm PBS but only for devices with 5 µm pitch. Bubbles formed along the IDE fingers, but metal was not delaminated. Figure 5-15. Optical micrographs of AdPro Plus® IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. 109 AdPro Plus® improved barrier properties for IDE devices with a 100 µm finger width, as a result, no bubbles were observed after a month in warm saline. Figure 5-16 shows the optical microscopy images. Figure 5-16. Optical micrographs of AdPro Plus® IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. 110 5.4.1.4 EGDA Devices Optical inspection revealed bubble formation after 2 weeks in warm saline. Figure 5-17 shows images where bubbles can be observed for IDE devices with 5, 10 and 25 µm pitch. This suggest that Parylene-to-metal ratio for this combination was an important factor in bubble formation. Figure 5-17. Optical micrographs of EGDA IDE devices (finger width = 10 µm) as a function of time soaked in warm saline, showing the appearance of bubbles along the electrode traces. Columns represent soaking time at different time points and rows represent the different electrode pitch. 111 EGDA also improved barrier properties of IDE devices with a 100 µm finger width. No bubbles or platinum delamination was observed over 1 month in warn saline. Figure 5-18 shows the optical microscopy images. Figure 5-18. Optical micrographs of EGDA IDE devices (finger width = 100 µm) as a function of time soaked in warm saline, showing no delamination. Columns represent soaking time at different time points and rows represent the different electrode pitch. 5.4.2 Electrical Testing 5.4.2.1 Control Devices The impedance (Figure 5-19a) of 10 µm IDEs decreased (<GΩ) after 7 days in warm solution, whereas the phase angle (Figure 5-19b) of these devices with pitch of 5 and 50 µm 112 showed no insulation failure for a week, then the phase angle started to increase. Both type of devices reached the same baseline after insulation failure. However, the impedance for 100 µm IDEs was higher than GΩ except at day 21 where it was lower than GΩ (Figure 5-19c); the phase angle started to increase after 5 days in warm PBS (Figure 5-19d). Figure 5-19. Measured impedance and phase angle at 1 Hz as function of soaking time for control devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. 5.4.2.2 Annealed Parylene Devices 10 µm annealed IDEs performed slightly better than 100 µm IDEs. The former devices with 10, 50 and 100 µm pitch showed high impedance (>GΩ) up to 28 days in warm PBS (Figure 5-20a). However, the angle phase started to increase after 1 day in warm saline for all pitch (Figure 5-20b). On the other hand, 100 µm devices showed a decreased in impedance (<GΩ) after 5 and 113 14 days in warm PBS (Figure 5-20c), whereas the phase angle increased after 1 day in warm PBS for most of the pitch combinations, except for 25 and 50 µm pitch (Figure 5-20d). Figure 5-20. Measured impedance and phase angle at 1 Hz as function of soaking time for annealed devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. 5.4.2.3 AdPro Plus® Devices The inclusion of AdPro Plus® maintained impedance stable of 10 µm IDEs during 7 days in warm saline, and after 14 days the impedance decreased (<GΩ) (Figure 5-21a). However, the phase angle of IDEs with 50 µm pitch started to increase after a day in warm PBS, whereas devices with 10 and 25 µm pitch started to increase after 5 days in warm saline (Figure 5-21b). The impedance was high (>GΩ) for 5 days in warm saline for 100 µm IDEs, and then it kept decreasing 114 over time (Figure 5-21c). The phase angle of these devices was stable for 5 days with a spike at day 2, and an increment after 6 days in warm PBS (Figure 5-21d). Figure 5-21. Measured impedance and phase angle at 1 Hz as function of soaking time for AdPro Plus® devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. 5.4.2.4 EGDA Devices Impedance was high (>GΩ) for 10 µm IDEs up to 7 and 21 days in warm saline for devices with 5 and 10 µm pitch, respectively. For the other pitch, the impedance started to decrease (<GΩ) and it was the lowest recorded of all the material combinations tested (Figure 5-22a). Same behavior was observed for the phase angle for IDEs with 5 and 10 µm pitch, where it was stable for 7 days in warm PBS and then it increased (Figure 5-22b). For 100 µm IDEs with 5, 25, 50 µm 115 pitch, the impedance was high (>GΩ) up to 21 days in warm PBS (Figure 5-22c) but the phase angle was approximately -80° up to 5 days for 50 µm pitch (Figure 5-22d). Figure 5-22. Measured impedance and phase angle at 1 Hz as function of soaking time for EGDA devices. (a) Impedance and (b) phase of 10 µm width IDEs; and (c) impedance and (d) phase of 100 µm width IDEs. Discussion Annealing treatment and two adhesion layers, AdPro Plus® and EGDA, as well as electrode designs were evaluated to identify fabrication and design parameters to improve insulation integrity while improving adhesion of Parylene-based devices. During optical testing, no bubbles or delamination were observed for IDEs with 100 µm finger width no matter the treatment or adhesion layer up to 2 months in warm saline. However, bubbles appeared, at different time points, for control, AdPro Plus® and EGDA devices with 10 µm finger width. 116 Interestingly, annealed devices remained fully and perfectly encapsulated after a 2-month soaking study. It is worth to mention that bubbles appeared for small pitch (5, 10 and 25 µm) for treated devices. This suggests that Parylene-to-platinum ratio plays an important role while using treatments, the larger the ratio the better the insulation and adhesion. The failure mechanism seems to be through the bulk material because only Parylene-platinum delamination was observed along the IDEs. Unfortunately, current application accelerated device failure of all the devices no matter what treatment, adhesion layer or electrode design was used. This disagrees with what was found in the mechanical study presented in Chapter 4 where the interposer layer improved adhesion of Parylene-platinum-Parylene interfaces. It is pointing out that the peeling study was conducted under passive conditions without any application of electrical current. In this electrical study, impedance dropped for most devices after 7 days in warm saline. However, some annealed devices maintained high electrochemical impedance for up to 28 days. These suggests that annealing improved adhesion and barrier properties of devices while the use of adhesion layers may improve adhesion but not barrier properties. Yasuda et al. coated aluminum sheets with 5 µm of Parylene with and without a plasma adhesive layer of trimethylsilane (TMS) to improve adhesion between the metal and the Parylene layer; they reported perfect insulation during 18 days in 0.9% NaCl when TMS was used as adhesion layer and sheets without any treatment started to fail after 1 day in the solution [1]. Li et al. studied the corrosion behavior of resistors insulated between Parylene layers of different thickness, 5 and 10 µm [5]. They observed bubbles and delamination in all samples after 1 day in 0.9% NaCl at 90 °C where no electric current was applied. However, holes were observed in the 5 µm Parylene film when electric current was applied; they suggest that the bubbles appeared where the magnetic field is concentrated due to defects or particles. On the other hand, resistors that were insulated with 10 µm layers of Parylene only showed bubbles due to moisture permeation. Seymour et al. used 2.5 µm Parylene layers to insulate IDEs with and without a plasma adhesion layer and an annealing treatment. All samples showed a dramatic decreased in impedance during the first hour of being soaked in PBS at 37 °C but annealed and devices with 117 plasma adhesive showed better impedance results due to improvements in adhesion over 60 days [7]. Hassler et al. used also silicon IDEs insulated with 10 µm of Parylene using O2 plasma, silane A-174, annealing and a combination of treatments to improve adhesion. The control device showed a decrease in impedance right after soaking it in warm PBS, whereas the silane sample showed a decrement after 38 days in warm PBS [8]. Von Metzen and Stieglitz used standard IDEs and perforated IDEs to observe the effects of Parylene insulation. IDEs were insulated between 10 µm Parylene layers and silane A-174 was used prior deposition of the second Parylene layer to improve adhesion. Finally, samples were annealed and then soaked in 0.9% NaCl solution at 37 °C. Un-annealed samples showed delamination after a few hours in the warm solution, whereas annealed samples at 200 °C lasted for 300 days in saline solution with a phase angle increment after 150 days. However, when samples were annealed at 300 °C showed stable results up to 320 days [10]. In most reports, impedance decreased after 1 day in warm saline solution no matter the treatment. However, Parylene thickness plays an important role since water diffuses through the bulk of the material rather than between the Parylene layers. All samples present delamination at the Parylene-metal-Parylene interface instead of in the Parylene-Parylene interface. Conclusions Insulation performance of Parylene-based devices was investigated by taking optical images and by measuring EIS up to 2 months in warm saline. Different adhesion layers, AdPro Plus® and EGDA, annealing treatment and different electrode designs were investigated. Bubbles were observed for all 10 µm finger width devices during optical testing, where as no delamination or bubble formation were observed for 100 µm finger width devices. The application of electrical current accelerated device failure for all type of devices within a week in warm saline. 118 References 1. Yasuda, H.; Yu, Q.; Chen, M. Interfacial factors in corrosion protection: An eis study of model systems. Progress in organic Coatings 2001, 41, 273-279. 2. Minnikanti, S.; Diao, G.; Pancrazio, J.J.; Xie, X.; Rieth, L.; Solzbacher, F.; Peixoto, N. Lifetime assessment of atomic-layer-deposited al2o3–parylene c bilayer coating for neural interfaces using accelerated age testing and electrochemical characterization. Acta biomaterialia 2014, 10, 960-967. 3. Chun, W.; Chou, N.; Cho, S.; Yang, S.; Kim, S. Evaluation of sub-micrometer parylene c films as an insulation layer using electrochemical impedance spectroscopy. Progress in Organic Coatings 2014, 77, 537-547. 4. Hsu, J.-M.; Tathireddy, P.; Rieth, L.; Normann, A.R.; Solzbacher, F. Characterization of a-sicx: H thin films as an encapsulation material for integrated silicon based neural interface devices. Thin solid films 2007, 516, 34-41. 5. Li, W.; Rodger, D.; Menon, P.; Tai, Y.-C. Corrosion behavior of parylene-metal-parylene thin films in saline. Ecs Transactions 2008, 11, 1-6. 6. Hsu, J.-M.; Rieth, L.; Normann, R.A.; Tathireddy, P.; Solzbacher, F. Encapsulation of an integrated neural interface device with parylene c. IEEE Transactions on Biomedical Engineering 2009, 56, 23-29. 7. Seymour, J.P.; Elkasabi, Y.M.; Chen, H.-Y.; Lahann, J.; Kipke, D.R. The insulation performance of reactive parylene films in implantable electronic devices. Biomaterials 2009, 30, 6158-6167. 8. Hassler, C.; von Metzen, R.P.; Ruther, P.; Stieglitz, T. Characterization of parylene c as an encapsulation material for implanted neural prostheses. Journal of Biomedical Materials Research Part B: Applied Biomaterials: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials 2010, 93, 266-274. 9. Xie, X.; Rieth, L.; Tathireddy, P.; Solzbacher, F. Long-term in-vivo investigation of parylene-c as encapsulation material for neural interfaces. Procedia Engineering 2011, 25, 483-486. 10. Von Metzen, R.P.; Stieglitz, T. The effects of annealing on mechanical, chemical, and physical properties and structural stability of parylene c. Biomedical microdevices 2013, 15, 727-735. 11. Lee, C.D.; Meng, E. Mechanical properties of thin-film parylene–metal–parylene devices. Frontiers in Mechanical Engineering 2015, 1, 10. 12. Forssell, M.; Ong, X.C.; Khilwani, R.; Ozdoganlar, O.B.; Fedder, G.K. Insulation of thin-film parylene- c/platinum probes in saline solution through encapsulation in multilayer ald ceramic films. Biomedical microdevices 2018, 20, 61. 13. Xie, X.; Rieth, L.; Caldwell, R.; Diwekar, M.; Tathireddy, P.; Sharma, R.; Solzbacher, F. Long-term bilayer encapsulation performance of atomic layer deposited al $ _ {\bf 2} $ o $ _ {\bf 3} $ and parylene c for biomedical implantable devices. IEEE Transactions on Biomedical Engineering 2013, 60, 2943-2951. 14. Lee, J.H.; Hwang, K.S.; Kim, T.S.; Seong, J.W.; Yoon, K.H.; Ahn, S.Y. Effect of oxygen plasma treatment on adhesion improvement of au deposited on pa-c substrates. Journal of the Korean Physical Society 2004, 44, 1177-1181. 15. Yu, Q.; Deffeyes, J.; Yasuda, H. Engineering the surface and interface of parylene c coatings by low- temperature plasmas. Progress in Organic Coatings 2001, 41, 247-253. 119 APPENDIX A Ion Transport through Parylene Films Sterilization: 1. Autoclave, for 15 minutes, the borosilicate electrochemical cells, 2 media storage glass bottles for Millipore water and PBS, respectively, and a graduate glass cylinder. 2. Let the glassware cool down for 10 minutes. Procedure: 1. Cut the Parylene film to be tested using a blade. 2. Place the Parylene film between the electrochemical cells using the Teflon O-ring and the flange with knuckle clamp. 3. Close the electrochemical cells and electrode ports with the caps. 4. Get fresh Millipore water in one of the storage glass bottles. 5. Make 30 mL of 1× concentration PBS using the fresh Millipore water and the sterilized graduate cylinder to measure the PBS and the Millipore water needed. Agitate the PBS solution for 5 minutes. 6. Pour 30 mL of Millipore in one of the electrochemical cells. 7. Place a Ag/AgCl reference electrode, through the electrode port in the electrochemical cell with water, and the counter and working electrodes. 8. Run EIS and record the water solution resistance at the appropriate frequency determined by minimizing the impedance phase. 9. Pour 30 mL of 1× PBS in the other electrochemical cell, and place another Ag/AgCl reference electrode, and the counter and working electrodes. 10. Run EIS and record the PBS solution resistance at the appropriate frequency determined by minimizing the impedance phase. 11. Take measurements every 15 minutes for 3 hours daily until the water resistance reaches 10% of its original value. 120 12. Remove Ag/AgCl reference electrodes every day and place them back the next day to continue testing. Clean up: 1. Remove counter and working electrodes, and rinse them with IPA and water. 2. Remove Ag/AgCl reference electrodes and rinse them with Millipore water. 3. Dump Millipore water and PBS solution in the sink. 4. Rinse electrochemical cells with Millipore water. 5. Remove the Parylene film, Teflon O-ring and clamping mechanism. 6. Let the electrochemical cells dry overnight. 121 APPENDIX B A-174 Silane Treatment Caution: Performed under the solvents Fume Hood only. Mixing steps: 1. Check the shelf-life tag on the A-174 bottle; if > 6 months, solution needs to be replaced. 2. Make 0.5% of A-174 in IPA and DI water. a. IPA: DI H20: A-174 = 500:500:5 ml. Stir for 30 seconds with a dedicated glass stirring rod. Make sure the solution is mixed well. b. Let it stand for 2 hours prior to use. c. Lifetime of mixed solution is 24 hours. 3. After 2 hours, submerge the cleaned wafer(s) in the solution for 15-30 min. 4. Air dry for 15-30 min or carefully blow dry or spin dry in spin dryer (wafer must be dry before proceeding to next step). a. Do not let surfaces to be coated to come into contact with dirty surfaces. 5. Immerse in IPA bath for 15-30 sec. Gently agitate wafers/parts in bath. 6. Air dry for 30-60 seconds or spin dry in spin dryer. 7. Bake dry to remove moisture in an oven for 15 min @ 100°C (wafer must be completely dry before Parylene coating). Clean up: 1. Treat used Silane A-174 as combustible/toxic waste. a. Collect the waste in a labeled waste glass container and store in flammable cabinet. b. Contact safety office (213-740-7215) for pick up when full. 2. Store A-174 back in the flammable cabinet. 3. Store A-174 dedicated glassware/labware back in the designated drawer in lab. 122 APPENDIX C DRIE Coupons Process Flow Figure C-1. Mask transparency. 1. A-174 silane treatment (see Appendix A) 2. Deposit Parylene (~12 µm, 27 g of dimer) 3. Pattern etch mask: a. AZ 4620 (~12 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 1200 rpm Softbake 115 °C, 3 minutes Hydration 60 minutes Exposure 550 mJ/cm 2 (25 mW/cm 2 , 22 sec) Development 4 minutes 30 sec b. KMPR 1025 etch mask (~25 µm thick) Pre spin 10 sec, 500 rpm Spin 45 sec, 4000 rpm Softbake 100 °C, 10 minutes Exposure 485 mJ/cm 2 (25 mW/cm 2 , 19.4 sec) Hardbake 100 °C, 2 minutes Development 2 minutes 4. Dice wafers. 123 APPENDIX D DRIE Profiles Effect of ICP Power Figure D-1. SEM cross-section images of the PDMS replica molds showing the effect of varying the ICP power on the sidewall angle. Scale bars are 10 µm in length. Effect of RF Power Figure D-2. SEM cross-section images of the PDMS replica molds showing the effect of varying the RF power on the sidewall angle. Scale bars are 10 µm in length. 124 Effect of Flow Rate of Gases Figure D-3. SEM cross-section images of the PDMS replica molds showing the effect of varying Ar and SF6 flow rates on the sidewall angle for the etch step. Scale bars are 10 µm in length. Figure D-4. SEM cross-section images of the PDMS replica molds showing the effect of varying Ar, for passivation and etch steps, and SF6, for etch step, flow rates on the sidewall angle. Scale bars are 10 µm in length. 125 Effect of Oxygen Flow Rate Figure D-5. SEM cross-section images of the PDMS replica molds showing the effect of varying O2 flow rate on the sidewall angle. Scale bars are 10 µm in length. 126 Effect of the Passivation Layer Figure D-6. SEM cross-section images of the PDMS replica molds showing the effect of varying the deposition time of C4F8 on the sidewall angle. Scale bars are 10 µm in length. Figure D-7. SEM cross-section images of the PDMS replica molds showing the effect of varying C4F8 flow rate on the sidewall angle. Scale bars are 10 µm in length. 127 APPENDIX E Dry Adhesion Forces for Parylene Interfaces Table E-1. Comparison of Dry Forces Reported in Literature Test Interface Adhesion Treatment Force Failure Interface Ref. Tensile Pt/PxN None ~5 Pa [14] GDMP ~7.5 Pa Si/PxC&PxC/Si 160 °C, 10 min, 1.5 MPa 3.45 MPa PxC/Si [15] 160 °C, 30 min, 1.5 MPa 5.38 MPa PxC/Si 160 °C, 10 min, 1.5 MPa, A-174 9.16 MPa PxC tore Si/PxC&PxC/Si 230 °C, 30 min, 800 N 3.60 MPa PxC-PxC [16] Si/PxC&PxC/Si A-174, 230 °C, 30 min, 800 N 1.12 MPa [17] Peel 90° Pt/PxC PPE 1,2 14210 mN/mm [18] PPE 1,2 , 125 °C, 1 h 21560 mN/mm PPE 2 35280 mN/mm PPM 18620 mN/mm PPM, 125 °C, 1 h 15190 mN/mm Ti/PxC PPE 1,2 34300 mN/mm PPE 2 24500 mN/mm PPM 9800 mN/mm PxN/Cu Plasma PIB 50-120 mN/mm [19] PxN/Cu Thermal evaporation Fail PxN/Cu [20] DC sputtering Fail PxN/Cu E-beam 50 mN/mm PxN/Cu PIB floating 50 mN/mm PxN/Cu PIB 100 mN/mm PxN/Cu Plasma PIB >900 mN/mm PxN Plasma PIB, 350 °C, 0.5 h 20 mN/mm Cr/PxC None 240 mN/mm Cr/PxC [7] 140 °C, 3 h >380 mN/mm Cr/PxC X 330 mN/mm X/PxC X, 140 °C, 3 h >430 mN/mm X/PxC or Cr/X Si/Pt/PxC None <5 mN/mm [21] O2 plasma <5 mN/mm A-174 ~100 mN/mm A-174 + plasma ~100 mN/mm A-174, 200 °C, 1 h ~100 mN/mm PxC/PxC None 480 mN PxC tore [22] 200 °C 790 mN PxC tore HF 590 mN PxC tore HF, 200 °C 880 mN PxC tore Hexane 630 mN PxC tore Hexane, 200 °C 830 mN PxC tore 128 Toluene 680 mN PxC tore Toluene, 200 °C 920 mN PxC tore P.C. 1030 mN PxC tore P.C., 200 °C 1060 mN PxC tore CF4 850 mN PxC tore CF4, 200 °C 970 mN PxC tore Ti/Pt/PxC Ar plasma 70 mN/mm [23] O2 plasma 90 mN/mm Ar plasma + TMS 240 mN/mm O2 plasma + TMS 340 mN/mm Ti/Au/PxC Ar plasma + TMS 230 mN/mm O2 plasma + TMS 320 mN/mm Pt/Ir/PxC None 6 mN/mm [24] Hatch 45° 400 mN/mm Hatch 2 um 4 mN/mm Hatch 5 um 310 mN/mm Au/PxC Ti Torn [25] Au/PxC DLC Torn Au/PxC Ti/SiOx Torn Au/PxC TMS Torn Delaminated Au/PxC A-174 Torn Peel 180° PxC/PxC None 170 mN/mm [7] 140 °C, 3 h >90 mN/mm X 1 mN/mm X 140 °C, 3 h 3 mN/mm PxC/PxC None ~100 mN/mm [21] O2 plasma ~100 mN/mm A-174 ~100 mN/mm A-174 + O2 plasma 1-100 mN/mm A-174, 200 °C, 1 h ~100 mN/mm Ti/PxC Ar plasma + TMS 540 mN/mm [23] O2 plasma + TMS 680 mN/mm PxC/PxC 440 mN/mm [24] PxC/Pt/PxC None 21900 mN/mm Top PxC [11] PxC/Pt/Al2O3/PxC Al2O3 ~12000 mN/mm PxC/Al2O3/PxC Al2O3 ~9000 mN/mm PxC/Al2O3/Pt/Al2O3/PxC Al2O3 ~5000 mN/mm PxC/Al2O3/Al2O3/PxC Al2O3 ~4000 mN/mm PxC/PxC None 430 mN/mm PxC/PxC None 0.08 MPa [26] A-174 15.25 MPa O2 plasma 32.7 MPa Pull-off Si/PxC/Cr/Au None 1.37 MPa A-174 1.23 MPa 129 O2 plasma 2.13 MPa GDMP = glow discharge polymerized methane P.C. = propylene carbonate PxC = Parylene PxN = Parylene N PPE = Plasma-polymerized ethane PPM = Plasma-polymerized methane TMS = trimethylsilane X = CH2NH2 or CHO, aldehyde and aminomethyl side group 1 = Washed with Micro brand cleaner, rinsed with water, isopropyl alcohol, Transene-100 2 = Metal surface treated with oxygen plasma, resulting film treated with an oxygen plasma References 1. Yasuda, H.; Yu, Q.; Chen, M. Interfacial factors in corrosion protection: An eis study of model systems. Progress in organic Coatings 2001, 41, 273-279. 2. Minnikanti, S.; Diao, G.; Pancrazio, J.J.; Xie, X.; Rieth, L.; Solzbacher, F.; Peixoto, N. Lifetime assessment of atomic-layer-deposited al2o3–parylene c bilayer coating for neural interfaces using accelerated age testing and electrochemical characterization. Acta biomaterialia 2014, 10, 960-967. 3. Chun, W.; Chou, N.; Cho, S.; Yang, S.; Kim, S. Evaluation of sub-micrometer parylene c films as an insulation layer using electrochemical impedance spectroscopy. Progress in Organic Coatings 2014, 77, 537-547. 4. Hsu, J.-M.; Tathireddy, P.; Rieth, L.; Normann, A.R.; Solzbacher, F. Characterization of a-sicx: H thin films as an encapsulation material for integrated silicon based neural interface devices. Thin solid films 2007, 516, 34-41. 5. Li, W.; Rodger, D.; Menon, P.; Tai, Y.-C. Corrosion behavior of parylene-metal-parylene thin films in saline. Ecs Transactions 2008, 11, 1-6. 6. Hsu, J.-M.; Rieth, L.; Normann, R.A.; Tathireddy, P.; Solzbacher, F. Encapsulation of an integrated neural interface device with parylene c. IEEE Transactions on Biomedical Engineering 2009, 56, 23-29. 7. Seymour, J.P.; Elkasabi, Y.M.; Chen, H.-Y.; Lahann, J.; Kipke, D.R. The insulation performance of reactive parylene films in implantable electronic devices. Biomaterials 2009, 30, 6158-6167. 8. Hassler, C.; von Metzen, R.P.; Ruther, P.; Stieglitz, T. Characterization of parylene c as an encapsulation material for implanted neural prostheses. Journal of Biomedical Materials Research Part B: Applied Biomaterials: An Official Journal of The Society for Biomaterials, The Japanese Society for Biomaterials, and The Australian Society for Biomaterials and the Korean Society for Biomaterials 2010, 93, 266-274. 9. Xie, X.; Rieth, L.; Tathireddy, P.; Solzbacher, F. Long-term in-vivo investigation of parylene-c as encapsulation material for neural interfaces. Procedia Engineering 2011, 25, 483-486. 10. Von Metzen, R.P.; Stieglitz, T. The effects of annealing on mechanical, chemical, and physical properties and structural stability of parylene c. Biomedical microdevices 2013, 15, 727-735. 11. Lee, C.D.; Meng, E. Mechanical properties of thin-film parylene–metal–parylene devices. Frontiers in Mechanical Engineering 2015, 1, 10. 12. Forssell, M.; Ong, X.C.; Khilwani, R.; Ozdoganlar, O.B.; Fedder, G.K. Insulation of thin-film parylene- c/platinum probes in saline solution through encapsulation in multilayer ald ceramic films. Biomedical microdevices 2018, 20, 61. 13. Xie, X.; Rieth, L.; Caldwell, R.; Diwekar, M.; Tathireddy, P.; Sharma, R.; Solzbacher, F. Long-term bilayer encapsulation performance of atomic layer deposited al $ _ {\bf 2} $ o $ _ {\bf 3} $ and parylene c for biomedical implantable devices. IEEE Transactions on Biomedical Engineering 2013, 60, 2943-2951. 14. Nichols, M.F.; Halm, A.W.; James, W.J.; Sharma, A.K.; Yasuda, H. Cyclic voltammetry for the study of polymer film adhesion to platinum neurological electrodes. Biomaterials 1981, 2, 161-165. 130 15. Noh, H.-s.; Moon, K.-s.; Cannon, A.; Hesketh, P.J.; Wong, C. In Wafer bonding using microwave heating of parylene for mems packaging, Electronic Components and Technology Conference, 2004. Proceedings. 54th, 2004; IEEE: pp 924-930. 16. Kim, H.; Najafi, K. Characterization of low-temperature wafer bonding using thin-film parylene. Journal of microelectromechanical systems 2005, 14, 1347-1355. 17. Shu, Q.; Huang, X.; Wang, Y.; Chen, J. In Wafer bonding with intermediate parylene layer, Solid-State and Integrated-Circuit Technology, 2008. ICSICT 2008. 9th International Conference on, 2008; IEEE: pp 2428-2431. 18. Yamagishi, F.G. Investigations of plasma-polymerized films as primers for parylene-c coatings on neural prosthesis materials. Thin Solid Films 1991, 202, 39-50. 19. Dabral, S.; Zhang, X.; Wang, B.; Yang, G.-R.; Lu, T.-M.; McDonald, J. In Metal-parylene interconnection systems, MRS Proceedings, 1995; Cambridge Univ Press: p 205. 20. Yang, G.-R.; Shen, H.; Li, C.; Lu, T.-M. Study of metal-polymer adhesion—a new technology: Cu plasma pib. Journal of electronic materials 1997, 26, 78-82. 21. Hassler, C.; von Metzen, R.P.; Ruther, P.; Stieglitz, T. Characterization of parylene c as an encapsulation material for implanted neural prostheses. Journal of Biomedical Materials Research Part B: Applied Biomaterials 2010, 93, 266-274. 22. Chang, J.H.-C.; Lu, B.; Tai, Y.-C. In Adhesion-enhancing surface treatments for parylene deposition, Solid- State Sensors, Actuators and Microsystems Conference (TRANSDUCERS), 2011 16th International, 2011; IEEE: pp 390-393. 23. Zeniieh, D.; Bajwa, A.; Ledernez, L.; Urban, G. Effect of plasma treatments and plasma‐polymerized films on the adhesion of parylene‐c to substrates. Plasma Processes and Polymers 2013, 10, 1081-1089. 24. Mueller, M.; Ulloa, M.; Schuettler, M.; Stieglitz, T. In Development of a single-sided parylene c based intrafascicular multichannel electrode for peripheral nerves, Neural Engineering (NER), 2015 7th International IEEE/EMBS Conference on, 2015; IEEE: pp 537-540. 25. Radun, V.; von Metzen, R.; Stieglitz, T.; Bucher, V.; Stett, A. Evaluation of adhesion promoters for parylene c on gold metallization. Current Directions in Biomedical Engineering 2015, 1, 493-497. 26. Xie, Y.; Pei, W.; Guo, D.; Zhang, L.; Zhang, H.; Guo, X.; Xing, X.; Yang, X.; Wang, F.; Gui, Q. Improving adhesion strength between layers of an implantable parylene-c electrode. Sensors and Actuators A: Physical 2017, 260, 117-123. 131 APPENDIX F Parylene T-Peel Test Structures Process Flow Figure F-1. Masks transparency. 1. Dehydration bake silicon wafers to remove moisture 110 °C, > 10 min 2. Deposit Parylene (~12 µm, 27 g of dimer) 3. Descum, O2 plasma 100 W, 100 mTorr, 1 min 4. Pattern AZ 4400 sacrificial layer (4 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 4000 rpm Softbake 90 °C, 2 min Exposure 150 mJ/cm 2 (10 mW/cm 2 , 15 sec) Development 50 sec Hardbake 90 °C, 3 min 5. Descum, O2 plasma 100 W, 100 mTorr, 1 min 6. Deposit adhesion promoter DLC or EGDA (if needed) 7. Deposit Parylene (~12 µm, 27 g of dimer) 8. Pattern AZ 4620 (double layer) etch mask (30 µm thick) a. Layer 1 Pre spin 5 sec, 500 rpm Spin 45 sec, 1200 rpm Softbake 90 °C, 10 min b. Layer 2 Pre spin 5 sec, 500 rpm Spin 45 sec, 1200 rpm Softbake 90 °C, 12 min 132 Rehydration > 1 hour Exposure 750 mJ/cm 2 (15 mW/cm 2 , 50 sec) Development 4 min 20 sec using two baths Hardbake using hotplate 100 °C, 15 min Hardbake using vacuum oven 90 °C, 15 min 9. Reactive Ion Etching 100 W, 100 mTorr, 5 min (26 cycles) 10. Release devices by cleaning the surface with acetone and IPA, and then peel them carefully while immersed in DI water 133 APPENDIX G Metal T-Peel Test Structures Process Flow Figure G-1. Masks transparency. 1. Dehydration bake silicon wafers to remove moisture 110 °C, > 10 min 2. Deposit Parylene (~12 µm, 27 g of dimer) 3. Deposit adhesion promoter AdPro Plus® (if needed) 4. Pattern AZ 5214-IR for lift-off (2 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 2000 rpm Softbake 90 °C, 70 sec Exposure 50 mJ/cm 2 (15 mW/cm 2 , 3.33 sec) IR bake 110 °C, 45 sec Global exposure 300 mJ/cm 2 (15 mW/cm 2 , 20 sec) Development (AZ 351 1:4 dilution) 22 sec 5. Descum, O2 plasma 100 W, 100 mTorr, 1 min 6. Platinum deposition 2000 Å (4 runs of 500 Å) 7. Lift-off in warm acetone 40 °C (gentle scrub if necessary) 8. Pattern AZ 4400 sacrificial layer (4 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 4000 rpm Softbake 90 °C, 2 min Exposure 150 mJ/cm 2 (10 mW/cm 2 , 15 sec) Development 50 sec Hardbake 90 °C, 3 min 9. Descum, O2 plasma 100 W, 100 mTorr, 1 min 134 10. Deposit adhesion promoter AdPro Plus® (if needed) 11. Deposit Parylene (~12 µm, 27 g of dimer) 12. Pattern AZ 4620 (double layer) etch mask (30 µm thick) a. Layer 1 Pre spin 5 sec, 500 rpm Spin 45 sec, 1200 rpm Softbake 90 °C, 10 min b. Layer 2 Pre spin 5 sec, 500 rpm Spin 45 sec, 1200 rpm Softbake 90 °C, 12 min Rehydration > 1 hour Exposure 750 mJ/cm 2 (15 mW/cm 2 , 50 sec) Development 4 min 20 sec using two baths Hardbake using hotplate 100 °C, 15 min Hardbake using vacuum oven 90 °C, 15 min 13. Reactive Ion Etching 100 W, 100 mTorr, 5 min (26 cycles) 14. Release devices by cleaning the surface with acetone and IPA, and then peel them carefully while immersed in DI water 135 APPENDIX H IDEs Process Flow Figure H-1. Masks transparency. 1. Dehydration bake silicon wafers to remove moisture 110 °C, > 10 min 2. Deposit Parylene (~10 µm, 31.5 g of dimer) 3. Dry wafer under vacuum 60 °C, > 30 min 4. Pattern AZ 5214-IR for lift-off (2 µm thick) Pre spin 5 sec, 500 rpm Spin 40 sec, 4000 rpm Softbake 100 °C, 1 min Exposure 42 mJ/cm 2 (20 mW/cm 2 , 2.1 sec) IR bake 110 °C, 63 sec Hydration 3 min Global exposure 280 mJ/cm 2 (20 mW/cm 2 , 14 sec) Development (AZ 351 1:4 dilution) 18 sec 5. Descum, O2 plasma 100 W, 100 mTorr, 1 min 6. Platinum deposition 2000 Å (4 runs of 500 Å) 7. Lift-off Room temperature 8. Descum, O2 plasma 100 W, 100 mTorr, 1 min 9. Deposit adhesion promoter AdPro Plus® or EGDA (if needed) 10. Deposit Parylene (~10 µm, 31.5 g of dimer) 11. Dry bake Parylene wafers under vacuum 60 °C, > 5 min 12. Pattern AZ 4620 first etch mask (15 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 1000 rpm 136 Softbake 90 °C, 5 min Dehydration 3 min Exposure 480 mJ/cm 2 (22.5 mW/cm 2 , 21.3 sec) Development 1 min 20 sec 13. Deep Reactive Ion Etching, O2 plasma 150 cycles Table H-1. DRIE recipe with O2 flow rate optimized Parameter Deposition Etch ICP power (W) 700 700 RF power (W) 10 20 O2 (sccm) 1 60 C4F8 (sccm) 35 1 Ar (sccm) 40 40 Time (sec) 3 10 14. Strip remaining etch mask with acetone, IPA and DI water 15. Pattern AZ 4620 second etch mask (15 µm thick) Pre spin 5 sec, 500 rpm Spin 45 sec, 1000 rpm Softbake 90 °C, 5 min Dehydration 3 min Exposure 480 mJ/cm 2 (22.5 mW/cm 2 , 21.3 sec) Development 1 min 20 sec 16. Deep Reactive Ion Etching, O2 plasma 150 cycles (see Table H-1) 17. Release devices by cleaning the surface with acetone and IPA, and then peel them carefully while immersed in DI water

Abstract (if available)

Abstract Parylene provides many advantages as a material for thin-film implantable devices, however, there is still much work to be done to develop reliable microfabrication protocols for Parylene that result in robust devices. With all of this in mind, it is the aim of the present work to accomplish the following research aim: ❧ To advance the development of technologies for Parylene-based devices by providing an optimum process technique to selectively remove Parylene, and the improvement of Parylene-Parylene and Parylene-metal-Parylene adhesions for dry and wet environments. ❧ This work focuses on the improvement of microfabrication processes to overcome the main drawbacks of Parylene. Two specific challenges are examined in the detail: the need to selectively remove Parylene to create more complex structures with a high-fidelity transferred pattern, and the need to improve dry and wet adhesion of Parylene to itself and metallic layers to improve long-term performance. These problems are significant obstacles for the implementation of Parylene-based medical implants.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Development of implantable Parylene-based MEMS technologies for cortical applications

PDF

Parylene C bioMEMS for implantable devices with electrochemical interfaces

PDF

Development of micromachined technologies for neural interfaces

PDF

Strategies for improving mechanical and biochemical interfaces between medical implants and tissue

PDF

Parylene-based implantable interfaces for biomedical applications

PDF

Penetrating parylene neural probe array for dense, in vivo, chronic recordings

PDF

The electrochemical evaluation of Parylene-based electrodes for neural applications

PDF

Fabrication and packaging of three-dimensional Parylene C neural interfaces

PDF

Parylene-based biomems sensors for multiple physiological systems

PDF

A wireless implantable MEMS micropump system for site-specific anti-cancer drug delivery

PDF

Fabrication of ultrasound transducer and 3D-prinitng ultrasonic device

PDF

Rapid microfluidic prototyping for biological applications via stereolithographic 3-D printing: a journey

PDF

A medical imaging informatics based human performance analytics system

PDF

Scale-up of vapor-phase deposition of polymers: towards large-scale processing

PDF

Implantable bioMEMS drug delivery systems

PDF

Thin-film impedimetric sensors for chronic in vivo use: design and application to hydrocephalus treatment

PDF

The roles of surface and pore properties in wetting resistance for membrane distillation membranes

PDF

High-resolution data acquisition with neural and dermal interfaces

PDF

Multi-region recordings from the hippocampus with conformal multi-electrode arrays

PDF

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

Asset Metadata

Creator Ortigoza-Diaz, Jessica Lizbeth (author)

Core Title Development of fabrication technologies for robust Parylene medical implants

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 04/26/2019

Defense Date 02/28/2019

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

Tag adhesion,medical devices,microfabrication,OAI-PMH Harvest,Parylene

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Meng, Ellis (committee chair), Gupta, Malancha (committee member), Zhou, Qifa (committee member)

Creator Email jess.ortigoza@gmail.com,jortigoz@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-148509

Unique identifier UC11660725

Identifier etd-OrtigozaDi-7276.pdf (filename),usctheses-c89-148509 (legacy record id)

Legacy Identifier etd-OrtigozaDi-7276.pdf

Dmrecord 148509

Document Type Dissertation

Format application/pdf (imt)

Rights Ortigoza-Diaz, Jessica Lizbeth

Type texts

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

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

Repository Name University of Southern California Digital Library

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