Citation
Design of Liquid-Assisted Self-Healing Metal-Matrix Composites

Material Information

Title:
Design of Liquid-Assisted Self-Healing Metal-Matrix Composites
Creator:
Fisher, Charles R
Publisher:
University of Florida
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
Myers, Michele V
Committee Members:
Fuchs, Gerhard E
Mecholsky, John J, Jr
Sinnott, Susan B
Mareci, Thomas H
Graduation Date:
8/10/2013

Subjects

Subjects / Keywords:
Alloys ( jstor )
Eutectics ( jstor )
Healing ( jstor )
Heat ( jstor )
Heat treatment ( jstor )
Liquids ( jstor )
Materials science ( jstor )
Mechanical properties ( jstor )
Self healing materials ( jstor )
Solid solutions ( jstor )
composite
design
metal-matrix
self-heal
City of Gainesville ( local )

Notes

General Note:
Advancements in materials science have enabled components to be fabricated lighter, stronger, and more functional than ever. However,damage mitigation in these materials still limits their usable lifetimes.Recently, materials with the ability to heal structural damage have shown promise as potential novel materials of the future. Within metallic-based systems, metal-matrix composites reinforced with shape memory alloys have the potential to demonstrate dramatic capabilities in damage mitigation and repair. Investigations of liquid-assisted self-healing in metal-matrix composites have centered on developing a high specific-strength matrix possessing a low-melting eutectic for use as the healing material. A systems design approach motivated by a thermodynamic-based methodology was used to determine appropriate matrix alloying elements for improved mechanical properties while maintaining healing capabilities. Initial research focused on improving a Sn-Bi matrix reinforced with commercial nickel-titanium shape memory alloy wires. Healing was established at over 94% retained strength post-healing. Recent developments in aluminum-based alloys have shown potential healing across several systems. This study will detail methodology used to design the prospective Al-based alloy systems and establish the efficacy of the design approach through prediction of strength, microstructure development, and shape memory alloy incorporation into the composite. In addition, mechanical behavior of several binary and ternary elements was investigated to determine how healing is affected by different alloying elements. Advancement of healing behavior in an oxygen-containing environment through reactive element addition was also investigated. Finally, areas to advance future matrix and composite designs will be brought forth.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Fisher, Charles R. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2015
Resource Identifier:
962765811 ( OCLC )

Downloads

This item has the following downloads:


Full Text

PAGE 1

1 DESIGN OF LIQUID ASSISTED SELF HEALING METAL MATRIX COMPOSITES By CHARLES ROBERT FISHER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGR EE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Charles Robert Fisher

PAGE 3

3 To my family & friends who have supported me on my adventures

PAGE 4

4 ACKNOWLEDGMENTS First, I need to thank my advisor, Dr Manuel, for her continued support through my four years at the University of Florida. From my first visit during graduate recruitment weekend when I was first introduced to the idea of self healing metals I knew she was someone who wou ld push me to my f ull potential during my time in Gainesville. From foreign conferences to new research tools to a multitude of national fellowships, she has been there to give helpful advice and direction to advance my career. For that, I sincerely thank you. Next, my comm ittee members who ha ve been there to answer my questions, give comments on my research and encourage me to pursue more knowledge. I appreciate the time you have given to assist me as I tried to understand self healing metallic composites. To my previous m entors, supervisors, and colleagues at Ames Laboratory, US Steel, Caterpillar, and the Naval Surface Warfare Center Carderock, I appreciate the time given to help mold a young, nave researcher into someone who can write a dissertation. Next, I must ackn owledge the contributions of the Department of the Navy via the Science, Mathematics, and Research for Transformation (SMART) Scholarship for sponsoring my final two years at the University of Florida. Research funds were provided in part through the Natio nal Science Foundation (NSF) via grant CMMI 0824352 and the National Aeronautics and Space Administration (NASA) via grant NNX12AQ42G and NNX12AP71A

PAGE 5

5 To the members of the Materials Design and Prototyping Laboratory, both past and present, I must thank th em for the countless hours they have given me. From acting as sounding boards for new research ideas to hours of deliberation over conference presentations, it has all been greatly appreciated But special thanks to Glenn, for assistance with sample fabric ation using a hot pressing technique and EDS analysis and Patterson, who assisted on the study of bonding pressure requirements in Sn Bi alloys and incorporation of reactive elements for their contributions to advancing self healing research I look forw ard to catching up everyone at conferences in the future! Next, to the members of my Gainesville family who have been there for me throughout my time here in Florida: thank you! To the members at First Lutheran Church, thank you for welcoming me into your lives as I continued along in my pursuit of deeper faith. To the Gainesville Rugby Football Club, thanks for the bruises and the beers l hoist a few beers to celebrate being young and alive! Please feel free to take the idea to wherever life takes you next. To all of my friends from back home Thanks for the words of encouragement throughout my time in graduate school, and the numerous calls, cards, and emails to stay in touch. I look forward to more RAGBRAIs, weddings, Homecomings, and random world travel with all of you. And to all of my new friends gained since ISU, come to Iowa and you will see w hy I love it so much! Finally, I need to thank my family from the bottom of my heart. To all of my extended family, I love coming home and seeing you all. Do not worry I will always be home for Christmas! To my brother, Andrew, for always being there for me when I

PAGE 6

6 needed you most, thanks for being the older brother I always wanted. To my sister, Liz, for her constant pushing to make me a better person, I love you the mostest! And finally, to my parents, for being the rocks keeping me grounded. Your consta nt encouragement and never ending love has helped shape me into the man I am today. I love you both. This degree is for you.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWL EDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 10 LIST OF FIGURES ................................ ................................ ................................ ........ 12 LIST OF ABBREVIATIONS ................................ ................................ ........................... 19 ABSTRACT ................................ ................................ ................................ ................... 23 CHAPTERS 1 INTRODUCTION ................................ ................................ ................................ .... 25 Motivation ................................ ................................ ................................ ............... 25 Potential Applications ................................ ................................ .............................. 27 Document Outline ................................ ................................ ................................ ... 29 2 BACKGROUND ................................ ................................ ................................ ...... 31 Self Healing in Metallic Systems ................................ ................................ ............. 31 Solid State Healing in Metallic Systems ................................ ................................ .. 32 Dynamic Preci pitation in Aluminum Alloys ................................ ........................ 32 Dynamic Precipitation in Fe based Alloys ................................ ........................ 33 Reactive Element Coatings ................................ ................................ .............. 34 Liquid Assisted Healing in Metallic Systems ................................ ........................... 36 Eutectic Gallium Indium ................................ ................................ .................... 36 Metal Matrix Composites ................................ ................................ .................. 37 Liquid Assisted Healing Methodo logy ................................ ................................ ..... 40 Shape Memory Alloys ................................ ................................ ............................. 41 Shape Memory Alloys in Composites ................................ ................................ ..... 45 Summary ................................ ................................ ................................ ................ 45 3 PROTOTYPE SELF HEALING COMPOSITE ................................ ........................ 47 Design Methodology ................................ ................................ ............................... 47 Sn Bi Matrix Alloy Characterization ................................ ................................ ........ 49 NiTi Wire Prepara tion ................................ ................................ .............................. 51 NiTi Mechanical Testing ................................ ................................ .......................... 58 Pressure Requirements for Healing ................................ ................................ ........ 60 Summary ................................ ................................ ................................ ................ 63 4 SYSTEM DESIGN METHODOLOGY ................................ ................................ ..... 65

PAGE 8

8 System Design Chart ................................ ................................ .............................. 65 Structure Property Model Development ................................ ................................ .. 66 Alloy Selection ................................ ................................ ................................ ........ 68 5 BINARY ALLOY DESIGN ................................ ................................ ....................... 72 Aluminum Tin ................................ ................................ ................................ .......... 72 Al Sn Fabrication ................................ ................................ .............................. 72 Al Sn Mechanical Testing ................................ ................................ ................. 74 Aluminum Copper ................................ ................................ ................................ ... 75 Al Cu Fabrication ................................ ................................ .............................. 76 Al Cu Mechanical Testing ................................ ................................ ................. 78 Aluminum Magnesium ................................ ................................ ............................ 80 Al Mg Fabrication ................................ ................................ ............................. 81 Al Mg Mechanical Testing ................................ ................................ ................ 82 Aluminum Silicon ................................ ................................ ................................ .... 86 Al Si Fabrication ................................ ................................ ............................... 88 Al Si Mechanical Testing ................................ ................................ .................. 88 Increased Volume Fraction NiTi ................................ ................................ ....... 93 NiTi Wire Properties ................................ ................................ ......................... 96 Binary Alloy Summary ................................ ................................ ............................. 99 6 TERNARY ALLOY DESIGN ................................ ................................ ................. 101 Aluminum Copper Silicon Matrix Design ................................ .............................. 101 Al Cu Si Matrix Properties ................................ ................................ .................... 106 Al Cu Si Composite Fabrication ................................ ................................ ............ 107 Al Cu Si Composite Properties ................................ ................................ ............. 108 NiTi Wire Properties ................................ ................................ .............................. 111 Strength Model Verification ................................ ................................ ................... 113 Solid Solution Strengthening Model ................................ ............................... 113 Composite Strengthening Model ................................ ................................ .... 116 Internal Stress in SMA Wires ................................ ................................ ................ 117 Compression of Matrix Alloys ................................ ................................ ......... 119 Transformation Temperatur es in NiTi Wires ................................ ................... 121 Optimization of NiTi Wire Fraction in Al Based Composites ........................... 123 Novel Processing Techniques ................................ ................................ .............. 125 Pre strain of SMA Wires ................................ ................................ ................. 125 Hot Compression for Composite Fabrication ................................ .................. 126 Summary ................................ ................................ ................................ .............. 129 7 REACTIVE ELEMENT ADDITIONS ................................ ................................ ...... 131 Interfacial Toughness ................................ ................................ ............................ 131 Matrix Design ................................ ................................ ................................ 133 Sample Fabr ication ................................ ................................ ........................ 135 Mechanical Testing ................................ ................................ ........................ 137

PAGE 9

9 Analysis ................................ ................................ ................................ .......... 140 Thermodynamic Driving Force ................................ ................................ 140 Sb Zn Specimens ................................ ................................ .................... 140 Sb Cu Specimens ................................ ................................ .................... 141 Reactive Elements in Sn Bi Matrix ................................ ................................ ........ 143 Summary ................................ ................................ ................................ .............. 146 8 CONCLUSIONS ................................ ................................ ................................ ... 147 9 FUTURE WORK ................................ ................................ ................................ ... 150 Matrix Alloy Deve lopment ................................ ................................ ..................... 150 Shape Memory Alloy Development ................................ ................................ ....... 150 Reactive Element Incorporation ................................ ................................ ............ 151 Healing Cycle Optimizat ion ................................ ................................ ................... 151 APPENDIX A MECHANICAL AND THERMAL TESTING DATA ................................ ................. 153 Sn Bi ................................ ................................ ................................ ..................... 153 Al Based Binary Systems ................................ ................................ ..................... 159 Al Based Ternary System ................................ ................................ ..................... 164 NiTi Wires ................................ ................................ ................................ ............. 165 Mechanical Properties ................................ ................................ .................... 165 Thermal Properties ................................ ................................ ......................... 167 Thermal Transitions in Al Cu Si ................................ ................................ ...... 170 B ALUMINUM MAGNESIUM BASED MATRIX DESIGN ................................ ......... 172 Al Mg Li Matrix ................................ ................................ ................................ ...... 172 Matrix Design ................................ ................................ ................................ 172 Composite Fabrication ................................ ................................ ................... 175 Mechanical Testing ................................ ................................ ........................ 175 Al Mg Si Matrix ................................ ................................ ................................ ..... 177 Summary ................................ ................................ ................................ .............. 179 LIST OF REFERENCES ................................ ................................ ............................. 180 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 195

PAGE 10

10 LIST OF TABLES Table page 3 1 Tensile test results for Sn 21 wt% Bi alloys ................................ ...................... 50 3 2 Average mechanical testing results for Sn 21 wt% Bi matrices reinforced with <0.5 vol% SMA wire reinforcements ................................ ........................... 54 3 3 Healing characteristics of Sn Bi composites ................................ ....................... 54 3 4 The austenite and martensite transition temperatures for heat treated NiTi SMA wires ................................ ................................ ................................ .......... 59 3 5 Mechanical testing results for heat treated NiTi SMA wires ............................... 59 3 6 Mechanical properties of wires removed from heat treated Sn Bi composite ..... 60 5 1 Mechanical testing results for Al Sn matrices reinforced with <2.0 vol% SMA wire reinforcements ................................ ................................ ............................ 74 5 2 Mechanical testing results for Al Cu matrices reinforced with 2 2.5 vol% SMA wire reinforcements ................................ ................................ ............................ 78 5 3 Mechanical testing results for Al Mg matrices reinforced with < 5 vol% SMA wire reinforcements ................................ ................................ ............................ 84 5 4 Mechanical testing results for Al Si matrices reinforced with 2 3 vo l% SMA wire reinforcements ................................ ................................ ............................ 89 5 5 Healing characteristics of Al Si composites ................................ ........................ 90 5 6 Healing characteristics of Al Si composites using healing efficiency calculations for elastic modulus, yield strength, and strain to failure. ................. 91 5 7 Mechanical testing results for Al Si matrices reinforced with 3.5 4.5 vol% SMA wire re inforcements ................................ ................................ ................... 93 5 8 Healing characteristics of Al Si composites with 3.5 4.5 vol% NiTi wires ........... 9 4 5 9 Average Thermal Properties o f NiTi Wires Subjected to Various Heat Treatments ................................ ................................ ................................ ......... 96 5 10 Tensile Properties of NiTi Wires Subjected to Various Heat Treatments ............ 99 6 1 Mechanical testing results for Al Cu Si alloys ................................ ................... 107 6 2 Mechanical testing results for Al Cu Si matrices reinforced with 2 3 vol% SMA wire reinforcements ................................ ................................ ................. 109

PAGE 11

11 6 3 Thermal Properties of NiTi Wires Subjected to Various Heat Treatments ........ 112 6 4 Tensile Properties of NiTi Wires Subjected to Various Heat Treatments .......... 112 6 5 Compression Testing of Different Al Cu Si Alloys ................................ ............ 114 6 6 0.2% Compressive Yield Strength at Elevated Temperatures of Al 4.1C u 2Si and Al 3Si (at%) Alloys ................................ ................................ ..................... 120 6 7 Parameters Used to Model Elevated Compression Testing in Al 4.1Cu 2Si and Al 3Si (at%) Alloys ................................ ................................ ..................... 120 6 8 Elevated Temperature Tensile Testing of NiTi BB Wires Heat Treated at 530C for 24 hours ................................ ................................ ........................... 123 6 9 Parameters of DOE for Hot Pressing Fabrication of Al Cu Si Composites ....... 126 6 10 Parameters of DOE for Hot Pressing Fabrication of Al Cu Si Composites ....... 127 7 1 Summary of Fracture Toughness Testing of Sb Cu and Sb Zn Alloys ............. 137 7 2 Measured bond area of post healing Sb 4Cu or 4Zn CNSB specimens ......... 139 A 1 Tensile testing results of Al 3.0 a t% Si (3.1 wt% Si) matrix alloys after heat treatment at 592C for 24 hours ................................ ................................ ....... 161 A 2 Thermal transition temperatures of BB 35 NiTi wires heat treated for 24 hours at 592C ................................ ................................ ................................ 169 A 3 Thermal transition temperatures of BB 35 NiTi wires heat treated for two cycles at 592C for 24 hours ................................ ................................ ........... 170 B 1 Mechanical testing results fo r Al Mg Li matrices reinforced with 2 3 vol% SMA wire reinforcements ................................ ................................ ................. 176

PAGE 12

12 LIST OF FIGURES Figure page 1 1 Schematic of the Damage Prevention and Da mage Management material paradigms.. ................................ ................................ ................................ ......... 26 2 1 Schematic showing research areas for solid state and liquid assisted self healing in different metallic systems ................................ ................................ ... 32 2 2 Schematic of liquid assisted healing in a MMC reinforced with SMA imbedded wires. ................................ ................................ ................................ 42 2 3 Schematic of phases present in NiTi based shape memory alloys. .................... 43 2 4 Schematic showing how the transition temperatures (M s and A s ) are changed as a result of applied stresses. ................................ ................................ .......... 44 3 1 Calculated Sn Bi phase diagram using NIST Pb free solder database. ............. 48 3 2 Phase fractions of Sn 21 wt% Bi at various temperatures. .............................. 49 3 3 Representative microstructure of Sn Bi composite. ................................ ............ 50 3 4 Ellingham diagram showing the relative free energy calculations for Sn, Bi, Ni, and Ti. ................................ ................................ ................................ ........... 51 3 5 I mages of mold used in casting metal matrix composites with SMA wire reinforcements ................................ ................................ ................................ .... 53 3 6 Mechanical testing results of Sn f wires = 0.37%) in virg in and healed condition.. ................................ ................................ ........................ 55 3 7 Image showing crack healing in a Sn Bi composite. ................................ ........... 55 3 8 Schematic of changes to the NiTi SMA w ires during a composite healing cycle. ................................ ................................ ................................ .................. 57 3 9 Representative tensile results for Sn Sn f wires = 0.37%). ................................ ............................. 57 3 10 Image of Sn Bi alloy loaded into the diffusion couple jig post heat treatment. .... 61 3 11 Comparison of tensile results from Sn Bi pieces compressed at v arious stresses during healing to simulate the clamping force from SMA wires. ........... 62 3 12 Microstructure along interface of healed Sn Bi alloy compressed at 30 MPa. .... 62

PAGE 13

13 3 13 Scanning electron microscopy (SEM) image of the crack face for Sn Bi alloys healed at a s tress of 20 MPa, 30 MPa, and 40 MPa after tensile testing. .......... 63 4 1 System design chart for matrix development to be utilized within a self healing metal matrix composite ................................ ................................ .......... 66 4 2 Ellingham diagram showing relative positions of Al 2 O 3 NiO, and TiO 2 on a free energy scale during the formation of oxides. ................................ ............... 69 4 3 Graph showing the eutectic temperature of various Al based binary alloys.. ..... 70 4 4 Graph of change in yield strength for binary Al alloys as a function of alloy content. ................................ ................................ ................................ ............... 71 5 1 Phase diagram of the Al Sn system. At 250C, the composition yielding 20% liquid was calculated to be Al 19.5 at% Sn (51.6 wt% Sn). .............................. 73 5 2 Representative microstructure of an Al 19.5 at% Sn (51.9 wt% Sn) composite after heat treatment at 250C for 4 hours. ................................ ......... 73 5 3 Tensile data of Al 19.5 at% Sn (51.9 wt% Sn) composite heat treatment at 250C for 4 hours. ................................ ................................ .. 74 5 4 Image of Al 19.5 at% Sn (51.9 wt% Sn) composite after t ensile testing showing cracking in the matrix caused by excess ductility. ................................ 75 5 5 Phase diagram of the Al Cu system below 40 at% Cu. ................................ ...... 76 5 6 Phase fraction of Al 4.5 at% Cu at various temperatures. ............................... 77 5 7 Representative microstructure of an Al 4.5 at% Cu (10 wt% Cu) matrix reinforced with NiTi SMA wire after heat tre atment at 566C for 24 hours and tensile testing. ................................ ................................ ................................ ..... 78 5 8 Tensile data of Al 4.5 at% Cu (10 wt% Cu) composite after heat treatment at 566C for 4 hours. ................................ ................................ .. 79 5 9 Image of tensile crack induced in Al 4.5 at% Cu composite ............................ 80 5 10 Phase diagram of the Al Mg system. ................................ ................................ .. 81 5 11 Phase fraction of Al 16.6 at% Mg at various temperatures. ............................. 82 5 12 Representative microstructure of an Al 16.6 at% Mg (15.2 wt% Mg) matrix reinforced with a N iTi SMA wire after heat treatment at 487C for 24 hours. ..... 83 5 13 Tensile data of Al 16.6 at% Mg (15.2 wt% Mg) composite after heat treatment at 487C for 24 hours. ................................ ........................ 83

PAGE 14

14 5 14 Image of Al 16.6 at% Mg composite pre heal and post heal.. ......................... 85 5 15 Image of Pyrex tubes holding Al Mg composite after heat treatm ent. ................ 85 5 16 Representative microstructure of Al Mg composite after heat treatment at 487C for 24 hours under vacuum.. ................................ ................................ .... 86 5 17 Phase diagram of Al Si below 50% Si. At a composition of Al 3.0 at% Si (3.1 wt% Si), a healing temperature of 592C will yield 20% liquid.. ................... 87 5 18 Phase fraction of Al 3.0 at% Si at vario us temperatures. ................................ 87 5 19 Representative microstructure of an Al 3.0 at% Si (3.1 wt% Si) composite after heat treatment at 592C for 24 hours. ................................ ........................ 88 5 20 Representative microstructure in an Al 3.0 at% Si (3.1 wt% Si) composite after heat treatment at 592C for 24 hours showing failure occurring along eutectic regions. ................................ ................................ ................................ 89 5 21 Comparison of the a) pre heal to b) post heal tensile bar in an Al 3.0 at% Si composite reinforced with 2.0 vol% NiTi SMA wire ................................ ............. 91 5 22 Comparison of the virgin and healed tensile beh avior of an Al 3.0 at% Si composite reinforced with 2.0 vol% NiTi SMA wire. ................................ ............ 92 5 23 Comparison of thepost tensile 1, p ost heal and post tensile 2 in an Al 3.0 at% Si composite reinforced wi th 2.4 vol% NiTi SMA wire ................................ 92 5 24 Virgin tensile behavior of an Al 3.0 at% Si composite reinforced with either 2.26 vol% or 4.43 vol% NiTi SMA wire. ................................ .............................. 95 5 25 Optical image showing debonding between an Al 3 at% Si matrix and NiTi wire (V f = 4.43%) following tensile testing. ................................ ......................... 95 5 26 Phase diagram of the Ni Ti system. ................................ ................................ ... 97 5 27 TTT diagram showing aging behavior in a Ti 52Ni alloy. ................................ .... 98 5 28 Comparison of the tensile results for different specimens from the Al Sn, Al Cu, Al Mg, and Al Si composite systems with 1.5 2.5 vol% NiTi wires.. ........... 100 6 1 Liquidus projection of the Al Cu Si system. ................................ ...................... 102 6 2 The percent l iquid in an Al Cu Si alloy comparing DSC data to calculations using Pandat software ................................ ................................ ...................... 103 6 3 Isotherm at 512 C in the Al Cu Si system. ................................ ....................... 104

PAGE 15

15 6 4 Pseudo isopleth of Al Cu Si system along the line from the ternary eutectic point to the closest Al solid solution phase region. ................................ ........... 105 6 5 Phase fractions of Al 4.1 at% Cu 2.0 at% Si at various temperatures.. ...... 105 6 6 Representative microstructure of an Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) composite after a heat treatment at 530C for 24 hours ........ 106 6 7 Image of graphite tensile bar mold used for composite fabrication ................... 108 6 8 Representative tensile results of an Al Cu Si compo site with 2.9 volume percent NiTi SMA wire reinforcements compared to an Al Cu Si matrix alloy. 109 6 9 Comparison of tensile results of an Al Cu Si matrix reinforced with increasing volume pe rcent NiTi SMA wire reinforcements ................................ ................. 110 6 10 Comparison of tensile results of Al Cu Si, Al Cu, and Al Si matrices reinforced with 2.2 2.4 volume percent NiTi SMA wire reinforcements ............ 111 6 11 Solid solution strengthening in Al for Cu and Si. ................................ .............. 115 6 12 Elevated temperature compressive yield strength for Al 4.1Cu 2Si and Al 3Si (at%) with an overlaid modified Johnson Cook model for fit. ............................ 120 6 13 Change in A s and A f for NiTi wires heat treated at 530 C for 24 hours.. ........... 121 6 14 Schematic of the stress temperature relationships for NiTi BB SMA wires. ..... 122 6 15 Schematic of the stress temperature relationships for NiTi BB SMA wires heat treate d at 530C for 24 hours with the stress in SMA wires found in an Al 4.1Cu 2.0Si (at%) composite with V f = 55% of NiTi wires. ........................... 123 6 16 Schematic of the stress temperature relationships for NiTi BB SMA wires heat treated at 592C for 24 hours with the stress in SMA wires found in an Al 3.0Si (at%) composite with V f = 25% of NiTi wires. ................................ ...... 124 6 17 EDS analysis of a representative micros tructure of a hot pressed Al Cu Si composites ................................ ................................ ................................ ....... 128 6 18 Microstructure of Al Cu Si composites pressed at 4 MPa for 4 hours and pressed at 4 MPa for 8 hours. ................................ ................................ .......... 128 6 19 Comparison of tensile behavior of a cast Al Cu Si composite (V f = 2.32%), a hot pressed Al Cu Si composite with 1 ply (V f = 1.37%), and a hot pressed Al Cu Si composite with 3 plies (V f = 1.67%) ................................ ................... 129 7 1 Schematic of a chevron notch short bar (CNSB) specimen. ........................... 133

PAGE 16

16 7 2 Ellingham diagram showing relative position of free energy for oxide formation for Cu, Sb, and Zn. ................................ ................................ ........... 134 7 3 Plot of percent liquid and percent solute in liquid for Sb Cu and Sb Zn alloys. 135 7 4 Phase diagrams of Sb Cu andSb Zn below 40 at% solute showing the healing tempe rature for each alloy ................................ ................................ ... 136 7 5 Steel holder used to clamp Sb alloys during a healing heat treatment. ............ 137 7 6 Representative microg raphs of Sb 4Cu and Sb 4Zn (at%) at the interface. ..... 138 7 7 Image along Sb 4 at% Zn healed interface showing presence of ZnO. ......... 139 7 8 Fracture surfaces of a) Sb 4Cu and b) Sb 4Zn (at%) healed samples. ............ 139 7 9 Schematic showing chevron notch crack front in relation to important mathematical factors ................................ ................................ ....................... 143 7 10 Representative micrographs of Sn Bi 0.2Li and Sn Bi 0.2Ce (at%) at the interface.. ................................ ................................ ................................ .......... 145 A 1 Mechanical testing results of Sn 21wt%Bi matrix alloys after heat treatment for 24 hours at 169C. ................................ ................................ ...................... 153 A 2 Mechanical testing results of Sn 21wt%Bi composites (V f = 0.3 0.4%) after heat treatment for 24 hours at 169C. ................................ .............................. 154 A 3 Mechanical testing comparison of Sn f = 0.31%) in the virgin and healed composite condition ................................ ....................... 154 A 4 Mechanical test ing comparison of Sn f = 0.40%) in the virgin and healed composite condition ................................ ....................... 155 A 5 Tensile testing results of as received NiTi BH 0075 wires ............................... 156 A 6 Tensile testing results of NiTi BH 0075 wires heat treated for 3 hours at 500C. ................................ ................................ ................................ .............. 156 A 7 Tensile testing results of NiTi BH 0075 wires heat tre ated for 3 hours at 500C and then again at 169C for 24 hours. ................................ ................... 157 A 8 Tensile testing results of NiTi BH 0075 wires removed from a Sn Bi composite sample which had been heat treated for 3 ho urs at 500C and 169C for 24 hours (2x). ................................ ................................ ................... 157 A 9 Thermal transition temperatures for a BH 0075 NiTi wire heat treated for 3 hours at 500C ................................ ................................ ................................ 158

PAGE 17

17 A 10 Thermal transition temperatures for BB 00755 NiTi wire heat treated for 3 hours at 500C and then 24 hours at 169C ................................ .................... 159 A 11 Mechanical testing results of Al 19.5 at% Sn (51.6 wt% Sn) composites (V f = 1.5 1.8%) after heat treatment for 4 hours at 250C. ................................ .... 159 A 12 Mechanical testing results of Al 4.5 at% Cu (10 wt% Cu) composites (V f = 2.2 2.4%) after heat treatment at 566C for 24 hours. ................................ ...... 160 A 13 Mechanical testing results of Al 16.6 at% Mg (15.2 wt% Mg) composites (V f = 1.9 4.0%) after heat treatment at 460C for 24 hours. ................................ .. 160 A 14 Mechanical testing results of Al 3.0 at% Si (3.1 wt% Si) matrix alloys after heat treatment at 592C for 24 hours. ................................ .............................. 161 A 15 Mechanical testing results of Al 3.0 at% Si (3.1 wt% Si) composites (V f = 2.0 2.4%) after heat treatment at 592C for 24 hours. ................................ ...... 162 A 16 Mechanical testing results of Al 3.0 at% Si (3.1 wt% Si) compo sites (V f = 3.7 4.4%) after heat treatment at 592C for 24 hours. ................................ ...... 162 A 17 Mechanical testing comparison of Al f = 2.0%) in the virgin and healed composite conditi on ................................ ....................... 163 A 18 Mechanical testing comparison of Al f = 2.4%) in the virgin and healed composite condition ................................ ................... 163 A 19 Mechanical testing results of Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) matrix alloys after heat treatment at 530C for 24 hours. ..................... 164 A 20 Mechanical testing result s of Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) composites (V f = 2.3 2.9%) after heat treatment at 530C for 24 hours. ................................ ................................ ................................ ............... 164 A 21 Tensile testing results of as received NiTi BB 35 wi res ................................ ... 165 A 22 Representative tensile results of NiTi BB 35 wires heat treated for 24 hours at 592C for either one cycle or two cycles. ................................ ...................... 166 A 23 Tensile testing results of NiTi BB 35 wires heat treated for 24 hours at 530C for one, two, or three cycles to simulate the entire healing process for Al Cu Si based composites. ................................ ................................ ..................... 166 A 24 R epresentative thermal transition temperatures for as received BB 35 NiTi wire. ................................ ................................ ................................ .................. 167 A 25 Thermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 530C ................................ ................................ ................................ .......... 168

PAGE 18

18 A 26 Thermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 530C two times ................................ ................................ .......................... 168 A 27 Representativ e thermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 592C ................................ ................................ .......... 169 A 28 Representative thermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 592C two times ................................ .......................... 170 A 29 Thermal transition temperatures for Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) matrix alloy after heat treatment at 530C for 24 hours. ........ 171 B 1 Liquidus projection of the Al Mg Li system.. ................................ ..................... 173 B 2 Isotherm at 445 C in the Al Mg Li ternary alloy system.. ................................ .. 174 B 3 Pseudo isopleth of Al Mg Li system along the line from the ternary eutectic point to the closest Al solid solution phase region. ................................ ........... 174 B 4 Represen tative microstructure of an Al 18.7 at% Mg 1.1 at% Li (17.3 wt% Mg 0.3 wt% Li) matrix exhibiting a non continuous eutectic phase surrounding Al solid solution phase. ................................ ................................ 175 B 5 Representati ve tensile results of an Al Mg Li matrix reinforced with 3.04 volume percent NiTi SMA wire reinforcements. ................................ ................ 176 B 6 Fracture surface of Al Mg Li composite showing oxidation on the crack interfa ce after a healing cycle.. ................................ ................................ ......... 177 B 7 Pseudo isopleth of Al Mg Si system along the line from the ternary eutectic point to the closest Al solid solution phase region.. ................................ .......... 178

PAGE 19

19 LIST OF ABBREVIATIONS One standard deviation C Degrees Celsius f Strain to failure a Applied s tress y Yield s t ress ACP Active corrosion protection A f Austenite finish temperature Ag Silver Al Aluminum A s Austenite start temperature at% Atomic percent Au Gold Avg. Average B Boron Bi Bismuth BN Boron Nitride CCC Chromate conversion coating Ce Cerium cm Centimeter Co Cobalt Cr Chromium CSIRO Commonwealth Scientific and Industri al Research Organization Cu Copper DOE Design of experiments

PAGE 20

20 DSC Differential scanning calorimetry E Elastic modulus EDS Energy dispersive spectroscopy EGaIn Eutectic gallium indium FCC Face centered cubic FEM Finite element modeling Ga Gallium Ge Germanium GPa Gigapascal Hf Hafnium ICP Inductively coupled plasma In Indium KSC Kennedy Space Center La L anthanum Li Lithium MDPL Materials Design and Prototyping Laboratory M f Martensite finish temperature Mg Magnesium MHz Megahertz mm Millimeter MMC Metal matrix composite Mn Manganese Mo Molybdenum MPa Megapascal M s Martensite start temperature

PAGE 21

21 N Nitrogen N ASA National Aeronautics and Space Administration Ni Nickel NiTi Nickel titanium shape memory alloy NSF National Science Foundation RE Rare earth S Siemens S Sulfur Sb Antimony SEM Scanning electron microscopy Si Silicon SMA Shape m emory a lloy Sn Tin Ta Ta ntalum Ti Titanium T room Room Temperature TTT Time Temperature Transformation UA Under aged UF University of Florida U TS Ultimate tensile strength UWM University of Wisconsin Milwaukee V Vanadium vol% Volume percent wt% Weight percent Y Yttrium

PAGE 22

22 Zn Zinc Z r Zirconium

PAGE 23

23 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DESIGN OF LIQUID ASSISTED SELF HEALING METAL MATRIX COMP OSITES By Charles Robert Fisher August 2013 Chair: Manuel, Michele Viola Major: Materials Science and Engineering Advancements in materials science have enabled components to be fabricated lighter, stronger, and more functional than ever. However, damage mi tigation in these materials still limits their usable lifetimes. Recently, materials with the ability to heal structural damage have shown promise as potential novel materials of the future. Within metallic based systems, metal matrix composites reinforced with shape memory alloys (SMA) have the potential to demonstrate dramatic capabilities in damage mitigation and repair. Investigations of liquid assisted self healing in metal matrix composites have centered on developing a high specific strength matrix possessing a low melting eutectic for use as the healing material. A systems design approach motivated by a thermodynamic based methodology was used to determine appropriate matrix alloying elements for improved mechanical properties while maintaining heal ing capabilities. Initial research focused on improving a Sn Bi matrix reinforced with commercial NiTi SMA wires. Healing was established at over 94% retained strength post healing. Recent

PAGE 24

24 developments in aluminum based alloys have shown potential healing across several systems. This study will detail methodology used to design the prospective Al based alloy systems and establish the efficacy of the design approach through prediction of strength, microstructure development, and SMA incorporation into the c omposite. In addition, mechanical behavior of several binary and ternary matrix alloy composite s w ere investigated to determine how healing is affected by different alloying elements in Al Advancement of healing behavior in an oxygen containing environmen t through reactive element addition was also investigated. Finally, areas to advance future matrix and composite designs will be brought forth.

PAGE 25

25 CHAPTER 1 INTRODUCTION Motivation Biological materials such as tissue, bone, and muscle are readily able to heal themselves through complex processes which involve microvascular networks, self assembling systems, and nanostructures ( 1 ) Recent literature has delved into the development of engineered biomimetic or bio inspired materials for use as structural component s ( 2 3 ) One potential aspect of biomimetic materials is the ability to repair damage forme ( 3 ) an active area of research at universities and government laboratories across the globe ( 4 9 ) Self healing materials represent a paradigm shift in traditional materials engineering. History is filled with high performance alloys designed to resist in service changes. With the ad vancement of self healing materials, however, there is the ability to shift from the Damage Prevention paradigm to one of Damage Management ( 10 ) Within the Da mage Prevention paradigm, stronger and/or more ductile materials are created in order to better sustain damage. In Figure 1 1 a, M aterial A represents a typical ductile material exposed to constant loads. While under relatively low stresses, i.e. below the yield stress, the material remains intact. However, with continued exposure to increased loads above the yield stress or high cycles to induce fatigue, damage begins to form. This damage will progress until failure of the component. Material B represents a stronger but brittle material, whereas M aterial C represents a stronger, more ductile material; both M aterial B & C were manufactured to resist more

PAGE 26

26 deformation during service. However, all of these materials (A C) trend in one direction with respect to d amage formation: either zero damage or positive damage : (Eq. 1 1) where t represents time. Figure 1 1 : Schematic of the a) Damage Prevention and b) Damage Management material paradigms For traditional materials in the Damage Prevention paradigm in a), Material A represents a low strength, ductile alloy, B is a higher strength, brittle material, and C a high strength, ductile alloy. For novel self healing materials in b), Material D rep resents a material able to heal itself once, whereas E is a material able to repeatedly heal itself. Adapted from van der Zwaag, Sybrand. 2007. Self Healing Materials: An Alternative Approach to 20 Centuries of Materials Science (page 4 and 6, Figures 1 an d 2). Springer, New York Self healing materials, on the other hand, represent the Damage Management paradigm where the damage is controlled in such a way that it can be reversed. Material D in Figure 1 1b represents such a self healing material. Here, the damage realized by the material during service is able to be reversed. Thus, self healing materials are instead represented by Equation 1 2 during a healing cycle: (Eq. 1 2 )

PAGE 27

27 where t i represents the initial time and t is th e change in time required for healing. An ideal self healing material is represented by something more akin to Material E; repeated healing over time to ensure damage in the material never reaches the failure point. This reversal of damage is unlike what i s found in traditional engineering materials that lie in the Damage Prevention paradigm, which only utilize higher strength or more ductile material s to prevent material failures. These new self healing materials all possess similar requirements, including a healing trigger, mobile atoms for the healing mechanism, and crack surfaces being in close contact ( 7 10 ) These requirements revea l that prospective self healing materials are more than a material, but ar e instead a self healing system similar to biological systems. The White and S ot tos groups at the University of Illinois were the first to exhibit appreciable self healing within a p olymer matrix system in 2001 ( 11 ) In the decade following, several more groups also began research across differe nt materials classes including polymers, ceramics, and metallic systems. Several review articles have been written covering all the different aspects for healing in polymer systems ( 8 12 13 ) and healing in ceramic/cementitious systems ( 14 16 ) in addition to textbooks on these topics ( 6 17 18 ) A review of self healing in metallic systems has not been completed at this time Potential Applications Applications for self healing materials can be found in the aer onautics and space industry ( 4 ) Due to the demanding requirements for materials in aerospace applications, materials are designed with extra safety factors in order to hand le the extreme environment. This leads to over design of the component. However, if

PAGE 28

28 designers are able to shift to a Damage Management paradigm, this would allow the reduction of volume and mass from components as the materials would enable healing of the structure without over designing the part through high safety factors or having to replace the component. One specific example would be meteorite damage on satellites. During service, space debris and meteors continually cause damage to spacecraft to such an extent that mission operations can be significantly degraded ( 19 ) Due to the extreme costs associated with space travel, replace ment of these damaged components is difficult. Therefore, self healing materials would be prime candidates to ensure continuation of the mission. Another example of applications of self healing materials would be for space exploration using rovers on plane t surfaces. In 2006, one of the wheels on the Mars rover Spirit stopped functioning properly ( 20 ) And in 2012, the landing of Curiosity on Mars was highly publicized because of the new landing system utilized ( 21 ) If the landing did not go smoothly, it was probable the rover would have crash landed, breaking required components, and been unusable. These cases s how the potential for self healing materials to be exploited for components in systems which require significant time and money in order to replace. Deep sea applications are another potential opportunity for self healing metals to be utilized. The depth and extreme pressures ensure that component repair is extremely difficult. However, an increase in pressure has been found to enhance healing ( 22 ) and thus it is a potential area to investigate. Brinker Technologies has started investigations using so

PAGE 29

29 ( 23 ) This technology is analogous to the human body sending platelets and clotting tissue to a wound using the vast vascular network known as our circulatory system. As described by van der Zwaag ( 10 ) healing. It is composite systems which will drive the viable application of self healing metals; graded microstructures fo r increased toughness and wear resistance, dispersed compounds for increased strength, coatings for high temperature capabilities, vascular networks consisting of liquid delivery systems via hollow tubes (or the even the designed microstructure itself) wit h shape memory reinforcements for crack size reduction. These multifunctional systems will help us realize a world of the future where a fender bender on your car can be fixed by using a heat gun to heat up the affected area, causing the metal to resume is previous form. Document Outline This document will seek to understand the underlying mechanisms affecting healing in a metal matrix composite (MMC) system pairing a solid solution strengthened matrix exhibiti ng non continuous eutectic phase with nickel titanium (NiTi) shape memory alloy (SMA) reinforcements. First, a review of the current topics in self healing metals will be discussed. Next, an investigation is presented on a prototype tin bismuth (Sn Bi) com posite system to elucidate how healing is affected by the connection between matrix and the wires in addition to how the pressure applied at the crack during a healing cycle changes post healing mechanical properties. To advance MMC technology, a methodolo gy is detailed to expand potential matrix alloys into higher strength materials specifically aluminum (Al) based alloys. Experiments involving potential binary alloys to serve as matrix materials lead into the design of a high

PAGE 30

30 strength ternary alloy. Th e next section investigates ways to improve healing in air via introduction of reactive elements to induce chemical reactions at the interface. Finally, a summary of the work is completed before future areas of development within liquid assisted self heali ng composites are detailed.

PAGE 31

31 CHAPTER 2 BACKGROUND Self Healing in Metallic Systems In 1967, an article appeared in Time the recently developed TRIP steels, which use a martensitic phase transformation to blunt small cracks ( 24 ) Since 199 3 the Ols o n group at Northwestern has been working to incorporate self healing characteristics into their iron (Fe) based metal matrix composites for structural applications ( 25 26 ) However, self healing in metallic systems has been relatively uninvestigated compared to polymeric and ceramic systems because of several disadvantages compared to other material classes: 1) stronger metallic bonds, 2) small atomic volume, and 3) relatively slow kinetics due to low diffusion rates of metals ( 3 ) Stronger metallic bonds mean the healing material requires higher activation energy in order to initialize the healing process. Unlike polymers w hich possess large atomic volumes, metal alloys are much smaller and thus it is more difficult to accumulate enough material to fill a crack. Also, low diffusion rates means healing reactions are slower. Despite these drawbacks, various examples of healing in metals have been documented in literature. Self healing in metallic systems can be grouped into two general types: solid state and liquid assisted healing ( 27 ) Solid state healing is associated with dynamic pre cipitation for creep and fatigue crack suppression and coatings to reduce corrosion. Liquid assisted healing has been demonstrated in metal matrix composites for structural use and in electrical applications. Each of these types will be broken down into th e one of two classes autonomous and non autonomous ( Figure 2 1). Autonomous healing is

PAGE 32

32 only present if there are no outside influences, e.g. heat or light, to trigger the healing process ( 7 ) Figure 2 1 : Schemat ic showing research areas for solid state and liquid assisted self healing in different metallic systems Solid State Healing in Metallic Systems Solid state healing can be found in both structural metals and coatings. For structural materials, research has been focused upon autonomous self healing through precipitation at high energy surfaces such as grain boundaries, dislocations and voids ( 28 ) Much of the research is focused on healing creep and fatigue damage in aluminum (Al) alloys and steel. This precipitation process reduces the maximum flaw size and thus solid state healing of microscale defects is feasible for structural metals. Dynamic Precipitation in Alumin um Alloys Al alloys have been utilized as structural materials following age hardening heat treatments since precipitation reactions were first detailed in 1911 for an Al copper (Cu), and magnesium ( Mg ) alloy ( 29 ) The basis for solid state self healing has been the result of similar research by collaborators at the Commonwealth Scientific and Industrial Research Organization (CSIRO) in Victoria, Australia investigating dynamic precipitation

PAGE 33

33 in under aged (UA) Al Cu Mg based alloys ( 9 28 30 35 ) Earlier work was performed by Ringer et al ( 36 ) Wang et al ( 37 ) and Skrotzki et al ( 38 ) In th e UA condition, supersaturated solutes are available to undergo dynamic precipitation throughout the matrix during creep or fatigue loading. A c ommercial alloy with manganese (Mn) additions, Al 2024 (Al 4.4Cu 1.5Mg 0.6 Mn ) in the UA condition during a creep test also displayed improved properties, including doubling the time to failure, delay of tertiary creep, and increase in strain at failure ( 30 ) The healing process was found to be a thermally activated process which corresponds to the Cu and Mg pipe diffusion in an Al matrix ( 39 41 ) Results from other studies have suggested the beneficial effects of under aging may be applied to fatigue failures as well since the pipe diffusion rate of Cu in Al is still 10 6 times faster than vacancy diffusion even at room tempera ture ( 32 ) While dynamic precipitation has the potential to delay fatigue crack initiation and the onset of crack propagation in Al alloys, Wanhill ( 42 ) detailed several limitations. First, healing is limited by a maximum allowable flaw size for repair Second, the amount of solute able to be precipitated results in another limitation o nc e the alloy reaches peak aging, dynamic precipitation is nullified. Finally, the major hurdle for in service use of self healing Al alloys is due to most failures in being initiated by surface defects. Because of oxidation issues and other surface effects, the current self healing technology for Al alloys would be inadequate to heal surface flaws. Despite th ese concerns current research is focused on the significance of the volume changes occurring during healing ( 2 8 ) Dynamic Precipitation in Fe based Alloys Shinya and colleagues ( 43 53 ) ha ve lead research successfully improving creep properties through autonomous self hea ling in heat resisting austenitic stainless steel s

PAGE 34

34 Creep fracture is known to be the result of nucleation, growth, and coalescence of creep induced cavities along grain boundaries M ost high strength heat resisting steels show this type of failure mechani sm ( 49 ) Creep cavities grow through diffusion along surfaces and grain boundaries Sulfur (S) contamination especially increases this rate due to its low melting point, thus increasing the surface diffusion rate ( 54 ) However, removal of S through additions of cerium (Ce), a reactive rare earth elemen t, or titanium (Ti) allows boron (B), Cu, and nitrogen (N) to preferentially segregate to creep cavity surfaces. This has shown that segregation of B or Cu and the precipitation of BN to the creep surfaces greatly decreases creep growth, thereby improving overall mechanical properties in modified type 304 and 347 stainless steel ( modifications include additions of Ce, Ti and/or B). This autonomous healing is accelerated at the elevated testing temperatures first filling the creep cavity and then blunting c reep cavity progression. Therefore, by decreasing the largest defect size within the material, the material life span is increased. In addition to the issues raised regarding self healing in Al alloys, dynamic precipitation for creep prevention in steels s uffers from another issue; the self healing ability of the modified steels does not prevent creep cavities from forming it only hinders their growth. With time, more cavities will be created and cause ultimate failure despite the healing capability of the steel alloys. Reactive Element Coatings Self healing for corrosion protection has been used extensively since the addition of large amounts of chromium (Cr) (>11%) was added to steel to form a protective oxide ( 55 ) Recent research has centered on passiv ation of metal surfaces using chromate

PAGE 35

35 based coatings for corrosion protection, especially in alloys; articles detailing potential re no later than the early 1950 s ( 56 ) In the decades following, advancement of chromate conversion coatings (CCC) continued into new chemistries, processes, and applicable metals available for coating; a thorough review was completed by Kendig and Buchheit ( 57 ) Unfortunately, the mobility of Cr 6+ in biological systems and reactivity with oxidation mediators, being linked with DNA damage and lung cancer, and has lead t o the regulation and elimination of CCCs and other Cr based chemicals across numerous industrial processes ( 58 ) In terms of non chromate corrosion prevention in metals, research can be grouped into three categories: 1) hypervalent transition metals [ e.g. molybdenum (Mo), vanadium (V) Mn] 2) transition metal [ e.g. zirconium (Zr) hafnium (Hf ) ] or covalent oxides [ e.g. silicon (Si) germanium (Ge ) ] and 3) rare earth metal coatings ( 57 ) Of these three, only the rare earth (RE) metal containing coatings have shown promise as healing ability, also known as active corrosion protection (ACP) ( 59 ) A review of the rare earth lanthanide series, especially Ce salts, was completed by Bethencourt et al ( 60 ) The latest research into reactive element additions (e.g. yttrium (Y) lanthanum (La) Zr Ce) has shown the most positive effects, including slowing scale growth and increasing oxide adhesion ( 61 ) The most widely accepted explanations for these reactive element effects are a high oxygen dissociation rate ( 62 ) and scavenging of coating impurities (namely S) after deposition ( 63 64 )

PAGE 36

36 Limitations of self healing metallic coatings are the continual reparation of the o xide surface protecting the underlying alloy. The loss of matter will slowly reduce the reservoir of the oxide former thus further use of the coating will continue to reduce the potential for healing within the coating. Liquid Assisted Healing in Metallic Systems Liquid assisted self healing has on going research in both electrical and structural composite applications. For electronic applications research into autonomous self healing has been realized through a low melting eutectic alloy of gallium indiu m For metal matrix composites, the focus has been along two fronts: 1) addition of thin tubes filled with low melting point alloys which break in service, allowing for material to flow out and fill a crack during a healing cycle and 2) shape memory alloy (SMA) reinforcements to enable crack closure during an elevated temperature healing cycle which allows for partial liquefaction of the matrix to enable healing. Eutectic Gallium Indium Electrical applications utilizing liquid assisted healing methods focus on eutectic gallium indium alloy (refer red to as EGaIn). This alloy, Ga 25wt%In possesses a high electrical conductivity, 3.4 x 10 4 S/cm which flows at a critical surface stress ( 65 ) Being liquid at room temperature (melting point = 16C), the alloy readily forms a thin, passivating oxide surface which enables its us e in forming metastable, non spherical structures. Early work using this alloy for electrical applications was conducted by researchers at Harvard University ( 66 67 ) Researchers at North Carolina State University have developed a self healing antenna based on EGaIn ( 68 ) The antenna is created by injecting EGaIn into micro channels within poly(dimethylsiloxane) (PDMS) which is used for structural stability.

PAGE 37

37 This composite structure enables the antenna to be mechanically tu ned through elongation, while also maintaining flexibility and a radiation range efficiency of 90% from 1910 1990 MHz. The team found the antenna possessed repeatable, autonomous self healing When cut with a razor blade, the electrical resistivity was ret urned to pre cut values upon the removal of the blade no matter how many times the antenna was cut. Electrical devices developed by collaborators at the University of Illinois (UI) have also used EGaIn to heal defects ( 65 ) They demonstrate d autonomous healing in less than one millisecond with over 99% recovery of condu ctivity using microencapsulated EGaIn layered above patterned gold ( Au ) lines on a glass substrate. During a crack event, decreasing conductivity of the device to near zero, the EGa In filled capsules are ruptured, allowing the liquid metal to flow into the crack and restore conductivity to pre crack levels T he physical characteristics post healing, specifically strain and fracture toughness values associated with in service device applic ation, have yet to be examined. Metal Matrix Composites Research into self healing of metal matrix composites first began at Northwestern University under the direction of Olson. This work was motivated by a hierarchical systems design methodology in an effort to elucidate processing structure property performance relationsh ips ( 69 70 ) Early research consisted of a ferrous m atrix aided by continuous SMA reinforcement. The composite, investigated by undergraduate design teams ( 26 71 72 ) consisted of a ferritic superalloy matrix reinforced with austenitic SMA wires. The austenitic SMA wires consisted precipitates in a face centered cubic (FCC) phase.

PAGE 38

38 Further research investigated a different Fe based self healing tank for turbine engines. A composition of Fe 12.55 wt% Ni 25 wt% Co 12.5 wt% Cr 2 wt% Al 0.9 wt% Ti was investigated ( 72 ) the running temperature of the engine results in severe cracking during service due to thermal fluctuations. A self healing composite tank could heal these c racks as the temperature increases because of the clamping force from SMA fibers embedded in the matrix. The doctoral student overseeing these undergraduate projects, Files, completed his thesis work on an Fe based B2 NiAl intermetallic superalloy matrix r einforced with austentic SMA wires ( 25 73 ) His thermodynamically designed Fe 27.6Ni 18.2Co 4.1Ti 1.6Al SMA wires were heat treated to obtain M s and A s temperatures near room temperature. Up to 5% strain, h e was able to demonstrate full recovery of the original length after unloadi ng and heating the SMA wire. His research also showed stable crack growth and crack bridging from the SMA wires in the Fe based composite. However, during all the Fe based composite research, complete healing was not realized; poor wettability, intergranul ar fracture in the SMA wire, and oxidation of the crack interface were blamed. Despite setbacks in ferrous based alloys, Files ( 25 ) developed a prototype Sn Bi matrix alloy reinforced with NiTi SMA composites which demonstrated over 80% healing. Further research into a Sn Bi In matrix was investigated by another undergraduate team in 1997 and crack closure and healing were realized ( 74 ) Because of wettability problems between the matrix and NiTi wires, Bernikowicz ( 75 ) focused on knots in the SMA wires to added mechanical adhesion. She also added flux to the crack

PAGE 39

39 interfaces to prevent oxidation during a healing cycle the composites display 88% recovery of strength. Other student projects investigated incorp oration of flux to the matrix to enable in situ fluxing during a healing cycle ( 76 77 ) Using both Sn Bi and Sn Bi In based composites, Knapp ( 78 ) optimized the liquid fraction in the matrix during healing. He concluded between 15 20% liquid in the matrix during healing balanced structural stability requirements while maintaining enough liquid phase to enable complete healing. Other undergraduate design teams investigate d addition of electronic systems to the Sn based composite systems ( 71 79 ) Manuel ( 80 ) was able to improve healing to over 94% in the Sn Bi prototype system by coating the NiTi SMA wires with gold to improve adhesion with the matrix. Finite element modeling (FEM) of this self healing composite system was completed by Burton et al ( 81 ) They demonstrated crack bridging and crack closure from the SMA wires through a one dimensional constitutive model of the NiTi wires. The model also confirmed the requirement to pre strain the SMA wires prior to heal ing in order to create a critical clamping force to enable healing. An investigation by another undergraduate team to design Mg and Al based matrices reinforced with high strength NiTi based SMA wires for use in self healing composites was also conducted ( 82 ) but samples were not fabricated. However, in Manuel ( 80 ) she investigated a Mg 5.7 at% Zn 2.7 at% Al matrix reinforced with both commercially available NiTi SMA wires and prototype, nanodispersion strengthened SMA wires developed by Jung ( 83 ) in his doctoral thesis. The prototype Ni 32 at% Ti 3 at% Al 15 at% Zr wires had similar processing temperatures to those found in the Mg based composites.

PAGE 40

40 a 160% increase in uniform duc tility in the composite and a 40% increase in strength over cast AZ91 commercial alloys ( 80 84 85 ) Healing behavior was limited due to the oxidation behavior of the Mg alloy in the liquid state even when samples were vacuum encapsulated during healing heat treatments. However, the strength and thermomechanical models developed will assist wit h design considerations for future composite development. A group of researchers at the University of Wisconsin Milwaukee (UWM) have also investigated healing in metal castings Initial investigations explored eutectic Sn Bi infiltrated alumina balloons ( 86 87 ) and an off eutectic healing model is based on non homogenous dendrites ( 88 ) with only minor success However, t he bulk of the UWM work has focused on filling hollow ceramic tubes with the l ow melting alloy and embedded these within an Al matrix to form a self healing composite ( 86 87 89 92 ) A crack is introduced in to the composite, breaking the ceramic fibers, before the composite is placed into a furnace and heated above the solder melting temperature. Upon cooling, the comp osite microstructure is investigated. Recently, a composite consisting of 206 Al alloy as a matrix with hollow quartz tubes filled with an 802 Al brazing alloy has shown self healing potential ( 87 ) but bonding behavior was limited between the low melting alloy and the stronger matrix alloys. Liquid Assisted Healing Methodology Liquid assisted healing using SMA reinforcements for crack closure has several advantages over other metals ba sed self healing techniques for structural components including faster healing kinetics, the ability to repair large scale defects, and the possibility of repeatable self healing. The research is based on the concept of crack

PAGE 41

41 healing through partial liqu ef ac tion of the matrix material in the presence of a clamping force applied by SMA wires embedded in the matrix. As shown in Figure 2 2, the composite, consisting of a matrix reinforced with longitudinally aligned SMA fibers, has a crack initiated. The matri x fails but the SMA wires stay intact due to their high strength relative to the matrix. Once the crack spans the entire width, the load is transferred completely to the SMA wires. This induces a local stress induced transformation at the crack site. Upon removal of the load, two different effects can occur within the SMA wires based on their microstructural state. If in the austenite phase, the wires will be pseudoelastic and immediately revert back to their original (shorter) shape, resulting in crack clo sure ( 93 ) If in the martensite phase, the wires will remain elongated until the composite temperature is raised above their austenite finish (A f ) temperature. Concurrently, the increase in temperature creates a partially liquefied matrix, enabling crack filling and matrix softening. This matrix softening enables the removal of plasticity in the composite acquired during fracture as the SMA wire s revert to their original length. Subsequently, cooling results in a solidified composite able to realize its pre cracked strength. A detailed description of the shape memory effect is provided in the following section. Shape Memory Alloys Shape memory al loys (SMA) benefit liquid assisted self healing composites because of their ability to apply a clamping force at a crack interface with elevated temperatures. SMAs are characterized by their ability to recover apparent permanent strains through thermoelast ic martensitic transformations ( 93 ) The material transforms its crystal lattice through a shearing deformation as a result of cooperative ato mic movement in response to applied loads.

PAGE 42

42 Figure 2 2: Schematic of liquid assisted healing in a MMC reinforced with SMA imbedded wires Starting at the top at T room a crack initiates in the matrix. Further propagation causes matrix failure and a stress induced transformation in the SMA wires. Increasing T > A s will initiate crack closure from the wires. Further increases will cause a clamping force to push the crack faces together while partial liquefaction of the matrix via low melting eutectic heals t he crack. Cooling enables solidification of the matrix and full functionality. SMAs possess two basic phases the cubic, high temperature austenite phase and the monoclinic, low temperature martensite phase ( 94 ) There are four characteristic temperatures in all SMAs: 1) martensi te start (M s ), the temperature at which the material initiates a martensitic phase transformation, 2) martensite finish (M f ), the temperature at which the martensite transformation is complete, 3) austenite start (A s ), the temperature

PAGE 43

43 at which the material begins transforming from martensite to austenite, and 4) austenite finish (A f ), the temperature at which the austenite transformation is completed. Depending on the starting phase of the alloy, austenite or martensite, the material will exhibit the pseudo elastic or shape memory effect, respectively. A schematic of phase changes in SMAs is shown in Figure 2 3 Figure 2 3: Schematic of phases present in NiTi based shape memory alloys. The red circles represent Ti and the blue circles represent Ni. The cyc le from 1 3 represents the shape memory effect and the cycle from 4 3 represents the pseudoelastic effect. Adapted from Wayman, C. M., and Duerig, T. W. 1990. Engineering Aspects of Shape Memory Alloys ( page 3 20) Butterworth Heinemann, U nited Kingdom. In the Figure 2 3 the austenite phase (in the absence of applied load) can be cooled (1) below its M f to the twined martensite phase; this results nearly zero volume or shape change ( 94 ) Deformation (2) in the twinned martensit e phase causes the twins to reorient along a dominant orientation, becoming detwinned martensite. Heating (3) the detwinned martensite above its A f temperature causes the reversion to the

PAGE 44

44 austenite phase in the same configuration the material started in T he cycle 1 2 3 is known as the shape memory effect. It has also been shown an increase in stress within the SMA at a given temperature will result in higher martensite and austenite transition temperatures ( 93 ) ( Figure 2 4 ) This is due to the equivalence between temperature and stress in stabilizing the martensite phase ( 95 ) Therefore, a mechanical transformation between austenite and martensite is also possible. Looking back at Figure 2 3 if the SMA is in the austenite phase an d a load being applied, a stress induced transformation (4) to martensite will result When the load is removed (3), the material immediately reverts back to its original austenite shape. This cycle, 4 3, is referred to as the pseudoelastic effect. Fig ure 2 4: Schematic showing how the transition temperatures (M s and A s ) are changed as a result of applied stresses. Adapted from Wayman, C. M., and Engineering Aspects of Shape Memory Alloys (page 3 20) Butterworth Heinemann, United Kingdom. Both of these behaviors, the shape memory effect and pseudoelasticity, have potential to be utilized to aid in self healing in metallic components ( 27 )

PAGE 45

45 Sh ape Memory Alloys in Composites Prior to their use in liquid assisted self healing MMCs, SMA reinforcements garnered significant attention because they have been found to increase the flow stress of the composite ( 96 ) Because of their inherent ability to transform shape, SMAs have also found use as reinforcements to enable crack closure ( 97 99 ) The role of interfacial debonding was investigated and important factors include composite temperature ( 100 ) and martensitic transformations as it relates to interfaci al shear stress ( 101 ) Early studies of SMA reinforcements in MMCs, especially Al based matrices, s howed similar strengthening and crack closure characteristics ( 102 104 ) The use of SMA reinforcement inclusion in polymer matrix composites (PMC) for enhanc ed healing properties revealed the need for optimization of fiber volume fraction and pre straining the SMA to decrease interfacial debonding ( 105 ) Recently, SMAs have also be en used to enable crack closure, thereby increasing crack healing ( 106 ) The SMA wires decrease the crack volume, thereby increasing the fill factor of the healing agents utilized in PMC. This is akin to the research in liquid assisted self healing in metals with SMA reinforcements. Summary There are t wo major types of self healing in metallic systems: solid state and liquid assisted. Solid state healing has been investigated for fatigue and creep crack reduction in Al and Fe based alloys, in addition to being utilized for self healing coatings. Liquid assisted healing has found applications in both electrical and composite materials. Solid state healing, however, is limited in applications due to defect size constraints and the reduction of healing material after each heal cycle. Liquid assisted healing however, has the advantage of faster healing kinetics due to the liquid phase.

PAGE 46

46 Also, within SMA reinforced metal matrix composites, the possibility to heal macro scale defects is realized using liquid assisted healing techniques because of the inherent s hape changing abilities of SMAs, making them the best candidate for further development. While the research at Northwestern advanced the field of liquid assisted self healing it also left several open questions to be addressed regarding the matrix: 1) how can oxidation be mitigated during a healing cycle, 2) the optimum time and temperature for a healing cycle, and 3) how percent liquid affects healing in non Sn based matrix systems. In regards to the incorporation of SMAs in the composite, the effect of v olume fraction of SMA wires on healing as well as the effect of casting and heat treatment on the SMA transition temperatures has also yet to be fully vetted. And finally, the most important question to be addressed, can self healing be demonstrated in a h igh strength composite system able to be used as a structural material?

PAGE 47

47 CHAPTER 3 PROTOTYPE SELF HEALING COMPOSITE Early work into self healing metal matrix composites was based on a prototype system developed by Files ( 25 ) and Manuel ( 80 ) which consists of a Sn Bi matrix con tinuously reinforced with commercially available NiTi shape memory alloy (SMA) wires This matrix alloy system was selected because it has a low cost, low melting temperature and good castability. In addition to being part of two doctoral students thesis work ( 25 80 ) several undergraduate stude nt projects were also conducted to investiga te healing within the Sn Bi composite system ( 26 71 77 7 9 82 ) This chapter will seek to improve upon this earlier work by investigating whether healing in Sn Bi can be established while using a mechanical means of attaching the matrix and wires. First, the design methodology for self healing in a Sn Bi composite is discussed, which lays the groundwork for the design of future higher strength alloys in subsequent chapters. Second, Sn Bi composites were fabricated into tensile test geometries and fractured un der different heat treatment conditions to demonstrate the effects of crack healing. Next, these composites are compared to Sn Bi alloy properties to understand the effect of SMA wire inclusion on the mechanical properties. The SMA wire mechanical and ther mal properties were also investigated to determine how the healing heat treatments modify the NiTi wire properties. Finally, an investigation was conducted to determine the clamping force required from the SMA wires to achieve healing in the Sn Bi matrix. Design Methodology Using the thermodynamic modeling software Pandat by CompuTherm, LLC ( 107 ) the Sn Bi phase diagram was calculated using the NIST Pb free solder database

PAGE 48

48 (Figure 3 1) ( 108 ) The calculated diagram was noted to be a simple binary eutectic, and matched experimental results available in the literat ure ( 109 ) Figure 3 1: Calculated Sn Bi phase diagram using NIST Pb free solder database ( 108 ) For Sn Bi alloys, the Sn rich side of the dia gram yields moderately ductile alloys with high castability ( 55 ) It can be ob served that the composition of Sn 21 wt% Bi (Sn 13.1 at% Bi) yields the largest temperature range between the eutectic temperature and the liquidus temperature. A large solidification range yields gradual melting during heat treatments. This ensures t hat small fluctuations of temperature are not likely to significantly modify the microstructure during heat treatment. At this composition, it was calculated that at 169C the alloy would be 20% liquid with the balance being Sn solid solution phase ( Figure 3 2). Therefore, 169 C was utilized as the healing temperature for this composition. It is important to note that previous studies have shown 20% liquid

PAGE 49

49 at the healing temperature results in a balance between enough healing agent, i.e. liquid eutectic ph ase, while still maintaining structural stability throughout the heat treatment ( 78 ) With a fluctuation of 5 C above or below 169C it was calculated the percent liquid could fluctuate between only 15.6 25.1%. Figure 3 2 : P hase fractions of Sn 21 wt % Bi at various temperatures. The heat treatment temperature of 169 C reveals a 20% liquid composition during the healing process. The box signifies the potential area of liquid % should the temperature fluctuate by 5 C during heat treatment. Sn Bi Matrix Alloy Characterization Alloys of Sn 21 wt% Bi were cast to investigate the matrix alloy mechanical properties. Appropriate amounts of tin (Sn shot, 99.8%, Alfa Aesar) and bismuth (Bi needles, 99.99%, Alfa Aesar) were melted in an open air furnace at 350 C until a liquid solution. The Sn Bi melt was poured into a graphite mold and allowed to cool. After casting, matrix compositions were verified via inductively coupled plasma (ICP)

PAGE 50

50 spectroscopy (Perkin Elmer Plasma 3200RL) and found to vary less than 2% f rom the nominal composition. The tensile bars were placed horizontally into a furnace for 24 hours at 169 C then air cooled ( Figure 3 3). An external thermocouple was utilized to monitor the stability of the furnace temperature. The heat treatment sets the matrix microstructure to ensure it is comparatively the same before and after a healing cycle. After air cooling post heat treatment, each alloy was polished to a 320 grit surface finish. Tensile testing was completed on an Instron 5582 until failure. Tab ulated r e sults of the testing are found in Table 3 1. Graphs of the tensile results can be found in Appendix A. Figure 3 3 : Representative microstructure of Sn Bi composite. The dark areas are the eutectic separating the lighter Sn solid solution phase. Table 3 1 : Tensile test results for Sn 21 wt% Bi alloys Specimen Modulus (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) 1 25.9 68 87.5 3.5 2 23.1 69 90.1 3.3 3 24.3 67 88.9 2.6 Avg. 24.4 68 88.8 3.1 1.4 1 1.3 0.5

PAGE 51

51 NiTi Wire Preparation The NiTi SMA wires, designated BH 0075 ( Ni 50.6 at% Ti, = 0. 18 mm cold drawn, Memry Corporation ) were prepared by vacuum encapsulating the wires and heat treating for 3 hours at 500 C. Previous research had shown the as received, cold drawn condition of these wires is inadequate for shape memory, therefore a heat treatment is required prior to casting ( 80 ) Since load transfer requires that the matrix be bonded to the NiTi SMA wires, chemically reactivity at the interface is an important criterio n. The driving forces for reactivity can be realized using an Ellingham diagram ( 110 ) An Ellingham diagram highlighting the constituent phases and elements at the interface are shown in Figure 3 4 It can be seen that the relative stability of TiO 2 in comparison to Bi 2 O 3 and SnO 2 illustrates the low reactivity between the Sn Bi matrix and NiTi wires. Figure 3 4 : Ellingham diagram showing the relative free energy calculations for Sn, Bi, Ni, and Ti. It is shown that any TiO 2 formed on the surface of the NiTi SMA wires is unlikely to be reduced by the Sn Bi matrix. Cur ves plotted with free energy data obtained from ( 110 )

PAGE 52

52 Previous re search has confirmed the poor bonding between the Sn Bi matrix and NiTi reinforcement ( 75 80 84 ) These earlier studies mitigated this issue by coating the NiTi wires with a thin layer of gold to improve adhesion with the matrix. However, the present research sought to establish a connection via mechan ical attachment, i.e. knots in the wires. Thus, in the area of the wires which would correspond to the grip section of a tensile bar, knots were tied to act as mechanical anchors and compensate for the low chemical bonding between the reinforcement and mat rix. With the NiTi wires prepared, a sub size tensile bar graphite mold, with measurements based on ASTM E 8M: Standard Test Methods for Tension Testing of Metallic Materials ( 111 ) with specially designed mold ends to anchor the wires was machined ( Figure 3 5 a ). The mold ends had holes cut through them to enable the incorporation of the SMA wires prior to casting. Appropriate amounts of tin (Sn shot, 99.8%, Alfa Aesar) and b ismuth (Bi needles, 99.99%, Alfa Aesar) were melted in an open air furnace at 350 C until a liquid solution. Three wires were placed into the mold and aligned longitudinally; each wire was clamped on the ends, outside the mold, to prevent the wires from sh ifting during casting. The Sn Bi melt was poured over the wires into the mold and allowed to cool ( Figure 3 5 b ). After casting, matrix compositions were verified via inductively coupled plasma (ICP) spectroscopy (Perkin Elmer Plasma 3200RL) and found to va ry less than 2% from the nominal composition. Each tensile bar was then placed horizontally into a furnace for 24 hours at 169 C then air cooled. An external thermocouple was utilized to monitor the stability of the furnace temperature. After air cooling, the virgin composites were polished to a surface finish of 320 grit. Each composite bar was tested in tension using an Instron 5582 mechanical testing

PAGE 53

53 Figure 3 5 : Images of mold used in casting metal matrix composites with SMA wire reinforcements: a) sp ecially designed mold end with holes for wire incorporation during casting and b) entire mold with Sn Bi matrix cooling after being poured over the wires. machine (strain rate = 1.0%/min) with an extensometer (Model 3542, Epsilon Technology Corp oration ) at tached to the gauge area. The bars were pulled to complete failure of the matrix while maintaining the integrity of the NiTi wires. Results for the virgin composite testing can be seen in Table 3 2 Curves of the representative stress strain behavior can b e found in Appendix A. The large standard deviations for the strength values are attributed to the poor wetting properties of the Sn Bi melt on the NiTi wires, which results in poor chemical bonding and thus low load transfer between the matrix and reinfor cement ( 112 ) This is despite the knots tied into the wires to maintain a mechanical attachment The composites were then placed on a steel plate and vacuum encapsulated. Each was laid horizontal ly in the furnace and heat treated again for 24 hours at 169 C Concurrently, the heat treatment causes the NiTi wires to begin to return to their previous shape and a partial liquefaction within the Sn Bi matrix to occur. Upon air

PAGE 54

54 Table 3 2 : Average me cha nical testing results for Sn 21 wt% Bi matrices reinforced with <0.5 vol% SMA wire reinforcements Specimen V f Wires (%) Modulus (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) A 0.31 23.6 71 90.7 4.3 B 0.37 22.3 59 72.5 5.9 C 0. 40 23.4 57 72.9 1.9 Avg. 0.36 23.1 62 78.7 4.0 0.04 0.7 7.6 10.4 2.0 cooling, it was observed the wires had returned back to their original length as the matrix was deformed back to its pre strained (virgin) length. The composites were tested in t ension again. Each composite was found to fail again at the same area as the original crack. The tension results were compared to the previous virgin composite tensile data to quantify the healing efficiency ( Table 3 3 ). Healing is established by comparing the ultimate tensile strength (UTS) of the healed and virgin composite tests via the following equation proposed by Manuel ( 80 ) : ( Eq. 3 1 ) Table 3 3 : Healing characteristics of Sn Bi composites Specimen UTS (MPa) % H eal A 73.0 80.5 B 68.2 94.1 C 49.6 68.1 Average 80.9 A representative tensile comparison of specimen B is shown in Figure 3 6 Crack healing in a Sn Bi composite is shown in Figure 3 7 The healed composites were found to retain similar strength v alues as the virgin composite, but none reached elongations above 1.5%. Using this method for Sn Bi composite development, it was established that over 94% healing was achievable in <0.5 vol% NiTi reinforced Sn Bi

PAGE 55

55 composite. These percent heal results were similar to those obtained by Files ( 25 73 ) and Manuel ( 80 84 ) Figure 3 6: Mechanical testing results of Sn Bi Composite B (V f wires = 0.37%) in virgin and healed condition. This healed condition was found to retain over 94% of the UTS of the virgin condition. Figure 3 7: Image showing crack healing in a Sn Bi composite. The top image is following tensile testing, the bottom image shows the same sample immediately following the healing heat treatment.

PAGE 56

56 When comparing Table 3 2 to Table 3 3 it was found that the specimens possess ing the greater failure strains were also found to obtain a greater healing percent. A schematic showing the different steps during a tensile test and healing as it relates to the NiTi wires is shown in Figure 3 8. In the figure, image A represents the com posite as it begins to be pulled in tension by the applied stress ( a ) at room temperature (T room ). This represents the normal length of the SMA wires when in the twinned martensite state. By image B the ductility of the composite has increased the length of the NiTi wires as the wires begin to reorient into detwinned mar tensite. At this point, a crack begins to initiate within the matrix. After full fracture of the matrix in image C, the load is removed after causing further reorientation of the NiTi into detwinned martensite. In image D, increasing the temperature of the composite above the A f temperature results in a clamping force exerted on the matrix from the NiTi wires as they revert to their previous shape. This compressive force, paired with softening of the matrix at elevated temperatures, begins to remove the exc ess length of the composite until the wires reach their initial length. Reducing the temperature back to T Room (which is below M f ) transforms the NiTi back into a twinned martensite state. From a comparison of the results in Table 3 1 for the matrix alloy to Table 3 2 for the composite specimens, it is shown that the modulus and failure strain of the alloys are statistically similar to those in the composite (using a standard t test with a 95% confidence). However, the alloys possessed strength values great er than the composites (Figure 3 9). This is most likely caused by the lack of bonding between the wires and the matrix.

PAGE 57

57 Figure 3 8: Schematic of changes to the NiTi SMA wires during a composite healing cycle. In A), the stress is applied ( a ). Image B ) shows lengthening of the SMA wires as a crack initiates in the matrix. Full fracture of the matrix occurs in C), resulting in reorientation of the NiTi to detwinned martensite. Increasing the temperature above A f in D) results in a compressive force from the NiTi wires as they begin to transfer to austenite. The initial length of the composite is recovered, as cooling of the composite in E) transforms the wires back to a twinned martensite state. Figure 3 9: R epresentative tensile results for Sn 21 wt% Bi alloy 2 compared to the Sn Bi composite B (V f wires = 0.37%)

PAGE 58

58 NiTi Mechanical Testing Previous research has shown the thermal and mechanical properties of NiTi SMAs can be modified through heat treatment ( 113 ) Therefore, an investigation was conducted on the NiTi SMA wires utilized for the Sn Bi composites to determine how the heat treatments affect the wire properties. Thermal properties to be investigated include the martensite and austenite transition temperatures to confirm the crystal structure of the NiTi wires throughout the healing process. Mechanical properties to be investigated include the plateau stress and t he failure stress. The plateau stress is the stress required to transform twinned martensite into detwinned martensite. The failure stress is the maximum stress the SMA wires can reach before wire failure. As the Sn Bi composites fail within the matrix but the wires remain intact this stress is an important design factor because the SMA wires must be strong enough to handle the load previously shared with matrix without breaking Pieces of the BH 0075 NiTi wire in the as received condition were placed in t o a Perkin Elmer 8000 Differential Scanning Calorimeter (DSC). Each was cycled 3 times from 50 200C at 10C/minute to investigate the austenite & martensite transition temperatures of the wire. No transformation was seen in the as received condition du e to the high level of cold work reducing twin boundary mobility, and thus the shape memory effect ( 94 ) Therefore, no austenite or martensite start/finish temperatures are reported. Other BH 0075 wires were vacuum encapsulated in Pyrex tubes and heat treated for 3 hours at 500C and then tested under similar conditions. Another set of wires were vacuum encapsulated in Pyrex tubes and heat treated for 3 hours at 500C, cooled, and then heat treated for 24 hours at 169C before being tested on the DSC to simulate Sn Bi composite hea t treatment conditions. The average and standard

PAGE 59

59 deviation for 3 wires under each condition are shown in Table 3 4. The transition temperatures as reported were calculated using the onset method ( 114 ) It was found that the low temperature (169C) required for healing in the Sn Bi composite had little effect on the NiTi wire martensitic transition temperatures Table 3 4: The austenite and martensite transitio n temperatures for heat treated NiTi SMA wires Transition Te Heat Treatment A s A f M s M f 500C for 3 hrs 113.8 0.3 130.7 1.8 69.8 0.1 59.3 0.7 500C for 3 hrs 169C for 24 hrs 92.3 0.7 107.6 0.9 67.9 0.1 60.6 0.6 Mechanical testing of similarly heat treated NiTi wires was also conducted. Similar heat treatments of the wires were conducted as above and pulled until failure on an Instron 5582. The plateau stress and failure stress were both noted ( Table 3 5). Representative curves of the tensile behavi or can be found in Appendix A. The initial heat treatment at 500 C was found to greatly decrease the failure stress, but enabled the appearance of mobile twin boundaries ( 113 ) and thus a martensitic transformation as shown by the plateau stress. Subsequent heat treatment at 169 C was found to have little effect on the plateau and failure stress of the NiTi wires. Representative cur ves of the tensile stress strain behavior can be found in Appendix A. Table 3 5: Mechanical testing results for heat treated NiTi SMA wires Avg. Heat Treatment Plateau Stress (MPa) Failure Stress (MPa) As Received N/A 1491 53 500C for 3 hrs 190 5 1183 5 500C for 3 hrs 169C for 24 hrs 190 5 1163 7

PAGE 60

60 To determine if the casting process had any effect on the mechanical properties of the SMAs, one of the Sn Bi composites with heat treated NiTi reinforcements was heat t reated for 24 hours at 169 C and air cooled, then heat treated again for another 24 hours at 169 C This was to match the wire mechanical properties found in a post healed Sn Bi composite. The composite was then placed into a solution of 2:1:1 water (H 2 O), nitric acid (HNO 3 ), and hydrochloric acid (HCl) until the Sn Bi matrix was completely dissolved away leaving only the NiTi wires. Each of the three wires was then pulled in tension until failure. Results are found in Table 3 6 whereas the stress strain cu rves can be found in Appendix A It was determined that the casting process and heat treatments had little effect on the NiTi wires other than a small increase in the plateau stress. Table 3 6: Mechanical properties of wires removed from heat treated Sn Bi composite Avg. Heat Treatment Plateau Stress (MPa) Failure Stress (MPa) 500C for 3 hrs 169C for 24 hrs (2x) 222 10 1187 25 Pressure Requirements for Healing In order to quantify the relationship between clamping force a nd crack healing ability, interface diffusion studies were conducted utilizing a Kovar diffusion couple setup. This work was performed in conjunction with an undergraduate senior thesis by W. Patterson Tuttle ( 115 ) Kovar was selected as the material for the diffusion couple jig because it exhibits very low thermal expansion (25 200 C = 5.5 x 10 6 / C ) over the temperature range st udied for self healing in Sn Bi ( 116 ) The Sn 21 wt% Bi alloy bar was cast and then heat treated in air for 24 hours at 169 C to set the microstructure. After cooling, pieces were cu t from the bar and polished to a 320 grit surface finish. A

PAGE 61

61 slice was cut through the center of the specimen to simulate a crack interface. Both sides of this interface were polished to a 4000 grit finish in order to ensure a smooth contact surface. After polishing, the pieces were loaded into the custom Kovar diffusion couple jigs which minimize pressure effects from the coefficient of thermal expansion caused by the jig itself ( Figure 3 10 ). Pressure was applied to the sample via a torque wrench. The torq ue wrench was calibrated to yield sp ecific forces at various torque levels thus allowing for varied pressures to be applied to the Sn Bi pieces. Figure 3 10 : Image of Sn Bi alloy loaded into the diffusion couple jig post heat treatment Note the centra l interface used to simulate a crack has been healed. Each jig and specimen setup was vacuum encapsulated in a Pyrex tube and heat treated for 24 hours at 169 C to heal the interface. After air cooling, the healed composite was pulled in tension via the In stron 5582 mechanical testing machine. Results are shown in Figure 3 1 1 using the metric defined in Equation 3 1 whe re the UTS virgin is 88.8 MPa. It is shown 30 MPa of stress is required to obtain the hig hest post healing strength this equates to slight ly less than 50% of the room temperature yield strength found in the Sn Bi alloys (Table 3 3). At greater stresses, the sample deformed and this is thought to have prevented the full compressive pressure from facilitating healing within the sample. It is l ikely the compressive yield strength of the Sn Bi matrix was reached at 40 MPa, the highest pressure, which would explain the

PAGE 62

62 Figure 3 1 2 shows the healed microstructure of the specimen compressed at 30 MPa Figure 3 1 1 : Comparison of tensile results from Sn Bi pieces compressed at various stresses during healing to simulate the clamping force from SMA wires. Figure 3 1 2 : Microstructure along interface of healed Sn Bi all oy compressed at 30 MPa Note the grain growth across the previous interface line. The black areas are pores likely caused by w edge type creep cavitation.

PAGE 63

63 As seen in Figure 3 1 2 p ores were evident along the healed interface. When investigating the fracture surfaces of the healed alloys, porosity was found to be present on every sample. The pores were also noted to increase in size ( Figure 3 1 3 ) b ut reduced in total area of the interface as the applied stress increased. At 20 MPa, they represented 23.4% of t he interface, at 30 MPa, 22.8%, and at 40 MPa 21.4%. Due to the nature of the cavities being present at a fixed stress in elevated temperatures and found at grain boundary triple points, they were determined to be wedge type (w type) creep cavities ( 117 ) These cavities represent areas of the sample which were not healed, thereby reducing the strength of the interface becaus e less material has joined together, and thus reducing the calculated percent healing due to this reduction in strength. Re calculating the percent heal while taking this reduction in area into ac count, the values all increase: 20 MPa = 46.0%, 30 MPa = 106 .5%, and 40 MPa = 73.3%. Figure 3 1 3 : Scanning electron microscopy (SEM) image of the crack face for Sn Bi alloys healed at a stress of a) 20 MPa, b) 30 MPa, and c) 40 MPa after tensile testing. Note how the size of individual cavities tends to increase with an increase in stress. Summary A methodology to design a self healing metal matrix composite was described. Sn 21 wt% Bi composites with < 0.5 vol% of SMA wires were fabricated and found to possess an average of 80% healing, with a high value of ove r 94% retained strength

PAGE 64

64 post healing. These composites were compared to Sn Bi matrix alloys and statistically found to possess similar mechanical properties. The NiTi wires were investigated to determine how casting and heat treatments varied their mechani cal properties. It was found that the healing cycle had only a minor effect on the mechanical and thermal properties of the heat treated SMA wires. Finally, it was established the SMA wires needed to provide 30 MPa of stress to obtain the greatest healing efficiency in a Sn Bi matrix. Due to the weight, strength, and usable temperatures for a Sn Bi matrix, the above experiments were not intended to establish a Sn Bi based composite for industrial application, but instead to serve as a prototype system. This Sn Bi system assisted in the development of a thermodynamics based methodology which can be utilized to design greater strength alloys for use as the matrix in a self healing metal matrix composite system. For the next iteration of matrix alloys, several characteristics are desired to enable self healing MMCs to be utilized for structural applications. First, increase the specific strength of the composite by lowering its overall density and increasing the strength of the matrix. The strength can be increa sed through using greater strength matrix materials and also through the formation of a chemical bond with the NiTi wires to ensure load transfer. Second, ensure the material has good castability to ensure wetting around the SMA wires. Finally, maintain th e high healing efficiency established by the Sn Bi composite system The latter point is the most important attribute for future iterations of self healing MMCs.

PAGE 65

65 CHAPTER 4 SYSTEM DESIGN METHODOLOGY In light of the positive results found in the Sn Bi matr ix reinforced with NiTi SMA wires as a self healing composite, further research into higher strength matrix materials was investigated. However, i dentifying potential matrix alloy systems that can be designed to exhibit self healing can be a tremendous and tedious undertaking due to the multi scale effects of the different variables involved When dealing with systems possessing multivariate interactions a systematic approach can be used to find the optimal combination of processing steps and microstructur es in order to achieve the desired behavior. T o facilitate the development of self healing composites, a systems design approach motivated by thermodynamics was used in determining element selection and processing conditions. System Design Chart The desig n methodology for matrix development utilized for these investigations was based upon systems design method put forth by Jenkins ( 118 ) He use d flow block diagrams to describe systems. These diagrams are broken do wn into sub systems which can be utilized to organize the overall objective for design. Modifying Jenkins model to illustrate the hierarchical nature of materials fabrication while incorporating the typical processing structure properties performance mater ials tetrahedron results in a system design chart. System design charts were first developed by Olson at Northwestern University ( 69 119 124 ) Figure 4 1 shows the design aspects for matrix development of a self healing metal matrix composite.

PAGE 66

66 Figure 4 1: System design chart for matrix development to be utilized within a self healing metal matrix composite The hierarchical aspect of the systems design chart is shown by having the most important feature higher on the chart in its respective category. The exception is i n the system, where the hierarchical nature shows the sequence of processi ng steps for composite fabrication. Each line connecting specific blocks to one another represent specific relationships between the sub systems. Those lines linking systems represent relationships that are thermodynamic o relationships can be described by empirical or physics based models. This chart will serve as the roadmap which details the direction for alloy design and experimental procedures, and be used to increase efficiency in designing the matrix alloys for use in a self healing metal matrix composite. Structure Property Model Development Figure 4 1 are physical in nature. The

PAGE 67

67 by the total composite mass (m c ) over the total composite volume (V c ) yielding the total composite density c ) as shown in Equation 4 1: ( Eq. 4 1 ) understood at this time. A senior thesis project by Steve Knapp at Northwestern University ( 78 ) found th at 15 20% eutectic was necessary for healing within a Sn 21 wt% Bi matrix alloy while still ensuring structural stability. However, it is not known if this will translate directly into other higher strength matrix alloy systems. The final relationship re ( 112 ) Equation 4 2 shows the volume averaged loads borne by the matrix and rei nforcement in the m = matrix, f = reinforcement fiber, and V f = volume fraction of the fiber: ( Eq. 4 2 ) The interface plays an important role as Equation 4 2 assumes perf ect bonding between the matrix and reinforcement. The concept of load sharing works best during elastic behavior, e.g. prediction of composite yield strength. A linear superposition proposed by Nembach ( 125 ) describes the effects of strengthening contributions to the alloy strength. This includes the strength of the main element ( Al ) the increase in solid solution strength from alloying elements ( SS ) and the strength of the eutectic phase ( Eut ) Thus, the strength of the matrix can be represented by:

PAGE 68

68 ( Eq. 4 3 ) w here the k exponent is a fit relatio nship based on modeling and has been found to be 1 < k < 2. Thus, the total strength of the Al based composite reinforced with SMA wires can be predicted by the following relationship: ( Eq. 4 4 ) This equation is simil ar to the one derived by Manuel for Mg based composite alloys ( 80 ) Alloy Selection To identify matrix alloys with self healing capabilities, the first step was the base alloy selection. As specific strength is a performance output (Figure 4 1), typical base metals w ould include Al, Mg, and Ti to obtain high strengths with low weight. However, because a reduction in eutectic temperature is desired in order to keep a low healing temperature, Al and Mg have advantages over Ti because of their lower melting points; 660 a nd 650C for Al and Mg, respectively ( 55 ) Mg requires special processing due to flammability issues, and was therefore removed from consideration as a base material. Al alloys, in addition to their low density and potential for high strengths, typically exhibit good castability, another beneficial characteristic for fabricating con tinuous fiber reinforced composites. The selection of Al as the base element for the matrix has an added benefit in that it will form a chemical bond with the surface of NiTi SMA wires. Unlike the Sn Bi matrix, Al 2 O 3 has lower formation energy than either the oxides of Ni or Ti ( Figure 4 2). Therefore during casting, the Al melt will reduce the TiO 2 on the surface of the NiTi

PAGE 69

69 wires, creating a strong chemical bond, relieving the need to create a mechanical attachment to incorporate the SMA wires (e.g. knott ing). Figure 4 2: Ellingham diagram showing relative positions of Al 2 O 3 NiO, and TiO 2 on a free energy scale during the formation of oxides. Curves plotted with free energy data obtained from ( 110 ) With the selection of Al as the main matrix element, the next step is the selection of alloying elements to obtain the eutectic microstructure required for healing. This investigation included alloying elements which would decrease the eutectic temperature in Al rich alloys to better facilitate efficient healing The eutectic temperatures of potential alloying elements to Al are shown in Figure 4 3. It is shown that Ge, Mg, Sn, and Zn are the only elements able to reduce the binary eutectic temperatures of Al rich binary alloys below 500C. The alloying elements added to the Al matrix must also be able to improve its yield strength of 12 MPa ( 55 ) The thermal cycles inherit in the healing performanc e

PAGE 70

70 necessitates alloying additions that exhibit good solid solution strengthening. The typical alloying additions to Al that exhibit solid solution strengthening are shown in Figure 4 4. Figure 4 3: Graph showing the eutectic temperature of various Al bas ed binary alloys. The red line denotes the melting point of pure Al ( 660C ). Graph plotted from phase diagram data found in ( 126 ) From a comparison of the results in Table 3 1 fo r the matrix alloy to Table 3 2 for the composite specimens, it is shown that the modulus and failure strain of the alloys are statistically similar to those in the composite (using a standard t test with a 95% confidence). However, the alloys possessed st rength values greater than the composites (Figure 3 9). This is most likely caused by the lack of bonding between the wires and the matrix. Note that Li extends out to 19 at% solute, but that data was not included for clarity. The trend lines were also inc luded to show differences at low atomic percent solute to help guide the eye and does not indicate nor imply a linear strengthening relationship.

PAGE 71

71 Figure 4 4: Graph of change in yield strength for binary Al alloys as a function of alloy content. The tren d lines of respective alloying additions are shown for clarity of information. Graph plotted with tabulated data found in ( 29 127 131 ) Combining the above information, there were four different binary alloys which were deemed appropriate for further consideration: Al Sn, Al Cu, Al Mg and Al Si. The first alloy that will be investigated, Al Sn, will explore an allo y system exhibiting the lowest eutectic temperature. Al Cu and Al Mg will be studied to investigate the effect of potent strengthening solutes which also decrease the binary eutectic temperature. The final alloy to be investigated will be Al Si because of its high castability, which is desirable for fabrication of a continuously reinforced composite.

PAGE 72

72 CHAPTER 5 BINARY ALLOY DESIGN Using the system design chart and the information gathered about binary aluminum alloys in the previous chapter, it was determi ned to investigate four different binary alloy systems: Al Sn, Al Cu, Al Mg, and Al Si. Aluminum Tin The first system investigated was Al Sn because it exhibited the lowest eutectic temperature at only 226 C As in Sn Bi, the Al Sn phase diagram is a eutec tic system ( Figure 5 1). Because of the large (Sn) + liquid region available for healing in the Al Sn system, a healing temperature of 250 C was selected. This would ensure any fluctuations in the furnace temperature would not result in the matrix moving i nto the solid (Al) + (Sn) region of the phase diagram and preventing liquid assisted healing. At 250 C, a composition of Al 19.5 at% Sn (51.6 wt% Sn) was calculated to have 20% liquid ( 132 ) Al Sn Fabrication Appropriate amounts of Al (Al shot, 99.99%, Alfa Aesar) and Sn (Sn shot, 99.8%, Alfa Aesar) were melted in an open air furnace at 750 C until a liquid solution. Three NiTi SMA wire s, designated BB 35 ( Ni 49.3 at% Ti, = 0. 87 mm Memry Corporation ) were laid horizontally in the graphite sub size tensile bar mold. The Al Sn melt was poured over the wires into the mold and allowed to cool. Each tensile bar was then placed horizonta lly into a furnace for 4 hour s at 250 C monitored by an external thermocouple, then air cooled to set the eutectic microstructure ( Figure 5 2). A study was conducted on Al 19.5 at% Sn and found that only 4 hours were needed to achieve the equilibrium mi crostructure; further heat treatment time had little effect.

PAGE 73

73 Figure 5 1: Phase diagram of the Al Sn system. At 250C, the composition yielding 20% liquid was calculated to be Al 19.5 at% Sn (51.6 wt% Sn). Adapted from Al Sn (Aluminum Tin) [database on the Internet]. ASM International. 1992 [ cited September 2012]. http://products.asminternational.org/hbk/index.jsp Figure 5 2: Representative microstructure of an Al 19.5 at% Sn (51.9 wt% Sn) composite after heat treatment at 250 C for 4 hours.

PAGE 74

74 Al Sn Mechanical Testing After heat treatment, the Al Sn composite bars were ground to a 320 grit surface finish before being tested in tension using the Instron 5582 machine at a rate of 1.0%/m in. Results of the testing are shown in Table 5 1. Because of the high Sn content, the composites were found to possess low yield and ultimate strengths (less than Sn 3. T he duct ility of the composite was evident in the sample as multiple cracks began forming along the gauge section, resulting in the dips seen in the tensile data. Table 5 1: Mechanical testing results for Al Sn matrices reinforced with <2.0 vol% SMA wire reinforce ments Specimen V f Wires (%) Modulus (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) A 1.54 49.6 23.9 47.1 5.8 B 1.51 49.9 23.9 37.0 7.4 C 1.78 44.2 23.1 39.8 8.6 Avg. 1.61 47.9 23.6 41.3 7.2 0.15 3.2 0.5 5.2 1.4 Figure 5 3: Tensile data of Al 19.5 at% Sn (51.9 wt% Sn) composite heat treatment at 250 C for 4 hours

PAGE 75

75 After initial tensile testing, each composite was placed into a furnace for 4 hours at 250 C to ind uce healing. Upon removal, large cracks were still evident and healing was not noted to occur. Figure 5 4 shows the extra cracking along the gauge section. These extra cracks did not have an SMA wire passing through them to enable crack closure during a he crack sites to assist with healing and the cracks remained open post healing. Necking was also noted in the samples and was thought to also contribute to the lack of healing. During necking, the deformation is concentrated to the necked region, decreasing the cross sectional area but still strain hardening the composite. It is thought the concentrated deformation may have caused permanent strains to occur in the SMA wires, decreasing their abi lity to revert back to their original length and thus complete crack closure to enable healing. Figure 5 4: Image of Al 19.5 at% Sn (51.9 wt% Sn) composite after tensile testing showing cracking in the matrix caused by excess ductility Aluminum Copper Al Cu was studied because it had the potential to display high strength increases with increasing solute concentration while simultaneously decreasing the eutectic temperature of the binary system. This is why it was selected over Al Mn which exhibited gr eat er strength increases per atomic percent solute but lacked the eutectic temperature reduction. Literature has shown compositions below 10 wt% Cu were subject to hot cracking during casting ( 133 ) therefore a composition of Al 4.5 at% Cu

PAGE 76

76 (10 wt% Cu) was selected. The phase diagram was calculated using Pandat software with the PanMagnesium database ( 107 ) ( Figure 5 5); it was found to closely match the phase diagram found in literature ( 134 ) It was calculated that at the composition of Al 4.5 at% Cu, the healing temperature would be 566 C for this system. Should fluctuations occur within the furnace, 566 C 5 C yields a range of only 18 21.8% liquid ( Figure 5 6). Figure 5 5: P hase diagram of the Al Cu system below 40 at% Cu. At a composition of Al 4.5 at% Cu (10 wt% Cu), a healing temperature of 566 C will yield 20% liquid in the matrix. Calculated usi ng Pandat ( 107 ) Al Cu Fabrication Al (Al shot, 99.99%, Alfa Aesar) and Cu (Cu shot, 99 .9%, Alfa Aesar) were melted in an open air furnace at 800 C until a liquid and cast into a graphite mold to create an ~30% Cu master alloy. After verifying the master alloy composition via ICP,

PAGE 77

77 Figure 5 6: Phase fraction of Al 4. 5 at% Cu at various temperatures. The heat treatment temperature of 5 66 C reveals a 20% liquid composition during the healing process. The box signifies the potential area of liquid % should the temperature fluctuate by 5C during heat treatment. the Al Cu master alloy and m ore pure Al were melted in an open air furnace at 750 C until a liquid solution. One NiTi SMA wire, designated BB 35 ( Ni 49.3 at% Ti, = 0. 87 mm Memry Corporation ) was laid horizontally in the graphite tensile bar mold. The mold was heated up to 350 C before casting to prevent cold shuts and air pockets from forming in the as cast microstructure. The Al Cu melt was poured over the wire into the mold and allowed to cool. Each tensile bar was then placed horizontally into a furnace for 2 4 hour s at 566 C monitored by an external thermocouple, then air cooled to set the eutectic microstructure ( Figure 5 7).

PAGE 78

78 Figure 5 7: Representative microstructure of an Al 4.5 at% Cu (10 wt% Cu) matrix reinforced with NiTi SMA wire after heat treatment at 566 C for 2 4 hours and tensile testing The lack of adhesion between the SMA wire and the matrix is attributed to debonding from mechanical testing. Al Cu Mechanical Testing After heat treatment, the Al Cu composite bars were ground to a 320 grit surface finish befo re being tested in tension using the Instron 5582 machine at a rate of 1.0%/min. Results of the testing are shown in Table 5 2. The Cu addition resulted in composites possessing greater modulus and ultimate strengths than the Al Sn composites ( Table 5 1). Be cause of the lack of ductility, the 0.2% yield stress was not able to be calculated (Figure 5 8) Table 5 2: Mechanical testing results for Al Cu matrices reinforced with 2 2.5 vol% SMA wire reinforcements Specimen V f Wires (%) Modulus (GPa) Ultimate Str ess (MPa) Failure Strain (%) A 2.4 67.6 88.1 0.22 B 2.2 80.1 104.7 0.2 Avg. 2.3 73.9 96.4 0.21 0.13 8.8 11.7 0.01

PAGE 79

79 Figure 5 8 : Tensile data of Al 4.5 at% Cu (10 wt% Cu) composite after heat treatment at 566 C for 4 hours. Following initial testing, the samples were encapsulated under vacuum and heat treated again for 24 hours at 566 C The composites were placed on flat graphite pieces and covered with tantalum foil to prevent oxide from forming along the crack interface which would prevent healing. Following air cooling, the tensile samples were tested in tension again for he aling, but little healing was found ( Figure 5 9 ). Small cracks were still noted on the outer edges post healing (Figure 5 9 c d). Wire shifting was also noted in the sample (Figure 5 9 d ), which is likely to be caused by a release of residual stress due to t he increase in healing temperature.

PAGE 80

80 Figure 5 9 : Image of tensile crack induced in Al 4.5 at% Cu composite: a) pre heal top view, b) pre heal bottom view, c) post heal top view, and d) post heal bottom view. Note the wire shifting during heat treatme nt resulted in a longitudinal crack along grain boundaries in d). Aluminum Magnesium Al Mg was studied because it has the potential of the greatest total strength increases through solid solution strengthening in binary aluminum alloys ( Figure 4 4). The ph ase diagram was calculated using Pandat software with the PanMagnesium database ( 107 ) ( Figure 5 10 ). The diagram nearly matches that found in literature ( 135 ) but has the Al 3 Mg 2 phase as a line compound with low solubility. Literature values indicate the Al 3 Mg 2 phase should be present from 38.5 40.3 at% Mg ( 135 ) T he composition of Al 16.6 a t% Mg ( 15.2 w t% Mg ) yields the largest freezing range between the eutectic temperat ure and the liquidus temperature allowing for a robust

PAGE 81

81 matrix should there be slight fluctuations in the furnace temperature during healing At this composition, the healing temperature was calculated to be 487 C Should fluctuations occur within the furn ace, 487 C 5 C yields a range of only 17 23% liquid ( Figure 5 1 1 ). Figure 5 10 : Phase diagram of the Al Mg system. At a composition of Al 16.6 at% Mg (15.2 wt% Mg), a healing temperature of 487 C will yield 20% liquid. Calculated using Pandat ( 107 ) Al Mg Fabrication Al (Al shot, 99.99%, Alfa Aesar) and Mg (Mg chips, 99.98%, Sigma Ald rich) were melted in a furnace at 750 C within an argon atmosphere until a liquid solution. One NiTi SMA wire, designated BB 35 ( Ni 49.3 at% Ti, = 0. 87 mm Memry Corporation ) was laid horizontally in the graphite tensile bar mold which was heated up t o 350 C before casting to prevent casting defects. The Al Mg melt was poured over the wire into

PAGE 82

82 Figure 5 1 1 : Phase fraction of Al 16.6 at% Mg at various temperatures. The heat treatment temperature of 487 C reveals a 20% liquid composition during the healing process. The box signifies the potential area of liquid % should the temperature fluctuate by 5C during heat treatment. the mold and allowed to cool. Each tensile bar was then encapsulated under vacuum in Pyrex tubes and placed into a furnace fo r 2 4 hour s at 487 C monitored by an external thermocouple. The bars were then air cooled to set the eutectic microstructure (Figure 5 1 2 ). A reaction zone was found at the wire/matrix interface likely caused by a reduction reaction between the matrix and the surface oxide present on the NiTi wire. The depth of this zone found it to be approximately deep on average. Al Mg Mechanical Testing After heat treatment, the Al Mg tensile composite bars were ground to a 320 grit surface finish and tested in te nsion using the Instron 5582 machine at a rate of 1.0%/min. Results of the testing are shown in Table 5 3. B ecause of the lack of ductility,

PAGE 83

83 Figure 5 1 2 : Representative microstructure of an Al 16.6 at% Mg (15.2 wt% Mg) matrix reinforced with a NiTi S MA wire after heat treatment at 487 C for 24 hours. The reaction zone surrounding the SMA wires was noted to be approximately deep. the 0.2% yield stress was not able to be calculated (Figure 5 13). While the modulus was similar for each composite, the ultimate stress had a wide range; increasing the volume fraction of SMA wires did not have an effect on the composite streng th. Figure 5 1 3 : Tensile data of Al 16.6 at% Mg (15.2 wt% Mg) composite after heat treatment at 487 C for 24 hours.

PAGE 84

84 Table 5 3: Mechanical testing results for Al Mg matrices reinforced with < 5 vol% SMA wire reinforcements Specimen V f Wi res (%) Modulus (GPa) Ultimate Stress (MPa) Failure Strain (%) A 1.91 53.6 107.2 0.19 B 3.97 56.8 62.1 0.11 C 2.12 52.3 83.3 0.15 Avg. 2.67 54.2 84.2 0.15 1.14 2.3 22.6 0.04 Following the virgin composite testing, the samples were encapsulated under vacuum again and heat treated a second time for 24 hours at 487 C Following air cooling, the Al Mg composites were tested in tension again for healing. L ittle evidence of healing was found in the composites due to the oxidation of the entire surface ( Figure 5 1 4 ), including the fracture interface. Despite the use of vacuum during heat treatment, it appears some oxygen remained to react with the liquefied A l Mg eutectic phases. Additionally, matrix vaporization was evident by the coating on the Pyrex tubes used during heat treatment (Figure 5 1 5 ). Looking at the microstructure, it was also evident that there was a loss of material during the heat treatment. Figure 5 1 6 shows a view of the edge of the composite where an oxide layer formed during heat treatment. It should be noted the black areas are pores where eutectic was thought to have previously occupied; these areas are evident throughout the matrix.

PAGE 85

85 Figure 5 1 4 : Ima ge of Al 16.6 at% Mg composite a) pre heal and b) post heal. Note the black oxide covering the entire surface of the tensile bar after the heal heat treatment 24 hours at 487 C under vacuum. Figure 5 1 5 : Image of Pyrex tubes holding Al Mg composite after heat treatment. The black color on the interior surface is likely vaporized magnesium from the liquid eutectic during the heat treatment.

PAGE 86

86 Figure 5 1 6 : Representative microstructure of Al Mg composite after heat treatment at 487 C for 24 hours under vacuum. This image is of the sample edge to show the oxide layer that forms, but the pores (black areas) are formed throughout the microstructure. The pores are areas previously occupied by eutectic which is thought to have liquefied du ring heat treatment and vaporized. Aluminum Silicon The Al Si system was studied because it represents an alloy system yielding moderate strength increases, a decrease in eutectic temperature, and the binary alloy known to have excellent castability ( 55 ) The phase diagram ( Figure 5 1 7 ) w as calculated using Pandat software ( 107 ) with the PanMagnesium database, and found to closely match that found in literature ( 136 ) Literature has shown compositions below 3 wt% Si are subject to hot cracking during casting ( 133 ) therefore a composition of Al 3.0 at% Si (3.1 wt% Si) was selected. At this composition, the healing temperature was calculated to be 592 C S hould fluctuations occur within the furnace, 592 C 5 C yields a range of only 18 22.6% liquid ( Figure 5 1 8 ).

PAGE 87

87 Figure 5 1 7 : Phase diagram of Al Si below 50% Si. At a composition of Al 3.0 at% Si (3.1 wt% Si), a healing temperature of 592 C will yield 20% liquid. Calculated using Pandat ( 107 ) Figure 5 1 8 : Phase fraction of Al 3.0 a t% Si at various temperatures. The heat treatment temperature of 592 C reveals a 20% liquid composition during the healing process. The box signifies the potential area of liquid % should the temperature fluctuate by 5C during heat treatment.

PAGE 88

88 Al Si Fabri cation Al (Al shot, 99.99%, Alfa Aesar) and Si (Si lump, 99.9999%, Alfa Aesar) were melted in a furnace at 850 C and cast into a graphite mold to create an approximately 15% Si master alloy. After verifying the composition via ICP, appropriate amounts of t he Al Si master alloy and pure Al were melted at 750 C until a liquid solution to obtain the desired Al 3.0 at% Si composition. One NiTi SMA wire, designated BB 35 ( Ni 49.3 at% Ti, = 0. 87 mm Memry Corporation ) was laid horizontally in the graphite tensile bar mold which was heated up to 350 C before casting to prevent casting defects. The Al Si melt was poured over the wire into the mold and allowed to cool. Each tensile bar was placed into a furnace for 2 4 hour s at 592 C and then air cooled to set the eutectic microstructure ( Figure 5 1 9 ). Figure 5 1 9 : Representative microstructure of an Al 3.0 at% Si (3.1 wt% Si) composite after heat treatment at 592 C for 24 hours Al Si Mechanical Testing After heat treatment, the Al Si tensile composite bar s were ground to a 320 grit surface finish and tested in tension using the Instron 5582 machine at a rate of

PAGE 89

89 1.0%/min. Results of the testing are shown in Table 5 4. T he Al Si composites showed moderate yield and ultimate stresses and moderate ductility in relation to the other binary alloys. The failures were found to occur along the eutectic regions as shown in Figure 5 20 Table 5 4: Mechanical testing results for Al Si matrices reinforced with 2 3 vol% SMA wire reinforcements Specimen V f Wires (%) Modul us (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) A 2.26 74.9 38.9 97.2 4.21 B 2.00 48.4 40.2 83.3 3.18 C 2.40 63.1 39.0 100.4 4.42 Avg. 2.22 62.1 39.4 93.6 3.94 0.20 13.3 0.7 9.1 0.66 Figure 5 20 : Representative micro structure in an Al 3.0 at% Si (3.1 wt% Si) composite after heat treatment at 592 C for 24 hours showing failure occurring along eutectic regions Following the virgin composite testing, the samples were encapsulated under vacuum (in Pyrex lying on a grap hite strip and wrapped in Ta foil) and heat treated a second time for 24 hours at 592 C Following air cooling, the Al Si composites were

PAGE 90

90 tested in tension again for healing. Using Equation 3 1, the percent healing was calculated for the samples exhibiting healing (Table 5 5). Table 5 5: Healing characteristics of Al Si composites Specimen Stress (MPa) % Heal A 88.1 90.6 B 93.7 112.5 C 72.0 71.7 Average 91.6 New ways to investigate the healing efficiency were also be performed usin g similar calculations for the elastic modulus (E), y f ). The new healing efficiency calculations are shown in Equations 5 1 to 5 3. T he healing efficiencies of the Al Si composites are shown in Table 5 6. It was found E and y were retained at averages of over 90%, similar to f was retained at over 88% on average. These results, paired with the UTS healing efficiencies, show nearly full healing of macro sized cracks occurred within the Al Si composite alloys. ( Eq. 5 1 ) ( Eq. 5 2 ) ( Eq. 5 3 ) It was noted that each specimen showed visible signs of healing (Figure 5 21 ); i.e. crack size reduction. A comparison between the virgin and healed tensile behavior 22 All three of the healed composites were found

PAGE 91

91 Table 5 6 : Healing characteristics of Al Si composites using healing efficiency calculations for elastic modulus, yield strength, and strain to failure Specimen E ( G Pa) % E Heal y ( M Pa) % y Heal f ( G Pa) % f Heal A 80.9 108.0 38.8 99.7 3.57 84.8 B 59.8 123.6 37.7 93.8 4.08 128.3 C 40.0 63.4 37.4 95.9 2.30 52.0 Average 98.3 96.5 88.4 to retain similar modulus and yield strength values as the virgin composite. For great er ultimate strength in the healed composite is attributed to healing a defect which resulted in the premature failure thereby allowing for a great er ultimate strength post hea ling and thus a healing efficiency over 100% the failure in the composite post healing actually occurred in a different place than the original crack (Figure 5 23 ) (bottom right side of Figure 5 23 ) after the initial tensile test in the grip section of the specimen. This crack was not healed during the healing cycle because the crack was opened in an area without a NiTi SMA wire, thus there was no clamping force imparted across t he crack. The lack of crack closure prevents the partially liquefied matrix alloy from completely filling the crack, and thus it was prevented from being healed. However, this crack did not serve as the final matrix failure position of the composite specim en. Figure 5 21 : Comparison of the a) pre heal to b) post heal tensile bar in an Al 3.0 at% Si composite reinforced with 2.0 vol% NiTi SMA wire in Table 5 5)

PAGE 92

92 Figure 5 2 2 : Comparison of the virgin and healed tensile behavior of an Al 3.0 at% Si composite reinforced with 2.0 vol% NiTi SMA wire. The sample was found to possess 90.6% retained tensile strength post healing heat treatment 5) Figure 5 2 3 : Comparison of the a) post tensile 1, b) post heal and c) post tensile 2 in an Al 3.0 at% Si composite reinforced with 2.4 vol% NiTi SMA wire (specimen 5) The arrow in a) shows the ori ginal composite failure and the arrow in b) shows the healed crack in the same location In c) the oval shows th e second failure that occurred during second tensile test.

PAGE 93

93 Increased Volume Fraction NiTi To study whether increasing the volume fraction of NiTi wires would increase healing, more Al Si composites were manufactured as above, but with an increased number o f NiTi wires. After a heat treatment of 592C for 24 hours, the Al Si tensile composite bars were ground to a 320 grit surface finish and tested in tension using the Instron at a rate of 1.0%/min. Results of the testing are shown in Table 5 7 The new Al S i composites with nearly double the volume fraction of SMA wires showed more variance in the resultant elastic modulus, yield and ultimate stresses, and strain to failure in relation to the lower volume fraction composites. However, the 0.2% yield strength was found to have increased over 20% on average. The other properties were found to be statistically similar using a t test with a 95% confidence interval. Table 5 7 : Mechanical testing results for Al Si matrices reinforced with 3.5 4.5 vol% SMA wire rein forcements Specimen V f Wires (%) Modulus (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) D 4.43 68.7 49.5 108.9 6.24 E 3.79 73.7 45.7 89.9 2.80 Avg. 4.11 71.2 47.6 99.4 4.52 0.45 3.5 2.7 13.4 2.43 Following the virgin composite testing, the greater volume fraction composite bars were encapsulated under vacuum (in Pyrex lying on a graphite strip and wrapped in Ta foil) and heat treated a second time for 24 hours at 592 C Following air cooling, the Al Si composites were tested in tension again for healing. Using Equation 3 1, the percent healing was calculated for the samples exhibiting healing (Table 5 8 ).

PAGE 94

94 Table 5 8 : Healing characteristics of Al Si composites with 3.5 4.5 vol% NiTi wires Specimen Ultimate Stress (MPa) % Heal D 35.1 32.2 E 50.2 55.8 Average 44.0 It was noted the Al Si composites with a great er volume fraction of NiTi wires did not heal as well as the lower volume fraction composites. Look ing at Figure 5 2 4 when comparing specimen A to specimen D, the increase yield strength is shown. But the tensile data for specimen E also shows numerous decreases in stress followed by recovery. These decreases were accompanied by a cracking sound during testing and are attributed to debonding between the NiTi fibers and Al Si matrix ( Figure 5 2 5 ). The debonding is seen by the decrease in stress at various strains and was accompanied by a cracking sound during tensile testing. This debonding is thought to be the reason for the decreased healing percentages found in the composites with a greater volume fraction of NiTi. Debonding would prevent the wires from pulling the matrix together when as the temperature is increased during a healing heat treatment. It is thought the increased debonding is a result of the extra turbulence during the casting process when the composite is originally fabricated. This extra turbulence could decrease how the molten alloy solidifies around the NiTi wires, therefore decreasing the bond strength at the interface.

PAGE 95

95 Figure 5 2 4 : Virgin tensile behavior of an Al 3.0 at% Si composite reinforced with either 2.26 vol% o r 4.43 vol% NiTi SMA wire. Great er volume fraction composites were shown to possess greater yield strengths, but exhibited debonding under tensile testing. Figure 5 2 5 : Optical image showing debonding between an Al 3 at% Si matrix and NiTi wire (V f = 4.43%) following tensile testing

PAGE 96

96 NiTi Wire Properties The properties of the NiTi wires were investigated to deter mine the effects of heat treatment on their output properties. Pieces of as received BB 35 wire were placed into a Perkin Elmer 8000 Differential Scanning Calorimeter (DSC) and cycled from 50 200 C at 20 C /minute. The austenitic and martensitic transiti on temperatures of the NiTi wire were calculated via the offset method ( 114 ) Another wire was cut from an as cast Al Si composite to determine whether the c asting process had any effect on the thermal properties. Other pieces of wire were encapsulated under vacuum in Pyrex tubes and heat treated either once or twice at 592 C for 24 hours to simulate the heat treatment of the Al Si composites before being test ed in the DSC. Results of the testing are found in Table 5 9 It was found that the casting process significantly increased all of the NiTi transition temperatures from the as received condition. The subsequent heat treatment ( 592C for 24 hrs) required to alter the Al Si composite matrix microstructure decreased these values, but the healing heat treatment ( a second cycle of 592C for 24 hrs) did not significantly change the transition temperatures. Table 5 9 : Average Thermal Properties of NiTi Wires Subj ected to Various Heat Treatments Temperature ( C) NiTi Wire Condition A s A f M s M f As Received 4.5 13.2 8.8 10.0 As Cast 54.8 72.7 38.3 15.3 592C for 24 hrs 40.0 53.4 15.8 3.4 592C for 24 hrs (2x) 40.0 53.5 15.4 4.2 The phase diagram of Ni Ti al loys ( 137 ) is shown in Figure 5 2 6 whereas a time temperature transformation (TTT) diagram ( 138 ) for a nickel rich Ni Ti SM A is shown in Figure 5 2 7 The heat associated with casting is thought to have caused formation of a metastable Ti 3 Ni 4 phase which has been shown to result in increases in martensite and

PAGE 97

97 austenite transformation temperatures ( 139 ) Further heat treatment at 592 C would enable a precipitation of the intermediate Ti 1 1 Ni 14 and Ti 2 Ni 3 phases, which would decrease the transformation temperatures ( 140 ) The next heat treatment cycle did not result is shifting these temperatures because the precipitation process requires more time to transform to TiNi and TiNi 3 equilibrium phases. Because the Ti 3 Ni 4 Ti 11 Ni 14 and Ti 2 Ni 3 phases are metastable, they do not show up on the phase diagram shown in Figure 5 2 6 Figure 5 2 6 : Phase diagram of the Ni Ti system The composition of the SMA BB wires used for self healing MMC fabrication is Ni 49.3 at% Ti Adapted from Ni Ti (Nickel Titanium) [database on the Internet]. ASM International. 1992 [cited May 2013]. Available from: http://prod ucts.asminternational.org/hbk/index.jsp

PAGE 98

98 Figure 5 2 7 : TTT diagram showing aging behavior in a Ti 52Ni alloy Adapted from Nishida, M., Wayman, C.M., and Honma, T. 1986. Precipitation Processes in Near Equiatomic TiNi Shape Memory Allo ys. Metallurgic al Transactions A, Vol. 17A (page 1507, Figure 3) Next, heat treated NiTi wires under similar conditions as above were tested in tension to investigate thermal effects on mechanical properties. The samples were pulled on the Instron mechanical testing mac hine to failure ( Table 5 10 ). As with the thermal properties, the mechanical properties significantly changed with heat treatment, decreasing the plateau stress over 85% and the failure stress by over 20%. However, a subsequent heat treatment to simulate h ealing of the composite had little effect on the mechanical properties. These results correlate with the thermal properties shown in Table 5 9 also showing little change between the initial heat treatment and the second one because of the precipitation pro cesses associated with the heat treatments.

PAGE 99

99 Table 5 10 : Tensile Properties of NiTi Wires Subjected to Various Heat Treatments NiTi Wire Condition Plateau Stress (MPa) Failure Stress (MPa) As Received 492 19 1368 13 592C for 24 hrs 67 3 1087 14 592C for 24 hrs (2x) 71 2 1091 18 Binary Alloy Summary Of the four different binary systems with potenti al for self healing studied, only one matrix, Al Si, was found to have significant healing. When comparing the different tensile results of the four systems ( Figure 5 2 8 ), it was noted that Al Si exhibited uniform elongation without necking. It is thought the moderate ductility allows for transformation in the SMA wires, which is analogous to pre straining the wires. Modeling of the self healing process has shown pre strain in the SMA wires is required in order to obtain the clamping force required for heal ing ( 81 ) The brittle Al Cu and Al Mg matrices do not allow this to occur as matrix failure occurred before transformation in the SMA wires began. The necking and multiple cracking in Al Sn is thought to cause permanent strains in the wires, preventing full shape reversion in the SMA wires, and thus lack of healing. Therefore, i t is postulated in order to create a MMC with self healing capabilities that moderate ductility in a matrix is required to ensure uniform elongation, enabling transformation in the SMA wires, but without necking to cause permanent deformation in the reinfo rcement.

PAGE 100

100 Figure 5 2 8 : Comparison of the tensile results for different specimens from the Al Sn, Al Cu, Al Mg, and Al Si composite systems with 1.5 2.5 vol% NiTi wires. Note that only the Al Si composites were found to heal appreciably.

PAGE 101

1 01 CHAPTER 6 TERNA RY ALLOY DESIGN Using the system design method for self healing metal matrix composite design laid out in Chapter 4 and the information gathered about binary aluminum alloys in Chapter 5, a great er strength aluminum based ternary alloy system was investiga ted. The investigation included design of the matrix, fabrication of composites for tensile testing, and healing the fractured composites. The models utilized for design of the matrix and composite were validated through additional testing. Due to issues w ith lack of ductility, different fabrication methods were investigated in an effort to increase healing. Aluminum Copper Silicon Matrix Design Looking at the investigated binary alloys Al Cu, Al Mg, Al Si, Al Sn self healing was only found in the Al Si system, but the greatest strength values were obtained from Al Cu composites. Literature relates how Cu additions to Al Si will increase castability ( 141 ) and increase solid solution strength ( 5 5 ) both desirable properties based on the system design chart (Figure 4 1). The ternary phase diagram is known to not form ternary compounds ( 141 142 ) From Chapter 4, it is known both Cu and Si reduce the eutectic temperature of Al (Figure 4 3), in addition to being solid solution strengtheners in Al (Figure 4 4). The Al Cu Si phase diagram was calculated using Pandat software with the PanMagnesium database ( 107 ) ( Figure 6 1); it was found be similar compositionally to the phase diagrams found in literature ( 143 144 ) but showed a lower eutectic temperature 512 C vs 524 C ( 145 ) Since the calculated ternary eutectic temperature

PAGE 102

102 was off by over 10 C from the accepted literature v alue, an experiment was undertaken Figure 6 1: Liquidus projection of the Al Cu Si system. The blue lines indicate isothermal lines to indicate the movement of the liquid across different compositions. The lowest melting eutectic point calculated to be 5 12 C at a composition of Al 15.0 Cu 7.3 Si (at%) as indicated by the arrow Calculated using Pandat ( 107 ) to investigate whether the difference affected the resultant percent liquid during the healing heat treatments. First, a piece of cast Al Cu Si matrix had its composition checked via inductively coupled plasma (ICP) spectroscopy on a Perkin Elmer Plasma 3200RL machine and found to vary less than 0.3% from the nominal composition. The Al Cu Si piece was placed into the DSC and cycled 3 times from 400 650 C at 10 C /minute to investigate the eutectic and solidus temperatures. Using a technique put forth by Chen and colleagues ( 146 ) the percent liquid was calculated for the Al Cu S i alloy. Upon comparison to the output from Pandat, the values were found

PAGE 103

103 to vary by less than 2% liquid up to 550 C ( Figure 6 2). A representative DSC curve can be found in Appendix A. Figure 6 2: The percent liquid in an Al Cu Si alloy comparing DSC da ta to calculations using Pandat software The microstructure of the Al 4.1Cu 2.0Si (at%) matrix was also investigated to validate the expected percentage of liquid. Because the heat treatments used in healing are in the solid + liquid phase region, a st rong correlation between percent liquid and percent eutectic in the microstructure is found. Therefore, a cast Al Cu Si sample was mounted in acrylic and polished to a final 0.03 ( 147 ) etchant 95 mL H2O, 2.5 mL, HNO3, 1.5 mL HCl, 1.0 mL HF was us ed to increase grain contrast by immersing the mounted sample for 15 seconds. Following etching the microstructure was imaged and uploaded to ImageJ software. Across several microstructural images, the percent eutectic was calculated to be 21 .1% 2.3% (A vg. ), which encompasses the goal of 20% liquid in the matrix during healing heat treatment.

PAGE 104

104 Studying the similarities between the calculated Pandat values compared to DSC results, in addition to the percent eutectic being near the 20% goal via microstruct ural analysis, the thermodynamic based methodology utilized for this research for the Al Cu Si system is validated. The next step in Al Cu Si matrix design was to calculate the isotherm at the ternary eutectic temperature of 512C (Figure 6 3 for the Al ri ch region). Creating a line between the eutectic point and the closest solid solution Al region, a pseudo isopleth was calculated (Figure 6 4). Note how closely it resembles the binary phase diagrams calculated in Chapter 5. At a healing temperature of 530 C, it was calculated that a composition of Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) would yield 20% liquid during a healing heat treatment. Structural stability of the composite during heat treatment was assured as fluctuation of 5C from the healing temperature resulted in a range of only 18 21% liquid in the matrix (Figure 6 5). Figure 6 3: Isotherm at 512 C in the Al Cu Si system. The dashed line represents the tie line between the ternary eutectic and nearest Al solid solution phase field. Calculated using Pandat ( 107 )

PAGE 105

105 Figure 6 4: Pseudo isopleth of Al Cu Si system along the line from the ternary eutectic point to the closest Al solid solution phase region. At a healing temperature of 530 C, a composition of Al 4.1 at% Cu 2.0 at% Si will yield 20% liquid in the matrix. The left side is at a composition of Al 16 .8 at% Cu 8.2 at% Si. Calculated using Pandat ( 107 ) Figure 6 5: Phase fractions of Al 4.1 at% Cu 2.0 at% Si at various temperatures. The heat treatment temperature of 530C reveals a 20% liquid composition during the healing process. The box signifies the potential area of liquid % should the temperature fluctuate by 5C during hea t treatment.

PAGE 106

106 Al Cu Si Matrix Properties The mechanical properties of the Al 4.1 at% Cu 2.0 at% Si matrix alloy were investigated via tensile testing. Appropriate amounts of Al (Al shot, 99.99%, Alfa Aesar), Cu (Cu shot, 99.9%, Alfa Aesar), and Sn (Sn s hot, 99.8%, Alfa Aesar) were melted in an open air furnace at 750 C until a liquid solution. The mold was heated up to 350 C before casting to prevent air pockets from forming in the as cast microstructure. The Al Cu Si melt was poured into the mold and al lowed to air cool. Each tensile bar was then placed horizontally into a furnace for 2 4 hour s at 530 C monitored by an external thermocouple, then air cooled to set the eutectic microstructure ( Figure 6 6 ). The results of testing are show in Table 6 1. Rep resentative curves of the tensile results can be found in Appendix A. Figure 6 6 : Representative microstructure of an Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) composite after a heat treatment at 530C for 24 hours

PAGE 107

107 Table 6 1 : Mechanical te sting results for Al Cu Si alloys Specimen Modulus (GPa) Ultimate Stress (MPa) Failure Strain (%) 1 66.0 99.0 0.24 2 84.1 109.3 0.19 3 78.6 129.9 0.30 Avg. 76.2 112.7 0.24 9.3 15.7 0.06 A commercial casting alloy of the similar composition was Al 213.0. This alloy possesses between 6 8 wt% Cu and 1 3 wt% Si. Literature shows a 0.2% offset yield strength in the as fabricated (F) state and T533 state of 186 MPa and UTS of 227 MPa and 213 MPa, respectively ( 148 ) It is thought the Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) investigate d in this study had decreased strengths caused by premature brittle failure from the greater eutectic percent, which is typically decreased to a minimum in traditional Al based alloy systems ( 149 ) Al Cu Si Composite Fabrication To investigate the self healing capabilities of the design Al Cu Si matrix, composites with NiTi SMA reinforcements were fabricated and mechanically tested. Appropriate amounts of Al (Al shot, 99.9 9%, Alfa Aesar), Cu (Cu shot, 99.9%, Alfa Aesar), and Sn (Sn shot, 99.8%, Alfa Aesar) were melted in an open air furnace at 750 C until a liquid solution. One or more NiTi SMA wire(s), designated BB 35 ( Ni 49.3 at% Ti, = 0. 87 mm Memry Corporation ) we re laid horizontally in the graphite tensile bar mold ( Figure 6 8 ). Because the heat of casting was found to induce shifting during the solidification process, the NiTi wires were heat treated for 24 hours at 530C to release residual stresses and create a martensitic state in the NiTi wires at room temperature; i.e. activating the shape memory effect (Figure 1 4). The mold was heated up to 350 C before casting to prevent cold shuts and air pockets from forming in the as

PAGE 108

108 cast microstructure. The Al Cu Si me lt was poured over the wire(s) in the mold and allowed to air cool. Each tensile bar was then placed horizontally into a furnace for 2 4 hour s at 530 C monitored by an external thermocouple, then air cooled to set the eutectic microstructure ( Figure 6 6 ). Figure 6 7 : Image of graphite tensile bar mold used for composite fabrication Al Cu Si Composite Properties After heat treatment, the Al Cu Si composite bars were ground to a 320 grit surface finish before being tested in tension using the Instron 5582 machine at a rate of 1.0%/min. Results of the testing are shown in Table 6 2 Because of the brittle nature of the composites, the 0.2% offset yield strength is not able to be reported. A tensile graph of Composite 3 is shown in Figure 6 8. The tensile da ta is compared to the tensile results for one of the Al Cu Si matrix alloys. T he small drops in stress in the composite data are attributed to debonding between the matrix and NiTi wires. Compared to the Al Cu Si matrix alloys in Table 6 1, the composite v alues are similar when using a t test with a 95% confidence interval. The apparent decrease in modulus for the composites is attributed to the lower elastic modulus found in the NiTi BB alloys. When heat treated

PAGE 109

109 for 24 hours at 530 C, the NiTi wires were f ound to possess an average E = 59.4 GPa. A rule of mixtures calculation can be performed: (Eq. 6 1) where E is elastic modulus, m is the matrix, f is the NiTi fibers, and X is the volume fract ion of the NiTi wires. Using the average value for the matrix (76.2 GPa) Al Cu Si specimens in Table 6 1 and the NiTi elastic modulus (59.4 GPa) the calculated average of the composite would be E = 75.75 GPa This decrease in modulus for the composite all oy from the matrix is consistent with the lower elastic modulus of the SMA reinforcements. Figure 6 8: Representative tensile results of an Al Cu Si composite with 2.9 volume percent NiTi SMA wire reinforcements compared to an Al Cu Si matrix alloy. Tabl e 6 2 : Mechanical testing results for Al Cu Si matrices reinforced with 2 3 vol% SMA wire reinforcements Specimen V f Wires (%) Modulus (GPa) Ultimate Stress (MPa) Failure Strain (%) 1 2.32 71.9 116.9 0.35 2 2.75 69.2 113.5 0.28 3 2.89 69.7 116.6 0.32 A vg. 2.65 70.3 115.7 0.32 0.30 1.4 1.9 0.04

PAGE 110

110 Additional Al Cu Si composites were fabricated in a similar manner, but with increased number of NiTi wires to increase the volume fraction of reinforcements, and thus increase the strength of the composite. A comparison of the tens ile properties of composites with increasing volume fraction of NiTi reinforcements is shown in Figure 6 9. It can be seen the increased volume fraction increases the yield strength and modulus, but there appears to be a limit to the increase. It was found that over 9 vol% NiTi in the cast composite sample introduced casting defects ( i.e pores) due to the increase in turbulence from extra wires during fabrication, thus the greatest volume fraction composites failed prematurely. Figure 6 9: Comparison of tensile results of an Al Cu Si matrix reinforced with increasing volume percent NiTi SMA wire reinforcements After initial composite testing, the fractured tensile bars were encapsulated under vacuum (in a Pyrex tube laying flat on graphite wrapped in Ta foil) and heat treated a second time for 24 hours at 530C to try and induce healing. Following air cooling, the Al Cu Si composites were tested in tension again for healing. However, no healing was detected above 32% as calculated by Equation 3 1; the ave rage healing was typically

PAGE 111

111 around 10%. Therefore, because of the casting issues with increased V f of NiTi wires both Al Si and Al Cu Si composites, future matrix development will require better fabrication techniques in order to increase the volume fractio n When comparing the Al Cu Si composite results to Al Cu and Al Si composites (Figure 6 10), Al Cu Si was found to show the greatest tensile strength of the three compositions. However, it was found the Al Cu Si composites were behaving in a brittle manne r much like the Al Cu composites. It is thought this brittle nature is not enabling forward transformation in the NiTi SMA wires, therefore there is not a clamping force during the second heat treatment to enable higher healing percentages. Figure 6 10: Comparison of tensile results of Al Cu Si, Al Cu, and Al Si matrices reinforced with 2.2 2.4 volume percent NiTi SMA wire reinforcements NiTi Wire Properties The properties of the NiTi wires (BB 35) were investigated to determine how the heat treatments fo r healing affect the output properties. Each wire was placed into a DSC and cycled from 50 200 C at 10 C /minute. The austenitic and martensitic

PAGE 112

112 transition temperatures were calculated via the offset method ( 114 ) The NiTi wires were encapsulated under vacuum in Pyrex tubes and heat treated either once or twice at 530 C for 24 hours to simulate the heat treatment of the Al Cu Si composites then placed in the DSC. Results of the testing compared to the previously determined as received properties (Table 5 8) are found in Table 6 3. As with the 592 C heat treatments, the 530 C heat treatment was found to significantly increase all of the NiTi transition tempera tures from the as received to initial heat treatment. However, a secondary heat treatment did not significantly change the transitions. Table 6 3: Thermal Properties of NiTi Wires Subjected to Various Heat Treatments NiTi Wire Condition A s A f M s M f As Rec eived 4.5 13.2 8.8 10.0 530C for 24 hrs 61.9 72.6 39.7 33.8 530C for 24 hrs (2x) 61.4 74.6 40.4 31.8 Next, NiTi wires under similar conditions as above were tested in tension to investigate thermal effects on mechanical properties. The samples were pulled on the Instron mechanical testing machine to failure ( Table 6 4). Representative curves showing the stress strain behavior can be found in Appendix A. As with the thermal properties, the mechanical properties significantly changed with heat treatme nt, decreasing the plateau stress nearly 75% and the failure stress by over 20%. However, subsequent heat treatment had little effect on the mechanical properties. Table 6 4: Tensile Properties of NiTi Wires Subjected to Various Heat Treatments Average NiTi Wire Condition Plateau Stress (MPa) Failure Stress (MPa) As Received 492 19 1368 13 530C for 24 hrs 131 5 1088 30 530C for 24 hrs (2x) 130 3 1097 1

PAGE 113

113 Strength Model Verification Looking at the design aspects found in Chapter 4, o ne of the major design goals was increasing the specific strength within the matrix alloy. Thus, verification of the strength models used in development of the ternary matrix was conducted. Solid Solution Strengthening Model Using a rule of mixtures method the strength of the Al Cu Si matrix alloy can be predicted by inserting the strength of phases present into the following equation: (Eq. 6 2 ) w here m is the strength of the matrix alloy, (Al) is the strength of the Al based solid solution eutectic strength of the eutectic phase, and X is the phase fraction of the eutectic phase As the eutectic phase of Al Cu Si alloys is extremely brittle ( 150 ) compression testing performed. First, the matrix alloy Al 4.1 Cu 2.0 Si (at%) was cast in a similar manner to the a bove composites, heat treated for 24 hours at 530 C and air cooled. A ternary eutectic composition (as calculated by Pandat) of Al 15.0 Cu 7.3 Si (at%) and a solid solution alloy of Al 1.85 Cu 0.74 Si (at%) were also cast; the solid solution alloy was calculated to have the highest alloying content at the heat treatment temperature. The solid solution alloy was heat treated at 530 C for 24 hours whereas the eutectic alloy was tested in the as cast condition. Each bar of the three compositions was l athed, cut, and polished to a 600 grit finish to fabricate cylindrical samples for compression testing on the Instron machine at 0.5 mm/min. A comparison of the results is found in Table 6 5.

PAGE 114

114 Table 6 5: Compression Testing of Different Al Cu Si Alloys Av e Composition (at%) 0.2% Compressive Yield Stress (MPa) Compressive Ultimate Stress (MPa) Al 1.85Cu 0.74Si (Solid Solution) 203 18 494 15 Al 15.0Cu 7.3Si (Eutectic) 365 6 536 26 Al 4.1Cu 2.0Si (Alloy) 267 4 424 9 I nserting in the average values into the rule of mixtures model ( Eq uation 6 2 ), the predicted matrix alloy yield strength with 20% eutectic phase is only 235 MPa. Compared to the compression data for the Al 4.1Cu 2.0Si alloy, this shows the model under pred icts the total matrix alloy compressive yield strength by 12%. The model, however, completely breaks down for the ultimate strength, as the matrix alloy was found to possess a lower compressive ultimate strength than both the solid solution and eutectic co mpositions. The predicted compressive ultimate stress is 502 MPa, which is over 80 MPa greater than measured. Therefore, it is posited the yield strength can be generally predicted by the rule of mixtures equation, but not the ultimate strength in this all oy. Another way to predict the solid solution strength in the matrix is through a strengthening model described by Nembach ( 125 ) ( Equation 4 3 ). This model relates how the additive strength of phases relat es to the alloy strength through constitutive phase strengthening. Knowing the strength of pure Al (99.9999%) is 12 MPa ( 55 ) the increase in solid solution strength from the 1.85 at % Cu and 0.74 at% Si additions maximum solubility of the alloys at the 530 C healing temperature must be calculated. Utilizing the solid solution strengthening in Al data for Cu and Si ( Figure 4 3), it can be seen the strengthening must be predicted using different models; Si by a linear

PAGE 115

115 fit and Cu usin g the classical Fleischer model ( 151 ) (Figure 6 11). The linear model can be found in Equation 6 2, whereas the Fleisher model is shown in Equation 6 3: (Eq. 6 3 ) (Eq. 6 4 ) w her e y is the change in yield strength from solute addition and the constants A and B are variable based on data fit. It was found A = 13.88 and B = 20, respectively, and therefore the increase in solid solution strength from 1.85 at% Cu and 0.75 at% Si in Al was found to be 37.5 MPa. This results in the predicted yield strength of between 31.5 MPa for k = 2 and 49.5 MPa for k = 1 at the solid solution composition. Further increases in yield strength can be associated with strengthening from the eutectic phase in the microstructure. Figure 6 11: Solid solution strengthening in Al for Cu and Si. The blue line shows a linear fit for the Al Si data, whereas the red line shows a Fleisher model for the Al Cu data. Graph plotted with tabulated data found in ( 127 130 131 )

PAGE 116

116 As the eutectic tensile yield strength is difficult to determine due to the brittle nature of the material, it is possible to estimate its contribution from the compressive data. Looking at Table 6 5, the compressive yield strength of the eutectic phase is 79.3% greater than the yield strength of the solid solution alloy in Al Cu Si. Therefore, it is postulated that the tensile yield strength of the eutectic phase will also be 79% greater than the tensile yield strength of the solid solution phase. Previously, the estimated solid solution strength was found to be 37.5 MPa, therefore the estimated tensile yield strength from the eutectic phase would be 67.1 MPa for use in the model. Inserting the estimated results just obtained Al = 12 MPa, SS = 37.5 MPa, Eut = 67.1 MPa into the Nembach model (Equation 4 3) results in an estimated matrix strength between 77.8 MPa and 116.6 MPa for k = 2 and k = 1, for an average value of 97.2 MPa. Compared to the strength values obtained for the Al 4.1Cu 2.0Si (at%) found in Table 6 1 this predicted strength falls at the low end of the range found experimentally: 112.7 15.7 MPa Using the rule of mixtures model (Equation 6 2 ) and inserting these estimated values SS = 37.5 MPa, Eut = 67 .1 MPa results in a predicted strength of 104.6 MPa which also falls within the experimental range. Therefore, the prediction of the matrix strength properties via models for the design of a self healing MMC may be reasonable to use Composite Strengthen ing Model Composite strengthening in a longitudinal, continuous fiber composite can be estimated using a rule of mixtures model incorporating the individual contributions to strength of the different phases ( 112 ) The general equation for strength prediction: ( Eq. 6 5 )

PAGE 117

117 where c is the strength of the composite, m is the strength of the matrix alloy, f is the strength of the fiber, and X is the phase fraction of the fibers. Using the average values for the Al Cu Si matrix UTS = 112.7 MPa (Table 6 1 ), the plateau strength of t he heat treated NiTi fibers = 131 MPa (Table 6 4), and a fiber volume percent of 2.65%, the predicted composite strength is 113.0 MPa. Compared to the average Al Cu Si composite values in Table 6 2 this predicted value falls within 3 MPa of the measured c omposite strength of 115.7 MPa. Another predictive model for composite strength (Equation 4 4) combines the Nembach model with the rule of mixtures. Using the values of the estimated matrix strength between 77.8 MPa and 116.6 MPa for k = 2 and k = 1 an d the plateau strength of NiTi and volume percent previously stated, the composite strength in the range of 79.2 117.0 MPa. Comparing this to the previous results shows that the prediction of strength via the Nembach based model works best with a value n ear k = 1. Internal Stress in SMA Wires In her dissertation, Manuel ( 80 ) introduced a model to determine the minimum volume fraction of SMA wires required to induce a clamping force at the crack interface during a healing cycle. The model is based on the transformati on temperatures of the SMA wires and the compressive yield strength of the matrix at elevated temperatures. The Johnson Cook relationship ( 152 ) is shown in Equation 6 6 : ( Eq. 6 6 ) flow = normalized plastic strain rate, T room = room temperature, T melt = melting temperature of alloy, and A, B, C, n, and m

PAGE 118

118 represent material constants. Each set of brackets repres ents a subset of work hardening, strain rate, and thermal effects. Because only temperature effects on stress are desired, Equation 6 6 can be simplified to: (Eq. 6 7 ) Because of the nature of the alloys, the eutectic temperatu re (T eut ) of the alloy was utilized as T melt due to the softening which occurs in a partially liquefied alloy. Thus, the modified Johnson Cook relationship can be represented by: ( Eq. 6 8 ) In an equilibrium conditio n, there are equal applied forces (F) when the matrix is being clamped back together by the SMA wires at elevated temperatures (Figure 3 8d). ) for the matrix and SMA wires: ( Eq. 6 9 ) If this relationship is divided by the cross sectional area, the following relationship is obtained: ( Eq. 6 10 ) where V f = volume fraction of SMA wires. During healing, the stress in the SMA wires R ) of the SMA wires and the stress in the matrix MCYS ) at elevated temperatures under compression. Thus, rearranging Equation 6 10 will yield:

PAGE 119

119 (Eq. 6 1 1 ) Equation 6 1 1 shows that the reversion stress of the SMA relates to the compressive yield strength of the alloy through an indirect relationship with the volume fraction of the SMA wires in the composite. Knowing that the flow stress of an alloy is equal to its yield strength while flow MCYS ), the model for optimal volume fraction of SMA wires in the composite is found by inputting the modified Johnson Cook model (Equation 6 8 ) into Equation 6 1 1 : (Eq. 6 1 2 ) This model describes the relationship to obtain the optimal volume fraction of SMA wires for a specific matrix alloy. To verify this model, both Al Cu Si and Al Si alloys were investigated to compare results. Compression of Matrix Alloys To obtain the elevated temperature com pressive yield stress of the alloys, Al 4.5Cu 2Si and Al 3Si (at%) bars were cast and heat treated at 530C for 24 hours and 5 92 C for 24 hours, respectively. Pieces were machined according to ASTM E209 00 : Standard Practice for Compression Tests of Metall ic Materials at Elevated Temperatures with Conventional or Rapid Heating Rates and Strain Rates ( 153 ) The samples were tested at a strain rate of 10 3 s 1 on an MTS 810 Material Testing Machine with an argon atmosphere furnace attachment. Each sample was tested at various temperatures to determine the compressive yield strength. The raw data for each of the alloys is found in Table 6 6.

PAGE 120

120 Table 6 6: 0. 2% Compressive Yield Strength at Elevated Temperatures of Al 4.1Cu 2Si and Al 3 Si (at%) Alloys 0.2% Yield Strength (MPa) Temperature (C) Al 4.1Cu 2Si Al 3Si 25 269 65 150 157 -200 -37 300 135 -400 21 9.5 500 10 -530 2 -550 -4 The data was plotted on a true stress vs. temperature graph and fitted to the modifi ed Johnson Cook model (Equation 6 8 ) with the parameters found in Table 6 7 used to determine fit (Figure 6 12) Figure 6 12: Elevated temperature compressive yield strength for Al 4.1Cu 2Si and Al 3Si (at%) with an overlaid modified Johnson Cook model f or fit. Table 6 7: Parameters Used to Model Elevated Compression Testing in Al 4.1Cu 2Si and Al 3Si (at%) Alloys Parameter Al 4.1Cu 2Si Al 3Si A 270 65 m 0.7 0.6 T room ( C) 25 25 T heal ( C) 592 530

PAGE 121

121 Transformation Temperatures in NiTi Wires To determin e how stress increases would shift the transformation temperatures in the NiTi SMA wires, BB wires were loaded on the Instron 5582 in tension at specific stress values The wires were previously heat treated at 530 C for 24 hours to simulate healing temper atures in Al Cu Si composites. The wire loaded at the specific stress was then held at the displacement this stress imparted for the remaining portion of the test. The temperature was increased on the wires under tension and the temperature was recorded wh ere transformation in the SMA wires was observed by an increase in stress. The results of the testing are shown in Figure 6 13. The black dots represent the A s and A f temperatures as determined previously from the DSC. The linear fit lines were found to in crease at a rate of 5.5 MPa/ C This aligns with previous research which has shown typical rates are from 2.5 15 MPa/ C ( 113 ) Figure 6 13: Change in A s and A f for NiTi wires heat treated at 530 C for 24 hours. The black dots are the A s and A f temperatures determined by DSC. The linear fit lines were found to increase at 5.5 MPa/ C

PAGE 122

122 Next elevated temperature testing of the SMA wires was conducted to determine the reorientation/yield temperatures. Olson and Cohen ( 154 ) described a stress temperature dependence for martensitic nucleation. Below a specific temperature and stress state called M s initial yielding of the wire occurs through a stress assisted transformation from twinned martensite to detwinned martensite; i.e. reorientation of the martensitic phase. However, above M s the strain induced nucleation of th e parent phase occurs through a slip; i.e. yielding of austenite. Using BB NiTi SMA wires which were heat treated at 530 C for 24 hours, elevated temperature tensile testing was conducted on the Instron 5582. Results of the testing can be found in Table 6 8. Overlaying this data on the previously determined A s and A f temperatures for similarly heated wires results in Figure 6 14. Figure 6 14: Schematic of the stress temperature relationships for NiTi BB SMA wires.

PAGE 123

123 Table 6 8: Elevated Temperature Tensil e Testing of NiTi BB Wires Heat Treated at 530 C for 24 hours Temperature (C) 0.2% Offset Stress (MPa) 25 85.6 50 108.1 100 166.8 150 166.8 200 140.4 Optimization of NiTi Wire Fraction in Al Based Composites Using the previously gathered NiTi stres s temperature data and overlaying the calculated stress in NiTi SMA wires within an Al 4.1Cu 2.0Si (at%) composite using Equation 6 12 results in the data shown in Figure 6 15. It was noted that the minimum NiTi fiber volume percent required was determined to be 55%, meaning the composite would possess more SMA wires than matrix alloy to be effective Figure 6 15: Schematic of the stress temperature relationships for NiTi BB SMA wires heat treated at 530 C for 24 hours with the stress in SMA wires found in an Al 4.1Cu 2.0Si (at%) composite with V f = 55% of NiTi wires Similar data was collected for SMA wires incorporated in an Al 3 at% Si composite (Figure 6 16). Here, the volume % of SMA wires was calculated to be 25%.

PAGE 124

124 However, as healing was found in Al Si composites at less than 2.5 vol% SMA wires, the Manuel model was found to be inadequate at determining minimum SMA volume fractions in Al based composites. This can be attributed to the assumptions in the model including equal strain and no plasticity in the composite. For composite models exhibiting equal strain, the ends of the reinforcements are fixed to the ends of the matrix and elongation results in equal length changes for both matrix and reinforcement, without sharing an equal stress. However, b ecause of the interfacial bonding, load sharing will occur, which would decrease the stress found in the NiTi wires. Also, plasticity was found to not be negligible in the Al Si composite with failure strains of nearly 4% on average. The increased toughnes s in the composite would also decrease the required NiTi volume fraction. Therefore, further investigations into more refined models will be required to optimize the volume fraction of SMA wires in future self healing MMCs. Figure 6 16: Schematic of the stress temperature relationships for NiTi BB SMA wires heat treated at 592 C for 24 hours with the stress in SMA wires found in an Al 3.0Si (at%) composite with V f = 2 5% of NiTi wires

PAGE 125

125 Novel Processing Techniques The healing of the Al Cu Si composites did not match the high healing percent results found in Sn Bi or Al Si composites. It is thought the lack of transformation of the SMA wires because of the lack of ductility in the matrix alloys prevented a clamping force from occurring during healing. This c lamping force has been shown to be required for healing (Chapter 3). Thus, different fabrication techniques were investigated to determine if healing of Al Cu Si composites could be improved. Pre strain of SMA Wires Previous research into SMA reinforcement for crack closure in composites had shown the need for pre straining the SMA wires to ensure a clamping force following a crack event ( 81 100 ) Therefore, an investigation into pre straining the NiTi SMA wires prior to casting was conducted. First, a NiTi BB wire was heat treated at 530 C for 24 hours to relieve residual stresses from fabrication and to move A s transformation temperatures above room temperature. This meant the wires were in the martensitic state at room temperature, and thus the SMA wires would retain any deformation until they were heated above their A s temperature. The heat treated NiTi wires were pulled in tension to 5% strain using the Instron 5582. Each wire was then place into a custom jig designed to hold the wire fast during the high temperature casting and not allow for transformation. Using simil ar casting techniques to those utilized in fabricating previous Al 4.1Cu 2.0Si (at%) composites, castings were poured with the pre strained SMA wires. Samples were tensile tested, and healed as previous, but little healing was noted. It was observed that t he elongated length of the SMA wires in the pre strained state was not retained following initial casting, so it is thought the SMA wires returned to their austenitic state during casting

PAGE 126

126 despite the jig to prevent this. Therefore, pre staining the wires v ia different means was continued. Hot Compression for Composite Fabrication Wrought products are known to have higher mechanical properties than traditional cast products because the deformation processes reduce defects. Therefore, a mechanical means of im proving ductility in the Al Cu Si composites was investigated in a joint project with another member of the Manuel group, Glenn Bean. First, an alloy of Al 4.1Cu 2.0Cu (at%) was cast into a bar and heat treated at 530 C for 24 hours to set the microstructu re. The bar was cut into thin plates (thickness ~ 1 mm) via electric Thin NiTi SMA wires (BH), were prepared similar to those for the Sn Bi composites: heat treating at 500 C for 3 hours. Two thin plates were then pressed around 3 aligned BH wires at 500 C at a strain rate of 10 4 s 1 to creep the matrix around the wires and pressed further at 530 C at different stresses for varying times. The pressing was completed in an argon atmosphere on t he MTS 810 testing machine. A design of experiments (DOE) was completed to determine if time or pressure was most important during the hot pressing fabrication ( Table 6 9). Table 6 9: Parameters of DOE for Hot Pressing Fabrication of Al Cu Si Composites Co ndition Force (MPa) Time (hr) Run Order A 2 4 1 B 2 8 3 C 4 4 2 D 4 8 4 Upon completion of fabrication, each bar was machined to a 320 grit finish and pulled in tension on the Instron 5582 at a rate of 1.0%/minute. Each composite was found to have a SMA vol% between 1.0 1.5%. Results of the virgin tensile testing can

PAGE 127

127 be found in Table 6 10. It was noted the highest yield strength was achieved in a composite pressed at 2 MPa for 8 hours, but the highest UTS and strain to failure was found in a specimen pressed at 4 MPa for 4 hours. Table 6 10: Parameters of DOE for Hot Pressing Fabrication of Al Cu Si Composites Condition V f Wires (%) 0.2% Yield Strength (MPa) Ultimate Strength (MPa) Failure Strain (%) Diffusion Layer A 1.40 89.0 89.5 0 .39 2.92 B 1.37 117.7 119.8 0.45 8.54 C 1.28 110.5 133.3 0.66 3.59 D 1.55 N/A 102.5 0.33 8.14 Further m icrostructural inspection of the cross section revealed a diffusion layer at the SMA/matrix interface. Using energy dispersive spectroscopy (EDS) an alysis, the diffusion was found to be mostly Al and Si into the NiTi wires (Figure 6 17 ). Results of the total thickness of the diffusion layer into the NiTi wires can be found in Table 6 10. The diffusion layer was found to be strongly dependent upon time but did not very with changes in stress. Images of the post hot pressing microstructure can be found in Figure 6 1 8 for specimens with the highest mechanical properties (condition B and C). Note how the grains have grown over the interface between the o riginal two matrix bars. After tensile testing, each bar was healed at 530 C for 24 hours under vacuum, but found to heal less than 15% in terms of strength retention (Equation 3.1). Another specimen was prepared with multiple plies of SMA wires (4 layers of matrix, 3 plies of wires) was also prepared in a similar manner to Condition B, which exhibited the highest yield strength. The multiple plies resulted in an increase to 1.67 vol% NiTi wires. Tensile testing was completed and a significant increase in s train to

PAGE 128

128 Figure 6 1 7 : EDS analysis of a representative m icrostructure of a hot pressed Al Cu Si composites : a) SEM image of the NiTi wire in the matrix, b) SEM image of the area in the red box in a) with the elemental analysis overlaid, and c) elemental analysis showing diffusion of Al and Si into NiTi SMA wire Cu was not found to diffuse into the NiTi. Figure 6 1 8 : Microstructure of Al Cu Si composites a) pressed at 4 MPa for 4 hours and b) pressed at 4 MPa for 8 hours.

PAGE 129

129 failure was noted ( Figure 6 19 ). Over the 1 ply hot pressed sample, the strain to failure was increased over 140%, but against a traditionally cast Al Cu Si composite (specimen 1), the strain to failure was increased over 210%. When a healing heat treatment was conducted vacuum encap sulated in Pyrex on a graphite sheet covered in Ta foil and heated at 530 C for 24 hours the 3 ply sample was found to exhibit over 37% healing. This leads to the conclusion that deformation based fabrication techniques may help increase healing in britt le matrix alloys. Figure 6 19 : Comparison of tensile behavior of a cast Al Cu Si composite (V f = 2.32%), a hot pressed Al Cu Si composite with 1 ply (V f = 1.37%), and a hot pressed Al Cu Si composite with 3 plies (V f = 1.67%). The hot pressed 3 ply compo site exhibited an increase of strain to failure of over 140% on the 1 ply specimen and over 210% on the cast specimen. Summary In an effort to produce a stronger matrix with self healing potential, an Al 4.1Cu 2.0Si (at%) composite was designed and fabrica ted. Tensile behavior showed a stronger composite, but lacked the strain to failure required to induce a pre strain in the

PAGE 130

130 NiTi reinforcements needed to cause a clamping force during a healing cycle. Increases in the vol% of SMA wires yielded great er modul i and strengths, but also lacked ductility. Various strength models used in the design methodology were verified, but a model used to optimize the vol% SMA wires was found to lack robustness for Al based matrices. Different fabrication techniques were inve stigated to induce higher failure strains and hot pressing while at the healing temperature was found to induce bonding across an interface while also increasing the strain to failure. Increasing the vol% through multiple plies resulted in the largest fail ure strains while also bringing healing of the composite up to over 37% retained strength.

PAGE 131

131 CHAPTER 7 REACTIVE ELEMENT ADDITIONS Bonding characteristics between liquid phases and solid surfaces is vital to the understanding healing in metal matrix compos ites. Surface oxides, however, have been found to decrease the bond strength at the solid/liquid interface. A n in situ fluxing technique inspired by lead free solder research has shown promise to improve healing capabilities in atmosphere. The method requi res a reactive element alloying addition to the matrix composition which possesses a lower free energy of oxide formation than the parent element, thus creating a strong chemical bond across the interface as the reactive element addition reduced the parent element oxide. This chapter will show how the reactive element addition maintains fracture toughness values measured using the chevron notch short bar technique after a bonding heat treatment by increasing the bond strength along the interface and deflec ting the crack into the matrix. Next, the methodology will be detailed for addition of reactive elements to the matrix alloys used in liquid assisted self healing in MMCs. Interfacial Toughness Traditional techniques for joining such as welding, soldering, and brazing utilize a liquid phase to bond two or more metallic components. Advanced processing methods, such as liquid phase sintering (LPS) ( 155 156 ) and transient liquid phase bonding (TLP) ( 157 158 ) also inherently use a liquid phase to enable joining of components. Often the substrate will possess a native oxide surface which hinder s the production of a strong bond. In order to increase the propensity for bonding between the liquid and metal surface, and thus the quality of the bond, fluxing agents are frequently used to

PAGE 132

132 breakdown or chemically reduce the surface oxide layer. This fluxing increases the availability of bondin g sites for the liquid to react with the underlying metal as well as increasing the wettability of the surface ( 159 160 ) To simplify this process, in situ fluxing techn iques have been developed utilizing self fluxing alloys which possess alloying additions to increase wettability and depress the melting temperature ( 161 ) Joining in the presence of a reducing environment has also been found to increase the wettability and bond efficiency ( 159 162 ) Investigations into lead free solders utilizing rare earth element additions to increase bond strength between metals and oxides have yielded positive results ( 163 165 ) It has been proposed that the reason for the increase in bond strength can be attributed to a strong chemical bond created at the solid/liquid interface through reduction of the metal oxide by the rare earth ele ment addition, which possess a highly negative free energy of formation ( 164 166 167 ) To expand on this research for self healing metal matrix composites an investigation into the thermodynamic principles of the bonding characteristics was undertaken. To study the effect of reactive elemental species durin g bonding, interfacial plane strain fracture toughness testing via the chevron notch short bar (CNSB) technique was utilized. The chevron notch technique has been found to yield reproducible interfacial fracture toughness results ( 168 170 ) and will therefore be utilized to obtain a quantitative measure of crack growth resistance at metal/oxide interfaces. Additionally, the short bar technique works well for alloy development because the specimen size is relatively small (10mm x 10mm x 14.5mm), fatigue pre cracking was not necessary due to the stress concentration and geometry to initiate a

PAGE 133

133 sharp crack at the tip of the chevron shaped notch ( 171 ) Furthermore, knowledge of the crack size is not required to determine the fracture toughness (K Ic ) because the critical point is constant for a specified geometry ( 168 ) A schematic of a representative CNSB specimen is found in Figure 7 1. Figure 7 1 : Schematic of a chevron notch short bar (CNSB) specimen. Adapted from ASTM Standard E 1304 97. 2008. Standard Test Method for Plane Strain (Chevron Notch) Fracture Toughness of Metallic Materials (Page 925, Figure 3) ASTM International, West Conshohocken, PA. Matrix Design To study how reactive elements could improve healing characteristics, p hase diagrams of two alloys, one containing a reactive species (relative to the primary element) and the other containing a no n reactive elem ent addition were investigated These prototype alloys must possess two characteristics. First, each system must exhibit a low temperature eutectic phase transformation akin to the previous Al and Sn based systems investigated for self heal ing Second, each system exhibited limited solubility of the reactive and non reactive species, translating to the additive element strongly partitioning to, and thus residing in, the liquid phase during heat treatment.

PAGE 134

134 The selected binary alloys meeting the above criteria were antimony copper (Sb Cu) and antimony zinc (Sb Zn). For this study, Sb will serve as the base element with Cu the thermodynamically less reactive and Zn the more reactive addition base d on free energies of formation as shown on the E llingham diagram ( Figure 7 2). The lower (more negative free energy) the element falls on the diagram, the greater the driving force at a given temperature for the oxide to oxidize any other element of higher energy in its presence Figure 7 2: Ellingha m diagram showing relative position of free energy for oxide formation for Cu, Sb, and Zn. Curves plotted with free energy data obtained from ( 110 ) The specific alloy composition utilized for testing was determined by plotting the percent liquid and percent solute in liquid versus temperature ( 172 173 ) For both systems, alloys containing 4 at% solute were chosen to ensure both alloys had a consistent liquid percent and percent solute in liquid during heat treatment. At 4 at% solute, both Sb Cu and Sb Zn alloys crossed at 20% ( Figure 7 3); this translates to heat

PAGE 135

135 treatment temperatures of 555C for Sb Zn and 574C for Sb Cu. These temperatures correspond to the solid + l iquid phase region ( Figure 7 4) for both of the respective binary alloys An alloy of the eutectic composition, Sb 37 at% Cu and Sb 33 at% Zn, was also selected for testing to quantify the fracture toughness of the solidified liquid phase for comparati ve purposes. Figure 7 3: Plot of percent liquid and percent solute in liquid for Sb Cu and Sb Zn alloys. At 20% liquid, heat treatment temperatures are 555C for Sb Zn and 574C for Sb Cu Data for Sb Zn alloys adapted from Sb Zn (Antimony Zinc) [databa se on the Internet]. ASM International. 1966 [cited October 2011]. Available from: http://products.asminternational.org/hbk/index.jsp Data for Sb Cu alloys adapted from Cu Sb (Copper Anti mony) [database on the Internet]. ASM International. 1990 [cited October 2011]. Available from: http://products.asminternational.org/hbk/index.jsp Sample Fabrication The 4 at% solute allo ys and eutectic alloys were prepared by placing appropriate amounts of Sb ( Sb shot, 99.999%, Alfa Aesar ) and Cu ( Cu shot, 99.9 %, Alfa Aesar ) or Zn ( Zn shot, 99.99% Alfa Aesar ) in graphite crucibles coated with boron nitride and heating at 750C in an argo n atmosphere until a liquid solution formed. Each alloy was

PAGE 136

136 Figure 7 4: Phase diagrams of a) Sb Cu and b) Sb Zn below 40 at% solute showing the healing temperature for each alloy. Data for Sb Zn alloys adapted from Sb Zn (Antimony Zinc) [database on the Internet]. ASM International. 1966 [cited October 2011]. Available from: http://products.asminternational.org/hbk/index.jsp Data for Sb Cu alloys adapted from Cu Sb (Copper Antimony) [da tabase on the Internet]. ASM International. 1990 [cited October 2011]. Available from: http://products.asminternational.org/hbk/index.jsp cast into a coated graphite bar mold and allowed to air cool. The alloys at the eutectic composition were heat treated at 400C in air for 24 hours to homogenize before machining. The Sb 4Zn and Sb 4Cu alloys were heat treated at 555C and 574C, respec tively, in air for 24 hours before air cooling All specimens were machined to approximately 10mm x 10mm x 14mm and polished to a 320 grit surface finish to reduce the possibility of surface cracks skewing results. One set of the Sb 4Zn and Sb 4Cu alloys post initial heat treatment, were cut in half and p olished to a 4000 grit finish on the inner surface. The finely polished surfaces of each half were then held together while the specimen was placed into a steel holder covered with carbon paper to prevent any chemical reaction between the steel and the sp ecimen during heat treatment. This holder was used to clamp to two halves together during heat treatment in an open air furnace for bonding ( Figure 7 5). These specimens underwent a second, healing heat treatment at identical temperatures as previous in

PAGE 137

137 Figure 7 5: Steel holder used to clamp Sb alloys during a healing heat treatment. order to compare heal ed and monolithic specimens with the same microstructure. After heat treatment, a chevron notch was machined into all Sb Zn and Sb Cu specimens for int erfacial fracture toughness testing as per ASTM Standard E 1304 ( 171 ) For the heal ed specimens, the notch was loca ted along the bonded interface. Mechanical Testing Interfacial fracture toughne ss testing was conducted on a T erraTek Model 4400A Fractometer. The data collected is summarized in Table 7 1. A t Method (assumes two populations with normal distributions but standard deviations probably unequal) with a 95% confidence in terval was run and found there was no statistical significance in the fracture toughness between the monolithic and heal ed specimens of either Sb 4Cu or Sb 4Zn. The Sb Cu eutectic alloys were found to possess greater fracture toughness values than eutectic Sb Zn. Table 7 1: Summary of Fracture Toughness Testing of Sb Cu and Sb Zn Alloys F Alloy Monolithic Healed Eutectic Sb Cu 2.08 0.22 1.59 0.39 1.93 0.22 Sb Zn 0.94 0.23 1.10 0.12 0.50 0.22

PAGE 138

138 Representative micrographs of the Sb 4Cu and Sb 4Zn alloys are found in Figure 7 6 at the fractu re interface. Both the monolithic and healed alloys exhibit coarse microstructures; on average, the Sb grain boundaries of the Sb 4Cu and Sb 4Zn monolithic and healed specimens was determined to be eutectic in nature. As seen in Figure 7.6, the eutectic phases can be observed along the interface in both the Sb Cu and Sb Zn bond specimens. Energy dispersive spectroscopy (EDS) analysis revealed the presence of ZnO at the bond interface as shown in Figure 7 7. C opper oxide was not found along the Sb Cu bonded interface; instead a Sb 2 O 3 phase was found on all of the specimen surfaces Figure 7 6: Representative micrographs of a) Sb 4Cu and b) Sb 4Zn (at%) at the interface. Healing was not evident in Sb 4Cu, but was evident in the Sb 4Zn alloy as seen by the crack line following eutectic lines in the bulk alloy and not at the healed interface. Representative images of the fracture surfa ces for both Sb 4Cu and Sb 4Zn healed alloys are shown in Figure 7 8. The dashed lines show the location of the chevron notch, whereas the solid line represents the end of the bonded area over which the fracture occurred. For the Sb 4Cu alloys, the remain ing, non bonded, area is

PAGE 139

139 Figure 7 7: Image along Sb 4 at% Zn healed interface showing presence of ZnO Figure 7 8: Fracture surfaces of a) Sb 4Cu and b) Sb 4Zn (at%) healed samples. Notice the greater fracture area for Sb Zn compared to Sb 4Cu. coat ed in a layer of Sb 2 O 3 (as confirmed through EDS analysis) from the oxidation of Sb during heat treatment. It was noted that the Sb Zn specimens exhibited more bond area than Sb Cu specimens as evidenced by the fracture surface area (Table 7 2). Table 7 2: Measured bond area of post healing Sb 4Cu or 4Zn CNSB specimens Alloy Bond Area (%) Sb 4 Cu 39 19 Sb 4 Zn 89 7

PAGE 140

140 Analysis Thermodynamic Driving Force During the heali ng heat treatment for the Sb 4Zn and Sb 4Cu alloys, the eutectic constit SbCu) liquefied as the temperature was raised above the eutectic temperature. Since Sb is known to oxidize readily at elevated temperatures, especially above 500C ( 55 ) the newly liquefied eutectic phases in the Sb Cu and Sb Zn alloys would immediately begin to form Sb 2 O 3 on the exposed surfaces, includ ing the cut and polished interface, in the following manner ( 110 ) : 4/3Sb (s) + O 2 (g) 2/3Sb 2 O 3 (s) [ = 332.1 kJ ] ( Eq. 7 1) Excess Sb 2 O 3 was also discovered in small spheres formed on the outside of each specimen ( Figure 7 5). These features were found on all of the bonded specimens after heat treatment. The presence of this surface oxide was noted on every Sb alloy and v erified through EDS analysis. Sb Zn Specimens At the heat treatment temperature of 555C, the reactive alloy heal specimen of Sb 4Zn has 20 at% Zn which has partitioned to the l iquid and is available to react with oxygen. The ZnO, possessing a lower Gibbs free energy of formation than Sb 2 O 3 is expected to reduce the Sb 2 O 3 found along the interface ( 110 ) : Sb 2 O 3 (s) + 3Zn (s) 3ZnO (s) + 2Sb (s) [ = 317.6 kJ ] ( Eq. 7 2) This formation of ZnO creates a chemical bond wel ding the two halves of the specimen together during the heat treatment. Although it was anticipated that the Sb 4Zn alloys would display an increase in interfacial fracture toughness over the monolithic value due to its increased propensity of forming a s trong bond the interface, it

PAGE 141

141 was noted that the fracture toughness did not deviate significantly after bonding. Close inspection of the crack path in Figure 7 6b shows that the crack progressed through the eutectic in the bulk of the specimen as opposed t o the strongly bonded interface. The low fracture toughness of the eutectic facilitated this failure mechanism as it Sb Cu Specimens The Sb 4Cu bond specimens have 20 at% Cu available to oxidize at their elevated bonding te mperature of 574C. However, since Cu 2 O possesses a greater Gibbs free energy of formation than Sb 2 O 3 it is not thermodynamically favorable for Cu to reduce the Sb 2 O 3 formed along the interface during heat treatment, as indicated by the positive free ener gy of formation shown below ( 110 ) : Sb 2 O 3 (s) + 6Cu (s) 3Cu 2 O (s ) + 2Sb (s) [ = 154.9 kJ ] (Eq. 7 3) There was no evidence of Cu 2 O bonding the halves together as there was ZnO in the Sb 4Zn alloys. The Sb 2 O 3 in Sb Cu bonded specimen passivated the crack surface preventing efficient bonding by the liquefied eutectic constituent. This is shown by the reduction of bonded surface area at the interface as shown in Figure 7 8. It is thought the Sb 2 O 3 would have started forming at the exterior edges of the interface first and continued inward to the cent er of the specimen. Therefore, the bonded area of the interface would have only been in the center of each specimen where the Sb 2 O 3 was not able to form. As the chevron notch was cut after the heat treatment was completed, the bonded area was near the tip of the chevron notch, locating the crack front at the center of the specimen. This unexpected processing condition would explain why the

PAGE 142

142 fracture area for the Sb Cu alloys was only at the interior of each of the chevron notch specimens. In comparing the h eal ed to monolithic specimens, it was still found that the facture toughness values in the Sb 4Cu alloys were similar. This is most likely caused by inflation of the fracture toughness values due to the influence of testing geometry. In the chevron notch t est, Newman ( 170 ) details how the fracture toughness value for an alloy can be determined when the crack length, a, is between a 0 and a 1 ( Figure 7 9). Using finite element analysis, it has been determined the crack length must reach a certain length or the fracture toughness results will be artificially inflated ( 174 ) This has been confirmed with experiment ( 170 ) Therefore, if the crack, a, did not reach its critical length, a c because the specimen was not bonded in that area, then a < a c resulting in an inflated fracture toughness value, K I > K IC This artificial inflation is th e most likely cause for the similar fracture toughness values for the Sb Cu bonded and monolithic alloys. The increase in bond area for the Sb Zn alloys over the Sb Cu alloys is also attributed to an increase in wettability. The addition of a reactive ele ment to an alloy has been shown to increase liquid metal wettability on metal oxide surfaces in other studies of metal ceramic interfaces ( 175 178 ) This increa se in wettability and bonding was similar to results found by several groups working on lead free solders ( 163 164 166 167 ) These groups have shown small additions of rare earth elements when added to lead free solders, in addition to increasing wettabili ty, also enables bonding to several classes of materials including steels, oxides and carbides.

PAGE 143

143 Figure 7 9: Schematic showing chevron notch crack front in relation to important mathematical factors. Adapted from Newman, J.C. 1984. A Review of Chevron No tched Fracture Specimens (Page 8, Figure 2). In: Underwood JH, Freiman SW, Baratta FI, editors. Chevron Notched Specimens: Testing and Stress Analysis, ASTM STP 855. American Society for Testing and Materials, Philadelphia. Reactive Elements in Sn Bi Matri x To advance this research to self healing MMCs using the Sn Bi prototype, a joint investigation with a senior research student, W. Patterson Tuttle was initiated ( 115 ) Previous attempts to heal the Sn Bi alloys in air resulted in an oxide surface at the crack interface, which prevented healing from occurring. It is thought a reactive element addition will reduce this oxide surfa ce and enable healing in open air. Potential candidates for the reactive element addition in a Sn Bi composite must have two characteristics. First, the free energy of oxide formation for the addition must be lower than both SnO 2 and Bi 2 O 3 ensuring the ab ility to reduce down the surface oxide at the crack interface. Next, the element must not possess any solubility with Sn in order to ensure it would partition towards the eutectic phase which liquefies during a

PAGE 144

144 healing cycle. Of the 20 elements which match ed these criteria, two were selected for further study. Ce was selected because it possesses the lowest melting point of the rare earth elements and Li because its melting point is below that of both Sn and Bi. From the work completed by Ramirez and collea gues ( 164 166 ) it was found that less than 2 wt% of rare earth elements were added to lead free solders to increase bonding to various materials. They also found these small alloying add itions did not significantly alter the melting point of the solder alloy. Therefore, a composition of Sn 13.0 at% Bi 0.2 at% Ce or Li would be investigated for healing potential. Master alloys of Sn 4 at% Ce and Sn 20 at% Li alloys were prepared by pla cing appropriate amounts of Sn ( Sn shot, 99. 8 %, Alfa Aesar ) and Ce ( Ce chips, 99.9 % Alfa Aesar ) or Li (Li granuales, 99%, Alfa Aesar) in graphite crucibles coated with boron nitride. The alloys were heated at 600 C for Sn Ce or 400 C for Sn Li in an argon atmosphere until a liquid solution formed. Each alloy was cast into a coated graphite bar mold and allowed to air cool. After checking composition via ICP analysis, appropriate amounts of the master alloys were added to Sn and Bi ( Bi needles, 99.9 9%, Alfa Aesar ) to create the Sn Bi 0.2 at% Ce or Li alloys. Each alloy was heated at 350 C in an argon atmosphere, cast into a coated graphite bar mold, and allowed to air cool. Each bar was then vacuum encapsulated in Pyrex glass tubes and heat treated for 24 ho urs at 169C. Following air cooling, each bar was machined to approximately 2mm thick wafers. Each wafer was polished to a 320 grit surface finish before being loaded into the steel holder utilized for the Sb alloy healing (Figure 7 5). The Sn Bi Ce or Li specimens underwent a second heat treatment at 169C for 24 hours in an open air furnace to investigate healing. Following air cooling, the wafers were mounted and

PAGE 145

145 polished to a 0.03 finish to look at the microstructure at the interface (Figure 7 10). In t he image, healing was not evident in the Sn Bi Li sample as shown by the gap at the interface between the two wafers. However, the Sn Bi Ce showed signs of healing via grain growth across the interface line. Therefore, Sn Bi Ce was selected to move forward and to determine its failure stress akin to the experiments conducted on Sn Bi alloys in Chapter 3 More Sn 13.0 at% Bi 0.2 at% Ce bars were cast as previous. After an initial heat treatment of 169C for 24 hours, pieces were machined into rectangular bars and Figure 7 10: Representative micrographs of a) Sn Bi 0.2Li and b) Sn Bi 0.2Ce (at%) at the interface. Healing was not evident in Sn Bi Li, but was in the Sn Bi Ce specimen as seen by the grain growth across the interface. polished to a 320 gri t surface finish. A slice was cut through the center of each bar and these interfaces were polished to a 4000 grit surface to ensure smooth contact. The pieces were then loaded into the Kovar diffusion couple jig and had a stress of 30 MPa applied via torq ue wrench. The sample and jig were then vacuum encapsulated and heat treated for 24 hours at 169C to heal the sample. Upon removal, the samples were attempted to load onto the Instron 5582 to be pulled in tension, but broke in handling, signaling lack of healing. Larger bars were attempted to increase the cross sectional

PAGE 146

146 area for healing, but this had little effect. Two higher composition Ce additions, 0.5 and 1.0 at%, were also attempted but lead to similar results. The samples all had a similar oxide sur face similar to that found on the un successful Sn Bi composites attempted to be healed in air. Looking back at Figure 7 10b, the small black circles were investigated and determined to be Ce 2 O 3 via EDS analysis. It is believed that the Ce in the melt is s o reactive it is creating Ce 2 O 3 during initial casting even though melting was completed in an argon atmosphere. Therefore, because Ce 2 O 3 formed in the melt, there is not any free Ce metal available during healing to reduce SnO 2 and healing is hindered. Fu ture development of the addition of reactive elements to self healing MMCs will need to account for this processing issue. Summary The healing heat treatment applied to Sb 4Zn alloys resulted in retention of fracture toughness when compared to a monolithi c sample. This is attributed to the reduction of Sb2O3 by Zn to form ZnO bonds along the healed interface. Sb 4Cu was also found to retain fracture toughness, but this is thought to be an artifact of the testing procedure. Addition of reactive element to a Sn Bi matrix was attempted in order to improve healing of a Sn Bi composite in air. Of the designed candidates, Ce was found to have the greatest potential. Unfortunately, difficulty in casting resulted in the oxidation of Ce before a healing cycle, and t hus full healing was not realized. Future matrix designs with reactive element additions must take this into consideration during casting. However, it is not believed that additional reactive element addition to the matrix alloys needs to only be limited t o rare earth elements; alkaline earth metals also have potential to be used.

PAGE 147

147 CHAPTER 8 CONCLUSIONS Self healing materials have shown potential to shift from a Damage Prevention paradigm of traditional materials to one of Damage Management, where the materi al itself is able to heal structural defects. Applications of this technology would enable improved properties in aeronautical, space, and underwater applications. Self healing in metallic systems is based on either solid state or liquid assisted healing. Solid state healing is found in fatigue crack and creep mitigation in aluminum and ferrous based alloys and protective coatings. Liquid assisted healing is found in electrical applications and metal matrix composites. A thermodynamic design based systems a pproach will enable faster development of metallic systems for use as structural materials. This methodology uses a systems design chart to act as a roadmap to elucidate relationships across the processing structure properties performance paradigm. This sy stems design method was used to develop a prototype Sn Bi matrix reinforced with commercially available NiTi SMA wires. Healing was performed by heat treating the composite at a temperature which would cause 20% of the matrix alloy to form into a liquid, a cting as the healing material while the NiTi wires clamped the crack faces together. For the Sn Bi system, average healing of 80% retained strength post healing cycle with over 94% possible. The pressure required to enable the highest healing was found to be 30 MPa at the crack interface. As the Sn Bi composites were heavy and lacked strength requirements, higher strength matrix alloy design lead to Al based binary alloys for use in the composite. Al Sn was investigated because it exhibited the lowest eutec tic temperature. Al Cu was

PAGE 148

148 investigated because it would show high strength increases per solute addition and decrease the eutectic temperature. Al Mg was investigated because it showed potential for the highest total solid solution strengthening of binary Al alloys. Al Si was investigated because it would decrease the eutectic temperature and increase strength moderately, but would exhibit high castability. Composite samples were fabricated with NiTi SMA wire reinforcements, mechanical testing was complete d in tension, and microstructures were examined via optical microscopy. Healing was performed on the cracked tensile samples. The Al Si was the only alloy system which exhibited appreciable healing. The Al Cu and Al Mg systems were too brittle to allow for transformation in the SMA wires, thus the required clamping force across the interface was not present. Al Sn samples were found to be too ductile, allowing for necking and permanent strains to deform the NiTi wires. In Al Si, the moderate ductility witho ut much necking allowed for pre straining of the SMA wires prior to healing, thus creating the clamping force at the interface to enable healing. Average healing was found to be over 91% retained strength post healing cycle. However, the Al Si system posse ssed high healing temperatures and low yield strengths. Therefore, a higher strength ternary alloy, Al Cu Si, was designed. Samples were fabricated and the strength and thermodynamic models utilized to design the alloy were validated. Healing, however, was found to be relatively lacking due to the brittle nature of the matrix alloy. Therefore, novel processing techniques were investigated, including wire pre straining prior to casting and hot pressing. Hot pressing was found to enable bonding across an inte rface, and enable multiple wires to be incorporated into

PAGE 149

149 the composite, which greatly increased the failure strain in the Al Cu Si composite. Further refinement of this technique is desired to achieve higher healing outputs. The healing of the Sn and Al b ased composites were all conducted under vacuum, a limitation to the ap plication of liquid assisted self healing MMCs. Addition of reactive elements to lead free solders has shown increase in bonding to oxide containing surfaces. Therefore, a study was con ducted to investigate a prototype Sb based system with reactive and non reactive additions. The fracture toughness of the reactive element addition specimen which was healed was found to match the monolithic specimen fracture toughness. Next, the addition of a reactive element to a Sn Bi X composite was investigated. However, oxidation of the reactive element during fabrication resulted in poor healing. Further refinement of the casting technique is required to enable incorporation of a reactive element to the Al based composites.

PAGE 150

150 CHAPTER 9 FUTURE WORK To advance the research of self healing MMCs, there are several areas of research with potential to delve into further investigations including 1) new matrix alloys, 2) new SMA reinforcements, 3) incorporatio n of reactive elements, and 4) optimization of healing cycles. Matrix Alloy Development While the initial work into different Al based binary alloys and one ternary alloy has shown promise, further development of Al based matrix alloys is needed. Increased yield strengths are needed to develop self healing composites which will be able to replace current traditional materials. Because of the methodology used in developing the self healing composites, it is thought the thermodynamic design can be used to adv ance other matrix alloys. One, Al Mg Li, shows potential for the highest specific strength via solid solution strengthening and low density. Another, Al Mg Cu, emulates 6000 series Al alloys with the ability for high strength increases through precipitatio n of secondary phases. Initial research into Al Mg Li and Al Mg Cu can be found in Appendix A. Other ternary alloy matrices with self healing potential include Al Mg Cu, Al Mg Zn, and Al Cu Li. These alloys exhibit a ternary eutectic point on the Al rich s ide near a region of Al solid solution. Shape Memory Alloy Development Currently, the SMA wires incorporated into the self healing composites are purchased commercially. These wires are not designed to be heated to the high temperatures involved with the healing process in Al based systems and their strengths

PAGE 151

151 show large decreases after a healing cycle from the as received values. Therefore, future designs must incorporate SMA wires able to withstand greater temperatures while maintaining their strength. Th is would enable higher yield strengths from the Al based composites and thus more application areas for liquid assisted self healing MMCs. Work by another member of the Manuel group, Derek Hsu, is currently pursuing high temperature SMA with his doctoral r esearch ( 179 ) Reactive Element Inco rporation Oxidation at the fracture surface has been found to prevent healing in Sn and Al based matrix alloys with self healing potential. Therefore, incorporation of reactive element additions similar to those designed in Chapter 7 has the potential to enable healing in an oxygen containing environment. This would be critical to improving self healing MMCs for use in terrestrial applications. With the success of the Sb based prototype alloys, the design of reactive elements to the Al based matrices must be pursued. Due to the reactivity of Al, potential additions are mostly limited to rare earth metals (e.g. Ce, Gd, Y) and alkaline earth metals (e.g. Ca, Y). The reactive element addition must oxidize preferentially over all other elements in the matrix al loy to ensure it will reduce any formation of surface oxides at the crack interface. The addition must also partition to the eutectic phase, to ensure it will be melted during the healing cycle, thus enabling faster reduction kinetics of surface oxides. He aling Cycle Optimization After more advanced composites are developed, optimization of the healing cycle is required. Currently, healing is conducted over a 24 hour period. However, a reduction of this time would enable incorporation of self healing MMCs i nto more applications.

PAGE 152

152 Optimization would not only include the time, but also the temperature required for healing, which directly affects the percent liquid. Previous work had shown 15 20% was optimal for Sn Bi alloys ( 78 ) but it is not known whether or not this translates directly to the Al based alloys instead. While Chapter 5 detailed how 20% liquid worked in an Al Si composite, a reduction of this percent would also decrease the healing temperature. However, optimization must be evaluated against the ability to heal

PAGE 153

153 APPENDIX A MECHANICAL AND THERMAL TESTING DATA For the matrix alloys, NiTi reinforcements, and composite specimens investigated for self healing capabilities, mechanical and thermal data was collected. Below are the tensile curves for all specimens included in this dissertation and thermal testing data for the NiTi wires and Al Cu Si matrix. Sn Bi For the Sn Bi system, tensile curves for the matrix alloy specimens can be found in Figure A 1. For each virgin composite specimen, tensile resul ts can be found in Figure A 2 Images of Sn Bi composite specimen A and C in the virgin and healed condition can be found in Figure A 3 and Figure A 4, respectively. Figure A 1 : Mechanical testing results of Sn 21wt%Bi matrix alloys after heat treatment for 24 hours at 169C.

PAGE 154

154 Figure A 2 : Mechanical testing results of Sn 21wt%Bi composites ( V f = 0.3 0.4%) after heat treatment for 24 hours at 169 C. Figure A 3 : Mechanical testing comparison of Sn V f = 0.31%) in the virgin and he aled composite condition

PAGE 155

155 Figure A 4 : Mechanical testing comparison of Sn V f = 0.40%) in the virgin and healed composite condition The NiTi reinforcements used were also investigated for tensile data. Figure A 5 shows tensile curv es for the as received NiTi BH 0075 wires ( Ni 50.6 at% Ti, = 0. 18 mm cold drawn, Memry Corporation). Figure A 6 shows tensile data after heat treatment of 500 C for 3 hours whereas Figure A 7 shows the wires after a heat treatment of 500C for 3 hour s then 169 C for 24 hours. Wires removed from a Sn Bi composite sample which had been heat treated for 500C for 3 hours, cast into the composite, then heat treated again for 169 C for 24 hours (2x) and removed from the matrix alloy by using acid (solution of 2:1:1 H 2 O, HNO 3 and HCl) to dissolve the matrix are shown in Figure A 8.

PAGE 156

156 Figure A 5 : Tensile testing results of as received NiTi BH 0075 wires Figure A 6 : Tensile testing results of NiTi BH 0075 wires heat treated for 3 hours at 500C

PAGE 157

157 Figure A 7 : Tensile testing results of NiTi BH 0075 wires heat treated for 3 hours at 500C and then again at 169 C for 24 hours Figure A 8 : Tensile testing results of NiTi BH 0075 wires removed from a Sn Bi composite sample which had been heat treated for 3 hours at 500C and 169 C for 24 hours (2x)

PAGE 158

158 The NiTi BH wires were also investigated for their thermal properties via DSC. Several wires were heat treated for 3 hours at 5 00 C and tested from 50 C to 200 C at a rate of 10 C A representative DSC curve aft er the 3 hour heat treatment is shown in Figure A 9. Additional BH wires were heat treated for 3 hours at 5 00 C then for 24 hours at 169 C and tested in a similar fashion. A representative DSC curve for these NiTi BH wires is shown in Figure A 10. The aver age values for the wires tested under both heat treatment conditions is shown in Table 3 4. Figure A 9 : Thermal transition temperatures for a B H 007 5 NiTi wire heat treated for 3 hours at 5 0 0 C

PAGE 159

159 Figure A 10 : Thermal transition temperatures for BB 0075 5 NiTi wire heat treated for 3 hours at 5 0 0 C and then 24 hours at 169 C Al Based Binary Systems For the Al Sn composite system, tensile curves for the composite specimens can be found in Figure A 11 Similar tensile tests for the Al Cu and Al Mg composite s can be found in Figure A 1 2 and Figure A 1 3 respectively. Figure A 11 : Mechanical testing results of Al 19.5 at% Sn (51.6 wt% Sn) composites ( V f = 1.5 1.8%) after heat treatment for 4 hours at 250 C.

PAGE 160

160 Figure A 1 2 : Mechanical testing results of Al 4.5 at% Cu (10 wt% Cu) composites ( V f = 2.2 2.4%) after heat treatment at 566 C for 24 hours . Figure A 1 3 : Mechanical testing results of Al 16.6 at% Mg (15.2 wt% Mg) composites ( V f = 1.9 4.0%) after heat treatment at 460 C for 24 hours For the Al S i system, tensile testing was completed for the matrix alloys specimens (Table A 1). The representative stress strain curves for the matrix alloy can be found in Figure A 1 4 Al Si composite specimen tensile data can be found in Figure

PAGE 161

161 A 1 5 Similar tensil e tests for Al Si composites containing increased NiTi reinforcements can be found in Figure A 1 6 Images of Al virgin and healed condition can be found in Figure A 1 7 and Figure A 1 8 respectively. Table A 1: Tensi le testing results of Al 3.0 at% Si (3.1 wt% Si) matrix alloys after heat treatment at 592C for 24 hours Specimen Modulus (GPa) 0.2% Yield Stress (MPa) Ultimate Stress (MPa) Failure Strain (%) 1 63.3 42.3 76.1 2.02 2 58.4 47.8 75.6 2.18 3 67.6 44.9 8 4.4 2.27 4 77.9 42.8 72.3 1.77 Average 66.8 44.5 77.1 2.06 8.3 2.5 5.2 0.22 Figure A 1 4 : Mechanical testing results of Al 3.0 at% Si (3.1 wt% Si) matrix alloys after heat treatment at 592C for 24 hours.

PAGE 162

162 Figure A 1 5 : Mechanical testing resu lts of Al 3.0 at% Si (3.1 wt% Si) composites ( V f = 2.0 2.4%) after heat treatment at 592 C for 24 hours Figure A 1 6 : Mechanical testing results of Al 3.0 at% Si (3.1 wt% Si) composites ( V f = 3.7 4.4%) after heat treatment at 592 C for 24 hours

PAGE 163

163 Fi gure A 1 7 : Mechanical testing comparison of Al V f = 2.0 %) in the virgin and healed composite condition Figure A 1 8 : Mechanical testing comparison of Al V f = 2 4 %) in the virgin and healed composite condition

PAGE 164

164 Al Based Ternary System For the Al Cu Si system, the tensile stress strain diagrams for the matrix alloy can be found in Figure A 1 9 The tensile data for the composite specimens is found in Figure A 20 Figure A 1 9 : Mechanical testing resul ts of Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) matrix alloys after heat treatment at 530 C for 24 hours. Figure A 20 : Mechanical testing results of Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) composites ( V f = 2.3 2.9%) after heat treatment at 530 C for 24 hours

PAGE 165

165 N iTi Wire s Mechanical Properties For the Al Si and Al Cu Si composites, type BB 35 NiTi wires ( Ni 49.3 at% Ti, = 0. 87 mm) were used as continuous reinforcements. Tensile properties of the NiTi wires in the as received condition are shown in Figure A 21 BB w ires which were heat treated for 24 hours at 5 92 C and pulled in tension are shown in Figure A 2 2 Tensile results of BB wires heat treated for 24 hours at 530 C once, twice, or three times to simulate casting and he aling processes in Al Cu Si composite manufacturing are shown in Figure A 2 3 Figure A 21 : Tensile testing results of as received NiTi BB 35 wires

PAGE 166

166 Figure A 2 2 : Representative t ensile results of NiTi BB 35 wires heat treated for 24 hours at 5 92 C for e ither one cycle or two cycles Figure A 2 3 : Tensile testing results of NiTi BB 35 wires heat treated for 24 hours at 5 30 C for one, two, or three cycles to simulate the entire healing process for Al Cu Si based composites

PAGE 167

167 Thermal Properties The NiTi BB wires were also investigated for their thermal properties via DSC. A sample DSC curve of the as received BB wires is shown in Figure A 2 4 Wires which were heat treated for 24 hours at 5 30 C for one cycle are shown in Figure A 2 5 and two cycles in A 2 6 A representative DSC curve for BB wires heat treated at 592 C for 24 hours is shown in Figure A 2 7 with all data included in Table A 2. BB wires put through two c ycles of 592 C for 24 hours are shown in Figure A 2 8 with all wire data included in Table A 3 Figure A 2 4 : Representative t hermal transition temperatures for as received BB 35 NiTi wire

PAGE 168

168 Figure A 2 5 : Thermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 5 30 C Figure A 2 6 : Thermal transition temperatures for BB 35 N iTi wire heat treated for 24 hours at 5 30 C two times

PAGE 169

169 Figure A 2 7 : Representative t hermal transition temperatures for BB 35 NiTi wire heat treated for 24 hours at 5 92 C Table A 2: Thermal transition temperatures of BB 35 NiTi wires heat treated for 24 hours at 5 92 C Temperature ( C) Number A s A f M s M f 1 39.2 49.9 16.4 6.4 2 40.4 53.9 15.6 2.8 3 40.3 56.4 15.4 1.0 Average 40.0 53.4 15.8 3.4 0.6 3.3 0.6 2.8

PAGE 170

170 Figure A 2 8 : Representative t hermal transition temperatures for BB 35 NiTi wire h eat treated for 24 hours at 5 92 C two times Table A 3: Thermal transition temperatures of BB 35 NiTi wires heat treated for two cycles at 5 92 C for 24 hours Temperature ( C) Number A s A f M s M f 1 40.0 53.4 15.6 3.8 2 39.8 53.4 15.2 4.6 3 40.1 53.7 15 .9 4.1 Average 40.0 53.5 15.5 4.2 0.2 0.2 0.4 0.4 Thermal Transitions in Al Cu Si The percent liquid in the Al Cu Si matrix alloy was investigated via DSC. The resultant data for transition temperatures and heat of fusion is shown in Figure A 2 9

PAGE 171

171 Figure A 2 9 : Thermal transition te mperatures for Al 4.1 at% Cu 2.0 at% Si (9.0 wt% Cu 1.9 at% Si) matrix alloy after heat treatment at 530 C for 24 hours

PAGE 172

172 APPENDIX B ALUMINUM MAGNESIUM BASED MATRIX DESIGN Future iterations of matrixes for self healing MMC must advance the strength of the composite in order to compete with traditional structural materials. To do this, different alloying additions and heat treatments must be attempted. This chapter details two such compositions which show promise for future developments Al Mg Li an d Al Mg Si. Al Mg Li alloys are known to possess high specific strength due to the solid solution strengthening abilities of Mg and Li in Al as well as the low density from these additions. Al Mg Si alloys are able to obtain high strength through precipita tion of second phases, and thus different heat treatments must be investigated to achieve the highest strength while maintaining its healing capabilities. Al Mg Li Matrix Looking at the Al based ternary matrix alloys, an alloy based on the Al Mg Li system has the potential reach the highest strength values achievable in solid solution ( Figure 4 4), while also decreasing the eutectic temperature (Figure 4 3) and decreasing composite density. These are all highly desirable properties based on the system desi gn chart for self healing MMCs (Figure 4 1) Matrix Design The Al Mg Li phase diagram was calculated using Pandat software with their magnesium database, PanMagnesium ( 107 ) ( Figure B 1); it was found be similar to the phase diagrams found in literature ( 180 ) The lowest melting eutectic found on the Al rich side was calculated to be Al 36.1 at% Mg 1.5 at% Li (34.1 wt% Mg 0.40 wt% Li) at a temperature of 445 C

PAGE 173

173 Figure B 1: Liquidus projection of the Al Mg Li system. The lowest melting eutectic point calculated to be 445 C at a composit ion of Al 36.1 Mg 1.5 Li (at%) as indicated by the arrow Calculated using Pandat ( 10 7 ) At the ternary eutectic temperature of 445 C, an isotherm was calculated ( Figure B 2 for the Al rich region). Drawing a line between the eutectic point and the closest solid solution Al region, a pseudo isopleth was calculated ( Figure B 3 ). At a heal ing temperature of 460 C, it was calculated that a composition of Al 18.7 at% Mg 1.1 at% Li (17.3 wt% Mg 0.3 wt% Li) would yield 20% liquid during a healing heat treatment. A fluctuation of 5 C from the healing temperature results in a range of onl y 17 22.5% liquid in the matrix.

PAGE 174

174 Figure B 2: Is otherm at 445 C in the Al Mg Li ternary alloy system. The dashed line represents the connection between the ternary eutectic and nearest Al solid solution phase field. Calculated using Pandat ( 107 ) Figure B 3: Pseudo isopleth of Al Mg Li system along the line from the ternary eutectic po int to the closest Al solid solution phase region. At a healing temperature of 460 C, a composition of Al 18.7 at% Mg 1.1 at% Li will yield 20% liquid in the matrix. Calculated using Pandat ( 107 )

PAGE 175

175 Composite Fabrication Appropriate amounts of Al (Al shot, 99.99%, Alfa Aesar), Mg (Mg pieces, 99.99%, American Elements), and Li (Li granules, 99%, Alfa Aesar) were melted in a furnace at 750 C contained within an argon atmosphere until a liquid solution. NiTi SMA wires, designated BB 35, were laid horizontally in the graphite tensile bar mold heated to 350 C The Al Mg Li melt was poured over t he wires in the mold and allowed to air cool. Each tensile bar was then placed horizontally into a furnace for 2 4 hour s at 460 C monitored by an external thermocouple, then air cooled to set the eutectic microstructure ( Figure B 4). Figure B 4: Repres entative microstructure of an Al 18.7 at% Mg 1.1 at% Li (17.3 wt% Mg 0.3 wt% Li) matrix exhibiting a non continuous eutectic phase surrounding Al solid solution phase Mechanical Testing After heat treatment, the Al Mg Li composite bars were ground t o a 320 grit surface finish before being tested in tension using the Instron at a rate of 1.0%/min. Results of the testing are shown in Table B 1. Because of the brittle nature of the

PAGE 176

176 composites, the 0.2% offset yield strength is not able to be reported. A tensile graph of Composite 2 is shown in Figure B 5 Table B 1: Mechanical testing results for Al Mg Li matrices reinforced with 2 3 vol% SMA wire reinforcements Specimen V f Wires (%) Modulus (GPa) Ultimate Stress (MPa) Failure Strain (%) 1 3.00 48.3 62. 3 0.12 2 3.04 53.7 126.5 0.31 3 2.29 52.6 41.9 0.11 Avg. 2.78 51.5 76.9 0.18 0.42 2.9 44.1 0.11 Figure B 5: Representative tensile results of an Al Mg Li matrix reinforced with 3.04 volume percent NiTi SMA wire reinforcements After tensile testing, the fractured composite bars were encapsulated under vacuum (in a Pyrex tube laying flat on graphite wrapped in Ta foil) and heat treated a second time for 24 hours at 460 C in an attempt to heal the crack. Healing was not found due to oxidation at the fracture interface as shown in Figure B 6. Casting defects such as oxide in clusions were also found, which is thought to have been the major

PAGE 177

177 factor in the composite displaying pre mature failure and lack of ductility. Investigation of into improved fabrication techniques, such as the hot pressing found in Al Cu Si, may enable to use of this matrix as a self healing MMC. Figure B 6: Fracture surface of Al Mg Li composite showing oxidation on the crack interface after a healing cycle. Note the presence of an oxide inclusion from casting which acted as the initiation point of fail ure. Al Mg Si Matrix The 6000 series of Al alloys based on the Al Mg Si system has potential to increase strength through precipitation of Mg 2 Si in the matrix. Paired with the pressing technique developed for Al Cu Si composites, this system has potential to improve the strength values achieved in self healing MMCs through extra fabrication steps; i.e. heat treatment and quenching. The matrix of Al Mg Si was designed in a similar manner to previous alloys through calculations using Pandat software with thei r magnesium database, PanMagnesium ( 107 ) The lowest melting eutectic found on the Al r ich side was calculated to be Al 5.7 at% Mg 1 2 .5 at% S i (5.2 wt% Mg 13.0 wt% Si) at a

PAGE 178

178 temperature of 556 C At the ternary eutectic temperature of 556 C, an isotherm was calculated to find the nearest solid solution Al region to the eutectic point. A s before, a pseudo isopleth was calculated between these two points ( Figure B 7). At a healing temperature of 565 C, it was calculated that a composition of Al 1.8 at% Mg 3.2 at% Si (1.6 wt% Mg 3.3 wt% Si) would yield 20% liquid during a healing heat treatment. A fluctuation of 5 C from the healing temperature results in a range of only 18.0 21.2% liquid in the matrix. Figure B 7: Pseudo isopleth of Al Mg Si system along the line from the ternary eutectic point to the closest Al solid solution ph ase region. At a healing temperature of 565 C, a composition of Al 1.8 at% Mg 3.2 at% Si will yield 20% liquid in the matrix. Calculated using Pandat ( 107 ) Unlike previous alloys, however, the Al Mg Si system would benefit the most from developing different thermal treatments to increase strength through precipitation. Thus to maximize the strength, different heat treatment sequences need to be investigated. Also, the healing cycle would also need to be paired with these different

PAGE 179

179 heat treatments in order to determine how the strengthening treatments for the matrix would affect eventual healing of the composite. Summary As shown above, there are more Al based matrix compositions which have only just begun to be investigated for self healing capabilities. Future studies must be completed to optimize healing cycles, matrix compositions, and precipitation heat treatments in order to obtain MMCs which rival traditional materials.

PAGE 180

180 LIST OF REFERENCES 1. Williams DF. On the Nature of Biomaterials. Biomaterials. [Leading Opinion]. 2009;30:5897 909. 2. Zhou BL. Bio inspired St udy of Structural Materials. Materials Science and Engineering C. 2000;11:13 8. 3. Zwaag SVd, Dijk NHv, Jonkers HM, Mookhoek SD, Sloof WG. Self Healing Behaviour in Man made Engineering Materials: Bioinspired but Taking Into Account Their Intrinsic Charact er. Philosophical Transactions of the Royal Society A. 2009;367:1689 704. 4. Carlson HC, Goretta KC. Basic Materials Research Program at the U.S. Air Force Office of Scientific Research. Materials Science and Engineering B. 2006;132:2 7. 5. White SR, Sotto s NR, Geubelle PH, Moore JS, Kessler MR, Sriram SR, Brown EN, Viswanathan S. Automatic Healing of Polymer Composites. Nature. 2001;409:794 7. 6. Nosonovsky M, Rohatgi PK. Biomimetics in Materials Science: Self Healing, Self Lubricating, and Self Cleaning M aterials: Springer; 2012. 7. Hager MD, Greil P, Leyens C, Zwaag SVd, Schubert US. Self Healing Materials. Advanced Materials. 2010;22:5424 30. 8. Kessler MR. Self healing: a new paradigm in materials design. Proceedings of the Institution of Mechanical Eng ineers, Part G: Journal of Aerospace Engineering. 2007;221(4):479 95. 9. Lumley RN. Self Healing in Metals: An Emerging Phenomenon in Metallurgy. Materials Forum. 2007;31:40 51. 10. Zwaag SVd. An Introduction to Materials Design Principles: Damage Preventi on vs. Damage Management. In: Zwaag Svd, editor. Self Healing Materials: An Alternative Approach to 20 Centuries of Materials Science: Springer; 2007. p. 1 18. 11. White SR, Sottos NR, Geubelle PH, Moore JS, Kessler MRS, S. R., Brown EN, Viswanathan S. Aut omatic Healing of Polymer Composites. Nature. 2001;409:794 7. 12. Wool RP. Self healing materials: A Review. Soft Matter. 2008;4(3):400.

PAGE 181

181 13. Wu DY, Meure S, Solomon D. Self healing Polymeric Materials: A Review of Recent Developments. Progress in Polymer S cience. 2008;33:479 522. 14. Wua M, Johannesson B, Geiker M. A review: Self healing in Cementitious Materials and Engineered Cementitious Composite as a Self healing Material. Construction and Building Materials. 2012;28:571 83. 15. Naslain R. Design, Prep aration and Properties of Non oxide CMCs for Application in Engines and Nuclear Reactors: An Overview. Composites Science and Technology. 2004;64:155 70. 16. Muynck WD, Beliea ND, Verstraeteb W. Microbial carbonate precipitation in construction materials: A review. Ecological Engieering. 2010;36:118 36. 17. Ghosh SK, editor. Self Healing Materials: Fundamentals, Design Strategies, and Applications: Wiley VCH; 2009. 18. Zwaag SVd, editor. Self Healing Materials: An Alternative Approach to 20 Centuries of Mat erials Science: Springer; 2007. 19. Peterson GE. Dynamics of Meteor Outbursts and Satellite Mitigation Strategies: Aerospace Press; 1999. 20. Spirit Continues Driving on Five Wheels. National Aeronautics and Space Administration; 2006 [May 7, 2013]; Availa ble from: http://www b.jpl.nasa.gov/missions/mer/images.cfm/images.cfm?id=1906 21. Amos J. NASA's Curiosity Rover Targets Smaller Landing Zone. BBC; 2012 [May 7, 2013]; Available from: http://www.bbc.co.uk/news/science environment 18401248 22. Karpov EG, Grankin MV, Liu M, Ariyan M. Characterization of Precipitative Self healing Materials by Mechanok inetic Modeling Approach. Journal of the Mechanics and Physics of Solids. 2012;60:250 60. 23. Ryan NJ, Evans K, Alexander P, Chirnside I, McEwan I, Stebbings T, editors. Using Platelet Technology to Seal and Locate Leaks in Subsea Umbilical Lines. Offshore Technology Conference; 2007; Houston, TX. 24. Self Healing Steel. Time. 1967 August 18, 1967:37. 25. Files BS. Design of a Biomimetric Self Healing Superalloy Composite. Evanston, IL: Northwestern University; 1997.

PAGE 182

182 26. Chu E, Fong D, Klapperich C, Yeh T. Terminator 3: Biomimetic Smart Steels. Evanston, IL: Northwestern University; 1993. 27. Manuel MV. Principles of Self Healing in Metals and Alloys: An Introduction. In: Ghosh SK, editor. Self Healing Materials: Fundamentals, Design Strategies, and Applicat ions: Wiley VCH; 2009. 28. Lumley RN. Self Healing in Aluminum Alloys. In: Zwaag SVd, editor. Self Healing in Materials: An Alternative Approach to 20 Centuries of Materials Science: Springer; 2007. p. 219 54. 29. Wilm A. Physical metallurgical Experiment s on Aluminum Alloys Containing Magnesium. Metallurgie. 1911;8:223. 30. Lumley RN, Morton AJ, Polmear IJ. Enhanced Creep Performance in an Al Cu Mg Ag Alloy Through Underageing. Acta Materialia. 2002;50:3597 608. 31. Lumley RN, Morton AJ, Polmear IJ. Enhan ced Creep Resistance in Underaged Aluminum Alloys. Materials Science Forum. 2000;331 337:1495 500. 32. Lumley RN, Polmear IJ, editors. Advances in Self Healing of Metals. First International Conference on Self Healing Materials; 2007; Netherlands: Springer 33. Lumley RN, Polmear IJ, Morton AJ. Development of Mechanical Properties During Secondary Aging in Aluminum Alloys. Materials Science and Technology. 2005;21(9):1025 32. 34. Lumley RN, Polmear IJ, Morton AJ. Interrupted Aging and Secondary Precipitatio n in Aluminum Alloys. Materials Science and Technology. 2003;19:1483 90. 35. Lumley RN, Schaffer GB. Precipitation Induced Densification in a Sintered Al Zn Mg Cu Alloy. Scripta Materialia. 2006;55:207 10. 36. Ringer SP, Yeung W, Muddle BC, Polmear IJ. Pre cipitate Stability in Al Cu Mg Ag Alloys Aged at High Temperatures. Acta Metallurgica et Materialia. 1994;42(5):1715 25. 37. Wang J, Wu X, Xia K. Creep Behaviour at Elevated Temperatures of an Al Cu Mg Ag Alloy. Materials Science and Engineering A. 1997;23 4 236:287 90. 38. Skrotzki B, Shiflet GJ, E.A. Starke J. On the Effect of Stress on Nucleation and Growth of Precipitates in an Al Cu Mg Ag Alloy. Metallurgical and Materials Transactions A. 1996;27A:3431 44.

PAGE 183

183 39. Hautakangas S, Schut H, Dijk NHv, Castillo PEJRDd, Zwaag SVd. Self Healing of Deformation Damage in Underaged Al Cu Mg Alloys. Scripta Materialia. 2008;58:719 22. 40. Hautakangas S, Schut H, Zwaag SVd, Castillo PEJRDd, Dijk NHv. Positron Annihilation Spectroscopy as a Tool to Develop Self Healing i n Aluminium Alloys. Physica Status Solidi C. 2007;4(10):3469 72. 41. Hautakangas S, Schut H, Zwaag Svd, Castillo PEJRDd, Dijk NHv, editors. The Role of Aging Temperature on the Self Healing Kinetics in an Underaged AA2024 Aluminum Alloy. First Internationa l Conference on Self Healing Materials; 2007; Netherlands: Springer. 42. Wanhill RJH, editor Fatigue Crack Initiation in Aerospace Aluminum Alloys, Components, and Structures. First International Conference on Self Healing Materials; 2007; Netherlands: Spr inger. 43. Laha K, Kyono J, Kishimoto S, Shinya N. Beneficial Effect of B Segregation on Ceep Cavitation in a Type 347 Austenitic Stainless Steel. Scripta Materialia. 2005;52:675 8. 44. Laha K, Kyono J, Sasaki T, Kishimoto S, Shinya N. Improved Creep Stren gth and Creep Ductility of Type 347 Austenitic Stainless Steel through the Self Healing Effect of Boron for Creep Cavitation. Metallurgical and Materials Transactions A. 2005;36A:399 409. 45. Laha K, Kyono J, Shinya N. An Advanced Creep Cavitation Resistan ce Cu Containing 18 Cr 12Ni Nb Austenitic Stainless Steel. Scripta Materialia. 2007;56:915 8. 46. Laha K, Kyono J, Shinya N. Some Chemical and Microstructural Factors Influencing Creep Cavitation Resistance of Austenitic Stainless Steels. Philosophical Mag azine. 2007;87(17):2483 505. 47. Laha K, Kyono J, Shinya N. Copper, Boron, and Cerium Additions in Type 347 Austenitic Steel to Improve Creep Rupture Strength. Metallurgical and Materials Transactions A. 2011;43A:1187 97. 48. Shinya N. Self Healing of Mech anical Damage in Metallic Materials. Advances in Science and Technology. 2008;54:152 7. 49. Shinya N. Self healing of Metallic Materials: Self healing of Creep Cavity and Fatigue Cavity/crack. In: Ghosh SK, editor. Self healing Materials: Fundamentals, Des ign Strategies, and Applications. Weinheim: Wiley VCH; 2009. p. 219 50.

PAGE 184

184 50. Shinya N, Kyono J. Effect of Boron Nitride Precipitation at Cavity Surface on Rupture Properties. Materials Transactions. 2006;47(9):2302 7. 51. Shinya N, Kyono J, Laha K. Self hea ling Effect of Boron Nitride Precipitation on Creep Cavitation in Austenitic Stainless Steel. Journal of Intelligent Material Systems and Structures. 2006;17:1127 33. 52. Shinya N, Kyono J, Laha K. Improvement of Creep Rupture Properties of Heat Resisting Steels. Materials Science Forum. 2003;426 432:1107 12. 53. Shinya N, Kyono J, Laha K, Masuda C, editors. Self Healing of Creep Damage through Autonomous Boron Segregation and Boron Nitride Precipitation During High Temperature Use of Austenitic Stainless S teels. First International Conference on Self Healing Materials; 2007; Netherlands: Springer. 54. White CL, Padgett RA, Swindeman RW. Sulfur and Phosphorus Segregation to Creep Cavities and Grain Boundaries in 304 SS. Scripta Materialia. 1981;15(7):777 82. 55. Russell AM, Lee KL. Structure Property Relations in Nonferrous Metals. Hoboken, NJ: John Wiley and Sons, Inc.; 2005. 56. Ostrander GW. Plating. 1951;38:1033. 57. Kendig MW, Buchheit RG. Corrosion Inhibition of Aluminum and Aluminum Alloys by Soluble C hromates, Chromate Coatings, anad Chromate Free Coatings. Corrosion. 2003;59(5):379 99. 58. O'Brien P, Kortenkamp A. The Chemistry Underlying Chromate Toxicity. Transition Metal Chemistry. 1995;20:636 42. 59. Buchheit RG, Mamidipally SB, Schmutz P, Guan H. Active Corrosion Protection in Ce Modified Hydrotalcite Conversion Coatings. Corrosion. 2002;58(1):3 14. 60. Bethencourt M, Botana FJ, Calvino JJ, Marcos M, Rodriguez Chacon MA. Lanthanide Compounds as Environmentally Friendly Corrosion Inhibitors of Alum inum Alloys: A Review. Corrosion Science 1998;40(11):1803 19. 61. Pint BA. Optimization of Reactive Element Additions to Improve Oxidation Performance of Alumina Forming Alloys. Journal of the American Ceramic Society 2003;86(4):686 95. 62. Hultquist G, Tveten B, Ho¨rnlund E, Limback M, Haugsrud R. Self Repairing Metal Oxides. Oxidation of Metals 2001;56(3/4):313 46.

PAGE 185

185 63. Bennett IJ, Kranenburg JM, Sloof WG. Modeling the Influence of Reactive Elements on the Work of Adhesion between Oxides and Metal Alloy s. Journal of the American Ceramic Society 2005;88(8):2209 16. 64. Bennett IJ, Sloof WG. Modelling the Influence of Reactive Elements on the Work of Adhesion Between a Thermally Grown Oxide and a Bond Coat Alloy. Materials and Corrosion 2006;57(3):223 9. 65. Blaiszik BJ, Kramer SLB, Grady ME, McIlroy DA, Moore JS, Sottos NR, White SR. Autonomic Restoration of Electrical Conductivity. Advanced Materials 2012;24:398 401. 66. Chiechi RC, Weiss EA, Dickey MD, Whitesides GM. Eutectic Gallium Indium (EGaIn): A Moldable Liquid Metal for Electrical Characterization of Self Assembled Monolayers. Angewandte Chemie International Edition 2008;47:142 4. 67. Dickey MD, Chiechi RC, Larsen RJ, Weiss EA, Weitz DA, Whitesides GM. Eutectic Gallium Indium (EGaIn): A Liquid Metal Alloy for the Formation of Stable Structures in Microchannels at Room Temperature. Advanced Functional Materials 2008;18:1097 104. 68. So J H, Thelen J, Qusba A, Hayes GJ, Lazzi G, Dickey MD. Reversibly Deformable and Mechanically Tunable Fluidic An tennas. Advanced Functional Materials 2009;19:3632 7. 69. Olson GB. Systems Design of Hierarchically Structured Materials: Advanced Steels. Journal of Computer Aided Materials Design 1997;4:143 56. 70. Olson GB. Computational Design of Hierarchically Str uctured Materials. Science 1997;277:1237 42. 71. Fine R, Heise M, Johnson M, Korpanty B. Terminator 3: A Brain for Frankensteel. Evanston, IL: Northwestern University; 1999. 72. McCallum M, McGeorge A, Witney A. Terminator 3+: The Biomimetic Smart Steel. Evanston, IL: Northwestern University; 1995. 73. Files B, Olson G, editors. Terminator 3: Biomimetic Self Healing Alloy Composite. SMST 97: Processings of the Second International Conference on Shape Memory and Superelastic Technologies; 1997; Pacific Grov e, CA: International Organization on Shape Memory and Superelastic Technologies.

PAGE 186

18 6 74. Forbell C, Barney M, Scharff C, Lai W. Tin Based Self Healing Alloy. Evanston, IL: Northwestern University; 1997. 75. Bernikowicz P. Terminator 3: Design of Tin Based Biom imetic Self Healing Alloy Tensile Specimens. Evanston, IL: Northwestern University; 1998. 76. Hughes J, Rellas D. Terminator III: Design of Tin Based Biomimetic Self Healing Alloy Tensile Specimens. Evanston, IL: Northwestern University; 1998. 77. Tao M, P rice E, Asahara J, Benes K, Key T. Terminator 3: Creating a Metal That Mimics Biological Healing. Evanston, IL: Northwestern University; 1998. 78. Knapp S. Frankensteel Revisited : Prototyping of Self Healing Metal. Evanston, IL: Northwestern University; 2 004. 79. Heneghan E, Lee D, Pasquini B, Ravi Shankar P. Terminator 3. Evanston, IL: Northwestern University; 2000. 80. Manuel MV. Design of a Biomimetic Self Healing Alloy Composite. Evanston, IL: Northwestern University; 2007. 81. Burton DS, Gao X, Brinso n LC. Finite element simulation of a self healing shape memory alloy composite. Mechanics of Materials 2006;38(5 6):525 37. 82. Kang J, Franks C, Baluch A, Weninger J, Chatterjee A. Terminator 4: Biomimetic Shape Memory Reinforced Alloy Composite for Mars Rover Self Healing Axles. Evanston, IL: Northwestern University; 2002. 83. Jung J. Design of Nanodispersion Strengthened TiNi Base Shape Memory Alloys. Evanston, IL: Northwestern University; 2003. 84. Manuel MV, Olson GB, editors. Biomimetic Self Healing Metals. First International Conference on Self Healing Materials; 2007; Netherlands: Springer. 85. Manuel MV, Olson GB, editors. Biologically Inspired Self Healing Metals. Second International Conference on Self Healing Materials; 2009; Chicago, IL. 86. Ru zek AC. Synthesis and Characterization of Metallic Systems with Potential for Self Healing. Milwaukee, WI: University of Wisconsin Milwaukee; 2009. 87. Martinez Lucci J. Heat Transfer and Fluid Flow Analysis of Self Helaing in Intermetallic Materials. Milw aukee, WI: University of Wisconsin Milwaukee; 2011.

PAGE 187

187 88. Martinez Lucci J, Ruzek A, Misra SK, Rohatgi PK, Amano RS, editors. Self Healing in Metal Castings. AFS 2011; 2011: American Foundry Society. 89. Martinez Lucci J, Amano RS, Rohatgi PK, Schultz B, edi tors. Experiment and Computational Analysis of Self Healing in an Aluminum Alloy. ASME International Mechanical Engineering Congress and Exhibition; 2008; Boston, MA: ASME. 90. Martinez Lucci J, Amano RS, Rohatgi PK, Schultz B, editors. Self Healing in an Aluminum Alloy Reinforced with Microtubes. 3rd Energy Nanotechnology International Conference; 2008; Jacksonville, FL. 91. Martinez Lucci J, Amano RS, Rohatgi PK, Schultz B, Ruzek A, editors. Design and Demonstration of Self Healing Behavior in a Lead Free Solder Alloy. 7th International Energy Conversion Engineering Conference; 2009; Denver, CO: American Institute of Aeronautics and Astronautics. 92. Nosonovsky M, Amano R, Luccia JM, Rohatgi PK. Physical Chemistry of Self Organization and Self Healing in M etals. Physical Chemistry Chemistry Physics 2009;11:9530 6. 93. Funakubo H, editor. Shape Memory Alloys. New York: Gordon and Breach Science Publishers; 1987. 94. Perkins J. Ti Ni and Ti Ni X Shape Memory Alloys. Metals Forum 1981;4(3):153 63. 95. Wayman CM, Duerig TW. An Introduction to Martensite and Shape Memory. In: Duerig TW, Melton KN, Stockel D, Wayman CM, editors. Engineering Aspects of Shape Memory Alloys. London: Butterworth Heinemann Ltd.; 1990. p. 3 20. 96. Yamada Y, Taya M, Watanabe R. Strengthening of Metal Matrix Composite by Shape Memory Effect. Materials Transactions, JIM 1993;34(3):254 60. 97. Araki S, Ono H, Saito K. Micromechanical Analysis of Crack Closure Mechanism for Intelligent Material Containing TiNi Fibers (1st Report Modeling of Crack Closure Mechanism and Analysis of Stress Intensity Factor). JSME International Journal Series A 2002;45(2):208 16. 98. Araki S, Ono H, Saito K. Micromechanical Analysis of Crack Closure Mechanism for Intelligent Material Containing TiN i Fibers (2nd Report, Numerical Calculation of Stress Intensity Factor in the Process of Shape Memory Shrinkage of TiNi Fibers. JSME International Journal Series A 2002;45(3):356 62.

PAGE 188

188 99. Hamada K, Kawano F, Asaoka K. Shape Recovery of Shape Memory Alloy F iber Embedded Resin Matrix Smart Composite after Crack Repair. Dental Materials Journal 2003;22(2):160 7. 100. Poon C, Zhou L, Jin W, Shi S. Interfacial Debond of Shape Memory Alloy Composites. Smart Materials and Structures 2005;14:N29 N37. 101. Payande h Y, Meraghni F, Patoor E, Eberhardt A. Debonding Initiation in a NiTi Shape Memory Wire Epoxy Matrix Composite Influence of Martensitic Transformation. Materials & Design 2010;31:1077 84. 102. Armstrong WD, Lorentzen T. Fiber Phase Transformation and M atrix Plastic Flow in a Room Temperature NiTi Shape Memory Alloy Fiber Reinforced 6082 Aluminum Matrix Composite. Scripta Materialia 1997;36(9):1037 43. 103. Furuya Y, Sasaki A, Taya M. Enhanced Mechanical Properties of TiNi Shape Memory Fiber / Al Matrix Composite. Materials Transactions, JIM 1993;34(3):224 7. 104. Lee JK, Taya M. Strengthening Mechanism of Shape Memory Alloys Reinforced Metal Matrix Composite. Scripta Materialia 2004;51:443 7. 105. Wang X. Shape Memory Alloy Volume Fraction of Pre stre tched Shape Memory Alloy Wire Reinforced Composites for Structural Damage Repair. Smart Materials and Structures 2002;11:590 5. 106. Kirkby EL, Rule JD, Michaud VJ, Sottos NR, White SR, Manson J AE. Embedded Shape Memory Alloy Wires for Improved Performan ce of Self Healing Polymers. Advanced Functional Materials 2008;18:2253 60. 107. Chen SL, Zhang F, Daniel S, Xie FY, Yan XY, Chang YA, Schmid Fetzer R, Oates WA. Calculating Phase Diagrams Using Pandat and PanEngine. Journal of Materials 2003;53:48 51. 1 08. NIST Pb Free Solder Database. Available from: http://www.metallurgy.nist.gov/phase/solder/solder.html 109. Bi Sn (Bismuth Tin) [database on the Internet]. ASM International. 1992 [cited February 2010]. Available from: http://products.asminternational.org/hbk/index.jsp 110. Reed TB. Free Energy of Formation of Binary Compounds: An Atlas of Charts for High Temperat ure Chemical Calculations. Cambridge, MA: MIT Press; 1971.

PAGE 189

189 111. ASTM. ASTM E8M. Standard Test Methods for Tension Testing of Metallic Materials [Metric]. West Conshohocken, PA: ASTM International; 2004. 112. Clyne TW, Withers PJ. An Introduction to Metal Matrix Composites. Davis EA, Ward IM, editors. New York: Cambridge University Press; 1993. 113. Melton KN. Ni Ti Based Shape Memory Alloys. In: Duerig TW, Melton KN, Stockel D, Wayman CM, editors. Engineering Aspects of Shape Memory Alloys. London: Butterw orth Heinemann Ltd.; 1990. p. 21 35. 114. ASTM. ASTM E794 06. Standard Test Method for Melting and Crystallization Temperatures By Thermal Analysis. West Conshohocken, PA: ASTM International; 2012. 115. Tuttle WP, Fisher CR, Manuel MV, editors. Correlat ion of Pressure to Bonding Capabilities Using Novel Heat Treatment Methods in Prototype Sn Bi Alloys. TMS Annual Meeting and Exhibition EMPMD Technical Division; 2013; San Antonio, TX: Poster. 116. ASTM. ASTM F15 04. Standard Specification for Iron Ni ckel Cobalt Sealing Alloy. West Conshohocken, PA: ASTM International; 2009. 117. Meyers MA, Chawla KK. Mechanical Metallurgy Principles and Applications: Prentice Hall Inc.; 1984. 118. Jenkins GM. The Systems Approach. In: Beishon RJ, Peters G, editors. Systems Behaviour. London: Open University Press; 1972. p. 56 79. 119. Olson GB. Systems Design of Materials. Encyclopedia of Materials Science and Engineering: Pergamon; 1993. p. 2041 5. 120. Olson GB. Designing a New Material World. In: Amato I, editor. Science Pathways of Discovery. New York: John Wiley & Sons; 2002. p. 91 108. 121. Olson GB. New Steels By Design. Journal of Materials Education 1989;11:515 28. 122. Olson GB. Computational Design of Hierarchically Structured Materials. Science 1997;277( 5330):1237 42. 123. Olson GB. Designing a New Material World. Science 2000;288:993 8.

PAGE 190

190 124. Olson GB. Beyond Discovery: Design for a New Material World. CALPHAD 2001;25(2):175 90. 125. Nembach E. Particle Strengthening of Metals and Alloys. New York: John Wiley & Sons, Inc.; 1997. 126. Volume 3: Alloy Phase Diagrams [database on the Internet]. ASM International. 2012 [cited June 2011]. Available from: http://products.asminternational.org/hb k/index.jsp 127. Dorn JE, Pietrokowsky P, Tietz TE. The Effect of Alloying Elements on the Plastic Properties of Aluminum Alloys. Transactions of the American Institute of Mining and Metallurgical Engineers 1950;188(7):933 43. 128. Noble B, Harris SJ, D insdale K. Yield Characteristics of Aluminum Lithium Allys. Metals Science 1982;16:425 30. 129. R.E. Sanders J, Baumann SF, Stumpf HC. Wrought Non Heat Treatable Aluminum Alloys. Treatise on Materials Science and Technology 1989;31:65 105. 130. R.E. Sand ers J, Baumann SF, Stumpf HC. Non Heat Treatable Aluminum Alloys. In: E.A. Starke J, T.H. Sanders J, editors. Aluminum Alloys Physical and Mechanical Properties 1986. p. 1441 84. 131. Starr CD, Dorn JE. Effect of Copper Additions on the Plastic Properti es of Aluminum Zinc Alloy. Transactions of American Society for Metals 1954;46. 132. Al Sn (Aluminum Tin) [database on the Internet]. ASM International. 1992 [cited September 2012]. Available from: http://products.asminternational.org/hbk/index.jsp 133. Jennings PH, Singer ARE, Pumphrey WI. Hot Shortness of Some High Purity Alloys in the Systems Aluminum Copper Silicon and Aluminum Magnesium Silicon. The Journal of the Institute of Metals 1948;74:227 48. 134. Al Cu (Aluminum Copper) [database on the Internet]. ASM International. 1992 [cited September 2012]. Available from: http://products.asminternational.org/hbk/index.jsp 135. Al Mg (Aluminum Magnesium) [database on the Internet]. ASM International. 1992 [cited May 2010]. Available from: http://products.asminternational.org/hbk/index.jsp

PAGE 191

191 136. Al Si (Alu minum Silicon) [database on the Internet]. ASM International. 1992 [cited September 2012]. Available from: http://products.asminternational.org/hbk/index.jsp 137. Ni Ti (Nickel Titanium) [database on the Internet]. ASM International. 1992 [cited May 2013]. Available from: http://products.asminternational.org/hbk/index.jsp 138. Nishida M, Wayman CM, Honma T. Precipitation Processes in Near Equiatomic TiNi Shape Memory Alloys. Metallurgical Transactions A 1986;17A:1505 15. 139. Zhang J, Cai W, Ren X, Otsuka K, Asai M. The Nature of Reversible Change in Ms Temperature of Ti Ni Alloys with Alternating Aging. Materials Transac tions of the Japanese Institute of Materials 1999;40(1367). 140. Otsuka K, Ren X. Physical Metallurgy of Ti Ni Based Shape Memory Alloys. Progress in Materials Science 2005;50:511 678. 141. Mondolfo LF. Aluminum Alloys: Structure and Properties. Boston: Butterworth & Co.; 1976. 142. Hatch JE, editor. Aluminum: Properties and Physical Metallurgy. Materials Park, OH: ASM International; 1984. 143. Wilson FH. The Copper Rich Corner of the Copper Aluminum Silicon Diagram. Transactions of the American Institute of Mining and Metallurgical Engineers 1948;175:262 73. 144. Chang YA, Neumann JP, Mikula A, Goldberg D. Phase Diagrams and Thermodynamic Properties of Ternary Copper Metal Systems. International Copper Research Association 1979. 145. Pan XM, Lin C, Morr al JE, Brody HD. An Assessment of Thermodynamic Data for the Liquid Phase in the Al Rich Corner of the Al Cu Si System and Its Application to the Solidification of a 319 Alloy. Journal of Phase Equilibiria and Diffusion 2005;26(3):225 33. 146. Chen S w, L in C C, Chen C M. Determination of the Melting and Solidification Characteristics of Solders Using Differential Scanning Calorimetry. Metallurgical and Materials Transactions A 1998;29A:1965 72. 147. VanderVoort GF. Metallography: Principles and Practice. Materials Park, Ohio: ASM International; 1999.

PAGE 192

192 148. Kaufman JG, Rooy EL. Aluminum Alloy Castings : Properties, Processes, and Applications Materials Park, OH ASM International; 2004. 149. Plotkowski AJ. Refinement of the Cast Microstructure of Hypereutec tic Aluminum Silicon Alloys with an Applied Electric Potential. Allendale, MI: Grand Valley State University; 2012. 150. Miura I, Hamanaka H. Mechanical Properties and Thermal Stabilities of Al CuAl2 Si, Al CuAl2 NiCu3Al6 and Al CuAl2 FeCu2Al7 Ternary Eute ctic Composites. Journal of the Japanese Society of Metals 1972;36:1224 31. 151. Fleischer RL. Solid Solution Hardening. In: Peckner D, editor. The Strengthening of Metals. Reinhold, New York: Champman & Hall; 1964. 152. Johnson GR, Cook WH. A Constitutiv e Model and Data for Metals Subjected to Large Strains, High Strain Rates, and High Temperatures. Seventh International Symposium on Ballistics; Hague, Netherlands1983. p. 541 7. 153. ASTM. E209 00. Standard Practice for Compression Tests of Metallic Mate rials at Elevated Temperatures with Conventional or Rapid Heating Rates and Strain Rates. West Conshohocken, PA: ASTM International; 2010. 154. Olson GB, Cohen M. Mechanism for Strain Induced Nucleation of Martensitic Transformations. Journal of Less Commo n Metals 1972;28(1):107 18. 155. German RM, Suri P, Park SJ. Review: Liquid Phase Sintering. Journal of Materials Science 2009;44:1 39. 156. Kingery WD. Densification during Sintering in the Presence of a Liquid Phase 1. Theory. Journal of Applied Phys ics 1959;30. 157. Gale WF, Butts DA. Transient liquid phase bonding. Science and Technology of Welding and Joining 2004;9(4):283 300. 158. Cook GO, Sorensen CD. Overview of Transient Liquid Phase and Partial Transient Liquid Phase Bonding. Journal of Mat erials Science 2011;46. 159. Kumar G, Prabhu KN. Wetting Behavior of Solders. Journal of ASTM International 2010;7(5). 160. Martin LF. Fluxes, Solders, and Welding Alloys. Park Ridge, NJ: Noyes Data Corporation; 1973.

PAGE 193

193 161. Otsubo F, Era H, Kishitake K. S tructure and Phases in Nickel Base Self Fluxing Alloy Coating Containing High Chromium and Boron. Journal of Thermal Spray Technology 2000;9(1):107 13. 162. Mackie AC. Reflow Atmospheres in the Lead Free Era. Circuits Assembly [serial on the Internet]. 20 03. 163. Dong W, Shi Y, Xia Z, Lei Y, Guo F. Effects of Trace Amounts of Rare Earth Additions on Microstructure and Properties of Sn Bi Based Solder Alloy. Journal of Electronic Materials 2008;37(7):982 91. 164. Mavoori H, Ramirez AG, Jin S. Lead Free Uni versal Solders for Optical and Electronic Devices. Journal of Electronic Materials2002;31(11). 165. Wu C. Properties of lead free solder alloys with rare earth element additions. Materials Science and Engineering: R: Reports 2004;44(1):1 44. 166. Ramirez AG, Mavoori H, Jin S. Bonding Nature of Rare Earth Containing Lead Free Solders. Applied Physics Letters 2002;80(3):398 400. 167. Wu CML, Wong YW. Rare earth additions to lead free electronic solders. Journal of Materials Science: Materials in Electronics 2006;18(1 3):77 91. 168. Chizhi W, Maochan Y, Tzeguang C. An Investigation on the Method for Determination of Fracture Toughness K Ic of Metallic Materials with Chevron Notched Short Rod and Short Bar Specimens. In: Underwood JH, Freiman SW, Baratta FI, e ditors. Chevron Notched Specimens: Testing and Stress Analysis, ASTM STP 855. Philidelphia, PA: American Society for Testing and Materials; 1984. p. 193 204. 169. Mecholsky JJ, Barker LM. A Chevron Notched Specimen for Fracture Toughness Measurements of Ce ramic Metal Interfaces. In: Underwood JH, Freiman SW, Baratta FI, editors. Chevron Notched Specimens: Testing and Stress Analysis, ASTM STP 855. Philadelphia: American Society for Testing and Materials; 1984. p. 324 36. 170. Newman JC, Jr. A Review of Chev ron Notched Fracture Specimens. In: Underwood JH, Freiman SW, Baratta FI, editors. Chevron Notched Specimens: Testing and Stress Analysis, ASTM STP 855. Philadelphia: American Society for Testing and Materials; 1984. p. 5 31. 171. ASTM. Standard E 1304 9 7 (Reapproved 2008). Standard Test Method for Plane Strain (Chevron Notch) Fracture Toughness of Metallic Materials. West Conshohocken, PA: ASTM International; 2008.

PAGE 194

194 172. Sb Zn (Antimony Zinc) [database on the Internet]. ASM International. 1966 [cited Oct ober 2011]. Available from: http://products.asminternational.org/hbk/index.jsp 173. Cu Sb (Copper Antimony) [database on the Internet]. ASM International. 1990 [cited October 2011]. Avail able from: http://products.asminternational.org/hbk/index.jsp 174. Beech JF, Ingraffea AR. Three dimensional Finite Element Calibration of the Short Rod Specimen. International Journal of Fracture 1982;18(3):217 29. 175. Finnis MW. The Theory of Metal Ceramic Interfaces. Journal of Physics: Condensed Matter 1996;8:5811 36. 176. Humenik JM, Kingery WD. Journal of the American Ceramic Society 1954;37. 177. Naidich JV. Progress of Surface Membrane Science 1981;14:353. 178. Peden CHF, Kidd KB, Shinn ND. Journal of Vacuum Science and Technology A 1991;9. 179. Hsu DHD, Sasaki TT, Thompson GB, Manuel MV. The Effect of Aluminum Additions on the Microstructure and Thermomechanical Behavior of N iTiZr Shape Memory Alloys. Metallurgical and Materials Transactions A 2012;43A(8):2921 31. 180. Goel NC, Cahoon JR. The Al Li Mg System (Aluminum Lithium Magnesium). Bulletin of Alloy Phase Diagrams 1990;11(6):528 46.

PAGE 195

195 BIOGRAPHICAL SKETCH The author was born in Waterloo, Iowa. Growing up outside Mt. Auburn, Iowa, he attended Vinton Shellsburg High School and graduated in 2004. Deciding on Iowa State University, he chose to pursue a Bachelor of Science in Materials Engineering At ISU, he spent one su mmer in London studying at Brunel University and a semester in Australia studying at the University of Wollongong. He also spent a summer interning for US Steel, a fall interning with Caterpillar, and over 4 years working as a research hourly for Scott Chu mbley at Ames Laboratory. He graduated magna cum laude in 2009. Following, he chose the University of Florida to pursue his graduate education in Materials Science & En gineering. He achieved a Master of Science and was awarded a Science, Mathematics, and Research for Transformation (SMART) Scholarship in 2011. Graduate R esearch Fellowship and named a winner of the National Defense Science & Engineering Graduate Fellowship. He w as awarded the Howard Hughes Medical Institute (HHMI) Science for Life Graduate Award and was a world finalist in the Institute He received his Doctor of Philosophy from th e University of Florida in the summer of 2013.