University of Southern California Dissertations and Theses

Synthesis and mechanical behavior of highly nanotwinned metals

(USC Thesis Other)

Synthesis and mechanical behavior of highly nanotwinned metals

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content SYNTHESIS AND MECHANICAL BEHAVIOR OF HIGHLY NANOTWINNED METALS by Timothy Allen Furnish A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (MECHANICAL ENGINEERING) May 16, 2014 i To my father, James Furnish, for your endless love and support ii (blank page) iii Acknowledgments I would first like to thank my advisor, Professor Andrea Hodge, for her continuous support, encouragement, guidance, and patience. Her vision, drive, and work ethic have truly been inspirational and this dissertation would not have been possible without her. I would also like to thank all of the staff and faculty at USC for their constant help, support, and encouragement throughout my time at the university. In particular, thank you to Professors Kassner, Nutt, and Eliasson for agreeing to serve as my dissertation committee and a special thanks to Samantha Graves and Silvana Martinez for their incredible dedication to the AME department and its graduate students. I am grateful to the Institute of Nanotechnology at Karlsruhe Institute of Technology for their hospitality, collaboration, and friendship. I will never forget the experience and opportunities they provided me during my time in Karlsruhe. Additionally, thank you to Lawrence Livermore National Laboratories, Northwestern University, and the Institute for Applied Materials at KIT for your collaborations and support throughout this dissertation. My gratitude and love for the Hodge Materials Nanotechnology Research Group is endless. We have become a family that I cherish, and I will never forget the relationships, discussions, laughter, challenges, and victories that took place in the lab every day. They have all been such blessings in my life! I would also like to thank my family and friends for their constant love, support, and guidance. In particular, thank you to my father, James Furnish, for showing me throughout my life how to love so completely, how to work hard, and to never give up hope. I know that any good in me comes straight from him and I am eternally grateful to be blessed with such a iv wonderful dad. Also, a special thank you to my Aunt Georgie and late Uncle Frank for always believing in me and encouraging me. Finally, I would like to thank the love of my life, Audrey Weldon, for her unfailing support, love, encouragement, and patience. She is truly the greatest blessing I could ever receive in this world. v Table of Contents List of Figures ........................................................................................................................... vii List of Abbreviations................................................................................................................xiii Abstract ..................................................................................................................................... xv Chapter 1 : Introduction .............................................................................................................. 1 Chapter 2 : Background .............................................................................................................. 3 2.1. Twins and Twin Boundaries ........................................................................................... 3 2.2. Types of Twins ................................................................................................................ 4 2.3. Mechanical Behavior ....................................................................................................... 8 2.4. Synthesis of Growth Twins ........................................................................................... 19 2.5. Magnetron Sputtering .................................................................................................... 22 2.6. Summary ....................................................................................................................... 28 Chapter 3 : Experimental Methods and Materials..................................................................... 31 3.1. Synthesis by Magnetron Sputtering .............................................................................. 31 3.2. Mechanical Testing ....................................................................................................... 45 3.3. Sample Characterization ............................................................................................... 54 3.4. Summary ....................................................................................................................... 63 Chapter 4 : Mechanical Behavior of Highly Nanotwinned Cu ................................................. 65 4.1. NT-Cu Synthesis ........................................................................................................... 66 4.2. As-sputtered Microstructure .......................................................................................... 66 4.3. Stress-Strain Response in Tension ................................................................................ 67 4.4. Microstructural Evolution and Shear Band Formation ................................................. 69 4.5. NT Stability, Detwinning, and Fracture ........................................................................ 83 4.6. Summary ....................................................................................................................... 86 Chapter 5 : Synthesis and Mechanical Behavior of other NT Metals ....................................... 88 5.1. NT-Ag ........................................................................................................................... 90 5.2. NT-CuAl ...................................................................................................................... 100 5.3. Summary ..................................................................................................................... 112 Chapter 6 : Conclusions and Future Work .............................................................................. 115 Appendix A : Summary of Sputtered Cu Samples.................................................................. 120 Appendix B : Summary of Sputtered Ag Samples ................................................................. 125 Appendix C : Summary of Sputtered CuAl Samples .............................................................. 127 References ............................................................................................................................... 131 vi (blank page) vii List of Figures Figure 1. Schematic representation [7] of the general twinning process in a cubic lattice where (a) shows a perfect lattice structure and (b) represents a twinned portion of the lattice .. 3 Figure 2. Schematic representation of stacking in an fcc metal for (a) perfect stacking and (b) twinned stacking.......................................................................................................................... 4 Figure 3. Comparison of typical twinned microstructures comprised of (a) mechanical twins, (b) annealing twins, and (c) growth twins. Figures (a) and (b) adapted from [8] and [9], respectively. ................................................................................................................................ 5 Figure 4. (a) Dark-field and (b) bright-field TEM images of deformation twins in a pure Cu foil after high pressure torsion [18]. ............................................................................................ 6 Figure 5. Annealing twins in annealed 70:30 brass [23] showing straight and parallel coherent twin boundaries (CT) and steps of incoherent twin boundaries (IT) at the CT ends. ................. 7 Figure 6. TEM micrograph representing a typical growth-twinned microstructure. Figure adapted from [31]. ....................................................................................................................... 8 Figure 7. True stress-strain plots of NT-Cu of varying thicknesses ranging from (a) 96 nm to 15 nm, and (b) 15 nm to 4 nm [4]. Also included are curves for coarse grained (cg) and ultra- fine grained (ufg) Cu for reference. .......................................................................................... 10 Figure 8. High resolution TEM images showing dislocations inclined to and passing through a the TB [46]. ............................................................................................................................ 13 Figure 9. Schematic representation of the “detwinning” process by (a) detwinning partials along the TBs and (b) the collective migration of Shockley partials (incoherent TBs). Figures adapted from [53]. ..................................................................................................................... 14 Figure 10. Schematic representation showing the effects of relative TB-load directions on the dislocation mechanisms. Generally, when the load direction is perpendicular (a) or parallel (c) to the TBs (shown by blue lines in the figures), dislocations travel inclined to the TBs. When the load direction is inclined to the TBs, dislocations typically travel parallel to or along the TBs (b). Adapted from [55] ...................................................................................... 16 Figure 11. Plot showing the theoretical (solid lines) and experimental (diamonds) yield strengths of NT-Cu as a function of twin thickness [42]. ......................................................... 17 Figure 12. Critical (transitional) twin thickness at which deformation mechanisms change as a function of the grain size in NT-Cu [42]. ............................................................................... 18 Figure 13. Cross-sectional micrograph of a pulsed electrodeposited NT-Cu sample, showing roughly randomly oriented and equiaxed grains containing NTs. ............................................ 20 viii Figure 14. Cross-sectional micrograph of a magnetron sputtered NT-Cu sample (note: the arrow represents the growth direction). .................................................................................... 21 Figure 15. (a) Schematic representation and (b) photograph of the magnetron sputtering process....................................................................................................................................... 23 Figure 16. Plots of the critical nucleation radius as a function of the (a) deposition rate and (b) twin boundary energy [61] .................................................................................................. 27 Figure 17. Plot showing the volume fraction of twinned grains and the average columnar grain size as a function of the deposition rate [29]. .................................................................. 28 Figure 18. Schematic representation of one of the magnetron sputtering chambers used in this study. ......................................................................................................................................... 32 Figure 19. Photograph of a planar magnetron sputtering source. Adapted from [72]. ........... 33 Figure 20. Photograph showing various targets used for planar magnetron sputtering. Note the ring-shaped trench that develops during sputtering. ........................................................... 34 Figure 21. Photograph showing the cooling system designed for the magnetron sputtering chamber. .................................................................................................................................... 35 Figure 22. Schematics of the substrate holders designed and used for the synthesis process of NT metals. ................................................................................................................................. 36 Figure 23. Schematic showing an interruption process in which the substrates are rotated using a turn-table system. ......................................................................................................... 41 Figure 24. Photograph showing the application of thermally conductive Ag paste between the substrate holder (a) and the substrate wafer back (b). .............................................................. 43 Figure 25. Dimensions (in mm) of the dogbone shaped specimens used for tensile testing of the NT metallic foils. ................................................................................................................ 47 Figure 26. Schematic of the tensile specimen die designed for cutting and shaping tensile dog-bone specimens. ................................................................................................................. 48 Figure 27. (a) as-sputtered 50mm dia foil cut into multiple strips for tensile dogbone specimen preparation using the tensile specimen die shown in Fig. 26. (b) Example of post- cut dogbone specimen. .............................................................................................................. 48 Figure 28. Photograph of the Instron tensile tester used for testing the NT foils. ................... 49 Figure 29. Custom jig/stage used to align and secure the NT foils to the testing machine fixtures. ..................................................................................................................................... 50 Figure 30. Screen-shot showing the tracking of paint markings by the video extensometer. . 51 ix Figure 31. (a) Schematic of the Vickers pyramidal diamond tip and (b) SEM image showing a set of Vickers indents on the surface of a NT foil. ................................................................. 52 Figure 32. Picture of the LECO-LM100 micro-indenter used in this study for the indentation of the NT foils. .......................................................................................................................... 53 Figure 33. Photographs showing the Ambios XP-2 used in the current study. ....................... 55 Figure 34. Step-profile used to measure the film/foil thickness directly from the profilometer. .............................................................................................................................. 56 Figure 35. Photograph of inside the Rigaku XRD system. ...................................................... 57 Figure 36. Example of a plot revealing various (111) out-of-plane textures in magnetron sputtered foils of different microstructures. .............................................................................. 58 Figure 37. FIB image of the surface of an ufg metallic sample showing the orientation contrast due to channeling effects of the ion beam. .................................................................. 60 Figure 38. FIB images of a cross-sectional trench (a) top-view and (b) cross-sectional view prepared for microstructural evaluation of the material cross-section. ..................................... 61 Figure 39. FIB channeling contrast image of the cross-section of a NT foil prepared with a high energy ion beam. ............................................................................................................... 61 Figure 40. (a) TEM specimen (highlighted by red box) prepared by FIB specimen preparation techniques; (b) channeling contrast image of the TEM specimen. ........................................... 62 Figure 41. (a) FIB cross-sectional micrograph [90] and (b) TEM cross-section of the as-sputtered highly NT-Cu microstructure [31]. Arrows show the growth direction. ............. 67 Figure 42. Engineering stress-strain curves of NT-Cu tested at various strain rates (10 -4 -10 -2 /s) at RT and 77K. ..................................................................................................... 68 Figure 43. Representative optical micrographs of fractured dogbone specimens tested in tension at (a) RT and (b) 77K at a rate of 10 -4 /s....................................................................... 68 Figure 44. Representative of a tensile specimen midsection deformed at 10 -4 /s at RT. Critical locations in the deformed sample were analyzed using cross-sectional FIB micrographs: (a) Region near the highest plastic deformation; (b) area close to the fractured deformation band showing slight grain rotation; and (c) area along a deformation band. Arrows in (a-c) show the direction of growth..................................................................................................... 71 x Figure 45. Representative midsection of a tensile specimen deformed with 10 -4 /s strain rate at 77K. Critical locations in the deformed sample were analyzed using cross-sectional FIB micrographs: (a) Region near the highest plastic deformation showing some grain growth; (b) area close to the fractured zone, showing slight grain rotation; and (c) area along a deformation band. Arrows in (a-c) show the direction of growth. .......................................... 72 Figure 46. (a) Optical micrograph of the gage section of a representative tensile sample tested at 10 -4 /s that was stopped immediately after the yield point. (b) A magnified optical micrograph of the shear band across the specimen width. (c) FIB cross-section micrograph of a section within the shear band region [location of FIB cut marked by double-headed arrow in (b)]............................................................................................................................................. 74 Figure 47. (a) cross-sectional FIB micrograph showing the original as-prepared microstructure; (b) schematic representation of the micro-focus X-Ray scans performed on the fractured samples; (c) representative diffraction pattern for the as-prepared sample (before tensile testing). .......................................................................................................................... 76 Figure 48. NT-Cu gauge section (one half shown) tensile tested at RT and 10 -4 /s. Representative microfocus X-ray DPs [(a1)-(c1)] and cross-sectional FIB micrographs [(a2)-(c2)] obtained at (a) 50 µm, (b) 150 µm, and (c) 1 mm from the fracture edge. Note that locations (a) and (b) are both in Region I. ................................................................................ 78 Figure 49. NT-Cu gauge section (one half shown) tensile tested at 77K and 10 -4 /s. Representative microfocus X-ray DPs [(a1)-(c1)] and cross-sectional FIB micrographs [(a2)-(c2)] obtained at (a) 50 µm, (b) 700 µm, and (c) 2 mm from the fracture edge. ............. 80 Figure 50. (a) Optical micrograph of a gage section after fracture for a representative tensile sample tested at 10 -4 /s. (b) Planar view of the FIB trench prepared for cross-section imaging. (c) FIB cross-section centered at the trench where a surface depression line near the fractured edge shows detwinning. (d) Representative magnified FIB image of a detwinned region. .... 85 Figure 51. Representative FIB cross-sectional micrographs of as-sputtered Ag foils containing approximately (a) 10%, (b) 40%, and (c) 100% highly NT columnar grains. ........ 92 Figure 52. Representative TEM cross-sectional image of a portion of a NT grain in the 100% NT grained Ag. ......................................................................................................................... 92 Figure 53. (a) SEM surface micrographs of Vickers indents and (b) FIB cross-sectional micrographs underneath the indents of the foils containing (1) 10%, (2) 40% and (3) 100% NT grains. ................................................................................................................................. 94 Figure 54. True stress-true strain curves of the Ag foils containing various volume fractions of highly NT grains. .................................................................................................................. 96 Figure 55. Representative optical micrograph of a tensile tested Ag sample showing fracture in the middle of the reduced section. ........................................................................................ 96 xi Figure 56. (a) top view and (b) cross-sectional view of the fractured dogbone specimens containing (1) 10%, (2) 40%, and (3) 100% volume fraction of highly NT grains. ................. 98 Figure 57. (a) FIB cross-sectional micrograph and (b) TEM micrograph of an as-sputtered highly NT-CuAl foil. Note, the arrow shows the growth direction. ...................................... 101 Figure 58. FIB cross-sectional micrograph of the as-sputtered ufg-CuAl foil. Note, the arrow shows the growth direction. .................................................................................................... 102 Figure 59. SEM surface images showing the Vickers indents in the (a) ufg-CuAl and (b) NT-CuAl foils. ................................................................................................................... 103 Figure 60. FIB cross-sectional micrograph underneath a Vickers indent in the highly NT-CuAl foil showing a detwinned crack network. ............................................................... 105 Figure 61. FIB cross-sectional micrograph underneath a Vickers indent in the ufg-CuAl.... 105 Figure 62. Engineering stress-strain curves for multiple NT-CuAl and ufg-CuAl foils. ....... 106 Figure 63. Representative optical micrograph of a tensile tested NT-CuAl sample showing fracture in the middle of the reduced section. ......................................................................... 106 Figure 64. Collection of data for flow stress as a function of grain size [129-134], including the ufg-CuAl and NT-CuAl foils in this study, where d=200nm is used for the grain size of the ufg-CuAl, and d=λ=3nm is used for the NT-CuAl. .......................................................... 109 Figure 65. FIB micrographs [(a) lower magnification and (b) higher magnification image] showing a cross-section trench prepared on the fracture surface of a tensile tested highly NT-CuAl foil. .......................................................................................................................... 110 Figure 66. FIB micrographs showing a cross-section trench prepared along a crack near the fractured surface of the tensile tested highly NT-CuAl foil. ................................................... 111 xii (blank page) xiii List of Abbreviations bcc body-centered cubic CTB coherent twin boundary DP diffraction pattern DPs diffraction patterns fcc face-centered cubic FIB focused-ion beam GB grain boundary GBs grain boundaries hcp hexagonal close packed HEMD high energy micro-diffraction ITB incoherent twin boundary MD molecular dynamics nc nanocrystalline NT nanotwinned NTs nanotwins RT room temperature SEM scanning electron microscope SFE stacking fault energy TBs twin boundaries TEM transmission electron microscope ufg ultra-fine grained XRD X-ray diffraction xiv (blank page) xv Abstract Metals containing high densities of nanoscale growth twins have shown potential as an alternative to nanocrystalline (nc) metals. Some reports on nanotwinned (NT) Cu, for example, have revealed that high strengths, comparable to nc metals, can be produced while improving other properties such as ductility, thermal stability, and mechanical stability. However, since the synthesis of highly NT metals is not yet fully understood, most studies have been limited to only NT-Cu; therefore, it is unclear if similar trends in the mechanical performance of NT-Cu can be extended to other NT systems. In addition, the majority of the work on determining the mechanical properties and deformation mechanisms of NT metals has been produced using nano-indentation; consequently, the larger-scale deformation mechanisms (i.e. those observed using μm to mm scale mechanical testing) of NT metals is still relatively unexplored. In this study, the synthesis of a variety of NT metals was investigated using magnetron sputtering, wherein the deposition conditions were systematically changed in order to produce highly columnar grained NT Cu, Ag, and Cu-6wt%Al (CuAl). The Cu and Ag served as representatives for moderate to low stacking fault energy (SFE) pure fcc metals, respectively, while the CuAl allowed for the study of NTs in low SFE alloyed fcc systems. It was observed that the relatively low SFE of these materials facilitated NT growth during sputtering. In addition, magnetron sputtering enabled the fabrication of relatively thick (26-170 μm) free- standing foils with low initial defects, internal stresses, and dislocation densities. The mechanical performance of the NT metals was evaluated using multiple testing techniques under various conditions, including tensile tests and micro-indentation. The NT xvi microstructures and deformation behaviors were evaluated using advanced characterization methods, including focused ion beam (FIB) and high energy microdiffraction (HEMD). Tensile results of the columnar grained NT-Cu tested at room temperature (RT) and 77K showed increases in both strength and ductility at 77K, compared to the foils tested at RT. Additionally, the deformation within the foils was observed to be primarily by shear banding at both testing temperatures. FIB and HEMD analysis led to the formulation of possible mechanisms leading to the shear band formation and the observed increase in ductility at 77K. The mechanical stability of the columnar grained NT structure was also examined under the different loading conditions. The Ag and CuAl foils were synthesized with varying volume fractions of NT columnar grains, ranging from 0% to 100% vol%. Results of tensile tests and Vickers micro- indentation for the Ag foils showed that the addition of NTs led to enhanced strengthening, similar to the NT-Cu foils; however, the ductility in the foils decreased with increasing volume fraction of NT grains. Overall, the deformation and NT stability within the 100% NT foils were consistent with the results of the NT-Cu. Conversely, the CuAl showed a general insensitivity to the presence of NTs, in which minimal differences in the mechanical properties between NT-CuAl and non-NT CuAl foils were found. Tensile tests and Vickers micro-indentation revealed similar yield strengths, hardness, and % elongation between the NT and non-NT CuAl. Plastic deformation was limited in both sets of CuAl foils, in which the samples showed semi-brittle fracture during tension tests and a general propensity for plastic instabilities and cracking. 1 Chapter 1 : Introduction Nanostructured materials have been the subject of extensive research during the last three decades, as their nano-scale features often lead to unique mechanical properties. It has generally been shown that when the microstructure of a material is altered, especially at the nano-scale, the material properties significantly change. For instance, some of the most commonly studied nanostructured materials, nanocrystalline (nc) metals, have shown superior strengths, up to 10-15 times that of their coarse-grained counterparts, when their grain size, d, is reduced to the nano-regime (d < 100 nm) [1]. However, the overall use of nc metals has been somewhat restricted, partly due to their poor thermal stability and limited ductility. For example, recrystallization and grain growth have been shown to occur at relatively low temperatures, and in some cases, even at room temperature (RT) [2]. In addition, the ductility in nc metals is typically confined to a few percent, even in metals that normally show a 40- 60% elongation-to-failure at conventional grain sizes [3]. One approach that has the potential to achieve benefits similar to those of nc metals, while improving on some of their limitations, is the use of nano-scale twin boundaries (TBs). Recent studies have revealed that introducing a high density of nanotwins (NTs) into pure Cu can lead to ultra high strengths, similar to those produced in nc-Cu, while maintaining other desirable properties, such as ductility [4] and thermal stability [5]. Although twinning phenomena in metals have been observed and studied in the past, particularly in the case of deformation twins, the development of nano-scale TBs that are grown into the material (i.e. growth-NTs) through bottom-up processing approaches is a relatively new topic. Therefore, the synthesis of NTs during processing and their corresponding effects on the material behavior are not yet fully understood. Thus far, the development of highly NT structures has 2 been mostly limited to Cu and much of the data on their mechanical properties have been obtained using small-scale testing techniques, such as nano-indentation. These limited, and sometimes disparate, data on the synthesis and deformation of NT metals have left many questions still unanswered. For example, can some of the trends observed for NT-Cu be repeated and extended to other metals? Are there practical means for bulk synthesis and control of NT structures? What is the overall role of the growth-NTs on the larger-scale deformation behavior and mechanical performance? The current work serves to address some of the above questions by examining the synthesis of NT structures in a variety of metals and exploring their overall deformation behavior and structure-property relationships. Specifically, processing conditions were systematically changed during magnetron sputtering to develop high densities of growth-NTs in Cu, Ag, and CuAl metallic foils. The foils were then mechanically tested under various conditions and the deformation, fracture, and microstructural stability was evaluated using advanced characterization and imaging techniques. This study has generated additional knowledge about the roles of growth-NTs in the overall deformation response of NT metals, and in controlling the overall NT distribution in order to tailor the mechanical properties. The enhanced understanding of these nanostructured metals will further develop their potential use for advanced engineering applications. 3 Chapter 2 : Background 2.1. Twins and Twin Boundaries Twinning is generally described as a portion of a crystal that takes on a definite, symmetrical orientation relative to the rest of the lattice, where the twinned portion is a mirror image of the parent crystal [6]. As represented in the schematic in Figure 1, the atoms in a section of a simple cubic lattice (Fig. 1a) have been shifted such that across a boundary, their orientation is a direct reflection of the original crystal (shown as the “twin” in Fig. 1b). The plane in which the shift has occurred is considered the twinning plane or the twin boundary (TB), and the “shifted crystal” is considered the twin. Figure 1. Schematic representation [7] of the general twinning process in a cubic lattice where (a) shows a perfect lattice structure and (b) represents a twinned portion of the lattice In a face-centered cubic (fcc) system, there are three unique positions in which a plane of atoms can occupy, commonly referred to as the A, B, and C positions. Assuming there is one A-plane of atoms, the atoms stacked on top of that plane can occupy either the B or C positions. In a perfect fcc crystal, as represented in Figure 2a, the order of the stacking is a repeated ABCABCABC… sequence. Conversely, Figure 2b shows an example of twinning in an fcc system, in which the stacking appears as ABCACBABC. In this case, the twin is 4 defined as the “ACBA” portion of the crystal and the twin boundaries (TBs) being the second and third A layers (from bottom-up), across which the shifts have occurred. The twin thickness in a material is defined as the distance between the two TBs. Figure 2. Schematic representation of stacking in an fcc metal for (a) perfect stacking and (b) twinned stacking 2.2. Types of Twins There are three different classifications of twins: mechanical (or deformation) twins, annealing twins, and growth twins. In each case, the twin boundaries and the atomic stacking of the twins across the boundaries may be identical, as described in Figure 2b for fcc metals. However, the difference lies in the means by which the twinning occurs (i.e. how the atomic “shifts” occur). As the names imply, mechanical/deformation twins are formed from plastic deformation, annealing twins are formed during annealing, and growth twins are formed during “ground-up” material growth. For a quick comparison, Figure 3 shows typical microstructures for all three types. Although the overall focus of this work is on growth- twinned materials, the other two mechanisms will be briefly discussed, followed by a more detailed description for growth twins. 5 Figure 3. Comparison of typical twinned microstructures comprised of (a) mechanical twins, (b) annealing twins, and (c) growth twins. Figures (a) and (b) adapted from [8] and [9], respectively. 2.2.1. Mechanical/Deformation Twins Mechanical twinning is a process of cooperative shear movement of atoms into the specific, low-energy twin arrangements [10] (i.e. atoms are physically sheared into twin positions). This type of deformation is in contrast to the much more commonly observed slip by dislocation movement. Typically, the stress required to cause twinning is greater than that to promote slip and therefore, slip is usually the preferred deformation mechanism. However, for certain metals and/or under certain conditions, the stress required for slip may be higher than that for twinning to occur, and twinning may become the dominant means of deformation. For example, the low symmetry associated with hexagonal close packed (hcp) metals makes it difficult for dislocations to move and higher stresses are typically required for slip [10]. This enables other mechanisms, such as deformation by twinning, to be activated in these materials and therefore, mechanical twins are commonly found in deformed hcp metals [11]. In addition, mechanical twins have been shown to develop as a means of plastic deformation in some body-centered cubic (bcc) metals when deformed at low temperatures [12] and/or high strain rates [13]. In the case of fcc metals, mechanical twinning is not as common as for hcp or bcc metals, although they have been observed in some fcc 6 metals after severe plastic deformation [12, 14-16]. Figure 4 shows transmission electron microscope (TEM) images of deformation twins [(a) dark-field, (b) bright field] in pure Cu deformed by high pressure torsion. The wedge-shaped, lenticular twins shown here are commonly observed in mechanical twinning [17]. Figure 4. (a) Dark-field and (b) bright-field TEM images of deformation twins in a pure Cu foil after high pressure torsion [18]. 2.2.2. Annealing Twins Annealing twins are those that develop during the various stages of the annealing process. Although there are many different theories and models to attempt to explain the twinning process during annealing [19-23], twinning can mostly be attributed to energy minimization, whether to reduce strain energy and dislocation densities [22] and/or to reduce the overall interfacial energies [19]. Annealing twins have been widely observed in many fcc metals, especially those of intermediate or low stacking fault energy (SFE) [23]. This type of twinning is strongly dependent on the material, inclusions, previous strain history, and the annealing conditions [24]. Figure 5 shows a detailed example of annealing twins in a 70:30 brass [23], in which straight and parallel coherent twin boundaries (CTBs) are found 7 throughout the center grain, with incoherent twin boundaries (ITBs) at the CTB ends (commonly observed in annealing twinned materials). Figure 5. Annealing twins in annealed 70:30 brass [23] showing straight and parallel coherent twin boundaries (CT) and steps of incoherent twin boundaries (IT) at the CT ends. 2.2.3. Growth Twins Growth twins are those which form during the growth of the material through various processing techniques such as physical vapor deposition and electrodeposition (both will be discussed in section 2.4). Generally, this type of twinning occurs due to a layer-by-layer deposition of <111> planes of atoms (i.e. close-packed planes) normal to the deposition direction [25]. As the planes of atoms are being stacked, as depicted previously in Figure 2, shifts in the normal fcc stacking may occur leading to twin boundary (TB) formation. Growth twinning is most commonly found for fcc metals, especially those of low to moderate SFE [26]. Since the study of growth twins in metals is relatively new, only a select amount of materials have been explored, such as Cu [27-28], stainless steel [29], and Ag [30]; thus, the 8 exact mechanisms leading to twin growth is still unclear. Figure 6 shows a typical example of a growth-twinned microstructure [31], where straight, parallel twins are found throughout the entire grain widths. The manuscript from this point forward will focus only on growth twins. Figure 6. TEM micrograph representing a typical growth-twinned microstructure. Figure adapted from [31]. The arrow shows the growth direction. 2.3. Mechanical Behavior Studies have shown that metals containing nanoscale growth twins may potentially possess a unique combination of properties, such as high strength, ductility, and thermal stability [4-5]. These attributes have generally been attributed to the symmetry and coherency of the TB. However, the detailed mechanisms involved and reasons for such properties are an ongoing topic and require much more work. In particular, although strengthening is often found in various nanotwinned (NT) metals, the combination of strength and ductility has, thus far, only been observed for NT-Cu. Therefore, it is currently unclear if these properties can 9 extend to other metals. In addition, until all deformation mechanisms are well understood and characterized, it will be difficult to model these systems to predict and control their mechanical behaviors. In the past few years, however, progress has been made on beginning to understand these materials. In this section, some of the trends in the mechanical performance previously observed for NT-Cu will be discussed in greater detail. Furthermore, a few of the deformation mechanisms and key microstructural features that have been linked to the enhanced properties will be introduced. 2.3.1. Strength and Ductility One of the most commonly observed trends in NT metals is a significant hardening/strengthening affect with the addition of high densities of TBs. Examples include studies on NT Cu [4, 28, 32], 330 stainless-steel [33], and Ni-Co [34], among others. In general, TBs can serve as dislocation obstacles due to the discontinuity of slip across the boundary, similar to traditional high angle grain boundaries [35]. Much like the enhanced strengthening with grain refinement in nc metals, NT metals have shown strengthening with decreasing twin thickness) [36]. In fact, a number of studies have revealed that the hardness of NT metals as a function of the twin thickness falls on the same Hall-Petch plot as their nc counterparts as a function of grain size [37]. One reason for this is that there is a limited capacity for dislocation multiplication and pileup within the grain and twin interiors, which increases the required stresses for slip to occur [38]. This mechanism is well-studied and relatively understood for both nc and NT metals for a certain range of grain size and twin thickness, respectively. 10 Another result observed in some NT-Cu is that the ductility can also increase concurrently with the strength as the twin thickness is reduced. Figure 7 shows tensile stress- strain curves produced by K. Lu et al. for NT-Cu samples with various twin thicknesses [4]. [Note: in this plot, the labels of “nt-x” denote the twin thickness, t, in nanometers (for example, “nt-15” for t = 15 nm, “nt-35” for t = 35 nm, etc).] For reference, tensile results for ultra-fine grained (ufg) and coarse grained (cg) Cu, both without NTs, are included. Figure 7. True stress-strain plots of NT-Cu of varying thicknesses ranging from (a) 96 nm to 15 nm, and (b) 15 nm to 4 nm [4]. Also included are curves for coarse grained (cg) and ultra-fine grained (ufg) Cu for reference. The yield strength of the NT-Cu increases with decreasing twin thickness from 96 nm to 15 nm, showing a maximum yield strength of almost 1 GPa at t = 15 nm (~10 times that of cg Cu); below 15 nm, however, softening occurs. Additionally, the ductility (measured, in this case, as the elongation-to-failure) increases monotonically with decreasing twin thickness across the full range from 96 nm to 4 nm. This result is in contrast to most nc metals, in which any strengthening that is produced by grain refinement is typically coupled with a 11 dramatic loss of ductility [39]. As mentioned earlier, the exact mechanisms responsible for these observed trends in strength and ductility, in addition to the softening effects below a certain thickness, are not fully understood. Many studies have been underway to explain this behavior and have generally revealed that the material properties are highly dependent on, and very sensitive to, many factors. These include the types of dislocations, types of twin boundaries (coherent or incoherent), and the respective orientations between dislocations and TBs; some of these will be summarized and discussed in the following section. 2.3.2. Dislocation Interactions with Twin Boundaries As mentioned earlier, many factors influence the mechanical properties of NT metals. Studies have shown that even down to very small twin thicknesses, on the order of a few nanometers, bulk dislocation activity continues to play an important role in the deformation, as opposed to any grain boundary (GB) mechanisms (e.g. GB sliding) [40-41]. There are many types of dislocations, slip systems, dislocation-dislocation interactions, dislocation- boundary interactions, nucleation mechanisms, surface affects, etc. that can alter the material behavior. Therefore, it is generally difficult to fully determine the dominant mechanisms in these materials. However, many experimental and molecular dynamic (MD) simulation studies have revealed multiple competing mechanisms that can dictate the overall deformation. Among these is the competition between dislocations oriented inclined to the TBs, and those parallel to them [42]. Dislocations inclined to TBs TBs can obstruct dislocation motion due to the discontinuity of the slip across the boundary, which has been described as one of the main reasons for the enhanced strengths in 12 NT metals. However, this particular obstruction can only exist if the dislocation is travelling in a slip system that is inclined to the TB, rather than parallel to it (i.e. if the dislocation direction is parallel to the TB, the slip will remain continuous until a different obstacle is encountered, e.g. a normal GB). An interesting observation found when inclined dislocations interact with a TB is that, although the dislocation motion is hindered, it is not completely stopped. Depending on the material, the type of dislocation, and the direction and magnitude of the applied stress, the dislocation can either cross-slip onto the TB or can transmit across it, thus further advancing plastic flow in the material [43-45]. In the NT systems most often studied, in particular NT-Cu, the most commonly observed mechanism when dislocations are inclined to the TBs is the transmission of partial dislocations across the boundary [40]. This was directly observed by Wang et al. [46] in an in situ high resolution TEM study on NT-Cu, which is represented in Figure 8. In this case, an extended dislocation approaching the TB (Fig. 8a) combined into a perfect dislocation and was absorbed into the TB (Fig. 8b). With further stress, the absorbed dislocation dissociated again and the slip continued into the adjacent twin (with the order of the partials reversed), and leaving behind a partial GB dislocation that then travelled along the TB (Fig. 8c). This process is schematically summarized in Figure 8d. This result has also been observed by other experimental and simulation works [41, 43, 47], and is significant because it confirms continued plasticity across TBs, even after the dislocation is initially blocked. Therefore, when dislocations inclined to the TBs are active, strengthening can be achieved due to the inhibition of dislocation motion, accompanied with some plasticity. 13 Figure 8. High resolution TEM images showing dislocations inclined to and passing through a the TB [46]. Dislocations parallel to TBs In most NT metals, the TB length (approximately equal to the grain size) is typically much larger than the twin thickness. Therefore, the strengthening that occurs when dislocations are inclined to the TBs may not be as prevalent when the dislocations are parallel to them, whereby a dislocation might travel relatively uninhibited [48]. However, since the travel of dislocations is primarily responsible for the total plastic strain, this relatively unrestrained motion leads to an increased overall plasticity (i.e. more ductility) in the NT metal. Furthermore, when dislocation motion is promoted along or parallel to the TBs, additional effects concerning the plasticity are observed, including the growing and shrinking of twins in the material. This “detwinning” process has been shown to significantly contribute to the plastic strain in NT-Cu, and is generally attributed to either the motion of 14 partial GB dislocations (a.k.a. “detwinning partials”) along the TBs [36, 48] and/or the collective migration of stacked Shockley partial dislocations (which make up ITBs), often found at TB ends [49-53]. The detwinning process is represented schematically in Figure 9 [53]. Figure 9. Schematic representation of the “detwinning” process by (a) detwinning partials along the TBs and (b) the collective migration of Shockley partials (incoherent TBs). Figures adapted from [53]. The deformation mechanisms that arise in NT metals from the different dislocation- TB orientations have been shown to greatly affect the plastic response and corresponding mechanical properties. As previously described, strengthening is often attributed to dislocations inclined to the TBs, while increased plastic flow and ductility may be promoted when dislocations are parallel to them. Therefore, in order to tailor NT metals for a desired material response, it is important to understand how these mechanisms might be controlled. Although this is an area of active research and requires further work, some key factors that have been shown to dictate the activation of one mechanism over the other will be discussed. 2.3.3. Controlling the Deformation Mechanisms A variety of dislocation-TB interactions can occur depending on their relative orientations with each other. In order to control the deformation behavior of NT metals, it is 15 first necessary to explore the means by which these mechanisms can be either promoted or suppressed. In general, studies have revealed that these mechanisms are very sensitive to factors such as the orientation of the TBs relative to the loading direction, the twin thickness, and the grain size. TB orientation vs. load direction In NT metals, it has been observed that the dislocation activity is highly dependent on the orientation of the TBs relative to the load direction [48, 54-55]. The observation of “hard” and “soft” load directions in NT metals has been explained by the well-known Schmid Law that relates the applied stress to the resolved shear stress:     cos cos  R Equation (1) where R  is the resolved shear stress, σ is the applied stress,  is the angle between the applied force and the normal to the slip plane, and λ is the angle between the slip plane direction and the applied force. When the term   cos cos (known as the Schmid Factor) is maximized, the resolved shear is also maximized and slip will primarily occur in that direction. In the case of NT metals, when the load direction is parallel or perpendicular to the TBs, the Schmid Factor is maximized in directions inclined to the TBs, and thus slip will occur in those directions – this is referred to as the “hard” direction, since more dislocation- TB interactions will occur. Conversely, when the load direction is inclined to the TBs, the maximum shear stress will be in directions parallel to the TBs and therefore, dislocations would be active in this “soft” direction. These mechanisms are summarized schematically in Figure 10. These results may lead to the possibility of tailoring the mechanical response and properties of NT by controlling the TB orientations relative to the load direction. 16 Figure 10. Schematic representation showing the effects of relative TB-load directions on the dislocation mechanisms. Generally, when the load direction is perpendicular (a) or parallel (c) to the TBs (shown by blue lines in the figures), dislocations travel inclined to the TBs. When the load direction is inclined to the TBs (b), dislocations typically travel parallel or along the TBs. Figure adapted from [55] Twin Thickness As was previously shown in Figure 7, the mechanical properties of NT metals can be very sensitive to the TB spacing. The underlying reasons for this include the promotion and/or suppression of certain deformation mechanisms (i.e. dislocations either inclined or parallel to the TBs) as the twin thickness is varied. It has been shown for NT-Cu that when the twin thickness is above some critical value, dislocations inclined to the TBs are most dominant and control the deformation [41]. At the same time, however, other mechanisms are still present but may have a lower influence on the deformation. For example, the intersections between TBs and grain boundaries (GBs) have been shown to be effective partial dislocation sources. Once generated, these partial dislocations can travel along the TBs, thus contributing to the overall plastic strain in the material. When the twins are thick (i.e. much greater than the critical spacing where mechanisms begin to change), the contributions of these partial TB dislocations are not particularly significant. However, as the twin thickness is reduced, the overall density of TBs within the grain is increased and more TB-GB dislocation sources are created [42]. The reduction in the twin thickness also reduces the space for dislocation multiplication within the twin interiors and, therefore, fewer inclined (a) (b) (c) 17 dislocations are generated [56]. This combination of effects leads to the transition from one mechanism to the other with decreasing twin thickness. Figure 11 shows this point, as the yield stress (obtained from both experiments and MD simulations) increases with decreasing twin thickness [42] down to a certain critical value. Below this critical thickness, softening occurs with further reduction of twin thickness due to the suppression of dislocations inclined to the TBs and the promotion of dislocations parallel to them. These results reveal that the TB spacing can also be used to dictate the dominant deformation mechanisms and may provide additional means to control the overall plastic response of NT metals. Figure 11. Plot showing the theoretical (solid lines) and experimental (diamonds) yield strengths of NT-Cu as a function of twin thickness [42]. Grain Size The grain size has also been found to affect the mechanical behavior of NT metals. Unlike nc metals, where the grain size is the major factor in the mechanical properties, the deformation behavior of NT metals is indirectly influenced by the grain size. It has been observed in NT-Cu that as the grain size is varied, the point at which one deformation mechanism becomes dominant over another also changes. For example, a few studies have 18 revealed that the critical twin spacing in which softening begins (as previously shown in Figs. 7 and 11), scales with the grain size [41-42, 56]. This is represented in Figure 12 in which the transitional twin thickness is shown as a function of the average grain size [42]. This study reveals that the critical twin thickness where the dominant deformation mechanism changes scales by the square-root of the grain size, i.e. d 1/2 , and that the maximum strength achievable for NT-Cu scales with d -1/2 . The primary reason for this is that the length of the TB is typically approximately equal to the grain size; thus, when dislocations traveling parallel or along the TBs are active, the grain size (or twin length) is the main limiting factor for plastic deformation. These results are important because the understanding of this indirect grain size effect on the mechanical properties may allow for the development of optimum grain-twin size relationships for a desired set of material properties. Figure 12. Critical (transitional) twin thickness at which deformation mechanisms change as a function of the grain size in NT-Cu [42]. In summary, some of the trends in the mechanical properties of NT-Cu have been discussed, particularly in terms of the potential combination of strength and ductility. These trends have led to many experimental, simulation, and modeling studies which have generally 19 revealed that dislocation activities play a major role in the deformation of NT metals, even down to twin thicknesses of a few nanometers. Various competitions in the dominant dislocation mechanisms in NT metals exist, one of which is the competition between dislocations travelling inclined or parallel to TBs. Depending on factors such as the TB orientation relative to the loading direction, twin thickness, and grain size, these mechanisms may be either promoted or suppressed. 2.4. Synthesis of Growth Twins The fabrication of growth-twinned metals is a relatively new topic and, therefore, there is a limited amount of data on the synthesis of these materials. This is likely due to the array of challenges associated with the processing, including the complexities of the growth twin development and the intricate interdependence of processing variables. These difficulties have mired a complete understanding and, to date, a comprehensive study concerning the processing of such materials has yet to be developed. Only two predominant approaches of developing highly growth-twinned metals have emerged, both with various levels of technical and practical difficulties; these are the methods of pulsed electrodeposition and magnetron sputtering. In this section, these two processing routes will be summarized, including any potential advantages or disadvantages related to the development of highly NT structures. 2.4.1. Growth Twins by Pulsed Electrodeposition Electrodeposition is a metallic coatings process in which metal ions (the coating species) in a solution are deposited onto an electrode (the substrate); the “growth” of the material is the build-up of the metal atoms on the substrate. In the case of pulsed 20 electrodeposition, the current is pulsed on and off in a controlled manner, which generally allows for large peak current densities and relatively high deposition rates [35, 57]. This approach has become one of the main means of manufacturing nc metals; however, its use in NT metals is somewhat of a new concept and is not yet fully understood. Nonetheless, this method has been adapted for the synthesis of certain NT metals, such as NT-Cu, and has produced some key experimental results. For example, the reports of enhanced strength and ductility by K. Lu et al., described earlier in section 2.3.1., used pulsed electrodeposition to create a variety of NT-Cu structures. Figure 13 shows one of these structures, which reveals a randomly oriented equiaxed ufg microstructure with high densities of embedded NTs. Figure 13. Cross-sectional micrograph of a pulsed electrodeposited NT-Cu sample, showing roughly randomly oriented and equiaxed grains containing NTs. There are a few advantages of using pulsed electrodeposition for development of NT metals, compared to other techniques. One of these advantages is that relatively high deposition rates can be achieved which may lead to higher probabilities of twin growth [57]. In addition, the randomly oriented microstructure may be useful for certain applications where an average isotropy in the material is desired. However, the details of NT growth using 21 pulsed electrodeposition are still not well understood and the choice of materials is often limited to mostly pure metals and only a few alloys [58]. 2.4.2. Growth Twins by Magnetron Sputtering Sputtering is a physical vapor deposition technique where a target material (the coating species) is bombarded by gas ions. Due to the collisions between the gas ions and target, the target atoms are ejected from the surface and produce a physical vapor of the target material is produced, which is then used to coat a substrate. The process of magnetron sputtering uses a set of permanent magnets mounted below the target surface, which increases the ionization rate of the gas and leads to a more efficient process [59]. Due to the wide variety of materials that can be sputtered, this process has recently been used for the synthesis of NT-Cu, 330 stainless-steel, Ag, and various NT multilayers [30, 32, 60-61]. The microstructures of NT metals produced by magnetron sputtering often consist of columnar- type grains that extend through the thickness of the material with embedded NTs oriented perpendicular to the growth direction, as represented in Figure 14. Figure 14. Cross-sectional micrograph of a magnetron sputtered NT-Cu sample (note: the arrow represents the growth direction). 22 The main advantages of using magnetron sputtering over electrodeposition techniques are: (i) there are few restrictions on the types of materials that can be sputtered and substrates that can be used. Since sputtering is purely a physical process, almost any metal (including alloys) or ceramic can be sputtered, and the stoichiometry and alloy content is typically preserved (relative to the target material) in sputtering alloys; (ii) the process is performed under high [10 -3 -10 -5 Pa (10 -5 -10 -7 Torr)] or ultra-high [10 -5 -10 -7 Pa (10 -7 -10 -9 Torr)] vacuum and thus, oxidation and other contamination is not a major concern; (iii) the highly aligned microstructures may allow for a detailed understanding of the orientation-dependent deformation mechanisms. In addition, the orientation of the TBs can be better controlled and adapted for expected loads in particular applications. Therefore, in the current study, magnetron sputtering was exclusively used for the synthesis of growth-NTs in a variety of metals. In the following section, magnetron sputtering is discussed in greater detail and some of the key factors in the synthesis of NT structures by this technique are described. 2.5. Magnetron Sputtering As previously mentioned, magnetron sputtering is a physical vapor process commonly used for thin film deposition. In this section, magnetron sputtering will be presented in more detail, including a more thorough description of the overall process. Additionally, some of the commonly used process variables and their potential effects on the formation of highly NT structures will be briefly discussed. A schematic representation and photograph of the process are shown in Figures 15a and b, respectively, which will be referenced throughout this section. 23 Figure 15. (a) Schematic representation and (b) photograph of the magnetron sputtering process. 2.5.1. Process Details 1. A cathode made up of the material to be sputtered (called the “target”) is placed in a high (or ultra-high) vacuum chamber [10 -3 -10 -7 Pa (10 -5 -10 -9 Torr)] which is then back-filled with a neutral gas (in most cases, Ar). 2. An electric potential is applied to the target until electrons are emitted from the target surface. These electrons are responsible for initially ionizing the neutral gas atoms through electron-electron collisions. This state of concentrated ionized gas is considered plasma. 3. As the potential is continued at the target, and as more free electrons from the ionization of the gas atoms are formed, a permanent magnetic field (represented by the blue arrows in Figure 15a) traps a high concentration of electrons near the target surface. 4. These trapped electrons then follow helical paths around the magnetic field lines and further ionize any neutral gas atoms in the vicinity (note: this is the key to magnetron sputtering as opposed to other sputtering techniques – the permanent magnetic field is 24 essential for keeping the electrons near the target surface, to be used to further ionize the gas in close proximity to the target, leading to a more dense plasma). 5. With the applied negative potential at the target surface, the positively charged gas ions are accelerated by the potential at the target surface. 6. The gas ions are injected into the target surface due to their large momentum, causing collision cascades within the material and eventual ejection of the target atoms from the surface. 7. Due to the high concentration of electrons (and hence, high ionization rates of the neutral gas atoms) near the surface, these collisions happen very quickly, continuously, and across the whole surface of the target. In addition, the target atoms that have been ejected maintain a neutral charge and are unaffected by the magnetic field. Therefore, a vapor made up of the target atoms is formed with each target atom traveling in the direction in which it was ejected (away from the target surface). 8. With the target material vapor formed, and the atoms travelling in directions away from the target, any object placed in the paths of the vapor will be coated. Therefore, a substrate is placed some distance away from the target to allow growth of the target material onto the substrate surface. 2.5.2. Process Parameters The characteristics of the deposited film (e.g. quality, composition, density, grain morphology, etc.) are highly dependent on the processing parameters. For example, classic work by Thornton et al. showed that factors such as the Argon pressure, substrate temperature, deposition rate, sputtering source type, and substrate surface roughness can all dramatically change the growth, structure, and morphology of the deposited films [62-64]. 25 Therefore, in order to fully control the growth of the material, and in the present context, to introduce a highly NT structure, it is important to explore some of these processing parameters. Some studies have revealed that the probability of forming twins during sputtering may be improved by increasing the deposition rate or decreasing the SFE [30]. This will be discussed in greater detail in section 2.5.3. The SFE is a material dependent property, and therefore, cannot be controlled during the process. On the other hand, the deposition rate is a function of many processing parameters that can be controlled. However, these parameters are often intricately interrelated to each other, and therefore, it can be practically difficult to change the deposition rate without affecting some other aspect of the growing film. For example, the deposition rate is a function of the power, gas pressure, working distance, and source diameter. These four variables can be individually, or collectively, varied in order to achieve a desired rate. However, some of these variables may also influence the substrate temperature [65-68], which may also affect the probability of forming NTs during growth. In addition, each of these might have adverse affects on the film quality, density, grain size, and morphology of the film. Therefore, careful attention must be paid to these variables, their relation to each other, and their overall effects on material growth. Although the development of growth twins during magnetron sputtering is a relatively new subject and requires further work, preliminary models relating some process parameters to the formation of NT structures have been developed. One of these models will be described and summarized in section 2.5.3 below. 26 2.5.3. Twin Nucleation and Growth during Sputtering Twin growth in sputtered metals has been observed in various fcc materials and under different sputtering conditions, e.g. Cu [60, 69-70], 330 stainless-steel [61], and Ni [70]. However, a comprehensive understanding of the formation of twins as they relate to processing conditions has not yet been developed. One model explored by Zhang et al. has made some advances in relating a material’s probability of forming twins during sputtering to the deposition rate and the SFE. A summary of this model [61] is given below: When atoms in the vapor phase condense on a substrate to form a solid, the nuclei that may form can either be perfect nuclei (free of planar faults), or may contain stacking faults and/or twins. The critical radius for a perfect disc-shaped nucleus with height, h, can be expressed as:                 s perfect P mkT J kT r   2 ln * Equation (2) and for a twinned nucleus:                  h P mkT J kT r t s twin    2 ln * Equation (3) where  is the surface energy, k is the Boltzmann constant, T is the substrate temperature,  is the atomic volume, J is the deposition flux, m is the atomic mass of the depositing species, P s is the vapor pressure above the solid, and t  is the TB energy. In comparing equations 2 and 3, it is clear that r* perfect < r* twin , implying that the nucleation of a perfect nucleus is more probable than that of a twinned nucleus. However, if the TB energy, t  , is very low and the 27 deposition flux (corresponding to the deposition rate), J, is very high, then the difference between r* perfect and r* twin is negligible. This is represented in Figure 16a and 16b, which show a decrease in the difference between r* perfect and r* twin with increasing deposition rate (a) and decreasing TB energy (b). Since the SFE is directly related to the TB energy [71], if a low SFE material is used, and/or the film is sputtered with a high deposition rate, the probability of forming twins will be increased. Figure 16. Plots of the critical nucleation radius as a function of the (a) deposition rate and (b) twin boundary energy [61] It has also been shown experimentally that in the case of NT-330 stainless-steel, the total volume fraction of twinned grains changes as a function of deposition rate. In this case, as shown in Figure 17, the amount of twinned grains developed during sputtering was proportional to the deposition rate, while the overall grain size remained unchanged [29]. 28 Figure 17. Plot showing the volume fraction of twinned grains and the average columnar grain size as a function of the deposition rate [29]. Although these models and observations do not account for all processing conditions and variables, they provide insight into the formation of growth-twinned structures during magnetron sputtering and warrant further study. In the current work, magnetron sputtering was selected as the processing route of choice to further explore the development of highly NT structures. The current work presented here begins to address some of these challenges and will add to the overall understanding of the synthesis of NT metals by magnetron sputtering. 2.6. Summary In this chapter, some of the commonly observed trends in NT metals, especially in terms of the potential for enhanced strength and ductility, have been discussed. It has generally been revealed that metals with high densities of TBs grown into the material are effective in strengthening the material due to the increase in dislocation obstacles. Dislocation activity is found to be the major deformation mechanism (as opposed to GB mediated mechanisms) in NT metals, even down to twin spacing of a few nm. In general, it has been shown that when the majority of dislocations travel inclined to the TBs, 29 strengthening is observed (this is the “hard” direction). Conversely, when dislocations along or parallel to the TBs are active, a “soft” mode is produced, and the material can exhibit much more plastic deformation. In order to tailor the response of NT metals, it is necessary to understand how these mechanisms can be promoted or suppressed. This is an area of ongoing work, but factors such as the TB orientation relative to the loading direction, twin thickness, and grain size have been explored as possible means to control these mechanisms. Due to the difficulties in the processing of NT metals, the data developed, thus far, regarding the deformation behavior, particularly at the μm-to-mm scale, and the synthesis of NT metals, are limited and mostly exists only for NT-Cu. Processing approaches such as pulsed electrodeposition and magnetron sputtering have recently been used in the development of highly NT structures, but a systematic and comprehensive understanding of the formation of NTs during growth is still lacking. Due to the versatility of the process, and the highly aligned NT microstructures that can be achieved, magnetron sputtering was selected for the current study for the synthesis of a variety of new, unique NT metals. In chapter 3, some of the experimental methods and materials pertaining to the current study will be discussed. Specifically, the magnetron sputtering system predominantly used for the processing, and the detailed methods for the characterization and mechanical testing of the NT metals in this study will be discussed. 30 (blank page) 31 Chapter 3 : Experimental Methods and Materials In this chapter, some of the key experimental procedures and materials used during this study will be discussed. In particular, the details of the magnetron sputtering system used to explore the synthesis of NT metals will be presented. Additionally, the methods used for the microstructural characterization and the mechanical testing of the NT foils will be shown. 3.1. Synthesis by Magnetron Sputtering In order to investigate the synthesis of NT metals, a magnetron sputtering chamber was adapted to sputter a range of metals under various conditions. In this section, some of the details of the process chamber, components, sputtering conditions, and general procedures related to the processing will be presented. 3.1.1. Chamber Configuration and Components Magnetron sputtering chambers are generally comprised of one or more magnetron sources mounted inside a high vacuum [10 -3 -10 -6 Pa (10 -5 -10 -7 Torr)] chamber. The atmospheric gases are pumped from the chamber, which is then back-filled with the working gas. Details of this process can be found in section 2.5.1. A schematic of one the magnetron sputtering chambers used in this study is shown in Figure 18. This chamber consists of a diaphragm pump and turbo-molecular pump that is capable of achieving high vacuum with pressures of 10 -4 -10 -5 Pa (10 -6 -10 -7 Torr) after about 24 hour pump time and pure (99.999%) Argon is typically used as the working gas. The current configuration allows for multiple sputtering sources of different sizes to be mounted to both sides and to the bottom of the chamber using 250 mm diameter high-vacuum CF flanges. In addition, there is an assortment of small flanges [33 mm and 84 mm diameter CF] which are generally used for 32 vacuum/deposition diagnostics and characterization, such as vacuum pressure gauges, probes for plasma analysis, thermocouples, and viewports. Also installed is a DC motor which drives a rotating substrate stage used for in-vacuum substrate rotation. Figure 18. Schematic representation of one of the magnetron sputtering chambers used in this study. Sputtering Sources and Targets The magnetron sputtering source uses a combination of permanent magnets and an applied voltage to ionize the working gas (see Ch. 2.5.1 for more details). There are various shapes and sizes of sources, e.g. rectangular, conical, cylindrical post, etc., which vary depending on the application and chamber configuration. One of the most commonly used types for thin film deposition is the planar magnetron sputtering source, as shown in Figure 19. In this case, the source is comprised of a set of water-cooled ring magnets below a Cu electrode (this assembly is considered the “magnetron”), which typically varies in 33 diameter from about 2.5 cm to 15 cm (although smaller and larger sources are possible). In the current study, the magnetron sources that were most widely used were Meivac MAK planar sputtering sources, similar to the one shown in Figure 19, with diameters (“d” in Fig. 19) of 3.3 cm, 5.1 cm, and 7.6 cm. Figure 19. Photograph of a planar magnetron sputtering source. Adapted from [72]. The sputtering target is made up of a disk of the material to be sputtered and is secured to the source electrode using either a magnetic keeper or a clamp. Examples of some sputtering targets are shown in Figure 20, which demonstrates a variety of metals of different diameters and thicknesses. In general, the sources are designed for a specific diameter target, but the thickness of the target for a given source can range from approximately 0.25 mm to 25 mm. As can be seen in Figure 20, the target in the upper-right of the photo is brand-new and has not been used for sputtering. The other targets, however, reveal a ring-shaped trench where the material has been sputtered away during the process. The formation of the trench is 34 due to the donut-shaped magnetic field lines above the target, where the ionization rate of Argon and subsequent sputtering rate is the greatest [73-74]. Figure 20. Photograph showing various targets used for planar magnetron sputtering. Note the ring-shaped trench that develops during sputtering. Cooling System Due to the sensitivity to heat of the magnets inside the magnetron sources, it is necessary to cool them using a constant flow of water or coolant at RT or below. This is usually accomplished using a recirculation water chiller which pumps chilled water through the cooling channels of the source. Each source has a specified coolant rate and temperature, depending on the design of the cooling channel, the coolant line diameters, and the sensitivity of the magnets. A picture of the cooling system used in the current chamber configuration is shown in Figure 21, which displays a water chiller (Thermo-Scientific brand, maintained at 16º C) connected to a custom-made manifold used to control and divert the chilled water to the various sources and turbo pump. In addition, various safety devices have been implemented, such as flow rate monitors and shutoff valves so that if water ceases to flow, an interlock to the source will be disconnected and source power will be turned off. 35 Figure 21. Photograph showing the cooling system designed for the magnetron sputtering chamber. Substrate Holders Many variations exist in the parts and methods used to secure substrates inside the chamber during sputtering. However, some general concerns related to design of substrate stages/holders include the minimization of shadowing effects on the substrate (i.e. the holding device cannot obstruct the path of the sputtered atoms), heat dissipation, and sample manipulation. Some of the specific design goals for the substrate holders used for this study were to (i) secure the sample with minimal force to allow for some deflection of the wafer during deposition; (ii) minimize shadowing while maximizing the coverage area; (iii) maximize heat conductance between the substrate and holder; and (iv) allow for sample rotation while under vacuum. With these considerations in mind, custom substrate holders of various sizes, shown below in Figure 22, were designed and machined using Cu (for heat 36 dissipation) and a base/cover system that secured the substrates without directly tightening them. Using this system, the wafers are gently placed in the counter-sunk recessed portions of the Cu base and a cover with diameter just smaller than the diameter of the wafer is attached to the base. The cover was also designed with a 45º taper to secure the substrate while minimizing shadowing affects. Figure 22. Schematics of the substrate holders designed and used for the synthesis process of NT metals. 3.1.2. Materials and Deposition Conditions As briefly discussed in Ch.2.5.2, the deposition conditions can greatly influence the characteristics of the growing film. For the synthesis of NT metals, in particular, typically a sufficiently high deposition rate and/or low SFE material must be used. Therefore, it is important to first understand the means in which these parameters can be controlled and to select optimum sputtering conditions that may lead to the desired microstructure. Other factors, such as film uniformity, purity, and density, should also be taken into account when choosing the appropriate processing parameters. A variety of conditions were used to explore 37 the synthesis of the highly NT metals in this study, which will be further discussed in this section. Stacking Fault Energy Since a low to moderate SFE is required to improve the chances of forming growth twins during the process, the choice of pure metals that are likely to twin is limited. These include metals such as Cu (~ 80 mJ/m 2 ), Ag (~ 25 mJ/m 2 ), and Au (~ 50 mJ/m 2 ) [6]. However, it has been generally established that by introducing alloying elements into the material, the SFE can be greatly reduced [75]. For example, by adding only 2wt% of Al into Cu, the SFE decreases from ~ 80 mJ/m 2 (in the pure Cu case) to ~ 25 mJ/m 2 [76]. The samples used for the current study are summarized in Table I, which shows the material and its corresponding SFE. Pure Cu and Ag were chosen as representative metals for exploring the synthesis of growth twins in moderate and low SFE pure metals, respectively. Additionally, Cu-6wt%Al was used as a representative binary alloy to investigate the effects of alloying on the NT formation. It should be noted that in Cu-6wt%Al, Al is in solid solution with Cu; therefore, the effects of multiple phases are avoided [77]. Table I. Summary of the metals used in the current study and their corresponding SFE. Material SFE (mJ/m 2 ) Reference Cu 78 - 80 [6, 76] Ag 17-25 [6, 30] Cu-6wt%Al 6 [76] 38 Vacuum (base) pressure In general, the contamination from the atmosphere, including air particles and moisture, can be significantly reduced by decreasing the vacuum pressure inside the chamber [78]. Thus, when attempting to sputter high quality, high purity metallic films, it is necessary to achieve the lowest pressures possible before sputtering. In the current chamber configuration, the turbo-molecular pump is capable of producing vacuum levels down to 2.7 x 10 -5 Pa (2.0 x 10 -7 Torr), where the contamination level is expected to be low [78]. Therefore, all NT films/foils presented here were sputtered after a vacuum base pressure of 10 -4 -10 -5 Pa (10 -6 -10 -7 Torr) was achieved. Argon Pressure During sputtering, the capability of forming and sustaining plasma is highly dependent on the number of collisions between electrons from the target and the gas atoms (i.e. the ionization rate of the gas). Therefore, the higher the gas pressure, the larger the rate of ionization. However, the deposition rate scales roughly by the inverse of the pressure; since the increased pressure leads to higher densities of gas atoms, many more interactions occur between the neutral sputtered atoms and the gas atoms. Thus, the kinetic energy of the sputtered atoms is reduced at higher pressured [79]. By using magnetron sputtering, the trapping of high densities of electrons near the target surface leads to enhanced ionization rates of the gas, so that the pressures required to generate and sustain plasma are an order of magnitude lower than that for traditional sputtering [80]. Therefore, much higher deposition rates can be achieved in magnetron sputtering. In addition, metals sputtered below approximately 1.3 Pa (10 mTorr) have been shown to contain much more densely packed 39 grains [63]. Since high deposition rates are optimum for the growth of NT structures, and fully dense structures are desired, the working gas pressure used for all samples in this study was 0.27 Pa (2 mTorr), which was the lowest achievable pressure using the current chamber configuration. Power As previously mentioned, the ionization of the working gas is essential to the sputtering process, since it is these ions that bombard the surface and eject the target atoms. If the number of ions created is increased, then the total number of collisions also increases and a greater material flux can be achieved (i.e. higher deposition rates). The flow of electrons (i.e. the current) around the magnetic field lines above the magnetron source is primarily responsible for the rate of gas ionization. In addition, the voltage applied to the target is related to the potential between the gas ions and the target surface, and therefore, the kinetics of the ions. Higher ion energy, produced by increasing the voltage to the target, generally leads to increased sputtering yields (i.e. how many target atoms are ejected per one ion collision) in the process, which also directly influences the deposition rate. Since the voltage and current are inter-related and often coupled to each other in the sputtering process, it is standard practice to define the power in the system, rather than the specific voltage or current. In general, there is a direct linear relationship between the deposition rate and the power, and therefore, adjusting the power is a common method to change the rate. However, it is also important to realize that increasing the power may have other effects on the process. For example, the substrate temperature has been shown to be directly related to the power used in the process [67]. In addition, each source has a specific threshold for power that can be 40 applied, which is mostly dependent on the cooling capabilities of the magnetron source [81-82]. In general, the range of powers that can be used depends on the source diameter. Therefore in practice, rather than defining a certain power, a power density (power divided by the target surface area) is typically defined. In the current study, a power was used such that the power density was 3-16 W/cm 2 . Target – substrate distance When the target to substrate distance is reduced, the target atoms travel a smaller distance before arriving at the substrate, and thus, fewer collisions between the target atoms and the gas atoms occur. This leads to faster and more energetic neutral atoms traveling toward the substrate, thus improving the deposition rate. In fact, the deposition rate has been shown to scale by the square of ratio of distances (e.g. if the distance is changed from 8 cm to 4 cm, the deposition rate will increase by a factor of 8 2 /4 2 = 4) [82]. In addition, the heat radiating from the target also scales roughly by the inverse square of the distance, which could lead to heating effects on the substrate and growing film. In order to maintain a moderate deposition rate while maximizing film coverage and minimizing overheating of the substrate, the distance used during the processing of the samples in this study was typically kept above approximately twice that of the target diameter. Interrupted sputtering parameters The approaches used for increasing the deposition rate, such as increasing the power, decreasing the working pressure, or decreasing the distance, may lead to undesirable effects, such as substrate heating. In general, it is difficult to increase the deposition rate by any method without significantly affecting some other area of the process. One method that has 41 been successful in minimizing some of these effects, while still producing a high deposition rate that can lead to twin growth, is using periodic interruptions during the process. In this case, the substrate is either rotated through the deposition area (typically at ~ 2-3 rev/min), as schematically represented in Figure 23, or is periodically blocked by an automated shutter system. Using this interrupted approach, high rates can be used that may lead to the formation of growth twins during the “on” time, while the “off” time may allow the internal stresses to relax [69], and allow the sample to briefly cool by dissipating heat to the surroundings [83]. The dissipation of heat during the process is essential so that overheating of the substrate and film, which can lead to recrystallization and grain growth during sputtering [63], can be minimized. In addition, the interruption process allows for thick films (i.e. > 100 μm) to be sputtered without high internal stresses that can cause internal cracking in the sample [69]. Figure 23. Schematic showing an interruption process in which the substrates are rotated using a turn-table system. 15 cm 42 This method was used successfully for the synthesis of the 170 µm thick highly NT Cu foils presented in this study, in which a deposition rate of approximately 1.8 nm/s was produced without excessive heating or internal stresses during the process. 3.1.3. General Procedures In this section, a more detailed description of the methods and procedures used for the synthesis of the NT metals in this study will be presented. Substrate and target preparation Si (100) single crystal wafers were used as the substrate material for the highly NT metals. Since the quality and structure of the growing film is also related to the initial substrate condition, the surface finish, flatness, roughness and cleanliness were taken into account before sputtering. The details of the substrates used here are presented in Table 1, which displays their initial condition prior to sputtering. Table II. Details of the Si (100) wafers used in this study for the synthesis of NT metals. Source Virginia Semiconductor Orientation <100> ± 0.9° Dopant Boron Diameter 50.8mm ± 25 μm Resistance 1-20 ohm-cm Primary Flat 15.88 ± 1.65mm @ <110> ± 0.9° Prior to loading the substrates into the vacuum chamber, they were also cleaned using a rinse procedure of (i) dish soap + water, (ii) regular (tap) water, (iii) DI water, and (iv) ethanol; finally, they were gently dried using compressed nitrogen gas. In addition, since heat dissipation is important during the growth of the films, and the convective heat transfer to the 43 surroundings is slight due to the low pressures in the chamber, conductive heat transfer to the substrate holder is critical. Therefore, a layer of thermally conductive Ag paste was spread between the substrate back and the substrate holder to ensure good thermal conductivity between the two. This is presented in the photograph shown in Figure 24, in which a uniform layer of paste can be seen on both the substrate holder (a) and wafer back (b). Figure 24. Photograph showing the application of thermally conductive Ag paste between the substrate holder (a) and the substrate wafer back (b). Similarly, the targets were cleaned thoroughly to reduce any contaminants during the process, and a generous amount of Ag paste was applied between the target back and the source electrode to ensure both thermal and electrical conductivity. The substrates and targets were then loaded into the vacuum chamber and the atmospheric gases were pumped until a minimum base pressure of 2.7 x 10 -4 Pa (2.0 x 10 -6 Torr) was achieved. Sputtering Procedure Once an appropriate base pressure was reached, the target/source was first cooled to approximately 16º C (about 5 minutes before sputtering), while ensuring proper water flow (minimum of 2.3 l/min) to the source. The desired working gas pressure was then set using a 44 closed-loop pressure controller and the substrate was either rotated out of the deposition area, or a shutter was put in place between the substrate and the target. This was performed in order to pre-sputter the target to remove any contaminates or oxides that may have formed on the surface during storage. The desired power was then set at the power supply and the source was turned on to generate the plasma and begin sputtering. A minimum of 2 minutes of pre-sputtering was performed before rotating the substrate into the deposition area, or opening the shutter to begin deposition on the substrate. The film/foil was sputtered for a predetermined time (depending on the desired thickness) based on previously calculated deposition rates. When interrupted sputtering was used, a similar procedure was followed with the exception of beginning the sample stage rotation after pre-sputtering. Note: the deposition rate was usually previously determined by sputtering onto a sacrificial glass substrate at a certain set of conditions and measuring the total thickness of the film by profilometry – the average rate was then calculated as the total thickness divided by the total time of deposition. Removal of foil from substrate After sputtering was completed, the sample was allowed to fully cool under vacuum (usually about 10-15 hours) before removing from the chamber. In order to produce a free- standing foil, the substrate back was scribed using a diamond-tip cutting tool and the substrate was intentionally broken, allowing the film to freely peel off of the substrate. Note: this method only works when there is a lack of adhesion between the substrate surface and foil (which was generally the case in the metallic foils used for this study) and when the foil is sufficiently thick (approximately 10 µm or thicker). Other methods, such as selective 45 chemical etching of the substrate or sputtering a thin under-layer of a different material can also be used to produce a free-standing film. Using the materials and methods described in this section, a variety of NT foils were synthesized, including highly NT Cu, Ag, and Cu-6wt%Al. In each of these metals, different twinned versions were produced by slightly varying the deposition conditions. However, the metallic foils that were ultimately mechanically tested and explored in greater detail were created using an optimum set of processing conditions, which produced the highest quality structures (i.e. dense, pure, highly twinned) possible, using the current chamber configuration. The details of these conditions can be found in later chapters that discuss the specific NT materials (Chapters 4 and 5). The following section summarizes the tests that were used to explore the mechanical behavior of the various NT foils. 3.2. Mechanical Testing An important aspect of NT metals is their potential for high strength while maintaining other useful properties, such as ductility (this was discussed in greater detail in Ch. 2). However, these reports have been mostly limited to Cu, which led to one of the objectives in this study of synthesizing other (non-Cu) NT metals. In addition, most of the data pertaining to possible enhanced properties of NT metals have been obtained using small-scale mechanical testing, in particular, nano-indentation. A wide-spread study and understanding of the mechanical behavior of NT metals across multiple length scales and under different deformation modes is still lacking. Therefore, in the current study, all NT metals were synthesized large enough such that larger-scale mechanical tests could be performed, such as tensile tests, micro-indentation, compression, torsion, etc. In the following section, some of 46 the techniques used to explore the mechanical properties and performance of the NT foils are presented. 3.2.1. Tensile Testing Tensile testing is a valuable tool for investigating the material’s properties and deformation behavior, especially pertaining to its strength and ductility. In a single tensile test, properties such as elastic modulus, yield strength, ultimate strength, work-hardening rate, and elongation-to-failure can be determined. In addition, the overall deformation modes in tension can be directly observed, such as permanent elongation, reduction of gauge width, strain localization, necking, fracture, etc. Therefore, in order to evaluate the overall tensile response, material properties, deformation, and fracture of the NT metals, tensile specimens of the various NT foils were prepared and tested under different conditions. Some of the details of the tensile tests and sample preparation will be discussed in this section. Tensile specimen preparation In order to perform the tensile tests, it was first necessary to cut the foils into dogbone shapes with a reduced area, similar to those used in traditional tensile tests [84]. Through analysis of various shapes and sizes for the tensile specimens, an optimum design was chosen such that the majority of the deformation would occur in the reduced gauge section. A dimensioned drawing of the dogbone shape used for the NT foils is shown in Figure 25. Note: the foil thickness ranged from 25 µm to 170 µm, depending on the material. 47 Figure 25. Dimensions (in mm) of the dogbone shaped specimens used for tensile testing of the NT metallic foils. A special specimen die, shown in Figure 26, was designed and machined so that the foils could be cut into the dogbone shapes with minimum handling. The foils were sandwiched between the two parts (labeled “1” and “2” in the schematic), which were then secured and aligned using press fit dowel pins (labeled “3” in the Fig. 26). The foils were either cut using a stainless steel blade or were polished down from the edges using Si-C paper. Once the foil was cut or shaped, it would be removed and loaded into the tensile tester stage (discussed in the following section). Figure 27 shows the initial sputtered foil cut into strips (a), and the final shaped specimen used for the tensile tests (b). 15 10.5 6 3 9 4.5 2 3 48 Figure 26. Schematic of the tensile specimen die designed for cutting and shaping tensile dog-bone specimens. Figure 27. (a) as-sputtered 50mm dia foil cut into multiple strips for tensile dogbone specimen preparation using the tensile specimen die shown in Fig. 26. (b) Example of post-cut dogbone specimen. 49 Tensile testing apparatuses and procedures The machine used for the tensile testing of the majority of the NT foils in this study was an Instron dual-column tabletop model 5965 with a load capacity of 5 kN. Non-contact strain measurements were made using an advanced video extensometer (Instron AVE 2663- 821) with a 60 mm field of view and 0.5 µm resolution. A photograph of the Instron machine is shown in Figure 28. Figure 28. Photograph of the Instron tensile tester used for testing the NT foils. 50 Since the NT specimens were sputtered to be high purity and relatively defect-free (i.e. low initial dislocation density, etc.), it was necessary to minimize the handling of the foils during installation into the tensile tester, which might introduce undesired deformation in the samples. Therefore, a special jig and adapter was designed and used in order to properly align and install the sample with limited handling. This is shown in Figure 29. In this setup, the specimen is first installed into a stage, which consists of two grips aligned and connected to each other by a backing plate. The sample is aligned in the axial direction using two dowel pins and is then clamped by grips using two screws on each end. The entire stage/specimen assembly is then fitted and secured to the machine fixtures using a clevis/pin system and the backing plate connecting the two grips is removed, leaving the foil unconstrained in the axial direction and ready to be tested. Figure 29. Custom jig/stage used to align and secure the NT foils to the testing machine fixtures. 10mm (a) (b) Alignment dowel pins Clevis/pin adapter to secure to tester fixtures Backing plate to stabilize sample during installation screw-grips 51 With the foils then installed in the testing fixtures, a slight pre-load was applied (approximately 5 N) in order to remove any slack in the specimen. The cross-head rate was used as the test control and was calculated as: Equation (4) where is the cross-head rate, is the desired strain rate, and L 0 is the original gauge length [85]. For example, if the desired strain rate was 10 -3 /s, and the gauge length measured by the video extensometer was 6mm, the cross-head rate would be set to a constant 0.006 mm/s. The strain measurements for the NT foils were performed using the video extensometer by first gently placing two white paint markings on the specimen, separated by the desired gauge length (approximately 6 mm in most cases). The two paint markings can be observed in the previous picture shown in Figure 29, and is further shown in the screen-shot taken of the video extensometer software in Figure 30, in which the two red boxes show the tracking of the two markings during the test. The extensometer measures directly the displacement of the two markings as the sample is pulled in tension and a strain is calculated by dividing the relative displacement by the original gauge length. Figure 30. Screen-shot showing the tracking of paint markings by the video extensometer. 6mm 52 3.2.2. Vickers Micro-Indentation The method of indentation is another valuable tool in assessing the micro- and macro- scale deformation of metals. In general, indentation is a method in which an indenter is pressed into the sample while accurately measuring the force. The hardness of the material is then calculated by the force divided by the indenter surface area. One of the most common types of indenters is the Vickers, which is defined by a square-based pyramidal shaped diamond indenter [86], as shown in Figure 31a. A scanning electron microscope (SEM) image of a set of Vickers indents on a metallic sample is shown in Figure 31b. There are two distinct test ranges using the Vickers tip, which are micro (10-1000 g) and macro (1-100 kg) indentation; however, the same tip is used across the entire force range and the test is considered mostly force independent (i.e. the same approximate hardness value will be obtained using a variety of forces). Further details of the Vickers indenter and standard procedures can be found in ASTM E384 [86]. Figure 31. (a) Schematic of the Vickers pyramidal diamond tip and (b) SEM image showing a set of Vickers indents on the surface of a NT foil. 53 For Vickers indentation, the hardness value, HV, is determined by measuring the average of the indent diagonals and using the equation Equation (5) where P is the force (in grams), A s is the contact area, α is the face angle (136° for Vickers), and d is the mean diagonal length of the indentation (in micrometers, μm). For Vickers measurements, this equation becomes Equation (6) Multiple indents are typically made for each sample (usually a minimum of 5) and the reported HV value and corresponding uncertainty is determined by averaging the values from all of the indents. In the current study, a LECO-LM100 Vickers micro-indenter was used, which is shown in Figure 32. Figure 32. Picture of the LECO-LM100 micro-indenter used in this study for the indentation of the NT foils. 18cm 54 In addition to using the tensile tests and Vickers micro-indentation for evaluating the deformation behavior and mechanical properties of the NT metals, other tests were performed on some of the samples, specifically, the NT-Cu foils in this study. This included compression, tension-tension fatigue, and high-pressure torsion to further evaluate the overall deformation and microstructural stability. However, these additional tests were performed in collaboration with other research groups, so the details will not be presented here. Some test details and results will be presented in the chapters corresponding to the NT metals on which the additional tests were performed. 3.3. Sample Characterization In order to evaluate the materials’ as-sputtered and post-tested microstructures, film quality, surface morphologies, etc., various characterization techniques were employed. In the current section, some of the methods commonly used in the analysis of the NT metals in this study will be summarized. 3.3.1. Film Thickness, Roughness, and Residual Stress The film thickness, surface roughness, and residual stresses in thin films are some of the first factors commonly characterized. All of these quantities were evaluated to determine process control parameters, such as deposition rate, as well as overall film quality. Using techniques such as stylus profilometry, the film thickness and roughness can be measured directly, while the residual stress can be indirectly measured and calculated. Profilometry is a method in which the vertical displacement is measured, typically with angstrom resolution. By moving the sample horizontally across the sample, the surface features and profile can be measured. In the current study, an Ambios XP-2 stylus profiler was used with a stage 55 diameter of 200 nm, maximum scan length of 50 mm, and a 0.1 nm vertical resolution using a 2.5 μm tip radius. This device is shown in detail in Figure 33. Figure 33. Photographs showing the Ambios XP-2 used in the current study. To measure the sample thickness, it is necessary to expose some surface of the underlying substrate and providing an abrupt step up to the film or foil. As can be seen in the photograph in Figure 33, the sample consists of a round wafer with a coating on its surface that is just smaller in diameter than the wafer. The stylus tip can then be moved along the 58 cm 36 cm 50 cm 56 sample, starting at the exposed portion of the substrate and ending on the film/foil. A step- profile can then be observed, as shown in Figure 34. The profile data can either be exported for further analysis, or the step height (the difference in the two plateau regions of the curve) can be measured and recorded directly. In addition to measuring the step height of the NT samples to determine overall film thickness and deposition rates, the profilometer was also used to calculate the surface roughness and the wafer profile before and after deposition to calculate the average residual stresses in the foils [87]. Figure 34. Step-profile used to measure the film/foil thickness directly from the profilometer. 3.3.2. Orientation and Texture Another commonly used technique in analyzing the structure of thin films is X-ray diffraction (XRD). This technique is very useful in determining grain orientations, grain sizes, and internal stresses; in addition, it can be used to determine an unknown material’s crystal structure and layer thickness in the case of multi-layered materials. For the current study, a Rigaku Ultima IV XRD system was used in analyzing the NT foils. 57 The internal setup of the Rigaku XRD is shown in Figure 35, in which the sample is laid flat on the stage and the angle between the X-ray source and detector is varied. The crystallographic planes that satisfy the Bragg equation [88] will allow the X-ray beam to enter the detector and some intensity count will be measured. In an out-of-plane measurement, the crystallographic planes that are parallel to the sample stage will be counted and the intensity is plotted as a function of the detector angle. Therefore, in viewing the plot of intensity vs. angle, one can easily observe either some texture in the out-of-plane (vertical direction), or a random orientation (if all peak intensities were roughly the same). Figure 35. Photograph of inside the Rigaku XRD system. In the case of magnetron sputtered NT samples, the microstructure often consists of columnar grains with TBs oriented parallel to the substrate. Since coherent TBs are always on (111) planes, measuring the out-of-plane (111) texture by XRD can give a qualitative measure of the amount of twinned grains in the sample. A plot of some NT samples with 8 cm 58 different amounts of NT grains is shown in Figure 36, which shows different intensities for the non-(111) peaks. This generally implies that the (111) texture is not as strong and the average grain orientation is slightly more random. Figure 36. Example of a plot revealing various (111) out-of-plane textures in magnetron sputtered foils of different microstructures. 3.3.3. Microstructural Imaging To directly observe and evaluate the surface morphology and microstructure of the NT foils, various imaging techniques were employed, including SEM, focused-ion beam (FIB), and TEM. Although all of these techniques were used in exploring the microstructures of the NT foils, the micro-machining capabilities and enhanced channeling contrast of the FIB enabled it to be the primary instrument used for the microstructural analysis of the samples in this study. Ag foil 1 Ag foil 2 Ag foil 3 59 Focused Ion Beam (FIB) The FIB used here uses a beam of Ga + gas ions to bombard the material, while a detector measures the intensity of the secondary electrons that are emitted due to the bombardment [89]. This process is similar to the SEM imaging, where instead of an ion beam, an electron beam is used. Since the Ga + ions are heavy compared to electrons, they travel deeper into the material and with more energy, thus producing secondary electrons from further within the material. Due to this effect, a different intensity signal will be produced, depending on the amount of atoms that the ions interact with (i.e. the sample’s crystallographic orientation relative to the ion beam). Therefore, FIB produces much better channeling contrast, compared to electron imaging, allowing for different grain orientations to be resolved. This effect can be seen in Figure 37, which shows an ufg metallic sample with differently oriented grains. The lighter colored grains in this sample are those that are oriented with their closed-pack directions parallel to the beam (i.e. the ions only travel a short distance due to the high density of atoms and, therefore, produce many secondary electrons near the surface – thus, a higher signal and brighter image is observed). Conversely, the darker grains are those with crystallographic orientations that allow the ions to travel deep into the material, in which the electron signal is much weaker, and a darker image (lower intensity) is produced. 60 Figure 37. FIB image of the surface of an ufg metallic sample showing the orientation contrast due to channeling effects of the ion beam. In addition to the grain orientation contrast produced by the ion beam, another useful effect of the FIB is that when higher energy ion beams are used, it is possible for the sample to be sputtered away. By carefully controlling the areas exposed to the ion beam, selective removal of the material can be achieved, leading to a micro-milling effect on the sample. This phenomenon has led to wide-spread use of the FIB as a micro-machining tool for various applications. One of the common applications is to cut a trench, such that the cross-section of the material is exposed, thus allowing for direct imaging of the microstructure by the various techniques (SEM, FIB, etc.). An example of this is shown in Figure 38. In Figure 39, a cross-sectional image is shown in which the cross-section was produced using a series of high energy ion beam and the actual imaging was performed using the FIB channeling contrast. 61 Figure 38. FIB images of a cross-sectional trench (a) top-view and (b) cross-sectional view prepared for microstructural evaluation of the material cross-section. Figure 39. FIB channeling contrast image of the cross-section of a NT foil prepared with a high energy ion beam. 62 Due to the micro-machining capabilities of the FIB, it is also often used for TEM specimen preparation. For TEM, it is typically required for samples to be on the order of tens of µm in height and width, and approximately 100 nm in thickness in order for the electron beam to transmit the sample. Using the FIB, the entire sample preparation can be performed in a few steps by essentially cutting a TEM specimen out from the material and using a micro- manipulator probe to move the specimen to a standard TEM grid. Figure 40. (a) TEM specimen (highlighted by red box) prepared by FIB specimen preparation techniques; (b) channeling contrast image of the TEM specimen. As shown, the FIB can be a valuable tool used for imaging of the sample microstructures; whether it’s by direct imaging using the channeling contrast produced by the FIB, or by preparing cross-sections and TEM specimens for imaging by other techniques, it has become a versatile imaging method. The current study used the FIB in a variety of ways, including as-sputtered sample characterization, post-testing microstructure analysis, fracture imaging, and TEM sample preparation. 100µm 3µm (a) (b) 63 3.4. Summary In this chapter, some of the details concerning the synthesis, mechanical testing, and characterization of the highly NT metals used for the current study were summarized and discussed. Specifically, magnetron sputtering was performed using a wide range of conditions in order to produce optimum NT structures (i.e. high purity, low initial defect density, high density of NTs) of Cu, Ag, and Cu-6wt%Al. The Cu and Ag served as representative pure moderate and low SFE, while the CuAl was used to investigate the role of alloying in reducing the SFE, thus further facilitating twin growth. Previous reports on the mechanical properties and deformation of NT metals were typically performed using smaller scale testing, such as nano-indentation. A systematic study of the mechanical performance of NT metals across multiple length scales and under a variety of modes is still lacking. Therefore, the samples in this study were synthesized large enough such that more traditional mechanical tests can be performed, such as tensile tests and micro-indentation. Various characterization methods were used, including profilometry, XRD, and FIB, to evaluate the initial and post-tested microstructures, deformation modes, and fracture of the highly NT foils. In chapter 4, a detailed study on the mechanical performance and deformation of highly NT-Cu tested under a variety of conditions will be presented. In addition, chapter 5 will present and discuss the synthesis and deformation behavior of the other (non-Cu) NT foils. 64 (blank page) 65 Chapter 4 : Mechanical Behavior of Highly Nanotwinned Cu Recent efforts in exploring the mechanical properties of nanotwinned (NT) metals have revealed that these materials possess a potential for strengthening while maintaining other properties, such as ductility. However, the NT-Cu samples that have been studied have mostly been limited in sample thickness (typically on the order of a few µm), due in part to processing difficulties. This has restricted mechanical testing to mostly small-scale approaches, in particular nanoindentation. Therefore, the overall mechanical behavior and deformation mechanisms across multiple length scales have, thus far, not been explored in great detail. Hodge et al. made progress in this area as they managed to synthesize thick (170 µm) highly NT-Cu foils [69]. Tensile tests were then performed at RT and liquid nitrogen temperature (77K) in order to investigate the larger-scale mechanical properties at both temperatures. The tests revealed that the samples tested at 77K showed both higher strength and ductility. Testing also showed that the formation of shear bands at both testing temperatures seemed to be the dominant deformation mechanisms. However, the exact role of the highly NT structure in the material behavior, the overall microstructural evolution during deformation, and the formation of the shear bands was unknown. The current study further explores the behavior of highly NT-Cu by focusing on the microstructural evolution and NT stability through the use of various characterization techniques. Thorough microstructure analysis of the post-tensile tested NT-Cu samples will be presented in this chapter, including a detailed discussion of the enhanced properties at low temperatures, deformation mechanisms within the NT structure, localization of plastic formation deformation by shear bands, and the microstructural events leading to fracture. 66 4.1. NT-Cu Synthesis The highly NT-Cu foils used here were synthesized from pure (99.999+%) Cu deposited onto Si (100) single-crystal wafers using an interrupted magnetron sputtering approach at Lawrence Livermore National Laboratories (LLNL). The large chamber configurations available at LLNL allowed for a rotating turn-table type sample stage to be used for continuous rotation of the substrates through the deposition area for long periods of time [69]. The general sputtering process was described earlier in chapter 3 and more detailed procedures can be found in section 3.1.2 and in Ref [69]. The interruption process enabled the use of high deposition rates (approximately 1.8 nm/s) while minimizing internal stresses and adverse temperature effects in the growing film. This allowed for the deposition of relatively stress-free, defect-free, thick (approximately 170 µm) free-standing Cu foils containing high densities of embedded growth-NTs. 4.2. As-sputtered Microstructure The initial microstructure of the NT-Cu foils consisted of highly aligned columnar grains with widths of 500-800 nm and an average TB spacing of 40 nm. The columnar grain boundaries were oriented parallel to the growth direction and the majority of TBs were oriented perpendicular to the growth direction. Figure 41 shows cross-sectional micrographs of the as-sputtered NT-Cu foil [(a) FIB micrograph and (b) TEM image]. 67 Figure 41. (a) FIB cross-sectional micrograph [90] and (b) TEM cross-section of the as-sputtered highly NT-Cu microstructure [31]. Arrows show the growth direction. 4.3. Stress-Strain Response in Tension The NT-Cu samples were tested in tension using similar methods described in chapter 3.2.1. The tests were performed at various strain rates (10 -4 -10 -2 /s) at RT and liquid-nitrogen temperature (maintained at 77K). (Note: these tensile tests were performed prior to this study at LLNL and the details of the tests and setup can be found in Ref [91]). Figure 42 shows the engineering stress-strain curves at the different strain rates and testing temperatures, while Figure 43 reveals the overall deformation of the samples tested at RT (a) and 77K (b). (a) (b) 68 Figure 42. Engineering stress-strain curves of NT-Cu tested at various strain rates (10 -4 -10 -2 /s) at RT and 77K. Figure 43. Representative optical micrographs of fractured dogbone specimens tested in tension at (a) RT and (b) 77K at a rate of 10 -4 /s. 69 The stress-strain curves in Figure 42, together with the micrographs showing the deformation in the samples in Figure 43, highlight four distinctive features: the observation of a yield peak at both testing temperatures due to the low initial dislocation density [91], the lack of strain hardening in all samples, deformation at both temperatures by shear band formation, and finally, the samples tested at 77K show both higher strength and ductility, compared to the samples tested at RT. Of these, the latter two are most puzzling. Why do the samples tested at 77K show both higher strength and ductility? What are the roles of the shear bands in the overall deformation? What are the contributions from the highly NT microstructure? In order to address some of these questions, the microstructural evolution and shear band formation were examined using detailed FIB microstructure analysis and high energy micro-diffraction (HEMD). In section 4.4, the results of the microstructural studies are discussed in terms of the enhanced strength and ductility at 77K; additionally, the formation of shear bands as the dominant deformation mechanism is further explored in the highly NT samples. Some of this work has been published in multiple peer-review journals, and further details can be found in Refs [18, 90, 92]. 4.4. Microstructural Evolution and Shear Band Formation In order to evaluate the microstructures of the post-tested NT-Cu samples, various characterization techniques were employed, including detailed FIB cross-sectional imaging and HEMD. In the FIB cross-sectional analysis, micrographs were produced in select locations within the tested sample to determine the modes and extent of localization of the deformation. In the HEMD study, Debye-Scherrer rings were obtained by scanning across the 70 deformation regions in order to further explore the origins and contributions of the observed shear bands. 4.4.1. Microstructure Analysis by FIB In this section, a summary of the FIB microstructure analysis in critical regions of the fractured tensile test specimens is presented. Further details can be found in the published version of this work [90]. Figure 44 shows a representative micrograph of the midsection of a tensile dogbone tested at RT at 10 -4 /s. The sample clearly shows two deformation bands which, once they meet, become a failure site. This shear band type of deformation is typically found in bulk metallic glasses and has also been reported in nc materials [38, 93-94]. In comparing the deformation micrograph to the stress-strain curves at RT, shown earlier in Figure 42, it is observed that one band forms and expands as the predominant deformation mode, while the second band is formed right after the yield peak. The shear band formation with respect to the yield peak is consistent with video observations by Carsley et al. on nanostructured Fe alloys [95-96]. This type of deformation was also observed by Wei et al. [97] on nc-Fe, where the behavior was described as “glasslike” due to the shear band formation and lack of strain hardening. Figures 44 (a-c) show critical locations in the deformed samples, which are similar to the observations in nc-Fe [95, 97]. Figure 44a is near the area of highest plastic deformation (several micrometers from fracture edge), where some grain growth (grain width ~ 1.2 µm) is observed, whereas the twin spacing is approximately the same size of the as-sputtered case. In Figure 44b some grain rotation is observed (grain width ~ 650 nm) and in Figure 44c, the grains and twin size do not change (grain width ~ 600 nm). In all three 71 figures, it is observed that the twins on either side of the boundary are out of registry, which seems to point to a build-up of dislocations, as shown by Shute et al. on fatigued NT-Cu samples [98]. Figure 44. Representative of a tensile specimen midsection deformed at 10 -4 /s at RT. Critical locations in the deformed sample were analyzed using cross-sectional FIB micrographs: (a) Region near the highest plastic deformation; (b) area close to the fractured deformation band showing slight grain rotation; and (c) area along a deformation band. Arrows in (a-c) show the direction of growth. In contrast, Figure 45 shows the sample tested at 77K, wherein the sample exhibits necking as well as multiple shear bands, which can also be observed in the shape of the stress strain curve (Fig. 42), in which a sequence of small yield peaks is associated with the sudden drops in the flow stress. The micrographs in Figures 45 (a-c) are similar to those at RT with the main difference being that out of registry NTs were not observed. 72 Figure 45. Representative midsection of a tensile specimen deformed with 10 -4 /s strain rate at 77K. Critical locations in the deformed sample were analyzed using cross-sectional FIB micrographs: (a) Region near the highest plastic deformation showing some grain growth; (b) area close to the fractured zone, showing slight grain rotation; and (c) area along a deformation band. Arrows in (a-c) show the direction of growth. The higher strength at 77K is expected since dislocations move slower at low temperatures and there is a lack of mobile dislocations [99]. However, based only on the microstructure analysis here, it is not clear why there is an increase in ductility at 77K. It should be noted that the increase in ductility comes mostly from post-necking elongation, in which the tests at 77K show a 3-5% increase in post-necking strain. In both the RT and 77K cases, the non-uniform plastic deformation occurred immediately after the yield peaks, and therefore there is no measurable change in the uniform plastic deformation before necking. In this case, there are several competing phenomena which are unique to this system due to the high purity, low initial dislocation density, and highly NT structure. First, it is observed that there are more deformation bands at 77K, which each can carry large amounts of deformation as compared to the RT tests. Here, it is proposed that these additional deformation bands can be adiabatic, unlike the bands formed at RT. Even though one might typically think of 73 adiabatic shear bands being formed as part of ballistic deformation, they are also prevalent at the low temperature regimes where the heat capacity of a metal is low [100]. However, copper is not prone to forming adiabatic shear bands [101], therefore, the observation of adiabatic shear bands at 77K is likely due to a combination of the effect of the NTs as well as the material’s decreased heat capacity at lower temperatures (for a detailed discussion, see [90]). In order to further explore the microstructural evolution and the role of NT’s in the observed localized plastic deformation, the microstructure within a shear band, prior to fracture, was analyzed in a sample (strain rate of 10 -4 /s) that was stopped immediately after the yield peak. Figure 46 shows an optical micrograph of the gauge section after the test was stopped, in which a clear shear band is observed across the entire width. A slight difference in the surface roughness is observed in the magnified micrograph in Figure 46b. A representative cross-section FIB micrograph within the shear band region is shown in Figure 46c, where there is little to no change in the microstructure compared to the as- prepared sample. It should be mentioned that FIB cross-section analysis of the microstructure was performed in multiple locations within the shear band region and at various orientations relative to the shear band direction. In all cases, the microstructure was similar to that seen in Fig. 46c. This observation implies that the shear band was formed prior to any major changes in microstructure (i.e. grain growth, refinement, detwinning, etc.). 74 Figure 46. (a) Optical micrograph of the gage section of a representative tensile sample tested at 10 -4 /s that was stopped immediately after the yield point. (b) A magnified optical micrograph of the shear band across the specimen width. (c) FIB cross-section micrograph of a section within the shear band region [location of FIB cut marked by double-headed arrow in (b)]. Although the macro-scale deformation is different in the samples tested at RT and 77K (i.e. different shear band densities, necking behavior, fracture, etc.), the FIB analysis revealed that the microstructural evolution seemed to be similar in all samples. In addition, the microstructure within the unfractured shear band was similar to that of the as-prepared, untested case. Therefore, using the FIB technique alone, a complete understanding of the shear band formation and the enhanced ductility at 77K could not be achieved. In order to further evaluate the origins of the shear bands, their role in the overall deformation and ductility, and the influence of the highly NT structure, a complimentary study was performed using HEMD. The following section will present details of the HEMD study and will further discuss the microstructural evolution, the formation of the shear bands, and their contribution to the deformation and ductility in the highly NT-Cu foils. 75 4.4.2. Shear Band Formation and Texture Analysis by HEMD The microstructure of the fractured tensile tested specimens was investigated by HEMD. Further details of this work can be found in the published version in Ref [92]. The HEMD study was accomplished using a micro-focus, monochromatic X-ray beam (beam size: 8 μm × 20 μm, energy: 69.7 keV) in transmission geometry at the HEMD endstation of beamline ID15A of the European Synchrotron Radiation Facility (ESRF). This type of study is particularly useful due to the ability to probe locally, given the relatively small beam size, and with good statistics since the beam transmits the entire sample. Diffraction patterns (DPs) comprising several complete Debye-Scherrer rings were obtained using an area detector (165 mm MAR CCD, Mar, Inc., Evanston, IL) and were taken every 20-30 μm along and across the shear bands and fractured edges. By analyzing the intensity variations in the Debye-Scherrer rings, the in-plane texture in the sample can be investigated. For example, a sample with completely random in-plane grain orientation will produce a set of homogenous concentric rings, whereas a sample with a pronounced texture, i.e. preferred grain orientations, will show intensity variations in different directions [102]. Cross-sectional FIB micrographs were also prepared at select locations in order to directly compare the DPs to the actual microstructure. Figure 47b shows a schematic representation of the scans performed on the fractured dogbone halves, where the arrows show the general scan locations and directions. A minimum of four separate scans were performed at various locations in the fractured samples. It was found that the scans running parallel to the tensile direction showed the most significant microstructural changes compared to the scans along the direction of fracture. In addition, the results for the parallel scans were similar to each other; therefore, sets of DPs and corresponding FIB micrographs from only one scan per sample are presented 76 in this study. For reference, the FIB cross-sectional micrograph of an as-prepared undeformed sample is shown in Figure 47a. Additionally, the DP for the as-prepared sample, showing a completely random in-plane texture (i.e. homogeneous concentric rings), is shown in Figure 47c. Figure 47. (a) cross-sectional FIB micrograph showing the original as-prepared microstructure; (b) schematic representation of the micro-focus X-Ray scans performed on the fractured samples; (c) representative diffraction pattern for the as-prepared sample (before tensile testing). Figure 48 shows an optical micrograph of one half of the fractured dogbone specimen tested at RT (arrow represents the X-ray scan in this case). The scan results can be separated into two distinct regions where noticeable differences in the DPs are found: x < 500 µm (Region I) and x > 500µm (Region II), where x is the distance from the fractured edge. Figures 48 (a1) and (b1) show DPs representative of Region I at x=50 µm and x=150µm, respectively, in which a pronounced texture (seen as the intensity variations around each ring) is observed after deformation. The texture consists of six-fold symmetries with maxima 77 intensities for (111), (200), (311), and (222) reflections in the tensile direction and minima in intensity for the (220) reflection. The development of texture in this region is indicative of dislocation-mediated plasticity [103-105] and in this case is likely formed by rotation of the columnar grains into directions of preferred slip [97, 105-106]. Cross-sectional FIB micrographs at the corresponding locations in Region I are shown in Figures 48 (a2) and (b2). Here, the twinned microstructure persists without any major changes in the twin spacing, while most noticeable changes occur at the columnar grain boundaries in which the boundaries are less well-defined and grains next to each other begin to merge. Since the twin boundaries are still intact, and no obvious detwinning in the form of thinning/thickening of the twins is observed, it is expected that dislocations were able to cut across the twin boundaries [41] or bow within the twin interiors [107] until eventually accumulating at the columnar grain boundaries, which resulted in disruption of the straight boundaries [31]. Beyond Region I, the DPs and FIB micrographs are similar to the undeformed case, further emphasizing the extent of the localization of plastic deformation. 78 Figure 48. NT-Cu gauge section (one half shown) tensile tested at RT and 10 -4 /s. Representative microfocus X-ray DPs [(a1)-(c1)] and cross-sectional FIB micrographs [(a2)-(c2)] obtained at (a) 50 µm, (b) 150 µm, and (c) 1 mm from the fracture edge. Note that locations (a) and (b) are both in Region I. Figure 49 shows the optical micrograph, representative X-ray DPs, and FIB micrographs for one half of a fractured dogbone specimen tested at 77K. In contrast to the sample tested at RT, three distinct deformation regions are observed: x < 120 µm (Region I), 120 µm < x < 1550 µm (Region II), and x > 1550 µm (Region III), where x is the distance from the fractured edge according to the arrow in Figure 49. For the three regions, specific changes in the DPs correspond with visible deformation features shown in the optical 79 micrograph in Figure 49 (i.e. the fractured edge, the area containing the shear bands, and outside of any major deformation, respectively). Overall, the DPs and FIB micrographs are similar to the RT sample, in which the same texture is observed within the shear band and necked regions. However, the obvious difference between the two samples is the additional region within 120 µm of the fractured edge, in which the DPs, as shown in Figure 49 (a1), display only individual spots without any observable in-plane texture. This “spotty” type of DP indicates a severely coarsened material [108], where only individual large grains fulfill the diffracting condition. As a consequence, no continuous Debye-Scherrer rings are obtained. The FIB micrograph, shown in Figure 49 (a2), confirms a completely detwinned, micro- grained structure, where none of the columnar grains or original NTs remained. In this sample the large grains are observed throughout the entire region of 120 µm from the edge. These grains are approximately 10 times the size of those found in the localized detwinned regions at RT. 80 Figure 49. NT-Cu gauge section (one half shown) tensile tested at 77K and 10 -4 /s. Representative microfocus X-ray DPs [(a1)-(c1)] and cross-sectional FIB micrographs [(a2)-(c2)] obtained at (a) 50 µm, (b) 700 µm, and (c) 2 mm from the fracture edge. It is interesting to note that in both samples, the same texture was developed within the shear banded areas. In the sample tested at 77K, the wider-dispersed zone of thin shear bands visible in the optical micrograph corresponds to the expanded textured region. A similar correlation was observed at RT, in which a more localized, narrow region of texture matched the singular, wider, shear band. It is reasonable to assume, therefore, that the shear bands formed as a result of dislocation-mediated grain rotation at both testing temperatures. Since 81 dislocation activity is expected to be much slower at low temperatures [91, 99], it is also anticipated that the rotation of the columnar grains, evidenced by the evolved texture, would also be slower. In general, if the grains were allowed to completely rotate to directions of maximum stress, the plastic flow of the material would likely be localized to only that region, similar to the predominant shear band that formed in the sample at RT. Once the strain had been effectively localized, the sample would continue to deform within these regions until fracture would inevitably occur. If, however, there was a resistance for complete grain rotation, such as that expected by the inhibited dislocation motion at low temperatures, then the amount of stress required to completely rotate the grains would be larger than to initiate partial rotation, thus leading to the formation of additional shear bands. Therefore, the total strain at 77K would comprise of more shear bands and smaller strains per shear band, which would delay the fracture. At this point, the origin of the abnormally large grains that completely replaced the original NT structure in Region I of the sample tested at 77K is unclear; however, observations from previous studies could elucidate on the interpretation of the observed microstructure. For example, Zhang et al. showed that abnormal grain growth was observed near the surface of an indent in nc-Cu indented at -190° C [33]. They reported that this growth was primarily stress driven and that the rate of growth also increased with the sample purity. Additionally, an in situ study by Wang et al. showed that changes in the NT-Cu microstructure occurred during testing at 120K [109]. However, it is also possible that the grains grew post-testing when being brought up to RT [110-112] due to the high strains in the localized region and the enhanced storage of dislocations inside the twins [35]. 82 Another possibility for the abnormal grain growth is that this structure formed as a result of the adiabatic nature of the localized deformation in this region. Therefore, the grain growth in the sample tested at 77K was a dynamic process, occurring during deformation. In copper, temperature rises in the range of 500-800K have been observed in adiabatic shear bands formed during high strain/high strain rate tests [113-114]. The localized temperature rise was thought to lead to static recrystallization [113] and grain coarsening [114]. Recently, a study by Zhao et al. on the thermal stability of NT-Cu samples [115] showed that at temperatures above 300° C, large grains, very similar to those found here, formed and began to replace the original NT structure. Consequently, if the temperature inside the adiabatic shear bands was to exceed this temperature, abnormal grain growth might occur. In order to explore this possibility, an equation predicting the temperature rise inside the adiabatic shear bands was used: Equation (7) where β is the fraction of work that is converted to heat, ρ is the density, σ is the flow stress, and ε is the strain [116]. For the sake of obtaining a maximum value, the temperature rise was calculated assuming that once all shear bands initially formed, the remaining global strain (observed in the stress-strain curve in Ref [91]) was localized to Region I. Also assuming that β=1, the temperature rise was calculated to be 1,340K. Although this value is clearly a theoretical maximum, its magnitude is much higher than was originally expected. Interestingly, if only half of the post-yielding strain was localized to this region, and a more realistic value of β=0.9 is used [117], the temperature rise would still be approximately 650K. It is still unclear if the large grains that formed aided in the enhanced ductility or not; 83 however, it is expected that once they formed, a continued softening would occur and catastrophic failure would ultimately result in this region. In this HEMD study, critical regions of localized deformation and microstructural changes in highly NT-Cu have been identified. Shear-banded areas exhibited a deformation texture, identifying dislocation-mediated plasticity and grain rotation as the governing deformation process. At RT, deformation is more localized than at 77K. The data acquired for the sample tested at 77K allowed for the formulation of several theories regarding the apparent increase in ductility due to the formation of multiple shear bands. 4.5. NT Stability, Detwinning, and Fracture After characterizing the microstructural evolution and shear band formation in the NT- Cu, previously presented in sections 4.3 and 4.4, a similar FIB study was performed with focus on the regions near and along the fractured edges. This was carried out in order to further investigate the overall stability of the columnar grained highly NT-Cu in the areas of highest strain, and to elucidate on the mechanisms involved in the eventual fracture. A collaborative effort was involved in this investigation, which included tests performed in tension, high pressure torsion, compression, and tension-tension fatigue. More detailed test descriptions and analysis can be found in the published version of this work in Ref [18]. Only the examination of the tensile tested NT-Cu samples will be presented here. The HEMD study revealed that a completely detwinned micro-grain structure replaced the original columnar grained highly NT structure within 50μm from the fracture edge. However, the DPs obtained in the samples tested at RT did not show any obvious loss of the NT structure or grain coarsening. It should be noted that the X-ray scans where the DPs were 84 obtained had a step size of 20-30 μm; therefore, if any NT instability or loss of the columnar structure occurred relatively close to the fractured edge (within approximately 30 μm), the X- ray scans might have missed these regions. Therefore, to further investigate the NT stability and fracture in the samples tested at RT, a FIB microstructure analysis, similar to that discussed in section 4.4.1, was performed within 30 μm from the fractured edge and along the actual fracture sites. Figure 50a shows an optical image of the gauge section of a fractured NT-Cu specimen, where the area of interest is highlighted by the black box. In this region, regularly spaced surface depression lines (vertical lines in Fig. 50b) were observed running parallel to the fractured edge (right side of Fig. 50b). In the center of Figure 50b, a FIB trench prepared for cross-section imaging is shown centered across a depression line that is approximately 30 µm away from the edge. In Figure 50c, the cross-section micrograph depicts a detwinned nano-grained structure (highlighted by the rectangle) which can clearly be seen under the surface depression. A representative micrograph of the nanograined structure can be seen in Figure 50d. It should be mentioned that FIB cross-section analysis was also performed directly along the fractured edge (within 1 µm from the edge) and the same nano-grained type structure was observed. It is interesting to note that on either side of the nano-grained regions (in between the surface depression lines), the NTs are mostly intact, and that the main change in the microstructure is a stretching/blending of the columnar grains, coupled with twin misregistry on either side of the boundaries. Similar observations were made by Shute et al. in which a NT-Cu sample fatigued to failure showed a nc grain structure underneath surface dips, while the majority of twins outside of the dips survived with approximately their original orientation [31]. 85 Figure 50. (a) Optical micrograph of a gage section after fracture for a representative tensile sample tested at 10 -4 /s. (b) Planar view of the FIB trench prepared for cross-section imaging. (c) FIB cross-section centered at the trench where a surface depression line near the fractured edge shows detwinning. (d) Representative magnified FIB image of a detwinned region. The microstructural observations in the tensile deformed NT-Cu samples imply that the growth-twinned microstructure remains relatively stable, even throughout the localized- strain regions in the developed shear bands. Eventually, however, there are severe changes in the twinned structure (observed as detwinned nano-grained structures) in the areas of highest strain, which seems to lead to the final fracture along these highly deformed regions. The exact mechanisms involved in the transformation from the columnar grained highly NT structure to the nano-grained equiaxed structure are still unclear. However, it is reasonable to 86 assume that the original formation of the shear bands by dislocation mediated grain rotation, discussed in section 4.4, was the dominant deformation mechanism that allowed the samples to produce the 7-8% elongation-to-failure at RT. Once the majority of the strain was effectively localized to these shear band regions, some additional plastic instability occurred in the form of the total loss of the NT structure, which led to the eventual fracture in these regions. 4.6. Summary In this chapter the mechanical response of highly NT-Cu foils was evaluated by detailed FIB microstructure analysis and HEMD. The synthesis of the NT-Cu was briefly summarized, with emphasis on the interrupted sputtering technique performed at LLNL that allowed for the development of thick (approx. 170 μm) highly NT columnar grained foils. The thick foils enabled larger-scale mechanical testing to be performed, as opposed to small- scale (e.g. nanoindentation) techniques, in order to explore the deformation behavior across multiple length scales. Tensile tests were performed on the highly NT-Cu foils at RT and 77K, and it was revealed that the samples tested at 77K showed both higher strength and ductility. The overall deformation behavior was observed to be localized in the form of shear bands shortly after yielding, with two predominant bands occurring at RT and multiple bands forming at 77K. The HEMD study, coupled with FIB microstructure analysis, led to the formulation of several theories regarding the formation of the shear bands and the apparent ductility in the samples, especially at 77K. It was shown that the microstructural texture that developed in the shear banded regions was due to dislocation-mediated plasticity, likely in the form of grain rotation. This rotation led to the shear bands, which each could carry a certain 87 amount of strain in the sample. The larger density of shear bands that developed at 77K allowed for a larger total strain to be produced before fracture. In the RT sample, fracture eventually occurred by first transforming the columnar grained NT structure into a less-stable nano/ultrafine grained structure. However, apart from the localized regions within the fractured areas, the NT structure remained stable and persisted even in areas of high strain. The samples tested at 77K revealed a coarsened grain structure in the regions of highest strain, which could be explained by the adiabatic nature of the shear bands at 77K. This work has led to a more thorough understanding of some of the deformation mechanisms and overall mechanical response of NT-Cu at lower temperatures. 88 (blank page) 89 Chapter 5 : Synthesis and Mechanical Behavior of other NT Metals Widespread efforts have been made to investigate the behavior of nanotwinned (NT) metals, due to their potential for high strengths while also maintaining ductility. For example, as shown in chapter 2, reports by K. Lu et al. on NT-Cu showed continuous increases in both yield strength and elongation-to-failure with decreasing twin spacing [4]. This result was in contrast to nc-Cu, which generally shows a dramatic loss of ductility with increasing strength as the grain size is reduced to the nano-regime [3]. Furthermore, results of some of the current work on highly NT-Cu, presented in chapter 4, showed increases in both strength and ductility with decreasing temperature. However, the majority of the work in trying to understand the unusual strength-ductility relationship in NT metals has generally been limited to NT-Cu; in addition, attempts to control the mechanical properties in NT metals have been mainly focused on changing the average twin spacing. Moreover, many of the mechanical data for highly NT metals, especially those for other (i.e. non-Cu) NT metals have been produced by nano-indentation [30, 61, 118-119], due in part to the challenges in producing thicker ( > 20 µm) samples. Therefore, many questions still exist regarding the potential for NT metals, such as whether or not the same trends observed in the strength and ductility for NT-Cu can be extended to other fcc metals. Additionally, are there other practical means of tailoring the mechanical response, in particular the strength and ductility, using other types of microstructural control rather than adjusting the twin spacing? Is there an optimum microstructural design for controlled mechanical deformation in NT metals? To explore some of these questions, 20-30 µm thick highly NT foils were produced in two different metals, Ag and Cu-6wt%Al, with various microstructures. The samples were mechanically tested by 90 microindentation and tensile tests and the overall deformation behavior, with emphasis on the strength and ductility, was compared to the highly NT-Cu foils presented chapter 4. 5.1. NT-Ag Ag serves as a good candidate for studying the effects and possible benefits of growth- NTs in other pure fcc metals, since it has low SFE (approx. 17 mJ/m 2 ) which facilitates NT growth during sputtering, as discussed earlier in section 2.5.3. In the present study, the NT- Ag foils were sputtered to have various volume fractions of highly NT grains, ranging from 10% to 100% vol%. This allowed for a direct comparison of the fully NT samples to the NT- Cu foils presented in chapter 4, while also exploring different means to tailor the mechanical properties using a bimodal microstructure approach. Additional details can be found in the published version of this work [120]. 5.1.1. Synthesis The Ag foils were synthesized by magnetron sputtering, as presented in chapter 2.5. For these samples, a static approach (non-interrupted sputtering) was used to directly sputter the Ag onto 50.8 mm diameter x 325 µm thick Si (100) single crystal wafers. A base pressure of approximately 2.7 x 10 -5 Pa (2.0 x 10 -7 Torr) was achieved prior to sputtering, and the deposition was performed using a 7.6 cm diameter x 12.7 mm thick pure (99.99+ %) Ag target mounted to a 76 mm diameter Meivac MAK magnetron sputtering source. The working pressure of the Ar gas was 0.27 Pa (2 mTorr) and a power of 200 W was used at a 14.6 cm working distance, which produced a deposition rate of approximately 1.5 nm/s. Although all of the Ag foils were sputtered under similar conditions, characterization of the sputtered microstructures revealed different volume fractions of highly NT grains. As 91 mentioned in chapter 2.5.2, the growth of NTs during deposition, especially in sputtering relatively thick foils, is very sensitive to the parameters and initial substrate conditions. Through analysis of the process, it was later revealed that the foils deposited onto the wafers with the largest initial curvature produced fewer NTs during growth. This was qualitatively attributed to a decreased thermal conductivity (and, thus, a larger buildup of film surface temperature) due to the decreased contact area between the substrate back and the substrate holders. However, the exact mechanisms of the effects of film temperature on the growing film and NT development are not fully understood and require further work. Nonetheless, the varying volume fraction of NT grains allowed for the exploration of a bimodal distribution of NT and non-NT grains and the overall effects on the mechanical behavior. 5.1.2. As-sputtered Sample Microstructures The 26 μm thick as-sputtered Ag foils consisted of varying volume fractions of highly NT ultra-fine columnar grains of 10%, 40%, and 100% vol%. The average columnar grain widths were 600 nm for the 100% twinned foil and 1µm for the other two samples, while the average twin spacing in the NT grains for each sample was approximately 10 nm. Figure 51 presents FIB cross-sectional micrographs of the as-sputtered samples showing the various amounts of highly NT grains, while Figure 52 shows a representative TEM image of one of the NT grains. Out-of-plane XRD measurements were also performed and generally showed that the fully NT foil consisted of a highly (111) out-of-plane texture, and a slightly more random average grain orientation was observed in the other samples. The extent of the random grain orientation was found to increase with decreasing volume fraction of NT grains (i.e. slightly lower (111) texture in the samples with fewer NT columnar grains). 92 Figure 51. Representative FIB cross-sectional micrographs of as-sputtered Ag foils containing approximately (a) 10%, (b) 40%, and (c) 100% highly NT columnar grains. Figure 52. Representative TEM cross-sectional image of a portion of a NT grain in the 100% NT grained Ag. 5.1.3. Mechanical Performance Two different tests were performed on the various Ag foils in order to evaluate the strength, ductility, and overall deformation behaviors. Vickers microindentation was used to initially estimate the hardness of the foils and a cross-sectional FIB microstructural analysis underneath the indents was performed. In addition, the samples were tested in tension and the stress-strain response, deformation modes, and fracture were analyzed by SEM. 93 Vickers Microindentation Vickers micro-indentation was performed on each foil using a LECO LM-100, as described in section 3.2.2. A force of 10 g was used in order to keep the indent depth below 1/10 of the foil thickness. Figures 53 (a1-a3) show representative SEM surface images of indents in the 10%, 40%, and 100% NT samples, respectively. The first observation was the decreasing indent size, corresponding to an increase in hardness, with increasing volume fraction of NT grains – the average values for the flow stress (approximated by the hardness divided by 3), averaged over at least 5 indents each, were 176 MPa, 286 MPa, and 399 MPa for the 10%, 40%, and 100% NT samples, respectively. The hardness results are not surprising as TBs have been shown to effectively block dislocations [43, 46]. The increase in NT grain fraction leads to a larger overall volume of TBs, and thus higher strengths were achieved. A similar result has been observed by nano-indentation on 330 stainless-steel containing different volume fractions of NT grains [29]. 94 Figure 53. (a) SEM surface micrographs of Vickers indents and (b) FIB cross-sectional micrographs underneath the indents of the foils containing (1) 10%, (2) 40% and (3) 100% NT grains. 95 The indent profiles in Figure 53 (a1-a3) also show a transition in shape from an almost perfect indent for the 10% NT sample, to a more rounded barrel-shape for the fully NT foil. This barrel-shaped impression in Vickers indentation typically occurs from the material flowing upward and piling against the indenter face [121], and is most commonly found in materials with limited work-hardenability [6, 122-123]. Since work-hardening in a material is often coupled with the ductility [27, 124], the results here qualitatively suggest that with increasing the volume fraction of NT grains, the ductility decreases. To further explore the deformation, FIB cross-sectional micrographs were also produced directly underneath each indent, shown in Figures 53 (b1-b3). In the 10% NT sample [Fig. 53 (b1)], FIB images reveal a contrast gradient within the majority of the grain interiors which is indicative of dislocation generation and accumulation (i.e. plastic deformation) within the individual grains [89, 125]. The fully NT sample, shown in Figure 53 (b3), however, does not produce this deformation contrast, but rather the majority of deformation is found in-between the columnar grains, in which columnar grains begin to bend (seen in the outer portions of the indent) and merge (directly underneath the indent). In the merged regions, the NTs on either side of the columnar boundaries have become out of registry and some slight thickening of the twins is observed. This type of deformation has been previously reported for columnar grained NT-Cu under various loading scenarios [18, 31, 90, 126-127], as discussed earlier in chapter 4, and was generally attributed to dislocation accumulation in or near the columnar boundaries [31]. In the sample containing 40% NT grains, shown in Figures 53 (a2) and (b2), a mixed mode of deformation is observed. In this case, a slight barrel-shaped indent profile occurs at the surface, and the microstructure reveals evidence of dislocation activity within 96 the non-NT grain interiors, in addition to similar bending of columnar grains and slight distortion of the grain boundaries in the highly NT grains. Tensile Tests To further evaluate the deformation behavior of NT-Ag, tensile tests were performed on dogbone specimens of 3 mm x 6 mm gauge sections at a strain rate of approximately 10 -3 /s using a constant crosshead speed, similar to the methods described in chapter 3.2.1. Figure 54 shows the true stress-true strain curves for the various samples while Figure 55 shows a representative optical micrograph of one of the fractured Ag foils. Figure 54. True stress-true strain curves of the Ag foils containing various volume fractions of highly NT grains. Figure 55. Representative optical micrograph of a tensile tested Ag sample showing fracture in the middle of the reduced section. 97 In the stress-strain curves in Figure 54, a monotonic increase in yield strength was observed with increasing fraction of highly NT grains. The average yield strengths (averaged over a minimum of three different tests for each sample) were 164 MPa, 215 MPa, and 500 MPa for the 10%, 40%, and 100% NT grained samples, respectively. These results agree with the trend of the flow stress values estimated from Vickers micro-indentation (see Table III for a summary of the properties). The stress-strain curves also reveal that the amount of uniform plastic strain decreased with increasing volume fraction of NT grains, from roughly 18% in the 10%NT sample to less than 1% for the fully NT sample. In addition, the average rate of work hardening tends to decrease as well. Table III. Summary of the microstructural feature sizes and mechanical properties of the Ag foils. Microstructure Vickers Tensile vol. fraction of NT grains d a [ μm ] λ b [nm] H/3 [MPa] σ y [MPa] 10% 1.3 10 176 ± 5 164 ± 2 40% 1.1 11 286 ± 16 248 ± 8 100% 0.6 9 399 ± 33 508 ± 23 a d is the average columnar grain width b λ is the average twin spacing within the NT grains Figures 56 (a1-a3) show SEM images of the top view of the fracture (only one half of the dogbone specimen shown here) in the 10%, 40%, and 100% NT grain Ag foils, respectively, while Figures 56 (b1-b3) show the cross-sectional view of the fracture surfaces. The overall fracture of the samples [Figs. 56 (a1-a3)] indicates that the behavior of the 10% and 40% NT grain foils are similar since both show reduction in gauge width, corresponding to the uniform plastic strains observed in the stress-strain curves, and a slightly serrated 98 fracture surface. In contrast, the fully NT sample [Fig. 56 (a3)] shows less uniform strain, in which there is only a slight reduction in width, in addition to a shear-band-like type of fracture, similar to NT-Cu foils presented previously in chapter 4. Furthermore, the 10% NT sample [Fig. 56 (b1)] shows much more necking in the thickness direction, while the fully NT foil [Fig. 56 (b3)] shows fracture mostly by void growth and coalescence [85] with little to no necking. The 40% NT grain sample displays a combination of some necking and void coalescence, indicated by the slight taper in the fracture surface and the dimples in the middle of the foil. Figure 56. (a) top view and (b) cross-sectional view of the fractured dogbone specimens containing (1) 10%, (2) 40%, and (3) 100% volume fraction of highly NT grains. The results of this study indicate an increase in strength with increasing volume fraction of NT grains in the Ag foils. However, the overall ductility is adversely affected by the highly NT grains. The Ag foils with 10% NT grains are more ductile since they show (i) evidence of dislocation activity/accumulation within grain interiors underneath micro- indents (seen by the FIB deformation contrast); (ii) more uniform plastic strain and some 99 work hardening in the stress-strain curves; and (iii) necking in both the width and thickness directions. In contrast, a change in the deformation behavior was observed in the fully NT samples, demonstrated by (i) the resistance to plastic flow in response to micro-indentation; (ii) limited uniform plastic strain and almost no work-hardening in tension; and (iii) little to no necking in either width or thickness directions, coupled with a shear-band type of fracture. The sample with 40% volume of NT grains showed a mixed behavior in which strengthening occurred while maintaining some ductility. The plastic response of the Ag foils in this study is in contrast to the original reports by K. Lu et al. for NT-Cu, wherein decreasing the twin spacing resulted in a continuous increase in ductility [128]. Based on that study, one might expect a similar result by increasing the volume fraction of highly NT grains, since the overall density of TBs in the sample would be greater. However, this was not observed in the current Ag samples, which could be due in part to the columnar structure and alignment of the TBs parallel to the loading direction [36, 48]. Nonetheless, it is shown here that the various mixtures of NT and non-NT grains led to different material responses and that a sample containing only 40% NT grains can provide higher strengths while still maintaining some ductility. Therefore, it is shown here that although NTs can be used to change the material behavior in Ag, it may be necessary to introduce a bi-modal microstructure in order to accomplish a combination of desired properties, such as strength and ductility. This approach may provide practical means to tailor material behavior and develop optimum microstructural designs for advanced applications. 100 5.2. NT-CuAl In order to evaluate the potential of using growth-NTs in alloyed materials, Cu-6wt%Al foils were synthesized, characterized, and mechanically tested. The composition of the Cu-6wt%Al foils was chosen for its low SFE ( ~ 6 mJ/m 2 ) [76] and its solid solution state, in which only a single fcc phase would develop [77]. Using magnetron sputtering, approximately 26 μm thick foils of both highly NT-Cu-6wt%Al (NT-CuAl) foils, similar to the NT-Cu and NT-Ag samples presented earlier in this work, and non-twinned ufg Cu-6wt%Al (ufg-CuAl) foils were synthesized. Vickers microindentation and tensile tests were performed on both sets of NT and ufg CuAl samples to investigate the mechanical performance and deformation behavior. 5.2.1. Synthesis The CuAl foils were synthesized by magnetron sputtering using a static (un- interrupted) sputtering approach, similar to the NT-Ag foils presented in section 5.1. A base pressure of approximately 1.3 x 10 -4 Pa (1.0 x 10 -6 Torr) was achieved prior to sputtering, and the deposition was performed using a 7.6 cm diameter x 6.35 mm thick Cu-6wt%Al target mounted to a 76 cm diameter Meivac MAK magnetron sputtering source. The working pressure of the Ar gas was 0.27 Pa (2 mTorr) and a power of 160 W was used at 12.7 cm working distance. The deposition rate was approximately 0.62 nm/s and the total foil thicknesses were approximately 26 μm. As was discovered in the synthesis of the NT-Ag foils, the development of the highly aligned NT microstructure was sensitive to the thermal conductivity between the substrate back and the substrate holder. It was determined here that by applying a uniform amount of thermal contact paste between the substrate back and the 101 substrate holder, the highly aligned NT structure would form. However, by omitting this step, it was also possible to sputter non-twinned ufg-CuAl with similar grain size as the NT-CuAl foils. This allowed for a direct comparison of the mechanical behaviors of the CuAl with and without NTs. 5.2.2. As-sputtered Sample Microstructure The as-sputtered NT-CuAl foils consisted of ultra-fine columnar grains with average grain widths of approximately 200 nm containing high densities of NTs oriented perpendicular to the growth direction, as shown in Figure 57a. The average twin spacing in the NT-CuAl was approximately 3 nm, as shown in the TEM micrograph in Figure 57b. Similar to the highly NT-Cu and NT-Ag foils, a (111) out-of-plane texture was observed by XRD measurements, revealing the extent of the (111) columnar grain texture in the growth direction. The ufg-CuAl foils, shown in Figure 58, consisted of approximately 200 nm grains, without a high density of NTs, and a slightly weaker (111) texture in the growth direction. Figure 57. (a) FIB cross-sectional micrograph and (b) TEM micrograph of an as-sputtered highly NT-CuAl foil. Note, the arrow shows the growth direction. 102 Figure 58. FIB cross-sectional micrograph of the as-sputtered ufg-CuAl foil. Note, the arrow shows the growth direction. 5.2.3. Mechanical Performance Vickers microindentation and tensile tests were performed on the NT and ufg CuAl foils in order to evaluate the strength, ductility, and overall deformation behaviors. The post- deformed microstructures were investigated using FIB cross-sectional imaging and SEM fracture surface analysis, similar to the previous studies on the NT-Cu and NT-Ag foils. Vickers Microindentation Vickers microindentation, as discussed in section 3.2.2, was performed on the CuAl foils using an indent force of 50 g and a dwell time of 10 s. The force was chosen to ensure that the indentation depth was less than 1/10 of the foil thickness. The indent profiles of the NT-CuAl and ufg-CuAl samples can be seen in Figure 59 a and b, respectively. 103 Figure 59. SEM surface images showing the Vickers indents in the (a) ufg-CuAl and (b) NT-CuAl foils. The first observation in comparing the two indent profiles is that the size of the indents is roughly the same, implying that the hardness of both materials is similar. In fact, the yield stresses of the NT-CuAl and ufg-CuAl were estimated (using σ y ≈ H/3, where H is the hardness) as 1.20 GPa and 1.17 GPa, respectively, with standard deviations of 0.04 GPa. Therefore, the two materials do not statistically differ in terms of the hardness and approximate yield stress. This result was unexpected, since the addition of NTs typically leads to much higher strength/hardness, as previously shown for both the NT-Cu and NT-Ag in this study. The second observation in comparing the two indent profiles is that a more rounded barrel-shaped indent developed in the ufg-CuAl foils, whereas a relatively perfect indent shape is found in the NT-CuAl samples. As discussed earlier in section 5.1.3, the barrel-shaped indent and material pile-up in Vickers indentation typically implies limited work hardenability in the material. Therefore, if only the shape of the Vickers indents is considered, these results might imply that the strengths of ufg-CuAl and NT-CuAl are similar, while the NT-CuAl foil is more ductile. However, to expand on this and to further investigate the deformation in the samples, FIB microstructure analysis underneath the indents was also performed. 104 Figure 60 shows the FIB cross-sectional micrograph of the NT-CuAl sample underneath a Vickers indent. In this case, the overall plasticity in the sample was very limited, as a network of cracks and severely detwinned regions were found throughout the cross-section. This result was different from the deformation in the NT-Cu and NT-Ag foils, presented earlier, in which some plastic deformation was observed without any crack formation [see, for example, Fig. 53 (b3)]. In those samples, evidence of dislocation activity was observed in the form of grain rotation and merging of columnar boundaries, which was not found here. In this case, the regions outside of the crack network were relatively unchanged, which further reveals the extent of the localization of the plastic instabilities in the NT-CuAl. The ufg-CuAl, shown in Figure 61, did not appear to develop the same type of crack network. Although the indent profiles shown in Figure 59 suggest more ductility in the NT-CuAl foils, the FIB micrographs under the indents suggest the opposite, in which the NT- CuAl was unable to accommodate the strain imposed by the indenter, thus forming the cracks. At this point, the origins and reasons for the plastic instabilities and cracking in the NT-CuAl foils are still unclear and require further work. In order to further investigate the mechanical properties, deformation behavior, and plasticity of the CuAl foils, tensile tests were also performed. 105 Figure 60. FIB cross-sectional micrograph underneath a Vickers indent in the highly NT-CuAl foil showing a detwinned crack network. Figure 61. FIB cross-sectional micrograph underneath a Vickers indent in the ufg-CuAl. 106 Tensile Tests Tensile tests were performed on dogbone specimens with 3mm x 6mm gauge sections at a strain rate of approximately 10 -3 /s using a constant crosshead speed, similar to the methods described in chapter 3.2.1. Figure 62 shows the engineering stress-strain curves for multiple NT and ufg CuAl tensile specimens, while Figure 63 shows a representative optical micrograph of one of the fractured NT-CuAl foils. It should be noted that all fractured tensile specimens for both NT and ufg CuAl were similar to that shown in Figure 63. Figure 62. Engineering stress-strain curves for multiple NT-CuAl and ufg-CuAl foils. Figure 63. Representative optical micrograph of a tensile tested NT-CuAl sample showing fracture in the middle of the reduced section. 107 The stress-strain curves in Figure 62 reveal that the tensile behavior of both sets of foils is similar in that both the yield strengths and percent elongation are almost identical. Both samples fractured in a semi-brittle fashion directly after yielding with a total plastic strain of less than 1%. The results for the yield stress agree well with the approximated yield strengths from the Vickers indentation, as summarized in Table IV. Contrary to most reports on pure NT metals, including the NT-Cu and NT-Ag foils in this study, the high densities of NTs in the CuAl foils did not enhance the strength. In addition, although the deformation by Vickers indentation showed a slight difference in the plastic deformation, the tensile tests revealed similar amounts of ductility, plastic flow, and overall fracture behavior. Table IV. Summary of the microstructural feature sizes and mechanical properties of the NT-CuAl and ufg-CuAl foils. Microstructure Vickers Tensile sample d a [ μm ] λ b [nm] H/3 [GPa] σ y [GPa] NT-CuAl 0.2 3 1.20 ± 0.04 1.33 ± 0.09 ufg-CuAl 0.2 - 1.17 ± 0.04 1.42 ± 0.09 a d is the average columnar grain width b λ is the average twin spacing within the NT grains These results are among the first for the mechanical properties of highly columnar grained CuAl foils containing high densities of NTs, and the data for NT alloyed materials, in general, is extremely limited. One report on NT-330 stainless-steel revealed increases in hardness with the addition of highly NT grains [29]; however, the columnar grain size in that study was an order of magnitude lower than the columnar grains here, and the film thickness was only 1μm. Therefore, the samples cannot be compared directly. In addition, since only nanoindentation was performed on the NT-330 stainless-steel films in that study, and only the 108 hardness values were reported, the deformation behavior and ductility was not discussed. The results of the NT-CuAl foils in this study are not fully understood and will likely require more in-depth studies, including MD or meso-scale simulations. However, in order to examine one possibility for the observed lack of strengthening in the NT-CuAl foils, the data for the flow stress of various CuAl foils (non-NT samples) from literature were plotted as a function of grain size [129-134], as shown in Figure 64. It should be noted that in comparing the Vickers hardness values to the hardness values determined by nanoindentation in the other studies, the Vickers hardness values for the CuAl samples was multiplied by 1.25, as a 25% difference between Vickers hardness and nano-indentation is commonly observed [135]. In this case, a grain size of 200 nm was used for the ufg-CuAl foils, while the twin spacing of 3 nm was used as the grain size in the NT-CuAl samples. Although the data for this alloy is limited, especially in ultra-fine and nano regimes, a possible softening effect is observed in the plot for grain sizes below 100 nm. This breakdown in the Hall-Petch (H-P) relationship has been observed for a variety of metals with reduction in the grain size to the nano- regime [38]. Therefore, if a softening effect occurs in this material, and if the TBs act similarly to grain boundaries as strengthening agents, which is commonly observed for other NT metals [35], only a moderate increase in the strength might be seen between the ufg and NT-CuAl foils. However, much more data is required to properly characterize the H-P plot for this material to determine the possible strengthening mechanisms below 100 nm. 109 Figure 64. Collection of data for flow stress as a function of grain size [129-134], including the ufg-CuAl and NT-CuAl foils in this study, where d=200nm is used for the grain size of the ufg- CuAl, and d=λ=3nm is used for the NT-CuAl. To further evaluate the deformation, NT microstructural stability, and fracture in the tensile tested foils, FIB cross-sectional imaging and SEM surface imaging were performed on the fractured samples. Figure 65a shows a representative cross-sectional FIB trench that was prepared on the fracture edge in the NT-CuAl foils. As can be seen in the higher magnification cross-section image in Figure 65b, minimal changes were observed in the overall NT columnar structure outside of fracture; in this case, the fracture was localized to one area and was found to cut through the columnar grains, without any obvious plastic deformation in the microstructure. [It should be noted that in Figure 65b, only some NT grains are seen; however, when the sample is rotated to produce a different orientation with respect to the ion beam, more NT grains are revealed and are found to be very similar to the 110 as-sputtered case]. Additionally, the fracture edge shown in Figure 65a reveals little evidence of ductile fracture in the form of ductile dimples and voids [85]. Figure 65. FIB micrographs [(a) lower magnification and (b) higher magnification image] showing a cross-section trench prepared on the fracture surface of a tensile tested highly NT-CuAl foil. In some regions near the fractured edge, well developed cracks were also observed, as shown in Figure 66. The microstructure within these cracked regions revealed two main deformation features. The first (shown in Fig. 66b) is the development of a detwinned vein that extended through the columnar structure at an angle of approximately 45° relative to the tensile direction. This feature is similar to the detwinned crack network observed underneath the Vickers indentation, previously shown in Figure 60. The other feature found in this region, shown in Figure 66c, is a well-developed crack with the same approximate angle relative to the testing direction, and parallel to the main fracture edge. In this case, the fracture appears to be somewhat stepped/serrated, in which the path of the crack seems to travel down the vertical columnar boundary and then traverse along a TB, and then down the opposite columnar boundary, etc. However, TEM work is required to further evaluate the crack development and propagation in these samples. 111 Figure 66. FIB micrographs showing a cross-section trench prepared along a crack near the fractured surface of the tensile tested highly NT-CuAl foil. The combination of the Vickers indentation, tensile tests, and microstructural investigation for the CuAl foils revealed a much different and somewhat unexpected deformation behavior compared to the pure NT metals (NT-Cu and NT-Ag). In particular, the lack of strengthening with the addition of NTs was different from most other reports for NT metals, and the overall plastic instability and the propensity for cracking in the highly NT foils has not been observed for other NT metals. Further work is still required to fully understand the effects of the NTs, the contributions of any possible solid solution effects, and the dominant deformation mechanisms (i.e. GB mediated mechanisms vs. dislocation mechanisms) in the NT-CuAl samples. This will likely require a combination of high resolution TEM studies, MD simulations, and meso-scale modeling in the future. 112 5.3. Summary In this chapter, the synthesis and deformation behavior of two additional (i.e. non-Cu) NT systems, NT-Ag and NT-CuAl, were presented in order to determine if the trends observed in the mechanical performance of NT-Cu can be extended to other metals. The Ag was chosen as a representative low SFE pure fcc metal, while the CuAl was selected to study NTs in an fcc binary alloy with low SFE. It was revealed that the low SFE of both materials facilitated NT growth during deposition but the effects of film temperature and thermal conductivity between the substrate holder and the substrate back were observed to change the overall volume fractions of highly NT grains. This allowed for direct comparison of the mechanical behavior of both highly NT and non-NT samples in both metals. Vickers microindentation and tensile tests revealed that the deformation mechanisms of highly NT-Ag were similar to those of highly NT-Cu. In this case, strengthening was achieved with the addition of the NTs, which was attributed to the strengthening mechanisms inherent to the twin boundaries. However, the overall ductility of the NT-Ag, compared to non-twinned ufg-Ag, was greatly reduced, which was the opposite trend for reports of concurrent increases in strength and ductility in NT-Cu. However, the introduction of a bimodal distribution of NT and non-NT grains was shown to produce a combination of some strengthening, compared to the non-NT sample, while maintaining some ductility. The bimodal approach was revealed to be a promising technique to potentially tailor the mechanical properties through microstructural control. In contrast, the NT-CuAl showed no difference in hardness, yield strength, or ductility when compared to the non-twinned ufg-CuAl. In addition, the NT-CuAl showed very limited plasticity and semi-brittle fracture, in which plastic instabilities and cracking were prevalent 113 underneath Vickers indents and in the tensile tested specimens. This result was different from most other reports for pure NT metals, including the NT-Cu and NT-Ag presented in this work. The exact mechanisms related to the plastic instabilities and lack of any strengthening or ductility with the addition of NTs is still not fully understood and will likely require more in-depth simulation and TEM studies. 114 (blank page) 115 Chapter 6 : Conclusions and Future Work In summary, various columnar grained highly NT metals were successfully synthesized using magnetron sputtering. The fabrication of relatively thick ( > 25 µm) NT foils allowed for different mechanical testing techniques to be performed, which enabled the deformation behavior across multiple length scales (i.e. beyond nano-scale deformation mechanisms) to be evaluated, with an emphasis on the contributions of the NTs on the strength, ductility, and mechanical stability. A variety of mechanical tests were performed on 170 µm thick ultra-fine columnar grained NT-Cu and the deformation was extensively investigated using FIB microstructural imaging and HEMD. Tensile tests performed at RT and 77K revealed many unique features, not commonly observed for pure Cu, such as observations of yield peaks in the stress-strain curves, deformation by shear banding, and enhanced strength and ductility at low temperatures. In the samples tested at 77K, the propensity for forming shear bands was observed to be much greater than at RT, in which multiple shear bands developed at 77K, compared to only two predominant bands at RT. HEMD coupled with FIB microstructure analysis allowed for the investigation of possible mechanisms responsible for the shear band formation and enhanced ductility at 77K. It was shown that dislocation-mediated grain rotation was likely the cause of the shear localization and a slower grain rotation, coupled with a decreased heat capacity at 77K, led to the formation of multiple adiabatic-type shear bands. The larger density of shear bands at 77K, compared to RT, enabled a larger total plastic strain to develop before fracture. Detailed microstructural imaging of the tensile samples also revealed an overall stability of the NT structure at both testing temperatures, even in areas of high plastic strain (e.g. within the localized shear bands). Eventually, 116 however, a severely detwinnned structure was observed to replace the original columnar grained NT structure, with nano-grains along the fracture in the RT samples, and coarsened micro-grains in the sample tested at 77K. It was proposed that the rotation of the columnar NT grains resulted in the formation of the shear bands, and once enough strain was accumulated within the shear bands, a transformation of the original structure to a less mechanically stable structure eventually led to fracture in the samples. In addition, 26-27 μm thick columnar grained NT-Ag and NT-Cu-6wt%Al foils were synthesized and mechanically tested using Vickers microindentation and tensile tests. The Ag served as a comparison of another pure fcc metal with low SFE, while the Cu-6wt%Al (CuAl) allowed for the exploration of the influence of NTs on the deformation behavior in alloyed fcc systems. In both samples, non-NT ufg foils were also produced in order to directly compare the mechanical behavior of the foils with and without high densities of NTs. It was revealed that the NT-Ag foils deformed similarly to the NT-Cu, with deformation occurring mostly by shear band-like behavior and fracture occurring along one predominant shear band. Additional evidence of some plasticity was also observed in the form of dislocation accumulation at the grain boundaries, slight rotation/tilting of the columnar grains, and some detwinning. When comparing the stress-strain response of the NT-Ag to that of the ufg-Ag, it was revealed that the strength increased by approximately 3 times with the addition of the NTs; however, the overall ductility and plastic strain in the NT-Ag foils was limited. It was also shown that by using a bimodal distribution of fully NT and non-NT grains in the Ag foils, strengthening was promoted, and some ductility was maintained. The deformation of the NT-CuAl foils was different from that of the NT-Cu and NT-Ag in that no change was found in the mechanical properties between the fully NT-CuAl 117 and the non-twinned ufg-CuAl. This result was among the first reports of the deformation behavior of a binary alloy with a high density of growth NTs, and it is in contrast to most other reports of enhanced properties, in particular the strength, with the addition of growth NTs. The microstructural investigation underneath Vickers indents and in the tensile tested samples revealed plastic instabilities in the NT-CuAl foils, with fracture occurring in a semi- brittle fashion with little evidence of uniform plastic deformation. The tested samples were shown to contain severely detwinned veins and crack networks with little change in the original columnar grained NT structure outside of the localized crack regions. The lack of strengthening with the addition of the NTs and the limited plasticity in the NT-CuAl is not yet fully understood; however, this study generally reveals that other mechanisms, such as solid solution effects, may be dominating the plastic behavior in the CuAl. This work has allowed for many questions that were still unanswered for the potential of metals containing high densities of growth NTs to be addressed. For example, are there practical means for bulk synthesis and control of highly NT structures? What is the larger scale (i.e. beyond nano-indentation) deformation behavior of NT metals? Can the trends observed in highly NT-Cu be extended to other fcc metals, including alloys? Throughout this work, it was shown that by using magnetron sputtering, including both static and interrupted sputtering techniques, highly NT thick foils (up to 170 μm and beyond) with very high (99.999%) purity, low initial defects and dislocation densities, and low internal stresses can be successfully produced in a variety of fcc metals. Using this technique, alloyed materials of low SFE can also be sputtered to produce highly NT structures, while maintaining the desired compositions. In addition, by adjusting the sputtering conditions, bimodal distributions of fully NT and non-NT grains can be achieved, 118 which may allow for practical means to control the mechanical response, in particular the strength and ductility. Recommendations for future work related to the synthesis include the further systematic study of the effects of processing conditions on the development of highly NT structures during sputtering. In particular, the detailed characterization of film temperatures during growth at various conditions may enhance the understanding of the growth process and may lead to increased control over the developing NT structures. In addition, alloys of low SFE should be further explored in order to determine the potential of using NTs as a means to control the mechanical behavior of alloyed systems. The larger-scale deformation mechanisms of NT-Cu, NT-Ag, and NT-CuAl were evaluated using μm- and mm-scale mechanical tests in order to explore the deformation behavior and mechanical stability across multiple length scales. It was revealed that the presence of the NTs can dramatically affect the mechanical properties and overall deformation and can promote mechanisms not commonly observed in the conventional non- twinned metals. High strengths (up to 10 times that of coarse-grained counterparts) were commonly observed in the NT metals, in addition to a general propensity for deformation by shear banding, which are not commonly observed for most pure fcc metals. In the case of the CuAl foils, however, the larger-scale mechanical tests revealed similar behavior between the fully NT-CuAl and non-twinned ufg-CuAl. Recommendations for future work regarding investigating the larger-scale deformation behavior of highly NT metals include a more detailed examination of shear banding in pure metals and semi-brittle behavior of the NT-CuAl by MD and atomistic simulations. The detailed characterization of the NT structures in this work highlights critical features that should be implemented in the simulations. By comparing simulated results for the mechanical behavior to actual observed 119 phenomena (e.g. grain rotation and shear localization), the simulations can be optimized and in turn, used to deepen our understanding of the underlying deformation mechanisms in NT metals. Finally, the question of whether or not similar trends in the mechanical performance observed for NT-Cu can be extended to other metals was addressed by the synthesis and mechanical testing of the NT-Ag and NT-CuAl foils. It was generally found that the behavior of the NT-Ag was similar to that for the NT-Cu, and that it is expected that Ag will continue to show similar trends in the mechanical performance. However, the CuAl foils produced different behavior from that of the NT-Cu and NT-Ag foils, in that very little difference was observed in the plastic responses of fully NT and non-NT CuAl. Recommendations for future work include the evaluation of the mechanical behavior of NT-CuAl with varying grain sizes (beyond 1 μm) and larger twin spacing (10 to 100 nm) to determine if different deformation mechanisms can be promoted and controlled using NTs. As mentioned earlier, large-scale MD simulations may also provide great insight into the controlling mechanisms in the NT alloyed systems. This study has enhanced the knowledge and understanding of the synthesis, mechanical properties, and deformation behavior of highly NT metals. These materials have shown a great potential as a viable solution for developing strong, ductile, and stable metals and should be studied further in order to explore their prospective use in advanced engineering applications. 120 (blank page) 121 Appendix A : Summary of Sputtered Cu Samples Cu samples produced using interrupted sputtering sample power [W] Ar pressure [mTorr] sputtering rate 1 [nm/s] layer thickness [nm] number of layers total thickness [μm] characterization 2 cu-cu-1 55 2 0.20 2.6 318 8.1 prof cu-cu-2 50 2 0.18 8.2 100 1.2 prof, FIB, SEM cu-cu-3 100 2 0.31 8.7 206 1.8 prof, FIB, SEM cu-cu-4 200 2 0.72 10.9 134 1.5 prof, FIB, SEM cu-cu-5 50 10 0.15 11.7 136 1.6 prof cu-cu-6 100 2 0.31 8.5 206 1.6 prof, FIB cu-cu-7 100 2 0.31 7.8 117 0.9 prof, FIB cu-cu-8 100 2 0.31 7.3 639 4.6 prof, FIB INT-ML-1 3 70 (RF) 2 0.21 10.3 300 3.1 prof, FIB INT-ML-2 300 (RF) 10 0.58 9.5 300 2.9 prof, FIB INT-ML-3 187 (RF) 2 0.54 8.9 300 2.7 prof, FIB INT-ML-5 292 (RF) 5 0.53 8.8 300 2.6 prof, FIB INT-ML-6 146 (RF) 10 0.24 12 300 3.6 prof, FIB INT-ML-7 77 (RF) 5 0.12 6.2 300 1.9 prof 1 defined as the rate of deposition during the "on" times (equivalent to static sputtering rate) 2 characterization: profilometry (prof), FIB cross-section (FIB), SEM surface (SEM), XRD 3 "INT" samples were sputtered at KIT in summer 2010 - interruptions were produced using pneumatic shutters 122 Cu samples produced using continuous sputtering (no interruptions) sample power [W] Ar pressure [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization 1 twinned? INT-THICK-1 2 300 (RF) 2 4 0.81 26.2 prof, XRD, FIB, SEM fully INT-THICK-2 2 300 (RF) 2 4 0.82 26.5 prof, FIB fully INT-2011-00 3 300 (RF) 2 4 0.85 10.7 FIB fully INT-2011-01 300 (RF) 2 4 0.84 24.9 FIB partially INT-2011-02 300 (RF) 2 4 0.85 25.3 FIB, SEM partially INT-2011-03 300 (RF) 2 4 0.85 22.0 FIB, SEM partially INT-2011-06 300 (RF) 2 4 0.85 20.3 FIB no INT-2011-07 300 (RF) 2 4 0.84 22.8 FIB partially INT-2011-08 300 (RF) 2 4 0.82 22.0 FIB no Cu-2011-001 200 2 5 0.97 10.3 prof, XRD, FIB partially Cu-2011-002 250 2 5 1.21 10.5 prof, XRD, FIB partially Cu-2011-003 250 2 6 0.74 8.9 prof, FIB partially Cu-2011-004 600 2 6 1.60 8.7 prof, XRD, FIB partially Cu-2011-005 600 2 6 1.80 8.1 prof, XRD, SEM, FIB no Cu-2011-006 300 2 6 0.99 10.2 prof, XRD, SEM, FIB no 1 characterization: profilometry (prof), FIB cross-section (FIB), SEM surface (SEM), XRD 2 sample were sputtered at KIT in summer 2010 3 all “INT-2011” samples were sputtered at KIT in summer 2011 123 Cu samples produced using continuous sputtering (cont’d) sample power [W] Ar pressure [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization twinned? Cu-2012-001 200 2 6 0.74 0.7 n/a - rate run - Cu-2012-002 200 2 6 0.74 0.7 n/a - rate run - Cu-2012-003 352 2 7 0.83 0.8 n/a - rate run - Cu-2012-004 470 2 7 0.95 4.7 prof, XRD, FIB partially Cu-2012-005 532 5 7 0.93 4.7 prof, XRD, FIB partially Cu-2012-006 532 5 7 0.91 4.6 prof, XRD, FIB partially Cu-2012-007 532 5 7 0.93 10.0 prof, XRD, FIB partially Cu-2012-008 800 5 7 1.52 6.8 prof, FIB partially Cu-2012-009 276 2 6 0.80 4.0 prof, XRD, FIB partially Cu-2012-010 584 2 6 1.65 11.9 prof, XRD, FIB no Cu-2012-011 300 2 6 0.98 10.5 prof, FIB partially Cu-2012-012 200 2 6 0.63 9.0 prof, FIB partially Cu-2012-013 200 2 5 0.94 10.2 prof, SEM, FIB partially Cu-2012-014 200 2 4 1.56 11.2 prof, FIB no 124 (blank page) 125 Appendix B : Summary of Sputtered Ag Samples Ag samples produced using continuous sputtering sample power [W] Ar pressure [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization 1 twinned? Ag-LLNL-run1 2 200 2 5.75 1.38 7.9 prof, XRD, FIB partially Ag-LLNL-run2 400 2 5.75 2.52 6.8 prof, XRD, FIB no Ag-LLNL-run3 3 200 2 5.75 1.50 27.0 XRD, FIB see note 3 Ag-LLNL-run4 200 2 5.75 1.50 27.0 XRD, FIB see note 3 Ag-LLNL-run5 200 2 5.75 1.50 27.0 XRD, FIB see note 3 Ag-LLNL-run6 200 2 5.75 1.50 27.0 XRD, FIB see note 3 TF-Ag-USC-01 4 200 2 5.75 1.45 26.2 XRD, FIB no TF-Ag-USC-02 5 200 2 4.75 - - FIB partially 6 TF-Ag-USC-03 200 2 5.75 1.45 26.2 FIB partially 6 TF-Ag-USC-04 7 200 2 5.75 1.45 26.2 XRD, FIB partially 6 1 characterization: profilometry (prof), FIB cross-section (FIB), XRD 2 all "LLNL" labeled samples were sputtered at LLNL using chamber M-12 3 in samples Ag-LLNL-run3 through Ag-LLNL-run6, 2" wafers and 1" half pieces were installed at the same 4 samples labeled "TF-Ag-USC" were sputtered using chamber at USC 5 foil popped off substrate before the thickness and rate could be determined 6 highly NT structure was observed in first 10 μm, then non-NT grains were observed throughout the rest 7 this sample was sputtered using interruptions to try to maintain the NT structure that occurs in the first 126 (blank page) 127 Appendix C : Summary of Sputtered CuAl Samples CuAl samples produced using continuous sputtering sample composition power [W] Ar pressure [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization 1 twinned? CuAl- 2012-001 Cu- 4.5wt%Al 532 5 7 0.89 8.0 prof, edx, XRD, SEM, FIB no CuAl- 2012-002 Cu- 4.5wt%Al 352 2 7 0.68 5.0 prof, edx, XRD, SEM, FIB partially 2 CuAl- 2012-003 Cu- 4.5wt%Al 100 2 4 0.51 3.7 prof, XRD, FIB no CuAl- 2012-004 Cu- 4.5wt%Al 128 2 4.5 0.65 4.8 prof, XRD, FIB no CuAl- 2012-005 Cu- 4.5wt%Al 256 2 4.5 1.18 9.2 prof, XRD, FIB no CuAl- 2012-006 Cu- 4.5wt%Al 532 2 7 1.02 7.4 prof, XRD, FIB no CuAl- 2012-007 Cu- 4.5wt%Al 352 2 6 0.94 6.8 prof, XRD, FIB no CuAl- 2012-008 Cu- 4.5wt%Al 352 2 7 0.62 22.3 FIB no CuAl- 2012-009 3 Cu- 4.5wt%Al 352 2 6 0.86 15.5 FIB no TF-CuAl- 010 Cu-6wt%Al 160 2 5 0.68 9.8 XRD, FIB fully TF-CuAl- 011 Cu-6wt%Al 160 2 5 0.62 22.3 XRD, FIB fully TF-CuAl- 012 Cu-6wt%Al 160 2 5 0.60 23.8 XRD, FIB partially 2 TF-CuAl- 013 Cu- 4.5wt%Al 160 2 5 0.69 10.3 XRD, FIB partially 2 TF-CuAl- 014 Cu- 4.5wt%Al 160 2 4 0.99 10.1 XRD, FIB partially 2 TF-CuAl- 015 Cu- 4.5wt%Al 120 2 4 0.73 9.2 XRD, FIB partially 2 1 characterization: profilometry (prof), FIB cross-section (FIB), XRD 2 NT columnar grains were observed on the substrate side but the grain/twin structure became more random through the thickness 3 sputtered onto two separate wafers – one with Carbon underlayer and the other with Cu underlayer 128 CuAl samples produced using continuous sputtering (cont’d) sample composition power [W] press [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization twinned? TF-CuAl- 016 Cu- 4.5wt%Al 120 2 4 0.73 9.2 FIB partially 2 TF-CuAl- 017 Cu- 4.5wt%Al 236 2 5 1.00 10.2 XRD, FIB partially 2 TF-CuAl- 018 4 Cu- 4.5wt%Al 120 2 4 0.73 9.5 XRD, FIB partially TF-CuAl- 019 Cu-6wt%Al 160 2 5 0.74 30.2 FIB partially 2 TF-CuAl- 020 Cu-6wt%Al 160 2 5 0.75 11.3 FIB partially 5 TF-CuAl- 021 Cu-6wt%Al 160 2 5 0.66 9.5 XRD, FIB fully TF-CuAl- 022 Cu-6wt%Al 160 2 5 0.66 28.4 XRD, FIB partially 2 TF-CuAl- 023 Cu-6wt%Al 160 2 5 0.63 12.0 XRD fully TF-CuAl- 024 Cu-6wt%Al 160 2 5 0.69 24.9 XRD, FIB fully TF-CuAl- 025 Cu-6wt%Al 160 2 5 0.69 25.0 XRD, FIB fully TF-CuAl- 026 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD, FIB partially 2 TF-CuAl- 027 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully TF-CuAl- 028 6 Cu-6wt%Al 160 2 5 0.68 24.6 XRD, FIB fully TF-CuAl- 029 6 Cu-6wt%Al 160 2 5 0.68 24.5 XRD fully TF-CuAl- 030 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully TF-CuAl- 031 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD, FIB fully 4 sample was sputtered using interruptions 5 thermal conductive paste was used on the back of the substrate – in areas where good paste made contact with substrate holder, sample was fully nanotwinned. In areas where poor contact was made, sample showed NTs on substrate side, then gradually lost the twinned structure 6 samples were sputtered with a 100-300nm underlayer of Cu in order to be able to remove from substrate 129 CuAl samples produced using continuous sputtering (cont’d) sample composition power [W] press [mTorr] approx. distance (in) sputtering rate [nm/s] total thickness [μm] characterization twinned? TF-CuAl- 032 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully TF-CuAl- 033 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD, FIB fully TF-CuAl- 034 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD, FIB partially TF-CuAl- 035 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully TF-CuAl- 036 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully TF-CuAl- 037 6 Cu-6wt%Al 160 2 5 0.69 25.0 XRD fully 130 (blank page) 131 References [1] C.C. Koch, Optimization of strength and ductility in nanocrystalline and ultrafine grained metals, Scripta Mater., 49 (2003) 657-662. [2] H. Gleiter, Nanocrystalline materials, Prog. Mater Sci., 33 (1989) 223-315. [3] C.C. Koch, D.G. Morris, K. Lu, et al., Ductility of nanostructured materials, MRS Bull., 24 (1999) 54-58. [4] L. Lu, X. Chen, X. Huang, et al., Revealing the maximum strength in nanotwinned copper, Science, 323 (2009) 607-610. [5] O. Anderoglu, A. Misra, H. Wang, et al., Thermal stability of sputtered Cu films with nanoscale growth twins, 103 (2008) 094322. [6] G.E. Dieter, Mechanical Metallurgy, Third ed., McGraw-Hill, Boston, 1986. [7] D.R. Askeland, P.P. Fulay, W.J. Wright, The Science and Engineering of Materials, 6 ed., Cengage Learning, Stamford, 2010. [8] R.F. Cochrane. Micrograph 427 [Micrograph of Annealing Twins]. 2002. University of Cambridge DoITPoMS Micrograph Library. Accessed 19 July 2013. <http://www.doitpoms.ac.uk/miclib/micrograph.php?id=427> [9] H. Idrissi, K. Renard, L. Ryelandt, et al., On the mechanism of twin formation in Fe –Mn – C TWIP steels, Acta Mater., 58 (2010) 2464-2476. [10] J.W. Christian, S. Mahajan, Deformation twinning, Prog. Mater Sci., 39 (1995) 1-157. [11] M. Cherkaoui, L. Capolungo, Atomistic and Continuum Modeling of Nanocrystalline Materials, Springer Science 2009. [12] S. Mahajan, D.F. Williams, Deformation twinning in metals and alloys, Int. Mater. Rev., 18 (1973) 43-61. [13] L.E. Murr, M.A. Meyers, C.S. Niou, et al., Shock-induced deformation twinning in tantalum, Acta Mater., 45 (1997) 157-175. [14] C.X. Huang, K. Wang, S.D. Wu, et al., Deformation twinning in polycrystalline copper at room temperature and low strain rate, Acta Mater., 54 (2006) 655-665. [15] W.S. Zhao, N.R. Tao, J.Y. Guo, et al., High density nano-scale twins in Cu induced by dynamic plastic deformation, Scripta Mater., 53 (2005) 745-749. 132 [16] K.S. Raju, M.G. Krishna, K.A. Padmanabhan, et al., Grain size and grain boundary character distribution in ultra-fine grained (ECAP) nickel, Materials Science and Engineering: A, 491 (2008) 1-7. [17] W.F. Hosford, Mechanical Behavior of Materials, Cambridge University Press, New York, 2005. [18] A.M. Hodge, T.A. Furnish, C.J. Shute, et al., Twin stability in highly nanotwinned Cu under compression, torsion and tension, Scripta Mater., 66 (2012) 872-877. [19] R.L. Fullman, J.C. Fisher, Formation of annealing twins during grain growth, J. Appl. Phys., 22 (1951) 1350-1355. [20] H. Gleiter, The formation of annealing twins, Acta Metall., 17 (1969) 1421-1428. [21] S. Mahajan, C.S. Pande, M.A. Imam, et al., Formation of annealing twins in f.c.c. crystals, Acta Mater., 45 (1997) 2633-2638. [22] M.A. Meyers, L.E. Murr, A model for the formation of annealing twins in F.C.C. metals and alloys, Acta Metall., 26 (1978) 951-962. [23] F.J. Humphreys, M. Hatherly, Recrystallization and Related Annealing Phenomena, Second ed., Elsevier Ltd, Kidlington, 2004. [24] B.B. Rath, M.A. Imam, C.S. Pande, Nucleation and growth of twin interfaces in fcc metals and alloys, Mater. Phys. Mech. 1, (2000) 61-66. [25] X.W. Zhou, H.N.G. Wadley, Twin formation during the atomic deposition of copper, Acta Mater., 47 (1999) 1063-1078. [26] E. El-Danaf, S. Kalidindi, R. Doherty, Influence of grain size and stacking-fault energy on deformation twinning in fcc metals, Metallurgical and Materials Transactions A, 30 (1999) 1223-1233. [27] E. Ma, Y.M. Wang, Q.H. Lu, et al., Strain hardening and large tensile elongation in ultrahigh-strength nano-twinned copper, Appl. Phys. Lett., 85 (2004) 4932-4934. [28] X. Zhang, H. Wang, X.H. Chen, et al., High-strength sputter-deposited Cu foils with preferred orientation of nanoscale growth twins, 88 (2006) 173116. [29] X. Zhang, O. Anderoglu, A. Misra, et al., Influence of deposition rate on the formation of growth twins in sputter-deposited 330 austenitic stainless steel films, Appl. Phys. Lett., 90 (2007) 153101. 133 [30] D. Bufford, H. Wang, X. Zhang, High strength, epitaxial nanotwinned Ag films, Acta Mater., 59 (2011) 93-101. [31] C.J. Shute, B.D. Myers, S. Xie, et al., Detwinning, damage and crack initiation during cyclic loading of Cu samples containing aligned nanotwins, Acta Materialia, 59 (2011) 4569- 4577. [32] X. Zhang, O. Anderoglu, R. Hoagland, et al., Nanoscale growth twins in sputtered metal films, JOM Journal of the Minerals, Metals and Materials Society, 60 (2008) 75-78. [33] X. Zhang, A. Misra, H. Wang, et al., Thermal stability of sputter-deposited 330 austenitic stainless-steel thin films with nanoscale growth twins, 87 (2005) 233116. [34] B. Wu, C. Schuh, P. Ferreira, Nanostructured Ni-Co alloys with tailorable grain size and twin density, Metallurgical and Materials Transactions A, 36 (2005) 1927-1936. [35] K. Lu, L. Lu, S. Suresh, Strengthening Materials by Engineering Coherent Internal Boundaries at the Nanoscale, Science, 324 (2009) 349-352. [36] I. Shabib, R.E. Miller, A molecular dynamics study of twin width, grain size and temperature effects on the toughness of 2D-columnar nanotwinned copper, Modell. Simul. Mater. Sci. Eng., 17 (2009) 055009. [37] Y.F. Shen, L. Lu, Q.H. Lu, et al., Tensile properties of copper with nano-scale twins, Scripta Mater., 52 (2005) 989-994. [38] M.A. Meyers, A. Mishra, D.J. Benson, Mechanical properties of nanocrystalline materials, Prog. Mater Sci., 51 (2006) 427-556. [39] K.S. Kumar, H. Van Swygenhoven, S. Suresh, Mechanical behavior of nanocrystalline metals and alloys, Acta Mater., 51 (2003) 5743-5774. [40] Y. Kulkarni, R.J. Asaro, Are some nanotwinned fcc metals optimal for strength, ductility and grain stability?, Acta Mater., 57 (2009) 4835-4844. [41] X. Li, Y. Wei, L. Lu, et al., Dislocation nucleation governed softening and maximum strength in nano-twinned metals, Nature, 464 (2010) 877-880. [42] Y. Wei, Scaling of maximum strength with grain size in nanotwinned fcc metals, Physical Review B, 83 (2011) 132104. [43] Y.T. Zhu, X.L. Wu, X.Z. Liao, et al., Dislocation –twin interactions in nanocrystalline fcc metals, Acta Mater., 59 (2011) 812-821. 134 [44] R.J. Asaro, Y. Kulkarni, Are rate sensitivity and strength effected by cross-slip in nano- twinned fcc metals, Scripta Mater., 58 (2008) 389-392. [45] T. Zhu, J. Li, A. Samanta, et al., Interfacial plasticity governs strain rate sensitivity and ductility in nanostructured metals, Proceedings of National Academy of Sciences of the United States of America, 104 (2007) 3031-3036. [46] Y.B. Wang, M.L. Sui, Atomic-scale in situ observation of lattice dislocations passing through twin boundaries, Appl. Phys. Lett., 94 (2009) 021909. [47] Z.H. Jin, P. Gumbsch, K. Albe, et al., Interactions between non-screw lattice dislocations and coherent twin boundaries in face-centered cubic metals, Acta Mater., 56 (2008) 1126-1135. [48] I. Shabib, R.E. Miller, Deformation characteristics and stress –strain response of nanotwinned copper via molecular dynamics simulation, Acta Mater., 57 (2009) 4364-4373. [49] N. Li, J. Wang, J.Y. Huang, et al., Influence of slip transmission on the migration of incoherent twin boundaries in epitaxial nanotwinned Cu, Scripta Mater., 64 (2011) 149-152. [50] N. Li, J. Wang, X. Zhang, et al., In-situ TEM study of dislocation-twin boundaries interaction in nanotwinned Cu films, JOM Journal of the Minerals, Metals and Materials Society, 63 (2011) 62-66. [51] B.P. Shu, L. Liu, B. Shen, et al., Deformation twinning and detwinning in the lamella copper observed by in-situ TEM, Mater. Lett., 106 (2013) 225-228. [52] J. Wang, O. Anderoglu, J.P. Hirth, et al., Dislocation structures of ?3 (112) twin boundaries in face centered cubic metals, AIP, 2009. [53] J. Wang, N. Li, O. Anderoglu, et al., Detwinning mechanisms for growth twins in face- centered cubic metals, Acta Mater., 58 (2010) 2262-2270. [54] Z. You, X. Li, L. Gui, et al., Plastic anisotropy and associated deformation mechanisms in nanotwinned metals, Acta Mater., 61 (2013) 217-227. [55] A. Stukowski, K. Albe, D. Farkas, Nanotwinned fcc metals: Strengthening versus softening mechanisms, Physical Review B, 82 (2010) 224103. [56] Y. Wei, The kinetics and energetics of dislocation mediated de-twinning in nano-twinned face-centered cubic metals, Materials Science and Engineering: A, 528 (2011) 1558-1566. [57] L. Lu, Y. Shen, X. Chen, et al., Ultrahigh strength and high electrical conductivity in copper, Science, 304 (2004) 422-426. 135 [58] F. Endres, A.P. Abbot, D.R. MacFarlane, Electrodeposition from Ionic Liquids Wiley- VCH, Weinheim, 2008. [59] P. Mitchell, in, Society of Manufacturing Engineers, Dearborn, 1996. [60] H. Jiang, T.J. Klemmer, J.A. Barnard, et al., Epitaxial growth of Cu(111) films on Si(110) by magnetron sputtering: orientation and twin growth, Thin Solid Films, 315 (1998) 13-16. [61] X. Zhang, A. Misra, H. Wang, et al., Enhanced hardening in Cu/330 stainless steel multilayers by nanoscale twinning, Acta Mater., 52 (2004) 995-1002. [62] J.A. Thornton, Influence of apparatus geometry and deposition conditions on structure and topography of thick sputtered coatings, Journal of Vacuum Science & Technology, 11 (1974) 666-670. [63] J.A. Thornton, Influence of substrate temperature and deposition rate on structure of thick sputtered Cu coatings, Journal of Vacuum Science and Technology, 12 (1975) 830-835. [64] J.A. Thornton, High rate thick film growth, Annu. Rev. Mater. Sci., 7 (1977) 239-260. [65] M. Andritschky, F. Guimarães, V. Teixeira, Energy deposition and substrate heating during magnetron sputtering, Vacuum, 44 (1993) 809-813. [66] A.N. Pargellis, Evaporating and sputtering: Substrate heating dependence on deposition rate, Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films, 7 (1989) 27-30. [67] S.S. Lau, R.H. Mills, Temperature rise during film deposition by rf and dc sputtering, Journal of Vacuum Science & Technology, 9 (1972). [68] R. Wendt, K. Ellmer, K. Wiesemann, Thermal power at a substrate during ZnO:Al thin film deposition in a planar magentron sputtering system, J. Appl. Phys., 82 (1997). [69] A.M. Hodge, Y.M. Wang, T.W. Barbee Jr, Large-scale production of nano-twinned, ultrafine-grained copper, Materials Science and Engineering: A, 429 (2006) 272-276. [70] S.D. Dahlgren, W.L. Nicholson, M.D. Merz, et al., Microstructural analysis and tensile properties of thick copper and nickel sputter deposits, Thin Solid Films, 40 (1977) 345-353. [71] J.P. Hirth, J. Lothe, Theory of Dislocations, Second ed., Wiley, New York, 1983. [72] Magnetrongun, in: http://en.wikipedia.org/wiki/File:Magnetrongun.jpg, 2006, pp. Magnetrongun.jpg. 136 [73] M. Ohring, Materials Science of Thin Films, Academic Press, San Diego 2002. [74] G.E. McGuire, D.C. McIntyre, S. Hofmann, in, Elsevier Science Publishing, New York, 1991, pp. 568. [75] J.A. Venables, The electron microscopy of deformation twinning, J. Phys. Chem. Solids, 25 (1964) 685-692. [76] A. Rohatgi, K.S. Vecchio, The variation of dislocation density as a function of the stacking fault energy in shock-deformed FCC materials, Materials Science and Engineering: A, 328 (2002) 256-266. [77] T.B. Massalski, Binary Alloy Phase Diagrams, American Society for Metals, 1986. [78] K. Jousten, in, Wiley-VCH, Weinheim, 2008, pp. 1017. [79] R.E. Somekh, The thermalization of energetic atoms during the sputtering process, Journal of Vacuum Science & Technology A, 2 (1984) 1285-1291. [80] P.J. Kelly, R.D. Arnell, Magnetron sputtering: a review of recent developments and applications, Vacuum, 59 (2000) 159-172. [81] R.K. Waits, Planar magnetron sputtering, Journal of Vacuum Science & Technology, 15 (1978) 179-187. [82] Practical Process Tips, Lesker Tech, 7 (2010). [83] A.L. Shull, F. Spaepen, Measurements of stress during vapor deposition of copper and silver thin films and multilayers, J. Appl. Phys., 80 (1996) 6243-6256. [84] ASTM, in: E8/E8M-11, 2012. [85] J.R. Davis, Tensile Testing, Second ed., ASM International Materials Park, 2004. [86] ASTM, in: E384-11, ASTM International 2013. [87] A.M. Hodge, R.J. Foreman, G.F. Gallegos, Residual stress analysis in thick uranium films, J. Nucl. Mater., 342 (2005) 8-13. [88] C. Suryanarayana, M.G. Norton, X-Ray Diffraction: A practical Approach, Plenum Press, New York, 1998. [89] L.A. Giannuzzi, F.A. Stevie, Introduction to Focused Ion Beams, Springer Science+Business Media, Inc., 2005. 137 [90] A.M. Hodge, T.A. Furnish, A.A. Navid, et al., Shear band formation and ductility in nanotwinned Cu, Scripta Mater., 65 (2011) 1006-1009. [91] A.M. Hodge, Y.M. Wang, T.W. Barbee Jr, Mechanical deformation of high-purity sputter-deposited nano-twinned copper, Scripta Mater., 59 (2008) 163-166. [92] T.A. Furnish, J. Lohmiller, P.A. Gruber, et al., Temperature-dependent strain localization and texture evolution of highly nanotwinned Cu, Appl. Phys. Lett., 103 (2013) 011904. [93] I.A. Ovid'ko, A.G. Sheinerman, Ductile vs. brittle behavior of pre-cracked nanocrystalline and ultrafine-grained materials, Acta Mater., 58 (2010) 5286-5294. [94] J. Carsley, A. Fisher, W. Milligan, et al., Mechanical behavior of a bulk nanostructured iron alloy, Metallurgical and Materials Transactions A, 29 (1998) 2261-2271. [95] J.E. Carsley, A. Fisher, W.W. Milligan, et al., Mechanical behavior of a bulk nanostructured iron alloy, Metall. Mater. Trans. A-Phys. Metall. Mater. Sci., 29 (1998) 2261- 2271. [96] J.E. Carsley, W.W. Milligan, X.H. Zhu, et al., On the failure of pressure-sensitive plastic materials .2. Comparisons with experiments on ultra fine grained Fe-10% Cu alloys, Scripta Materialia, 36 (1997) 727-732. [97] Q. Wei, D. Jia, K.T. Ramesh, et al., Evolution and microstructure of shear bands in nanostructured Fe, Appl. Phys. Lett., 81 (2002) 1240-1242. [98] C.J. Shute, B.D. Myers, S. Xie, et al., Detwinning, damage and crack initiation during cyclic loading of Cu samples containing aligned nanotwins, Acta Materialia, 59 (2010) 4569–4577. [99] W.G. Johnston, J.J. Gilman, Dislocation velocities, dislocation densities, and plastic flow in lithium fluoride crystals, J. Appl. Phys., 30 (1959) 129-144. [100] H.C. Rogers, Adiabatic plastic-deformation, Annu. Rev. Mater. Sci., 9 (1979) 283-311. [101] R.W. Armstrong, F.J. Zerilli, Dislocation mechanics aspects of plastic instability and shear banding, Mechanics of Materials, 17 (1994) 319-327. [102] O. Engler, V. Randle, Introduction to Texture Analysis: Macrotexture, Microtexture, and Orientation Mapping, Second ed., Taylor & Francis Group, Boca Raton, 2010. [103] W. Gambin, Plasticity and textures, Kluwer Academic Publishers, Dordrecht, 2001. [104] J. Lohmiller, M. Grewer, C. Braun, et al., submitted to Acta Materialia. 138 [105] G.J. Fan, L.F. Fu, Y.D. Wang, et al., Uniaxial tensile plastic deformation of a bulk nanocrystalline alloy studied by a high-energy x-ray diffraction technique, Appl. Phys. Lett., 89 (2006) 101918. [106] E. Ma, Watching the nanograins roll, Science, 305 (2004) 623-624. [107] Z.S. You, L. Lu, K. Lu, Tensile behavior of columnar grained Cu with preferentially oriented nanoscale twins, Acta Mater., 59 (2011) 6927-6937. [108] P.E.J. Flewitt, R.K. Wild, Physical methods for materials characterisation, 2nd ed., Institute of Physics, Bristol, 2003. [109] Y.-D. Wang, W. Liu, L. Lu, et al., Low temperature deformation detwinning —A reverse mode of twinning, Adv. Eng. Mater., 12 (2010) 906-911. [110] T.N. Kon’kova, S.Y. Mironov, V.N. Danilenko, et al., Effect of low-temperature rolling on the structure of copper, Phys. Metals Metallogr., 110 (2010) 318-330. [111] C. Saldana, T.G. Murthy, M.R. Shankar, et al., Stabilizing nanostructured materials by coherent nanotwins and their grain boundary triple junction drag, Appl. Phys. Lett., 94 (2009) 021910-021913. [112] T.H. Blewitt, R.R. Coltman, J.K. Redman, Low-temperature deformation of copper single crystals, J. Appl. Phys., 28 (1957) 651-660. [113] U. Andrade, M.A. Meyers, K.S. Vecchio, et al., Dynamic recrystallization in high- strain, high-strain-rate plastic deformation of copper, Acta Metallurgica et Materialia, 42 (1994) 3183-3195. [114] Y.S. Li, N.R. Tao, K. Lu, Microstructural evolution and nanostructure formation in copper during dynamic plastic deformation at cryogenic temperatures, Acta Mater., 56 (2008) 230-241. [115] Y. Zhao, T.A. Furnish, M.E. Kassner, et al., Thermal stability of highly nanotwinned copper: The role of grain boundaries and texture, J. Mater. Res., 27 (2012) 3049-3057. [116] Q. Wei, L. Kecskes, T. Jiao, et al., Adiabatic shear banding in ultrafine-grained Fe processed by severe plastic deformation, Acta Mater., 52 (2004) 1859-1869. [117] Y. Bai, B. Dodd, Adiabatic Shear Localization, First ed., Pergamon Press, Oxford, 1992. [118] O. Anderoglu, A. Misra, H. Wang, et al., Epitaxial nanotwinned Cu films with high strength and high conductivity, 93 (2008) 083108. 139 [119] J. Bezares, S. Jiao, Y. Liu, et al., Indentation of nanotwinned fcc metals: Implications for nanotwin stability, Acta Mater., 60 (2012) 4623-4635. [120] T.A. Furnish, A.M. Hodge, On the mechanical performance and deformation of nanotwinned Ag, APL Materials, submitted (2014). [121] A.E. Giannakopoulos, S. Suresh, Determination of elastoplastic properties by instrumented sharp indentation, Scripta Mater., 40 (1999) 1191-1198. [122] J. Alcalá, A.C. Barone, M. Anglada, The influence of plastic hardening on surface deformation modes around Vickers and spherical indents, Acta Mater., 48 (2000) 3451-3464. [123] M.M. Chaudhri, M. Winter, The load-bearing area of hardness indentation, J. Phys. D: Appl. Phys., 21 (1988) 370-374. [124] Y.M. Wang, E. Ma, Three strategies to achieve uniform tensile deformation in a nanostructured metal, Acta Mater., 52 (2004) 1699-1709. [125] L.V. Olivante, Materials Science Research Trends, Nova Science Publishers, New York, 2008. [126] C.J. Shute, B.D. Myers, Y. Liao, et al., High-pressure torsion of copper samples containing columns of highly aligned nanotwins, Scripta Mater., 65 (2011) 899-902. [127] C.J. Shute, B.D. Myers, S. Xie, et al., Microstructural stability during cyclic loading of multilayer copper/copper samples with nanoscale twinning, Scripta Mater., 60 (2009) 1073- 1077. [128] L. Lu, R. Schwaiger, Z.W. Shan, et al., Nano-sized twins induce high rate sensitivity of flow stress in pure copper, Acta Mater., 53 (2005) 2169-2179. [129] S. Xia, J. Wang, Mechanical behavior of different constitutional regions characterized by nanoindentation in a bimodal Cu-Al alloy, Materials Transactions, 51 (2010) 36-38. [130] A. Rohatgi, K.S. Vecchio, G.T. Gray III, The Influence of Stacking Fault Energy on the Mechanical Behavior of Cu and Cu-Al Alloys: Deformation Twinning, Work Hardening, and Dynamic Recovery, Metallurgical and Materials Transactions A, 32A (2001) 135-145. [131] K. Nakanishi, H. Suzuki, Analysis of the grain size dependence of the yield stress in copper-aluminum and copper-nickel alloys, Transactions of the Japan Institute of Metals, 15 (1974) 435-440. 140 [132] C.X. Huang, W. Hu, G. Yang, et al., The effect of stacking fault energy on equilibrium grain size and tensile properties of nanostructured copper and copper –aluminum alloys processed by equal channel angular pressing, Materials Science and Engineering: A, 556 (2012) 638-647. [133] X.H. An, Q.Y. Lin, S.D. Wu, et al., The influence of stacking fault energy on the mechanical properties of nanostructured Cu and Cu –Al alloys processed by high-pressure torsion, Scripta Mater., 64 (2011) 954-957. [134] X.H. An, Q.Y. Lin, S.D. Wu, et al., Significance of stacking fault energy on microstructural evolution in Cu and Cu –Al alloys processed by high-pressure torsion, Philosophical Magazine, 91 (2011) 3307-3326. [135] T. Sawa, in: IMEKO 2010 TC3, TC5, TC22 Conferences, Pattaya, Chonburi, Thailand, 2010.

Abstract (if available)

Abstract Metals containing high densities of nanoscale growth twins have shown potential as an alternative to nanocrystalline (nc) metals. Some reports on nanotwinned (NT) Cu, for example, have revealed that high strengths, comparable to nc metals, can be produced while improving other properties such as ductility, thermal stability, and mechanical stability. However, since the synthesis of highly NT metals is not yet fully understood, most studies have been limited to only NT-Cu

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Mechanical behavior and deformation of nanotwinned metallic alloys

PDF

A comprehensive study of twinning phenomena in low and high stacking fault energy metals

PDF

Structure and behavior of nano metallic multilayers under thermal and mechanical loading

PDF

The role of nanotwins and grain boundary plane in the thermal, corrosion, and sensitization behavior of nanometals

PDF

Mechanical behavior and microstructure optimization of ultrafine-grained aluminum alloys and nanocomposites

PDF

Synthesis, characterization, and mechanical properties of nanoporous foams

PDF

Shock wave response of in situ iron-based metallic glass matrix composites

PDF

Understanding the formation and evolution of boundaries and interfaces in nanostructured metallic alloys

PDF

Ageing and mechanical failure of fiber reinforced polymers

PDF

Tailoring compositional and microstructural complexity in nanostructured alloys

$X-ray microbeam diffraction measurements of long range internal stresses in equal channel angular pressed aluminum; & Mechanical behavior of an Fe-based bulk metallic glass$

PDF

X-ray microbeam diffraction measurements of long range internal stresses in equal channel angular pressed aluminum; & Mechanical behavior of an Fe-based bulk metallic glass

PDF

Mechanical behavior of materials in extreme conditions: a focus on creep plasticity

PDF

Microstructural evolution and mechanical properties of a copper-zirconium alloy processed by severe plastic deformation

PDF

Synthesis and properties study of Q1D semiconductor nanostructures

PDF

Investigation of the influence of severe plastic deformation on the microstructure and mechanical properties of Al-7136 alloy

PDF

Processing, mechanical behavior and biocompatibility of ultrafine grained zirconium fabricated by accumulative roll bonding

PDF

Synthesis and mechanical evaluation of micro-scale truss structures formed from self-propagating polymer waveguides

PDF

Exploring the thermal evolution of nanomaterials: from nanometallic multilayers to nanostructures

PDF

Development and characterization of hierarchical cellular structures

PDF

Corrosion studies of engineered interfaces

Asset Metadata

Creator Furnish, Timothy Allen (author)

Core Title Synthesis and mechanical behavior of highly nanotwinned metals

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Mechanical Engineering

Publication Date 04/11/2014

Defense Date 03/11/2014

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

Tag deformation,mechanical properties,nanostructured metals,nanotwin,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hodge, Andrea M. (committee chair), Eliasson, Veronica (committee member), Kassner, Michael E. (committee member), Nutt, Steven R. (committee member)

Creator Email tfurnish@usc.edu,tfurnish22@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-376553

Unique identifier UC11296564

Identifier etd-FurnishTim-2344.pdf (filename),usctheses-c3-376553 (legacy record id)

Legacy Identifier etd-FurnishTim-2344.pdf

Dmrecord 376553

Document Type Dissertation

Format application/pdf (imt)

Rights Furnish, Timothy Allen

Type texts

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

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

Repository Name University of Southern California Digital Library

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