University of Southern California Dissertations and Theses

Development of ceramic-to-metal package for BION microstimulator

(USC Thesis Other)

Development of ceramic-to-metal package for BION microstimulator

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content DEVELOPMENT OF CERAMIC-TO-METAL PACKAGE FOR BION® MICROSTIMULATOR by Guangqiang Jiang A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (BIOMEDICAL ENGINEERING) August 2005 Copyright 2005 Guangqiang Jiang R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. UMI Number: 3196824 Copyright 2005 by Jiang, Guangqiang All rights reserved. INFORMATION TO USERS The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleed-through, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. ® UMI UMI Microform 3196824 Copyright 2006 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Acknowledgements ii I would like to express my deepest gratitude to m y advisors, Dr. Tom Hedman and Dr. Joseph H. Schulman for their guidance throughout this study. I also thank members of my dissertation committee, who have spent time reading this thesis. They are Dr. David Z. D'Argenio and Dr. M ichael C.K. Khoo from the Department of Biomedical Engineering and Dr. Florian B. M ansfeld from the Department of M aterials Science and Engineering of University of Southern California. I would also like to thank the Alfred M ann Foundation for Scientific Research for the financial support of both, my education at University of Southern California and the research work. My sincere thanks to Jon Phil Mobley for his encouragement and consideration while I am pursuing my degree. I thank A ttila Antalfy, Brain Dearden, Dr. David Zhou, Dr. Kate Purnell, Chuck Byers, Roy Callam , Richard Nelson, Delta M ishler and Dr. Ross Davis for their help on some of the experiments. My thanks also go to Dr. Gary Schnittgrund, Donald Knaepple for their time reading and proofing this dissertation. I also feel a debt of gratitude to m y fellow students and personal friends who made me feel comfortable being here. Finally I would like to express my sincere gratitude to my family, especially to my wife, Qingfang Yao, for the encouragement and help. W ithout her help, this study would have been much more difficult. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. iii TABLE OF CONTENTS Acknowledgements................................................................................................ ii List of Tables.......................................................................................................... vii List of Figures........................................................................................................ viii Abstract.................................................................................................................... xiv Chapter 1: Introduction..................................................................................... 1 1.1 Background of the M icrostim ulator.................................................... 1 1.2 Goal of Present Research W o rk ............................................................ 6 1.3 R eferences................................................................................................. 8 Chapter 2: Background of the Microstimulator Package............................ 10 2.1 Requirements for the BION® M icrostimulator P ack ag e................. 10 2.2 Materials S elections................................................................................. 11 2.2.1 Titanium and Titanium Alloys Chosen for Metal Components ... 11 2.2.2 3Y-TZP Chosen for Ceramic C om ponents......................................... 13 2.3 Bonding Methods for Metal to Ceramic Jo in ts................................ 20 2.4 Objectives of the R esearch .................................................................... 23 2.5 R eferences................................................................................ 24 Chapter 3: Development of the 3Y-TZP to Ti-6A1-4V Braze Joint 28 3.1 Introduction................................................................................................ 28 3.2 Material Specification ............................................................................ 32 3.3 Experimental P rocedure.......................................................................... 34 3.3.1 Brazing P ro cess........................................................................................ 34 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. iv 3.3.2 Flexural Strength T e s t.............................................................................. 35 3.3.3 Scanning Electron M icroscopy - Energy Dispersive X -Ray M icro Analysis and X- Ray Diffraction A n aly sis.......................... 36 3.4 Results and D iscussions........................................................................... 37 3.4.1 Effect of Brazing Temperature on the Flexural Strength of Brazed Jo in ts............................................................................................. 37 3.4.2 Effect of Holding Time on the Flexural Strength of Brazed J o in ts........................................................................................................... 41 3.4.3 Interfacial Microstructures of Tini-50® Brazed J o in ts ................... 44 3.4.4 Interfacial M icrostructures of Nickel Brazed J o in ts ......................... 47 3.4.5 Bonding M echanism of Tini-50® Brazed J o in ts............................... 51 3.4.6 Bonding M echanism of Nickel Brazed J o in ts ................................... 53 3.4.7 Fracture of Brazed Jo in ts ........................................................................ 55 3.5 C onclusions................................................................................................ 59 3.6 R eferences.................................................................................................. 60 Chapter 4: Biocompatibility of 3Y-TZP to Ti-6A1-4V Brazed Join t 62 4.1 Introduction................................................................................................ 62 4.2 Materials and E xperim ents..................................................................... 64 4.2.1 SEM-EDS Analysis and X-Ray Diffraction A n aly sis...................... 64 4.2.2 Electrochemistry T e s t.............................................................................. 64 4.2.3 Immersion Test ........................................................................................ 66 4.2.4 Implantation and Gross P athology....................................................... 67 4.3 Results and D iscussions.......................................................................... 68 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. V 4.4 C onclusions.................................................................................................. 73 4.5 R eferences................................................................................................. 75 Chapter 5: Low Temperature Degradation of 3Y-TZP and Life Prediction of the BION® Microstimulator Package................ 76 5.1 Introduction............................................................................................. 76 5.2 Experimental M eth o d s......................................................................... 83 5.3 Results and D iscussions....................................................................... 84 5.3.1 Surface Finish of Different C eram ics............................................... 84 5.3.2 Microstructural Observation of Different C eram ics..................... 85 5.3.3 X-ray Diffraction R e su lts.................................................................... 87 5.3.4 Transformed Layer Thickness M easurem ent.................................. 89 5.3.5 Ceramic Aging in Liquid W ater and in S te a m ................................ 93 5.3.6 Arrhenius Factor Qio C alculation....................................................... 94 5.3.7 Bend Strength Reduction of In-Vitro Aged Ceram ic s .................. 98 5.3.8 Bend Strength Reduces with Increase of Transformed Monoclinic Layers ................................................................................. 102 5.3.9 Geometrical F a c to r................................................................................ 103 5.3.9.1 2mm Diameter Ceramic A Tubes Soaking in 127°C S te a m 103 5.3.9.2 3mm Diameter Ceramic A Tubes Soaking in 127°C S team 106 5.3.10 Hot Isostatic Pressed C eram ics.......................................................... 109 5.3.11 Protection Effect of Ion Beam Assisted Deposited AI2O3 C oating.................................................................................................... 115 5.3.12 Estimated Lifetime of BION® M icrostimulator P ack ag e............. 119 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. vi 5.4 C onclusions............................................................................................... 119 5.5 R eferences.................................................................................................. 121 Chapter 6: Conclusions.................................................................................... 125 Bibliography........................................................................................................... 128 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. vii LIST OF TABLES Table 2.1: Breakdown Potential of Metallic Biomaterials in Hank’s Solution and Repassivation in 0.9% NaCl (pH = 7.4)............................................... 12 Table 2.2: M echanical Properties of Metallic Biomaterials ..................................... 13 Table 2.3: Characteristics of Some Ceramics for Biomedical A pplications 13 Table 2.4: Organic Adhesives Used in Ceramic/metal Jo in in g ................................ 23 Table 3.1: The Chemical Composition and M echanical Properties of 3Y-TZP Ceramic and Ti-6A1-4V B a rs ........................................................................ 33 Table 3.2: Chemical Composition of Various Phases Observed in Figure 3 .8 (a ).................................................................................................................. 46 Table 3.3: Chemical Composition of Various Phases Observed in Figure 3.10 ... 48 Table 4.1: Chemical Composition of Phosphate Buffered Saline (Sigma Chemical Co.) Used in Both Electrochemistry and Corrosion Tests ... 65 Table 4.2: Nickel/titanium Ion Release Rates (pg/cm 2 /week) in 87°C PBS S o lu tio n .............................................................................................................. 69 Table 4.3: Chemical Composition of Various Phases Observed in Figure 4 .4 ...... 70 Table 4.4: Corrosion Potentials of Pure Nickel, TiNi-50® Clad Filler Material, Ti-6A1- 4V End Cap, Nitinol and Ti-6A1-4V End Cap Braze Joint in PBS Solution .................................................................................................... 72 Table 5.1: Arrhenius Factors (Qios) Calculated from Monoclinic Phase Rate C o nstants........................................................................................................... 97 Table 5.2: Arrhenius Factor Qio Values S um m ary...................................................... 99 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. viii LIST OF FIGURES Figure 1.1: The Package of the First Generation BION® Micro-stimulation M o d u le ............................................................................................................... 4 ( R ) Figure 1.2: Potential Clinical Applications Overview of the BION M icrostim ulator.............................................................................................. 5 Figure 2.1: The Concept of the Ceramic-to-metal P ack ag e..................................... 11 Figure 2.2: Representation of Stress-induced Transformation Toughening P ro cess.............................................................................................................. 16 Figure 2.3: High Zirconia Part of Zirconia-yttria Phase D iag ram .......................... 17 Figure 2.4: Fracture Toughness vs. Yttria C o n ten t..................................................... 17 Figure 2.5: Ceramic-to-metal Joining P rocesses........................................................ 21 Figure 3.1: The Flexural Strength Test S e t-u p ............................................................ 36 Figure 3.2: Ti-Ni Binary Phase D iag ram ...................................................................... 37 Figure 3.3: Flexural Strength of 0.05mm Tini-50® Brazed Joints Changes with Brazing T em perature.................................................................................... 39 Figure 3.4: Flexural Strength of Brazed Joints Changes with Brazing Temperature; Joints Brazed with 0.015mm Thick Pure Nickel Filler M e ta l................................................................................................................. 40 Figure 3.5: Flexural Strength of 0.05mm Thick Pure Nickel Brazed Joints Changes with Brazing T em perature........................................................... 41 Figure 3.6: Flexural Strength of Brazed Joints Changes with Holding Time; Joints Brazed with 0.05mm Thick Tini-50® Filler M aterial................. 42 Figure 3.7: Flexural Strength of Brazed Joints Changes with Holding Time; Joints Brazed with 0.05mm Thick Pure Nickel Filler Metal ........... 43 Figure 3.8: M icrostructure of A Joint Brazed at 1035°C for 5 Minutes with 0.05 mm Thick Tini-50® Filler Material (a) and A Dark Oxide Line at the Interface ( b ) ....................................................................................... 46 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 3.9: M icrostructures of Joint Brazed at (a) 1035°C for 300 Minutes and (b) 1100°C for 5 M inutes with 0.05mm Thick Tini-50® Filler M aterial ............................................................................................................ 47 Figure 3.10: M icrostructures of A Joint Brazed at 1010°C for 5 M inutes with 0.05mm Thick Pure Nickel Filler M e ta l..................................................... 48 Figure 3.11: M icrostructures of joints brazed at (a)1035°C for 5 minutes and (b) 1175°C for 5 minutes with 0.05mm thick pure nickel filler m e ta l.... 50 Figure 3.12: Microstructures of joints brazed at (a)1035°C for 5 minutes, (b)1035°C for 60 minutes and (c)1035°C for 30 minutes with 0.05mm thick pure nickel filler m e ta l...................................................... 51 Figure 3.13: Phases of Ti, Ti2Ni and N i2 Ti40 Identified on Titanium Side Fracture Surface from a Joint Brazed at 1035°C- 5 minutes with 0.05mm Thick Tini-50® Clad Filler M aterial........................................... 52 Figure 3.14: Phases of Z r0 2, Zr0.92Y0.0sO1.96, Yo.1 5 Zro.s5O 1 .93, Ti, Ti2Ni and Ni2 Ti4<3 Identified on Ceramic Side Fracture Surface from a Joint Brazed at 1035°C- 5 minutes with 0.05mm Thick Tini-50® Clad Filler M aterial................................................................................................. 53 Figure 3.15: Phases of Ti, Ti2Ni and N i2Ti40 Identified on Titanium Side Fracture Surface of Nickel Brazed Joint at 1035°C-5 m inutes 54 Figure 3.16: Phases of Z r0 2, Zro.92Yo.osO1.96, Yo.15Zro.g5O1.93, Ti, Ti2Ni and Ni2Ti40 Identified on Ceramic Side Fracture Surface of Nickel Brazed Joint at 1035°C-5 m in u tes............................................................ 54 Figure3.17: Fracture Lines in: (a) Joint Brazed at 1035°C-60 minutes; (b) Joint Brazed at 960°C-5 minutes; (c) Joint Brazed at 1100°C-5 minutes and (d) Close-up of Area A; All Joints Brazed with 0.05mm Thick Tini-50® Clad Filler M aterial................................................................ 56 Figure3.18: Fracture Lines in: (a) Joint Brazed at 960°C-5 minutes; (b) Joint Brazed at 1035°C-20 minutes; (c) Joint Brazed at 1175°C-5 minute and (d) Joint Brazed at 1100°C-5 minutes; All Joints Brazed with 0.050mm Thick Pure Nickel Filler M e ta l.............................................. 57 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. X Figure 4.1: Electrochemical Test Cell S e tu p ................................................................. 65 Figure 4.2: EG&G 273 Potentiostat Controlled by A Personal Computer Running the M 270 Electrochem ical Analysis S oftw are........................ 66 Figure 4.3 Locations of Implanted Stimulators in S h eep .......................................... 68 Figure 4.4: SEM Image of Interface of A TZP/Ti-6Al-4V Joint Brazed at 1035°C for 5 M inutes and Elemental Line Profile Across the Interface............................................................................................................. 70 Figure 4.5: Polarization Behavior of Pure Nickel, Tini-50® Clad Filler Material, Ti-6A1-4V End Cap, Nitinol and Ti-6A1-4V End Cap Braze Joint in PBS Solution at Scan rate of 0.5mV/s, Logarithmic Current Scale .... 72 Figure 4.6. Photomicrographs of (a) Sheep 005 (Examined after 163 days of Implantation): Anterior Right Microstimulator, Trichrome Stain; (b) Sheep 305 (Examined after 51 days of implantation): Anterior Left Microstimulator, H em atoxylin and Eosin (H & E); (c) Sheep 310 (Examined after 113 days of implantation): Posterior Right Microstimulator, H&E Stain; (d) Sheep 003 (Examined after 121 days of Implantation): Posterior Right, Siever-Munger Stain, Nerve ............................................................................................................................... 74 Figure 5.1: XRD Profile of A Ceramic Containing Both Tetragonal and Monoclinic P h a se s.......................................................................................... 80 Figure 5.2: (a) Two BION® M icro-stimulators Using B Ceramic Implanted in (R) Sheep and (b) Two BION Micro-stimulators Using C Ceramic Implanted in A R a t........................................................................................... 84 Figure 5.3: Ceramic A S u rface......................................................................................... 85 Figure 5.4: Ceramic B S u rface.......................................................................................... 85 Figure 5.5: Ceramic C S u rface.......................................................................................... 85 Figure 5.6: Grains of Ceramic A (Average Grain Size=0.50m icrons).................... 86 Figure 5.7: Grains of Ceramic B (Average Grain Size =0.59 microns) .... 87 Figure 5.8: Grains of Ceramic C (Average Grain Size =0.56 m icrons).... 87 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Figure 5.9: Plot of ln(ln (l/(l-f))) Versus lnt for the D eterm ination of n (slope) and lnb (Ordinate Origin) in the MAJ E q u atio n ...................................... 88 Figure 5.10: Plot of lnb Versus 1/T for the Determination of the Apparent Activation Energy for the T-M Phase Transformation Process ........................................................................................................................... 89 Figure 5.11: Transformed Layer Growth in A Ceramic in 107°C Aging T e s t 90 Figure 5.12: Transformed Layer Growth in A Ceramic in 127°C Aging T e s t.... 90 Figure 5.13: Transformed Layer Growth Rates of Ceramic A Soaking in 107 and 127°C S tea m ......................................................................................... 91 Figure 5.14: Transformed Layer Growth in Ceramic B in 87°C Aging T e s t 91 Figure 5.15: Transformed Layer Growth in Ceramic B in 107°C Aging T e s t 92 Figure 5.16: Transformed Layer Growth Rates of Ceramic B Soaking in 87 and 107°C W a te r.................................................................................................. 92 Figure 5.17: T-M Phase Transformation Occurred at Same Rate in Equal Temperature Liquid W ater and S tea m ................................................ 93 Figure 5.18: Monoclinic Phase Percentage (f) Increases over Tim e for Both A and B Ceramics in 67°C Soaking T e s t................................................. 95 Figure 5.19: M onoclinic Phase Percentage (f) Increases over Tim e for Both A and B Ceramics in 87°C Soaking T e s t................................................. 96 Figure 5.20: Monoclinic Phase Percentage (f) Increases over Tim e for Both A and B Ceramics in 107°C Soaking T e s t.............................................. 96 Figure 5.21: Monoclinic Phase Percentage (f) Increases over Time for Both Ceramics A and B in 127°C Soaking T e s t......................................... 97 Figure 5.22: Cracks in Transformed Layer of Ceramic B after 42 Days of Soaking in 87°C W a te r........................................................................... 99 Figure 5.23: Cracking Load of 127°C Aged 2mm Diam eter Ceramic A Brazed Cases in Three-point Bend T e s t............................................................... 100 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. xii Figure 5.24: Cracking Load of 2mm Diameter Ceramic B Tubes in Three-point Bend Test Degrades After Soaking in 87°C W a te r........................... 101 Figure 5.25: Self-destructed 2mm Diameter Ceramics B After Excessive T ransform ation........................................................................................... 101 Figure 5.26: Cracking Load Reduces with Increase of Transformation Layer Thickness (127°C Aged 2mm D iam eter Ceramic A Brazed Cases) 102 Figure 5.27: Cracking Load Reduces with Increase of Transformation Layer Thickness (2mm Diam eter Ceramic B Tubes Soaked in 127°C S team ).......................................................................................................... 102 Figure 5.28: Transformed Layer on Outer Surfaces of Regular Processed 2mm Diameter Ceramic A Tubes in 127°C Soaking T e s t........................... 104 Figure 5.29: Transformed Layer on Inner Surfaces of Regular Processed 2mm Diameter A Ceramic Tubes in 127°C Soaking T e s t.......................... 105 Figure 5.30: Transformed Layer Growth Rates of Regular Processed 2mm Diameter Ceramic A T u b e s ...................................................................... 105 Figure 5.31: Transformed Layer Growth at Inner Surfaces of 3mm Diameter Ceramic A in 127°C Aging T e s t............................................................. 106 Figure 5.32: Transformed Layer Growth at Outer Surfaces of 3mm Diameter Ceramic A in 127°C Aging T e s t............................................................. 10^ Figure 5.33: The T-M Transformation on Outer Surface W as Faster Than That 108 on Inner Surface Due To Geometrical Self-restriction.................. Figure 5.34: HIP Processed 2mm Diameter A Ceramics Has Slower T-M Transformation Rate Than Regular Processed 2mm Diameter Ceramics A ................................................................................................ HO Figure 5.35: Cracking Forces in Three-point Bending Test of Regular and HIPed 2mm Diameter Ceramic A Tubes Decreased in 127°C Soaking T e s t................................................................................................. Figure 5.36: Transformed Layer on Outer Surfaces of HTP Processed 2mm ^ Diameter Ceramic A Tubes in 127°C Soaking T e s t..................... R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. xiii Figure 5.37: Transformed Layer on Inner Surfaces of HIP Processed 2mm D iam eter Ceramic A Tubes in 127°C Soaking T e s t......................... 114 Figure 5.38: Transformed Layer Growth Rates of HIP Processed 2mm D iam eter Ceramic A T u b e s ................................................................ 114 Figure 5.39: Schematic of IB AD Experimental A pparatus................................... 116 Figure 5.40: XRD Profiles of AI2O3 Coated 3Y-TZP Ceramic Prior to and After 20, 85, 137 and 201 Hours of Soaking in 127°C S te a m 118 Figure 5.41: Comparison of Phase Transformation of T-ZrC> 2 w ith and without AI2O3 C o atin g ............................................................................................. 118 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. Abstract The goal of the present research work is to develop a bonding method to produce quality ceramic-to-metal joints suitable for long term implantable biomedical devices, such as BION® microstimulator where ceramic provides a transparent window for AC magnetic. A series of zirconia (3Y-TZP) ceramic to Ti-6A1-4V alloy braze runs have been carried out in vacuum using Tini-50® (50 wt.% Titanium / 50 wt. % Nickel), and pure nickel filler metals. The evolution of microstructures was studied using Scanning Electron M icroscopy (SEM) and Energy Dispersive M icroanalysis (EDS). Reaction products, such as T i2Ni and N i2Ti4 0 , were found on the fracture surfaces, as determined by X-Ray D iffraction (XRD) analysis. The interfacial reaction product Ni2Ti40 is responsible for bond development. Temperature and holding time strongly influenced joint strength. Three fracture models were identified and the exact failure model is a function of the brazing parameter. The biocompatibility of the bond has been evaluated by means of electrochemical testing, immersion testing and gross pathology. Results demonstrated good corrosion resistance and safe biocompatibility of Tini-50® brazed joints. The same results can be expected from implantation utilizing nickel-brazed ceramic-to-metal joints, because microstructures in nickel brazed joints are identical to that of Tini-50® brazed joints. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. It has been learned that 3Y-TZP ceramic are subject to moisture-induced tetragonal-to-monoclinic (T-M) phase transformation at body temperature resulting in degradation of its mechanical properties. To ensure the integrity of the BION® microstimulator package over the projected lifetime, in-vitro accelerated life tests (ALT) at various temperatures and in-vivo aging tests in sheep and rats have been conducted. A reliable method to predict the implantation lifetime of biomedical devices containing 3Y-TZP has been developed. The BION® microstim ulator package utilizing the current 3Y-TZP ceramic is adequate for up to 70 years of implantation in a human body. The successful development of 3Y-TZP ceramic to Ti-6A1-4V braze joint provides a promising packaging approach for many long-term implantable biomedical devices. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 1 Chapter 1: Introduction 1.1 Background of the Microstimulator Since the experiments of Luigi Galvani more than two centuries ago, it has been known that electrical current can be used to simulate muscle contraction. Achieving functionally useful movements of limb has been more elusive. During the past fifty years, engineers and neurophysiologists started to fully recognize and address the demands that such movements require but with little success. Now it appears that this may be achievable by im planting and integrating sensors, stimulators, and control systems within moveable limbs. It also appears feasible to graft robotic and electrophysiological instrumentation onto a biological system to repair it, or to restore functionality, however this will require many channels of bi directional information transmission. These channels must be installed, and function safely and reliably, in one of the most challenging environments conceivable: the human body. Based on the nature of the interaction between the patient and the technology, three classes of treatment [1] can be considered: 1. Therapeutic Electrical Stimulation (TES) - electrically induced exercise in which the beneficial effect occurs primarily offline as a result of trophic effects on muscles and perhaps the Central Neural System (CNS); R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 2 2. Neuromodulatory Stimulation (NMS) - preprogrammed stimulation that directly triggers or m odifies a function without ongoing control or feedback; 3. Functional Electrical Stimulation (FES) - precisely controlled muscle contractions that produce specific movements required by the patient to perform a task. There have been many attempts to use noninvasive interfaces with nerves and muscles based on skin-surface electrodes, and more recently, magnetic fields [2-4], W hile such approaches minimize expense and avoid the potential morbidity of surgery, they impose significant daily burdens on the patient, who must carefully don and maintain the interface as well as manage various external leads and connections in order to achieve the desired function [1]. This burden is particularly onerous for patients with the very sensorimotor disabilities that the interface is intended to address. Surgically implanted electrodes and stimulators can overcome the technical limitations and daily use problems required by external stimulation approaches, however the costs are high and may not be justified if clinical success is limited or uncertain. Thus, one of the key objectives has been the development of a technology that would combine the reliability and simplicity of use of an implanted interface with the low cost and low morbidity of a non-surgical approach [1]. As a minimally invasive technology, BION®s offer an advantage over functional-electrical-stimulation systems that require surgical implantation of a R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 3 stimulator with attached wires, or that apply electrical currents at the surface of the skin. The use of BION®s can enable therapists to apply currents directly to one or more muscles at widely varying levels of intensity, depending on the clinical need. W hen the BION® is fully charged, it’s electrodes discharge through a controlled current source producing well-controlled stimulation pulses of any intensity, which are able to stimulate a peripheral nerve or even a relatively large muscle (via its motor axons). BION®s can be modified to produce direct currents, which are used in experimental methods to promote soft tissue, nerve, and bone regeneration. They also serve as a flexible means for introducing electrical currents, in many forms of treatment. The idea of an injectable microstimulator for paralyzed muscles was first proposed in a paper [5] by Dr. W illiam Heetderks at National Institute of Health (NIH). In 1989, The Alfred Mann Foundation in Los Angeles, CA received a contract from NIH to develop a microstimulator using the principles shown in Heetderks paper. Pritzker Institute (PI) in Chicago, IL and Q ueen’s University (QU), Kingston, Ontario, Canada were subcontractors on that project. Joseph Schulman was the principal investigator. Phil Troyk and Gerald Lobe were co-principle investigators, respectively. In 1992, this group demonstrated the first working microstimulators. Each microstimulator consisted of a cylindrical glass capsule with a rigidly mounted electrode on each ends [6]. The most important requirement of the package is to protect the electronic circuitry from the deleterious effect of water. The hermeticity of the glass-to-metal seals in the BION depends on the chemical R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. bonding between the borosilicate glass (Kimbel N51A) and the native oxide on the tubular Tantalum metal electrode stem feedthrough, as shown in Figure 1.1 [7]. It is very difficult to obtain a reliable seal, as the hermeticity of the seal is very sensitive to the oxidation at the interface. The hermeticity of these seals can be lost if excess residual oxygen or longitudinal grooves are left in the Tantalum feedthrough metal from the drawing process. Besides these challenges, the produced seals which pass initial hermeticity testing, tended to fail catastrophically during prolonged soaking and temperature cycling in saline, due to the differences in the coefficient of thermal expansion between the glass (3.5 x 10'6 /°C) and Tantalum (6.5 x 10'6/°C) [1], Ta stem Schottky^diode wirebonds from ASIC top ferrite Ta tube final seal anodized sintered Ta( electrode 7 " 7 — \ A glass Ptfr alumina copper glass • / moisture glass Iridium bead washer PCB coil capsule " 9\ getter bead electrode V -W 1 2mm soldered coil terminations bottom ferrite 16mm Figure 1.1: The Package of the First Generation BION® M icrostimulator The mechanical integrity of the glass B IO N R package itself is another concern. The glass capsule is most susceptible to breakage by three-point bending over its long axis; and it will fail catastrophically if forces greater than 2 kg are applied at rigid fixation points [1]. Another problem which appeared later was the R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 5 susceptibility of the tantalum to glass seal to leak when the electrodes were subject to perpendicular cycling load. Though, the glass BION® package meets the requirements for some certain applications, where surrounded by a fair amount of muscle or soft tissue, its applications at severe-stress locations are greatly limited. BION®s have been in clinical trials since December 1999, when a stroke patient was implanted for the first tim e by Gerald Loeb at QU for shoulder sublaxation. Since then, BION® m icrostim ulators have been considered or utilized for more applications, such as sleep apnea, urinary incontinence, foot drop, knee osteoarthritis, wrist and finger contractures, pressure ulcers, and so on [8,9], Figure 1.2 is an overview of the BION®s potential clinical applications. In many cases, the implanted BION®s are subject to severe mechanical stresses and use of the glass-to- metal BION®s in these applications is not recommended. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. ' f ,.r ; ] : V '/'% "YVj I ' / / • i i n u . N - s » a f c W I . , , B I O N - i v « ™ t i . n n f ; ■ / ' • ■ - > V!> bleep A p n e a > i H I O N ’M V<w»I C o n tr o l . « BION 1 P rn c n n o n uf ; S lim iW crSuhluw tii.n u H IN - A ^i-H d ! ! < % ' f I - / i \ ' ’ f . ; i \ f ; A . \ . B U >V V G astric F.nifming i ‘...V / F t / ■ ■ 1 ' ' ! Bt<)!Vf * P ain C m trrn i / ' a > \ r ' BTON! V * A ssisted C««igb j / I I I O N >HK cbH biliU ilun s a f te r N ip S w r u e n • I * I I = B I O S 5" P re v e n tio n o f ItalMirtj!. j h l O N J f I riniirv O . n l i u r m r B IO N ' 1 C ard in resp iran m C onditioning ; ■ \ f \ j v / BI0 IN ’ ! ’ y Reh»l>iliii»ik?« ( | ' i' 4 . | I::.- ~\ a f t e r K n e e N « r » e r y i / i f f ! v 1 / 1 / . | & j & - # i c i t y O o f t t r t i l ; . ; BION 1 Prevention of ? Deep Vtsm T h rd b ilto sis , ■ \ ""vf : /- '■ ' j ’ f fMON1 / * 1 Prevent!op ; I t;:jr I I t ofI?p*>ftJrop. '> Figurel.2: Potential Clinical Applications Overview of BION® M icrostimulator [7] 1.2 Goal of Present Research Work The objective of the present work is to develop a more robust and reliable, biocompatible package for the second generation BION® microstim ulator [10]. This BION® microstimulator will be suitable for long-term implantation at almost any location in the human body to restore neural / muscular functionalities. A ceramic- to-metal brazed case package has been designed, developed, and tested for this purpose, and will be discussed. This work will focus on three main parts: 1). Development of a 3 Y-TZP ceramic to Ti-6A1-4V brazed joint 2). Biocompatibility of the 3Y-TZP ceramic to Ti-6A1-4V brazed joint R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 7 3). Life prediction of the BION® m icrostim ulator package Several hermetic packaging approaches have been developed for implantable devices, such as: glass-to-metal package and alumina-to-metal package, etc. As discussed above, glass-to-metal package is lack of reliability and robustness for long term implantation in server-stressed conditions. Alumina is often brazed to niobium to provide a hermetic package. Besides the poor machinability and high cost of niobium, alumina ceramic is brittle and possesses low mechanical properties. Above limits further miniaturization of hermetic packages. The successful development of 3Y-TZP ceramic to Ti-6A1-4V braze joint will provide a promising alternative packaging approach not only for the BION® microstimulator, but for all long-term implantable biomedical devices where a non-conductive, non-magnetic case is required for internal electronics. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 8 1.3 References 1. Loeb GE and Richm ond FJR, BION® Implants for Therapeutic and Functional Electrical Stimulation, Neural Prostheses for Restoration of Sensor and M otor Function, eds. J.K. Chapin, K.A M oxon and G. Gaal, CRC Pres, Boca Raton, FL 2000. 2. Galloway NTM, El-Galley RES, Sand PK, Appell RA, Russel HW, and Carlan SJ, Extracorporeal Magnetic Innervation Therapy for Stress Urinary Incontinence, Urology, 53, 1108, 1999. 3. Ishikawa N, Suda S, Sasaki T, Yamanishi T, Hosaka H, Yasuda K, and Ito H, Development of A Non-invasive Treatment System for Urinary Incontinence Using A Functional Continuous Magnetic Stim ulator (FCMS), Medical & Biological Engineering & Computing, November, 704, 1998. 4. Lin VW, Hsieh C, Hsiao IN, and Canfield J, Functional M agnetic Stimulation of Expiratory Muscles: A Noninvasive and New M ethod for Restoring Cough, Journal of Applied Physiology, 84,1144, 1998. 5. Heetderks W J, RF Powering of Millimeter- and Submillimeter-Sized Neural Prosthetic Implants, IEEE Trans. Biomed. Eng. 35: 323-327, May 1988. 6 . Loeb GE, Zamin CJ, Schulman JH, and Troyk P, Injectable M icrostimulator for Functional Electrical Stimulation, North Sea: Transducers and Electrodes, November, 1991. 7. http://ami .usc.edu/projects/Bion 8. Singh J, Peck RA and Loeb GE, Development of BION Technology for functional electrical stimulation: Hermetic Packaging, Proc. IEEE-EMBS (Istanbul, Turkey), 2001. 9. Dupont AC, Bagg SD, Chun S, Creasy JL, Romano C, Romano D, Waters RL, Wederich CL, Richm ond FJR and Loeb GE. Clinical Trials of BION™ Microstimulators, Proc. IFESS (LJUBLJANA, SLOVENIA), 2002. 10. Dupont AC, Loeb GE, Richmond FJR, Bagg SD, Creasy JL, Romano C, Romano D., Clinical trials of BION injectable neuromuscular stimulators, Proc IFESS (Cleveland, Ohio), 2001. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 11. Arcos I, Davis R, Fey K, M ishler D, Sanderson D, Tannacs C, Vogel MJ, W olf R, Zilberman Y, and Schulman J., Second-Generation Microstimulator. Artificial Organs, 26 (3): 228-231, 2001. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 10 Chapter 2: Background of the Microstimulator Package 2.1 Requirements for the BION® Microstimulator Package A reliable and functional package is essential for the implantable BION® microstimulator. First, the package of the BION® microstimulator needs to be biocompatible, as it’s intended use is for implantation. Second, it must be cylindrical in shape, and must keep the size as small as possible, for a minimally invasive implantation procedure. Third, the package has to provide a hermetic housing for the electronics module. Fourth, the package must be transparent to AC magnetic fields. Fifth, the device has to pass electrical current from the metal into surrounding tissue without exhibiting electrolytic corrosion, or contaminating the body with corrosion products. Another important requirement is the robustness of the package. This package design not only has to protect the inside electronics from the deleterious effects of water, but must also be able to resist damage from surgeon’s tools and other accidental stresses that might occur prior to, during and after the implantation. Finally, the BION® microstimulator is aimed for long-term implantation (up to 80 years). To meet all above requirements, a ceramic-to-metal brazed case concept has been considered. The ceramic-to-metal package is composed of a ceramic tube, a metal end cap at one end, and a metal ferrule at the other. These three components will be hermetically joined together with bonding media. The open-ended ferrule will be sealed at final assembly after the electronics are inserted. Figure 2.1 shows the concept of the ceramic-to-metal package. R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 11 Bonding M edia Ceramic Tube M etal Ferrule Metal End cap Figure 2.1: The Concept of the Ceramic-to-metal Package 2.2 Materials Selections 2.2.1 Titanium and Titanium Alloys Chosen For Metal Components Under normal conditions, the pH value of the human body fluid, i.e. a solution of about 0.9% NaCl, which is strongly buffered at 7.4 pH. In this medium the most corrosion resistant materials are titanium and its alloys, niobium, and tantalum followed by wrought and cast vitallium, and by stainless steel 316L [1], Breakdown potential measurements of various implant materials in H ank’s solution also showed a clear order of precedence. Usually, 100 m illiliter H ank’s solution contains 0.80 gram NaCl, 0.04 gram KC1, 0.014 gram CaCh, 0.01 gram of MgS0 4 .7H 2 0 , 0.01 gram M gCl2.6H 2 0 , 0.006 gram Na2HP0 4 .2H 2 0 , 0.006 gram KH2PO4, 0.1 gram Glucose and 0.035 gram NaHCCU A few m etallic materials used as biomaterials are limited to the following groups [2]: • Stainless steel, e.g. X 2CrN iM ol812 (316L) R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 12 • CoCr-based alloys (vitallium) in the as-cast condition, e.g. CoCr30Mo6, or in the as-wrought conditions, e.g. CoNi35Cr20 • Commercial pure (cp)-titanium and titanium alloys, e.g. Ti-6A1-4V • Cp-niobium, • Cp-tantalum, • Platinum and its non-corroding alloys, and • Iridium and non-corroding alloys W hile cp-Ti and Ti-6A1-4V had high breakdown potentials of 2.4 and 2.0 V, respectively, the values for stainless steel and CoCr-alloys (cast and wrought) are only 0.2 and 0.42 V, respectively, as shown in Table 2.1 [3], Table 2.2 shows typical values of the mechanical properties of metallic biomaterials [1]. Table 2.1: Breakdown Potential of M etallic Biomaterials in H ank’s Solution and ___________________ Repassivation in 0.9% NaCl (pH=7.4)___________________ Breakdown Repassivation time (msec) potential te t0.o s (V)____________ -0.5V +0.5V -0.5V +0.5V SS 316L 0.2 - 0.3 >72000 35 >>72000 >6000 CoCr (as cast) +0.42 44.4 36 » 6 0 0 0 >6000 CoNiCr (as wrought) +0.42 35.5 41 >6000 5300 Ti-6A1-4V +2.0 37 41 43.4 45.8 Cp-Ti +2.4 43 44.4 47.4 49 Cp-Ta +2.25 41 40 43 45 Cp-Nb +2.5 47.6 43.1 47 85 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 13 Table 2.2: M echanical Properties of Metallic Biomaterials E (GPa) YS (MPa) a f (MPa) %EL SS316L 210 450 250 40 CoCr (as cast) 200 500 300 8 CoNiCr (as wrought) 220 850 500 20 Ti-6A1-4V 105 900 550 13 Ti-5Al-2.5Fe 105 900 550 15 Cp-Ti 100 300 200 40 Cp-Ta 200 300 200 40 Cp-Nb 120 250 150 70 2.2.2 3Y-TZP Chosen for Ceramic Components Aluminum oxide (AI2O3) or alumina is one of the most versatile refractory ceramic oxides and finds use in a wide range of applications. It is the close packing of the aluminum and oxygen atoms within this structure that leads to its good mechanical and thermal properties. However, compared to zirconia, especially yttria stabilized zirconia, characterized by fine grained microstructures known as tetragonal zirconia polycrystals (TZP), alumina is weaker in both flexural strength and fracture toughness. Table 2.3 shows characteristics of some ceramics used for biomedical applications. Table 2.3: Characteristics of Some Ceramics used for Biomedical Applications Property Units Alumina2 Mg-PSZc Y-TZPC Ce-TZPd Chemical 99.9% ZrO 2 / Z r0 2 13 Z r0 2 / composition a i 2o 3/ 8 -1 0 mol mol % 12-14 mol MgO % MgO y 2o 3 % C e 0 2 Density g em '3 >3.97 5.74-6 >6 6 .1-6.2 Porosity % <0.1 - <0.1 - Grain size ftm 7 50b 0.4-0.6 1-3 R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 14 Flexural strength M Pa >500 450-700 900-1200 400-800 Y oung’s modulus GPa 380 200 200 200 Fracture toughness M P am 1 4 7-15 7-12 6-10 Thermal expansion coefficient K 1 8 x 10’6 7 -1 0 x l0 6 llx lO " 6 10-12 xlO" 6 Thermal conductivity W m K' 1 30 2 2 1.9 Notes: a. ISO 6474:1981 requirements for surgical grade alumina. b. 50 jiim cubic grains containing -0 .1 5 /xm metastable tetragonal precipitates. c. From reference 4. d. From references 5-11. Zirconia is a well-known polymorph that occurs in three forms: monoclinic (M), cubic (C) and tetragonal (T). Pure Zirconia is monoclinic at room temperature. This phase is stable up to 1170°C. Above this temperature it transforms into tetragonal and then into cubic phase at 2370°C. During cooling, a T-M transformation takes place at a temperature range of 970° to 1070°C. The phase transformation taking place during cooling is associated with a volume expansion of approximately 3 - 4 %. Stresses generated by the expansion initiates micro-cracks in the pure zirconia, which after sintering between 1500° -1700 °C, break into pieces at room temperature. The addition of ‘stabilizing’ oxides, like CaO, M gO, Ce02 and Y2O 3 to pure zirconia allow the generation of multiphase materials known as Partially Stabilized Zirconia (PSZ), whose microstructure at room temperature generally consists [12] of cubic zirconia as the major phase, with monoclinic and tetragonal zirconia R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission. 15 precipitates as the minor phase. The development of zirconia as an engineering material was marked by Garvie et al. [13], who in their paper ‘Ceramic Steel?’ showed that tetragonal metastable precipitates finely dispersed within the cubic matrix were able to be transformed into the monoclinic phase when the constraint exerted on them by the matrix was relieved, i.e. by a crack advancing in the material. We learned that a 3 - 4% volume expansion takes place during the T-M phase transformation. In that case, the stress field associated with expansion due to the phase transformation acts in opposition to the stress field that promotes the propagation of the crack. An enhancement in mechanical strength and toughness is obtained, because the energy associated with crack propagation is dissipated both in the T-M transformation and in overcoming the compression stresses due to the volume expansion. A schematic representation of this phenomenon is shown in Figure 2.2. The development of such tetragonal metastable precipitates may be obtained by the addition of sufficient MgO (about 8 mol %) to ZrOi- PSZ can also be obtained in the Zr0 2 -Y203 system (Figure 2.3). However, within this system it is also possible to obtain ceramics formed at room temperature with a tetragonal phase only, called TZP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 process zone go"€<fc£% °o§toc>8 o ovoo o o o o P o ° ° . X . « . c j transform ing untrans formed transformed nnrticle particle particle ^ Figure 2.2: Representation of Stress-induced Transformation Toughening Process [14] (Energy of the advancing crack is dissipated in phase transformation and in overcoming the matrix constraint by transforming grains) TZP materials, containing approximately 2 - 3 % mol Y2O3, are completely constituted by tetragonal grains with sizes of the order of hundreds of nanometers. The fraction of T-phase retained at room temperature is dependent on a variety of factors: the size of the grains, on the yttria content, on the grade of constraint exerted on them by the matrix, etc. M echanical properties of TZP ceramics, such as fracture toughness (Figure 2.4), depend on such parameters. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 2500 2000 1500 P S Z 1000 TZP 500 2 0 6 8 10 4 Y 2Oj content (mol%) Figure 2.3: High Zirconia Part of Zirconia-yttria Phase Diagram [15] (Shaded regions indicate commercial PSZ and TZP composition and processing temperatures) Tetragonal Phase Volume Fraction i.a 0.2 0.0 0.1 06 B « 8 fib s TETRAGONAL C U B I C V2O3 content (mol%) Figure 2.4: Fracture Toughness vs. Yttria Content [16] Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 Presently, zirconia is in clinical use for total hip replacement (THR), but developments are in progress for application in other medical devices. Concerns have been raised when zirconia is used as biomaterial because of two of its properties: radioactivity and phase transformation. a) Radioactivity can be caused by the presence of traces of Th and U-series elements, originating from the raw material mineral used for the manufacturing of zirconia feedstock. Values of under lOBq per kg of zirconia give rise to radiation activities of anywhere from 2 to 11 mSv/yr, which is well below the natural radiation exposure of 20 mSv/yr. The effective doses of radiation for various zirconia products are between 0.13 and 0.53 mSv/yr, as com pared to 0.16 mSv/yr for high purity Alumina. For these specific activities, the estimated effective dose is well below the dose limit of 1.0 mSv/yr defined for the general public [17]. Zirconia can therefore be regarded as causing no potential radiation threat to the bearer of the prosthesis. In the case of the BION® microstimulator, the amount of ceramic materials in the package is so small compared to other zirconia products, such as femoral head for total hip replacement (THR), that the threat to the patient is negligible. b) Phase transformation: As it has been discussed, the T-M phase transformation in the ceramic during a loading process will enhance its mechanical strength and fracture toughness, and is so called stress-induced phase transformation. On the other hand, a progressive spontaneous Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 transformation of the metastable tetragonal phase into the monoclinic phase while water is present will degrade the mechanical property of the zirconia, known as ‘aging’, or so called moisture-induced T-M phase transformation. This behavior is well known in the presence of water vapor at temperatures above 200°C [18, 19]. W hile the degradation of Y-TZP ceramic at 37°C has not been fully assessed, it has been preliminarily investigated. For example, Thompson and Rawlings [20] assessed the mechanical properties of Y-TZP in a simulated body solution (Ringers solution). The result was that Y-TZP demonstrated a significant strength decrement when aged for long periods in Ringer’s solution and was therefore unsuitable as implant material. Drummond [21,22] reported that yttria-stabilized zirconia demonstrated low-temperature degradation at 37°C with a significant decrement in strength in as a short period as 140 days in deionized water, saline, or Ringer’s solution. Similar observation have also been reported by others [23-28], where yttria-stabilized zirconia demonstrated a strength decrement in water vapor, room temperature water, body temperature water, R inger’s solution, hot water, boiling water, in-vivo and post in-vivo aging. Even though dense, polycrystalline, Y-TZP degrades in an aqueous environment at elevated temperatures; the mechanisms of T-M phase transformation in TZP are still in dispute. Some models have been proposed to explain the spontaneous T-M transformation in TZP which are based on the formation of zirconium hydroxides [18, 19, 29] or yttrium hydroxides [30] promoting phase Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 transition for local stress concentration, or variation of the yttrium/zirconium ratio. In contrast, Kimel and Adair showed in a recent report [31] on the corrosion of a commercial Y-TZP powder, that the m echanism for T-M transformation was attributable to yttrium leaching. M echanical properties of Zirconia relate to its fine-grained, metastable microstructures. The stability of this structure during the lifetime of TZP components is the key point to attain the expected performances. It is extremely important to be able to predict the lifetime of the TZP-containing device. However, since there is no reliable method available to predict the performance of TZP ceramics, this is one of the areas that this research will focus on. 2.3 Bonding Methods for Metal To Ceramic Joints With the possible metallic and ceramic candidates in mind, it is time to consider how to obtain a hermetic joint between them. Joining dissimilar materials has been long investigated, not only for ceramic-metal, but also for a number of other combinations, such as glass/metal and glass-ceramic/metal. Glass/metal joining, for example, dates back to the invention of the electric light bulb in the early 1800’s. Most of the time, joining different materials is not an easy task. Atoms, ions, or molecules in materials of different classes - ceramics, metals, or polymers - are joined together in different ways, and are therefore characterized by particular combinations of physical-chemical and mechanical properties. A variety of metal- ceramic joining processes, whose use mainly depend on the base materials to be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 bonded and the application of the joined component, are either currently available or under development. Ceramics and metals can be joined together by mechanical, direct, or indirect processes (Figure 2.4) [32, 33]. M tc ta n iic a l .1in n in g Direct Joining SIKTAL/CERAM IC JO IN IN G Figure 2.5: Ceramic-to-metal Joining Processes M echanical joining approaches generally can’t provide hermetic sealstDKi], Typical mechanical strengths of the joints vary from 10 to 50 M Pa. Stress concentration areas (especially in the ceramic counterpart) and design limitation are among the major restrictions of those methods [32-34]. Indirect and direct joining refers to whether an intermediate material has been used, such as a filler material to promote physical or chemical bonding between the counterparts. Charge or mass transfer can take place in either case [32], Solid-state diffusion is an example of a direct ceramic-to-metal joining process. This process can present difficulty in obtaining a hermetic seal from a small bonding area where two m irror surfaces are often required. A slight imperfection will compromise the bond. Because a high Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 pressure load is often needed during the diffusion bonding process, which can deform the metal components and result in out-of-spec dimensions of as-joined parts. This is not a mass production process. Adhesive joining using organic interlayers offers suitable mechanical strength below 250 °C (Table 2.4), but the joint is not hermetic suitable for implantable applications. Adhesive joint often degrades and fails in both hermeticity and strength in relatively short period of time. Brazing is by far the most widely used joining process when mechanically reliable vacuum tight joints are required. Brazing allows low-cost large-scale joining of intricate geometries and is not necessarily restricted to flat surfaces. Brazing has been defined by the American W elding Society as a joining process that takes place above 450°C using filler metals or alloys which flow by capillary forces and whose melting temperature is lower than the solidus temperature of the base materials. Filler metals or alloys should adhere to the surface of the base materials. Besides the hermeticity and strength requirements, other consideration needed to be given to the bonding process include: biocompatibility and resistance to corrosion. This requirement suggests that further biocompatibility investigation of the ceramic-to-metal seal should be performed, because there is no data in the bulk of literature indicating a safe utilization of such ceramic-to-metal seal for implantation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 3 Table 2.4: Organic Adhesives Used in Ceramic/metal Joining [32] Adhesive Setting Maximum h ikin^ Temperature (”C) Epoxy Hot 170-220 Polyurethane Hot t cold! 120-180 Silicon Cold. 180-220 Cyanocritate Cold 150-230 Elastomer Hot 90-110 2.4 Objectives of the Research W hile attempting to develop a braze technique to provide a hermetic, strong, and reliable ceramic-to-metal seal for the implantable BION® microstimulator, this work will also investigate some fundamental aspects of brazing 3Y-TZP ceramic to Ti-6A1-4V using both Tini-50® and nickel filler metals. The bonding mechanisms of 3Y-TZP to Ti-6A1-4V joints, the evolution of the microstructure in the joint and the effects of brazing parameters (brazing temperature and holding time) on the strength of the brazed joints will be investigated. Because nickel is one of the elements in the braze filler material, there is concern about potential allergic reactions. The electrochemical stability of nickel in the ceramic-to-metal seal will be evaluated in phosphate buffered saline (PBS) by meaning of electrochemistry testing. Long-term passive corrosion test will be carried out with the ceramic-to-metal braze joint to obtain quantitative information of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 4 nickel ion release rate in PBS solution. The data will be discussed and compared to Nitinol, a famous biocompatible material that will be tested as control. Implantation in animals will also be carried out to evaluate the biocompatibility of the BION® microstimulator package. To ensure the integrity of the BION® microstim ulator package over the projected lifetime, in-vitro accelerated life tests (ALT) at various temperatures, and in-vivo aging tests in sheep and rats will be conducted. This work will also try to develop a reliable method to predict the lifetime of a biomedical device containing 3Y-TZP ceramic. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 5 2.5 References 1. Leyens C and Peters M, Titanium and Titanium Alloys: Fundamentals and Applications, 457-498. W iley-VCH, 2003. 2. Zitter H and Plenk HJ, The Electrochemical Behavior of M etallic Im plant M aterials as An Indicator of Their Biocompatibility, J. Biomed. Mater. Res. 21, 881, 1987. 3. Fraker AC, Ruff AW, Sung P, Orden ACV, and Speck KM, Surface Preparation and Corrosion Behavior of Titanium Alloys for Surgical Implants. T i’80 Science and Techn., Plenum Press, 2447, 1980. 4. Piconi C and M accauro G, Review: Zirconia as A Ceramic Biomaterial, 1-25, Biomaterials 20, 1999. 5. Tsukuma K and Shimada M, Strength, Fracture Toughness and Vickers Hardness of Ce0 2 -stabilized Tetragonal Z r0 2 Polycrystals (CeTZP), Journal of Materials Science; Volum e 20, Pg 1178-1184, 1985. 6 . Schneider SJ (ed), Ceramics and Glasses, Engineered Materials Handbook, Volume 4, ASM International, 1991. 7. Schwartz MM, Handbook of Structural Ceramics, McGraw-Hill Publishers, USA, 1992. 8. Guillou MO, Henshall JL, H ooper RM and Carter GM, Indentation Fracture Testing and Analysis, and its Application to Zirconia, Silicon Carbide and Silicon Nitride Ceramics, Journal of Hard Materials; Volume 3, Pg 421-434, 1992. 9. W hitney ED, Ceramic Cutting Tools, Pg 166, Chapter 7, Noyes Publications, USA, 1994. 10. Griffin EA, Mumm DR and M arshall DB, Rapid Prototyping of Functional Ceramic Composites, American Ceramic Society Bulletin; Volume 75, 65-68,1996. 11. Cambridge Materials Selector, A M aterials Database - Version 2.04, Granta Design Utd, UK. 1997. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 12. Subbarao EC, Zirconia - An Overview. In: Heuer AH, Hobbs LW, editors. Advances in Ceramics, vol. 3:1-24. Science and Technology of Zirconia. Amsterdam: Elsevier, 1981. 13. Garvie RC, Hannink RH, Pascoe RT., Ceramic Steel? Nature; 258:703-4, 1975. 14. Bulter EP, Transformation Toughened Zirconia Ceramics, Mat Sci Tech: 1:417-32, 1985. 15. Scott HG, Phase Relationship in Zirconia-yttria Systems, J. M ater Sci. 10:1527-35, 1975. 16. Lange FF, Transformation Toughening. Part 3 - Experimental Observations in the Zr0 2 -Y203 system. J, M ater Sci; 17:240-6, 1982. 17. Porstendorfer J, Reineking A, W illert HG, Radiation Risk Estimation Based on Activity M easurements of Zirconium Oxide Implants, J. Biomedical Materials Research, 32:4, 663-667, 1996. 18. Sato T, Shimada M., Control of the Tetragonal-to-monoclinic Phase Transformation of Yttria Partially Stabilized Zirconia in Hot Water, J. M ater Sci; 20:3899-992, 1985. 19. Sato T, Shimada M, Transformation of Yttria-doped Tetragonal Zr02 Polycrystals by Annealing in Water, J Amer Ceram Soc, 68(6) 356-9, 1985. 20. Thomson I and Rawlings RD, M echanical Behavior of Zirconia and Zirconia-toughened Alumina in a Simulated Body Environment, Biomaterials, 11: 505-8, 1990. 21. Drummond JL, In Vitro Aging of Yttria-stabilized Zirconia, J Amer Ceram Soc, 72(4): 675-6, 1989. 22. Drummond JL. Effects of In Vitro Aging of M agnesia-stabilized Zirconia, J Amer Ceram Soc, 75(50) 1278-80, 1992. 23. Drouin JM, Cales B., Yttria-stabilized Zirconia Ceramic for Improved Hip Joint Head In: Anderson OH, Yli-Urpo A, editors. Bioceramics 7. London: Butterworth-Heinemann Publ., 387-94, 1994. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 7 24. Shimizu K, Oka M and Kumar P, Time-dependent Changes in the M echanical Properties of Zirconia Ceramic, J Biomed M at Res., 27:729- 34,1993. 25. Fujisa A, Shimotoso T, M asuda S, Makinouchi k., The D evelopm ent of Zirconia Balls for THR with A High M echanical Strength, Low Phase Transformation. In: Kokubo T, Nakamura T, Miyaji F, editors. Bioceramics 9. Amsterdam: Elsevier Science Publ., 503-6, 1996. 26. Cales B, Stefani Y., M echanical Properties and Surface A nalysis of Retrieved Zirconia Femoral Hip Joint Heads After An Im plantation Time of Two to Three Years. J M ater Sci M ater Med, 5: 376-378, 1994. 27. M urray MGS, Pryce AW and Stuart JW. A Comparison of Zyranox Zirconia Femoral Heads Before and 1-1/2 Years After Implantation. In: Ravaglioli A, editor. 4th Euro-Ceramics, vol. 8, Bioceramics. Faenza (Italy): Faenza Editrice Publ. 37-44, 1995 28. M urray MGS, Pryce AW. A Physical, Chemical and M echanical Evaluation of A Retrieved Zyranox Zirconia Femoral Heads Before and 11/2 Years After Implantation. In: Trans, 5th W orld Biom aterials Congress. Toronto, Canada, 1996. 29. Yoshimura M, Noma T, Kawabata K, Somiya S., Role of H 2O on the Degradation Process of Y-TZP, J M ater Sci Lett, 6:465, 1987. 30. Lange FF, Dunlop GL, Davis BI. Degradation During A ging of Transformation-toughened Zr0 2 -Y203 Materials at 250°C, J Amer Ceram Soc. 69:237, 1986. 31. Kimel RA and Adair JH. Aqueous Degradation and Chem ical Passivation of Yttria Tetragonally Stabilized Zirconia, J Amer Ceram Soc., 85[6], 403-408, 2002. 32. Martinelli AE, Diffusion Bonding of Silicon Carbide and Silicon Nitride to Molybdenum. PhD Thesis, McGill University, M ontreal, Canada, 1996. 33. Suganuma K, Mater. Res. Soc. Symp. Proc. 314, 51, 1993. 34. Klomp JT, With GD, Mater. & Manuf. Proc. 8, 2: 129, 1993. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 Chapter 3: Development of the 3Y-TZP Ceramic to Ti-6A1-4V Braze Joint 3.1 Introduction It has been discussed in the previous chapters that yttria stabilized zirconia tetragonal polycrystals (Y-TZPs) have several advantages, such as high flexural strength and fracture toughness, over other ceramics due to the transformation toughening mechanisms. They also demonstrated good biocompatibility [1], Titanium alloys, especially Ti-6A1-4V (wt.%), also have excellent mechanical properties, high corrosion resistance and favorable biocompatibility. Since both of these materials have been widely used in biomedical applications, 3Y-TZP ceramic and Ti-6A1-4V have been chosen as the materials for BION® microstimulator package. Brazing is by far the most widely used joining process when mechanically reliable water-tight joints are required. Filler alloys are classified into two categories: active and non-active. Active filler alloys include the presence of an active element, such as Ti, Al, Hf, Zr or Nb in their composition. Non-active filler alloys often require prior metallization of the ceramic substrate to provide for enough wetting, so an interface (usually reactive) is formed. Physical vapor deposition (PVD), chemical vapor deposition (CVD) or mechanical metallization can be used to deposit metallic films, such as molybdenum, manganese, tungsten or their combination onto ceramic surfaces prior to brazing, however, this metallization process can complicate the brazing process and makes quality control of the joint Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 9 more difficult. Brazing with active filler materials is a relatively simple method. Active filler alloys for direct ceramic/metal brazing should possess some essential features in order to improve interfacial microstructure, such as [2-5]: 1. M elting point or melting range compatible with those of the base materials; 2. M oderated fluidity at the brazing temperature, promoting capillarity and uniform distribution over the joint, but preventing infiltration into base materials (both metals and ceramics); 3. Homogeneous composition and stability to minimize constituent separation or segregation upon melting and solidification (brazing cycle); 4. Thermodynamic compatibility with the base metal surfaces, thus promoting wetting; 5. Limited tendency to brittle phase formation (usually intermetallics) and 6 . Compatibility with the working temperature, mechanical loading, environment, and intended life span of the joint. Some earlier studies have selected active filler alloy brazing as the joining route for zirconia to Ti-6A1-4V, and have successfully brazed the zirconia to Ti-6A1- 4V with Ag-Cu series filler materials. For example, zirconia was brazed to Ti-6A1- 4V alloy by a Cu-40Ag-5Ti (wt.%) alloy at 850° - 890°C for 5 minutes, leading to formation of Cu2(Ti, A l^O at the interface [6], This mixed oxide reaction layer was the weakest area within the joint and resulted in an average tensile strength of 150 (±45MPa). Elemental Ag and Cu were found within the joint area. Agathopoulos Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 0 [7] et al. analyzed the joint of Zirconia to Ti and Ti-6A1-4V brazed with Ag-35Cu-1.65Ti filler alloy at 900°C in vacuum. In TZP/Ti joints, a complex interface microstructure comprising Ti2 0 , C uiT^O , Ag[Cu], the formation of a solid solution and Cu-Ti intermetallics was dependent on the time duration of the various process steps. In TZP/Ti-6Al-4V joints, the evolution of microstructure with brazing time depends on diffusion of Ti, Al and V to the interface forming Cu2(Ti,Al,V)4 0 , and high joint strengths of over 400 M Pa have been obtained by utilizing short brazing times. Zirconia has also been brazed to titanium using Ag-30Cu-10Sn (wt.%) filler material [8], Enhanced wetting was obtained by coating the ZrC> 2 surface with Ti prior to brazing by RF sputtering or electron beam evaporation. ZrC> 2 to Ti joints were obtained with shear strength of 140 MPa. The Ag-Cu filler material features two particular characteristics: (a) excellent gap filling ability and (b) extensive evaporation of Ag at >1100°C under vacuum. Although this method seems very promising in terms of interfacial strength, it might meet some resistance with regard to biomedical applications as far as the possible toxicity effect of Cu is concerned. Thus, the use of noble metals, such as Pt, Au or Ag has been investigated. Correia et al. have thoroughly investigated the potential use of Pt as interlayer (25 pm) between TZP and Ti-containing blocks within a wide temperature range [9], Though the chemical reaction is strong, the interfaces are rather weak, actually failing at the interface between the TZP and Pt-rich zone. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 Since Au dissolves up to 11% (atomic %) Ti at 1115°C, a coating technique (i.e. sputtering) was employed to produce thin interlayers (< l|im ) in the form of single or multi-layer coating either on TZP or on Au thin (25pm ) foils in order to retain Au/Ti atomic ratio between 99/1 to 92/8. These interlayers were used to investigate the possibility of production of TZP-TZP joints by brazing at 1065- 1100°C under vacuum [10]. The most characteristic problem in this system is the poor wettability resulting in an inadequate filling of the gap between the two work pieces, which is more pronounced if ceramic surfaces bear some surface curvature. In TZP/(Au-25 pm)/Ti joints (under vacuum), the infinite supply of Ti to the interface through the liquid Au results in a continuous interface without gaps. However, the Ti-Au intermetallics formed at the interface do not hinder Ti-diffusion towards the ceramic to form unfavorable Ti oxides and TixA uy intermetallics. Silver was also tested in form of thin foil (35pm) for the production of TZP/Ti joints at 980°C under vacuum. The interfacial reaction seemed stronger than in Au case. A thick zone of a Ti-Oxide (assigned to T i302) formed at the interface featuring large holes. In order to avoid the involvement of foreign elements, zirconia-titanium and zirconia-titanium alloy joints were attempted by diffusion bonding [7,9,11]. At bonding temperatures below 1245°C in vacuum or Ar environment, TZP/Ti interfaces failed, while no evidence of chemical reactions was observed, except the zirconia blackened, which was attributed to oxygen depletion resulting in Zr02-X . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 2 Further increasing the bonding tem perature to 1328°C and 1494°C, interfaces are strong enough to barely survive ceram ographic preparation. Despite the strong reaction, the joints featured negligible fracture strength, probably due to the Ti embrittlement caused by the enlargem ent o f Ti-grains by two orders of magnitude. Diffusion bonding TZP and Ti with a Zirconium interlayer (30 pm ) inserted between has been attempted by Agathopoulos et al. [11], however no successful result has been reported. Recently, Lasater [12] disclosed a method to produce hermetical sealing zirconia to titanium alloys using titanium -nickel alloy filler material. Fey and Jiang [13] discovered that zirconia could be joined to titanium alloys using pure nickel brazing filler material. However, no report has been seen on the bonding mechanism and interfacial microstructures of this ceramic-to-metal system. Zirconia has been brazed to Ti-6A1-4V under vacuum using titanium-nickel clad and pure nickel filler metals. The effects of brazing parameters, such as brazing temperature and holding time on flexural strength are studied and discussed. The bonding mechanism and interfacial microstructures are investigated by SEM, EDS and XRD analyses. 3.2 Material Specification Both 3Y-TZP ceramic and Ti-6A1-4V bars used in this work are 3.2 mm (height) by 3.2 mm (width) by 15.5mm (length). The chemical composition and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 3 mechanical properties of 3Y-TZP ceramic and Ti-6A1-4V alloy are shown in Table 3.1. Table 3.1: The Chemical Composition and M echanical Properties of 3Y-TZP Ceramic and Ti-6A1-4V Bars Property Units 3 Y-TZP3 Ti-6Al-4Va Chemical composition Y 20 3 - 5.15% A120 3 - 0.246% N a20 - 0.004% S i0 2 - 0.006% Fe20 3 - 0.003% Z r0 2 - Balance C - <0.08% Fe - <0.25% N 2-< 0 .0 5 % O2- < 0 .2 % Al - 5.5-6.76% V - 3.5-4.5% H2( b a r ) - <0.0125% Ti - Balance Density g e m '3 6.05 4.42 Hardness GPa 13.75 HV 300-400 Flexural strength M Pa 1343 - Tensile strength M Pa - 1000 Thermal expansion b 200-500 °C 11.0 300-500 °C 9.4 10'6 /°C 600-1000 °C 12.0 650-900 °C 14.9 200-1000 °C 11.5 200-1000 °C 11.9 Fracture toughness K jc MPa*m1 /2 13 - a — Data are from material analysis and certificates provided by the vendor b — Data obtained from testing per standard test method for linear thermal expansion of solid materials with a vitreous silica dilatometer, ASTM document number: E228-95. ( R ) Two types of brazing filler materials are used in this study: Tini-50 , and plain nickel. Tini-50® brazing filler material is lam inated in a sequence of nickel/titanium/nickel, with a Ti/Ni weight ratio of 1/1, and has a liquidus and solidus of 1032 °C and 960 °C respectively. The nickel filler metal is a minimum purity of 99.6 wt. % nickel. Both braze filler materials have been cut into 3.2 x 3.2 mm squares to be used for brazing. The thickness of Tini-50® filler material is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 4 0.05mm; while two thicknesses, 0.015 and 0.050mm, of nickel filler metals were studied. Rectangular bars, instead of hollow tubes as used in the microstimulator packages, are used in the present work for the purposes of easy assembly for brazing process and to meet the requirement for the later flexural strength test per ASTM - C 1161-02c (Standard Test M ethod for Flexural Strength of Advanced Ceramics at Ambient Temperature). The conclusions will be applicable to hollow tubes. 3.3 Experimental Procedure 3.3.1 Brazing Process To reveal the effect of brazing temperature on flexural strength, joints were fabricated at various brazing temperatures while all other process parameters were kept constant. The brazing process was carried out in a vacuum furnace at a vacuum level of less than 1.3xl0'5 mbar. A brazing fixture was designed to align the samples, and to provide 0.5 M Pa of pressure on the joint throughout the brazing process. The assembled braze samples were heated up to 600°C at a ramp rate of ~ 20°C/min and held at temperature for 10 minutes in order to guarantee a uniform temperature distribution over the braze samples and fixture. Then, the set up was heated again to the preset temperatures at the ramp rate of 12.5°C/min, and held at these temperatures for various periods of time. After brazing, the samples were held Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 5 in vacuum until cooled to 200°C, then were quenched to room temperature using inert gas. The braze process was carried out in a furnace equipped with a programmable logic controller (PLC) for fully automatic and precise operation of the processing cycles. The temperature accuracy in the furnace throughout the brazing process was ±1°C. M inimum of six specimens were produced at each brazing condition. One specimen from each condition was subjected to cross sectioning and SEM-EDS analysis. The remaining specimens from each condition were subjected to four-point flexural strength testing. Some fractured surfaces from the flexural strength test were subject to SEM-EDS and XRD analyses. 3.3.2 Flexural Strength Test Flexural strength test according to ASTM - C 1161-02c (Standard Test Method for Flexural Strength of Advanced Ceramics at Ambient Temperature) was carried out on the joints brazed at various brazing parameters. The flexural strength test set-up is shown in Figure 3.1. The outer span is 20mm and inner span is 10 mm. Instron Model 5544 was used to apply load with a crosshead speed of 0.5mm/minute. The data collection rate was set to 50/second. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 6 ,_ p J Ti-6A1-4V Supporting base Loading fixture M 3Y-TZP Supporting ixture Figure 3.1: Flexural Strength Test Set-up 3.3.3 Scanning Electron Microscopy - Energy Dispersive X-Ray Micro Analysis and X-Ray Diffraction Analysis Braze joint cross-sections were prepared for microanalysis by abrasive sectioning and polishing using a Struers sectioning/polishing machine. Hitachi S-570 Scanning Electron M icroscope (SEM) equipped with the Integrated Microanalyzer for Images and X rays (IMIX) software from Princeton Gamma-Tech (PGT) was utilized for the SEM examination and Energy Dispersive M icroanalysis (EDS). The acceleration voltage for SEM analysis and EDS was 20 KV and the probe size for the EDS analysis is estimated to be less than 2 pm. An X-Ray Diffractometer (Siemens D-500) was used to scan the fracture surfaces to reveal the reaction products at the braze interfaces. Cu K a radiation was used for the X-Ray diffraction analysis. X-Ray pattern processing and identification Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 7 was carried out with Jade 6.0 software from Materials Data, Inc. The typical current and accelerated voltage setups for XRD analysis are 30mA and 40KV. 3.4 Results and Discussions 3.4.1 Effect of Brazing Temperature on Flexural Strength of Brazed Joints Titanium and nickel have a eutectic point of 942°C with a Ti/Ni weight ratio of 72/28, as shown in Figure 3.2. At temperatures above this eutectic point, a liquid alloy of Ti and Ni with the eutectic ratio forms, and the activity of liquid Ti enhances dramatically from its solid state and promotes its reaction to ceramic. The liquidus and solidus of Tini-50® are 1032 °C and 960 °C, respectively. Atomic percent nickel 80 90 100 80 30 1800 1700 1600 Liquid 1500 1400 9 1300 2 2 1200 (Ni) 2 a . 1 1 0 0 E £ 1000 T i N i m ) 800 700 600 90 100 80 30 40 50 60 70 20 0 10 Ti Weight percent nickel N i Figure 3.2: Ti-Ni Binary Phase Diagram [10] Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 8 A series of brazing runs were carried out with brazing temperatures close to this range using 0.05mm thick Tini-50® preforms. At a load pressure of 0.5 MPa, braze joints have been fabricated at temperature range o f 930 to 1110°C, and using a holding time of five minutes. Six samples were produced at each brazing temperature. Upon completion of the fabrication steps, five joints were flexural tested. The effects of various temperatures on flexural strength are plotted in Figure 3.3, and the data show that the flexural strength of the brazed joints increases first as the temperature increases up to 1035°C, then decreases as the brazing temperature further increases. The joints brazed at 1035°C demonstrated a maximum flexural strength of 195.0 M Pa, with a standard deviation of 44.3 MPa. Joints produced below 945°C demonstrated zero flexural strength. Analysis of variance test (ANOVA) suggests that the collected data is statistically significant (significance level P=0.0001). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 9 300 — 250 (0 Q . s ~ 200 U ) £ 150 5) | 100 x 0 > 50 0 920 940 960 980 1000 1020 1040 1060 1080 1100 1120 Brazing Temperature (°C) Figure 3.3: Flexural Strength of 0.05mm Thick Tini-50® Brazed Joints Changes with Brazing Temperature Braze joints were also fabricated using 0.015mm thick pure nickel filler material at a temperature range of 930 to 1110°C, and with a holding time of five minutes. A minimum of six samples were produced at each brazing temperature. U pon completion of the fabrication steps, at least five joints were flexural tested to failure. Figure 3.4 indicates that the flexural strength of the brazed joints increases first as the temperature increases to 1010°C, then decreases as the brazing temperature further increases. Joints brazed at 1010°C demonstrated a maximum flexural strength of 175 MPa, with a standard deviation of 34 M Pa. Joints produced below 945°C demonstrate no bonding and no flexural strength. ANOVA test suggests that the collected data is statistically significant (P=0.0095). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 0 300 -S' 250 ( 0 Q . s ~ 200 4 -* o > £ 150 H - * C O 2 100 3 X _ 0 ) U - 50 0 940 960 980 1000 1020 1040 1060 1080 1100 1120 1140 1160 Brazing Temperature (°C) Figure 3.4: Flexural Strength of 0.015mm Thick Pure Nickel Brazed Joints Changes with Brazing Temperature A series of brazing runs were also carried out with 0.05mm thick pure nickel preforms. At a load pressure of 0.5 MPa, joints have been fabricated at brazing temperatures of 930 to 1175°C and holding for five minutes. Six samples were produced at each brazing temperature. Upon completion of the fabrication step, five joints were flexural tested to failure. Flexural strength data showed that joints brazed with 0.050 mm nickel demonstrated a maximum strength of 227 M Pa at a brazing temperature of 1105°C, with a standard deviation of 46.6 MPa. The flexural strength test results are shown in Figure 3.5. Joints fabricated below 960°C demonstrated no flexural strength. ANOVA test also suggests that the collected data is statistically significant (P=0.027). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 300 ^ 250 n T C L §- 200 £ 4 -* O ) § 150 w CO 5 100 x < D u- 50 0 900 950 1000 1050 1100 1150 1200 Brazing Temperature (°C) Figure 3.5: Flexural Strength of 0.05mm Thick Pure Nickel Brazed Joints Changes with Brazing Temperature 3.4.2 Effect of Holding Time on Flexural Strength of Brazed Joints Six specimens were fabricated at each holding times of 5, 20, 60, 120 and 300 minutes at 1035°C using 0.05m m Tini-50® clad performs. The load pressure for all the samples was 0.5MPa. Flexural strength test were performed to evaluate the influence of holding time to joint strength. Figure 3.6 shows the strength changes with the increase of holding time. The joints brazed at 1035°C holding for 20 minutes illustrated a m aximum flexural strength of 252 M Pa with a standard deviation of 59 MPa. ANOVA test also suggests that the collected data is statistically significant (P=0.033). -------------- --------- - - - I - 1- 1 — ♦- —♦— _ ? 1 1 ■ ■ Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 2 350 300 S. 250 S C < D i- ” 150 ( 0 3 ® 100 U . 50 0 Figure 3.6: Flexural Strength of 0.05mm Thick Tini-50® Brazed Joints Changes with Flolding time 5m in 20m in 60 min 120 min 300 min Six specimens were fabricated at each holding times of 5, 20, 60, 120 and 300 minutes at 1035°C using 0.05mm thick pure nickel preforms. Figure 3.7 shows that the joint strength varied with the increase of holding time. The load pressure was also 0.5MPa for all the joints tested. The joints brazed at 1035°C holding for 60 minutes had a maximum flexural strength of 257.3 M Pa with a standard deviation of 67.7 MPa. ANOVA test again suggests that the collected data is statistically significant (P=0.024). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 3 350 300 "S ' | 250 .C g 200 a > « 150 ( 0 x 100 a ) ui 50 0 Figure 3.7: Flexural Strength of 0.05mm Thick Pure Nickel Brazed Joints Changes with Holding Time Figures 3.4 and 3.5 reveal that for joints brazed using 0.05mm thick pure nickel filler metal a higher brazing temperature, consequently a longer time above the liquidus, is needed to provide maximum flexural strength than for joints using 0.015mm thick pure nickel filler metal. This is because thicker filler metal takes longer time to be consumed and form a liquid alloy to bond to the ceramic. It is also found that, a longer holding time is needed for brazed joints using 0.05mm thick pure nickel filler metal than that using same thickness Tini-50® filler material, to reach maximum flexural strength, as shown in Figures 3.6 and 3.7. As described in section 3.2, Tini-50® brazing filler material is laminated in a Nickel/Titanium/Nickel sequence. W hen assembled with ceramic and Ti-6A1-4V base metal, there are three interfaces where titanium and nickel have direct contact to form Ti-Ni eutectic alloy. W hile a joint brazing with pure nickel, there is only one 5 min 20 min 60 min 120 min 300 min Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 4 interface where titanium directly contacts nickel. So, shorter time is needed for the Tini-50® brazed joint to consume the filler material, form adequate liquid Ti-Ni alloy and to react with the ceramic to form a strong bond than that for brazed joint using pure nickel. 3.4.3 Interfacial Microstructures of Tini-50® Brazed Joints It is obvious that interfacial microstructures of Tini-50 filler metal brazed joints are brazing parameter dependent. Figure 3.8 (a) is a backscatter electron (BSE) image of a joint brazed at 1035°C and held for 5 minutes. Table 3.2 shows the chemical composition of the various phases observed in the brazed joint area. Phase 1 is in the ceramic bulk and consists of zirconium, yttrium, hafnium and oxygen. EDS on phase 2 revealed that the atomic ratio of Ti/Ni in this phase is very close to 2:1, with very small amounts of aluminum and vanadium. According to the Ti-Ni binary phase diagram [10], the phase is presumed to be Ti2 Ni intermetallics ((Ti,Al,V)2Ni to be exact), which is verified by XRD analysis. Phase 3 is a Ti rich layer, where small amounts of A1 and V can be found. Phase 4 has a similar composition as phase 2, but with more A1 and V. Because phase 4 is adjacent to Ti- 6A1-4V alloy where is the source of A1 and V. Phase 4 is presumed to be Ti2 Ni ((Ti,Al,V)2 Ni to be exact) as well. Phase 5 contains 1.60 atomic % (1.83 wt. %) nickel, besides the elements from the Ti-6A1-4V base metal: titanium, aluminum and vanadium. Phase 6 contains more nickel (9.58 atomic % (11.57 wt. %)) than Phase Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 5 5. Phase 7 contains Ti (84.60 at. %), A1 (12.76 at. %) and V (2.45 at. %), and no nickel is found. Phase 7 is determined to be a-T i phase. The extent of nickel diffusion in the Ti-6A1-4V bulk is dependent on the brazing process. Nickel diffused into the Ti-6A1-4V bulk deeper at a higher brazing temperature and/or longer holding time. Nickel does not uniformly distribute in the Ti-6A1-4V alloy but preferably exists along the phase boundaries. At higher magnification, a dark layer can be distinguished at the interface of the ceramic bulk and T i2 Ni phase, as seen in Figure 3.8(b). This layer is arbitrarily presumed as an oxide phase Ni2Ti40 (Ni2(Ti,Al,V)40 to be exact) and has been identified with XRD analysis. The thickness of this layer is less than 1.0 pm and its chemical composition is difficult to obtain due to the spatial resolution limit of the EDS. It is found that joints brazed at higher temperatures (above 1010°C) or long holding time (greater than 5 minutes at 1035°C) featured two characteristics: 1) a Ti2 Ni layer at the ceramic to filler material interface, and 2) a Ti rich layer next to the T i2 Ni phase. As the filler metal is laminated in Ni/Ti/Ni sequence, liquid alloying of Ti and Ni with composition close to the eutectic point (Ti 72 wt. % and Ni 28 wt. %) starts to form at Ti and Ni interfaces when the joint is heated above the Ti-Ni eutectic point (942°C). The sandwiched Ti and Ni are gradually consumed with increase of brazing temperature or holding time. As elemental Ti has a high affinity for oxygen, melting Ti tends to segregate at the ceramic to filler metal interface. Ti2 Ni forms when the joint cools below 984°C (the liquidus of Ti2Ni) and phase Ni2Ti40 also forms at adjacency to the Y-TZP ceramic. If the joint is held at a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 6 temperature above 984°C, the nickel besides combining with segregating Ti at the ceramic surface, it also diffuses away from the filler metal area into the Ti-6A1-4V bulk preferably along the phase boundaries and a Ti rich layer forms. This Ti rich layer is presumed to be a-T i phase and the thickness of this layer increases with increase of brazing temperature and holding time. For instance, the Ti rich layer was about 10-pm thick in a joint brazed at 1035°C for 5 minutes. The same layer was about 10 times thicker in a joint produced at 1035°C for 300 minutes (Figure 3.9(a)). About a 25-pm thick Ti rich layer was observed in a joint brazed at 1100°C for 5 minutes (Figure 3.9(b)). Oxide line TO microns Figure 3.8: (a) Microstructure of A Joint Brazed at 1035 °C for 5 M inutes and (b) A Dark Oxide Line at the Interface Table 3.2: Chemical Composition of Various Phases Observed in Figure 3.8(a) Elem ents Phase 1 Phase 2 at. %( wt. at. %(wt. %) %) Phase 3 at. %(wt. %) Phase 4 at. %(wt. %) Phase 5 at. %(wt. %) Phase 6 at. %(wt. %) Phase 7 at. %(wt. %) Zr 30.82(67.35)— — — — — — Y 2.43(5.18) — — — — — — H f 0.48(2.08) — — — — — — 0 66.27(25.39)— — — — — — Ti 64.71 87.92 62.13 96.33 84.39 84.60 (61.05) (91.70) (59.66) (96.48) (83.15) (89.43) Mi 31.25 29.33 1.60 9.58 IN I (36.14) (34.52) (1.83) (11.57) 2.63 9.78 6.05 0.89 2.15 12.76 A1 ~ d -40) (5.75) (3.28) (0.51) (1-19) (7.59) Y — 1.40 2.30 2.48 1.20 3.89 2.65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 7 Zone ZrO? (1.41) (Ti, Al, V)2Ni (2.55) a-T i [A1,V] (2-54) (Ti, Al, V )2 Ni ( 1-18) (4.09) P- Ti[Ni,Al,V] Ti[Ni,Al,V] (2.98) a-T i [A1,V] Figure 3.9: M icrostructure of Joints Brazed at: (a) 1035 °C -300 minutes; (b) 1100 °C - 5 minutes with 0.05mm Thick Tini-50® Clad Filler Material 3.4.4 Interfacial Microstructures of Nickel Brazed Joints It is found that interfacial microstructures in nickel brazed joints are also brazing param eter dependent. Figure 3.10 is a backscatter electron (BSE) image of a joint brazed at 1010°C and held for 5 minutes using 0.05mm thick nickel filler metal. Table 3.3 shows the chemical composition of the various phases observed in Figure 3.10. Phase 1 is the ceramic bulk. Phases 2 and 4 are Ti2Ni intermetallics and Phase 3 is a transitional phase containing chemical com position close to Ti-Ni eutectic composition with small amounts of Al and V. Phase 5 is [3 — Ti phase, where Ni is rich and Phase 6 is the a -T i phase. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 3.10: M icrostructure of A Joint Brazed at 1010 °C for 5 M inutes with 0.05mm Thick Pure Nickel Filler Metal Table 3.3: Chemical Composition of Various Phases Observed in Figure 3.10 Elements Phase 1 at. %( wt. %) Phase 2 Phase 3 at. %(wt. %) at. %(wt. %) Phase 4 at. %(wt. %) Phase 5 at. %(wt. %) Phase 6 at. %(wt. %) Zr 28.83(62.02) — — — — — Y 4.18(8.76) — — — — — Hf 1.03(4.32) — — — — — O 65.98(24.90) — — — — — T i 59.57 80.26 62.69 75.55 82.10 1 1 (54.69) (81.45) (59.59) (77.51) (87.02) 39.12 8.26 31.25 8.03 0.90 Ni ----- (44.01) (10.28) (36.42) (10.10) (1.18) A 1 0.19 8.11 4.50 10.75 13.88 A l (0.20) (4.64) (2.40) (6.21) (8.29) V 1.13 3.36 1.58 5.67 3.12 V (1-10) (3.63) (1-59) (6.18) (3.51) Zones Z r02 (Ti, Al, V)2Ni a-T i [A1,V] (Ti, Al, V )2 Ni P-Ti[Ni,Al,V] a-T i [A1,V] Two Ti2 Ni intermetallics layers were observed in Figure 3.10, with one layer located at the ceramic to filler metal interface, and the other existing at the filler metal to Ti-6A1-4V base metal interface. This provides strong evidence o f titanium segregation, because initially, no titanium is present at the ceramic side interface. Titanium and nickel start to react and form Ti-Ni liquid alloy at the filler m etal to Ti- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 9 6A1-4V interface when the joint is heated. Titanium then diffuses and reaches the ceramic-to-filler metal interface as the brazing process goes on. More and more titanium segregates at this interface, while the joint is kept at the elevated brazing temperature due to titanium ’s high affinity to oxygen from the ceramic. Ti2 Ni and Ni4Ti20 form at the ceramic to filler metal interface and T i2 Ni forms at the Ti-6A1- 4V to filler metal interface during the cooling period. In the joints that are brazed at higher temperature (>1010°C at 5minutes) or longer holding time (>5 minutes at 1035°C), the Ti2 Ni phase at the Ti-6A1-4V base metal side disappears, as shown in Figures 3.11 (a) and (b) and Figures 3.12 (a) (b) and (c). This is because nickel seems to diffuse away from this Ti2Ni phase and into the higher titanium-contained phase, Ti-6A1-4V bulk. Figures 3.11 (a) and (b) are microstructural photos of joints brazed at 1035 °C for 5 minutes and 1175 °C for 5 minutes with 0.050mm thick pure nickel. Figure 3.12 (a) (b) and (c) are microstructures of joints brazed at 1035 °C for 5 minutes, 1035 °C for 60 minutes and 1035 °C for 300 minutes with 0.050mm thick nickel, respectively. Another observation is that the T i2 Ni phase is much denser at relatively low brazing temperature and short holding time; it becomes sparser while the joints are kept at brazing temperature for a prolonged period of time. It is also found that grains of Ti-6A1-4V base metal increase dramatically with both increase of brazing temperature and holding time. The average spacing of plate-like a -T i phase was about 2 microns for the Ti-6A1-4V base metal prior to brazing process; it widens to about 8 microns after the 1035°C for 5 minutes brazing process. After Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 0 1175°C for 5 minutes or 1035°C for 60 minutes brazing process, the width is about 15 microns. The width is about 30 microns after brazing at 1035°C and holding for 300 minutes. Normally, titanium alloys with coarser grains result in poorer mechanical properties. However, in the case of ceramic to titanium braze joints; this detrimental effect was not evident because the strength of coarse-grained titanium base metal is still higher than the brazed joints. Figure 3.11: Microstructures of Joint Brazed at (a) 1035 °C for 5 minutes and (b) 1175°C for 5 minutes with 0.050mm Thick Pure Nickel Metal As it has been observed in Tini-50® brazed joints, Ti-rich phase (presumed to be a -T i phase) can also be found in nickel-brazed joints that were produced at relatively high brazing temperature or long holding time. This layer grows with the increase of brazing temperature and holding time. The thickness of this layer is about 50 microns in the joint brazed at 1175°C for 5 minutes. This layer hasn’t appeared in joints brazed at 1035°C for 5 minutes. However, this layer grew to about 40 microns thick when the joints were held for 60 minutes at 1035°C and almost 90 microns thick after 300 minutes of holding tim e at 1035°C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 Figure 3.12: M icrostructures of Joints Brazed at (a) 1035 °C for 5 minutes, (b) 1035 °C for 60 minutes and (c) 1035 °C for 300 minutes with 0.050mm Thick Pure Nickel Metal 3.4.5 Bonding Mechanism of Tini-50® Brazed Joints XRD analysis identified phases of Z r0 2, Yo.1 5 Zro.85O 1.93, Zr0.93Y0.0sO1.96, Ti2Ni, Ni2Ti40 and Ti on the ceramic side fractured surfaces and Ti2 Ni, Ni2 Ti40 and Ti on titanium side from a joint brazed at 1035°C- 5 minutes using Tini-50® braze material, as shown in Figures 3.13 and 3.14. These results suggest that when the joint is heated above the eutectic point of Ti-Ni alloy, the liquidized TiNi alloy chemically reacts with the ceramic, seizes oxygen from Z r0 2 and leads to the formation of Z r0 2_ x and N i2Ti40 at the bonding interface upon cooling. Although the nature of the Z r0 2_ x region of the Zr-O phase diagram for temperatures <1500°C Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 2 has not been reliably established [15], at 1000°C “x” was found to range between 0 and 0.17 [16]. However, based on the experimentally observed thickness of the oxide layer (<1 pm) and the darkened layers of ceramic (for example, about 8m m at 1010°C), the value of “x” was estimated to be at the order of 10'3. This is way below the limitation of EDS. a 2000 Figure 3.13: Phases of Ti, Ti2 Ni and N i2 Ti40 Identified on Metal Side Fracture Surface from A Joint Brazed at 1035°C- 5 M inutes with 0.05mm Thick Tini-50® Clad Filler Material Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 1000 a 7 5 0 J j 500 250 0 j I L 1 1 \ I i 1 h 4 S - 0 2 2 4 > Z fii i i ' l V O . 0 3 0 1 . ' I t ■ m r i c - f r i Z B - c - '/ r i H j m O x i d s • t o - I S 4 2 ' Z r Q 2 • ( j - . i d e 3 D - t 4 6 8 > Y O . 1 5 Z r O . S 5 G l 8 3 • Y t t r i u m Z i r c o n i u m O x i d e , ! 1 . , , , , 1 7 2 - 0 4 4 3 > t J i 2 T i 4 d • N i c k * T i t a n i u m O x i d e 1 1 1 1 8 - D 8 © 8 > T O N i • N i c k e l T i t a n i u m i ...............................I ................. I . . . ............... 1 I 4 4 - 1 2 9 4 . T t - T i t a n i u m 30 40 50 60 70 80 90 100 2-Theta(°) Figure 3.14: Phases of Z r0 2, Zr0.92Y0.0gO1.96, Yo.15Zro.g5O 1 .93, Ti, Ti2Ni and Ni2Ti40 Identified on Ceramic Side Fracture Surface from a Joint Brazed at 1035°C- 5 Minutes with 0.05mm Thick Tini-50® Clad Filler Material 3.4.6 Bonding Mechanism of Pure Nickel Brazed Joints Fractured surfaces from the pure nickel brazed joint showed almost identical XRD profiles as those from the Tini-50® brazed joints, as shown in Figures 3.15 and 3.16, and on all of the fractured surfaces, no elemental nickel was found. The bonding mechanism for the joint brazed with nickel is considered to be the same as the Tini-50® brazed joints. The Ti-6A1-4V base metal is the only source of titanium for the nickel brazed joints; while both Ti-6A1-4V alloy base metal and the clad filler metal are the titanium source for the Tini-50® brazed joints. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 4 o 40 50 60 70 60 90 100 Figure 3.15: Phases of Ti, Ti2Ni and Ni2Ti40 Identified on Titanium Side Fracture Surface of Nickel Brazed Joint at 1035°C-5 M inutes tZRC2N101 .M 011 7.'Q'?-Ni-D1 <?Tr0V0 W .___f t ------- — .a „ A ________________ ________ J ^ ' U L ..................../"L -'vvA .................. ..../ A ______ __A -M * 70-1408> YO.lSZrO 8501 93- Yttrium Zirconium 0*id« j | 1 7I-D443> NiiTiiO • | 1S-D S98> T2Ni - Nicks! Titanijrr. I I I .........................................................................I .......................................... I , 44-1264 Ti-Tsaswjrr. 30 40 50 60 70 80 90 100 2 -T h e ta O Figure 3.16: Phases of Z r0 2, Zr0.92Y0.0sO1.96, Yo.15Zro.s5O1.93, Ti, Ti2Ni andN i2 Ti40 Identified on Ceramic Side Fracture Surface of Nickel Brazed Joint at 1035°C-5 Minutes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 3.4.7 Fracture of Brazed Joints Three failure modes were observed for Tini-50® brazed joint. Figure 3.17 (a) represents failure mode A, where the fracture line goes along the ceramic to filler metal interface and into the Ti2Ni phase region. This type of failure provides the highest facture strength. Figure 3.17 (b) shows failure mode B. Here, the fracture line lies right at the ceramic to filler metal interface. This type of failure demonstrates low fracture strength and is often observed in joints brazed at low temperature or short holding time, where inadequate chemical reaction occurred. Figures 17 (c) and (d) show failure mode C, where fracture occurs both, in the ceramic bulk and in the a-T i layer. This failure often happens in joints brazed at high temperatures. All failure modes have also been observed in nickel brazed joints. Figures 18 (a), (b) and (c) are fracture interfaces from nickel brazed joints. Joints produced at low brazing temperature often fail in failure mode A, i.e. fractures right at the ceramic to filler metal interface at relative low stress due to inadequate reaction, as shown in Figure 18 (a). W ith increase of brazing temperature or holding time, adequate interface reaction occurs and a strong joint is obtained. Joints brazed at higher brazing temperature, or longer holding time, often fail in failure mode B, where fracture happens both at the ceramic to filler metal interface and in the filler metal bulk, as shown in Figure 18 (b). W ith further increase of brazing temperature or holding time, joints fracture in both the ceramic bulk and the brittle a -T i layer, as shown in Figure 18 (c). This type of failure resulted in relative low strength. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 Fracture line 3Y-T/11 Ceramic 3Y-TZP ! . ■ ■ f Ti-6A1-4V 1M B Fracture line A l Ti-6A14V _ t " ^ (c) Figure 3.17: Fracture Lines in: (a) Joint Brazed at 1035°C-60 minutes; (b) Joint Brazed at 960°C-5 minutes; (c) Joint Brazed at 1100°C-5 minutes and (d) Close-up of Area A; All Joints Brazed with 0.05mm Thick Tini-50® Clad Filler Material Fracture line N 3Y-TZP ' * # '- A \ *** » • * > .Jl'-Z’ f**"" "V"'- ’ V ** Ii-oAMV Ceramic ( b k H m m 3 Y - T / . p S h H B Ccrami Ceramic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 7 Figure 3.18: Fracture Lines in: (a) Joint Brazed at 960°C-5 minutes; (b) Joint Brazed at 1035°C-20 minutes; (c) Joint Brazed at 1100°C-5 minutes; All Joints Brazed with 0.050mm Thick Pure Nickel Filler Metal Generally, besides the properties of the ceramic, metal component and braze filler material used, the joint strength is thought to be influenced by at least three other basic factors: the extent of interfacial chemical reaction, the degree of residual stresses in the joint, and the formation of brittle reaction products. These factors are dependent on brazing parameters. The activity of active elements, such as Ti, increases with the increase of brazing temperature. It dramatically increases from solid state to liquid state. The interfacial reaction becomes sufficient with increase of brazing temperature and holding time initially, the interfacial bonding strength increases. Therefore, the flexural strength gradually increases. However, with further increase of brazing temperature, a-T i phase layer forms and grows in the joint area, which is brittle and it will have negative impact on flexural strength. So, the flexural strength is lower when the joint is overheated. Moreover, though Y-TZP ceramic has a similar Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 coefficient of thermal expansion (CTE) as Ti-6A1-4V over the room temperature to brazing temperature range, the difference in CTE between the Z r0 2 and the reaction products (TiiNi, Ni2Ti40 and a-T i) is significant, producing localized residual stress in the Z r0 2 near the brazed surface, resulting in fractures in ceram ic bulk, as seen in Figures 3.17 (c) and (d) and 3.18 (c). Therefore, the flexural strength decreases with increase of brazing temperature. All these factors together determined the m axim um flexural strength characteristics for both Tini-50® and pure nickel brazed joints as a function of brazing temperature as shown in Figures 3.3 through 3.5. Similarly, it was determined by these three factors as well, that the flexural strength changes with the variation of holding time, as previously shown in Figures 3.6 and 3.7. At short holding times (less than 20 minutes for 0.05m m thick Tini-50® brazed joints and less than 60 minutes for 0.05mm thick pure nickel brazed joints), the chemical reaction extent increases with the increase of holding time, the flexural strength of the joints increases. Though the brittle layer (a-T i) thickness increases will reduce the flexural strength, it is not predominate at the beginning. As the brittle layer thickness further increases with the increase of holding time, its detrimental effects on flexural strength gradually emerges, the strength of joint goes downward as the holding time further increases. Also, the thicker the a -T i layer is, the higher the localized residual stress due to CTE mismatch, which reduces the flexural strength of the brazed joints. Overall, a maximum flexural strength characteristics Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 can be observed as a function of holding time for both Tini-50® and pure nickel brazed joints. 3.5 Conclusions Strong joints can be produced with both Tini-50® and pure nickel filler metals. Brazing parameters have a strong influence on the flexural strength of both types of joints. This has been demonstrated experimentally and the effects of the three main factors were explained. Further, three failure modes were identified, and the exact failure mode has been shown to be a direct function of the brazing parameters. The highest fracture strength of 257 MPa with a standard deviation of 67 M Pa was obtained with joints brazed at 1035°C for 60 minutes using 0.050mm thick pure nickel filler metal. The Ni2 Ti40 phase that formed at the ceramic to metal interface is responsible for the bond development. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 0 3.6 References 1. Gogotsi GA, Lamonova EE, Furmanov YA, Savitskaya IM, Zirconia Crystals Suitable for Medicine: 1 Implants, Ceramics Int., 20, 343-348, 1994. 2. American Welding Society, W elding Handbook, 7th Ed., Vol. 2, 1978. 3. J. A. Hamill Jr, Int. J., Powder M etal. 27, 4, 363, 1991. 4. R. Arroyave, T. W. Eagar, Acta Mater. 51,4871, 2003. 5. M. M. Schwartz, Brazing, ASM Int., 1995. 6. C. Peytour, P. Berthet, F. Barbier, A. Revcolevschi, Interface Microstructure and M echanical Behavior of Brazed Ti6A14V/Zirconia Joints, J. Mater. Sci. Lett, 9, 1129-31, 1990. 7. Agathopoulos S, M oretto P, Peteves SD, Emiliano JV, Correia RN, Brazing of Zirconia to Ti and Yi6A14V. In 1996 Am er Ceram. Soc. Meeting, Indianapolis, USA 1996, Ceramic Joining, Ceramic Transactions. Indianapolis, Vol. 77, P. 75-82, 1997. 8. Santella M L and Pak JJ, Brazing Titanium-Vapor-Coated Zirconia, W elding Research Supplement, p 165-172, 1993. 9. Correia RN, Emiliano JV, and M oretto P, M icrostructure of Diffusional Zirconia-Titanium and Zirconia-Ti6A14V Alloy Joints, J. M aterial Sci., Vol. 33, pp215-221, 1998. 10. Agathopoulos S, Correia RN, Joanni E and Fernandes JRA, Interactions at Zirconia-Au-Ti Interfaces at High Temperatures, Key Eng M ater, 206- 213 487-90, 2002. 11. Agathopoulos S, Pina S and Correia RN, A Review of Recent Investigations on Zirconia Joining for Biomedical Applications, Ceramic Transactions, Vol. 138, pl35-147, 2002. 12. Lasater BJ, Methods for hermetically sealing ceramic to m etallic surfaces and assemblies incorporating such seal, US Patent: 6,221,513 B l, 2001. 13. Fey KE and Jiang G, Application and manufacturing m ethod for a ceramic to metal seal, US Patent: 6,521,350 B2, 2003. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 14. Baker H, et al. ASM Handbook, 3 Alloy Phase Diagrams, 319, 1992. 15. Rauh EG and Garg SP, The Zr02-X (cubic)-Zr02-x (cubic + tetragonal) Phase Boundary, J. Am. Ceram. Soc., 63 [3-4]. 239-240, 1980. 16. Kubasckewski O and Dench WA, The Dissociation Pressures in the Zirconium-Oxygen System at 1000°C, J. Inst. Metals, 84, 440-444, 1995. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 2 Chapter 4: Biocompatibility of 3Y-TZP-to-Ti-6Al-4V Brazed Joint 4.1 Introduction Both 3Y-TZP and Ti-6A1-4V are considered to be biocompatible materials. 3Y-TZP has been used as one of two main ceramic materials of the ball head for Total Hip Replacement (THR), and Ti-6A1-4V has also been widely used for m any implantable biomedical devices and surgical tools. The present work will focus on the biocompatibility of the 3Y-TZP-to-Ti-6Al-4V brazed joint using Tini-50® and nickel filler materials. The chemical toxicity of a metal inside the body is closely related to the concentration of released ions and wear particles, the toxicity of these elements, and the toxicity of the form ed compounds. Even a poisonous substance has no toxic effects in small concentrations, while nutritious substances cause adverse responses if present in excessive amounts. It is difficult to know the exact concentrations of metallic compounds released from implanted materials, because there are many factors affecting them, such as implantation time and local conditions (PH, fretting, etc.). It is known that nickel toxicity is rated differently, depending on w hether it was metal, an insoluble compound, a soluble compounds, or in as a form ation of some other volatile nickel compounds. For implants of pure nickel, Uo et al. showed that they can lead to the release of nickel ions in the vicinity of the im plant, resulting in strong inflammation [1], However, it is well known that Nitinol, consisting of about 49% Titanium and 51% Nickel, has excellent biocompatibility, at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 3 least comparable or better than stainless steel 316L [2-4], where nickel content is only about 10-14 %. The excellent biocompatibility of Nitinol is mainly because the tenacious nature of the TiC> 2 oxide layer on the surface protects nickel from leaching out. It is learned that nickel concentration in nickel-containing devices can’t be used as indicator of biocompatibility, especially when different types of nickel-containing metals are involved. This suggests that the harmful effects of nickel-containing devices toward humans and other animals is not dependant on the amount of nickel or nickel compounds in the device, but rather on the nickel release rate from the devices and the subsequent accumulation of nickel in the tissue. To evaluate the toxicity of nickel in the brazed cases, it is necessary to characterize the solubility or stability of nickel elements (passivation) in the braze joint. The stability of the braze joint can be evaluated by means of electrochemistry testing [5] and nickel release rate measurement in immersion test. Twenty-eight BION® microstimulators, utilizing the 3Y-TZP-to-Ti-6Al-4V joints brazed with Tini-50® filler material, were implanted in seven sheep for various periods of time (up to 163 days), in a study to stimulate the hypoglossal nerve (HGN). Gross inspection at necropsy, and histological examination of tissue and nerve, provided additional information on the biocompatibility of the BION® microstimulators [6]. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 4 4.2 Materials and Experiments 4.2.1 SEM-EDS and X-Ray Diffraction Analysis Braze joint cross-sections were prepared for microanalysis by abrasive sectioning and polishing using a Struers sectioning/polishing machine. A Hitachi S-570 Scanning Electron M icroscope (SEM) equipped with the Integrated Microanalyzer for Images and X rays (IMIX) software from Princeton Gamma-Tech (PGT) was utilized for the SEM examination and Energy Dispersive Microanalysis (EDS). The acceleration voltage for SEM analysis and EDS was 20 KY and the probe size for the EDS analysis was estimated to be less than 2 pm. 4.2.2 Electrochemistry Test Phosphate buffered saline (PBS) solution for the electrochemical test was prepared with one pouch of phosphate buffered saline (Sigma Chemical Co.) diluted to 1 liter with de-mineralized water and had a pH of 7.25 at 22.3°C. Chemical composition of the PBS is shown in Table 4.1. The current and voltage measurements were conducted per standard reference methodology using the EG&G 273 potentiostat that was controlled by a personal computer running the M270 electrochemical analysis software. A three-electrode electrochemical cell was used with an Ag/AgCl reference electrode and a Pt counter electrode. Figure 4.1 showed the three-electrode electrochemical cell setup. Figure 4.2 is the EG&G 273 potentiostat with the assisting PC setup for the electrochemistry test. Five samples, including pure nickel, Tini-50® clad braze filler material, Ti-6A1-4V end cap, Nitinol Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 5 sample, and brazed Ti-6A1-4V end cap joint, were prepared for the electrochemical analysis. The brazed Ti-6A1-4V end cap joint in test was produced in vacuum at 1035°C, and held for 5 minutes, using 0.05m m thick Tini-50® clad filler material. Table 4.1 Chemical Composition of Phosphate Buffered Saline (Sigma Chemical ___________ Co.) Used in Both Electrochem istry and Corrosion Tests___________ Component Percent Component Percent Sodium Chloride (NaCl) 83.8% Potassium Phosphate, Monobasic (KH2P 0 4) 2.0% Di-Sodium Hydrogen Phosphate Anhydrous (Na2 H P 0 4) 12.0% Potassium Chloride (KC1) 2.0% Figure 4.1: Electrochemical Test Cell Setup Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 66 Figure 4.2: EG&G 273 Potentiostat Controlled by A Personal Computer Running the M 270 Electrochemical Analysis Software 4.2.3 Immersion Test Fractured brazed joints and brazed joints were used for a prolonged immersion test. Additionally, Nitinol sample bars were obtained from Nitinol Devices and Components Inc., and were used with the as-ground surface finish. Brazed joints used for the immersion test have the same configuration as described in section 3.2 and were brazed at 1035°C, holding for 5 minutes, with 0.05mm thick Tini-50® clad filler material. Fractured brazed joints have more exposed joint area than that of brazed joints. As all nickel concentrates in the joint area, using the fractured brazed joint in this test presents the worst situation in terms of nickel Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 release. All samples were ultrasonically cleaned for 5 minutes in successive baths of acetone, methanol, and distilled water and dried in air after each step. Samples were then submerged in a 15 ml jar filled with PBS solution and tested at 87°C, in incremental periods, for three months. The chemical composition of the PBS solution was the same as that used for electrochemistry test. The specimens were removed after each immersion time period, and the elemental (nickel and titanium) compositions of the solutions were analyzed with a VG Plasma Quad inductively coupled plasma-mass spectroscopy (ICP-MS), model PQS. The lowest reporting limit for these samples was 0.01 mg/L. 4.2.4 Implantation and Gross Pathology A minimally invasive surgical insertion technique was used to implant 28 BION® microstimulators in seven sheep. H alf of the stimulators were inserted by an anterior approach and half by a posterior approach for stimulation of the Hypoglossal Nerve (HGN). Figure 4.3 shows the locations of im planted stimulators in sheep. Stimulating was performed on the day of surgery, when time permitted, to determine thresholds. Thereafter, stimulation was performed in one period of 90 minutes to three hours, twice per month maximum. Necropsies were performed in all seven sheep, and the gross pathology was reported. Tissue surrounding the microstimulators was explanted and sent for histopathologic evaluation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. S u tm ilaim l> m '•*54 Stimulator A Figure 4.3 Locations of Implanted B IO N R M icrostimulators in Sheep 4.3 Results and Discussions The results from immersion test indicated that fractured brazed joints have higher nickel and titanium release rates than Nitinol does, while brazed joints have the lowest nickel and titanium release rates. Nickel ion release rate is about twice of that of titanium for both fractured and intact brazed joint. Though nickel release rate from a fractured braze joint is almost three times of Nitinol, intact brazed joint nickel release rate is only about a quarter of that of Nitinol. All specimens have nickel release rates well below 0.5pg/cm /week - the safe level restricted by European Parliament and Council Directive 94/27/EC, even at the accelerated temperature of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 9 87°C. Table 4.2 shows the nickel and titanium ion release rates from the samples tested. The ion release rate was calculated with Equation 4.1: Ion Release Rate (pg/cm 2 /week) = Amount of ion (pg) / Total exposure area (cm2) / Immersion time (week) (4.1) Table 4.2: Nickel/titanium Ion Release Rates (pg/cm 2/week) in a 87°C PBS Solution Immersion time (days) Nitinol Fractured Brazed Joints Brazed Joints 7 Below detection Below detection Below detection limit limit limit 14 Below detection Below detection Below detection limit limit limit 45 0.019/0.0025 0.052/0.022 0.005/0.003 91 0.019/0.0035 0.068/0.032 0.004/0.002 Figure 4.4 presents elemental line profiles on a cross-section of a brazed joint. The joint was brazed at 1035°C for 5 minutes using Tini-50® filler material. Table 4.3 shows the chemical composition of the various phases observed in Figure 4.4 using EDS analysis. The analysis revealed that the maximum nickel concentration is located in the Ti2 Ni phase, which can also be seen from the nickel elemental line profile. The maximum nickel concentration is 32.91 at. % (37.58 wt. %), which is well below 50 at. % (56 wt. %), the biocompatible limit of nickel content in TiNi alloy reported by Bogdanski et al. [7] from in vitro cell culture experiments, where Ti-Ni alloys were assessed. Though, as discussed previously, nickel concentration in nickel-containing devices can’t reliably be used as indicator Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 0 of biocompatibility, it is reasonably to cite the results of reference 7, as both alloys in comparison are Ti-Ni alloys and experienced similar metallurgical process and post treatment. Figure 4.4: SEM Image of Interface of A TZP/Ti-6Al-4V Joint Brazed at 1035°C for 5 Minutes and Elemental Line Profile Across the Interface Table 4.3: Chemical Composition of Various Phases Observed in Figure 4.4 Elements Spot 1 Spot 2 Spot 3 Spot 4 at. % ( wt. %) at. %(wt. %) at. %(wt. %) at. %(wt. %) Zr 31.36 (68.89) — — — Y 2.04 (4.36) — — — H f 0.28(1.19) — --- --- O 66.33 (25.56) — --- --- Ti — 65.50 (61.04) 90.10(88.98) 96.33 (96.48) Ni — 32.91 (37.58) 6.53 (7.86) 1.60(1.83) Al — 0.44 (0.24) 0.70 (0.40) 0.89 (0.51) V — 1.15 (1.24) 2.66 (2.76) 1.20(1.18) XRD analyses were conducted on fracture surfaces. 3Y-TZP (Z r0 2, Yo.15Zro.85O1.93, Zro.93Yo.08O1.96), T i2 Ni, Ni2 Ti40 and Ti were found on the ceramic Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 71 side fracture surface, while the m ating surface contained Ti2Ni, Ni2Ti40 and Ti. No elemental nickel was found on either fracture surfaces, which is consistent with the data from EDS analysis. Based on the results from the SEM-EDS and XRD analyses, it is presumed that one part of the nickel from the braze filler material bonded to Ti, Al and V to form T i2(NiAlV) inter-metallic, another part is formed as nickel titanium oxides (Ni2Ti4 0 ) at the ceramic-to-braze interface, and the remainder diffused into the bulk of Ti-6A1-4V. Figure 4.5 represents polarization curves of five tested samples: pure nickel, Tini-50® braze filler material, Ti-6A1-4V end cap, Nitinol, and a brazed Ti-6A1-4V end cap joint, in PBS solution at a scan rate of 0.5 mV/s. The sharp rise of current density on the TiNi clad braze filler material and pure nickel at the potential of about 0.3V, with respect to the Ag/AgCl reference electrode, corresponds to active nickel dissolution. This current rise is not observed on the Ti-6A1-4V end cap or the end cap joint brazed with the Tini-50® clad braze filler material. This indicates free nickel is absent on both samples tested. The corrosion potential of the brazed Ti- 6A1-4V end cap joint is lower than the Ti-6A1-4V end cap, which is possibly due to the brazing process changing the Ti-6A1-4V end cap surface characteristics. The polarization curve of brazed Ti-6A1-4V end cap joint lies between Ti-6A1-4V end cap and Nitinol, indicating a good corrosion resistance of the braze joint, as both Ti- 6A1-4V and Nitinol are considered biocompatible materials. Summarized in Table 4.4 are the corrosion potentials of all tested samples. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 2 E E & m 09 o < / ) c a > Q •4-P c £ 3 a O m M ? " * < • a P u r e Nickel -TiNi50 C la d Filler M aterial Ti-6AI-4V E n d C a p D Nitinol E— TiNi50 B r a z e d E n d C a p Joint Potential, V Figure 4.5: Polarization Behavior of Pure Nickel, Tini-50® Clad Filler Material, Ti-6A1- 4V End Cap, Nitinol and Ti-6A1-4V End Cap Braze Joint in PBS Solution at Scan rate of 0.5mV/s, Logarithmic Current Scale Table 4.4: Corrosion Potentials of Pure Nickel, TiNi-50® Clad Filler Material, Ti- 6Al-4Y End Cap, Nitinol and Ti-6A1-4V End Cap Braze Joint in PBS Solution Tested Samples Corrosion potential (V) TiNi Clad Filler Material +0.3 Pure Nickel +0.3 Ti-6A1-4V End Cap Braze Joint +0.9 Nitinol +0.8 Ti-6A1-4V End Cap +1.3 Seven sheep were successfully implanted with four Tini-50® brazed joint BION® microstimulators each. There was no post-operative hemorrhage or infection. Gross pathology on all sheep at necropsy showed that the tissues near the microstimulators and the hypoglossal nerves were found to be unremarkable with no evidence of any inflammation or previous hemorrhage. Thirteen of the fourteen Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 73 HGNs in the seven sheep were prepared, sectioned, and examined histologically by a pathologist, who reported that all nerve samples were normal. The capsules formed around each of the 28 microstimulators was identified during the initial extraction when the microstimulators were removed. The pathologist prepared, sectioned and examined 22 of these capsules. The histology of the 22 tissue samples showed normal mature fibrous connective tissue encapsulation (varying from 50 to 500 microns in thickness (mean thickness of 0.163 mm, 44 sampled sites) with focal to multi-focal regions of minimal to m ild mononuclear inflammatory cells. Representative photomicrographs are presented in Figure 4.6. 4.4 Conclusions SEM-EDS, XRD and electrochemical analyses indicated, that nickel from Tini-50® filler material has bonded to other elements and that no elemental nickel is present in the brazed joint after the 1035°C and holding for 5 minutes brazing process. Both immersion and electrochemistry tests indicated good corrosion resistance of brazed joints, better than Nitinol. Tissue examination by gross inspection at necropsy and histologically revealed thin fibrous tissue encapsulation of the microstimulators and no histological damage to the adjacent nerve, indicating safe biocompatibility. The nickel brazed joint has identical microstructures to Tini- 50® brazed joint, therefore a safe biocompatibility of the nickel braze joint is expected, too. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 4 Figure 4.6. Photomicrographs of (a) Sheep 005 (examined after 163 days of implantation): Anterior Right microstimulator, Trichrome stain; (b) Sheep 305 (examined after 51 days of implantation): Anterior Left microstimulator, Hematoxylin and Eosin (H & E); (c) Sheep 310 (examined after 113 days of implantation): Posterior Right microstimulator, H&E Stain; (d) Sheep 003 (examined after 121 days of implantation): Posterior Right, Siever-M unger Stain, Nerve Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 5 4.5 References 1. Uo M, Watari F, Yokoyama A, M atsuno H and Kawasaki T., Dissolution of Nickel and Tissue Response Observed by X-ray Scanning Analytical Microscopy. Biomaterials, 20: 747-55, 1999. 2. Assad M, Lemieux N, Rivard CH, Yahia L., Comparative In Vitro Biocompatibility of Nickel-titanium, Pure nickel, Pure titanium, and Stainless Steel: Genotoxicity and Atomic Absorption Evaluation. Biomed M ater Eng 9:1-12, 1999. 3. Kpanen A, Ryhanen J, Danilov A, Tuukkanen J., Effect of Nickel-titanium Shape M emory Metal Alloy on Bone Formation. Biomaterials, 22:2475-80, 2001. 4. W ever DJ, Veldhuizen AG, Sanders MM, Schakenraad JM, van Horn JR. Cytotoxic, Allergic and Genotoxic Activity of A Nickel-titanium Alloy. Biomaterials, 18:1115-20, 1997. 5. Rocha LA, Ariza E, Costa AM, Oliveira FJ and Silva RF., Electrochemical Behavior of Ti/A1203 Interfaces Produced by Diffusion Bonding. Mater Res, 6:4 439-444, 2003. 6. Jiang G, M ishler DR, Davis R, M obley JP, Schulman JH, Zirconia to Ti- 6A1-4V Braze Joint for Implantable Biomedical Device, J. of Biomed M ater Res, Part B: Appl Biomater, 72B: 316-321, 2005. 7. Bogdanski D, Koller M, M uller D, M uhr G, Bram M, Buchkrem er HP, Stover De, Choi J and Epple M., Easy Assessment of the Biocompatibility of Ni-Ti Alloys by In Vitro Cell Culture Experiments on a Functionally Graded Ni-NiTi-Ti Material. Biomaterials, 23: 45 b, 49-4555, 2002. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Chapter 5: Low Temperature Degradation and Life Prediction of 3Y-TZP Ceramics Package 5.1 Introduction The flexural strength of 3Y-TZP ceramic is about three times higher than that of alumina ceramics. However, 3Y-TZP can suffer from low temperature degradation (LTD) in moist environments. This aging phenomenon is caused by the transformation of the crystalline structure from the tetragonal (T) phase to the monoclinic (M) phase, resulting in decrease in strength and toughness, along with micro- and macro-cracking [1-3], which limits 3Y-TZP’s long-term applications. It was considered that such aging-induced phase transformation could be suppressed by decreasing the grain size of 3Y-TZP [4-7]; however, the fracture toughness and flexural strength decrease with further decrease of the grain size [8], W atanabe et al. [4] identified a critical grain size below which no transformation occurred during 1000 hours of aging at 300 °C. On the other hand, Li [9] reported that the low temperature phase transformation becomes more difficult with the increase in grain size. The effect of grain size on transformation is complicated by many factors, resulting in a “U ” type relationship between the ease of transformation and grain size. However, too big grain size will also compromise the mechanical properties of the 3Y-TZP. Unfortunately, the characteristics of 3Y-TZP ceramic are that it demonstrates superior mechanical properties at the same average grain size of 0.4- 0.5 (am around which the aging problem will occur. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. It is learned that the degradation rate is not the same for all TZP ceramics. One can observe that there is experimental evidence that Y-TZP ceramic is able to maintain good mechanical properties in moist environments for extended periods, but general conclusions about the stability of Y-TZP m ust be avoided, as this behavior is peculiar to each material and to manufacturing technique. Swab studied seven Y-TZP materials from different commercial companies, and they all behaved quite differently [10]. Variability in aging behavior is related to the differences in equilibrium of micro-structural parameters, like the concentration and distribution of stabilizer, grain size, flaw population and distribution, residual stress, density, etc. Thus, it is necessary to conduct separate aging experiments on individual ceramics, in order to obtain relevant information on aging behavior. There is limited information available on the degradation of the ceramic exposed to water at body temperature, especially from in-vivo studies. W hile the degradation of Y-TZP ceramic at 37°C has not been fully assessed, it has been preliminarily investigated. For example, Thompson and Rawlings [11] assessed the mechanical properties of Y-TZP in a stimulated body solution (Ringer’s solution). The result was that Y-TZP demonstrated a significant strength decrem ent when aged for long periods in the Ringer’s solution, and was therefore unsuitable as implant material. Drummond [12] reported that yttria-stabilized zirconia demonstrated low- temperature degradation at 37°C with a significant decrement in strength in as short a period as 140 to 302 days in deionized water, saline, or Ringers solution. Drummond also reported on similar observation by others, where yttria-stabilized Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 zirconia demonstrated a strength decrement in water vapor, room temperature water, Ringers solution, hot water, boiling water, and post-zn vivo aging. Several in vivo or body temperature in vitro studies [13-18] have reported various results on aging behavior of Y-TZP ceramic. Researchers in the biom edical industry perform accelerated life tests (ALT) to estimate the lifetime of a device by using the Arrhenius factor to reduce the cost and testing time [19]. For example, one can test a device in a physiology-mimic saline at an elevated temperature, for exam ple 87°C to accelerate its failure; and then use the AF between the elevated temperature and body temperature (assuming there is an Arrhenius relationship for this system) to extrapolate the lifetime of the device at body temperature. For the BION® microstimulator package, there are two possible failure modes: One mode could be the loss of the hermeticity of the package due to corrosion occurred in the ceramic-to-metal seal. The second could be the degradation of the ceramic results in surface spallation or self-destruction. The second mode is the more likely one when a qualified ceramic-to-metal joint is provided. To establish the long-term implantation performance of the implant packaging containing the Y-TZP ceramics, it is desirable first to verify that the LTD process obeys Arrhenius law and then to employ the Arrhenius factor to predict the time course of LTD. Svante Arrhenius (1859-1927, Sweden, Nobel Prize in 1903) studied the dependence of the reaction rate versus temperature and proposed a phenomenological law. This is a nice example used by the scientists to model Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 experimental phenomena. In many cases, a chemical reaction does not start even if its thermodynamic parameters are favorable (AG<0): for instance, paper or wood do not ignite spontaneously at ambient temperature, but need to be lit before burning. The same applies to hydrogen peroxide, which does not break up spontaneously to evolve oxygen and produce water, or to permanganate in aqueous solution, which does not oxidize water at ambient temperature. In all cases, and without exception, one observes that the rate of chemical reactions increases when the temperature of the reactants is raised. The temperature-dependence of the reaction rate is specific to each chemical reaction. Looking at the above-mentioned phenomenological facts, Arrhenius acquired the conviction that a chemical reaction needs a “start up” . Energy is needed to “prepare” the reactants for reaction. For instance, they must collide with an adequate orientation so that the collision results in a weakening of the chemical bonds to be broken. Arrhenius consequently proposed a law containing two temperature-independent parameters: The activation energy Ea expressed in kJ-m of1, and the frequency (or pre-exponential) factor A expressed in the same units as k, as shown in Equation 5.1: k = A* exp(-EA /RT) (5.1) In the case of moisture-induced T-M phase transformation, as tetragonal and monoclinic phases have different crystal structures, peak intensities of tetragonal phase, T ( l l l ) at 20=30.2°, and monoclinic phase, M(T11) at 20=28.2° and M ( lll) at 20 = 31.3° shall be identified by X-ray diffraction (Copper K a radiation) analysis Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 0 to calculate percent of monoclinic phase by the modified Garvie-Nicholson Equation 5.2. Figure 5.1 is a typical XRD profile of a ceramic containing both T and M phases. Monoclinic phase % = M(T11)+M(111)/(T(111)+M(T11)+M(111)) (5.2) TKUt M ££ A dw .5.1 1 B S M (+ll) MCI ll) Figure 5.1: XRD Profile of A Ceramic Containing Both Tetragonal and Monoclinic Phases Degradation rate can be demonstrated by a rate constant. Arrhenius factor can be determined from the quantitative XRD information obtained from the ceramics aged at different temperature. It has learned that the aging kinetics of the T-M phase transformation are first order with respect to the amount of tetragonal phase available for transformation. Plots of -ln (l-f) versus aging time for the Y-TZP materials in which the T-M phase transformation occurred can be made. The value f is the concentration of monoclinic phase expressed as a fraction of the maximum concentration of monoclinic phase formed; (1-f) thus corresponds to the remaining fraction of transformable tetragonal phase available for transformation. The rate Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 constants can be calculated by linear regression analysis. So the Arrhenius relationship between various temperatures shall be obtained using the rate constant. The transformed M phase can be distinguished from the original tetragonal bulk, and a SEM can be used to measure the transformed layer thickness. The monoclinic layer growth rates at various accelerated aging temperatures can be calculated and a set of Arrhenius factors can be determined from the resulting monoclinic phase layer growth rates at various temperatures. Another approach to obtain the Arrhenius factor (AF) is based on its activation energy Ea of the ceramic, which is also based on XRD data. M any studies have suggested that the relationship between the amount of monoclinic phase and the aging time could be expressed by the M ehl-Avrami-Johnson (MAJ) law [20, 21] i.e., Equation 5.3: where f is the transformation fraction, t is the time and b and n are constants. In the MAJ theory summarized by Christian [21], it is shown that the n exponent, which can be derived from the slope of the In (In (l/(l-f))) versus In t plot, is related to nucleation and growth conditions; and that b is a param eter giving the apparent activation energy Ea by Equation 5.4. where bo is a constant, R the gas constant and T the absolute temperature; and where b can be calculated for each temperature from the In (In (l/(l-f))) versus lnt plot. (1-f) corresponds to the remaining fraction of tetragonal phase available for f = 1- exp [-(b*t)A n] (5.3) b=b0 *exp [-Ea/(RT)] (5.4) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2 transformation. The apparent activation energy for the T-M phase transformation process can then be determined by plotting lnb versus 1/T. Equation 5.5 can be used to calculate AF from the Ea. Ln(A F)=-Ea/R*(l/T2-l/T i) (5.5) Arrhenius factor Q 10, also called temperature coefficient, is to determine the ratio of the rate of activity at one temperature, to the rate of activity at a temperature 10°C lower. Chevalier conducted aging tests on zirconia hip joint heads in steam, at a temperature of 134 °C, under a pressure of 2 bars (-0.2 M Pa). The estimated activation energy (Ea) is 113.2KJ/mol [22], The corresponding Arrhenius factor (AF) Qio values are from 3.22 to 2.05 at the temperature range of 37 °C to 134 °C. The Arrhenius factors (Qio) obtained from three 3Y-TZP ceramics aging in water at 83 °C to 120 °C by Sato range from 1.28 to 2.40 [2] by calculation from the tetragonal-to- monoclinic (t-m) phase transformation rate constants. Their activation energy values for the three 3Y-TZP ceramics are 76.2, 78.7 and 93.8 KJ/mol, respectively. Predicting lifetime of Y-TZP materials used in implant packaging is done by determining the Arrhenius factor of the system based on quantitative information obtained from the accelerated aging tested dummy micro-stimulators and in-vivo studies carried out with BION® microstimulator implants in sheep and rats. The SEM micrographs will provide visible evidence of T-M phase transformation and quantitative proof of the first order characteristics of the T-M phase transformation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 The relationship between bend strength and transformed layer thickness is also discussed. 5.2 Experimental Methods Ceramics from three different vendors, designated as A, B and C, have been studied. Aging treatments were carried out in temperature-controlled ovens and in autoclaves. Average grain size of each ceramic has been measured by SEM using the line-intercept technique [17]. The average transformed layer thickness was measured and calculated from SEM pictures using the Coating Thickness Module in the PGT X-Ray and Imaging M icroanalysis Systems software statistically. LTD was quantified by both the monoclinic fraction (f) and the transformed layer thickness measured by SEM; while f was measured using an XRD technique and calculated from the modified Garvie-Nicholson equation, i.e., f=Im /(Im +If). Im is the peak intensity of monoclinic phase on XRD profile and \t is the peak intensity of tetragonal phase. DataScan and Jade 6.0 software from Materials Data Inc. were used to collect and process the XRD data. Based on the XRD data and the approach of Chevalier [16], activation energy can be obtained. Aged samples were obtained and subjected to three-point bending strength test to reveal the remaining strength of the ceramic after various aging time intervals. The span of the 3-point bend test was 0.40 inches. Two BION® microstimulators using ceramic B were implanted in a sheep for 454 days. Two BION® microstimulators using ceramic C were implanted in a rat. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 One was explanted from rat after 206 days and the other after 238 days. Figure 5.2 (a) and (b) are X-ray pictures of the samples in the sheep and a rat. All units were XRD analyzed before implantation and after explant to obtain the f. Figure 5.2: (a) Two BION® M icrostimulators Using Ceramic B Implanted in Sheep and (b) Two BION® M icrostimulators Using Ceramic C Implanted in A Rat 5.3 Results and Discussions 5.3.1 Surface Finish of Different Ceramics Ceramic surfaces have been observed using SEM. The surface finish of the ceramics appeared quite different from each other, which can be attributed to different machining methods and parameters. Ceramic A had the smoothest surface while ceramics B and C both had rough finishes, as shown in Figures 5.3 through 5.5. It is known that post-machining induces a T-M transformation near the surface, at a level depending on the stress applied (feed rate, cutting speed, etc.). Ceramic A has 0.5% monoclinic on the surface originally, while ceramic B has 2.2%, and ceramic C has 3.0%. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 Figure 5.3: Ceramic A Surface Figure 5.4: Ceramic B Surface Figure 5.5: Ceramic C Surface 5.3.2 Microstructural Observation of Different Ceramics The grain size was estimated by the line intercept method from the scanning electron micrographs of the air-etched fracture surfaces. The samples were etched in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 6 air by heating up to 1300 °C and holding for 1 hour. This thermal treatment is believed to have no grain growth influence [23], SEM micrographs of ceramics A, B, and C are shown in Figures 5.6, 5.7 and 5.8. The average grain size for each material varies, as well as the grain size distribution. The average grain size of ceramic A is 0.50 microns with a very small standard deviation. Ceramic B has an average grain size of 0.59 microns, but with a huge variation. Some grains are as small as 0.2 microns while some are almost 3.0 microns. Ceramic C has a 0.56 microns average grain size. The grain size variation is bigger than ceramic A but smaller than ceramic B. The average grain sizes of all the materials are above the critical grain size (about 0.3 microns for 3 mol % Y-TZP materials [4]) of LTD. T- M phase transformation is expected for all the ceramic samples. Figure 5.6: Grains of Ceramic A (Average Grain Size=0.50microns) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 Figure 5.7: Grains of Ceramic B (Average Grain Size =0.59 microns) Figure 5.8: Grains of Ceramic C (Average Grain Size = 0.56 microns) 5.3.3 X-Ray Diffraction Results Ceramic A tubes were soaked in water and in steam at temperatures of 67 °C to 127°C. The relationship between the amount of monoclinic phase and the aging time at various temperatures was sigmoidal, in agreement with previous studies conducted at high temperatures [24, 25-27]. These studies suggested that the relationship between the amount of monoclinic phase and the aging time could be expressed by the Mehl-Avrami-Johnson (MAJ) law [26, 27] (Equation 5.3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In the MAJ theory summarized by Christian [21], it is shown that the n exponent, which can be derived from the slope of the In (In (l/(l-f))) versus In t plot, is related to nucleation and growth conditions. Since b is a param eter giving the apparent activation energy E a by Equation 5.3, b can be calculated for each temperature from the In (In (l/(l-f))) versus In t plot, shown in Figure 5.9. In our soaking tests with ceramic A, the monoclinic phase content only reaches 73% even after a very long duration; f is the fraction in this saturated amount of monoclinic phase. Thus (1-f) corresponds to the remaining fraction of tetragonal phase. Figure 5.10 gives activation energy of 69.9 kJ/mol of A ceramics tested with excellent reliability (correlation coefficient of 99.42%), which verified that the low temperature degradation process of this system did follow the Arrhenius law. Figure 5.9: Plot of ln(ln (l/(l-f» ) Versus lnt for the Determination of n (slope) and lnb (Ordinate Origin) in the MAJ Equation 3 y4 = 0.9757x- 6.9294 y3 = 0.9357x-7.4028 y2 = 0.8889x- 7.9156 y1 =0.8934x-9.3976 ♦ 87C(y2) 107C(y4) 95C(y3) • 127C(y5) y5 = 1.0892x-5.6015 ■ 67C(y1) -5 0 2 4 ln(t) 6 8 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. -5.5 — y — --0415 +T 5:37 8 -----------^ 0 .9042---- -6.5 -■ f -7.5 - -9.5 -■ 0.0028 0.003 0.0032 0.0024 0.0026 1J T (1Jk) Figure 5.10: Plot of lnb Versus 1/T for the Determination of the Apparent Activation Energy for the T-M Phase Transformation Process 5.3.4 Transformed Layer Thickness Measurement Ceramic A was soaked in 107°C and 127°C, and soaking tests on ceramic B were performed at 87°C and 107°C. After the destructive strength tests of both ceramics, SEM observations on the fractured surfaces were conducted. The transformed portion can be distinguished from the original tetragonal bulk and can be measured under SEM as shown in Figures 5.11, 5.12, 5.14 and 5.15. This is apparent and visible evidence of T-M transformation. SEM observation found that the transformed phase gets deeper with the increase of soaking time. It is also found that there is a linear relationship between the thickness of transformed layer and the aging time, as shown in Figures 5.13 and 5.16. The transformed layer growth rates of ceramic A are 0.0168 micron/hour and 0.0566 micron/hour at 107°C and 127°C, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 0 respectively. The transformed layer growth rates of ceramic B are 0.0177 and 0.0714 micron/hour at 87°C and 107°C, respectively. It was noticed that the growth rate of ceramic B at 107°C is about 4.25 times of that of ceramic A at the same temperature, which is related to the differences in equilibrium of micro-structural parameters of the ceramic. Figure 5.11: Transformed Layer Grows in Ceramic A in 107°C Aging Test Figure 5.12: Transformed Layer Grows in Ceramic A in 127°C Aging Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 35 - ■ y = 0.0566k, 30 - ■ 15 - ■ 10 -■ ♦ 107C *1270 800 1000 1200 200 400 600 0 Aging time (hours) Figure 5.13: Transformed Layer Growth Rates of Ceramic A Soaking in 107 and 127°C Steam 600houi Figure 5.14: Transformed Layer Grows in Ceramic B in 87°C Aging Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 92 Figure 5.15: Transformed Layer Grows in Ceramic B in 107°C Aging Test 8 40 c u y = 0.0714x 35 - ■ 30 - ■ 15 - ■ 10 - ♦ 87C a 107 C l- 2000 2500 1500 1000 500 0 Aging tim e (hours) Figure 5.16: Transformed Layer Growth Rates of Ceramic B Soaking in 87 and 107°C W ater Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5.3.5 Ceramic Aging in Liquid Water and in Steam Both X-Ray diffraction results from aged ceramic A tubes and transformed layer growth rates of aged ceramic Bs suggest that the mechanism of the moisture- induced T-M phase transformation doesn’t change the existing states of water (liquid and vapor in this test). To further verify this finding, aging test of ceramic B in both 127 °C liquid water and steam were carried out. Two ceramic B tubes were put in sterilization bags. Two were placed in a sealed jar filled with water. Both samples were placed in a 127°C autoclave. Figure 5.17 showed that the monoclinic phase percentage on ceramic B surfaces in both 127 °C liquid water and steam increased at the same speed, which suggested same temperature liquid w ater and steam have equal effect as far as the m oisture-induced T-M phase transformation is concerned. 45 40 35 £ 3 0 8 20 i « 10 5 0 0 20 40 60 80 100 120 Soaking Time (hours) Figure 5.17 T-M Phase Transformation Occurred at Same Rate in Equal Temperature Liquid W ater and Steam ♦ In Steam ■ In Liquid Water Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 4 5.3.6 Arrhenius Factor Qio Calculation As discussed previously, there are several ways to obtain the Arrhenius factor. The most popular AF or temperature coefficient, designated as Qio, is the ratio of the rate of activity at one temperature, to the rate of activity at a temperature 10°C lower. The activation energy given by Figure 5.10 corresponds to Arrhenius factor Qio values of 2.33 to 1.72 in the temperature range of 37°C to 127°C by applying to the equation 5.5. The transformed layer growth rates extrapolated from Figures 5.13 and 5.16 showed that the transformation rate of ceramic A at 127°C is about 3.8 times of that at 107°C. By V an’t Floff Equation [28] (Equation 5.6), K2=K 1 *Q1 0 a(10/(T2-T1 )) (5.6) where Ti and T2 are the absolute temperatures and Ki and K 2 are the corresponding reaction rates, the calculated Qio value for ceramic As is 1.84. It is similarly calculated that the transformed layer growth rate of ceramic B in 107°C aging test is about 4.03 times that of 87°C, with a calculated Qio value of 2.01 in the corresponding temperature range. The linear growth of monoclinic layer on the ceramic surface indicated a lst-order kinetic of the T-M transformation, which has also been suggested by Sato [2], Alternatively, AF Qio can be determined by using the monoclinic phase fraction (f) acquired from XRD analysis. By plotting -In (1-f) versus aging time (t), both for A and ceramic Bs at various temperatures, rate constants of transformation can be obtained. Figures 5.18 to 5.21 showed changes of monoclinic phase Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 percentage (f) over time for both A and ceramic Bs at 67°C to 127°C soaking tests. The Arrhenius factor Qio values calculated from these rate constants are summarized in Table 5.1, assuming Qio is constant within the two temperatures 20°C apart. As indicated theoretically, Qio is temperature dependent: Qio value is larger at low temperature range and smaller at high temperature range. The Qio data tendency shown in Table 5.1 agrees well to theory prediction. 67C Soaking 0.3 0.25 y = 0.00022x + 0.025 + ^ 0.2 c _i ■ a 0.05 0 500 1000 1500 2000 2500 3000 3500 soaking tim e (hrs) Figure 5.18: Monoclinic Phase Percentage (f) Increases Over Time for Both A and B Ceramics in 67°C Soaking Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 96 87 C soaking c-' 5 y= 0.0012x + 0.115 4 2 1 0 1000 2000 3000 4000 5000 0 soaking time ( hrs) B *A Figure 5.19: M onoclinic Phase Percentage (f) Increases Over Time for Both A and B Ceramics in 87°C Soaking Test 107C Soaking 2.5 = 0.0058x - 0.1084 9 c _l 0.5 800 1000 1200 200 400 600 -0.5 soaking tim e (hrs) ♦ b ■ A Figure 5.20: Monoclinic Phase Percentage (f) Increases over Time for Both A and B Ceramics in 107°C Soaking Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 127C soaking *B ■ A 0 soaking time (hrs) Figure 5.21: Monoclinic Phase Percentage (f) Increases over Time for Both A and B Ceramics in 127°C Soaking Test Table 5.1: Arrhenius Factors (Qios) Calculated from M onoclinic Phase Rate Constants Materials 67 °C 87°C 107°C 127°C A R.C. (/hr) 4.0e-5 2.1e-4 1.0e-3 0.0043 Qio 2.29 2.18 2.07 B R.C. (/hr) 2.2e-4 1.2e-3 5.8e-3 0.022 Qio 2.33 2.20 1.95 (R) M onoclinic phase of ceramic B brazed BION microstimulators prior to implantation was 0%. After 454 days implantation in sheep, the monoclinic phase fraction of one unit is 0.7% and it is 0.8% for another one, indicating the T-M transformation did occur in sheep body. Sheep body temperature was around 39°C. The same transformation was observed from the rat explants, where ceramic C was used. Rat body temperature was around 36°C. The monoclinic fraction increases to ♦ ^ y = O .Q22x - 0.1868 y = U.0043X - O .U Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 10.4% and 11.1% respectively, after 206 and 238 days implantation, from the original 3.0 % prior to implantation. Using the rate constants obtained from the in- vivo implants, compared to comparable units which experienced 87°C accelerated aging tests, Arrhenius factor Q i0 values can be calculated. For ceramic B, the Q i0 values are 2.05 and 2.40 for the two units, respectively. For ceramic C units, the Qio values are 2.02 and 2.05. Table 5.2 summarized Q i0 values obtained from both in- vitro and in-vivo tests via various methods. It is shown that the Qio values obtained fall into a relatively tight range for the ceramics tested. Now we can predict the lifetime of an implantable device containing these ceramic materials by conducting accelerated life testing and using the obtained Arrhenius factors. Table 5.2: Arrhenius Factor Q iq Values Summary MATER IALS Activation energy (Ea) approach Transformed layer growth rates approach Rate constants approach via XRD data from ALT Rate constants approach via XRD data from In-vivo 37°C units and 87°C ALT units A 2.33 ~ 1.72 1.84 (127°C / 2.29-2.07 (37~127°C) 107°C) (67~127°C) B 2.04 (107°C 2.33-1.95 2.05 and 2.40 /87°C) (67-127°C) (87°C/37°C) C 2.02 and 2.05 (87°C/37°C) 5.3.7 Bend Strength Reduction of In-Vitro Aged Ceramics As discussed in previous chapter, T-M phase transformation induces micro- or macro-cracking on the surface. W ater or water vapor can penetrate through the transformed layer and into the bulk so to cause further T-M transformation. The T- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 M phase transformation degrades mechanical properties of the ceramic not only because the transformed monoclinic phase possesses poorer mechanical properties than that of tetragonal phase, but also the transformation often induces micro or macro cracks due to the accumulated stress because of the volum e change associated with the T-M phase transformation that could be the initial point of fracture during bending test. Figure 5.22 showed a crack at the transform ed and untransformed ceram ic interface. Transformed Ceramic Layer Tetragonal Ceramic Bulk Crack Figure 5.22: Cracks in Transformed Layer of Ceramic B A fter 42 Days of Soaking in 87°C W ater Three-point bend test performed on 2mm diameter ceramic A dummy units revealed that the bend strength degrades with the increase of aging time in 127°C steam as shown in Figure 5.23. The average cracking load for non-aged 2mm diameter ceramic A brazed units is about 38 lbs, which is equivalent to 1124 MPa of bend strength. After 1368 hours aging, the cracking load reduced to 19 lbs. Which Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 0 is about a half of its original, but it is still stronger than alumina ceramic. 2mm diameter ceramic B tubes demonstrated the same behavior: cracking load decreases with the increase of soaking time, as shown in Figure 5.24, but degrade at much faster rates because ceramic B has lower aging resistance, and aging process occurred on both inner and outer ceramic surfaces. Ceramic B tubes also started to flake after a short period of time, and self-destructed due to the stress accumulation from the T-M transformation. Micrographs (Figure 5.25) show flaking-off and self- destructed ceramic B samples (both brazed cases and bare tubes) after a period of time of equivalent 15 years of implantation aging where excessive phase transformation occurred. 50 i 10 -■ m o -I- - - - - - - - - - - - - - - - - , - - - - - - - - - - - - - - - - - 1 - - - - - - - - - - - - - - - - - 0 500 1000 1500 Soaking time (hours) Figure 5.23: Cracking Load of 127°C Aged 2mm D iam eter Ceramic A Brazed Cases in Three-point Bend Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 1 50 ~ 4 5 £ 40 ~ 3 5 = 20 1 15 o 10 5 0 0 500 1000 1500 2000 2500 Soaking Time (hours) Figure 5.24: Cracking Load of 2mm Diam eter Ceramic B Tubes in Three-point Bend Test Degrades After Soaking in 87°C W ater Figure 5.25: Self-destructed 2mm Diameter Ceramic B After Excessive Transformation Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 2 5.3.8 Bend Strength Reduces with Increase of Transformed Monoclinic Layers It is demonstrated that three-point bend strength of ceramic A decreased with increase of transformation layer thickness, as shown in Figure 5.26. Figure 5.27 shows the same tendency of 2mm Diameter Ceramic B soaked in 127°C steam. (A n ■ o ro o O ) C o (0 O 45 40 35 30 25 20 15 10 5 0 10 20 30 40 Transformation layer thickness (microns) 50 Figure 5.26: Cracking Load Reduces with Increase of Transformation Layer Thickness (127°C Aged 2mm Diameter Ceramic A Brazed Cases) 10 20 30 40 50 60 70 Transformed layer thickness (morons) 80 90 Figure 5.27: Cracking Load Reduces with Increase of Transformation Layer Thickness (2mm Diameter Ceramic B Tubes Soaked in 127°C Steam) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 5.3.9 Geometrical Factor 5.3.9.1 2mm Diameter Ceramic A Tubes Soaking in 127°C Steam Regular processed 2mm diameter ceramic A tubes have been soaked in 127°C up to 903 hours. SEM observation was performed on the fractured surfaces of the aged ceramic tubes. It was found that the ceramic materials on the inner and outer surfaces transformed at different rates, as shown in Figure 5.28 through 5.30, which is possibly due to the ceramic materials on these two surfaces being under opposite stress constrained conditions. The ceramic grains on the inner surfaces are under compression while the grains close to outer surface are under tension. It was further learned that the moisture-induced T-M phase transformation will result in a 3-5 % volume expansion. The compressive stresses on the inner surface restrained the spontaneous moisture-induced t-m phase transformation so as to slow down the T-M transformation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 X500 50 M m 0 0 0 0 4: 800 hours Figure 5.28: Transformed Layer Grows on Outer Surfaces of Regular Processed 2mm Diameter Ceramic A Tubes in 127°C Soaking Test i.n 0 liouo (hi 2t)2 lioms Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 (d) 610 hours Figure 5.29: Transformed Layer Grows on Inner Surfaces of Regular Processed 2mm Diameter Ceramic A Tubes in 127°C Soaking Test V=0 .0 9 8 1 x - 10.389 0.9976 ♦ Inner Surface I Outer Surface 1000 800 600 400 200 0 Soaking time (hours) Figure 5.30: Transformed Layer Growth Rates of Regular Processed 2mm Diameter Ceramic A Tubes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 0 6 5.3.9.2 3mm Diameter Ceramic A Tubes Soaking in 127°C Steam Figures 5.31 and 5.32 are transformed layers on the inner and outer surfaces of 3mm Diameter ceramic A, respectively, after soaking in 127°C steam for various time periods. Figure 5.31: Transformed Layer Grows at Inner Surfaces of 3mm Diameter Ceramic A in 127°C Aging Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 Figure 5.32: Transformed Layer Grows at Outer Surfaces of 3mm Diameter Ceramic A in 127°C Aging Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 Figure 5.33 shows that the transformation layer growth rate on outer surface (0.0381 microns/hour) is about 1.8 times of that on inner surface (0.0217 microns/hour). v > E U Q 3 « ■3.S ■O c 0 ) a v > E s I- 70 60 50 40 30 B z, 20 + Outer Surface ■ Inner Surface 10 0 500 1000 Soaking Time (hrs) 1500 2000 Figure 5.33: The T-M Transformation of Outer Surface Was Faster Than That of Inner Surface Due to Geometrical Self-restriction Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 09 5.3.10 Hot Isostatic Pressed Ceramics One approach to mitigating the T-M phase degradation is to utilize Hot Isostatic Pressing (HIP), which is a process where as sintered ceramic material is pressurized in an inert gas environment at a pressure of at least 100 MPa and heated up to 1200 to 1450°C [29], It minimizes the porosity percentage of the as-sintered ceramic, increases its density to further improve the mechanical properties of the ceramic. As sintered TZP material has a density of about 5.8 ~ 6.0 g/cm3 depending on sintering parameters, such as the sintering temperature or sintering time. Y-TZP that has undergone HIP is denser than the as-sintered TZP material, with a measured 3 3 bulk density of about 6.05 g/cm , compared to6.0 g/cm , the bulk density of non-HIP processed material. It was learned that the T-M transformation occurs on the surfaces and proceeds into the ceramic bulk. The T-M phase transformation induces micro or macro cracks, which are the migration paths for water vapor and results in further transformation. It is believed that HIP enhances the density of ceramics and could thus hinder the w ater from further penetration into the ceramic bulk so to reduce the degradation of the ceramic materials. To evaluate the possible inhibition effect of HIP on the T-M transformation, HIP processed 2mm diameter ceramic A tubes, which were obtained from the same lot of regularly processed ceramic tubes and then HIP processed, were soaked in 127°C steam for up to 800 hours along with regularly processed ceramic A tubes. XRD analysis performed on the ceramic surfaces revealed that the HIP processed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 ceramic A tubes have a lower T-M phase transformation rate constant than the regularly processed ceramic A tubes, as shown in Figure 5.34. This indicates a higher degradation resistance for the HIP processed ceramic A tubes. HIPped and Regular Ceramics Soaking in 127C 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 y"="0‘ .0‘ 08‘ 2x R = 0.9696 ■ Regular ♦ HIPped 0 100 200 300 400 500 Soaking Time (Hours) Figure 5.34: HIP processed 2m m D iam eter Ceramic A Has A Slower T-M Transformation Rate Than the Regularly Processed 2mm Diameter Ceramic A Aged HIP processed ceramic tubes were strength evaluated by three-point bending and compared to the regularly processed ceramic tubes. Ten samples were tested at 0 hour soaking for both groups. Five samples were evaluated at each soaking time intervals of 200, 400, 600 and 800 hours. Three-point bending test results revealed that HIP processed 2mm diameter ceramic A tubes have a much higher cracking load (about twice) than regularly processed ceramic tubes initially, which is possibly due to HIP eliminating surface voids or cracks from the post machining process, and consequently requiring a much higher load to induce cracking on the ceramic surface. However, the cracking load of HIP processed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I l l ceramic A tubes decreased dramatically after about 200 hours of soaking and reached the same level of that of regularly processed ceramic A tubes, as shown in Figure 5.35. HIP processed and regularly processed ceramic A tubes have almost the same cracking loads after over 200 hours soaking in 127°C steam and almost constant after that. This could probably be attributed to the fact that the cracking load needed to initiate micro cracks and induce macro-cracks on the ceramic surface is determined by the condition of the surface layer of the ceramic. Surfaces of both HIP processed and regularly processed ceramic A tubes after 200 hours soaking in 127°C steam have become so porous that a monoclinic phase containing micro cracks develops, as seen in Figures 5.36 (b) and 5.28 (b), so relatively low loads are needed to initiate micro-cracks and induce macro-cracking. w Td) y ■ o c 3 100 1 o a. 80 4-* C o „ 11 60 Q . O ) 40 c IS pan o (0 O n l ♦ Ffegular ■ HFped 100 200 300 400 500 600 Soaking Time (hours) 700 800 900 Figure 5.35: Cracking Forces in Three-point Bending Test of Regular and HIP processed 2mm Diameter Ceramic A Tubes Decreased in 127°C Soaking Test Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 Figures 5.36 and 5.37 are transform ed layers on outer and inner surfaces of HIP processed ceramic A tubes, after soaking in 127°C steam. Figures 5.38 summarized the transformed layer growth of HIP processed ceramic A tubes soaking in 127°C steam. It is demonstrated that the outer surface transformed at a slightly higher rate than the inner surface, same as previously found on regularly processed 2mm diameter ceramic A tubes. The explanation for this finding is that the inner surface has a higher geometric self-restriction than the outer surface. It was found that the transformed layer growth rates in HIP processed 2mm diameter A ceramic tubes are slightly slower than that of regularly processed ceramic A tubes, as shown in Figure 5.38, which is consistent w ith the conclusion drawn from the XRD data shown in Figure 5.34. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 113 (d) 610 hours (e) 800 hours 1 5 k U X 5 0 0 5 y M m Q 0 W b J L + ti i o S E I Figure 5.36: Transformed Layer on Outer Surfaces of HIP processed 2mm Diameter Ceramic A Tubes in 127°C Soaking Test (b) 202 hours (a) 0 hours l S f c U X 5 0 0 5 0 » m 0 0 0 9 4 2 Z S ' S E I Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 4 (d) 610 hours (c) 403 hours (e) 800 hours Figure 5.37: Transformed Layer on Inner Surfaces of HIP processed 2mm Diameter Ceramic A Tubes in 127°C Soaking Test y=0.0668x-16.229 B F T = 0 . 9 6 0 5 / y=0.0754<-16.758 FT=09606 ♦ Inner Surface ■ Cuter Surface 800 icjoo 400 600 200 -10 Soaking time (hours) Figure 5.38: Transformed Layer Growth Rates of HIP processed 2mm Diameter Ceramic A Tubes Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 115 HIP processing enhances the density of ceramics and inhibited water from further penetration into the ceramic bulk, and so reduces the degradation of the ceramic materials; however, the anti-degradation improvement is limited. On one hand, the regularly processed ceramic A tubes have a density of about 6.0 gram/cm , which is close to (about 98.3%) the specific density. HIP processing doesn’t increase the density too much, to about 6.05 grams/cm3 (99.2% of specific density). The effect of hindering water from penetrating is insignificant, so its anti-degradation is limited. On the other hand, as the cracking load during a three-point bending test is mainly determined by the surface condition of the ceramic, HIP processing eliminates surface micro-cracks induced by machining and dramatically increases the cracking load value for the non-aged samples. However, after 200 hours of soaking, a thin layer of transformed monoclinic phase appeared on both HIP processed and regular process ceramics and the cracking loads for both ceramics become comparable. 5.3.11 Protection Effect of Ion Beam Assisted Deposited AI2 O3 Coating It has been shown that the low temperature degradation of 3Y-TZP will occur only where aqueous solution is present [30-32], A concept of applying an insulative and biocompatible coating on 3Y-TZP surfaces to prevent 3Y-TZP from degradation has been realized [33]. As discussed previously, AI2O3 is a biocompatible material and has become a candidate for the coating. A coating of Ion beam assisted Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 deposition (IBAD) AI2O3 has been preliminarily investigated [34], Figure 5.39 presents a schematic illustration of an IB AD setup. Component (s) Holder & M onitor I r i t i ; r . i c . * i » ' t 5 il»J e-beam Evaporator Figure 5.39: Schematic of IB AD Experimental Apparatus The base vacuum level for the IBAD process was about 1.33xlO" 5 mbar. The working pressure of argon and oxygen was 4x10" mbar. The flow rates of Argon to the ion gun and IBAD chamber were 10 and 3.5 standard cubic centimeters per minute (seem), respectively. The flow rate of Oxygen channel to the IBAD chamber was 5.5 seem, and the substrate temperature during IBAD process is about 300°C. The electron beam evaporation source is a sapphire AI2O3 solid block with a purity of 99.99 at. %. The IBAD deposition rate is 1.5A/s. The ion beam bombarding energy was set at lOOOeV and the current was 26mA. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A 1.6-micron thick alumina coating was applied by IBAD on a sealed dummy ceramic case comprised of ceramic B. X-ray diffraction analysis was performed on this unit prior to and after 20, 85, 137, and 201 hours soaking in 127°C steam. The X-rays penetrate the thin alumina layer and allow peak detection of 2 Theta angles of 28.2, 30.2 and 31.3 degrees, as shown in Figure 5.40. The monoclinic phase fraction of the IBAD processed part was calculated by the modified Garvie-Nicholson equation. As presented in Figure 5.41, the monoclinic phase percentage changes with the increase of soaking time. Initially the phase transformation rate of the ceramic coated with 1.6 microns alumina was much slower compared to ceramic without alumina coating. After 150 hours, the monoclinic phase increased abruptly. The possible reason for this sudden increase of monoclinic phase is that the moisture finally penetrated through the A120 3 barrier after 150 hours soaking. The alumina coated ceramic self-destructed after 201 hours of soaking when the monoclinic phase was 49%, short of the 70% saturation level. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 Figure 5.40: XRD Profiles of AI2O3 Coated 3Y-TZP Ceramic Prior to and After 20, 85, 137 and 201 Hours of Soaking in 127°C Steam 0 50 1 0 0 1 5 0 200 250 Soakign time (Hours) Figure 5.41: Comparison of Phase Transformation of T-ZrC> 2 with and without AI2O3 Coating It has been shown, that IBAD processed A120 3 coating has a protective effect in terms of preventing 3Y-TZP ceramic from moisture-induced degradation. Further investigation is needed to address the remaining questions, such as what is the optimal thickness of the AI2O3 coating so that the coating can provide effective barrier for water or water vapor from penetrating through but the coating itself w on’t ♦ With A 12CG coating ■ Without AI206 coating Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 1 9 flake, and why the coated 3Y-TZP self destructed when the monoclinic phase was only 49%, short of the 70% saturation level, which has not been seen on the ceramic tubes without coating. 5.3.12 Estimated Lifetime of BION® Microstimulator Package Besides the flaking-off issue due to LTD of ceramics, the package’s lifetime is considered to be exceeded if the sealed case fails hermeticity or is unable to sustain a 15 lb 3-point bend load. No flaking-off was observed on dummy units brazed with ceramic A after 1368 hours in 127°C steam, and they all remained hermetic and sustained 19 lbs or more of bend load. 1368 hours aging in 127°C steam is equivalent to 70 years implantation in 37°C body based on the Arrhenius relationship obtained via the activation energy (Ea) approach. 5.4 Conclusions The T-M transformation did occur on all ceramics tested in moist environment at body temperature. Arrhenius factors obtained through different approaches with three different materials agreed well. BION® microstimulators with different ceramics will have different implantation lifetimes. Ceramic A has the best anti-degradation properties. It is concluded that implants with ceramic A will remain hermetic and retain the ability to withstand a minimum of 15 pounds of cracking load in three point bending test after 70 years of implantation in a human body. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 HIP process had some impact on the T-M phase transformation of ceramics and a slightly longer implantation lifetime is expected for the devices that are manufactured with HIP processed ceramic material. IBAD processed A120 3 coating demonstrated protective effect initially in terms of preventing ceramic from T-M phase transformation. Further investigation is recommended. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 121 5.5 References 1. Tsukuma K, K ubota Y, and Tsukidate T, Advances in Ceramics, Vol. 12, Science and Technology of Zirconia II. Edited by N. Claussen, M. Ruhle, and A. H. Heuer. American Ceramic Society, Columbus, OH, p. 382, 1984. 2. Sato T, and Shimada M, Transformation of Yttria-doped Tetragonal Zr02 Polycrystals by A nnealing in Water. J A m Cer. Soc., 68 (6): 356-369, 1985. 3. Zirconia Ceramics, 9, edited by S. Somiya and M. Yoshimura, Uchida Rokakuho Publishing Co., Ltd, Tokyo., 1987. 4. W atanabe M, Iio S, and Fukuura I, Advances in Ceramics, Vol. 12, Science and technology of Zirconia II. Edited by N. Claussen, M. Ruhle, and A. H. Heuer. American Ceramic Society, Columbus, OH, p. 391, 1984. 5. Sato T and Shimada M., J. Mater. Sci., 20, 3988, 1985. 6. Lange FF, Dunlop GL and Davis BI, Degradation During Aging of Transformation-Toughened Z r02-Y 203 M aterials at 250°C, J. Am., Ceram. 69, 237-240, 1986. 7. Tsukuma K and Shimada M, J. Mater. Sci. Lett., 49, 857, 1985. 8. Chevalier J, Olagnon C, and Fantozzi G, Subcritical Crack Propagation in 3Y-TZP Ceramics, J. Am. Ceram. Soc., 82 [11] 3129-38 1999. 9. Li J and W atable R, Phase Transformation in Y203-Partially-Stabilized Z r02 Polycrystals of Various Grain Sizes during Low-Temperature Aging in Water, J. Am. Ceram. Soc., 81 [10] 2687-91, 1998. 10. Swab JJ, Low Temperature Degradation of Y-TZP Materials, Journal of Materials Sciences, 26 (1991) 6706-6714.J. L. Drummond, In-vitro Aging of Yttria-Stabilized Zirconia. J. Amer. Ceram. Soc., 72 (4): 675-6, 1989. 11. Thomson I and Rawlings RD, M echanical Behavior of Zirconia and Zirconia-toughened Alumina in A Simulated Body Environment, Biomaterials, 11: 505-8, 1990. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 12. Drummond JL, In Vitro Aging of Yttria-stabilized Zirconia, J Amer Ceram Soc, 72(4): 675-6, 1989. 13. Drouin JM, Cales B., Yttria-stabilized Zirconia for Improved Hip Joint Head. Anderson OH, Yli-Urpo A, editors. Bioceramics 7. London: Butterworth-Heinemann Publ., 387-94, 1994. 14. Shimizu K, Oka M, Kumar P et al, Time-dependent Changes in the M echanical Properties of Zirconia Ceramic. J Biom ed Mat Res, 27:729- 34, 1993. 15. Fujisa A, Shimotoso T, Masuda S, M akinouchi k., The Development of Zirconia Balls for THR with a High M echanical Strength, Low Phase Transformation. In: Kokubo T, N akam ura T, Miyaji F, editors. Bioceramics 9. Amsterdam: Elsevier Science Publ., 503-6, 1996. 16. Cales B, Stefani Y., Mechanical Properties and Surface Analysis of Retrieved Zirconia Femoral Hip Joint Heads after An Implantation Time of Two to Three Years. J. Of M ater Sci: M ater in Med, 5:376-80, 1994. 17. M urray MGS, Pryce AW, Stuart JW. A Comparison of Zyranox Zirconia Femoral Heads Before and 11/2 Years After Implantation. Ravaglioli A, editor. 4th Euro-Ceramics, vol. 8, Bioceramics. Faenza (Italy): Faenza Editrice Publ., 37-44, 1995. 18. M urray MGS, Pryce AW. A Physical, Chemical and Mechanical Evaluation of A Retrieved Zyranox Zirconia Femoral Heads Before and 1-1/2 Years after implantation. In: Trans, 5th W orld Biomaterials Congress. Toronto, Canada, 1996. 19. Jiang G, Purnell K, Mobley P and Schulman J, Accelerated Life Tests and In-Vivo Test of 3Y-TZP Ceramics, Proceedings of Materials & Processes for Medical Devices Conference, Anaheim, CA, 477-82, Sept. 2003. 20. Johnson W A and Mehl RF, Reaction Kinetics in Processes of Nucleation and Growth, Trans. Am. Inst. Min., Metall. Pet. Eng., 135, 416-41, 1939. 21. Christian JW, The Theory of Transformations in Metals and Alloys, 2n d ed.; ppl-19. Pergamon Press, Oxford, U.K., 1965. 22. Chevalier J, Drouin, JM and Cales B, Low Temperature Aging Behavior of Zirconia Hip Joint Head; pp. 135-38 in Bioceramics, Vol. 10, Proceedings of the 10th international Symposium on Ceramics in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 M edicine (Paris, France, Oct. 5-9, 1997), International Society of ceramics in Medicine, San Bernardino, CA, 1997. 23. M endelson MI, Average Grain Size in Polycrystalline Ceramics, J. Am. Ceram. Soc., 52 [8] 443-46, 1969. 24. Chevalier J, Low Temperature Aging of Y-TZP Ceramics, J. Am. Ceram. Soc., 82 [8] 2150-54, 1999. 25. Tubakino H, Hamamoto M, and Nozato R, Tetragonal-to-M onoclinic Phase Transformation during Thermal Cycling and Isothermal Aging in Partially Stabilized Zirconia, J. Mater. Sci., 26, 5521-26, 1991. 26. Tubakino H, Hamamoto M , and Nozato R, M artensite Transformation Behavior during Isothermal Aging in Partially Stabilized Zirconia with or without Alumina addition, J. Mater. Sci., 12, 196-98, 1993. 27. Zhu W, Lei T and Zhou Y, Time Dependent Tetragonal-to-M onoclinic Transition in Hot-Pressed Zirconia Stabilized with 2 mol.% Yttria, Mater. Chem. Phys., 34, 317-20, 1993. 28. Belehradek J, Temperature and Living Matter. Protoplasma- Monographien. Vol. 8, Bomtraeger, Berlin, 1935. 29. Jiang G and Purnell K, M ethod of M inimizing Low-temperature Degradation of Yttria-Stabilized Zirconia, US Patent Application, Pending. 30. Yoshimura M, Nom a T and Kawabata K, Role of H 2 0 on the Degradation Process of Y-TZP, J. Mater. Sci. Letters, 6:4 465-467, 1987. 31. Ho FY and Wei W CJ, Dissolution of Yttrium Ions and Phase Transformation of 3Y-TZP Powder in Aqueous Solution, J. Am. Ceram. Soc., 82:6 1614-16, 1999. 32. Kim YS, Jung CH and Park JY, Low Temperature Degradation of yttria- stabilized Tetragonal Zirconia Polycrystals Under Aqueous Solutions, J. of Nuclear Mater. 209:326-331, 1994. 33. Koh YH, Kong YM, Kim S and Kim HE, Improved Low-Temperature Environmental Degradation of Yttria-stabilized tetragonal Zirconia Polycrystals by Surface Encapsulation, J. Am. Ceram. Soc., 82:6 1456- 58, 1999. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 34. Jiang G, Material and M ethod to Prevent Low Tem perature Degradation of Zirconia in Biomedical Implants, US Patent Application, Pending. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 Chapter 6: Conclusions This work has focused on a portion of the development process of the BION® microstimulator package, and consists of three main parts: 1). Development of a 3Y-TZP ceramic to Ti-6A1-4V brazed joint 2). Biocompatibility of the 3 Y-TZP ceramic to Ti-6A1-4V brazed joint 3). Low temperature degradation of 3Y-TZP ceramics and life prediction of the BION® microstimulator package Strong joints have been produced with both Tini-50® filler material and pure nickel filler metal. Brazing parameters have a strong influence on the flexural strength of the both types of joints. An attempt was made to explain the dependence of flexural strength to brazing parameters. Three brazing param eter dependent factors: the extent of chemical reaction at the interface, the residual stress in the joint and the amount or thickness of brittle reaction products were proposed and discussed. Three failure modes were identified during the four-point flexural strength test. The exact failure mode is a function of the brazing parameters for both types of brazed joints. The highest fracture strength of 257 M Pa with a standard deviation of 67 M Pa was obtained with joints brazed at 1035°C for 60 minutes using 0.050mm thick pure nickel filler metal. The N i2 Ti40 phase form ed at the ceramic to metal interface is responsible for the bond development. SEM-EDS, XRD and electrochemical analyses indicated, that nickel from ( R ) either Tini-50 filler material or pure nickel filler metal has bonded to other elements Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 and that no elemental nickel is present in the brazed joint after the 1035°C for 5 minutes brazing process. Immersion test revealed that the nickel release rate from the brazed joints is slower than the as-ground Nitinol sample, which provides evidence to believe that the brazed joints have excellent corrosion resistance in PBS solution and are considered to be biocompatible. Animal trials were conducted using ( r ) BION microstimulators containing the 3Y-TZP-to-Ti-6Al-4V brazed joints with Tini-50® filler configuration, and which provided further evidence of the safe biocompatibility of the brazed joint. BION® microstimulators utilizing Tini-50® brazed cases have been used in some implantation applications and many clinical trials. Nickel brazed joints have identical microstructures to Tini-50® brazed joints, thus safe biocompatibility is also expected. This work has verified that the moisture-induced T-M phase transformation obeys Arrhenius law. A reliable lifetime prediction method has been developed for the BION® microstimulator package and for an implantable biomedical device containing 3Y-TZP ceramic. It is concluded that the BION® microstimulator package utilizing the current 3Y-TZP ceramic is adequate for up to 70 years of implantation. The HIP process increased the density of the ceramic material further, and provided for stronger ceramic materials. However, accelerated aging test with HIP processed and regularly processed ceramics in 127°C steam revealed that the HIP process only slowed the phase transformation slightly, possibly because the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill regular processed ceramic is already close to its theoretical density, thus the anti degradation im provem ent using HIP processed ceramic is limited. Ion beam assisted deposited AI2O3 coating has been preliminary investigated, and it has been shown that an IBAD produced AI2O3 coating has a protective effect in terms of preventing the 3Y-TZP ceramic from the moisture-induced degradation. Further investigation is needed to completely understand some remaining questions and bring this process to practical application. The successful development of 3Y-TZP ceramic to Ti-6A1-4V braze joint provides a promising packaging approach for many long-term implantable biomedical devices. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1 28 Bibliography Agathopoulos S, Correia RN, Joanni E and Fernandes JRA, Interactions at Zirconia- Au-Ti Interfaces at High Temperatures. K ey Eng Mater, 206-213 487-90, 2002. Agathopoulos S, Moretto P, Peteves SD, Emiliano JV, Correia RN, Brazing of Zirconia to Ti and Ti6A14V. 1996 Am. Ceram. Soc. M eeting, Indianapolis, USA 1996, Ceramic Joining, Ceramic Transactions. Indianapolis, Vol. 77:75-82, 1997. Agathopoulos S, Pina S, and Correia RN, A Review of Recent Investigations on Zirconia Joining for Biomedical Applications. Ceramic Transactions, Vol. 138:135- 147, 2002. American W elding Society, W elding Handbook, 7th Ed., Vol. 2, 1978. Arcos I, Davis R, Fey K, M ishler D, Sanderson D, Tannacs C, Vogel MJ, W olf R, Zilberman Y, and Schulman J., Second-Generation M icrostimulator. Artificial Organs, 26 (3): 228-231, 2001. Arroyave R, Eagar TW, Metal Substrate Effects on the Thermochemistry of Active Brazing Interfaces. Acta Mater. 51: 4871-4880, 2003. Assad M, Lemieux N, Rivard CH, Yahia L., Comparative In Vitro Biocompatibility of Nickel-titanium, Pure nickel, Pure titanium, and Stainless Steel: Genotoxicity and Atomic Absorption Evaluation. Biomed. Mater. Eng. 9:1-12, 1999. Baker H, et al. ASM Handbook, 3 Alloy Phase Diagrams, 319, 1992. Belehradek J, Temperature and Living Matter. Protoplasma-Monographien. Vol. 8, Bomtraeger, Berlin, 1935. Bogdanski D, Koller M, M uller D, M uhr G, Bram M, Buchkremer HP, Stover De, Choi J andEpple M., Easy Assessment of the Biocompatibility of Ni-Ti Alloys by In Vitro Cell Culture Experiments on A Functionally Graded Ni-NiTi-Ti Material. Biomaterials 23: 4549-4555, 2002. Bulter EP, Transformation Toughened Zirconia Ceramics. M at Sci Tech, 1:417-32, 1985. Cales B and Stefani Y., Mechanical Properties and Surface Analysis of Retrieved Zirconia Femoral Hip Joint Heads After An Implantation Time of Two to Three Years. J. Of Mater Sci: Mater in Med, 5:376-80, 1994. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 Cambridge M aterials Selector, A Materials Database - V ersion 2.04, Granta Design Ltd, UK, 1997. Chevalier J, Low Temperature Aging of Y-TZP Ceramics. J. Am. Ceram. Soc., 82:8, 2150-54, 1999. Chevalier J, Olagnon C, and Fantozzi G, Subcritical Crack Propagation in 3Y-TZP Ceramics. J. Am. Ceram. Soc., 82 [11] 3129-38, 1999. Chevalier J, Drouin JM, and Cales B., Low Temperature Aging Behavior of Zirconia Hip Joint Head, 135-38 in Bioceramics, Vol. 10, Proceedings of the 10th International Symposium on Ceramics in Medicine (Paris, France, Oct. 5-9, 1997), International Society of ceramics in Medicine, San Bernardino, CA, 1997. Christian JW, The Theory of Transformations in Metals and Alloys. 2nd ed., 1-19. Pergamon Press, Oxford, U.K., 1965. Correia RN, Emiliano JV and M oretto P, Microstructure of Diffusional Zirconia- Titanium and Zirconia-Ti6A14V Alloy Joints. J. M aterial Sci., Vol. 33, 215-221, 1998. Drouin JM, Cales B., Yttria-stabilized Zirconia for Im proved Hip Joint Head. In: Anderson OH, Yli-Urpo A, editors. Bioceramics 7:387-94. London: Butterworth- Heinemann Publ., 1994. Drummond JL, In Vitro Aging of Yttria-stabilized Zirconia. J Amer Ceram Soc, 72(4): 675-6, 1989. Drummond JL, Effects of In Vitro Aging of M agnesia-stabilized Zirconia. J. Amer. Ceram. Soc., 75:50, 1278-80, 1992. Dupont AC, Bagg SD, Chun S, Creasy JL, Romano C, Romano D, Waters RL, W ederich CL, Richmond FJR and Loeb GE, Clinical Trials of BION™ Microstimulators, Proc. IFESS, LJUBLJANA, SLOVENIA, 2002. Dupont AC, Loeb GE, Richmond FJR, Bagg SD, Creasy JL, Romano C, Romano D., Clinical Trials of BION Injectable Neuromuscular Stimulators, Proc. IFESS Cleveland, Ohio, 2001. Fey KE and Jiang G, Application and Manufacturing M ethod for A Ceramic to M etal Seal, US Patent: 6,521,350 B2. 2003. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 Fraker AC, Ruff AW, Sung P, Orden ACV, and Speck KM, Surface Preparation and Corrosion Behavior of Titanium Alloys for Surgical Implants. T i’80 Science and Techn., Plenum Press, 2447, 1980. Fujisa A, Shimotoso T, M asuda S, M akinouchi K, The Development of Zirconia Balls for TFCR with A High M echanical Strength, Low Phase Transformation. In: Kokubo T, Nakamura T, M iyaji F, editors. Bioceramics 9:503-6. Amsterdam: Elsevier Science Publ., 1996. Galloway NTM, El-Galley RES, Sand PK, Appell RA, Russel FTW, and Carlan SJ, Extracorporeal Magnetic Innervation Therapy for Stress Urinary Incontinence, Urology, 53:1108, 1999. Garvie RC, H anninkR H , Pascoe RT., Ceramic Steel? Nature 258:703-4, 1975. Gogotsi GA, Lamonova EE, Furmanov YA and Savitskaya IM, Zirconia Crystals Suitable for Medicine: 1 Implants. Ceramics Int., 20:343-348, 1994. Griffin EA, Mumm DR and M arshall DB, Rapid Prototyping of Functional Ceramic Composites. Am. Ceram. Soc. Bulletin, Volume 75:65-68, 1996. Guillou MO, Henshall JL, Hooper RM and Carter GM, Indentation Fracture Testing and Analysis, and Its Application to Zirconia, Silicon Carbide and Silicon Nitride Ceramics. J. Of Hard Materials; Volume 3:421-434, 1992. Hamill JA Jr, Int. J. Powder Metal. 27:4, 363, 1991. Heetderks W J, RF Powering of M illimeter- and Submillimeter-Sized Neural Prosthetic Implants, IEEE Trans. Biomed. Eng. 35: 323-327, May 1988. Ho FY and Wei WCJ, Dissolution of Yttrium Ions and Phase Transformation of 3Y- TZP Powder in Aqueous Solution, J. Am. Ceram. Soc., 82:6 1614-16, 1999. Ishikawa N, Suda S, Sasaki T, Yamanishi T, Hosaka H, Yasuda K, and Ito H, Development of A Non-invasive Treatment System for Urinary Incontinence Using A Functional Continuous Magnetic Stimulator (FCMS), Medical & Biological Engineering & Computing, November, 704, 1998. Jiang G, Material and method to Prevent Low Temperature Degradation of Zirconia in Biomedical Implants, US Patent Application, Pending. Jiang G and Pumell K, M ethod of M inimizing Low-temperature Degradation of Yttria-Stabilized Zirconia. US Patent Application, Pending. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 Jiang G, Purnell K, M obley P and Schulman J, Accelerated Life Tests and In-Vivo Test of 3Y-TZP Ceramics, Proceedings of M aterials & Processes for Medical Devices Conference, Anaheim, CA, 477-82, Sept. 2003. Jiang G, M ishler DR, Davis R, M obley JP, Schulman JH, Zirconia to Ti-6A1-4V Braze Joint for Implantable Biomedical Device, J. of Biomed M ater Res, Part B: Appl Biomaterial, 72B: 316-321, 2005 Johnson W A and M ehl RF, Reaction Kinetics in Processes of Nucleation and Growth, Trans. Am. Inst. Min., Metall. Pet. Eng., 135, 416-41, 1939. Kimel RA and Adair JH, Aqueous Degradation and Chemical Passivation of Yttria Tetragonally Stabilized Zirconia. J Amer Ceram Soc., 85[6], 403-408, 2002. Kim YS, Jung CH and Park JY, Low Temperature Degradation of yttria-stabilized Tetragonal Zirconia Polycrystals Under Aqueous Solutions, J. of Nuclear Mater., 209:326-331, 1994. Klomp JT, W ith GD, Mater. & M anuf. Proc. 8:2, 129, 1993. Koh YH, Kong YM, Kim S and Kim HE, Improved Low-Temperature Environmental Degradation of Yttria-stabilized tetragonal Zirconia Polycrystals by Surface Encapsulation, J. Am. Ceram. Soc., 82:6 1456-58, 1999. Kpanen A, Ryhanen J, Danilov A, Tuukkanen J., Effect of Nickel-titanium Shape Memory M etal Alloy on Bone Formation. Biomaterials 22:2475-80, 2001. Kubasckewski O and Dench W A, The Dissociation Pressures in the Zirconium- Oxygen System at 1000°C. J. Inst. Metals, 84, 440-444, 1995-56. Lange FF, Transformation toughening. Part 3 - Experimental Observations in the Z r0 2-Y20 3 System. J. M ater Sci., 17:240-6, 1982. Lange FF, Dunlop GL and Davis BI, Degradation During Aging of Transformation- Toughened Zr0 2-Y203 M aterials at 250°C. J. Am., Ceram. 69, 237-240, 1986. Lasater BJ, Methods for Hermetically Sealing Ceramic to Metallic Surfaces and Assemblies Incorporating Such Seal, US Patent: 6,221,513 B l, 2001. Leyens C and Peters M, Titanium and Titanium Alloys: Fundamentals and Applications, 457-498. W iley-VCH, 2003. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 Li J and W atable R, Phase, Transformation in Y 203-Partially-Stabilized Z r0 2 Polycrystals of Various Grain Sizes during Low-Temperature Aging in Water. J. Am. Ceram. Soc., 81[10] 2687-91,1998. Lin VW, Hsieh C, Hsiao IN, and Canfield J, Functional M agnetic Stimulation of Expiratory Muscles: A Noninvasive and New M ethod for Restoring Cough. Journal of Applied Physiology, 84, 1144, 1998. Loeb GE, Zamin CJ, Schulman JH, and Troyk P, Injectable M icrostimulator for Functional Electrical Stimulation, North Sea: Transducers and Electrodes, November, 1991. Loeb GE and Richmond FJR, BION® Implants for Therapeutic and Functional Electrical Stimulation, Neural Prostheses for Restoration of Sensor and M otor Function, eds. J.K. Chapin, K.A. M oxon and G. Gaal, CRC Pres, Boca Raton, FL, 2000. Martinelli AE, Diffusion Bonding of Silicon Carbide and Silicon Nitride to Molybdenum. Ph.D. Thesis, McGill University, Montreal, Canada, 1996. Mendelson MI, Average Grain Size in Polycrystalline Ceramics. J. Am. Ceram. Soc., 52:8,443-46,1969. Murray MGS, Pryce AW, A Physical, Chemical and M echanical Evaluation of A Retrieved Zyranox Zirconia Femoral Heads Before and 1-1/2 Years After Implantation. In: Trans, 5th W orld Biomaterials Congress. Toronto, Canada, 1996. Murray MGS, Pryce AW and Stuart JW , A Comparison of Zyranox Zirconia Femoral Heads Before and One and A H alf Years After Implantation. In: Ravaglioli A, Editor. 4th Euro-Ceramics, vol. 8:37-44, Bioceramics. Faenza (Italy): Faenza Editrice Publ., 1995. Peytour C, Berthet P, Barbier F, Revcolevschi A, Interface M icrostructure and Mechanical Behavior of Brazed Ti6A14V/Zirconia Joints, J. Mater. Sci. Lett., 9: 1129-31, 1990. Piconi C and Maccauro G, Review: Zirconia as a Ceramic Biomaterial. Biomaterials 20:1-25, 1999. Porstendorfer J, Reineking A, W illert HG, Radiation Risk Estimation Based on Activity Measurements of Zirconium Oxide Implants, J. Biomedical Materials Research, 32:4, 663-667, 1996. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 133 Rauh EG and Garg SP, The Zr02-x(cubic)-Zr02-x(cubic+tetragonal) Phase Boundary, J. Am. Ceram. Soc., 63 [3-4], 239-240, 1980. Rocha LA, Ariza E, Costa AM, Oliveira FJ and Silva RF., Electrochemical Behavior of Ti/Al203 Interfaces Produced by Diffusion Bonding. M ater Res, 6:4 439-444, 2003. Santella M L and Pak JJ, Brazing Titanium-Vapor-Coated Zirconia. W elding Research Supplement, 165-172, 1993. Sato T, Shimada M., Control of the Tetragonal-to-monoclinic Phase Transformation of Yttria Partially Stabilized Zirconia in Hot Water. J. M ater Sci 20:3899-992, 1985. Sato T, Shimada M., Transformation of Yttria-doped Tetragonal Z r0 2 Polycrystals by Annealing in Water. J Amer Ceram Soc 68(6) 356-9, 1985. Schneider SJ (ed.), Ceramics and Glasses, in Engineered Materials Handbook, Volume 4, ASM International, 1991. Schwartz MM, Brazing, ASM Int., 1995. Schwartz MM, Handbook of Structural Ceramics, McGraw-Hill Publishers, USA 1992. Scott HG, Phase Relationship in Zirconia-yttria Systems. J. M ater Sci 10:1527-35, 1975. Shimizu K, Oka M, Kumar P et al., Time-dependent Changes in the M echanical Properties of Zirconia Ceramic. J Biomed M at Res, 27:729-34, 1993. Singh J, Peck RA and Loeb GE, Development of BION Technology for functional electrical stimulation: Hermetic Packaging , Proc. IEEE-EMBS (Istanbul, Turkey), 2001. Somiya S. and Yoshimura M. editor, Zirconia Ceramics, 9, Uchida Rokakuho Publishing Co., Ltd, Tokyo. 1987. Subbarao EC, Zirconia - An Overview. In: Heuer AH, Hobbs LW, editors. Advances in Ceramics. Vol.3:1-24. Science and Technology of Zirconia. Amsterdam: Elsevier, 1981. Suganuma K, Mater. Res. Soc. Symp. Proc., 314:51,1993. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 Swab JJ, Low Temperature Degradation of Y-TZP M aterials. Journal of Materials Sciences, 26:6706-6714, 1991. Thomson I and Rawlings RD, Mechanical Behavior of Zirconia and Zirconia- toughened Alumina in A Simulated Body Environment. Biom aterials 11: 505-8, 1990. Tsukuma K, Kubota Y, and Tsukidate T, Advances in Ceramics, Vol. 12:382, Science and Technology of Zirconia II. Edited by N. Claussen, M. Ruhle, and A. H. Heuer. American Ceramic Society, Columbus, OH, 1984. Tsukuma K. and Shimada M., J. Mater. Sci. Lett., 49:857, 1985. Tsukuma K and Shimada M, Strength, Fracture Toughness and Vickers Hardness of Ce02-stabilized Tetragonal Z r0 2 Polycrystals (CeTZP), Journal of M aterials Science 20:1178-1184, 1985. Tubakino H, Hamamoto M, and Nozato R, Tetragonal-to-M onoclinic Phase Transformation during Thermal Cycling and Isothermal Aging in Partially Stabilized Zirconia, J. Mater. Sci., 26, 5521-26, 1991. Tubakino H, Hamamoto M, and Nozato R, M artensite Transformation Behavior during Isothermal Aging in Partially Stabilized Zirconia with or without Alumina Addition, J. Mater. Sci., 12, 196-98, 1993. Uo M, Watari F, Yokoyama A, Matsuno H and Kawasaki T., Dissolution of Nickel and Tissue Response Observed by X-ray Scanning Analytical Microscopy. Biomaterials 20: 747-55, 1999. Watanabe M, Iio S, and Fukuura I, Advances in Ceramics, Vol. 12:391, Science and Technology of Zirconia II. Edited by N. Claussen, M. Ruhle, and A. H. Heuer. American Ceramic Society, Columbus, OH, 1984. W ever DJ, Veldhuizen AG, Sanders MM, Schakenraad JM, van Horn JR., Cytotoxic, Allergic and Genotoxic Activity of A Nickel-titanium Alloy. Biomaterials 18:1115- 20, 1997. W hitney ED, Ceramic Cutting Tools, 166, Chapter 7, Noyes Publications, USA, 1994. Yoshimura M, Noma T, Kawabata K, Somiya S., Role of H20 on the Degradation Process of Y-TZP, J Mater Sci Lett 6:465, 1987. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Zhu W, Lei T and Zhou Y, Time dependent Tetragonal-to-M onoclinic Transition in Hot-Pressed Zirconia Stabilized with 2 mol.% Yttria, Mater. Chem. Phys., 34, 317- 20, 1993. Zitter H, Plenk Jr, The Electromechanical Behavior of M etallic Implant Materials as an Indicator of their Biocompatibility. J. Biomed. Mater. Res. 21:881, 1987. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Effects of prenatal cocaine exposure in quantitative sleep measures in infants

PDF

English phoneme and word recognition by nonnative English speakers as a function of spectral resolution and English experience

PDF

Cardiorespiratory interactions in sleep apnea: A comprehensive model

PDF

Cellular kinetic models of the antiviral agent (R)-9-(2-phosphonylmethoxypropyl)adenine (PMPA)

PDF

Head injury biomechanics: Quantification of head injury measures in rear-end motor vehicle collisions

PDF

A model of upper airway dynamics in obstructive sleep apnea syndrome

PDF

Destructive and non-destructive approaches for quantifying the effects of a collagen cross-linking reagent on the fatigue resistance of human intervertebral disc

PDF

Comparisons of deconvolution algorithms in pharmacokinetic analysis

PDF

Hyperspectral optical imaging for detection, diagnosis and staging of cancer

PDF

Finite element analysis of the effects of stem geometry, surface finish and cement viscoelasticity on debonding and subsidence of total hip prosthesis

PDF

A model of cardiorespiratory autoregulation in obstructive sleep apnea

PDF

Evaluation of R.F. transmitters for optimized operation of muscle stimulating implants

PDF

The effects of the vitreous state on the ocular pharmacokinetics following intravitreal drug delivery: an evaluation with quantitative magnetic resonance imaging

PDF

Characteristics and properties of modified gelatin cross-linked with saline for tissue engineering applications

PDF

Bayesian inference using Markov chain Monte Carlo methods in pharmacokinetic /pharmacodynamic systems analysis

PDF

Design of a portable infrared spectrometer: application to the noninvasive measurement of glucose

PDF

Systematic identification of potential therapeutic targets in cancers using heterogeneous data

PDF

Cardiorespiratory variability in normal and sleep disordered breathing states: a comprehensive model

PDF

A finite element model of the forefoot region of ankle foot orthoses fabricated with advanced composite materials

PDF

Propofol Effects On Eeg And Levels Of Sedation

Asset Metadata

Creator Jiang, Guangqiang (author)

Core Title Development of ceramic-to-metal package for BION microstimulator

School Graduate School

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

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

Tag engineering, biomedical,engineering, materials science,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Hedman, Tom (committee chair), Schulman, Joseph H. (committee chair), D'Argenio, David (committee member), Khoo, Michael C.K. (committee member), Mansfield, Florian (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-453098

Unique identifier UC11340782

Identifier 3196824.pdf (filename),usctheses-c16-453098 (legacy record id)

Legacy Identifier 3196824.pdf

Dmrecord 453098

Document Type Dissertation

Rights Jiang, Guangqiang

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 au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA