University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Structure and Behavior of Nano Metallic Multilayers under Thermal
and Mechanical Loading

Mikhail N. Polyakov

Dr. Andrea M. Hodge, Advisor

A Dissertation Presented to the Faculty of the USC Graduate School
In Partial Fulfillment of the Requirements for the Degree
Doctor of Philosophy
(Mechanical Engineering)

University of Southern California

December 2015

I

Acknowledgements
Many people have contributed to me conducting and completing the work contained in this
dissertation, and while there is not space enough to thank everyone here, I would like to thank at
least some of them.
Firstly, I would like to thank my parents for all of the love, time, and care they put into me
throughout my life. They are largely responsible for the person I am today, and I am eternally
grateful. I love you, mama and papa.
In addition, all of my siblings and extended family gave me a warm, loving environment in which
I was nurtured while growing up. I was not aware of what a blessing this was at the time; though
now I am starting to see what a treasure I had in those years and still have now.
Several of my professors at the University of Portland were also instrumental in pushing me and
challenging me when I wasn’t willing to push myself. Most notably, I would like to thank Dr.
Timothy Doughty, Dr. Kenneth Lulay, and Dr. V. Dakshina Murty for helping me get to graduate
school.
Once I made it to graduate school, my Ph.D. adviser, Dr. Andrea Hodge, was critical to nearly
everything I was able to accomplish. Her firm and occasionally gentle hand pushed me through
times in which I despaired of ever finishing. She took the time to help me grow as a person and a
scientist, in a manner well above and beyond what would be expected. Thank you, Andrea. I also
appreciate all of the time which Dr. Oliver Franke gave to me. I enjoyed the classes which I took
with him, and his desire for us to truly learn the material was always tangible. He also cared for
II

me personally, whether during training on equipment at USC or showing me around Germany.
Thank you, Oliver.
Through all the highs and lows, my labmates helped me remain sane throughout the entire Ph.D.
process, and I would like to thank them for the support, laughter, and shared struggles that I
experienced. Five years would have felt even longer without them there. Thank you all – Tim,
Leo, Yifu, I-Chung, Thien, Teri, Nate, Jianfeng, Sebastian, Andrew, Lyssa, Georg, Tim V., and
Anahita.
My research collaborators also made much of the work possible, by allowing me to tap into their
expertise and produce research that I could not have achieved on my own. Therefore, I would like
to acknowledge my collaborators from KIT - Dr. J. Lohmiller, Dr. P. A. Gruber, Dr. E. Courtois-
Manara, Dr. D. Wang, Dr. K. Chakravadhanula, and Dr. C. Kübel, from MIT - Dr. T. Chookajorn
and Dr. C. A. Schuh, and from USC - Dr. M. Mecklenburg. In addition, John Curulli assisted in
much of the characterization work and made it much smoother than it would have been otherwise.
I would also like to thank those who sat on my qualifying exam and dissertation defense
committees. Thank you for taking the time out of your busy schedules to help me at this critical
point in my learning process – Dr. Hodge, Dr. Franke, Dr. Kassner, Dr. Meng, Dr. Goo, Dr.
Eliasson, and Dr. Ravichandran.
Finally, I would like to thank God for getting me to this place by the power of his life and spirit
within me. His church sustained me both growing up and at USC. Thank you, Lord Jesus.

III

Table of Contents
ACKNOWLEDGEMENTS ........................................................................................................................................... I
TABLE OF CONTENTS ............................................................................................................................................. III
LIST OF FIGURES ................................................................................................................................................... VII
LIST OF TABLES ..................................................................................................................................................XVII
LIST OF ABBREVIATIONS ..................................................................................................................................... XIX
1 ABSTRACT ..................................................................................................................................................... 1
2 INTRODUCTION ............................................................................................................................................ 2
3 BACKGROUND .............................................................................................................................................. 4
3.1 NANOSTRUCTURED MATERIALS ............................................................................................................................... 4
3.1.1 Material Properties ................................................................................................................................ 4
3.1.2 Thermal Stability .................................................................................................................................... 6
3.1.3 Methods for Improving Thermal Stability .............................................................................................. 7
3.2 NANO METALLIC MULTILAYERS ............................................................................................................................. 11
3.2.1 Microstructure ..................................................................................................................................... 12
3.2.2 Fabrication Methods ............................................................................................................................ 15
3.2.3 General Properties ............................................................................................................................... 20
3.2.4 Mechanical Properties ......................................................................................................................... 21
3.3 THERMAL STABILITY ............................................................................................................................................ 28
3.4 SUMMARY ........................................................................................................................................................ 29
4 PRELIMINARY WORK .................................................................................................................................. 31
4.1 SPUTTERING CHAMBER MODIFICATION .................................................................................................................. 31
4.2 TEMPERATURE MEASUREMENTS DURING SPUTTERING ............................................................................................... 33
IV

4.3 QUENCH VACUUM FURNACE ................................................................................................................................ 35
5 CHARACTERIZATION ................................................................................................................................... 36
5.1 ATOMIC FORCE MICROSCOPY ............................................................................................................................... 36
5.2 TRANSMISSION ELECTRON MICROSCOPY ................................................................................................................. 38
5.2.1 Automated Crystallographic Orientation Mapping ............................................................................. 38
5.2.2 Electron Energy Loss Spectroscopy ...................................................................................................... 39
5.2.3 TEM Nanotwin Imaging ....................................................................................................................... 40
5.2.4 TEM Imaging of Al 5456 Beta-Phase Growth ...................................................................................... 42
5.3 X-RAY DIFFRACTION ............................................................................................................................................ 43
6 MICROSTRUCTURAL VARIATIONS IN CU/NB AND AL/NB NANO METALLIC MULTILAYERS .......................... 45
6.1 OVERVIEW ........................................................................................................................................................ 45
6.2 MICROSTRUCTURE .............................................................................................................................................. 47
6.3 MECHANICAL BEHAVIOR - NANOINDENTATION ........................................................................................................ 52
7 LOAD SHARING PHENOMENA IN NANOSCALE CU/NB MULTILAYERS .......................................................... 54
7.1 OVERVIEW ........................................................................................................................................................ 54
7.2 IN-SITU X-RAY DIFFRACTION ................................................................................................................................ 55
7.3 SAMPLE FABRICATION AND ARCHITECTURE .............................................................................................................. 56
7.4 DEFORMATION BEHAVIOR .................................................................................................................................... 59
7.5 CONCLUSIONS ................................................................................................................................................... 66
7.6 FUTURE RESEARCH DIRECTIONS ............................................................................................................................ 66
8 HF-TI THERMAL STABILITY .......................................................................................................................... 68
8.1 SELECTION OF BINARY SYSTEMS ............................................................................................................................ 68
8.2 METHODS ......................................................................................................................................................... 69
8.2.1 Experimental ........................................................................................................................................ 69
8.2.2 Modeling .............................................................................................................................................. 70
V

8.3 RESULTS AND DISCUSSION .................................................................................................................................... 72
8.3.1 Microstructure and Grain Growth ....................................................................................................... 72
8.3.2 Kinetics ................................................................................................................................................. 77
8.3.3 Ti Segregation ...................................................................................................................................... 80
8.3.4 Modeling .............................................................................................................................................. 83
8.4 CONCLUSIONS ................................................................................................................................................... 85
9 CONCLUSIONS AND FUTURE WORK ............................................................................................................ 87
10 REFERENCES ............................................................................................................................................... 90
11 ADDITIONAL BIBLIOGRAPHY ..................................................................................................................... 101
APPENDIX A. IN-SITU XRD TENSILE TESTING ................................................................................................ 102
APPENDIX A.1 GENERAL NOTES FOR SPUTTERING ONTO POLYIMIDE .................................................................................... 102
APPENDIX A.2 PROCEDURE FOR SPUTTERING ONTO POLYIMIDE .......................................................................................... 104
APPENDIX A.3 ADDITIONAL SAMPLES ............................................................................................................................ 105
APPENDIX A.4 OTHER NMM SYSTEMS SPUTTERED ONTO POLYIMIDE ................................................................................. 108
APPENDIX A.5 IN-SITU XRD TENSILE TESTING NOTES ....................................................................................................... 109
APPENDIX B. VARIOUS SPUTTERING NOTES ................................................................................................ 111
APPENDIX B.1 VARIATION IN THICKNESS ACROSS SUBSTRATE POSITIONS .............................................................................. 111
APPENDIX B.2 SUBSTRATE HOLDER FOR POLYIMIDE SUBSTRATES ........................................................................................ 112
APPENDIX B.3 CHECKING OF PHASES BY XRD ................................................................................................................. 112
APPENDIX B.4 SPUTTERING AMORPHOUS ALLOYS ............................................................................................................ 116
APPENDIX C. RESIDUAL STRESSES ................................................................................................................ 118
APPENDIX D. TEM IMAGING OF NMMS ....................................................................................................... 120
APPENDIX E. SEM AND FIB IMAGING OF NMMS.......................................................................................... 123
APPENDIX E.1 CROSS-SECTIONAL .................................................................................................................................. 123
VI

APPENDIX E.2 TOPOGRAPHICAL .................................................................................................................................... 124
APPENDIX F. ADDITIONAL XRD INVESTIGATIONS ........................................................................................ 127
APPENDIX G. CO-SPUTTERING INVESTIGATION ............................................................................................ 131
APPENDIX G.1 EXISTING FLANGES ................................................................................................................................. 131
APPENDIX G.2 CO-SPUTTERING FLANGE......................................................................................................................... 133
APPENDIX H. ADDITIONAL THERMAL STABILITY RESEARCH ......................................................................... 135
APPENDIX H.1 OXIDATION CONSIDERATIONS .................................................................................................................. 135
APPENDIX H.2 REMOVING SAMPLES FROM SUBSTRATES ................................................................................................... 136
APPENDIX H.3 QUENCH VACUUM FURNACE SPECIFICS ..................................................................................................... 137
APPENDIX H.4 ADDITIONAL HEAT-TREATED SYSTEMS ....................................................................................................... 138
APPENDIX I. LIST OF SAMPLES .................................................................................................................... 145
APPENDIX I.1 MONOLITHIC SINGLE-ELEMENT FILMS ON SI AND GLASS SUBSTRATES ............................................................... 145
APPENDIX I.2 MONOLITHIC MULTI-ELEMENT FILMS ON SI AND GLASS SUBSTRATES ............................................................... 151
APPENDIX I.3 MULTILAYERED SAMPLES ON SI AND GLASS SUBSTRATES ................................................................................ 153
APPENDIX I.4 SAMPLES ON POLYIMIDE SUBSTRATES ......................................................................................................... 157
APPENDIX I.5 HEAT-TREATED SAMPLES .......................................................................................................................... 160

VII

List of Figures
FIGURE 1: SCHEMATIC OF DISLOCATION PILE-UP AT AN OBSTACLE. THE APPLIED STRESS IS Τ, THE DISTANCE FROM THE OBSTACLE IS XI, AND
THE STRESS ON DISLOCATION “0” INCREASES WITH EVERY DISLOCATION PILED UP BEHIND IT. [7] ............................................... 5
FIGURE 2: PLOT SHOWING THE THREE TYPICAL REGIMES FOR FLOW STRESS AS A FUNCTION OF GRAIN SIZE IN MATERIALS WITH EQUIAXED
GRAINS. IN THE GRAIN SIZE > 100 NM REGIME, HALL-PETCH STRENGTHENING IS OCCURRING, WITH A SLOPE OF K. AT GRAIN SIZES
OF 100 NM AND 10 NM, THE CHANGES IN SLOPE INDICATE DEFORMATION MECHANISM CHANGES. [8] ...................................... 5
FIGURE 3: BRIGHT FIELD TRANSMISSION ELECTRON MICROSCOPE IMAGES SHOWING GRAIN GROWTH IN NI AT 420 °C. THE LABELS
INDICATE THE HEAT TREATMENT TIME FOR EACH IMAGE. THE GRAIN GROWTH BEGINS ALMOST INSTANTLY (NOTE THE SCALE BAR
CHANGES AT 1 S AND 11 H). [16] ................................................................................................................................ 6
FIGURE 4: BINARY (A) AND TERNARY (B, C) HEAT TREATED MATERIALS. THE PROPOSED STABILIZATION MECHANISM, TRIPLE JUNCTION
FORMATION, IS ILLUSTRATED IN (D) FOR TERNARY SYSTEMS. [19] ....................................................................................... 7
FIGURE 5: GRAIN SIZE DISTRIBUTIONS FOR NANOCRYSTALLINE CU WITH TA ADDITIONS HEAT-TREATED FOR 4 HOURS AT (A) 540 °C, (B)
770 °C, (C) 900 °C, AND (D) 1040 °C. THE GRAIN SIZE INCREASED WITH INCREASING HEAT TREATMENT TEMPERATURE, BUT IT
REMAINED IN THE NANO AND ULTRA-FINE GRAIN REGIMES, EVEN FOR THE HIGHEST HEAT TREAT TEMPERATURE. [24] ................... 8
FIGURE 6: GRAIN SIZE AS A FUNCTION OF THE ANNEALING TIME FOR (A) AN FE-CR ALLOY AND (B) AN FE SAMPLE. THE DOTS ARE
EXPERIMENTAL DATA AND THE LINES ARE BASED ON MODELS. NOTE THE TEMPORARY PLATEAUS IN THE GRAIN SIZE. [29] .............. 9
FIGURE 7: PLOT SHOWING WHICH MINOR ALLOYING ELEMENTS SHOULD PRODUCE NANOCRYSTALLINE STABLE (GREEN CIRCLE AREA) AND
BULK STABLE (RED CROSS AREA) MATERIALS WHEN ALLOYED WITH W. THE ERROR BARS ON TI INDICATE TYPICAL UNCERTAINTIES FOR
THE CALCULATIONS USED TO GENERATE THIS PLOT. [31] ................................................................................................. 10
FIGURE 8: (A) GRAIN SIZE DISTRIBUTIONS FOR AS-MILLED W-TI, ANNEALED W-TI, AND ANNEALED PURE W SAMPLES. ANNEALING TOOK
PLACE AT 1100 °C FOR A WEEK. THE ANNEALED W-TI MATERIAL RETAINED NEARLY THE SAME GRAIN SIZE, WHILE THE ANNEALED
PURE W SAMPLE SHOWED SUBSTANTIAL GRAIN GROWTH. TEM IMAGES INDICATING GRAIN SIZES FOR (B) AS-MILLED W-TI , (C)
ANNEALED PURE W, AND D) ANNEALED W-TI REFLECT THE GRAIN SIZE DISTRIBUTIONS IN (A). THE INSETS IN (B) AND (D) ARE DARK-
FIELD TEM IMAGES THAT BETTER SHOW SOME OF THE INDIVIDUAL GRAINS. [31] ................................................................. 11
FIGURE 9: CROSS-SECTIONAL TRANSMISSION ELECTRON MICROSCOPE IMAGE OF CU/NB NMM. ..................................................... 12
VIII

FIGURE 10: (A) CROSS-SECTIONAL TEM IMAGE SHOWING THE COLUMNAR MICROSTRUCTURE IN A CU/NB NMM. THE 2 NM-THICK CU
AND NB LAYERS ARE NOT APPARENT AT THIS MAGNIFICATION [35], (B) CROSS-SECTIONAL SEM IMAGE OF A MG/NB NMM
WITHOUT A CLEAR COLUMNAR MICROSTRUCTURE. MG LAYERS ARE DARK, NB LAYERS ARE BRIGHT. [36].................................. 12
FIGURE 11: SCHEMATIC OF NMM GRAIN BOUNDARY STRUCTURES, WITH MATERIALS Α AND Β, ARRANGED BY THERMAL STABILITY, FROM
(A) MOST STABLE TO (D) LEAST STABLE. THE VERTICAL LINES ARE GRAIN BOUNDARIES, THE HORIZONTAL LINES ARE INTERFACES
BETWEEN LAYERS. [37] ............................................................................................................................................ 13
FIGURE 12: TWO COHERENT INTERFACES ARE SHOWN: (A) PERFECT COHERENT INTERFACE, (B) TILTED COHERENT INTERFACE. [38] ........ 13
FIGURE 13: SCHEMATIC SHOWING PARTIAL COHERENCE BETWEEN CRYSTALLOGRAPHICALLY MISMATCHED MATERIALS, RESULTING IN A
COMMENSURATE INTERFACE. THE TWO MATERIALS ARE BROUGHT INTO COHERENCY BY THE TILTING OF ONE MATERIAL RELATIVE TO
THE OTHER. THE FILM LATTICE CONSTANT IS DF AND THE SUBSTRATE LATTICE CONSTANT IS DS. THE FILM MILLER INDICES ARE H, K,
AND L, WHILE M AND N INDICATE THE NUMBER OF LATTICE SPACINGS BETWEEN COINCIDENT LATTICE POSITIONS. [39] ................ 14
FIGURE 14: SCHEMATICS OF DIFFERENT INTERFACE TYPES: (A) STRAINED COHERENT INTERFACE WITH THE TOP MATERIAL IN TENSION AND
THE BOTTOM IN COMPRESSION, (B) SEMI-COHERENT INTERFACE WITH DISLOCATIONS TO RELIEVE THE COHERENCY STRAIN, AND (C)
INCOHERENT INTERFACE WITH DISLOCATIONS THAT ARE COMPLETELY OVERLAPPED, AND, THUS, INDISTINGUISHABLE. [38] .......... 15
FIGURE 15: ACCUMULATIVE ROLL-BONDING SCHEMATIC SHOWING THE STEPS NECESSARY TO FABRICATE AN NMM BY THIS METHOD. [40]
........................................................................................................................................................................... 16
FIGURE 16: SCHEMATIC SHOWING THE MAIN STEPS OF MAGNETRON SPUTTERING: 1) AN AR GAS IS INTRODUCED AND IS IONIZED BY THE
NEGATIVELY BIASED TARGET, 2) THE AR IONS STRIKE THE TARGET, CAUSING THE EJECTION OF TARGET ATOMS, 3) THE TARGET ATOMS
DEPOSIT ONTO THE SUBSTRATE. ................................................................................................................................. 19
FIGURE 17: SEM IMAGES OF 800 NM DIAMETER CU/ZR PILLARS IN THE AS-MILLED STATE (LEFT) AND COMPRESSED STATE (RIGHT). THE
LAYER THICKNESSES OF THE SAMPLES ARE (A, B) 100 NM, (C, D) 50 NM, AND (E, F) 20 NM. EXTRUSION OF THE CU LAYERS IS
EVIDENT IN (B), WHILE (D) AND (F) SHOW SHEARING OF THE PILLARS. [55] ......................................................................... 22
FIGURE 18: HARDNESS VALUES FOR DIFFERENT NMM SYSTEMS AS A FUNCTION OF THE INVERSE SQUARE OF LAYER THICKNESS. [60] ...... 23
FIGURE 19: ILLUSTRATIVE PLOT SHOWING THE THREE STRENGTH REGIMES OBSERVED IN NMMS AT DIFFERENT LAYER THICKNESSES. THE
CORRESPONDING DEFORMATION MECHANISMS ARE SHOWN FOR THE DIFFERENT REGIMES. [63] ............................................. 24
FIGURE 20: HIGH-RESOLUTION TRANSMISSION ELECTRON MICROSCOPE IMAGE SHOWING A CONTINUOUS FCC STRUCTURE AT AN AL/TI
INTERFACE. THIS RESULTS IN A COHERENT INTERFACE THAT WILL BE AMENABLE TO DISLOCATION CROSSING. [69] ...................... 26
IX

FIGURE 21: STRESS-STRAIN CURVES FOR MONOLITHIC AG, MONOLITHIC CU, AND AG/CU NMMS WITH DECREASING LAYER THICKNESS.
[75] ..................................................................................................................................................................... 27
FIGURE 22: DUCTILE TENSILE DEFORMATION OF A CU/CUZR NMM, COMPARED TO A CU/304 SS NMM AND PURE CU. [76]............. 28
FIGURE 23: (A) SPUTTERING CHAMBER SCHEMATIC SHOWING THE DIFFERENT SOURCES WHICH ALLOW FOR SPUTTERING OF MULTIPLE
MATERIALS AND THE MOTOR FOR SPINNING THE SUBSTRATE TO FACE THE DIFFERENT SOURCES, (B) PICTURE OF THE SPUTTERING
PROCESS TAKING PLACE, WITH THE COMPONENTS LABELED. ............................................................................................. 31
FIGURE 24: CO-SPUTTERING FLANGE SCHEMATIC. THE TWO PORTS WITH 2.75” CONFLAT (CF) FLANGES ARE FOR SPUTTERING GUNS, THE
TWO 2.12” CF FLANGES ARE FOR SHUTTERS, AND THE CENTRAL 1.33” CF FLANGE IS A MULTIPURPOSE PORT FOR PROBES. THE
2.12” AND 2.75” CF FLANGES ARE INCLINED 12° TO THE VERTICAL; THE 2.12” AND 2.75” CF FLANGES IN EACH PAIR (RIGHT AND
LEFT SIDE OF THE IMAGE) ARE PARALLEL TO ONE ANOTHER. THE 2.75” CF PORTS ARE ALIGNED TO INTERSECT 291 MM BELOW THE
FLANGE. ................................................................................................................................................................ 32
FIGURE 25: 51 MM DIAMETER SI SUBSTRATE WITH EMBEDDED THERMOCOUPLE FOR TEMPERATURE MEASUREMENTS DURING SPUTTERING.
........................................................................................................................................................................... 33
FIGURE 26: TEMPERATURE PLOTS DURING SPUTTERING AS A FUNCTION OF TIME FOR (A) DIFFERENT POWERS (0.27 PA AR PRESSURE, 152
MM SPUTTERING DISTANCE, 51 MM DIAMETER AL SPUTTERING TARGET) AND (B) DIFFERENT ARGON PRESSURES (300 W
SPUTTERING POWER, 152 MM SPUTTERING DISTANCE, 51 MM DIAMETER AL SPUTTERING TARGET). ....................................... 34
FIGURE 27: SAMPLE COOLING CURVE UNDER VACUUM DIRECTLY AFTER SPUTTERING. ..................................................................... 35
FIGURE 28: 3-D AFM PLOTS OF NB FILMS SPUTTERED USING DIFFERENT PARAMETERS; THE SAMPLES ARE (A) SAMPLE 31 AND (B) SAMPLE
43 IN TABLE 2 ABOVE. ............................................................................................................................................. 37
FIGURE 29: SAMPLE NED PATTERNS FROM AN AL/NB NMM SAMPLE. THE DIFFERENT DIFFRACTION PATTERNS INDICATE DIFFERING GRAIN
ORIENTATIONS. ....................................................................................................................................................... 38
FIGURE 30: ACOM IMAGES OF (A) FINE-GRAINED PD AND (B) TENSILE-STRESSED AL. THE ORIENTATIONS NORMAL TO THE VIEWING
DIRECTION ARE INDICATED BY THE COLORS OF THE STEREOGRAPHIC TRIANGLES. [92] ............................................................ 39
FIGURE 31: TYPICAL EELS SPECTRUM SHOWING ZERO-LOSS PEAK, LOW-LOSS EDGE, AND CORE-LOSS EDGES. THE PRESENCE OF THE CORE-
LOSS EDGES CAN BE USED TO GENERATE COMPOSITION MAPS OF THE SAMPLE. NOTE THE MULTIPLICATION OF THE SPECTRUM VALUE
AT APPROXIMATELY 10 EV AND 240 EV, WHICH IS NEEDED DUE TO THE STEEP DECREASE IN SIGNAL WITH INCREASING ENERGY LOSS.
[93] ..................................................................................................................................................................... 40
X

FIGURE 32: TEM IMAGES OF A NANOTWINNED CUAL SAMPLE AT (A) LOW AND (B) HIGH MAGNIFICATION. MULTIPLE HIGHLY TWINNED
GRAINS ARE EVIDENT IN (A). IN (B), THE INDIVIDUAL TWINS CAN BE MEASURED AND COUNTED. IN (C) AN SAED PATTERN IS GIVEN
FROM THE SAME SAMPLE AND IS CHARACTERISTIC OF A HIGHLY NANOTWINNED STRUCTURE ORIENTED TO A [110] ZONE AXIS. ...... 41
FIGURE 33: TWIN SPACING DISTRIBUTIONS FOR CUAL ALLOYS (A, B) AND CU (C). THE GREEN LINES ARE CURVE FITS. [99] .................... 42
FIGURE 34: CROSS-SECTIONAL TEM IMAGES SHOWING MG GRAIN GROWTH AT THE GRAIN BOUNDARIES OF A HEAT-TREATED AL 5456
SAMPLE (A, B). THE ORIENTATIONS OF GRAIN BOUNDARIES WHICH WERE EXAMINED FOR MG GRAIN GROWTH ARE SHOWN IN (C).
[100] ................................................................................................................................................................... 42
FIGURE 35: XRD SCAN OF A CU/NB NMM SHOWING MULTIPLE SUPERLATTICE PEAKS. AN XRD SCAN FROM A NON-NMM SAMPLE
CONTAINING NB AND CU WOULD HAVE HAD ONLY TWO PEAKS IN THIS SCAN RANGE: NB AT 38.5° AND CU AT 43.3°. ................ 43
FIGURE 36: A LOW-ANGLE XRD SCAN OF A CU/TA NMM SHOWING MULTIPLE PEAKS, WHICH CAN BE USED TO ESTIMATE LAYER THICKNESS
AND UNIFORMITY. ................................................................................................................................................... 44
FIGURE 37: SCHEMATIC OF CU/NB AND AL/NB NMMS SPUTTERED FOR THIS STUDY WHICH CONSIST OF THICK NB LAYERS (100 NM) AND
THIN AL OR CU LAYERS (1-20 NM). ............................................................................................................................ 46
FIGURE 38: REPRESENTATIVE CROSS-SECTIONAL STEM IMAGES OF AL/NB MULTILAYERS WITH (A) 20 NM AND (B) 2 NM AL LAYER
THICKNESSES. CORRESPONDING CROSS-SECTIONAL ACOM IMAGES FOR THE (C) 20 NM AND (D) 2 NM AL SAMPLES. THE AL LAYERS
ARE ALIGNED FOR THE STEM AND ACOM IMAGES. THE GRAIN ORIENTATIONS NORMAL TO THE IMAGES FOR (C) AND (D) ARE
INDICATED BY THE STEREOGRAPHIC TRIANGLE INSET. THE ARROWS IN (B) INDICATE THE AL LAYERS. THE GROWTH DIRECTION IS
INDICATED ON THE LEFT. [107] .................................................................................................................................. 47
FIGURE 39: CROSS-SECTIONAL STEM IMAGE (A), ACOM IMAGE (B), AND ACOM PHASE MAP (C) OF A CU/NB MULTILAYER SAMPLE WITH
20 NM CU LAYERS. IN (B) THE GRAIN ORIENTATIONS NORMAL TO THE IMAGE ARE INDICATED BY THE STEREOGRAPHIC TRIANGLE
INSET. THE NED PATTERNS TO THE LEFT INDICATE DISTINCT LOCAL ORIENTATIONS IN THE CU AND NB LAYERS. THE BOXED AND
CIRCLED AREAS HIGHLIGHT CRYSTALLINE AND AMORPHOUS CU REGIONS, RESPECTIVELY. FOR (C), THE COLORS INDICATE STRUCTURE:
RED-BCC NB, GREEN-FCC CU, BLUE-AMORPHOUS AND SEMI-AMORPHOUS, BLACK-POOR INDEXING. THE REPRESENTATIVE NED
PATTERNS ON THE RIGHT INDICATE CRYSTALLINE, SEMI-AMORPHOUS, AND AMORPHOUS CU REGIONS. THE GROWTH DIRECTION IS
INDICATED ON THE LEFT. [107] .................................................................................................................................. 49
FIGURE 40: CROSS-SECTIONAL STEM IMAGE (A), ACOM IMAGE (B), AND ACOM PHASE MAP (C) OF A CU/NB MULTILAYER SAMPLE WITH
2 NM CU LAYERS. THE ARROWS IN (A) INDICATE THE CU LAYERS. IN (B), THE GRAIN ORIENTATIONS NORMAL TO THE IMAGE ARE
XI

INDICATED BY THE STEREOGRAPHIC TRIANGLE INSET. FOR THE ACOM PHASE MAP IN (C), THE COLORS INDICATE STRUCTURE: RED-
BCC NB, GREEN-FCC CU, BLACK-POOR INDEXING. NOTE THAT AMORPHOUS AND SEMI-AMORPHOUS REGIONS WERE NOT
OBSERVED. THE GROWTH DIRECTION IS INDICATED ON THE LEFT. [107] ............................................................................ 51
FIGURE 41: HARDNESS PLOTS, AS MEASURED BY NANOINDENTATION, FOR (A) CU/NB AND (B) AL/NB NMMS AS A FUNCTION OF CU AND
AL LAYER THICKNESS, WITH CONSTANT NB LAYER THICKNESSES OF 100 NM. THE BLUE SQUARES AND RED DIAMONDS CORRESPOND
TO SAMPLES THAT HAD SLOW AND FAST CU AND AL SPUTTERING RATES, RESPECTIVELY (SLOW DEPOSITION RATES – CU: 0.15 NM/S,
AL: 0.07 NM/S; FAST DEPOSITION RATES – CU: 0.69 NM/S, AL: 0.29 NM/S; NB DEPOSITION RATE - 0.12 NM/S). THE ERROR
BARS INDICATE THE STANDARD DEVIATION. ................................................................................................................... 52
FIGURE 42: SCHEMATIC OF IN-SITU SYNCHROTRON MEASUREMENTS DURING TENSILE TESTING. AS THE RECTANGULAR SAMPLE ON THE
LEFT IS BEING LOADED, THE X-RAY BEAM TRAVELS THROUGH THE SAMPLE, AND THE DIFFRACTION CONES FROM THE X-RAY BEAM
APPEAR AS DIFFRACTION RINGS ON THE CAMERA. THESE RINGS CORRESPOND TO DIFFERENT LATTICE PLANES IN THE MATERIALS. .. 55
FIGURE 43: PICTURE OF A TENSILE TESTED CU/NB NMM SAMPLE ON A POLYIMIDE SUBSTRATE. THE SPECKLE PATTERNS ON THE RIGHT
AND LEFT SIDES ARE FOR OPTICAL TRACKING OF THE OVERALL SAMPLE STRAIN. ..................................................................... 57
FIGURE 44: TEST SPECIMEN SCHEMATIC (SIDE VIEW). SAMPLES CONSISTING OF ALTERNATING NB AND CU LAYERS WERE SPUTTERED ONTO
POLYIMIDE SUBSTRATES. TEST SPECIMENS WERE THEN CUT AND TENSILE TESTED PARALLEL TO THE LAYER DIRECTION. (IMAGE NOT TO
SCALE) [126] ......................................................................................................................................................... 57
FIGURE 45: BRIGHT-FIELD CROSS-SECTIONAL TEM IMAGES OF AS-SPUTTERED CU/NB SAMPLES ON POLYIMIDE SUBSTRATES. THE NOMINAL
THICKNESSES ARE 10 NM FOR ALL NB LAYERS, AND (A) 20 NM, (B) 10 NM, (C) 5 NM, AND (D) 2 NM FOR THE CU LAYERS. DARKER
LAYERS ARE NB AND BRIGHTER LAYERS ARE CU. [126] .................................................................................................... 59
FIGURE 46: STRESS-STRAIN DATA COLLECTED DURING IN-SITU XRD TENSILE TESTING OF CU/NB MULTILAYER SAMPLES ON POLYIMIDE
SUBSTRATES. SAMPLE NAMES STATE THE THICKNESSES OF THE DIFFERENT LAYERS IN NM. THE STRESS-STRAIN CURVES FOR THE (A)
CU AND (B) NB LAYERS WITHIN THE SAMPLES ARE PLOTTED. THREE POINTS OF FAILURE ARE INDICATED BY ARROWS FOR THE SAMPLE
WITH 2 NM CU LAYERS IN (A) AND (B). SAMPLE FAILURE STRAINS ARE PLOTTED AS A FUNCTION OF THE CU LAYER THICKNESS IN (C).
MAXIMUM LATTICE STRAINS, WHICH CORRESPOND TO THE STRENGTHS, ARE PLOTTED FOR THE CU AND NB LAYERS AS A FUNCTION
OF THE CU LAYER THICKNESS IN (D). [126] ................................................................................................................... 60
XII

FIGURE 47: XRD PEAK BREADTH DATA COLLECTED DURING IN-SITU XRD TENSILE TESTING OF CU/NB MULTILAYER SAMPLES ON POLYIMIDE
SUBSTRATES. SAMPLE NAMES STATE THE THICKNESSES OF THE DIFFERENT LAYERS IN NM. PEAK BREADTHS CORRESPONDING TO THE
(A) CU AND (B) NB LAYERS ARE PLOTTED. THE THREE ARROWS IN (A) INDICATE SAMPLE FAILURE STRAINS. [126] ....................... 64
FIGURE 48: SEM IMAGES OF THE POST-MORTEM SURFACES OF TENSILE-TESTED CU/NB MULTILAYER SAMPLES ON POLYIMIDE SUBSTRATES.
THE NB LAYERS ARE 10 NM THICK FOR ALL SAMPLES, AND THE CU LAYERS ARE (A) 20 NM, (B) 10 NM, (C) 5 NM, AND (D) 2 NM
THICK. THE SAMPLES WITH THICKER CU LAYERS, (A) AND (B), HAVE CRACKS WHICH START AND STOP WITHIN THE SAMPLES, AS
POINTED OUT BY ARROWS IN (A). THE CRACKS IN THE SAMPLES WITH THINNER CU LAYERS, (C) AND (D), SPANNED THE ENTIRE
SAMPLE WIDTH. THE INSETS SHOW THAT THE SAMPLES WITH THICKER CU LAYERS SHOWED MORE TORTUOUS CRACK PATHS. THE
BLACK SCALE BARS IN THE INSETS ARE ONE ΜM LONG. ..................................................................................................... 65
FIGURE 49: DIFFRACTION RING INTENSITY AT DIFFERENT ANGLES FOR THREE NMM SAMPLES: CU20/NB10 (BLUE, MIDDLE), CU5/NB10
(RED, TOP), AND CU2/NB10 (GREEN, BOTTOM). .......................................................................................................... 67
FIGURE 50: CROSS-SECTIONAL TEM IMAGES OF AS-SPUTTERED NANOMETALLIC MULTILAYERED COMPOSITE SAMPLES. SAMPLE A CONSISTS
OF 15 NM HF-TI/2 NM TI LAYERS, SEE (A) AND (B), AND SAMPLE B CONSISTS OF 40 NM HF-TI/5 NM TI LAYERS, SEE (C) AND (D).
IN THE BRIGHT FIELD TEM IMAGES, (A) AND (C), THE TI LAYERS ARE BRIGHT AND THE HF-TI LAYERS ARE DARK. THE DARK-FIELD
TEM IMAGES, (B) AND (D), DEMONSTRATE THE COLUMNAR STRUCTURES OF THE SAMPLES, WITH THEIR RESPECTIVE SAED
PATTERNS SHOWING STRONG HF (002) TEXTURE. THE GROWTH DIRECTION IS INDICATED ON THE LEFT, AND THE DIAMETERS OF
SEVERAL GRAINS ARE MARKED WITH YELLOW ARROWS IN (D). .......................................................................................... 72
FIGURE 51: CROSS-SECTIONAL TEM IMAGES OF HEAT-TREATED (800 °C, 96 HOURS) NANOMETALLIC MULTILAYERED COMPOSITE
SAMPLES. THE FIRST SAMPLE IS HT-SAMPLE A, WHICH ORIGINALLY CONSISTED OF 15 NM HF-TI/2 NM TI LAYERS, SEE (A), (B), AND
(C). THE SECOND SAMPLE, HT-SAMPLE B, ORIGINALLY CONSISTED OF 40 NM HF-TI/5 NM TI LAYERS, SEE (D), (E), AND (F). THE
BRIGHT-FIELD TEM IMAGES IN (A) AND (D) SHOW THE NANOCRYSTALLINITY OF THE SAMPLES, WITH INSET SAED PATTERNS (SCALE
BAR = 5/NM) AND THE GROWTH DIRECTION INDICATED ON THE LEFT. DARK-FIELD TEM IMAGES, (B) AND (E), SHOW INDIVIDUAL
GRAINS. THE ANNULAR DARK-FIELD STEM IMAGES, (C) AND (F), DISPLAY THE HF AND TI SEGREGATION, WITH HF-RICH REGIONS
APPEARING BRIGHT AND TI-RICH REGIONS APPEARING DARK. ........................................................................................... 74
FIGURE 52: CROSS-SECTIONAL DARK-FIELD TEM IMAGE (A) OF AS-SPUTTERED SAMPLE C (CO-SPUTTERED HF-TI SAMPLE, 20 AT.% TI).
THE GROWTH DIRECTION IS INDICATED ON THE LEFT. ANNULAR DARK-FIELD STEM (B) AND DARK-FIELD TEM (D) IMAGES ARE
SHOWN OF THE HEAT-TREATED SAMPLE (800 °C, 96 HOURS), HT-SAMPLE C. DARK AREAS ARE HF-RICH AND BRIGHT AREAS ARE
XIII

TI-RICH IN (B). THE GRAIN SIZES BEFORE AND AFTER HEAT TREATMENT ARE GIVEN IN (C), WITH AVERAGE GRAIN SIZES OF 57 NM
AND 196 NM BEFORE AND AFTER HEAT TREATMENT, RESPECTIVELY. .................................................................................. 76
FIGURE 53: GRAIN SIZE DISTRIBUTIONS FOR AS-SPUTTERED AND HEAT-TREATED (800 °C, 96 HOURS) NANOMETALLIC MULTILAYERED
COMPOSITE SAMPLES, SAMPLE A (15 NM HF-TI/2 NM TI) AND SAMPLE B (40 NM HF-TI/5 NM TI). THE AVERAGE GRAIN SIZES
WERE AS FOLLOWS: (A) SAMPLE A: 25 NM AND 50 NM BEFORE AND AFTER HEAT TREATMENT, RESPECTIVELY, AND (B) SAMPLE B:
40 NM AND 52 NM BEFORE AND AFTER HEAT TREATMENT, RESPECTIVELY. .......................................................................... 77
FIGURE 54: CALCULATED HF DIFFUSIVITIES AND 96-HOUR DIFFUSION DISTANCES AT 800 °C FOR TRIPLE JUNCTIONS, GRAIN BOUNDARIES,
AND BULK, AS WELL AS OVERALL HF-TI INTERDIFFUSION, AS A FUNCTION OF GRAIN SIZE. ....................................................... 80
FIGURE 55: CROSS-SECTIONAL HF (A) AND TI (B) COMPOSITIONAL MAPS RECORDED BY EDX IN A TEM FOR HT-SAMPLE B. THE SAMPLE
ORIGINALLY CONSISTED OF ALTERNATING 40 NM HF-TI/5 NM TI LAYERS AND WAS HEAT TREATED AT 800 °C FOR 96 HOURS. AN
ANNULAR DARK-FIELD STEM IMAGE IS SHOWN FROM THE SAME AREA IN (C). A LINESCAN COMPOSITIONAL PROFILE (D) IS SHOWN
FOR HF AND TI, ALONG THE A-B LINE LOCATED IN (C). .................................................................................................... 81
FIGURE 56: CONVERGENT BEAM ELECTRON DIFFRACTION (CBED) PATTERNS FROM A TI-RICH (A) AND A HF-RICH (C) REGION OF HT-
SAMPLE B (A HEAT-TREATED 40 NM HF-TI/5 NM TI MULTILAYER SAMPLE). THE LOCATIONS WHERE THESE PATTERNS WERE
RECORDED ARE POINTED OUT BY ARROWS IN THE CROSS-SECTIONAL TEM IMAGE IN (B). THE SCALE BAR IN (B) IS 20 NM LONG. ... 82
FIGURE 57: MONTE CARLO MODELS OF HF-TI STRUCTURES (23 AT.% TI) AT 800 °C. A FORCED BULK STRUCTURE IS SHOWN IN (A) AS A
HOMOGENEOUS SOLID SOLUTION, WITH TI ATOMS PRESENTED IN BLACK. THE FULLY EQUILIBRATED STRUCTURE IS SHOWN IN (B),
WHERE TI ATOMS ARE BLACK AND THE COLORS INDICATE DIFFERENT GRAINS. THE STRUCTURE SHOWN IN (C) IS FOR A MULTILAYERED
SAMPLE, WITH BLACK TI LAYERS AND A COLUMNAR GRAIN STRUCTURE REPRESENTED BY THE GRAIN NUMBER MAP OF BLUE AND RED
IN THE BACKGROUND. A TABLE OF THE CALCULATED INTERNAL ENERGIES FOR DIFFERENT CONFIGURATIONS IS SHOWN IN (D). ...... 83
FIGURE 58: SYNCHROTRON XRD PATTERN FROM A MULTILAYERED 10 NM CU/10 NM NB SAMPLE PRIOR TO TENSILE TESTING. THE CU
(111) AND NB (110) RINGS ARE IDENTIFIED. THE INNER RINGS ARE FROM THE POLYIMIDE SUBSTRATE. ................................. 110
FIGURE 59: VARIATION IN SAMPLE THICKNESS MEASURED AT MULTIPLE POINTS ALONG A SPUTTERED 20 MM DIAMETER CU FILM DEPOSITED
AT A SPUTTERING DISTANCE OF 152 MM. ................................................................................................................... 111
FIGURE 60: POLYIMIDE SUBSTRATE HOLDER FOR SPUTTERING - (A) ASSEMBLED VIEW, (B) EXPLODED VIEW. THE RECESS CUT INTO THE BASE
ACCOMMODATES A 50 MM X 60 MM POLYIMIDE SUBSTRATE. ........................................................................................ 112
XIV

FIGURE 61: XRD PLOTS FOR AS-SPUTTERED CU/NB SAMPLES ON SI (100) SUBSTRATES WITH NB AND CU SPUTTERING RATES OF
APPROXIMATELY 0.12 NM/S AND 0.16 NM/S, RESPECTIVELY. THE NB LAYER THICKNESSES ARE 100 NM AND THE CU LAYER
THICKNESSES ARE GIVEN ON THE RIGHT FOR THE DIFFERENT SAMPLES. CU AND NB PEAKS ARE INDICATED. SI PEAKS WERE REMOVED
FROM THE CURVES. ................................................................................................................................................ 113
FIGURE 62: XRD PLOTS FOR AS-SPUTTERED CU/NB SAMPLES ON SI (100) SUBSTRATES WITH NB AND CU SPUTTERING RATES OF
APPROXIMATELY 0.12 NM/S AND 0.70 NM/S, RESPECTIVELY. THE NB LAYER THICKNESSES ARE 100 NM AND THE CU LAYER
THICKNESSES ARE GIVEN ON THE RIGHT FOR THE DIFFERENT SAMPLES. CU AND NB PEAKS ARE INDICATED. SI PEAKS WERE REMOVED
FROM MOST OF THE CURVES. ................................................................................................................................... 114
FIGURE 63: XRD CURVES FOR SPUTTERED PURE TA SAMPLES. THE SAMPLE NAMES, STARTING BASE PRESSURES, SUBSTRATE TYPES, AND
INTERRUPTION DURING SPUTTERING ARE INDICATED ON THE RIGHT FOR THE DIFFERENT CURVES. THE RELEVANT ALPHA AND BETA TA
PEAKS ARE IDENTIFIED. SI PEAKS HAVE BEEN REMOVED FOR TA-2. .................................................................................. 115
FIGURE 64: XRD CURVES FOR SPUTTERED 1 NM CU/10 NM TA NMM SAMPLES. THE SAMPLE NAMES, STARTING BASE PRESSURES, AND
SUBSTRATE TYPES ARE INDICATED ON THE RIGHT FOR THE DIFFERENT CURVES. THE RELEVANT ALPHA AND BETA TA PEAKS ARE
IDENTIFIED. SI PEAKS HAVE BEEN REMOVED FOR CU-TA 1 AND CU-TA 3. ......................................................................... 116
FIGURE 65: XRD PATTERNS SHOWING THE AMORPHOUS STRUCTURES OF CO-SPUTTERED SAMPLES (A) CUZR-1 AND (B) TACU-15. ...... 117
FIGURE 66: TEM IMAGE OF 100 NM NB/2 NM CU NMM. THE GROWTH DIRECTION IS VERTICAL. THE CU LAYER LOCATION IS SHOWN BY
THE WHITE LINES. THE NB GRAIN ORIENTATIONS ARE CHANGED BY THE CU LAYER. ............................................................. 120
FIGURE 67: TEM IMAGE OF 100 NM NB/5 NM CU NMM. THE GROWTH DIRECTION IS VERTICAL. THE CU LAYER IS ROUGHLY DELINEATED
BY THE WHITE LINES. THE SPACING OF TEN LATTICE PLANES IS SHOWN FOR NB AT THE TOP AND CU IN THE MIDDLE. ................. 121
FIGURE 68: TEM IMAGES OF A 100 NM NB/5 NM CU NMM, WITH A CU LAYER IN THE MIDDLE. THE GROWTH DIRECTION IS VERTICAL.
IMAGE (A) IS THE RAW TEM IMAGE. THAT IMAGE WAS PROCESSED BY FFT, THE CU RINGS WERE ISOLATED, AND AN INVERSE FFT
WAS PERFORMED, AND THIS IS SHOWN IN (B). THE CU LAYER SHOWS PERIODICITY, WHILE THE NB REGIONS HAVE A RANDOM
MOSAIC PATTERN. ................................................................................................................................................. 122
FIGURE 69: SEM IMAGES OF FRACTURED CROSS-SECTIONS OF AS-SPUTTERED 20 NM CU/100 NM NB NMM SAMPLES, (A) CU/NB 29
AND (B) CU/NB 43. THE FRACTURE SURFACE OF (A) APPEARS TO BE MORE BRITTLE THAT THAT OF (B). ................................. 123
XV

FIGURE 70: (A) FIB CROSS-SECTIONAL IMAGES OF AN AS-SPUTTERED 20 NM CU/100 NM NB NMM SAMPLE (CU/NB 1). THE SEM
IMAGE (B) SHOWS DAMAGE INCURRED DURING FIB IMAGING (THAT CROSS-SECTIONAL SURFACE WAS FLAT PRIOR TO FIB IMAGING).
......................................................................................................................................................................... 124
FIGURE 71: SEM IMAGES OF THE SURFACES OF TWO CO-SPUTTERED HF-TI SAMPLES, (A) HF-TI 5 AND (B) HF-TI 9. .......................... 125
FIGURE 72: SEM IMAGES OF THE SURFACE OF A CO-SPUTTERED HF-TI SAMPLE (HF-TI 5) SHOWING DAMAGE PRODUCED DURING IMAGING.
......................................................................................................................................................................... 125
FIGURE 73: TWO SURFACE SEM IMAGES OF (A) A HEAT-TREATED HF-TI SAMPLE (HF-TI 9 HT2) AND (B) AN AS-SPUTTERED TA-CU SAMPLE
(TA-CU 5). THE SMALL DOTS APPEAR TO BE SOME FORM OF OXIDE OR PRECIPITATE. .......................................................... 126
FIGURE 74: XRD SCANS OF SI (100) SUBSTRATES AT DIFFERENT ORIENTATIONS. ......................................................................... 127
FIGURE 75: XRD CURVES FOR SPUTTERED MONOLITHIC TI FILMS. THE SAMPLE NAMES AND NOMINAL THICKNESSES ARE GIVEN ON THE
RIGHT. ................................................................................................................................................................ 128
FIGURE 76: XRD SCANS OF SPUTTERED MONOLITHIC NB FILMS. THE SAMPLES NAMES AND NOMINAL THICKNESSES ARE GIVEN ON THE
RIGHT. ................................................................................................................................................................ 129
FIGURE 77: XRD SCANS OF SPUTTERED MONOLITHIC CU FILMS. THE SAMPLE NAMES AND NOMINAL THICKNESSES ARE GIVEN ON THE
RIGHT. ................................................................................................................................................................ 130
FIGURE 78: XRD SCANS OF MULTILAYERED AL/NB SAMPLES. THE NB LAYER THICKNESSES ARE 100 NM, AND THE AL LAYER THICKNESSES
AND RELATIVE AL SPUTTERING RATES (FAST = 0.29 NM/S, SLOW = 0.07 NM/S) ARE INDICATED ON THE RIGHT OF EACH PLOT. THE
CURVES HAVE BEEN FLATTENED FROM 60°-80° TO REMOVE THE LARGE SI (4 0 0) PEAK. ..................................................... 130
FIGURE 79: SCHEMATIC OF THE EXISTING SPUTTERING CONFIGURATION. ALL FOUR POSSIBLE SPUTTERING GUN POSITIONS ARE SHOWN IN
(A), WITH THE SUBSTRATE HOLDER IN THE MIDDLE. THE SUBSTRATE HOLDER CAN BE MADE TO FACE ANY OF THE FOUR GUNS. TWO
GUNS ARE SHOWN IN (B) AND THEY ARE EACH INCLINED 45° TO THE SUBSTRATE HOLDER. .................................................... 131
FIGURE 80: CHART SHOWING NECESSARY TEMPERATURES AND PARTIAL PRESSURE OF WATER VAPOR NECESSARY TO PREVENT OXIDATION.
[191] ................................................................................................................................................................. 136
FIGURE 81: SCHEMATIC OF QUENCH VACUUM FURNACE. (NOT TO SCALE) .................................................................................. 138
FIGURE 82: CROSS-SECTIONAL TEM IMAGE OF A HEAT-TREATED SPUTTERED HF SAMPLE. THE BRIGHT AREAS ARE CRACKS. ................ 139
FIGURE 83: CROSS-SECTIONAL STEM IMAGE OF A HEAT-TREATED CO-SPUTTERED TA-CU SAMPLE (800 °C, 4 DAYS). THERE ARE SUB-100
NM FEATURES, BUT THE COMPOSITION OF THOSE FEATURES IS NOT YET KNOWN. ................................................................ 142
XVI

FIGURE 84: CROSS-SECTIONAL TEM IMAGE OF A HEAT-TREATED MULTILAYERED MO-AU SAMPLE (800 °C, 4 DAYS). THE DIFFERENT
VIEWS, (A) AND (B), ARE FROM DIFFERENT SIDES OF THE SAMPLE. NOTE THAT (A) SHOWS MUCH LARGER GRAINS, WHILE (B) SHOWS
A PARTIAL RETENTION OF THE MULTILAYER STRUCTURE AT THE BOTTOM OF THE IMAGE. THE GROWTH DIRECTION IS VERTICAL.... 143

XVII

List of Tables
TABLE 1: POLARIZATION DATA OBTAINED IN A 3.5 WT.% NACL SOLUTION FROM A BARE NDFEB SPECIMEN, NDFEB SPECIMENS COATED
WITH AL AND TI SINGLE LAYERS, AND NDFEB SPECIMENS COATED WITH TI/AL MULTILAYERS [47] .......................................... 20
TABLE 2: DESCRIPTION OF NB SAMPLES AND CORRESPONDING SURFACE ROUGHNESS VALUES AS MEASURED BY AFM. ......................... 37
TABLE 3: DESCRIPTION OF CU/NB AND AL/NB NMM SAMPLES SPUTTERED ONTO SI (100) SUBSTRATES. ......................................... 46
TABLE 4: TENSILE TEST SPECIMEN DESCRIPTIONS* ................................................................................................................... 58
TABLE 5: SELECTED SYSTEMS FOR NMM THERMAL STABILITY STUDY ........................................................................................... 68
TABLE 6: MULTILAYERED HF-TI SAMPLE DESCRIPTIONS ............................................................................................................ 69
TABLE 7: DIFFUSION COEFFICIENTS AND CALCULATED DIFFUSION DISTANCES ................................................................................. 79
TABLE 8: SPUTTERED CU/NB NMMS WITH VARYING NB LAYER THICKNESSES............................................................................. 106
TABLE 9: SPUTTERED CU/ZR NMMS WITH VARYING CU AND ZR LAYER THICKNESSES .................................................................. 107
TABLE 10: SPUTTERED CU/CUZR NMM SAMPLES ON POLYIMIDE SUBSTRATES WITH VARYING CU AND CUZR LAYER THICKNESSES ........ 108
TABLE 11: RESIDUAL STRESSES IN SPUTTERED CU/NB NMMS*............................................................................................... 118
TABLE 12: RESIDUAL STRESSES IN SPUTTERED AL/NB NMMS* ............................................................................................... 119
TABLE 13: RESIDUAL STRESSES IN MONOLITHIC FILMS ............................................................................................................ 119
TABLE 14: MONOLITHIC AL SAMPLES ................................................................................................................................. 145
TABLE 15: MONOLITHIC AU SAMPLES ................................................................................................................................ 145
TABLE 16: MONOLITHIC AL-5456 SAMPLES ........................................................................................................................ 145
TABLE 17: MONOLITHIC CU SAMPLES ................................................................................................................................. 146
TABLE 18: MONOLITHIC HF SAMPLES ................................................................................................................................. 147
TABLE 19: MONOLITHIC NB SAMPLES................................................................................................................................. 148
TABLE 20: MONOLITHIC TA SAMPLES ................................................................................................................................. 149
TABLE 21: MONOLITHIC TI SAMPLES .................................................................................................................................. 150
TABLE 22: MONOLITHIC ZR SAMPLES ................................................................................................................................. 150
TABLE 23: MONOLITHIC CUAL SAMPLES ............................................................................................................................. 151
TABLE 24: MONOLITHIC CUZR SAMPLES ............................................................................................................................. 151
XVIII

TABLE 25: MONOLITHIC HFTI SAMPLES .............................................................................................................................. 151
TABLE 26: MONOLITHIC MOAU SAMPLES ........................................................................................................................... 152
TABLE 27: MONOLITHIC TACU SAMPLES ............................................................................................................................. 152
TABLE 28: MONOLITHIC TAHF SAMPLES ............................................................................................................................. 152
TABLE 29: MONOLITHIC TIHF SAMPLES* ............................................................................................................................ 153
TABLE 30: MULTILAYERED AL/NB SAMPLES ......................................................................................................................... 153
TABLE 31: MULTILAYERED CU/NB SAMPLES* ...................................................................................................................... 154
TABLE 32: MULTILAYERED CU/TA SAMPLES* ....................................................................................................................... 155
TABLE 33: MULTILAYERED HF/TI SAMPLES .......................................................................................................................... 156
TABLE 34: MULTILAYERED HFTI/TI SAMPLES ....................................................................................................................... 156
TABLE 35: MULTILAYERED MOAU/AU SAMPLES .................................................................................................................. 156
TABLE 36: MULTILAYERED AL/NB SAMPLES ON POLYIMIDE SUBSTRATES ................................................................................... 157
TABLE 37: CU SAMPLES ON POLYIMIDE SUBSTRATES .............................................................................................................. 157
TABLE 38: MULTILAYERED CU/CUAL SAMPLES ON POLYIMIDE SUBSTRATES ............................................................................... 157
TABLE 39: MULTILAYERED CU/CUTA SAMPLE ON A POLYIMIDE SUBSTRATE ................................................................................ 157
TABLE 40: MULTILAYERED CU/CUZR SAMPLES ON POLYIMIDE SUBSTRATES*.............................................................................. 158
TABLE 41: MULTILAYERED CU/NB SAMPLES ON POLYIMIDE SUBSTRATES ................................................................................... 158
TABLE 42: MULTILAYERED CU/ZR SAMPLES ON POLYIMIDE SUBSTRATES* .................................................................................. 159
TABLE 43: MULTILAYERED CU/NB SAMPLES ON POLYIMIDE SUBSTRATES TESTED BY IN-SITU XRD TENSILE TESTING ............................ 160
TABLE 44: CONDUCTED HEAT TREATMENTS ......................................................................................................................... 160
TABLE 45: HEAT-TREATED SAMPLES ................................................................................................................................... 161

XIX

List of abbreviations
ACOM Automated crystallographic orientation mapping
ADF Annular dark field
AFM Atomic force microscopy
ARB Accumulative roll bonding
BCC Body-centered cubic
BFTEM Bright-field transmission electron microscopy
CBED Convergent beam electron diffraction
CLS Confined layer slip
CVD Chemical vapor deposition
DFTEM Dark-field transmission electron microscopy
EBSD Electron back-scattered diffraction
EDX, EDS Energy-dispersive X-ray spectroscopy
EELS Electron energy-loss spectroscopy
FCC Face-centered cubic
FIB Focused ion beam
HCP Hexagonal close-packed
HRTEM High-resolution transmission electron microscopy
IBS Inter-boundary slip
MBE Molecular beam epitaxy
nc nanocrystalline
NED Nano electron diffraction
NMM Nano metallic multilayer
PLD Pulsed-laser deposition
PVD Physical vapor deposition
RT Room temperature
SAED, SAD Selected-area electron diffraction
SEM Scanning electron microscopy
STEM Scanning transmission electron microscopy
TEM Transmission electron microscopy
XRD X-ray diffraction

1

1 Abstract
Nano metallic multilayers (NMMs) are nanostructured composite materials which consist of thin
(< 100 nm) alternating layers of metals. They have many attractive mechanical, electromagnetic,
and optical properties. These properties can be tailored by adjusting the layer and interface
parameters of the NMMs. In addition, the ability to tailor these sample characteristics allows for
NMMs to be used as model systems in the study of a variety of nanostructure phenomena. In the
studies described in this dissertation, the structure and behavior of NMMs under mechanical and
thermal loading were investigated. Several general results are as follow: 1) A novel transmission
electron microscopy technique revealed how 20 nm thick layers of Cu and Al more greatly refined
the overall structure of Cu/Nb and Al/Nb NMMs than 2 nm layers, 2) The load-sharing behavior
among sub-20 nm thick layers of Cu and Nb within Cu/Nb NMMs was deconvoluted, and, by
tracking the changes in deformation behavior for varying layer thicknesses, a competing interplay
of strengthening and weakening was demonstrated for the different materials, 3) The high-
temperature behavior of nanostructured Hf-Ti was investigated, with an observed increased
thermal stability for Hf-Ti NMMs as compared to monolithic Hf-Ti films, showing that the starting
configuration may strongly influence the stability of a nanostructured system. Overall, these
studies demonstrated the ability to more fully characterize NMMs and configure them for
optimized behavior under mechanical and thermal loading.

2

2 Introduction
Nano metallic multilayers (NMMs) are nanostructured composite materials which consist of thin
(< 100 nm) alternating layers of metals. These materials have been shown to have exceptional
properties such as high strength, corrosion resistance, radiation damage resistance, and high fatigue
strength, thus holding promise for applications ranging from protective coatings to load-bearing
films. In addition, their electromagnetic properties give them applications in such fields as data
storage and x-ray optics.
Their exceptional properties are seen to stem from two main sources: 1) the nanometer thicknesses
of the layers themselves, whose behavior will vary with thickness, material, crystal structure, and
uniformity, and 2) the large volume of interfaces, whose behavior will vary with sharpness,
roughness, and coherence. Adjusting these layer and interface characteristics tailors the NMM
properties; however, many NMM material systems, configurations, and loading conditions require
further research before they can be applied in industry. In addition, besides their potential
industrial applications, the ability to tailor microstructural features and local compositions allows
NMMs to be used as model systems for the study of a variety of nanostructure phenomena.
The present study focused on using advanced characterization techniques to image as-synthesized
and mechanically or thermally loaded samples, in order to both measure their properties and
investigate nanostructure stability. Cu/Nb and Al/Nb NMMs were imaged by a novel nanoscale
orientation mapping transmission electron microscope technique in order to probe the structure of
nanometer thick layers of Cu, Al, and Nb within the multilayer samples. In-situ X-ray diffraction
tensile testing was then performed on Cu/Nb NMM samples to deconvolute the deformation
behaviors of Cu and Nb layers at different thicknesses and determine the load-sharing behavior of
3

the multilayer samples. Finally, a series of NMM systems, most notably Hf/Ti, were sputtered in
different configurations in order to investigate their thermal stability and explore the predictions
of a thermodynamic nanograin stability model.
By means of these investigations, the structure and behavior of NMMs under mechanical and
thermal loading could be made more evident. These data will aid the identification of optimal
NMM configurations for load-bearing applications and advance the search for thermally stabilized
nanostructured materials.

4

3 Background
3.1 Nanostructured Materials
3.1.1 Material Properties
Nanostructured materials have characteristic dimensions under 100 nm and have been shown to
have many attractive properties, resulting in a variety of applications [1-6]. From a structural
standpoint, the most important properties may be their high strengths, which are generally modeled
by the Hall-Petch equation:
𝜎 𝑦 = 𝜎 0
+ 𝑘 𝐷 −1/2
Eq. 1

The yield strength, σy, is seen to increase with decreasing grain size, D, at a slope of k, and σ0 is
the frictional stress required to move dislocations [7]. For this equation, dislocations in a coarse-
grained material are modeled as piling up behind one another. The pile up of multiple dislocations,
as shown in Figure 1, increases the stress on the leading dislocation (labeled “0”), forcing it to
interact with the grain boundary by transmission, absorption, or nucleation of other dislocations in
neighboring grains [7]. The stress increase with increasing number of piled up dislocations is
given by
𝜏 ∗
= 𝑛𝜏 Eq. 2

where τ is the applied stress, τ
*
is the stress on the leading dislocation, and τ
*
increases with the
number of piled up dislocations, n [7]. Therefore, if the grain size is small and fewer dislocations
can pile up in a single grain, the leading dislocation will have less stress acting on it. Thus, a larger
applied stress will be needed to cause the leading dislocation to interact with the boundary.
5

Figure 1: Schematic of dislocation pile-up at an obstacle. The applied stress is τ, the distance
from the obstacle is x i, and the stress on dislocation “0” increases with every dislocation piled
up behind it. [7]

Figure 2 illustrates a typical change in strength with decreasing grain size in a material with
equiaxed grains [8]. The first part of the curve corresponds to Hall-Petch strengthening. Once the
grain size falls below approximately 100 nm into the nanocrystalline (nc) regime, the slope of the
curve is seen to decrease, corresponding to a lack of dislocation pile-up [9]. The slope can even
become negative below a grain size of 10 nm, where deformation mechanisms such as grain
boundary-sliding may begin operating [9]. Different deformation mechanisms that are activated
at these nanoscale regimes will be explained in Section 3.2.4.2 in the case of nano metallic
multilayers (NMMs).

Figure 2: Plot showing the three typical regimes for flow stress as a function of grain size in
materials with equiaxed grains. In the grain size > 100 nm regime, Hall-Petch strengthening is
occurring, with a slope of k. At grain sizes of 100 nm and 10 nm, the changes in slope indicate
deformation mechanism changes. [8]

6

3.1.2 Thermal Stability
Though nanostructured materials have many attractive properties, they generally exhibit poor
thermal stability. One reason for this is the energy stored in their many interfaces, which drives
grain growth and loss of nanostructure [3, 10]. The interfaces also provide fast diffusion paths
through the material, allowing grain growth to occur even at room temperature in some cases [11-
15]. At elevated temperatures, grain growth can happen very quickly, as shown in Figure 3 for Ni
at 420 ºC, where the nc Ni shows grain growth after only seconds (note the scale bar changes) [16].
On the other hand, some studies have shown thermal stability for up to an hour for nc Au (497 ºC)
and nc Cu (500 ºC) [17, 18].

Figure 3: Bright field transmission electron microscope images showing grain growth in Ni at
420 °C. The labels indicate the heat treatment time for each image. The grain growth begins
almost instantly (note the scale bar changes at 1 s and 11 h). [16]

as-deposited 30 s
1 h 11 h
1 s
120 h
7

3.1.3 Methods for Improving Thermal Stability
A variety of approaches to producing thermally stable nanograined materials have been used. One
approach is the use of structured grain boundaries, as shown in Figure 4, where triple junctions
were observed to stabilize the nanograins in a ternary Co-Cu-Ag system [19].

Figure 4: Binary (a) and ternary (b, c) heat treated materials. The proposed stabilization
mechanism, triple junction formation, is illustrated in (d) for ternary systems. [19]

Another stabilization method is the addition of alloying elements which are immiscible in the
matrix and preferentially segregate to the grain boundaries [20-28]. The addition of these types of
solutes results in the slowing of grain boundary motion and reduction of grain growth. Such
methods have met with partial success, as illustrated in Figure 5, where the grain growth of nc Cu
is slowed by the addition of Ta [24]. The average Cu grain size grows to only 166 nm, even after
a heat treatment of 4 hours at 1040 °C. Another study looked at nanograin growth in Fe and the
Fe-Cr system, and temporary plateaus in the grain size were observed during heat treatment, as
8

shown in Figure 6 [29]. These plateaus were attributed to solute segregation to grain boundaries,
which slowed grain growth. However, the eventual precipitation of these solutes allowed grain
boundary motion and resulted in additional grain growth. Overall, these grain growth inhibition
methods attempt to kinetically stabilize the nanostructures.

Figure 5: Grain size distributions for nanocrystalline Cu with Ta additions heat-treated for 4
hours at (a) 540 °C, (b) 770 °C, (c) 900 °C, and (d) 1040 °C. The grain size increased with
increasing heat treatment temperature, but it remained in the nano and ultra-fine grain
regimes, even for the highest heat treat temperature. [24]

(a)
(d)
(b)
(c)
9

Figure 6: Grain size as a function of the annealing time for (a) an Fe-Cr alloy and (b) an Fe
sample. The dots are experimental data and the lines are based on models. Note the
temporary plateaus in the grain size. [29]

In addition to kinetic stability models, several models for thermodynamic stability have been
advanced [20, 21, 30], where the material systems once again contain minor alloying elements that
segregate to the grain boundaries. However, in contrast to kinetic stabilization, these alloying
elements do not simply slow the grain boundary migration - rather, they reduce the energy at the
grain boundaries to the point where the material is at a lower energy state with the grain boundaries
than without them. If enough grain boundary volume can be stabilized, then the nanostructure will
be stable.
One model put forth by Chookajorn et al. examines the “free energy surface in composition-grain
size space” [31]. Minima are sought on that free energy surface and are thought to determine the
thermodynamically favored grain sizes for certain parameter sets. A plot of different W alloys,
with W being the major alloying element, was generated based on this model and is shown in
Figure 7 [31]. The x-axis is the enthalpy of mixing and the y-axis is the enthalpy of segregation.
The diagonal line separates the alloy compositions which are expected to be bulk stable (bottom
right) and nanostructure stable (top left).
10

Figure 7: Plot showing which minor alloying elements should produce nanocrystalline stable
(green circle area) and bulk stable (red cross area) materials when alloyed with W. The error
bars on Ti indicate typical uncertainties for the calculations used to generate this plot. [31]

Chookajorn et al. experimentally investigated their model by examining a nanostructured W-Ti
sample. W and Ti are soluble at the selected heat treatment temperature (1100 °C), and thus do
not meet the immiscibility criterion which is generally set forth as being necessary for stabilization
of nanostructures. However, it was shown that the ~20 nm grains of the nc W-Ti alloy were
maintained after being heat treated for a week at 1100 °C, Figure 8 [31]. The nc W without Ti
additions, on the other hand, increased in grain size to ~600 nm, Figure 8c. The preservation of the
nanograins in the W-Ti alloy is encouraging, as it suggests thermodynamic stability in the
nanostructured material. This thermodynamic stability model was subsequently applied to other
systems, and both expected bulk stable and nanostructure stable systems were identified [31-34],
though these predictions still require validation.
Overall, nanograined materials have many attractive properties, but they require much more study
in order to maximize their potential. Such studies can be facilitated by using tailored
11

nanostructures, such as nano metallic multilayers, which allow for more control of starting
configurations.

Figure 8: (a) Grain size distributions for as-milled W-Ti, annealed W-Ti, and annealed pure W
samples. Annealing took place at 1100 °C for a week. The annealed W-Ti material retained
nearly the same grain size, while the annealed pure W sample showed substantial grain growth.
TEM images indicating grain sizes for (b) as-milled W-Ti , (c) annealed pure W, and d) annealed
W-Ti reflect the grain size distributions in (a). The insets in (b) and (d) are dark-field TEM
images that better show some of the individual grains. [31]

3.2 Nano Metallic Multilayers
A nano metallic multilayer (NMM) is a specific type of nanostructured material which consists of
alternating layers of metals, with the individual layer thicknesses on the order of nanometers. An
NMM is shown in Figure 9 which consists of alternating 100 nm layers of Nb and 20 nm layers of
Cu. The behavior of the Cu and Nb layers will change with varying layer thickness; additionally
the interfacial content will increase rapidly with decreasing layer thickness. For example, if an
interface is assumed to be 0.5 nm thick and the layer thicknesses are 10 nm, then 5 percent of the
12

sample may behave similar to an interface. If the layer thickness is decreased to 2 nm, then 25
percent of the sample may show the behavior of an interface. Thus, the interfacial properties are
much more important in NMMs than in standard coarse-grained materials.

Figure 9: Cross-sectional transmission electron microscope image of Cu/Nb NMM.

3.2.1 Microstructure
A range of microstructures can be achieved in NMMs by means of different fabrication methods
and parameters, as illustrated in Figure 10a (columnar structure) and Figure 10b (non-columnar
structure) [35, 36]. Grain sizes, grain boundaries, and grain orientations within the layers can be
altered, with corresponding changes in material behavior. For example, in Figure 11 several NMM
microstructure schematics are arranged from most to least thermally stable [37].

Figure 10: (a) Cross-sectional TEM image showing the columnar microstructure in a Cu/Nb
NMM. The 2 nm-thick Cu and Nb layers are not apparent at this magnification [35], (b) Cross-
sectional SEM image of a Mg/Nb NMM without a clear columnar microstructure. Mg layers are
dark, Nb layers are bright. [36]
(a) (b)
13

Figure 11: Schematic of NMM grain boundary structures, with materials α and β, arranged by
thermal stability, from (a) most stable to (d) least stable. The vertical lines are grain
boundaries, the horizontal lines are interfaces between layers. [37]

The microstructure will often be determined by the interface types which are present: coherent,
semi-coherent, and incoherent. A coherent interface occurs when the positions of the atoms in the
two layers match at the interface plane. Both of the interfaces in Figure 12 are coherent, but in
Figure 12b the lattice is tilted to maintain coherency [38]. While tilting may not result in perfect
matching of the lattice sites, it may still give enough coincidence sites to be favored, as in Figure
13 [39]. Such a situation allows for partial coherence between crystallographically mismatched
materials, and the interface can be referred to as being commensurate if the matching sites at the
interface are periodic.

Figure 12: Two coherent interfaces are shown: (a) perfect coherent interface, (b) tilted
coherent interface. [38]

(a)
(b)
14

Figure 13: Schematic showing partial coherence between crystallographically mismatched
materials, resulting in a commensurate interface. The two materials are brought into
coherency by the tilting of one material relative to the other. The film lattice constant is d f and
the substrate lattice constant is d s. The film Miller indices are h, k, and l, while m and n indicate
the number of lattice spacings between coincident lattice positions. [39]

Another way to maintain coherency in spite of an imperfect match is by straining at the interface,
which is accompanied by coherency stress. A strained coherent interface is shown in Figure 14a,
where the bottom crystal is in compression and the top one is in tension [38]. Different levels of
coherency are possible by tilting and straining the crystals, and the optimal arrangement will
depend upon the crystal structures and misorientations of the crystals.

15

Figure 14: Schematics of different interface types: (a) strained coherent interface with the top
material in tension and the bottom in compression, (b) semi-coherent interface with
dislocations to relieve the coherency strain, and (c) incoherent interface with dislocations that
are completely overlapped, and, thus, indistinguishable. [38]

A semi-coherent interface, Figure 14b, consists of an interface in which some atomic positions
match both materials, but there are also misfit dislocations which relieve the coherency stresses in
the layers. When the mismatch is great enough, the dislocation cores begin to overlap, and the
interface becomes incoherent, Figure 14c.
3.2.2 Fabrication Methods
The structures of the interfaces and layers in NMMs will greatly depend upon their fabrication
methods, which can be grouped into two categories: top-down and bottom-up. Top-down
approaches consist of starting with a bulk material and then deforming the material in such a way
(a)
(b)
(c)
16

that the resulting structure is that of an NMM, while bottom-up approaches consist of building the
NMM layer-by-layer.
3.2.2.1 Top-Down
The main top-down technique used to fabricate NMMs is accumulative roll-bonding (ARB). ARB
consists of starting with several alternating layers of the desired materials arranged in a sandwich
structure, with layer thicknesses on the order of millimeters. The sandwich is then rolled, which
reduces the layer thickness. Once a certain number of roll passes have occurred, the sample is cut
in half, one half is placed on top of the other, and the rolling process is repeated. The rolling,
cutting, and stacking steps are repeated until the desired layer thickness is reached. As needed,
annealing steps may also be incorporated to improve bonding or reduce the stresses which
originate from the rolling procedure. The ARB process is illustrated in Figure 15 [40].

Figure 15: Accumulative roll-bonding schematic showing the steps necessary to fabricate an
NMM by this method. [40]

17

Another top-down approach which is used to produce NMM-like structures is the wire drawing
technique. Here, different materials are formed into a rod, and then drawn or extruded to form a
sample with small characteristic dimensions [41].
3.2.2.2 Bottom-Up
Bottom-up approaches consist of organizing the material in a layer-by-layer fashion. Several
techniques are used for this process, among them chemical vapor deposition (CVD),
electrodeposition, and physical vapor deposition (PVD).
CVD techniques consist of starting with a precursor gas, and then, by means of a chemical reaction,
causing the precursor to dissociate and deposit material onto a substrate. A variety of precursors
(e.g. silanes and organometallics), plasmas, and heating techniques can be used, which result in a
multitude of CVD methods.
Electrodeposition consists of first placing a substrate into a bath of dissolved metal ions. The
dissolved metal is then deposited onto the substrate by means of applied electric fields. Multilayers
can be fabricated by using either single-bath techniques or dual-bath techniques [42, 43]. Single-
bath techniques require multiple types of metal ions in one bath. One of the single-bath techniques
consists of long voltage pulses, which first result in the preferred deposition of material A onto the
sample. Then, since material A has been preferentially deposited and is depleted directly around
the sample, material B is more abundant for the remaining duration of the voltage pulse. This
results in a graded composition in the layers.
Another single-bath technique uses multiple pulse types, with varying intensities, shapes, and
durations, in order to preferentially deposit the different materials. Overall, single-bath techniques
do not produce sharp interfaces or layers consisting of purely one material. Dual-bath techniques
18

can give sharper interfaces, but, since they consist of moving the sample between different baths,
they are generally more complicated than single-bath techniques [43].
PVD techniques consist of depositing a material which starts in a solid form and ends in a solid
form, without intermediate chemical reactions. There are a variety of PVD techniques, but only
molecular beam epitaxy (MBE), e-beam evaporation, pulsed-laser deposition (PLD), and
sputtering will be briefly explained here. All of these techniques share the common characteristic
of being performed in a vacuum environment.
MBE consists of starting with a solid material and heating it. The heating causes the material to
sublimate in the vacuum chamber and deposit onto the substrate. The deposition of alternating
layers of different materials can be achieved by shutters or by movement of the substrate between
the different materials being deposited. MBE is generally performed in ultra-high vacuum
conditions, and thus produces films with very little contamination. However, MBE has relatively
low deposition speeds.
E-beam evaporation is similar to MBE in that a solid material is made to evaporate and deposit
onto a substrate. E-beam evaporation differs in that it uses an electron beam to evaporate atoms
from the surface of the source material. PLD is a similar technique, but it uses laser pulses instead
of an electron beam to evaporate the materials. Once again, use of shutters, activation of different
sources, or movement of the sample between different materials can be used to form the different
layers.
3.2.2.3 Sputtering
Sputtering was used in this study and is shown schematically in Figure 16. The method consists
of introducing a sputtering gas, Ar in this case, into a vacuum chamber. A bias voltage is then
19

applied to the target (source material), which causes the Ar to become ionized, forming plasma.
The Ar ions are attracted to the target and strike it. The impact of the Ar ions on the target surface
results in the ejection of the target atoms, which travel across the vacuum chamber and deposit
onto the substrate.

Figure 16: Schematic showing the main steps of magnetron sputtering: 1) an Ar gas is
introduced and is ionized by the negatively biased target, 2) the Ar ions strike the target,
causing the ejection of target atoms, 3) the target atoms deposit onto the substrate.

In our case, a strong magnet was used to better concentrate the Ar plasma; this technique is referred
to as magnetron sputtering. Magnetron sputtering increases the sputtering rates, and, in turn,
deposition rates. Magnetron sputtering generally has faster deposition rates than the MBE or e-
beam techniques. Fast deposition ( > 0.1 nm/s) is critical for the fabrication of samples which are
thick enough to mechanically test ( > 1 μm for nanoindentation, > 10 μm for microtensile testing).

Ar ions
(plasma)
Negatively
biased
target
Ejected
target
atoms
20

3.2.3 General Properties
NMMs show many greatly improved properties compared to bulk materials. For example, Cu/Nb
and Cu/V NMMs have been shown to be radiation-damage resistant [44, 45]. This resistance to
damage is attributed to the presence of a multitude of interfaces, which causes a decrease in
diffusion distance for defects induced by radiation.
The corrosion resistance is also seen to improve over that of bulk materials [46, 47]. As shown in
Table 1, the much smaller corrosion currents of the multilayers indicate that they have a lower
corrosion rate than the single layers of Ti and Al [47]. Some explanations for the improved
corrosion resistance are 1) the possibility of different compositions and phases in some layers, 2)
changes in dielectric properties with layer thickness changes, and 3) the torturous corrosion path
which must be followed in an NMM when the interface boundaries between the layers are not
aligned, resulting in deflection of the corrosion cracks by the different layers [46, 47].
Table 1: Polarization data obtained in a 3.5 wt.% NaCl solution from a bare NdFeB specimen,
NdFeB specimens coated with Al and Ti single layers, and NdFeB specimens coated with Ti/Al
multilayers [47]

The electromagnetic properties of NMMs have also been studied, revealing that they have
applications as TES microbolometers and x-ray mirrors [48-51]. A giant magnetoresistance effect
is seen in some material systems such as Cu/Co and Fe/Cr, which is useful in the field of data
storage [52-54].
21

3.2.4 Mechanical Properties
In addition to the properties mentioned above, NMMs have attractive mechanical properties. A
variety of testing techniques have been used to measure these properties and determine which
deformation mechanisms are operating.
3.2.4.1 Mechanical Testing Methods
Much of the mechanical testing of NMMs has been conducted by nanoindentation.
Nanoindentation is a technique which consists of indenting a surface using a very small tip (tip
radius ~200 nm). It is similar to other hardness testing techniques such as Rockwell and Vickers
hardness testing, but it differs in the resolution that it provides. By virtue of being able to give
quantitative data with indent depths of only 100 nm, this technique lends itself to the study of thin
films, such as NMMs, while allowing the samples to remain attached to substrates. However,
indentation techniques do not reveal much about the ductility of the materials.
Another compression testing method is deformation of nanometer-sized pillars. The samples are
milled by a focused ion beam (FIB) and compressed with a flat-punch nanoindenter tip, as shown
in Figure 17 [55]. While this testing method is closer to a standard compression test than
nanoindentation, the samples will have some FIB damage and the bottoms of the pillars are still
attached to the substrate, both of which introduce testing artifacts.
22

Figure 17: SEM images of 800 nm diameter Cu/Zr pillars in the as-milled state (left) and
compressed state (right). The layer thicknesses of the samples are (a, b) 100 nm, (c, d) 50 nm,
and (e, f) 20 nm. Extrusion of the Cu layers is evident in (b), while (d) and (f) show shearing of
the pillars. [55]

If the samples are thick enough to become freestanding films, then standard tensile tests can be
performed. Samples can also be tested while attached to soft substrates such as polyimide, and the
mechanical contributions of the polyimide substrates can be subtracted, giving stress-strain curves
for just the NMM films [56, 57]. Though tensile tests give both strength and ductility information,
it is not possible to deconvolute the behaviors of the different layers in order to determine which
material layers are bearing which proportion of the load.
23

Nanoscale tensile tests and nanoindentation tests can also be executed in-situ while performing
TEM imaging, and such tests have recorded NMM deformation behaviors such as dislocation
channeling within individual layers and nanoscale crack propagation [58, 59]. However, such tests
are currently more qualitative than quantitative.
3.2.4.2 Strength Improvement
For a variety of testing techniques, the strength of NMMs increases with decreasing layer size.
This is illustrated in Figure 18, where the hardness increases from 2-3.5 GPa to 4-7 GPa with a
decrease in layer thickness from 200 nm to 1 nm [60]. The fact that this trend holds true for a
variety of material systems indicates that this behavior is a result of the layered structure, not just
a particular material system. In addition to having high strengths, such materials are also strong
in fatigue loading [61, 62], holding promise for protective coatings and structural applications.

Figure 18: Hardness values for different NMM systems as a function of the inverse square of
layer thickness. [60]

NMM strengthening behavior is illustrated in Figure 19 and can be compared to the grain size
strengthening in Figure 2 (note that the x-axes are reversed, with smaller layer thicknesses being
on the left in this case) [63]. The difference between NMMs and nc materials is that the
24

characteristic dimension in NMMs can be the layer thickness rather than the grain size. The grain
sizes may still govern the deformation of NMMs if they are comparable to the layer thicknesses,
but that will vary with the fabrication and processing [64].

Figure 19: Illustrative plot showing the three strength regimes observed in NMMs at different
layer thicknesses. The corresponding deformation mechanisms are shown for the different
regimes. [63]

While NMMs initially show Hall-Petch strengthening with decreasing layer thickness (right side
of Figure 19), a change in dislocation mechanism is observed when the layer thicknesses decrease
below 50-100 nm, because the layer thicknesses are too small to allow for dislocation pile-up.
Instead, a confined-layer slip (CLS) mechanism is observed, which operates by the bowing of
dislocations within a single layer, as shown in the middle of Figure 19 [58, 65]. The CLS
mechanism has been experimentally observed by TEM and shows a different slope in strength
with decreasing layer thickness than the regime with layer thicknesses above 100 nm [58].
With a further decrease in layer thickness below about 5 nm, there is another change in deformation
mechanism. The decreasing layer thickness forces the bowing dislocations to assume very small
25

bowing radii, requiring higher stresses. With increasing stress necessary to bow the dislocations,
the mechanism switches from CLS to inter-boundary slip (IBS). In this regime, the necessary
stress for a dislocation to cross a boundary is lower than that required for further dislocation
bowing. Therefore, dislocations are no longer confined to one layer, and the strength of the
samples is generally seen to either stay the same with decreasing layer thickness, or even decrease,
as shown in the IBS regime in the left section of Figure 19.
3.2.4.3 Interface Strength
The ability of dislocations to cross the interface and show IBS deformation depends upon the
interface strength [66]. The first component of the interface strength is the Koehler stress. This
stress is due to the mismatch in the elastic moduli of the two layers, which inhibits crossing of the
dislocations from the softer to the harder layer [67].
The second component is due to the coherency stresses at coherent and semi-coherent interfaces.
The coherency stresses may repel or attract dislocations, depending upon their sign [63, 66].
Dislocations at semi-coherent interfaces will also affect dislocation crossing, since the interface
dislocations have their own stress fields.
Chemical mismatch affects boundary strength, since the passage of a dislocation changes the
nature of the boundary and will create more material A-material B bonds [66]. Whether or not A-
B bonds are favored over A-A and B-B bonds depends upon the miscibility of the materials. The
miscibility of the materials will also affect the sharpness of the interfaces which are produced, as
miscible materials are more likely to form a diffuse interface, which would allow for easier
dislocation transmission [68].
26

Misorientation between the slip planes of the two layers greatly affects the strength of the interface
[66]. An interface which has well-aligned slip planes between the two materials, as is true for the
epitaxial Ti and Al layers in Figure 20, will allow for easy dislocation transmission [69]. However,
an interface between two highly misaligned slip planes will require a higher stress to transmit the
dislocation, and there will be more dislocation debris left at the interface after the dislocation is
transmitted [70].

Figure 20: High-resolution transmission electron microscope image showing a continuous FCC
structure at an Al/Ti interface. This results in a coherent interface that will be amenable to
dislocation crossing. [69]

Finally, the strength of the boundary in shear must be considered. Coherent boundaries will be
strong in shear than incoherent boundaries [71, 72]. The boundaries which are weak in shear will
hinder dislocation crossing, because the dislocation cores will more easily dissociate in the
interface [73, 74]. This dissociation makes it difficult for the dislocations to re-form on the other
side of the interface, in effect trapping the dislocations in the interface.
3.2.4.4 Deformation Behaviors
In addition to the large amount of strengthening that has been reported for NMMs with decreasing
layer thicknesses, other deformation characteristics of NMMs have also been studied. Based on
27

tensile testing, NMM ductility is generally poor and decreases with increasing strength, as seen in
the stress-strain curves for Ag, Cu, and Ag/Cu samples in Figure 21 [75]. However, some material
systems, such as the crystalline Cu/amorphous Cu-Zr system, show elongation above 10% at
certain layer thicknesses, Figure 22 [76].

Figure 21: Stress-strain curves for monolithic Ag, monolithic Cu, and Ag/Cu NMMs with
decreasing layer thickness. [75]

28

Figure 22: Ductile tensile deformation of a Cu/CuZr NMM, compared to a Cu/304 SS NMM and
pure Cu. [76]

Changes in deformation behavior with varying layer thicknesses are suggested by compression
testing as well. For example, hardness testing of NMMs with thinner layers can result in indents
that have shear bands which are not present for NMMs with thicker layers [77]. In some systems,
there is evidence of more prevalent deformation in the softer layers, and this is seen in nanopillar
compression testing, Figure 17 [78, 79]. These and a variety of other changes in deformation have
been observed to take place when the material systems and layer thicknesses were varied.
3.3 Thermal Stability
Many applications demand that NMMs not only be strong, but also thermally stable. For example,
energy production often requires operating conditions of hundreds of degrees Celsius. If the
structures of the materials used are not properly tailored, the material properties will degrade with
time. This must be considered for all nanostructured materials if they are to be used industrially.
29

NMMs have many of the same thermal stability issues as nanostructured materials (Section 3.1.2).
Multiple studies have examined the stability of NMMs, with varying results [52, 80-84]. Some
systems maintain their layered structure up to 800 °C [85]. Other systems show a large change in
structure and decrease in properties with just a 400 °C heat treatment [86].
The loss of the layered structure is undesirable, because the properties of NMMs stem directly
from the presence of small layers and many interfaces. However, the interfaces are detrimental
with regard to the thermal stability of the NMMs, because they are both the source of the excess
energy in the materials and the preferred paths for diffusion. The excess energy stems from the
stresses, additional free volume, and chemical interactions at the interface. Different interfaces
are expected to have different levels of stability, with sharp, coherent, immiscible interfaces and
vertically aligned grain boundaries, Figure 11, being more stable [37, 87]. In addition to these
considerations, the thermodynamic stability models discussed in Section 3.1.3 may prove useful
in the investigation of NMM thermal stability, since NMMs are multi-element materials. If such
models can aid the development of thermally stable NMMs, the field of possible NMM
applications would be greatly expanded.
3.4 Summary
NMMs are nanostructured materials which consist of alternating sub-100 nm layers of metals.
These materials have been shown to be strong, radiation-damage resistant, corrosion-resistant, and
electromagnetically useful. However, their behavior depends greatly on their microstructures, and
imaging the nanoscale features of NMMs requires specialized characterization techniques. In
addition, their thermal stability is not sufficient for many applications. Therefore, the studies
30

described in the following sections sought to address the issues of microstructural control,
mechanical deformation, and high-temperature behavior of NMMs.

31

4 Preliminary Work
A variety of preliminary work was conducted to adapt equipment and investigate the use of certain
characterization techniques for the specific needs of the current study.
4.1 Sputtering Chamber Modification
The purpose of the chamber modification was to improve the control available for sputtering very
thin layers of material. Figure 23a is a schematic of the chamber used in this study. The substrate
can be rotated by a motor in order to face the different sources for the sputtering of different
materials. The modifications to this chamber consisted of installation of a shutter in the sputtering
chamber, design and installation of a co-sputtering flange, addition of a larger power supply,
addition of heating tapes, and installation of a new argon flow system.

Figure 23: (a) Sputtering chamber schematic showing the different sources which allow for
sputtering of multiple materials and the motor for spinning the substrate to face the different
sources, (b) picture of the sputtering process taking place, with the components labeled.

The shutter that was installed above one of the sputtering sources is shown in Figure 23b. This
shutter is controlled by a set of automatic pneumatic valves and controls the amount of material

Shutter
Substrate
holder
Source
material
Plasma
(a)
(b)
32

that is deposited onto the substrate for each of the individual layers, allowing for precise control
of the layer thicknesses.
In order to allow for co-sputtering of two different materials, a custom-designed flange was
installed onto the chamber. The flange schematic is shown in Figure 24, with two big ports for
sputtering guns, two medium ports for shutters, and a small port for an electrical feedthrough. The
gun ports were aligned to meet at the center of the sputtering chamber, where the substrates are
placed. Using this flange, two materials can be simultaneously sputtered onto the substrate.

Figure 24: Co-sputtering flange schematic. The two ports with 2.75” ConFlat (CF) flanges are for
sputtering guns, the two 2.12” CF flanges are for shutters, and the central 1.33” CF flange is a
multipurpose port for probes. The 2.12” and 2.75” CF flanges are inclined 12° to the vertical;
the 2.12” and 2.75” CF flanges in each pair (right and left side of the image) are parallel to one
another. The 2.75” CF ports are aligned to intersect 291 mm below the flange.

A new Pinnacle Series 6 kW DC power supply was also installed, which allows for sputtering at
higher powers and, in turn, higher deposition rates (Advanced Energy). Since the film
characteristics can vary greatly with deposition conditions, the film structures could thus be more
tightly controlled. The addition of heating tapes allowed for baking of the chamber during pump
down, resulting in lower vacuum levels (24 h pumpdown – 7x10
-5
Pa), which can strongly affect
the microstructure of sputtered films. For example, the presence of high (2x10
-5
Pa) and low (2x10
-
3
Pa) vacuum prior to deposition has been shown to strongly influence the formation of alpha and
33

beta tantalum [88]. The heating tapes were used on the vacuum chamber at various temperatures,
heating durations, and pump down times in order to determine recipes for reaching different
vacuum levels.
A new argon flow system was also installed in the chamber and allowed for tighter control of the
Ar flow (models 250E-4-D, 0248A-00200SV, 627DU5TCD1B, from MKS Instruments). Argon
is used as the sputtering gas in the chamber and different Ar pressures will change the
microstructures of the films, according to the Thornton sputtering model [89]. Thus, the new argon
flow system allowed for better control of the film microstructure.
4.2 Temperature Measurements during Sputtering
Based on the Thornton model mentioned above, the microstructure will also vary with temperature
[89], where higher temperatures will allow for more surface diffusion and favor a more
thermodynamically stable structure. Thus, the temperature must be well-characterized in order to
control the microstructure. Temperature measurements were taken during sputtering using a
substrate with an embedded thermocouple that allows for live temperature monitoring during
sputtering, Figure 25 (KLA Tencor).

Figure 25: 51 mm diameter Si substrate with embedded thermocouple for temperature
measurements during sputtering.
34

The temperature of the substrate was monitored at different sputtering powers, pressures, and
distances from the source to the substrate. The temperature is shown as a function of time for
different powers in Figure 26a, and there is a substantial increase in temperature with increasing
sputtering power. On the other hand, the temperature is inversely related to the sputtering pressure,
as shown in Figure 26b, where the 0.13 Pa sputtering pressure produced the highest temperatures,
while the 1.33 Pa sputtering pressure showed the lowest temperatures.

Figure 26: Temperature plots during sputtering as a function of time for (a) different powers
(0.27 Pa Ar pressure, 152 mm sputtering distance, 51 mm diameter Al sputtering target) and (b)
different argon pressures (300 W sputtering power, 152 mm sputtering distance, 51 mm
diameter Al sputtering target).

In addition to experiencing heating during sputtering, the samples also experienced cooling periods
as they were being rotated to face the different material sources. An example of a cooling curve
is shown in Figure 27, where a quick initial decrease in temperature is evident. Given this heating
and cooling data and the small layer thicknesses used in this study, the majority of samples were
not expected to exceed 100 °C, since they did not have time for substantial heating during the
sputtering of one layer, and they quickly cooled as the substrate was being rotated.
35

Figure 27: Sample cooling curve under vacuum directly after sputtering.

4.3 Quench Vacuum Furnace
A quench vacuum furnace was designed and assembled and is described in Appendix H.3. This
furnace allowed for materials to be heat-treated and quenched under vacuum in order to retain the
high-temperature microstructure of the materials while limiting sample oxidation. Overall, the
modification of the sputtering chamber and design of a quench vacuum furnace allowed for tighter
control during sputtering, more flexibility with regard to the obtained microstructures, and heat
treatment and quenching of samples in a vacuum environment.

36

5 Characterization
Several specialized characterization techniques that were used are described below, as well as
some of the imaging that was performed for samples not directly related to the current study.
5.1 Atomic Force Microscopy
As will be shown later for the Al/Nb samples in Figure 38, roughness can develop in the films
during sputtering. The presence of such roughness in NMM samples will result in projection of
the different layers during TEM imaging, which complicates the interpretation of the images. A
reduction in roughness can be achieved by varying sputtering parameters, among them sputtering
power, sputtering pressure, base pressure, and sputtering distance [90, 91].
In order to characterize the roughness produced by different sputtering parameters, atomic force
microscopy (AFM) was performed. This is a technique in which a sharp tip is brought close to a
surface, and the tip-surface interactions are used to characterize the sample. As the tip is scanned
across the surface, the surface features cause deflections of the tip. A piezoelectric transducer is
used to maintain the spacing between the sample and the sharp tip. The contraction and expansion
of the transducer allows for the height of the sample at different points to be accurately determined.
The basic AFM imaging mode is contact mode, in which the tip is moved across the sample while
in contact with the surface. The imaging mode used in this study was tapping mode, wherein the
tip oscillates above the surface, and the piezoelectric transducer contracts and expands in such a
way that the amplitude of oscillation remains constant. This imaging mode is slower than contact
mode, but it does less damage to the surface.
37

Other imaging modes can be used which detect electric properties, magnetic properties, elastic
modulus variations, and other properties, but these were not used in this study.
Multiple Nb samples were sputtered at different sputtering pressures, powers, and distances with
thicknesses between 1 and 1.3 μm. The sputtering conditions and roughness values are shown in
Table 2. The table shows that a large difference in roughness can be produced by the use of
different sputtering conditions. The illustrative 3-D plots in Figure 28, which are from Sample 31
and Sample 43 in Table 2, illustrate this difference in roughness. Once the roughness values were
determined for different sputtering parameters, that information was used to sputter NMM samples
that provided flatter layers and sharper interfaces for TEM imaging.
Table 2: Description of Nb samples and corresponding surface roughness values as measured by
AFM.
Sample
Number
Sputtering
Power (W)
Ar
Pressure
(Pa)
Sputtering
Distance
(mm)
Roughness
(Rq, nm)
31 100 0.67 152 11.4
33 100 0.40 152 3.25
35 50 0.40 152 3.85
37 200 0.40 152 3.21
39 200 0.67 127 5.90
43 200 0.40 127 2.79
45 250 0.67 127 3.69
47 250 0.40 118 3.35
49 250 0.27 118 2.85

Figure 28: 3-D AFM plots of Nb films sputtered using different parameters; the samples are (a)
Sample 31 and (b) Sample 43 in Table 2 above.

(a) (b)
38

5.2 Transmission Electron Microscopy
A variety of standard TEM imaging modes are used for NMM characterization, among the more
common being bright (BF) and dark-field (DF) TEM, scanning transmission electron microscopy
(STEM), energy dispersive x-ray spectroscopy (EDX or EDS), and selected area electron
diffraction (SAED). Additional techniques such as automated crystallographic orientation
mapping (ACOM) and electron energy loss spectroscopy (EELS) provide complementary data to
these established techniques. These were used in the present study and are explained below.
5.2.1 Automated Crystallographic Orientation Mapping
ACOM produces an orientation map of the sample, similar to electron back-scattered diffraction
(EBSD). The technique consists of first capturing nano electron diffraction (NED) patterns, such
as those shown in Figure 29, at every point as the beam is rastered across the sample. The NED
patterns are then compared to a database of expected patterns for the system under study. This
allows for the orientation to be determined at each individual point in the sample.

Figure 29: Sample NED patterns from an Al/Nb NMM sample. The different diffraction patterns
indicate differing grain orientations.

The spatial resolution of this technique is on the order of one or two nanometers, which is much
finer than that available by traditional EBSD (tens of nanometers). Since nanograin diameters and
NMM layer thicknesses can be on the order of 10 nm or less, ACOM is necessary for the
39

determination of their orientation. The fine resolution of the technique can be seen in Figure 30a,
where even the 5 nm grains are clearly distinguished and indexed [92]. If needed, ACOM can be
used for larger scan areas, as shown in Figure 30b. Overall, this technique allows for the
orientations of many individual grains to be mapped, giving much information about an NMM’s
texture, which strongly influences its properties.

Figure 30: ACOM images of (a) fine-grained Pd and (b) tensile-stressed Al. The orientations
normal to the viewing direction are indicated by the colors of the stereographic triangles. [92]

5.2.2 Electron Energy Loss Spectroscopy
Electron energy loss spectroscopy (EELS) is a technique which determines elemental and chemical
information about a sample. It does so by measuring the different losses in energy as electrons are
inelastically scattered. Figure 31 shows an EELS spectrum, which displays the number of
electrons that passed through the sample with different amounts of energy loss: zero-loss, low-
loss, and core-loss [93]. It is evident that the majority of the electrons are in the “zero-loss” peak,
which corresponds to the electrons passing through the sample with minimal energy loss. Since
the low-loss and core-loss edges are much smaller than the zero-loss peak, the spectrum value in
Figure 31 is multiplied at the places where “x50” and “x10
6
” are indicated.
40

Figure 31: Typical EELS spectrum showing zero-loss peak, low-loss edge, and core-loss edges.
The presence of the core-loss edges can be used to generate composition maps of the sample.
Note the multiplication of the spectrum value at approximately 10 eV and 240 eV, which is
needed due to the steep decrease in signal with increasing energy loss. [93]

In addition to the zero-loss peak, there is a low-loss “edge,” corresponding to plasmon interactions.
Finally, there are multiple core-loss edges, such as those labeled as corresponding to the C K-edge
and the Mn L-edges, indicating the presence of these two elements in the spot from which this
spectrum was obtained. By determining which parts of the samples produce edges corresponding
to certain elements, the composition of the material can be mapped. Therefore, this technique can
yield elemental data similar to EDX, but it is faster and has higher energy resolution (it can resolve
elements that have similar characteristic X-ray peaks).
5.2.3 TEM Nanotwin Imaging
TEM imaging was initially performed with nanotwinned CuAl samples, which have features which
are similar to those of NMMs and thus have comparable TEM imaging requirements. Similar to
NMMs, nanotwinned materials have small characteristic dimensions and a multitude of interfaces,
in this case, twin boundaries [94]. While nanotwinned samples consist of a bulk material, not
alternating layers of materials, they still exhibit some of the same behavior as NMMs. The
41

coherency at the twin interface and the discontinuity in the slip systems across the interface cause
the twins to mechanically behave similar to the layers in the NMMs, with nanotwinned materials
showing exceptional strengthening with decreasing twin spacing [94-96].
TEM imaging of CuAl nanotwinned structures was performed in order to determine grain sizes
and twin spacings in the materials. Some of the sample cross-sectional images which were taken
are shown in Figure 32. Figure 32a shows a low-magnification BF-TEM image in which the
columnar grains have many twins. Figure 32b shows a higher magnification BF-TEM image in
which individual nanotwins can be clearly distinguished. A typical SAED pattern indicating a
highly twinned structure is shown in Figure 32c [97, 98].

Figure 32: TEM images of a nanotwinned CuAl sample at (a) low and (b) high magnification.
Multiple highly twinned grains are evident in (a). In (b), the individual twins can be measured
and counted. In (c) an SAED pattern is given from the same sample and is characteristic of a
highly nanotwinned structure oriented to a [110] zone axis.

This TEM work contributed to a study which looked at the control of twin spacing in CuAl alloys
[99]. That study found that the twin spacing of sputtered Cu and CuAl samples could be controlled
by the Al composition and the sputtering rate. The twin spacing distributions for three of those
samples are shown in Figure 33, with a variation in average twin spacing from ~2 nm to ~10 nm
[99]. The TEM imaging modes used for these samples can be used to determine layer thicknesses
and preferred orientations in NMMs.
(a) (b) (c)
42

Figure 33: Twin spacing distributions for CuAl alloys (a, b) and Cu (c). The green lines are curve
fits. [99]

5.2.4 TEM Imaging of Al 5456 Beta-Phase Growth
TEM imaging was also performed for Al5456 samples in order to determine their susceptibility to
sensitization after heat treatment. Cross-sectional imaging identified regions of Mg grain growth
at Al grain boundaries in coarse-grained samples, Figure 34 [100]. It was observed that there was
a lack of the Mg grain growth in sputtered samples, indicating the benefit of sputtered materials
for improved corrosion resistance [100]. Similar elemental mapping was subsequently applied to
NMM samples.

Figure 34: Cross-sectional TEM images showing Mg grain growth at the grain boundaries of a
heat-treated Al 5456 sample (a, b). The orientations of grain boundaries which were examined
for Mg grain growth are shown in (c). [100]
43

5.3 X-ray Diffraction
X-ray diffraction (XRD) is a characterization technique which uses x-rays to determine the
structure of samples. It is generally used to identify composition, grain sizes, and strains in a
sample. In addition to these standard uses, XRD can be used for coherent multilayers to identify
characteristic superlattice peaks, which appear when NMMs have very small layer thicknesses, as
shown in Figure 35. While only two peaks are expected for this sample (a Cu/Nb NMM) in the
30-50° scan range (Nb at 38.5° and Cu at 43.3°) there are at least six peaks present. The additional
peaks are the superlattice peaks, and they allow for the determination of the distinctness of the
layers and calculations of the layer thicknesses [101-103].

Figure 35: XRD scan of a Cu/Nb NMM showing multiple superlattice peaks. An XRD scan from a
non-NMM sample containing Nb and Cu would have had only two peaks in this scan range: Nb
at 38.5° and Cu at 43.3°.

Grazing-incidence XRD is used for thin films, because it probes only the surface (tens of
nanometers) of a sample. Low-angle XRD scans also provide useful NMM information; an
example is given in Figure 36. This is a scan a Cu/Ta NMM, and the spacing between the peaks
can be used to determine layer thickness and uniformity [104-106].
44

Figure 36: A low-angle XRD scan of a Cu/Ta NMM showing multiple peaks, which can be used to
estimate layer thickness and uniformity.

Overall, a variety of characterization techniques were explored and applied to preliminary
investigations of nanostructured materials, with the goal of preparing for the NMM investigations
which are the focus of this study.

45

6 Microstructural Variations in Cu/Nb and Al/Nb Nano Metallic Multilayers
While Cu/Nb and Al/Nb NMMs have been the subject of previous studies, the effects of their local
microstructural variations on their mechanical behavior require further investigation. In this study,
Cu/Nb and Al/Nb microstructures with different layer thicknesses were examined by a specialized
TEM technique and their mechanical behavior was investigated by nanoindentation.
6.1 Overview
NMMs with alternating Cu and Nb layers (Cu/Nb) and Al and Nb layers (Al/Nb) were sputtered
onto Si (100) wafers. The descriptions of the samples used in this study are given in Table 3. The
samples had a constant Nb layer thickness of 100 nm, while the Cu and Al layer thicknesses were
varied from 1 nm to 20 nm, as shown schematically in Figure 37. The overall thicknesses of the
samples varied from 1.1 to 1.3 μm. Each of these samples was sputtered in two manners: one with
the Cu or Al sputtered at 50 W (slow deposition rates – Cu: 0.15 nm/s, Al: 0.07 nm/s) and another
with the Cu or Al sputtered at 250 W (fast deposition rates – Cu: 0.69 nm/s, Al: 0.29 nm/s). The
Nb layers were kept at a constant thickness and sputtering power (100 W, 0.12 nm/s deposition
rate) in order to better isolate the effects of the thin Cu and Al layers. The samples will be referred
to by the thickness of the thinner layers, i.e., the sample with 100 nm Nb layers and 20 nm Cu
layers will be called 20 nm Cu.

46

Table 3: Description of Cu/Nb and Al/Nb NMM samples sputtered onto Si (100) substrates.
Cu/Nb Al/Nb
Sample Nb layer
thickness
(nm)
Cu layer
thickness
(nm)
Fast/slow
Cu rate
Total
thickness
(μm)
Nb layer
thickness
(nm)
Al layer
thickness
(nm)
Fast/slow
Al rate
Total
thickness
(μm)
1a 100 1 fast 1211 100 1 fast 1211
1b 100 1 slow 1211 100 1 slow 1211
2a 100 2 fast 1222 100 2 fast 1222
2b 100 2 slow 1222 100 2 slow 1222
5a 100 5 fast 1255 100 5 fast 1255
5b 100 5 slow 1255 100 5 slow 1255
10a 100 10 fast 1310 100 10 fast 1310
10b 100 10 slow 1310 100 10 slow 1310
20a 100 20 fast 1300 100 20 fast 1300
20b 100 20 slow 1300 100 20 slow 1300

Figure 37: Schematic of Cu/Nb and Al/Nb NMMs sputtered for this study which consist of thick
Nb layers (100 nm) and thin Al or Cu layers (1-20 nm).

NMMs are generally characterized by standard TEM techniques such as high-resolution TEM
(HRTEM) and selected area electron diffraction (SAED). HRTEM gives much information about
the grains and microstructure, but it is a very local technique and is highly dependent upon sample
orientation. SAED is a global technique which gives overall texture information, but it lacks local
information about single grains, unless very many small area SAED patterns are taken.
Automated crystallographic orientation mapping (ACOM), on the other hand, gives the local
texture on a global scale, i.e. orientations of individual grains throughout a substantial portion of
47

the material (see Section 5.2.1), and thus was used in this study. The spatial resolution for the
images taken is about 1-2 nm, allowing for the identification of very fine structures.
6.2 Microstructure
The samples in Table 3 with 2 and 20 nm Cu and Al layers with fast sputtering rates (Cu: 0.69
nm/s, Al: 0.29 nm/s) were selected for microstructural characterization. STEM and ACOM images
of the 20 and 2 nm Al samples are given in Figure 38, with the growth direction indicated on the
left [107]. The STEM image of the 20 nm Al sample in Figure 38a shows the Nb layers (bright)
and Al layers (dark) in this sample. The presence of a layered structure is evident, even though
there is roughness in the layers. A STEM image of the 2 nm Al sample is shown in Figure 38b.
The Al layers are present, as pointed out by the arrows, though there is overlap with the Nb grains.
For both samples, the grain structure is not very clear and must be imaged by other techniques
such as ACOM.

Figure 38: Representative cross-sectional STEM images of Al/Nb multilayers with (a) 20 nm and
(b) 2 nm Al layer thicknesses. Corresponding cross-sectional ACOM images for the (c) 20 nm
and (d) 2 nm Al samples. The Al layers are aligned for the STEM and ACOM images. The grain
orientations normal to the images for (c) and (d) are indicated by the stereographic triangle
inset. The arrows in (b) indicate the Al layers. The growth direction is indicated on the left.
[107]
48

The ACOM images of the 20 nm Al and 2 nm Al samples are shown in Figure 38c and Figure 38d,
respectively. The different colors indicate the different orientations of the grains in the direction
normal to the image, with the colors corresponding to those in the orientation triangle inset in
Figure 38d. The STEM and ACOM images are aligned so that the Al layers are at the same levels
in both images. The Al layers for the 20 nm Al sample, Figure 38c, are not very clear, which is
due to the roughness mentioned earlier that results in overlap between the Nb and Al signal.
However, the location of the Al layers is still made evident by the Nb grain orientation changes at
the expected 100 nm spacings, shown by the color changes of the Nb grains above and below the
Al layers.
Another point of note in Figure 38c is that both crystalline and amorphous regions were observed
in the Al layers. The amorphous state is indicated in this ACOM image by the presence of many
pixels of different colors in a small area. However, due to the overlap of the Al layers with the Nb
layers in projection, amorphous and crystalline regions will be further discussed for the 20 nm Cu
samples, which had smoother layers. What is interesting is that both the amorphous and crystalline
regions of the Al layers effectively interrupted the orientations of the Nb grains.
On the other hand, the 2 nm Al layers failed to interrupt the Nb grain orientations, as shown by the
continuous Nb grains in Figure 38d. Since the layers are visible in the STEM images, which
emphasize atomic number contrast, but not in the ACOM images, which detect lattice spacings
and angles, this is an indication of epitaxial growth in the layers. Epitaxial growth is coherent
growth of one metal on top of a dissimilar metal (see Section 3.2.1). Such coherent growth would
allow for the transmission of orientation information across the Al layer as we see in this image.
However, since the layer size is so small, there is probably some overlap of the Al layers with the
49

Nb layers, which causes the Nb to dominate the NED patterns and does not allow us to verify the
structure of the Al layers.
Cu/Nb samples were compared with the Al/Nb samples, because Cu/Nb is an immiscible system,
while Al/Nb is a miscible system. The STEM and ACOM images for the 20 nm Cu sample are
shown in Figure 39. The STEM image (Figure 39a) shows flat and sharp interfaces between Cu
(dark) and Nb (bright) layers, which are typical of the Cu/Nb multilayer system and make it ideal
for this ACOM study. The ACOM image in Figure 39b clearly shows the Cu layers, with the grain
orientations normal to the image once again being indicated by the colors in the triangle inset.
Some sample NED patterns and the areas from which they are taken are shown to the left of the
ACOM image.

Figure 39: Cross-sectional STEM image (a), ACOM image (b), and ACOM phase map (c) of a
Cu/Nb multilayer sample with 20 nm Cu layers. In (b) the grain orientations normal to the
image are indicated by the stereographic triangle inset. The NED patterns to the left indicate
distinct local orientations in the Cu and Nb layers. The boxed and circled areas highlight
crystalline and amorphous Cu regions, respectively. For (c), the colors indicate structure: red-
BCC Nb, green-FCC Cu, blue-amorphous and semi-amorphous, black-poor indexing. The
representative NED patterns on the right indicate crystalline, semi-amorphous, and amorphous
Cu regions. The growth direction is indicated on the left. [107]

Similar to the 20 nm Al layers, the 20 nm Cu layers were found to consist of both crystalline
(boxed area) and amorphous (circled area) regions. However, in this case, the crystalline Cu
regions were found to result in maintained orientation between Nb grains in different layers, while
50

the amorphous regions interrupted the Nb grain orientations, as shown by Nb grain color
differences below and above the amorphous Cu regions. Amorphous regions are generally not
observed in pure sputtered metals, although they are common in sputtered oxides and metal alloys
[108-110].
The presence of amorphous and crystalline regions in the 20 nm Cu sample is more clearly shown
in the phase map, Figure 39c. Red regions correspond to body-centered cubic (BCC) Nb, green to
face-centered cubic (FCC) Cu, and blue to semi-amorphous and amorphous regions. It is clear
that there are multiple amorphous regions in the Cu layers. The level of structure is evident in the
sample NED patterns on the right. The top NED pattern is from a crystalline Cu region, which has
a very distinct pattern. The middle NED pattern is from a semi-amorphous region as it has only a
few outer diffraction spots. The bottom NED pattern corresponds to a fully amorphous region,
with no outer diffraction spots.
As a comparison to the 20 nm Cu sample, the 2 nm Cu sample images are shown in Figure 40.
The STEM image, Figure 40a, shows the presence of the Cu layers, as pointed out by the arrows.
The layers are flatter than the 2 nm Al layers, but the 2 nm Cu layers are still not visible in the
ACOM image, Figure 40b. Apart from a few Cu crystallites, the ACOM image indicates
continuous Nb grains. This is also seen in the phase map, Figure 40c, where a few green Cu
crystallites are observed, but the vast majority of the sample corresponds to a BCC Nb structure.
51

Figure 40: Cross-sectional STEM image (a), ACOM image (b), and ACOM phase map (c) of a
Cu/Nb multilayer sample with 2 nm Cu layers. The arrows in (a) indicate the Cu layers. In (b),
the grain orientations normal to the image are indicated by the stereographic triangle inset.
For the ACOM phase map in (c), the colors indicate structure: red-BCC Nb, green-FCC Cu, black-
poor indexing. Note that amorphous and semi-amorphous regions were not observed. The
growth direction is indicated on the left. [107]

This is an indication of epitaxial growth, as it was for the 2 nm Al layers. Cu has been reported to
grow pseudomorphically in a BCC structure, rather than in its native FCC structure, on top of Nb,
and that may in fact be what is taking place in these samples [111-113]. However, though the Cu
layers are less rough than the Al layers, there is still some layer roughness, which results in overlap
with the Nb grains. Also, the 2 nm Cu layer thickness is approaching the resolution of the
technique. Therefore, we cannot state for certain the structure of the Cu layer.
Overall, from this microstructural study we were able to see that both the layer thickness and
crystallinity of the Al and Cu grains affected the continuity of the Nb grains. The Nb grain
continuity will greatly affect sample properties, because corrosion paths, diffusion paths, slip
system continuity, and other sample characteristics will greatly alter the sample response to outside
loading and treatment. This study also observed both crystalline and amorphous regions in a single
layer of an NMM, which is not typical. These findings produce a better understanding of the
different structures which are possible in NMMs, which will in turn guide the fabrication of future
NMM materials.
52

6.3 Mechanical Behavior - Nanoindentation
In addition to the above microstructural study, which focused only on the samples with 2 and 20
nm Cu and Al layers with fast sputtering rates, all of the Cu/Nb and Al/Nb samples in Table 3
were tested by nanoindentation. This was to determine whether there is a change in mechanical
behavior with decreasing layer thickness or sputtering rate. The plots of the hardness values of
the samples with decreasing Cu and Al layer thickness are shown in Figure 41.

Figure 41: Hardness plots, as measured by nanoindentation, for (a) Cu/Nb and (b) Al/Nb NMMs
as a function of Cu and Al layer thickness, with constant Nb layer thicknesses of 100 nm. The
blue squares and red diamonds correspond to samples that had slow and fast Cu and Al
sputtering rates, respectively (slow deposition rates – Cu: 0.15 nm/s, Al: 0.07 nm/s; fast
deposition rates – Cu: 0.69 nm/s, Al: 0.29 nm/s; Nb deposition rate - 0.12 nm/s). The error bars
indicate the standard deviation.

For both the Cu/Nb and Al/Nb systems, the peak hardness occurs at an intermediate value, with
the Cu/Nb sample hardness peaking at a Cu layer thickness of 5 nm, and the Al/Nb sample hardness
peaking at an Al layer thickness of 10 nm. This trend was present for both fast and slow sputtering
rates, even though the hardness values varied for the different sputtering rates. The hardness
decreased for both systems when the thicknesses of the Cu and Al layers decreased to 2 nm. This
can be partially explained by the ACOM images in Section 6.2, where the Nb grain orientations
were uninterrupted by the 2 nm Cu and Al layers. This would result in easy dislocation travel
between the Nb layers, and, thus, a softer material. However, it is not clear why the strength
53

peaked at 5 and 10 nm for the Cu/Nb and Al/Nb samples, respectively. Further investigation of
the mechanical behavior of these types of NMMs was therefore conducted by means of an in-situ
XRD tensile testing technique, in order to determine the behavior of Cu and Nb layers individually
within Cu/Nb NMMs, and this work is presented below in Section 7.

54

7 Load Sharing Phenomena in Nanoscale Cu/Nb Multilayers
7.1 Overview
The attractive mechanical properties exhibited by NMMs are generally attributed to the fact that,
as their layer thicknesses decrease, the individual layer behaviors change and the interface volume
in the material increases (see Section 3.2.4) [44, 63]. However, as the individual layer thicknesses
decrease below 20 nm, the deformation behavior becomes more difficult to observe. While in-situ
transmission electron microscopy (TEM) studies allow for direct imaging of the deformation and
have demonstrated several localized behaviors in NMMs, such as confined layer slip within
individual layers and crack deflection mechanisms, other in-situ techniques are needed to
characterize the global behavior of these materials [58, 59]. One such technique is in-situ X-ray
diffraction (XRD) during tensile testing, which allows for directly measuring the lattice spacings
of the individual metallic layers within the NMMs and probes large sample volumes (tens of square
μm) [114-116]. Several NMMs have been previously tested with this method, but those studies
generally concentrated on the technique itself or investigated relatively thick layers [115, 117-
122]. NMMs with layer thicknesses of approximately 20 nm have been previously studied,
although the influence of layer thickness was not investigated, while a study of W/Cu NMMs with
varying layer thicknesses focused on the elastic properties and interfacial mixing in the samples
[123-125]. In contrast, in this study we explored the deformation of ultrathin (2-20 nm thick)
layers as a function of the layer thickness, with the aim of isolating the mechanical contribution of
each material in the NMM system [126].

55

7.2 In-Situ X-ray Diffraction
In order to perform the in-situ XRD, we collaborated with researchers at the Karlsruhe Institute of
Technology (KIT), using a setup similar to the one described by Lohmiller et al. [114, 127]. To
perform the measurements, a sample was first placed in the path of an X-ray beam, as shown in
Figure 42. The sample was then loaded perpendicularly to the X-ray beam. The interaction of the
beam with the sample caused a series of diffraction rings to be formed on a CCD camera. The
camera records full Debye-Scherrer rings from diffracting planes which are highly inclined to the
sample normal, due to the high energy of the X-rays. Owing to these large scattering vectors, the
analysis of the radial distance variation of different sections of the Debye-Scherrer rings allows
for determination of lattice strains in various orientations, such as the loading and transverse
directions. Since the rings correspond to different diffraction planes within the materials, the
stresses in the nanometer-thick Cu and Nb layers could be tracked separately [123].
Simultaneously with the XRD lattice strain measurements, the overall sample strains were
measured by optical methods [127, 128].

Figure 42: Schematic of in-situ synchrotron measurements during tensile testing. As the
rectangular sample on the left is being loaded, the X-ray beam travels through the sample, and
the diffraction cones from the X-ray beam appear as diffraction rings on the camera. These
rings correspond to different lattice planes in the materials.

56

7.3 Sample Fabrication and Architecture
The samples for this study were Cu/Nb NMMs deposited onto 50 μm thick Kapton E polyimide
substrates (DuPont). Polyimide substrates were used because they are relatively X-ray transparent
and stable at 200 °C (important for sputtering). The substrates were cleaned in the manner
described by Lohmiller et al. and then heated on a hot plate for 15 minutes at 200 °C immediately
prior to being introduced into the sputtering chamber [129]. The base chamber pressure was
between 1.9 and 2.3 E-4 Pa for the different samples. DC sputtering was performed at 100 W for
the Cu layers and 200 W for the Nb layers, with an Ar pressure of 0.4 Pa, giving sputtering rates
of 0.29 nm/s and 0.35 nm/s for the Cu and Nb, respectively. These sputtering parameters were
chosen to minimize the residual stresses, which can result in curling of the samples.
After sputtering, the samples showed strong (111) Cu and (110) Nb out-of-plane textures, with
random in-plane textures (not shown), which is typical for sputtered Cu/Nb NMMs [84, 130].
Rectangular strips with dimensions of 6 mm x 35 mm were deformed in tension ( ) with
simultaneous in-situ transmission XRD data collection (every 45 s, except Cu2/Nb10: 75 s) at the
PDIFF beamline of the Angströmquelle Karlsruhe (ANKA), Germany. A picture of one of the
tested specimens is shown in Figure 43. The speckle patterns on the right and left sides of the
specimen are for the optical tracking of the sample to determine overall sample strains. TEM
specimens of the as-sputtered samples were prepared by focused ion beam (FIB) liftout using a
JIB-4500 dual beam FIB (JEOL) and TEM imaging was performed on a JEM-2100F TEM (JEOL).
1 5
10
 
 s  
57

Figure 43: Picture of a tensile tested Cu/Nb NMM sample on a polyimide substrate. The
speckle patterns on the right and left sides are for optical tracking of the overall sample strain.

A schematic of the sample architecture is given in Figure 44. The NMMs were sputtered directly
onto the polyimide substrates, with Nb being both the initial and capping layer. Sputtering was
performed with the goal of obtaining samples which had constant Nb layer thicknesses (10 nm)
and varying Cu layer thicknesses (2-20 nm), while keeping a total Cu thickness of 60 nm. Full
sample descriptions, including layer thickness, number of layers, and total Cu/Nb thickness, are
given in Table 4 for the different samples. The naming refers to the thicknesses of the Cu and Nb
layers, e.g. the sample with three 20 nm Cu layers and four 10 nm Nb layers is called Cu20/Nb10.
An additional sample with a total Cu thickness of 40 nm, Cu5/Nb10 (b), was fabricated in order to
determine whether the deformation is being controlled by the layer thicknesses or the total Cu/Nb
thicknesses. The results presented here will focus on the first four samples in Table 4.

Figure 44: Test specimen schematic (side view). Samples consisting of alternating Nb and Cu
layers were sputtered onto polyimide substrates. Test specimens were then cut and tensile
tested parallel to the layer direction. (image not to scale) [126]

58

Table 4: Tensile test specimen descriptions*

Representative TEM images are given in Figure 45 in order of decreasing Cu layer thickness, from
Cu20/Nb10 in Figure 45a to Cu2/Nb10 in Figure 45d [126]. Note that the polyimide substrates
are not atomically flat, and this waviness caused some overlap of the different layers in cross-
sectional imaging. In spite of this overlap, the layered structure is evident for all samples, and, in
a previous study, we demonstrated layer thickness control down to 2 nm in the sputtering of NMMs
[107].
59

Figure 45: Bright-field cross-sectional TEM images of as-sputtered Cu/Nb samples on polyimide
substrates. The nominal thicknesses are 10 nm for all Nb layers, and (a) 20 nm, (b) 10 nm, (c) 5
nm, and (d) 2 nm for the Cu layers. Darker layers are Nb and brighter layers are Cu. [126]

7.4 Deformation Behavior
XRD data collected in-situ during the tensile tests are shown in Figure 46 [126]. Figure 46a shows
the measured lattice strains in the (111) Cu planes (measured by XRD) as a function of the overall
sample strains (measured by optical methods) [114]. The lattice strains shown are in the tensile
direction, and they correspond to the stresses in the Cu layers. Thus, these curves mimic stress-
strain plots for just the Cu layers within the NMMs. It must be noted that the measurements with
the current transmission geometry setup only yield relative lattice strain changes. Therefore, the
lattice strain values do not correspond directly to absolute strength values for the layers.
60

Figure 46: Stress-strain data collected during in-situ XRD tensile testing of Cu/Nb multilayer
samples on polyimide substrates. Sample names state the thicknesses of the different layers in
nm. The stress-strain curves for the (a) Cu and (b) Nb layers within the samples are plotted.
Three points of failure are indicated by arrows for the sample with 2 nm Cu layers in (a) and (b).
Sample failure strains are plotted as a function of the Cu layer thickness in (c). Maximum lattice
strains, which correspond to the strengths, are plotted for the Cu and Nb layers as a function of
the Cu layer thickness in (d). [126]

Figure 46b shows the Nb (110) lattice strains with increasing sample strain. As in Figure 46a,
these are also similar to stress-strain plots, but Figure 46b isolates the Nb behavior. In Figure 46a
and Figure 46b, failure was defined as the sample strain at which the first lattice strain drop
occurred. Note that, for each sample, failure clearly occurred at the same strain for both the Cu
and Nb layers, except for Cu20/Nb10. In that sample, the Cu lattice strain showed a gradual
decrease after around 0.054 sample strain, which is prior to the first clear Cu load drop at 0.059
sample strain. On the other hand, the Nb curve does not have a very clear load drop. Therefore,
the sample strain at the clear Cu load drop was taken as the sample failure strain. For all samples,
61

even though the Cu and Nb layers experienced failure, the polyimide substrates prevented sample
rupture.
In Figure 46a, the Cu curves for the samples with different Cu layer thicknesses deviate from each
other well before failure, with only the 2 nm Cu layers showing a lack of plasticity prior to failure.
On the other hand, in Figure 46b, the Nb curves all initially overlap, which is expected for samples
having the same nominal Nb layer thicknesses. Since, prior to failure, the Cu curves deviate from
each other and the Nb curves do not, this indicates that, while the Cu and Nb layers are rigidly
bonded to each other, the deformation mechanisms of the Nb layers appear to be independent of
those in the Cu layers. The initial isolation of the deformation of the Cu and Nb layers appears to
match the confined layer slip model of dislocation motion which was explained by Embury and
Hirth for multilayers and has been recently demonstrated by modeling and in-situ TEM [58, 131-
133]. In the confined layer slip model, the dislocation motion is limited to individual layers, since
the interfaces are impenetrable to dislocations, giving independent deformation of the different
materials. Earlier studies of Cu/Nb composites provided evidence of impenetrable Cu/Nb
interfaces at length scales > 20 nm, which seems to correspond to the observed deformation in the
current study [58, 120]. However, after the initial deformation, the varying thicknesses of the Cu
layers begin to affect the deformation, resulting in different failure strains for the different samples.
The sample failure strains are plotted in Figure 46c. As expected, the samples with thicker Cu
layers showed higher failure strains. There appears to be a shielding mechanism operating, as
described in other studies [134-136]. In this mechanism, defects are thought to be already present
in the brittle Nb layers or are generated there by dislocation motion. However, the ability of a
defect to propagate through the sample is determined by whether the Cu layers are thick enough
to shield its propagation. Thus, the samples with thicker Cu layers should allow the Nb layers to
62

reach higher stresses prior to failure, which was observed in this study. Such a mechanism would
also result in failure of all Cu and Nb layers once cracks are able to cross between Nb layers. This
is in fact indicated in Figure 46a and Figure 46b, where the lattice strain drops (effectively stress
drops) occur at the same sample strains for the Cu and Nb layers. For example, in the Cu2/Nb10
sample, both the Cu and Nb layers experience lattice strain drops at sample strains of
approximately 0.02, 0.04, and 0.065, as pointed out by arrows in Figure 46a and Figure 46b.
The lattice strain at the first lattice strain drop was defined as the maximum lattice strain. This is
plotted in Figure 46d and corresponds to the strength of the layers. There is an increase in
maximum Cu lattice strain of more than 50 percent from Cu20/Nb10 to Cu2/Nb10. This increasing
strength is in the layer thickness regime where a strength plateau or even reverse Hall-Petch effect
has been reported [63, 70]. Others have observed continued strengthening with layer thicknesses
in the range of 20 nm to 2 nm, but not as high as the 50 percent increase measured here [63, 130].
The maximum lattice strains of the Nb layers, on the other hand, decreased by about 40 percent
with decreasing Cu layer thickness, Figure 46d. The mechanism appears to be load-sharing
behavior which is assumed for such systems, with the Nb layers taking a smaller proportionate
amount of the load as the Cu layers get stronger with decreasing thickness. A previous study using
a single set of layer thicknesses (26.5 nm Cu/28.5 nm Nb) investigated the load-sharing among
different Cu and Nb crystal orientations and found that the stiffer Nb (110) and (211) orientations
took more of the load from the Cu (220) and Nb (200) orientations with increasing stress [123].
The present study extends beyond this by demonstrating that the increasingly stronger Cu layers
take more of the load from the Nb layers with decreasing Cu layer thickness. The load sharing
may explain the unusually large increases in strength that we observed for the Cu layers. Even
though the Cu layers were becoming much stronger with decreasing thickness, the samples were
63

failing at lower Nb stresses. Therefore, the net behavior of such NMMs as measured by other
techniques would be either a modest increase or even a decrease in the overall sample strength
with decreasing Cu layer thickness. However, by using this in-situ XRD technique, a large strength
increase in the ultrathin Cu layers could be observed.
To verify that the deformation is being controlled by the layer thickness, not the total Cu/Nb
thickness, an additional test was performed using the Cu5/Nb10 (b) sample (see Table 4). When
this sample was tested, the Cu and Nb deformation curves followed those of the Cu5/Nb10 sample,
Figure 46, even though the two samples varied in total Cu/Nb thickness. Therefore, the differences
in deformation observed in this study are in fact due to changes in the Cu layer thickness.
It should be noted that the lattice strain for all of the Cu layers had a general downward trend after
the first lattice strain drop, Figure 46a. The opposite trend is observed for the Nb layers, Figure
46b; this is most clearly seen for the Cu2/Nb10 curve, which ends with a Nb lattice strain
substantially higher than that seen at the first lattice strain drop. These opposing behaviors indicate
that, at the first lattice strain drop, true failure is occurring in Cu layers, since they are permanently
weakened, while the Nb layers can continue to strengthen with increasing strain. This result is
supported by the plots in Figure 47, which show the Cu and Nb XRD peak breadths as a function
of the sample strain. The main sources of larger peak breadths are smaller grains, microstrain, and
accumulated plastic deformation [137]. For example, the Cu2/Nb10 sample should have the
smallest Cu grains (generally on the order of the layer thickness for sputtered NMMs), and
therefore has a larger Cu peak breadth than the other samples, Figure 47a. During deformation,
the Cu peak breadth curves reached plateaus at strains that were close to the failure strains, as
pointed out by arrows for Cu2/Nb10, Cu5/Nb10, and Cu10/Nb10. This indicates no net increase
in microstrain or number of dislocations in the Cu during further straining. Additionally, the sharp
64

increases in Cu peak breadths which occurred just prior to the points of failure have been
previously reported to correspond to macroscopic cracking [138]. Contrary to the Cu, the Nb peak
breadth curves in Figure 47b continued to increase with sample strain, indicating ongoing
accumulation of microstrain and/or dislocations within the Nb layers. Also, note that the higher
the lattice strain in the Nb layers (increasing Cu layer thickness), the stronger is the increase in
peak breadth. This is in contrast to an earlier Cu/Nb study, which did not observe significant XRD
peak broadening for 20 nm thick Nb layers after deformation, though the loading was compressive
in that case and the measurements were performed ex-situ [124]. The consistently increasing Nb
peak breadth may not be made apparent in freestanding NMMs, since the failure of the Cu layers
would result in failure of the entire sample. However, the support of the samples by the polyimide
substrates prevents catastrophic failure and strain localization, and thus allows this deformation
behavior to be made apparent.

Figure 47: XRD peak breadth data collected during in-situ XRD tensile testing of Cu/Nb
multilayer samples on polyimide substrates. Sample names state the thicknesses of the
different layers in nm. Peak breadths corresponding to the (a) Cu and (b) Nb layers are plotted.
The three arrows in (a) indicate sample failure strains. [126]

SEM images of the NMM sample surfaces after tensile testing are given in Figure 48. The images
are arranged in order of descending Cu layer thickness. The fracture changes noticeably, with the
samples having thin Cu layers, Cu5/Nb10 (Figure 48c) and Cu2/Nb10 (Figure 48d), showing very
65

brittle fracture. The cracks in these samples were observed to travel across the entire 6 mm width
of the tested samples. On the other hand, the cracks in the Cu20/Nb10 sample often stopped and
started in the middle of the sample, as pointed out by arrows in several places in Figure 48a. The
relative ductility of the Cu20/Nb10 sample is also made evident by the representative crack that is
inset in Figure 48a and follows a more tortuous path than that seen in the insets for the other
samples. These behaviors match previous testing of NMMs attached to polyimide [56, 117].

Figure 48: SEM images of the post-mortem surfaces of tensile-tested Cu/Nb multilayer samples
on polyimide substrates. The Nb layers are 10 nm thick for all samples, and the Cu layers are (a)
20 nm, (b) 10 nm, (c) 5 nm, and (d) 2 nm thick. The samples with thicker Cu layers, (a) and (b),
have cracks which start and stop within the samples, as pointed out by arrows in (a). The cracks
in the samples with thinner Cu layers, (c) and (d), spanned the entire sample width. The insets
show that the samples with thicker Cu layers showed more tortuous crack paths. The black
scale bars in the insets are one μm long.

66

7.5 Conclusions
In this study, the deformation behaviors of ultrathin metal layers within NMMs were characterized
by using an in-situ XRD tensile testing technique. The results matched the confined layer slip
model in that they indicated independent deformation mechanisms in the Cu and Nb layers prior
to failure. However, contrary to some previous studies, the Cu strength continued increasing
significantly with decreasing thickness, down to 2 nm Cu layers. This may be due to the fact that
the load-bearing contributions of the different materials were isolated. Specifically, while it was
observed that, with decreasing Cu layer thickness, the Cu layers were showing higher lattice strains
at failure, the Nb layers were actually showing lower lattice strains at failure. Without separating
the behavior of the two materials, the NMMs may have been shown to weaken with decreasing Cu
layer thickness, but that would have been an incomplete description of the NMM behavior.
Quantifying the load-sharing behavior of different NMM systems in future studies will make
clearer the behavior of individual materials and will facilitate choices with regard to layer thickness
and material selection for strength optimization.
7.6 Future Research Directions
Based on this study, several future research directions have been identified. One avenue of
research is testing different material systems, most notably those containing different crystal
structures. Several additional systems (Cu/Zr and Cu/CuZr) have been selected for study, and a
summary of the samples which have been fabricated is given in Appendix A.3.
In addition, the effects of residual stresses can also be investigated, which will allow for calculation
of absolute stress values. These calculations are complicated by the fact that the stress state is not
purely uniaxial, due to the fact that the polyimide substrate has a Poisson’s ratio which is different
67

from that of the NMM film. In addition, the diffraction rings developed a texture in some cases,
further complicating the stress state, as presented below in Figure 49, where the Cu diffraction
rings from three different samples Cu20/Nb10 (blue, middle), Cu5/Nb10 (red, top), and Cu2/Nb10
(green, bottom) show different intensities at different angles. Further investigation of the
developed textures may reveal the types of deformation mechanisms that are taking place and the
slip systems which are being preferentially activated [139-143].

Figure 49: Diffraction ring intensity at different angles for three NMM samples: Cu20/Nb10
(blue, middle), Cu5/Nb10 (red, top), and Cu2/Nb10 (green, bottom).

Finally, there have been reports of variations in elastic modulus for NMMs with different layer
thicknesses, and there may be different slopes for the elastic regimes of the Cu curves in Figure 46
[144, 145]. This could be further investigated by also checking for a variation in elastic modulus
for the Nb or Zr layers with varying thicknesses in the Cu/Nb and Cu/Zr NMMs.

68

8 Hf-Ti Thermal Stability
The thermal stability of nanostructured materials is generally poor (see Section 3.1.2). However,
several models have been put forth for achieving thermodynamically stabilized nanostructures,
and one of these models was investigated by using NMM configurations.
8.1 Selection of Binary Systems
As explained in Section 3.1.3, the thermodynamic nanostructure stability model by Chookajorn et
al. has been used to classify a variety of binary systems according to predicted bulk stability or
nanostructure stability [31-34]. Several of these systems were selected for further investigation,
and they are given in Table 5. The first two columns identify the major and minor elements, while
the last two columns state whether the system is expected to be nanostructure or bulk stable and
whether the elements are soluble at 800 °C (the heat treatment temperature which was selected for
this study). Of note, Hf-Ti is expected to be nanostructure stable when Hf is the major element
and bulk stable when Ti is the major element, indicating that solubility is not the determining factor
for this model with regard to nanostructure stability.
Table 5: Selected systems for NMM thermal stability study
Material 1
(major element)
Material 2
(minor element)
Nanostructure or
bulk stable
Solubility (at 800 °C)
Hf Ti Nanostructure Fully soluble
Mo Au Nanostructure Insoluble
Ta Cu Nanostructure Insoluble
Ta Hf Nanostructure Insoluble
Ti Hf Bulk Fully soluble

Preliminary investigations were conducted on the Mo-Au, Ta-Cu, Ta-Hf, and Ti-Hf systems, and
these are explained in Appendix H. However, the Hf-Ti system showed the most interesting
behavior, and this section will focus on the behavior of a variety of Hf-Ti configurations.
69

While the thermal stabilities of ternary systems containing Hf and Ti have been investigated for
shape memory alloys and metallic glasses, an investigation into the nanograin stability of the
binary Hf-Ti system at high temperatures is lacking [146-148]. To control the initial nanostructure,
the Hf-Ti samples were sputtered as nanometallic multilayers (NMMs), with alternating nanometer
thick layers of pure Ti and co-sputtered Hf-Ti. The use of sputtering avoids some of the
microstructural defects and chemical contaminants that can be introduced by some of the other
nanostructure fabrication methods, such as ball-milling [149, 150]. The use of this multilayered
geometry also allows for control of the grain size (which correlates to the layer thickness) and
provides the ability to distribute the Ti in nanoscale proximity to the Hf-rich regions, which should
lower the kinetic barrier between the initial and expected equilibrium microstructures.
8.2 Methods
8.2.1 Experimental
Samples were synthesized using DC magnetron sputtering, with two sputtering sources being used
to deposit Hf and Ti onto (100) Si substrates. The sample descriptions are shown in Table 6, where
Samples A and B consist of alternating pure Ti and co-sputtered Hf-Ti layers. The following
sputtering conditions were used: 60 W for Ti, 200 W for Hf, and a 0.67 Pa Argon pressure, with
the on-times for the Hf source being varied to produce samples with differing layer thicknesses.
Table 6: Multilayered Hf-Ti sample descriptions
Sample name
Hf-Ti layer
thickness (nm)
Ti layer
thickness (nm)
Overall sample
thickness ( μm)
Composition
(at.% Ti)
Sample A 15 2 1.6 23
Sample B 40 5 2.1 24

70

An XP-2 profilometer (AMBiOS) was used to measure as-sputtered sample thicknesses, after
which the samples were removed from the Si substrates. The samples were then subjected to an
800 °C heat treatment for 96 hours under vacuum (pressure ≈ 4 x 10
-4
Pa) in a GSL1100X tube
furnace (MTI Corporation). The heat treated samples will be referred to as HT-Sample A and HT-
Sample B. At the end of the heat treatment, quenching was performed in low vapor pressure oil
(Invoil 705, Inland Vacuum Industries) without breaking the vacuum. The samples were then
removed from the oil and cleaned with ethanol. Samples for transmission electron microscopy
(TEM) were prepared by Focused Ion Beam (FIB) liftout using a JIB-4500 FIB (JEOL). TEM,
scanning TEM (STEM), and Energy Dispersive X-ray spectroscopy (EDX) were performed in a
JEM-2100F transmission electron microscope (JEOL). A JSM-7001 scanning electron
microscope (SEM) was used to determine global compositions by EDX (JEOL).
8.2.2 Modeling
In collaboration with the Schuh group at MIT, a lattice Monte Carlo method was used to simulate
the equilibrium structure of a binary alloy at finite composition and temperature by considering
both chemical and topological arrangements. Each lattice site maintains two state variables: the
chemical type, which can be assigned as a solvent or solute atom, and the grain number, which
denotes the group of atoms that belong to the same grain. The model accounts for mixing
interactions inside granular and intergranular regions using two different interaction energies, and
the details on bond energy calculations can be found in Ref. [34]. The structure evolves by changes
in the chemical distribution and grain structure. A chemical switch selects two lattice sites at
random and exchanges their atom types. A grain switch allows a lattice site to change its grain
allegiance to one of its neighbor’s across a grain boundary or nucleate a new grain with a distinct
grain number. A switch that evolves the system to a lower energy state is always executed, while
71

a switch that evolves the system to a higher internal energy E2 relative to the initial internal energy
E1 is accepted at a probability 𝑒 −(𝐸 2
−𝐸 1
)
𝑘𝑇
, where k is the Boltzmann constant and T is the absolute
temperature.
The structure is initialized in a randomized state, with regard to both chemical and grain number
distributions, at 10,000 K and slowly cooled at a rate
–(𝑇 step
−𝑇 final
)
1000
until the final temperature Tfinal
= 1073 K is reached. During each Monte Carlo step, one switch per atom is attempted on average
across the whole system at the intermediate temperature Tstep. A total of 100,000 Monte Carlo steps
are performed, which is found to achieve an equilibrated state (i.e., slower cooling produces the
same essential state) for the present conditions. Periodic boundary conditions were imposed on the
three principal axes, and all simulations employ a hexagonal close packed lattice with 500 x 500
x 6 sites, which is equivalent to 160 nm in width and encompasses a total of 1,500,000 atoms.
For the simulation of the Hf-Ti system studied here, the bond energies used are 𝐸 c
HfHf
= 1.61 eV,
𝐸 c
TiTi
= 1.25 eV, 𝐸 c
HfTi
= 1.41 eV, 𝐸 gb
HfHf
= 1.58 eV, 𝐸 gb
TiTi
= 1.22 eV, and 𝐸 gb
HfTi
= 1.42 eV, where
the subscript indicates a bond in a crystal (c) or grain boundary (gb) region, and the superscript
denotes the chemical pair. Other relevant material parameters are the atomic volume Ω
Hf
= 13.44
cm
3
/mol, Ω
Ti
= 10.64 cm
3
/mol, hexagonal close-packed (HCP) coordination number of 12, and
grain boundary thickness of 0.5 nm. These bond energies yield reasonable values for the pure
component grain boundary energies γ
0,Hf
= 0.72 J/m
2
, γ
0,Ti
= 0.68 J/m
2
.

72

8.3 Results and Discussion
8.3.1 Microstructure and Grain Growth
The microstructures of the as-sputtered samples are shown in the cross-sectional TEM images in
Figure 50. Figure 50a and Figure 50c are bright-field TEM (BF-TEM) images of Sample A and
B, respectively. The bright layers are Ti, the dark layers are Hf-Ti, and the growth direction is
vertical, as indicated by the arrow. EDX scans by SEM showed overall compositions of
approximately 23 at.% Ti for Sample A and 24 at.% Ti for Sample B. Figure 50b and Figure 50d
are dark-field TEM (DF-TEM) images of Sample A and B, respectively. There is a columnar
structure in the as-sputtered samples, and a preferred Hf (002) orientation, as indicated by the inset
diffraction patterns. The average grain diameters are larger for Sample B (40 nm) than for Sample
A (25 nm), and their grain size distributions will be discussed later.

Figure 50: Cross-sectional TEM images of as-sputtered nanometallic multilayered composite
samples. Sample A consists of 15 nm Hf-Ti/2 nm Ti layers, see (a) and (b), and Sample B
consists of 40 nm Hf-Ti/5 nm Ti layers, see (c) and (d). In the bright field TEM images, (a) and
(c), the Ti layers are bright and the Hf-Ti layers are dark. The dark-field TEM images, (b) and (d),
demonstrate the columnar structures of the samples, with their respective SAED patterns
showing strong Hf (002) texture. The growth direction is indicated on the left, and the
diameters of several grains are marked with yellow arrows in (d).
73

In order to investigate the thermal stability of both structures, the samples were heat-treated as
explained in Section 2.1, and cross-sectional TEM images of the heat-treated samples are shown
in Figure 51 in three different TEM imaging modes, with HT-Sample A on the left and HT-Sample
B on the right. The BF-TEM images, Figure 51a and Figure 51d, show the nanocrystallinity of
the samples, as well as loss of the layered structure (the growth direction is indicated by the arrow
on the left). The inset diffraction patterns indicate that, though the samples were heat treated in
vacuum, there is a mixture of Hf and HfO2. The DF-TEM images, Figure 51b and Figure 51e,
show that the grains are relatively equiaxed and that the initial columnar structure has been lost.
Finally, the annular dark-field (ADF) STEM images, Figure 51c and Figure 51f, show a strong
segregation into Ti-rich regions (dark), and Hf-rich regions (bright), which was confirmed by EDX
mapping (shown later in this section). Note that the structures of HT-Sample A and HT-Sample
B are similar in all of the images.
74

Figure 51: Cross-sectional TEM images of heat-treated (800 °C, 96 hours) nanometallic
multilayered composite samples. The first sample is HT-Sample A, which originally consisted of
15 nm Hf-Ti/2 nm Ti layers, see (a), (b), and (c). The second sample, HT-Sample B, originally
consisted of 40 nm Hf-Ti/5 nm Ti layers, see (d), (e), and (f). The bright-field TEM images in (a)
and (d) show the nanocrystallinity of the samples, with inset SAED patterns (scale bar = 5/nm)
and the growth direction indicated on the left. Dark-field TEM images, (b) and (e), show
individual grains. The annular dark-field STEM images, (c) and (f), display the Hf and Ti
segregation, with Hf-rich regions appearing bright and Ti-rich regions appearing dark.

Since the columnar structure as well as the chemical layering was lost after the heat treatments, an
additional sample was co-sputtered (referred to as Sample C) in order to investigate whether the
Ti interlayers were responsible for the formation of equiaxed grains. Sample C is a monolithic
Hf-Ti film, with a slightly lower Ti concentration (20 at.% Ti) than the multilayered samples.
75

Sample C initially had columnar grains (Figure 52a, DF-TEM cross-sectional image), but, after a
96 hour, 800 °C heat treatment and quenching, the columnar structure was lost, and equiaxed
grains resulted. This is shown in Figure 52b and Figure 52d, which are cross-sectional ADF-
STEM and DF-TEM images, respectively. In Figure 52b, Hf-rich regions are dark and Ti-rich
regions are bright; this reversal in contrast compared to Figure 51c and Figure 51f is due to the
differences in thicknesses of the FIB-prepared samples. Based on this heat-treated monolithic
sample, Sample C, the thin Ti interlayers do not appear to be responsible for the loss of the
columnar grains in the multilayered samples, and it would appear that this system has an intrinsic
preference to evolve to an equiaxed grain structure with Ti situated around Hf-rich grains,
independently of the initial structure. This evolution of the structure to the equiaxed configuration
will be studied in future work; this study focused on the equiaxed structure itself, and its energetics.
76

Figure 52: Cross-sectional dark-field TEM image (a) of as-sputtered Sample C (co-sputtered Hf-Ti
sample, 20 at.% Ti). The growth direction is indicated on the left. Annular dark-field STEM (b)
and dark-field TEM (d) images are shown of the heat-treated sample (800 °C, 96 hours), HT-
Sample C. Dark areas are Hf-rich and bright areas are Ti-rich in (b). The grain sizes before and
after heat treatment are given in (c), with average grain sizes of 57 nm and 196 nm before and
after heat treatment, respectively.

DF-TEM images were used to measure the grain sizes of the samples before and after heat
treatment. By assuming a cylindrical grain geometry, the grain sizes of the as-sputtered samples
were measured as diameters at several points on each columnar grain, as indicated by the yellow
double-sided arrows on Figure 50d. For the heat-treated samples, the areas of individual grains
were determined from DF-TEM images, such as in Figure 51b and Figure 51e, and equivalent
diameters were calculated assuming spherical grains. The grain size distributions for the samples
are given in Figure 53. For Sample A, the average columnar grain diameter was initially 25 nm,
which evolved to roughly equiaxed grains of 50 nm diameter after heat treatment, while for Sample
77

B the average grain size grew from 41 nm to 52 nm. Overall, HT-Sample A and HT-Sample B
showed similar average grain sizes and grain size distributions.

Figure 53: Grain size distributions for as-sputtered and heat-treated (800 °C, 96 hours)
nanometallic multilayered composite samples, Sample A (15 nm Hf-Ti/2 nm Ti) and Sample B
(40 nm Hf-Ti/5 nm Ti). The average grain sizes were as follows: (a) Sample A: 25 nm and 50 nm
before and after heat treatment, respectively, and (b) Sample B: 40 nm and 52 nm before and
after heat treatment, respectively.

Based on the observation that the average grain sizes grew to only ~50 nm after 96 hours at 800
°C, it appears that grain growth is inhibited in the multilayered samples. In contrast, for Sample
C, the average grain sizes grew from 57 nm to 196 nm, Figure 52c, further suggesting that the
distribution of Ti as interlayers in Samples A and B indeed limits grain growth. In addition, since
both the multilayered and monolithic samples experienced some oxidation during heat treatment,
it is not likely that oxidation is solely responsible for the limited grain growth in the multilayered
samples. In order to investigate the role of kinetics in the grain growth during heat treatment of
these samples, diffusion calculations are presented in the next section.
8.3.2 Kinetics
An estimation of the achievable diffusional distances during heat treatment in Samples A and B
was performed by assuming a 3-D polycrystalline grain geometry. Although the samples started
78

with columnar microstructures, they ended with mostly equiaxed grains, so this assumption should
be sufficient for order-of-magnitude calculations. Since nanocrystalline materials have a
substantial volume of grain boundaries and triple junctions, in order to determine the overall
sample diffusivity, the diffusion through these areas must be taken into account. For this purpose,
the following equation was used to estimate the self-diffusivity of Hf (the majority element) [151]:
𝐷 eff
= 𝐷 b
+ 𝑔 (𝑑 ) [
2𝐻 gb
𝛿 𝑑 (𝐷 gb
− 𝐷 b
) +
𝐻 tj
𝛿 2
𝑑 2
(𝐷 tj
− 𝐷 b
)] 𝐸𝑞 . 3
where Deff is the effective diffusivity of a polycrystalline structure, Db is the bulk diffusivity, Dgb
is the grain boundary diffusivity, Dtj is the triple junction diffusivity, δ is the grain boundary
thickness (taken to be 0.5 nm), and g(d) is a geometrical parameter which can be determined by
the following equation:
𝑔 (𝑑 ) =
2𝐻 gb
𝑑 3
+ 2𝐻 tj
𝛿 𝑑 2
6𝐻 gb
𝑑 3
− (2𝐻 gb
2
− 3𝐻 tj
)𝛿 𝑑 2
− 3𝐻 gb
𝐻 tj
𝛿 2
𝑑 − 𝐻 tj
2
𝛿 3
)
𝐸𝑞 . 4
where d is the grain size and Hgb = 2.9105 and Htj = 2.5259 are structure constants characteristic
of a log-normal Voronoi polyhedral structure, corresponding to the grain boundaries and triple
junctions, respectively [151].
By assuming ideal solution behavior, the Hf self-diffusivity can be combined with the Ti self-
diffusivity, using Darken’s equation [152, 153]:
𝐷 𝑖 nt
= 𝑋 Hf
𝐷 Ti
+ 𝑋 Ti
𝐷 Hf
𝐸𝑞 . 5
where 𝐷 int
is the overall sample interdiffusivity and 𝑋 Hf
and 𝑋 Ti
are the mole fractions of Hf and
Ti. Table 7 shows the different values that were used and their sources. The bulk diffusivity
values 𝐷 b
= 𝐷 0
𝑒 −
𝑄 𝑅𝑇
were calculated at T = 800 °C, where 𝐷 𝑜 is the pre-exponential rate constant,
79

Q is the activation energy, and R is the ideal gas constant [154-157]. The grain boundary
diffusivity was calculated based on the grain boundary diffusion parameter in the reference given
and a grain boundary width of 0.5 nm, while the triple junction diffusivity was assumed to be
approximately 100 times as large as the grain boundary diffusivity, which is on the lower end of
what has been observed [151, 158-161].
Table 7: Diffusion coefficients and calculated diffusion distances
Species Path D0 (m
2
/s) Q (kJ/mol) D (m
2
/s) L (nm) Ref.
Hf bulk 5.40 x 10
-6
323 1 x 10
-21
20 [154]
Hf grain boundary 3 x 10
-14
1 x 10
5
[158]
Hf triple junction 3 x 10
-12
1 x 10
6
[151]
Hf in Ti bulk 4.00 x 10
-6
254 2 x 10
-18
8 x 10
2
[155]
Ti bulk 1.35 x 10
-3
303 2 x 10
-18
9 x 10
2
[156]
Ti in Hf bulk 2.70 x 10
-5
322 6 x 10
-21
40 [157]

Using these diffusivity values, the overall sample interdiffusivity was plotted for a temperature of
800 °C as a function of the grain size in Figure 54. Also included in the plot are the individual
diffusivity contributions from the Hf bulk, Hf grain boundaries, and Hf triple junctions. The
diffusivity can be used to approximate the chemical diffusion distance, L, according to the
following equation:
𝐿 ≈ (𝐷𝑡 )
1 2 ⁄
𝐸𝑞 . 6
where t is the time, taken here as 96 hours to match the experiments. The chemical diffusion
distances are shown on the right y-axis of Figure 54, and they are also given in Table 7. At a grain
size of 50 nm (the approximate post-heat-treatment grain size of Sample A and B), the
interdiffusivity is 8 x 10
16
m
2
/s and the chemical diffusion distance is 16 μm. This shows that the
diffusion length scales that are achievable after 96 hours at 800 °C are sufficiently large to allow
for significant grain growth and chemical redistribution. The significant restructuring of the
80

samples confirms this expectation, and what is more, the final structure of these samples is much
finer than could have, in principle, been accessed kinetically.

Figure 54: Calculated Hf diffusivities and 96-hour diffusion distances at 800 °C for triple
junctions, grain boundaries, and bulk, as well as overall Hf-Ti interdiffusion, as a function of
grain size.

8.3.3 Ti Segregation
HT-Sample A and B showed strong segregation of Hf-rich and Ti-rich regions (Figure 51c and
Figure 51f). EDX imaging was performed in a TEM to investigate this phenomenon for both
samples, and since they showed very similar structures after heat treatment, results are presented
only for HT-Sample B in Figure 55. While oxygen was detected, and the inset diffraction patterns
in Figure 51 indicate that there is a mix of Hf and HfO2 in the samples, EDX is not well-suited for
quantitative measurement of oxygen, and the results are presented on a metals-only basis. The
bright regions in the ADF-STEM image (Figure 55c) correspond to Hf-rich regions (with Hf
mapping in Figure 55a), while the dark regions correspond to Ti-rich regions (with Ti mapping in
Figure 55b). A Hf-Ti composition profile obtained by a line scan is shown in Figure 55d along
the direction indicated in Figure 55c by the line A-B. There is large change in the Ti content, with
a maximum of ~75 at.% in some dark regions and a drop to nearly 0 at.% in the middle of the
81

bright Hf-rich regions. Additionally, these composition values were obtained by scanning the
sample in projection, so the measured degree of separation is likely to be an underestimate.

Figure 55: Cross-sectional Hf (a) and Ti (b) compositional maps recorded by EDX in a TEM for
HT-Sample B. The sample originally consisted of alternating 40 nm Hf-Ti/5 nm Ti layers and was
heat treated at 800 °C for 96 hours. An annular dark-field STEM image is shown from the same
area in (c). A linescan compositional profile (d) is shown for Hf and Ti, along the A-B line located
in (c).

In considering these large composition variations, it must be kept in mind that though Ti is the
minor alloying element, it is still a significant constituent (∼23 at.%). While some of the thinner
Ti-rich regions appear to be just boundary layers around the Hf-rich grains (suggestive of grain
boundary segregation), there are also larger Ti-rich regions which appear to be grains of their own.
However, under high-resolution TEM (HRTEM) imaging, none of the large Ti-rich regions
showed lattice fringes, though lattice fringes were readily apparent in the Hf-rich grains. This
suggests an amorphous Ti-rich structure, which is also indicated by the convergent beam electron
82

diffraction (CBED) pattern shown in Figure 56a, which is from a typical Ti-rich region and has an
indistinct ring. In contrast, the pattern on the right, Figure 56c, was taken from the crystalline Hf-
rich region specified in Figure 56b.

Figure 56: Convergent beam electron diffraction (CBED) patterns from a Ti-rich (a) and a Hf-rich
(c) region of HT-Sample B (a heat-treated 40 nm Hf-Ti/5 nm Ti multilayer sample). The locations
where these patterns were recorded are pointed out by arrows in the cross-sectional TEM
image in (b). The scale bar in (b) is 20 nm long.

It is intriguing that these Ti-rich regions developed in this system. Based on the Hf-Ti phase
diagram, at 800 °C and 23 at.% Ti, the system is expected to show full miscibility [162]. What is
more, the diffusion distances from Figure 54 and Table 7 show that the elements should have been
able to kinetically mix during heat treatment of the multilayers, but they did not. Even more
counterintuitive is the result for Sample C, where the elements were already mixed and evolved to
an unmixed configuration after annealing
Since we observed some oxidation of the samples, the hafnia-titania phase diagram was also
examined, but it does not predict the degree of segregation shown in Figure 55 [163]. Strong
segregation was reported in previous work with nanostructured hafnia-titania samples [163-169],
indicating that the oxidation may influence the Hf and Ti segregation, but it does not explain the
differences in grain growth between the multilayered and monolithic samples. In order to better
understand the slowed grain growth in the multilayered samples, Monte Carlo simulations were
conducted for the Hf-Ti system.
83

8.3.4 Modeling
As a starting point for thermodynamic analysis, the expected bulk equilibrium structure of the Hf-
23 at.% Ti system at 800 °C was simulated by the Monte Carlo method explained in Section 8.2.2;
a forced single crystal structure (with the same grain number for all lattice sites) was simulated by
prohibiting grain boundary formation, and the result is a random solid solution of Hf and Ti (23
at.%) atoms, shown in Figure 57a as a single grain number map in solid green with an overlay of
Ti atom positions as black dots. The images in Figure 57 are 2-D cross sections of the full 3-D
structure.

Figure 57: Monte Carlo models of Hf-Ti structures (23 at.% Ti) at 800 °C. A forced bulk structure
is shown in (a) as a homogeneous solid solution, with Ti atoms presented in black. The fully
equilibrated structure is shown in (b), where Ti atoms are black and the colors indicate different
grains. The structure shown in (c) is for a multilayered sample, with black Ti layers and a
columnar grain structure represented by the grain number map of blue and red in the
background. A table of the calculated internal energies for different configurations is shown in
(d).

In contrast to the above bulk equilibrium structure, when the same alloy was allowed to fully
equilibrate without the bulk constraint, and permitting intergranular configurations using the
84

procedure described in Section 8.2.2, the resulting structure is shown in Figure 57b. Again the
black dots are Ti atoms and the colored background represents the grain structure. The
polycrystalline structure is of lower free energy than the bulk equilibrium one, because the grain
boundaries in the polycrystal are decorated preferentially with Ti atoms. The grain sizes in Figure
57b vary from approximately 10 to 80 nm in diameter, which is a good qualitative match for our
observed grain sizes in the experiments, Figure 53. Figure 57b does not show the large Ti-rich
regions that we observed in the heat-treated samples, e.g. Figure 55. However, it must be kept in
mind that the model is for pure Hf-Ti and does not account for oxides.
Multilayered Hf-Ti samples were also explicitly constructed to evaluate their internal energies for
comparison with those of both the bulk structure, Figure 57a, and the fully equilibrated structure,
Figure 57b. An example of such a model having 40 nm thick Hf-Ti layers is shown in Figure 57c,
where Ti atoms are black and the columnar grain structure is represented by the grain number map
of blue and red in the background.
The structures in Figure 57 can be compared on the basis of their internal energy, which is
approximately equal to the enthalpy. Although the MC algorithm produced the structures of Figure
57a and Figure 57b using a thermal process that accounts for entropy, we do not have a quantitative
measure of the configurational entropy of any of these structures, and thus compare only their
internal energies, which are given in Figure 57d. The structures of lowest internal energy are those
of the multilayers, which are fully phase separated, and thus which are in the classical ground state
for the Hf-Ti system, which has a positive heat of mixing. It is the contribution of entropy that
permits the bulk solid solution to become the favored structure at 800 °C. More interestingly,
though, the bulk solid solution is of higher internal energy than the grain boundary segregated
85

polycrystal; even without the contribution of entropy this system would appear to favor the
presence of segregated grain boundaries.
These results therefore suggest that the equiaxed polycrystalline structure is favored at 800 °C in
a thermodynamic sense, with a very strong enthalpic contribution to the stability of that structure.
This not only aligns with the alloy configurations observed in the experiments, it also helps to
explain how several very different initial structures, both mixed and unmixed, evolve to a very
similar nanostructured state with nominally unexpected titanium partitioning around hafnium-rich
grains. Although the presence of oxygen in the experiments is a confounding factor not captured
in the simulations, the present results suggest that interfacial alloy thermodynamics play a strong
role in stabilizing a nanostructure in this alloy.
8.4 Conclusions
Sputtered multilayer and monolithic samples were used to investigate the predicted high-
temperature nanograin stability of the Hf-Ti system. TEM imaging of heat-treated samples
indicated strong Hf and Ti segregation, which was not expected based on the Hf-Ti phase diagram.
There is an indication of grain boundary segregation of Ti around Hf-rich grains, as well as some
Ti-rich regions of the samples that appear to be amorphous. In addition, the multilayered samples
exhibited more limited grain growth than the monolithic sample.
Monte Carlo simulations were conducted to better understand this phenomenon, and it was found
that the nanograined configurations were obtained from a full thermodynamic equilibration and
have lower internal energies than the solid solution Hf-Ti phase that is expected from the phase
diagram, which helps explain the observed nanograin stability.
86

The main advantages of the multilayer sputtering technique for the study of nanograin stability are
control of grain size, global composition, and elemental distribution. This control allows for
tailored starting configurations that may be energetically favorable, as was observed for the
multilayered samples in this study. Thus, sputtered multilayers are useful for continuing
investigations of nanograin stability and could be a route to produce stable nanograin structures.

87

9 Conclusions and Future Work
Nano metallic multilayers (NMMs) have many attractive properties and behaviors. However,
much additional testing and characterization is needed to optimize them for widespread
application. Thus, this study looked at their microstructures in as-sputtered, deformed, and heat-
treated states, in order to better understand their behavior.
In the first part of this study, Cu/Nb and Al/Nb NMMs were examined by a high-resolution
orientation mapping TEM technique (ACOM), and the crystal structures and orientations of sub-
20 nm layers was ascertained. It was revealed that thin (2 nm) layers of Cu and Al failed to
interrupt the Nb grain orientations. On the other hand, 20 nm Cu and Al layers showed a mix of
amorphous and crystalline regions, with the amorphous regions interrupting the Nb grain
orientations and crystalline regions generally allowing transmission of the Nb grain orientation.
The presence or lack of orientation interruption is significant, since such interruptions will refine
the microstructure and result in changes in the material behavior, due to factors such as more
difficult dislocation transmission across layer boundaries and more tortuous corrosion paths.
Based on this study, the ACOM technique is seen to be critical for a deeper understanding of NMM
behavior, as it gives local orientation information on a global scale. Such imaging will allow for
better control of texture, interfacial epitaxy, and microstructural refinement, all of which need to
be controlled to achieve true optimization of NMM configurations.
The next part of this study focused on the mechanical behavior of NMMs. The goal was to
deconvolute the load-sharing behavior among materials made of nanometer-thick layers. This was
achieved by depositing the films onto polymer substrates and conducting in-situ tensile tests,
wherein the lattice strains in the different materials were tracked separately by XRD. In this
88

manner, the deformation behaviors of 2-20 nm layers of Cu and Nb were made evident. It was
found that, as expected, the Cu layers became stronger with decreasing layer thickness. However,
it was also found that the Nb layers failed at lower stresses with decreasing Cu layer thickness.
Thus, while standard tensile testing of these Cu/Nb NMMs may have indicated a bulk weakening
with decreasing Cu layer thickness, it was shown that there are competing behaviors among the
Cu and Nb that oppose one another with regard to the overall strength of the material. It was also
found that though the deformation behaviors appear to be independent in the different materials
prior to failure, the point of failure will be a coupling of the behavior of the two materials.
Such testing of NMMs is critical for optimization of materials and layer thicknesses when
designing NMMs, since this testing allows for the influence of individual material layers to be
isolated, and NMMs can thus be configured as necessary to achieve certain properties. Without
such separation of individual layer behavior, much more extensive testing would be necessary to
explore all of the possible configurations and determine which is in fact best suited to a certain
application.
Two promising avenues for further understanding the isolated deformation behaviors are 1) testing
multiple material systems, in order to determine the effect of different interface types (FCC/BCC,
FCC/HCP, FCC/amorphous, etc.) on the deformation of the individual layers, and 2) incorporating
residual stress and texture information, so that the stress and load-bearing in the individual
materials can be fully quantified.
The final part of this study focused on the application of NMMs in thermal stability investigations.
The Hf-Ti system had been predicted to be stable at high temperatures, and to investigate this
prediction Hf-Ti samples were sputtered in monolithic and multilayer configurations. After an
89

extended heat treatment, an unexpectedly large degree of segregation was observed in the samples,
and the multilayered Hf-Ti samples showed limited grain growth compared to the monolithic Hf-
Ti sample. Thus, it was demonstrated that the Hf-Ti system does appear to show some thermal
stabilization, and, based on a difference in behavior between the multilayered and monolithic
samples, we saw that starting configuration is important when conducting thermal stability
investigations. NMMs lend themselves to such studies, since they allow for a large amount of
microstructural control. In addition, the relative stability of the multilayered Hf-Ti samples
indicates that NMMs may be a route to producing thermally stabilized nanostructured materials.
Many other system have been predicted by modeling to exhibit thermal stabilization, and these
systems can be further studied by the fabrication and heat treatment of NMMs in various
configurations. This will turn elucidate the effect of grain size, composition, and other parameters
on the high-temperature behavior of these systems.
Overall, the studies which were conducted showed 1) the importance of nanoscale orientation
characterization of NMMs, 2) the ability to decouple deformation behavior of layers within
NMMs, and 3) that NMMs are promising for future thermal stability studies and may be an avenue
to producing thermally stabilized nanostructured materials. The conducted studies have opened
multiple questions and provide a clear path to continue answering these questions surrounding the
structure and behavior of NMMs and other nanostructured materials.

90

10 References
[1] C. Suryanarayana. Nanocrystalline Materials, Int. Mater. Rev. 40 (1995) 41-64.
[2] C. Suryanarayana. Recent developments in nanostructured materials, Adv. Eng. Mater. 7
(2005) 983-992.
[3] H. Gleiter. Nanocrystalline Materials, Progress in Materials Science 33 (1989) 223-315.
[4] P. Moriarty. Nanostructured materials, Reports on Progress in Physics 64 (2001) 297-381.
[5] D.E. Aston, J.R. Bow, D.N. Gangadean. Mechanical properties of selected nanostructured
materials and complex bio-nano, hybrid and hierarchical systems, Int. Mater. Rev. 58 (2013) 167-
202.
[6] J.H. Lee, J.P. Singer, E.L. Thomas. Micro-/Nanostructured Mechanical Metamaterials,
Adv. Mater. 24 (2012) 4782-4810.
[7] M. Meyers, K. Chawla. Mechanical Behavior of Materials. 2nd ed., Cambridge University
Press, 2009.
[8] G.J.O. K.O. Sanusi. Effects of grain size on mechanical properties of nanostructured
copper alloy by severe plastic deformation (SPD) process, Journal of Engineering, Design and
Technology 7 (2009) 335-341.
[9] G.E. Fougere, J.R. Weertman, R.W. Siegel, S. Kim. Grain-size dependent hardening and
softening of nanocrystalline Cu and Pd, Scr. Metall. Materialia 26 (1992) 1879-1883.
[10] H. Gleiter. Nanostructured materials: Basic concepts and microstructure, Acta Materialia
48 (2000) 1-29.
[11] T. Suzuoka. Exact Solutions of Two Ideal Cases in Grain Boundary Diffusion Problem and
the Application to Sectioning Method, Journal of the Physical Society of Japan 19 (1964) 839-
851.
[12] J.S. Lee, C. Minkwitz, C. Herzig. Grain boundary self-diffusion in polycrystalline tungsten
at low temperatures, Phys. Status Solidi B-Basic Res. 202 (1997) 931-940.
[13] B. Gunther, A. Kumpmann, H.D. Kunze. Secondary recrystallization effects in
nanostructured elemental metals, Scr. Metall. Materialia 27 (1992) 833-838.
[14] P.G. Sanders, G.E. Fougere, L.J. Thompson, J.A. Eastman, J.R. Weertman. Improvements
in the synthesis and compaction of nanocrystalline materials, Nanostruct. Mater. 8 (1997) 243-
252.
[15] J.A. Haber, W.E. Buhro. Kinetic instability of nanocrystalline aluminum prepared by
chemical synthesis; Facile room-temperature grain growth, J. Am. Chem. Soc. 120 (1998) 10847-
10855.
[16] G.D. Hibbard, J.L. McCrea, G. Palumbo, K.T. Aust, U. Erb. An initial analysis of
mechanisms leading to late stage abnormal grain growth in nanocrystalline Ni, Scripta Materialia
47 (2002) 83-87.
[17] J.M. Tao, X.K. Zhu, R.O. Scattergood, C.C. Koch. The thermal stability of high-energy
ball-milled nanostructured Cu, Mater. Des. 50 (2013) 22-26.
[18] S. Okuda, M. Kobiyama, T. Inami, S. Takamura. Thermal stability of nanocrystalline gold
and copper prepared by gas deposition method, Scripta Materialia 44 (2001) 2009-2012.
[19] Y. Chen, Y. Liu, F. Khatkhatay, C. Sun, H. Wang, X. Zhang. Significant enhancement in
the thermal stability of nanocrystalline metals via immiscible tri-phases, Scripta Materialia 67
(2012) 177-180.
[20] J. Weissmuller. Alloy effects in nanostructures, Nanostruct. Mater. 3 (1993) 261-272.
91

[21] P.C. Millett, R.P. Selvam, A. Saxena. Stabilizing nanocrystalline materials with dopants,
Acta Materialia 55 (2007) 2329-2336.
[22] P. Choi, M. da Silva, U. Klement, T. Al-Kassab, R. Kirchheim. Thermal stability of
electrodeposited nanocrystalline Co-1.1at.%P, Acta Materialia 53 (2005) 4473-4481.
[23] K.A. Darling, B.K. VanLeeuwen, C.C. Koch, R.O. Scattergood. Thermal stability of
nanocrystalline Fe-Zr alloys, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 527
(2010) 3572-3580.
[24] K.A. Darling, A.J. Roberts, Y. Mishin, S.N. Mathaudhu, L.J. Kecskes. Grain size
stabilization of nanocrystalline copper at high temperatures by alloying with tantalum, J. Alloy.
Compd. 573 (2013) 142-150.
[25] J.W. Cahn. Impurity-drag effect in grain boundary motion, Acta Metallurgica 10 (1962)
789-798.
[26] C.C. Koch, R.O. Scattergood, M. Saber, H. Kotan. High temperature stabilization of
nanocrystalline grain size: Thermodynamic versus kinetic strategies, Journal of Materials
Research 28 (2013) 1785-1791.
[27] E. Nes, N. Ryum, O. Hunderi. On the Zener drag, Acta Metallurgica 33 (1985) 11-22.
[28] H. Sun, C. Deng. Direct quantification of solute effects on grain boundary motion by
atomistic simulations, Comput. Mater. Sci. 93 (2014) 137-143.
[29] Z. Chen, F. Liu, X.Q. Yang, C.J. Shen, Y.M. Zhao. A thermokinetic description of nano-
scale grain growth under dynamic grain boundary segregation condition, J. Alloy. Compd. 608
(2014) 338-342.
[30] R. Kirchheim. Grain coarsening inhibited by solute segregation, Acta Materialia 50 (2002)
413-419.
[31] T. Chookajorn, H.A. Murdoch, C.A. Schuh. Design of Stable Nanocrystalline Alloys,
Science 337 (2012) 951-954.
[32] H.A. Murdoch, C.A. Schuh. Estimation of grain boundary segregation enthalpy and its role
in stable nanocrystalline alloy design, Journal of Materials Research 28 (2013) 2154-2163.
[33] H.A. Murdoch, C.A. Schuh. Stability of binary nanocrystalline alloys against grain growth
and phase separation, Acta Materialia 61 (2013) 2121-2132.
[34] T. Chookajorn, C.A. Schuh. Thermodynamics of stable nanocrystalline alloys: A Monte
Carlo analysis, Physical Review B 89 (2014).
[35] F. Wang, L.F. Zhang, P. Huang, J.Y. Xie, T.J. Lu, K.W. Xu. Microstructure and Flow
Stress of Nanoscale Cu/Nb Multilayers, J. Nanomater. (2013).
[36] B. Ham, X. Zhang. High strength Mg/Nb nanolayer composites, Mater. Sci. Eng. A-Struct.
Mater. Prop. Microstruct. Process. 528 (2011) 2028-2033.
[37] H.B. Wan, Y. Shen, X. He, J. Wang. Modeling of Microstructure Evolution in Metallic
Multilayers with Immiscible Constituents, Jom 65 (2013) 443-449.
[38] D.A. Porter, K.E. Easterling, M.Y. Sherif. Phase Transformations in Metals and Alloys.
3rd ed., CRC Press, Taylor & Francis Group, Boca Raton, FL, 2009.
[39] S.C. Erwin, C.X. Gao, C. Roder, J. Lahnemann, O. Brandt. Epitaxial Interfaces between
Crystallographically Mismatched Materials, Phys. Rev. Lett. 107 (2011).
[40] I.J. Beyerlein, N.A. Mara, J. Wang, J.S. Carpenter, S.J. Zheng, W.Z. Han, R.F. Zhang, K.
Kang, T. Nizolek, T.M. Pollock. Structure-Property-Functionality of Bimetal Interfaces, Jom 64
(2012) 1192-1207.
92

[41] S.I. Hong, M.A. Hill. Microstructural stability of Cu-Nb microcomposite wires fabricated
by the bundling and drawing process, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process.
281 (2000) 189-197.
[42] B.P. Shu, L. Liu, Y.D. Deng, C. Zhong, Y.T. Wu, B. Shen, W.B. Hu. An investigation of
grain boundary diffusion and segregation of Ni in Cu in an electrodeposited Cu/Ni micro-
multilayer system, Mater. Lett. 89 (2012) 223-225.
[43] C.A. Ross. Electrodeposited multilayer thin-films, Annu. Rev. Mater. Sci. 24 (1994) 159-
188.
[44] A. Misra, M.J. Demkowicz, X. Zhang, R.G. Hoagland. The radiation damage tolerance of
ultra-high strength nanolayered composites, Jom 59 (2007) 62-65.
[45] X. Zhang, E.G. Fu, N. Li, A. Misra, Y.Q. Wang, L. Shao, H. Wang. Design of Radiation
Tolerant Nanostructured Metallic Multilayers, J. Eng. Mater. Technol.-Trans. ASME 134 (2012).
[46] S. Yogesha, R.S. Bhat, K. Venkatakrishna, G.P. Pavithra, Y. Ullal, A.C. Hegde.
Development of Nano-structured Zn-Ni Multilayers and their Corrosion Behaviors, Synth. React.
Inorg. Met.-Org. Nano-Metal Chem. 41 (2011) 65-71.
[47] T.T. Xie, S.D. Mao, C. Yu, S.J. Wang, Z.L. Song. Structure, corrosion, and hardness
properties of Ti/Al multilayers coated on NdFeB by magnetron sputtering, Vacuum 86 (2012)
1583-1588.
[48] M. Parra-Borderias, I. Fernandez-Martinez, L. Fabrega, A. Camon, O. Gil, R. Gonzalez-
Arrabal, J. Sese, J.L. Costa-Kramer, B. Warot-Fonrose, V. Serin, F. Briones. Thermal stability of
Mo/Au bilayers for TES applications, Supercond. Sci. Technol. 25 (2012).
[49] T.C. Chen, F.M. Finkbeiner, A. Bier, B. DiCamillo. Molybdenum-gold proximity bilayers
as transition edge sensors for microcalorimeters and bolometers, Supercond. Sci. Technol. 12
(1999) 840-842.
[50] T.W. Barbee, S. Mrowka, M.C. Hettrick. Molybdenum-silicon multilayer mirros for the
extreme ultraviolet, Appl. Optics 24 (1985) 883-886.
[51] G.C. Hilton, J.A. Beall, S. Deiker, L.R. Vale, W.B. Doriese, J. Beyer, J.N. Ullom, C.D.
Reintsema, Y. Xu, K.D. Irwin. X-ray microcalorimeter arrays fabricated by surface
micromachining, Nucl. Instrum. Methods Phys. Res. Sect. A-Accel. Spectrom. Dect. Assoc.
Equip. 520 (2004) 435-438.
[52] D. Bisero, G. Bordin, M. Minelli, F. Ronconi, F. Spizzo, A. Baraldi, S. Lizzit, G. Paolucci,
L. Pareti, G. Turilli. Effects of atomic diffusion processes in Co-Cu multilayer granular films,
Nanostruct. Mater. 11 (1999) 769-774.
[53] M. Hecker, J. Thomas, D. Tietjen, S. Baunack, C.M. Schneider, A. Qiu, N. Cramer, R.E.
Camley, Z. Celinski. Thermally induced modification of GMR in Co/Cu multilayers: correlation
among structural, transport, and magnetic properties, J. Phys. D-Appl. Phys. 36 (2003) 564-572.
[54] M.N. Baibich, J.M. Broto, A. Fert, F.N. Vandau, F. Petroff, P. Eitenne, G. Creuzet, A.
Friederich, J. Chazelas. Giant magnetoresistance of (001) Fe/(001) Cr magnetic superlattices,
Phys. Rev. Lett. 61 (1988) 2472-2475.
[55] S. Lei, J.Y. Zhang, J.J. Niu, G. Liu, X. Zhang, J. Sun. Intrinsic size-controlled strain
hardening behavior of nanolayered Cu/Zr micropillars, Scripta Materialia 66 (2012) 706-709.
[56] X.F. Zhu, G.P. Zhang. Tensile and fatigue properties of ultrafine Cu-Ni multilayers, J.
Phys. D-Appl. Phys. 42 (2009).
[57] J.Y. Zhang, P. Zhang, R.H. Wang, G. Liu, G.J. Zhang, J. Sun. Enhanced mechanical
properties of columnar grained-nanotwinned Cu films on compliant substrate via multilayer
scheme, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 554 (2012) 116-121.
93

[58] N. Li, J. Wang, A. Misra, J.Y. Huang. Direct Observations of Confined Layer Slip in
Cu/Nb Multilayers, Microsc. microanal. 18 (2012) 1155-1162.
[59] K. Hattar, A. Misra, M.R.F. Dosanjh, P. Dickerson, I.M. Robertson, R.G. Hoagland. Direct
Observation of Crack Propagation in Copper-Niobium Multilayers, J. Eng. Mater. Technol.-Trans.
ASME 134 (2012).
[60] E.G. Fu, N. Li, A. Misra, R.G. Hoagland, H. Wang, X. Zhang. Mechanical properties of
sputtered Cu/V and Al/Nb multilayer films, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct.
Process. 493 (2008) 283-287.
[61] Y.C. Wang, A. Misra, R.G. Hoagland. Fatigue properties of nanoscale Cu/Nb multilayers,
Scripta Materialia 54 (2006) 1593-1598.
[62] M.R. Stoudt, R.C. Cammarata, R.E. Ricker. Suppression of fatigue cracking with
nanometer-scale multilayered coatings, Scripta Materialia 43 (2000) 491-496.
[63] A. Misra, J.P. Hirth, R.G. Hoagland. Length-scale-dependent deformation mechanisms in
incoherent metallic multilayered composites, Acta Materialia 53 (2005) 4817-4824.
[64] S.P. Wen, R.L. Zong, F. Zeng, Y.L. Gu, Y. Gao, F. Pan. Thermal stability of microstructure
and mechanical properties of Ni/Ru multilayers, Surf. Coat. Technol. 202 (2008) 2040-2046.
[65] A. Misra, J.P. Hirth, H. Kung. Single-dislocation-based strengthening mechanisms in
nanoscale metallic multilayers, Philos. Mag. A-Phys. Condens. Matter Struct. Defect Mech. Prop.
82 (2002) 2935-2951.
[66] S.I. Rao, P.M. Hazzledine. Atomistic simulations of dislocation-interface interactions in
the Cu-Ni multilayer system, Philos. Mag. A-Phys. Condens. Matter Struct. Defect Mech. Prop.
80 (2000) 2011-2040.
[67] J.S. Koehler. Attempt to design a strong solid, Physical Review B 2 (1970) 547-551.
[68] X. Chu, S.A. Barnett. Model of superlattice yield stress and hardness enhancements, J.
Appl. Phys. 77 (1995) 4403-4411.
[69] R. Banerjee, R. Ahuja, H.L. Fraser. Dimensionally induced structural transformations in
titanium-aluminum multilayers, Phys. Rev. Lett. 76 (1996) 3778-3781.
[70] A. Misra, H. Kung, D. Hammon, R.G. Hoagland, M. Nastasi. Damage Mechanisms in
Nanolayered Metallic Composites, Int. J. Damage Mech. 12 (2003) 365-376.
[71] R.G. Hoagland, R.J. Kurtz, C.H. Henager. Slip resistance of interfaces and the strength of
metallic multilayer composites, Scripta Materialia 50 (2004) 775-779.
[72] H.M. Zbib, C.T. Overman, F. Akasheh, D. Bahr. Analysis of plastic deformation in
nanoscale metallic multilayers with coherent and incoherent interfaces, Int. J. Plast. 27 (2011)
1618-1639.
[73] Y. Kim, A.S. Budiman, J.K. Baldwin, N.A. Mara, A. Misra, S.M. Han. Microcompression
study of Al-Nb nanoscale multilayers, Journal of Materials Research 27 (2012) 592-598.
[74] A. Misra, M. Verdier, Y.C. Lu, H. Kung, T.E. Mitchell, N. Nastasi, J.D. Embury. Structure
and mechanical properties of Cu-X (X = Nb,Cr,Ni) nanolayered composites, Scripta Materialia 39
(1998) 555-560.
[75] H.B. Huang, F. Spaepen. Tensile testing of free-standing Cu, Ag and Al thin films and
Ag/Cu multilayers, Acta Materialia 48 (2000) 3261-3269.
[76] Y. Wang, J. Li, A.V. Hamza, T.W. Barbee. Ductile crystalline-amorphous nanolaminates,
Proceedings of the National Academy of Sciences 104 (2007) 11155-11160.
[77] F. Wang, P. Huang, M. Xu, T.J. Lu, K.W. Xu. Shear banding deformation in Cu/Ta nano-
multilayers, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 528 (2011) 7290-7294.
94

[78] P. Dayal, M.Z. Quadir, C. Kong, N. Savvides, M. Hoffman. Transition from dislocation
controlled plasticity to grain boundary mediated shear in nanolayered aluminum/palladium thin
films, Thin Solid Films 519 (2011) 3213-3220.
[79] J.Y. Zhang, S. Lei, J. Niu, Y. Liu, G. Liu, X. Zhang, J. Sun. Intrinsic and extrinsic size
effects on deformation in nanolayered Cu/Zr micropillars: From bulk-like to small-volume
materials behavior, Acta Materialia 60 (2012) 4054-4064.
[80] D.L. Beke, G.A. Langer, M. Kis-Varga, A. Dudas, P. Nemes, L. Daroczi, G. Kerekes, Z.
Erdelyi. Thermal stability of amorphous and crystalline multilayers produced by magnetron
sputtering, Vacuum 50 (1998) 373-383.
[81] D.L. Beke, Z. Erdelyi, G.A. Langer, A. Csik, G.L. Katona. Diffusion on the nanometer
scale, Vacuum 80 (2005) 87-91.
[82] J.S. Carpenter, S.J. Zheng, R.F. Zhang, S.C. Vogel, I.J. Beyerlein, N.A. Mara. Thermal
stability of Cu-Nb nanolamellar composites fabricated via accumulative roll bonding,
Philosophical Magazine 93 (2013) 718-735.
[83] A. Ullrich, M. Bobeth, W. Pompe. Monte Carlo investigation of the thermal stability of
coherent multilayers, Scripta Materialia 43 (2000) 887-892.
[84] P. Troche, J. Hoffmann, K. Heinemann, F. Hartung, G. Schmitz, H.C. Freyhardt, D.
Rudolph, J. Thieme, P. Guttmann. Thermally driven shape instabilities of Nb/Cu multilayer
structures: instability of Nb/Cu multilayers, Thin Solid Films 353 (1999) 33-39.
[85] A. Misra, R.G. Hoagland, H. Kung. Thermal stability of self-supported nanolayered Cu/Nb
films, Philosophical Magazine 84 (2004) 1021-1028.
[86] D.R. Economy, B.M. Schultz, M.S. Kennedy. Impacts of accelerated aging on the
mechanical properties of Cu-Nb nanolaminates, J. Mater. Sci. 47 (2012) 6986-6991.
[87] C. Lin, G.W. Yang, B.X. Liu. Prediction of solid-state amorphization in binary metal
systems, Physical Review B 61 (2000) 15649-15652.
[88] A.A. Navid, A.M. Hodge. Nanostructured alpha and beta tantalum formation-Relationship
between plasma parameters and microstructure, Mater. Sci. Eng. A-Struct. Mater. Prop.
Microstruct. Process. 536 (2012) 49-56.
[89] J.A. Thornton, D.W. Hoffman. Stress-related effects in thin-films, Thin Solid Films 171
(1989) 5-31.
[90] I. Levine, A. Yoffe, A. Salomon, W.J. Li, Y. Feldman, A. Vilan. Epitaxial two dimensional
aluminum films on silicon (111) by ultra-fast thermal deposition, J. Appl. Phys. 111 (2012).
[91] D.L. Rode, V.R. Gaddam, J.H. Yi. Subnanometer surface roughness of dc magnetron
sputtered Al films, J. Appl. Phys. 102 (2007).
[92] P. Moeck, S. Rouvimov, E.F. Rauch, M. Veron, H. Kirmse, I. Hausler, W. Neumann, D.
Bultreys, Y. Maniette, S. Nicolopoulos. High spatial resolution semi-automatic crystallite
orientation and phase mapping of nanocrystals in transmission electron microscopes, Cryst. Res.
Technol. 46 (2011) 589-606.
[93] Scanning Transmission Electron Microscopy: Imaging and Analysis, Springer
Science+Business Media, LLC, 2011.
[94] L. Lu, Y.F. Shen, X.H. Chen, L.H. Qian, K. Lu. Ultrahigh strength and high electrical
conductivity in copper, Science 304 (2004) 422-426.
[95] A.M. Hodge, Y.M. Wang, T.W. Barbee Jr. Mechanical deformation of high-purity sputter-
deposited nano-twinned copper, Scripta Materialia 59 (2008) 163-166.
[96] D. Bufford, H. Wang, X. Zhang. High strength, epitaxial nanotwinned Ag films, Acta
Materialia 59 (2011) 93-101.
95

[97] D. Bufford, Z. Bi, Q.X. Jia, H. Wang, X. Zhang. Nanotwins and stacking faults in high-
strength epitaxial Ag/Al multilayer films, Applied Physics Letters 101 (2012).
[98] K. Han, K. Yu-Zhang. Transmission electron microscopy study of metallic multilayers,
Scripta Materialia 50 (2004) 781-786.
[99] L. Velasco, M.N. Polyakov, A.M. Hodge. Influence of stacking fault energy on twin
spacing of Cu and Cu-Al alloys, Scripta Materialia 83 (2014) 33-36.
[100] Y. Zhao, M.N. Polyakov, M. Mecklenburg, M.E. Kassner, A.M. Hodge. The role of grain
boundary plane orientation in the β phase precipitation of an Al–Mg alloy, Scripta Materialia 89
(2014) 49-52.
[101] D. Neerinck, J.P. Locquet, L. Stockman, Y. Bruynseraede, I.K. Schuller. Lattice mismatch
and interfacial disorder in superlattices, Phys. Scr. 39 (1989) 346-350.
[102] D.M. Vardanyan, H.M. Manoukyan. The dynamic theory of X-ray-diffraction by the one-
dimensional ideal superlattice .1. Diffraction by the arbitrary superlattice, Acta Crystallogr. Sect.
A 41 (1985) 212-217.
[103] L. Haggstrom, I. Soroka, S. Kamali. Thickness dependent crystallographic transition in
Fe/Ni multilayers. in: Muller H, Reissner M, Steiner W, Wiesinger G, (Eds.). International
Conference on the Applications of the Mossbauer Effect, vol. 217. Iop Publishing Ltd, Bristol,
2010.
[104] B.C. Kang, H.Y. Kim, O.Y. Kwon, S.H. Hong. Bilayer thickness effects on
nanoindentation behavior of Ag/Ni multilayers, Scripta Materialia 57 (2007) 703-706.
[105] E.E. Fullerton, W. Cao, G. Thomas, I.K. Schuller, M.J. Carey, A.E. Berkowitz.
Quantitative characterization of epitaxial superlattices by X-ray-diffraction and high-resolution
electron-microscopy, Applied Physics Letters 63 (1993) 482-484.
[106] J.P. Locquet, D. Neerinck, L. Stockman, Y. Bruynseraede, I.K. Schuller. Discrete and
continuous disorder in superlattices, Physical Review B 39 (1989) 13338-13342.
[107] M.N. Polyakov, E. Courtois-Manara, D. Wang, K. Chakravadhanula, C. Kubel, A.M.
Hodge. Microstructural variations in Cu/Nb and Al/Nb nanometallic multilayers, Applied Physics
Letters 102 (2013).
[108] R.L. Zong, S.P. Wen, F. Lv, F. Zeng, Y. Gao, F. Pan. Amorphous-CuTa/amorphous-
CoZrNb multilayers: Structure, mechanical properties and thermal stability, Surf. Coat. Technol.
202 (2008) 4242-4247.
[109] V. Mikhelashvili, G. Eisenstein. Composition, surface morphology and electrical
characteristics of Al2O3-TiO2 nanolaminates and AlTiO films on silicon, Thin Solid Films 515
(2006) 346-352.
[110] J.B. Vella, A.B. Mann, H. Kung, C.L. Chien, T.P. Weihs, R.C. Cammarata. Mechanical
properties of nanostructured amorphous metal multilayer thin films, Journal of Materials Research
19 (2004) 1840-1848.
[111] H. Kung, Y.C. Lu, A.J. Griffin, M. Nastasi, T.E. Mitchell, J.D. Embury. Observation of
body centered cubic Cu in Cu/Nb nanolayered composites, Applied Physics Letters 71 (1997)
2103-2105.
[112] J. Wang, R.G. Hoagland, A. Misra. Phase transition and dislocation nucleation in Cu-Nb
layered composites during physical vapor deposition, Journal of Materials Research 23 (2008)
1009-1014.
[113] Z.Q. Wang, S.H. Lu, Y.S. Li, F. Jona, P.M. Marcus. Epitaxial-growth of a metastable
modification of copper with body-centered-cubic structure, Physical Review B 35 (1987) 9322-
9325.
96

[114] J. Lohmiller, R. Baumbusch, O. Kraft, P.A. Gruber. Differentiation of Deformation Modes
in Nanocrystalline Pd Films Inferred from Peak Asymmetry Evolution Using In Situ X-Ray
Diffraction, Phys. Rev. Lett. 110 (2013).
[115] S. Djaziri, P.O. Renault, F. Hild, E. Le Bourhis, P. Goudeau, D. Thiaudiere, D. Faurie.
Combined synchrotron X-ray and image-correlation analyses of biaxially deformed W/Cu
nanocomposite thin films on Kapton, J. Appl. Crystallogr. 44 (2011) 1071-1079.
[116] S. Frank, U.A. Handge, S. Olliges, R. Spolenak. The relationship between thin film
fragmentation and buckle formation: Synchrotron-based in situ studies and two-dimensional stress
analysis, Acta Materialia 57 (2009) 1442-1453.
[117] P.A. Gruber, E. Arzt, R. Spolenak. Brittle-to-ductile transition in ultrathin Ta/Cu film
systems, Journal of Materials Research 24 (2009) 1906-1918.
[118] S. Djaziri, D. Faurie, E. Le Bourhis, P. Goudeau, P.O. Renault, C. Mocuta, D. Thiaudiere,
F. Hild. Deformation modes of nanostructured thin film under controlled biaxial deformation, Thin
Solid Films 530 (2013) 30-34.
[119] P.A. Gruber, J. Bohm, F. Onuseit, A. Wanner, R. Spolenak, E. Arzt. Size effects on yield
strength and strain hardening for ultra-thin Cu films with and without passivation: A study by
synchrotron and bulge test techniques, Acta Materialia 56 (2008) 2318-2335.
[120] L. Thilly, P.O. Renault, V. Vidal, F. Lecouturier, S. Van Petegem, U. Stuhr, H. Van
Swygenhoven. Plasticity of multiscale nanofilamentary Cu/Nb composite wires during in situ
neutron diffraction: Codeformation and size effect, Applied Physics Letters 88 (2006).
[121] E. Le Bourhis, D. Faurie, B. Girault, P. Goudeau, P.O. Renault, P. Villain, F. Badawi.
Mechanical properties of thin films and nanometric multilayers using tensile testing and
synchrotron X-ray diffraction, Plasma Process. Polym. 4 (2007) 311-317.
[122] S. Djaziri, D. Faurie, P.O. Renault, E. Le Bourhis, P. Goudeau, G. Geandier, D. Thiaudiere.
Yield surface of polycrystalline thin films as revealed by non-equibiaxial loadings at small
deformation, Acta Materialia 61 (2013) 5067-5077.
[123] C.C. Aydiner, D.W. Brown, N.A. Mara, J. Almer, A. Misra. In situ x-ray investigation of
freestanding nanoscale Cu-Nb multilayers under tensile load, Applied Physics Letters 94 (2009).
[124] A.S. Budiman, S.M. Han, N. Li, Q.M. Wei, P. Dickerson, N. Tamura, M. Kunz, A. Misra.
Plasticity in the nanoscale Cu/Nb single-crystal multilayers as revealed by synchrotron Laue x-ray
microdiffraction, Journal of Materials Research 27 (2012) 599-611.
[125] P. Goudeau, P. Villain, T. Girardeau, P.O. Renault, K.F. Badawi. Elastic constants
investigation by X-ray diffraction of in situ deformed metallic multi-layers, Scripta Materialia 50
(2004) 723-727.
[126] M.N. Polyakov, J. Lohmiller, P.A. Gruber, A.M. Hodge. Load Sharing Phenomena in
Nanoscale Cu/Nb Multilayers, Adv. Eng. Mater. (2014) n/a-n/a.
[127] J. Lohmiller, A. Kobler, R. Spolenak, P.A. Gruber. The effect of solute segregation on
strain localization in nanocrystalline thin films: Dislocation glide vs. grain-boundary mediated
plasticity, Applied Physics Letters 102 (2013).
[128] C. Eberl, D.S. Gianola, K.J. Hemker. Mechanical Characterization of Coatings Using
Microbeam Bending and Digital Image Correlation Techniques, Exp. Mech. 50 (2010) 85-97.
[129] J. Lohmiller, N.C. Woo, R. Spolenak. Microstructure-property relationship in highly
ductile Au-Cu thin films for flexible electronics, Mater. Sci. Eng. A-Struct. Mater. Prop.
Microstruct. Process. 527 (2010) 7731-7740.
[130] T.E. Mitchell, Y.C. Lu, A.J. Griffin, M. Nastasi, H. Kung. Structure and mechanical
properties of copper/niobium multilayers, J. Am. Ceram. Soc. 80 (1997) 1673-1676.
97

[131] I.N. Mastorakos, H.M. Zbib, D.F. Bahr. Deformation mechanisms and strength in
nanoscale multilayer metallic composites with coherent and incoherent interfaces, Applied Physics
Letters 94 (2009) 173114.
[132] I.N. Mastorakos, A. Bellou, D.F. Bahr, H.M. Zbib. Size-dependent strength in
nanolaminate metallic systems, Journal of Materials Research 26 (2011) 1179-1187.
[133] J.D. Embury, J.P. Hirth. On dislocation storage and the mechanical response of fine-scale
microstructures, Acta Metall. Mater. 42 (1994) 2051-2056.
[134] J.Y. Zhang, X. Zhang, R.H. Wang, S.Y. Lei, P. Zhang, J.J. Niu, G. Liu, G.J. Zhang, J. Sun.
Length-scale-dependent deformation and fracture behavior of Cu/X (X = Nb, Zr) multilayers: The
constraining effects of the ductile phase on the brittle phase, Acta Materialia 59 (2011) 7368-7379.
[135] J.Y. Zhang, X. Zhang, G. Liu, G.J. Zhang, J. Sun. Dominant factor controlling the fracture
mode in nanostructured Cu/Cr multilayer films, Mater. Sci. Eng. A-Struct. Mater. Prop.
Microstruct. Process. 528 (2011) 2982-2987.
[136] K. Hu, L.J. Xu, Y.Q. Cao, G.J. Pan, Z.H. Cao, X.K. Meng. Modulating individual thickness
for optimized combination of strength and ductility in Cu/Ru multilayer films, Mater. Lett. 107
(2013) 303-306.
[137] C. Hammond. The Basics of Crystallography and Diffraction. 3rd ed., Oxford University
Press Inc., New York, 2009.
[138] J. Lohmiller, R. Spolenak, P.A. Gruber. Alloy-dependent deformation behavior of highly
ductile nanocrystalline AuCu thin films, Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct.
Process. 595 (2014) 235-240.
[139] J. Lohmiller, M. Grewer, C. Braun, A. Kobler, C. Kübel, K. Schüler, V. Honkimäki, H.
Hahn, O. Kraft, R. Birringer, P.A. Gruber. Untangling dislocation and grain boundary mediated
plasticity in nanocrystalline nickel, Acta Materialia 65 (2014) 295-307.
[140] E. Ma. Watching the Nanograins Roll, Science 305 (2004) 623-624.
[141] G.J. Fan, L.F. Fu, Y.D. Wang, Y. Ren, H. Choo, P.K. Liaw, G.Y. Wang, N.D. Browning.
Uniaxial tensile plastic deformation of a bulk nanocrystalline alloy studied by a high-energy x-ray
diffraction technique, Applied Physics Letters 89 (2006) 101918.
[142] T.A. Furnish, J. Lohmiller, P.A. Gruber, T.W. Barbee, A.M. Hodge. Temperature-
dependent strain localization and texture evolution of highly nanotwinned Cu, Applied Physics
Letters 103 (2013).
[143] W. Gambin. Plasticity and Textures, Kluwer Academic Publishers 2001.
[144] B. Girault, V. Vidal, L. Thilly, P.O. Renault, P. Goudeau, E. Le Bourhis, P. Villain-Valat,
G. Geandier, J. Tranchant, J.P. Landesman, P.Y. Tessier, B. Angleraud, M.P. Besland, A. Djouadi,
F. Lecouturier. Small scale mechanical properties of polycrystalline materials: in situ diffraction
studies, Int. J. Nanotechnol. 5 (2008) 609-630.
[145] P. Villain, P. Goudeau, P.-O. Renault, K.F. Badawi. Size effect on intragranular elastic
constants in thin tungsten films, Applied Physics Letters 81 (2002) 4365-4367.
[146] P.P. Choi, J.S. Kim, H.S. Choi, D.H. Kwon, Y.S. Kwon. Production of Cu-Hf-Ti bulk
glassy composites by mechanical alloying and spark-plasma sintering. in: Sung JH, Lee CG, You
YZ, Lee YK, Kim JY, (Eds.). Heat Treatment of Materials, vol. 118. Trans Tech Publications Ltd,
Stafa-Zurich, 2006. pp. 655-659.
[147] V. Ronto, F. Tranta, J. Solyom, A. Kovacs, P. Pekker, Iop. Investigation of Solidification
Behaviour in Cu-based Cu-Hf-Ti alloy system, 3rd International Conference on Advances in
Solidification Processes 27 (2012).
98

[148] X.L. Meng, W. Cai, K.T. Lau, L.M. Zhou, L.C. Zhao. Phase transformation and thermal
stability of aged Ti-Ni-Hf high temperature shape memory alloys, J. Mater. Sci. Technol. 22 (2006)
691-695.
[149] H.J. Fecht, E. Hellstern, Z. Fu, W.L. Johnson. Nanocrystalline metals prepared by high-
energy ball milling, Metallurgical Transactions a-Physical Metallurgy and Materials Science 21
(1990) 2333-2337.
[150] A. Revesz, T. Ungar, A. Borbely, J. Lendvai. Dislocations and grain size in ball-milled
iron powder, Nanostruct. Mater. 7 (1996) 779-788.
[151] Y. Chen, C.A. Schuh. Geometric considerations for diffusion in polycrystalline solids, J.
Appl. Phys. 101 (2007).
[152] R.E. Rudd, W.H. Cabot, K.J. Caspersen, J.A. Greenough, D.F. Richards, F.H. Streitz, P.L.
Miller. Self-diffusivity and interdiffusivity of molten aluminum-copper alloys under pressure,
derived from molecular dynamics, Phys. Rev. E 85 (2012).
[153] L.S. Darken. Diffusion, mobility and their interrelation through free energy ing binary
metallic systems, Transactions of the American Institute of Mining and Metallurgical Engineers
175 (1948) 184-201.
[154] C. Herzig, Y. Mishin, S. Divinski. Bulk and interface boundary diffusion in group IV
hexagonal close-packed metals and alloys, Metall. Mater. Trans. A-Phys. Metall. Mater. Sci. 33
(2002) 765-775.
[155] R.A. Pérez, F. Dyment, G.G. Bermúdez, H. Somacal, D. Abriola. Measurements of Hf
diffusion in α-Ti by HIRBS, Journal of Nuclear Materials 207 (1993) 221-227.
[156] M. Köppers, C. Herzig, M. Friesel, Y. Mishin. Intrinsic self-diffusion and substitutional Al
diffusion in α-Ti, Acta Materialia 45 (1997) 4181-4191.
[157] M. Köppers, C. Herzig, U. Södervall, A. Lodding. Volume Diffusion of Titanium in α-
Hafnium Single Crystals - A Combined Study of SIMS and Tracer Experiments. Defect and
Diffusion Forum, vol. 95: Trans Tech Publications, 1993. p.783-792.
[158] F. Guthoff, Y. Mishin, C. Herzig. Self-diffusion along stationary and moving grain-
boundaries in alpha-Hf, Z. Metallk. 84 (1993) 584-591.
[159] B. Bokstein, V. Ivanov, O. Oreshina, A. Peteline, S. Peteline. Direct experimental
observation of accelerated Zn diffusion along triple junctions in Al, Mater. Sci. Eng. A-Struct.
Mater. Prop. Microstruct. Process. 302 (2001) 151-153.
[160] L.M. Klinger, L.A. Levin, A.L. Petelin. The model of triple junction diffusion, Defect and
Diffusion Forum 143 (1997) 1523-1526.
[161] A. Peteline, S. Peteline, O. Oreshina. Triple junction diffusion: Experiments and models,
Scitec Publications Ltd, Uetikon-Zuerich, 2001.
[162] H. Okamoto. Hf-Ti (hafnium-titanium), J. Phase Equilib. 18 (1997) 672-672.
[163] J.P. Coutures, J. Coutures. The system HfO2-TiO2, J. Am. Ceram. Soc. 70 (1987) 383-
387.
[164] M. Popovici, A. Delabie, S. Van Elshocht, S. Clima, G. Pourtois, L. Nyns, K. Tomida, N.
Menou, K. Opsomer, J. Swerts, C. Detavernier, D. Wouters, J.A. Kittl. Growth and Material
Characterization of Hafnium Titanates Deposited by Atomic Layer Deposition, J. Electrochem.
Soc. 156 (2009) G145-G151.
[165] R. Ruh, G.W. Hollenberg, E.G. Charles, V.A. Patel. Phase Relations and Thermal
Expansion in the System HfO2-TiO2, J. Am. Ceram. Soc. 59 (1976) 495-499.
[166] D.H. Triyoso, R.I. Hegde, S. Zollner, M.E. Ramon, S. Kalpat, R. Gregory, X.D. Wang, J.
Jiang, M. Raymond, R. Rai, D. Werho, D. Roan, B.E. White, P.J. Tobin. Impact of titanium
99

addition on film characteristics of HfO2 gate dielectrics deposited by atomic layer deposition, J.
Appl. Phys. 98 (2005).
[167] M.C. Cisneros-Morales, C.R. Aita. Crystallization, metastable phases, and demixing in a
hafnia-titania nanolaminate annealed at high temperature, J. Vac. Sci. Technol. A 28 (2010) 1161-
1168.
[168] M.C. Cisneros-Morales, C.R. Aita. Intrinsic metastability of orthorhombic HfTiO4 in thin
film hafnia-titania, Applied Physics Letters 98 (2011).
[169] M.C. Cisneros-Morales, C.R. Aita. Mixed cation phases in sputter deposited HfO(2)-
TiO(2) nanolaminates, Applied Physics Letters 93 (2008).
[170] X.L. Zhou, C.Q. Chen. Molecular dynamic simulations of the mechanical properties of
crystalline/crystalline and crystalline/amorphous nanolayered pillars, Comput. Mater. Sci. 101
(2015) 194-200.
[171] K. Wu, J.Y. Zhang, G. Liu, P. Zhang, P.M. Cheng, J. Li, G.J. Zhang, J. Sun. Buckling
behaviors and adhesion energy of nanostructured Cu/X (X = Nb, Zr) multilayer films on a
compliant substrate, Acta Materialia 61 (2013) 7889-7903.
[172] J.Y. Zhang, Y. Liu, J. Chen, Y. Chen, G. Liu, X. Zhang, J. Sun. Mechanical properties of
crystalline Cu/Zr and crystal-amorphous Cu/Cu-Zr multilayers, Mater. Sci. Eng. A-Struct. Mater.
Prop. Microstruct. Process. 552 (2012) 392-398.
[173] J.J. Niu, P. Zhang, R.H. Wang, J.Y. Zhang, G. Liu, G.J. Zhang, J. Sun. Formation of
multiple twins and their strengthening effect in nanocrystalline Cu/Zr multilayer films, Mater. Sci.
Eng. A-Struct. Mater. Prop. Microstruct. Process. 539 (2012) 68-73.
[174] T.G. Nieh, T.W. Barbee, J. Wadsworth. Tensile properties of a free-standing Cu/Zr
nanolaminate (or compositionally-modulated thin film), Scripta Materialia 41 (1999) 929-935.
[175] W. Guo, E. Jagle, J.H. Yao, V. Maier, S. Korte-Kerzel, J.M. Schneider, D. Raabe. Intrinsic
and extrinsic size effects in the deformation of amorphous CuZr/nanocrystalline Cu
nanolaminates, Acta Materialia 80 (2014) 94-106.
[176] J.Y. Zhang, Y.Q. Wang, G. Liu, J. Sun. Plastic deformation characteristics of Cu/X (X=Cu-
Zr, Zr) nanolayered materials, Appl. Surf. Sci. 321 (2014) 19-23.
[177] J.Y. Zhang, G. Liu, S.Y. Lei, J.J. Niu, J. Sun. Transition from homogeneous-like to shear-
band deformation in nanolayered crystalline Cu/amorphous Cu-Zr micropillars: Intrinsic vs.
extrinsic size effect, Acta Materialia 60 (2012) 7183-7196.
[178] T.G. Nieh, J. Wadsworth. Bypassing shear band nucleation and ductilization of an
amorphous-crystalline nanolaminate in tension, Intermetallics 16 (2008) 1156-1159.
[179] Y.M. Wang, A.V. Hamza, T.W. Barbee. Incipient plasticity in metallic glass modulated
nanolaminates, Applied Physics Letters 91 (2007).
[180] M. Elbatanouny, M. Strongin. Structural and electronic trends in the growth of Cu
overlayers on the Nb(110) surface, Physical Review B 31 (1985) 4798-4801.
[181] H. Wormeester, E. Huger, E. Bauer. hcp and bcc Cu and Pd films, Phys. Rev. Lett. 77
(1996) 1540-1543.
[182] F. Zeng, Y. Gao, L. Li, D.M. Li, F. Pan. Elastic modulus and hardness of Cu-Ta amorphous
films, J. Alloy. Compd. 389 (2005) 75-79.
[183] H.R. Gong, L.T. Kong, W.S. Lai, B.X. Liu. Atomistic modeling of solid-state
amorphization in an immiscible Cu-Ta system, Physical Review B 66 (2002).
[184] F.W. Saris, L.S. Hung, M. Nastasi, J.W. Mayer, B. Whitehead. Failure temperature of
amorphous Cu-Ta alloys as diffusion-barriers in Al-Si contacts, Applied Physics Letters 46 (1985)
646-648.
100

[185] J.Y. Zhang, G. Liu, J. Sun. Self-toughening crystalline Cu/amorphous Cu-Zr
nanolaminates: Deformation-induced devitrification, Acta Materialia 66 (2014) 22-31.
[186] G.G. Stoney. The tension of metallic films deposited by electrolysis, Proc. R. soc. Lond.
Ser. A-Contain. Pap. Math. Phys. Character 82 (1909) 172-175.
[187] A. Hodge, R. Foreman, G. Gallegos. Residual stress analysis in thick uranium films,
Journal of Nuclear Materials 342 (2005) 8-13.
[188] S. Timoshenko. Analysis of bi-metal thermostats, J. Opt. Soc. Am. Rev. Sci. Instrum. 11
(1925) 233-255.
[189] J. Chakraborty, K. Kumar, R. Ranjan, S.G. Chowdhury, S.R. Singh. Thickness-dependent
fcc-hcp phase transformation in polycrystalline titanium thin films, Acta Materialia 59 (2011)
2615-2623.
[190] K.S.S. Harsha. Principles of Physical Vapor Deposition of Thin Films, Elsevier, San Diego,
CA, 2006.
[191] N. Bredzs, Tennenho.Cc. Metal-metal oxide-hydrogen atmosphere chart for brazing or
bright metal processing, Weld. J. 49 (1970) S189-S193.
[192] D.W. Umrath. Fundamentals of Vacuum Technology, 1998.
[193] C.B. Alcock, V.P. Itkin, M.K. Horrigan. Vapor-pressure equations for the metallic
elements - 298-2500-K, Can. Metall. Q. 23 (1984) 309-313.
[194] Q.M. Wei, A. Misra. Transmission electron microscopy study of the microstructure and
crystallographic orientation relationships in V/Ag multilayers, Acta Materialia 58 (2010) 4871-
4882.
[195] Y. Kaneko, H. Sakakibara, S. Hashimoto. Dependence of Vickers hardness on annealing
temperature at Co/Cu multilayered films. in: Chang YW, Kim NJ, Lee CS, (Eds.). Pricm 6: Sixth
Pacific Rim International Conference on Advanced Materials and Processing, Pts 1-3, vol. 561-
565. Trans Tech Publications Ltd, Stafa-Zurich, 2007. pp. 2399-2402.
[196] M.A. Monclus, M. Karlik, M. Callisti, E. Frutos, J. Llorca, T. Polcar, J.M. Molina-
Aldareguia. Microstructure and mechanical properties of physical vapor deposited Cu/W
nanoscale multilayers: Influence of layer thickness and temperature, Thin Solid Films 571 (2014)
275-282.
[197] S.J. Zheng, J.S. Carpenter, J. Wang, N.A. Mara, I.J. Beyerlein. An interface facet driven
Rayleigh instability in high-aspect-ratio bimetallic nanolayered composites, Applied Physics
Letters 105 (2014).

101

11 Additional Bibliography
[1] K.S. Sree Harsha. Principles of Vapor Deposition of Thin Films, 2006.
[2] Phase transformations in metals and alloys, 3d ed, Ringgold, Inc, 2009.
[3] Y. Leng. Materials Characterization: Introduction to Microscopic and Spectroscopic
Methods, 2010.
[4] G.F. Vander Voort. Metallography, principles and practice, McGraw-Hill, New York,
1984.
[5] D. Hull. Introduction to dislocations, Pergamon Press, Oxford ;New York, 1975.
[6] J. Weertman, J.R. Weertman. Elementary dislocation theory, Oxford University Press,
New York, 1992.
[7] D.M. Mattox. Handbook of Physical Vapor Deposition (Pvd) Processing, ElsevierEBL -
Ebook Library [Distributor], London : London, 2010.
[8] G. Dieter, D. Bacon. Mechanical Metallurgy. SI Metric ed., 1988.
[9] P.G. Shewmon. Diffusion in solids, Minerals, Metals & Materials Society, Warrendale, Pa,
1989.
[10] M.A. Meyers, K.K. Chawla. Mechanical behavior of materials, Cambridge University
Press, Cambridge ;New York, 2009.
[11] L.V. Azároff. Elements of X-ray crystallography, McGraw-Hill, New York, 1968.

102

Appendix A. In-Situ XRD Tensile Testing
As described in Section 7, in-situ tensile testing was conducted in a synchrotron beamline in order
to track the lattice strains in the different materials of an NMM independently. Several additional
considerations for this form of testing are given in this section, as well as a description of samples
which have been prepared and await further testing.
Appendix A.1 General Notes for Sputtering onto Polyimide
The in-situ testing was conducted with NMMs sputtered onto polyimide substrates. Since
polyimide is a polymer, this presents several complications that are not present when sputtering
onto Si. The first is the thermal stability of polyimide. Though polyimide is stable up to at least
200 °C, which is higher than many other polymers, this temperature can be exceeded during
extended sputtering (see Figure 26). Therefore, sputtering using high powers, low Ar pressures,
and short distances between the substrate and sputtering gun (all of which result in higher
temperatures) is not feasible in some cases. If high-temperature sputtering conditions are
necessary, then interrupted sputtering may be used, so that the substrate is allowed cool between
layers. However, interrupted sputtering conditions must be chosen judiciously, because extended
interruptions may result in contamination by oxygen, nitrogen, or other gases on the freshly
sputtered surfaces.
When sputtering onto polyimide, the pliability of the substrate may result in curling of the samples
upon removal from the substrate holder after sputtering. This was observed to occur for the first
11 Cu/Nb samples sputtered onto polyimide substrates and made those samples of little use for
tensile testing, since the curling of the films resulted in deformation and cracking upon removal
from the sputtering chamber. This was addressed by varying the Ar sputtering pressure. It was
103

found that an Ar pressure of ~0.4 Pa produced the lowest amount of curling for the Cu/Nb layer
thicknesses tested, with lower Ar pressures resulting in compressive stresses and higher Ar
pressures resulting in tensile stresses. However, the optimal Ar pressure increased to 0.47 Pa after
we installed a new Ar flow system. The optimal pressure will be different for other systems and
layer thicknesses. For example, for the optimized Cu/Zr and Cu/CuZr samples, 0.47 Pa was used
for Cu layers, 1.33 Pa for Zr layers, and 2.67 Pa for CuZr layers.
The curling of the films can be reduced by using thicker substrates, since the moment of inertia
will vary with the cube of the substrate thickness. However, we used the 50 μm-thick polyimide
substrates rather than another polyimide substrate which is 125 μm thick, because the surface finish
of the thinner polyimide film was superior [129]. The surface finish is important, since our layer
thicknesses are on the order of nanometers and overall sample thicknesses are on the order of
hundreds of nanometers, meaning that even small surface imperfections may greatly alter the
deformation behavior. (On a side note, the two sides of the polyimide substrates have different
surface roughnesses, so it is important to sputter onto the smoother sides of the substrates.) Since
we used the thinner substrates, we sputtered fewer total layers, in order to minimize the overall
sample thickness and reduce film curling. However, this resulted in longer collection times being
necessary during the in-situ XRD data collection, since our scattering volumes were smaller than
they would have been with more layers.
In addition to surface roughness, surface contamination will also affect the film deformation,
because the contaminants will act as stress concentrations and sources of crack initiation. The
result will be that non-intrinsic deformation modes will be activated, as observed by Lohmiller et
al. in the case of AuCu films [129]. For this reason, a cleaning procedure (Appendix A.2) was
followed that was based on the work in Ref. [129].
104

Appendix A.2 Procedure for Sputtering onto Polyimide
Below is the procedure for cleaning the polyimide substrates which was followed prior to
sputtering.
1) Cutting samples
a) Cut five pieces of lint-free nylon wipes for each polyimide substrate that will be cleaned
(these pieces should fit into the membrane box (60 mm x 60 mm membrane area))
b) Lay polyimide substrate sheet face-down on full size (uncut) clean room wipe
c) Make sure to mark some way to determine if the substrate is face-up or face-down (if you
are cutting polyimide while it’s face-down, make sure to mirror the mark placement). This
is because the bottom of the substrate is rough (currently, the polyimide in storage has the
tissue paper in contact with the top face (the smoother side)). You can check which face
is up by marking the polyimide with a sharpie on both sides and comparing the smoothness
of the sharpie mark under a microscope. 50x mag works well on the Nikon camera (5x
nosepiece lens).
d) Measure rectangles which are 5 cm x 6 cm on the polyimide sheet
e) Cut rectangular pieces with scissors or Exacto knife
f) Store in membrane boxes with nylon wipes sandwiching the polyimide pieces or with the
polyimide pieces taped down in boxes with nothing touching their surfaces
2) Acetone cleaning
a) Put acetone in beaker (600 ml beaker, 175 ml acetone)
b) Place polyimide substrate face up in acetone, cover beaker with Parafilm, secure the
Parafilm with tape
c) Place beakers in ultrasonic cleaner, separate with a clean wipe, weigh them down with
some weight on top
d) Fill the ultrasonic cleaner with new cold water to the “Operating Level” line (make sure to
use cold water, if the water is lukewarm, the sample may overheat during ultrasonic
cleaning, and a film of some sort will form on the substrate)
e) Clean in the ultrasonic cleaner for 45 min
f) Rinse polyimide first with acetone and then isopropanol, blow with air, press between two
nylon wipes, blow with air until dry
g) Store cleaned polyimide pieces
h) Change water in bath to cool the ultrasonic cleaner (will replace this water before starting
a new sample)
i) Pour used acetone into a waste container, wash the inside of the 600 ml beakers with dish
soap (make sure to do this before each sample)
j) Repeat for each sample
3) Isopropanol cleaning
a) Needed supplies for isopropanol cleaning: polyimide substrates, 2” (51 mm) Si wafer,
removable double-sided tape, scissors, tweezers, membrane boxes, isopropanol, gloves,
facemask, Kimwipes/clean wipes, permanent marker, safety glasses
b) Mount the 2” (51 mm) Si wafer on the resist spinner
c) Clean the 2” (51 mm) Si wafer with isopropanol and a clean wipe
105

d) Place small (0.3 cm square) piece of removable double-sided tape in the middle of the resist
spinner, press down with tweezers (if not well pressed down, it will come off with the
substrate)
e) Place polyimide substrate face up and centered on the Si wafer
f) Press down lightly with blunt object on polyimide to adhere it to the tape (only a very slight
push is needed)
g) Start spin cycle (1 min at 3000 rpm, 1 min at 5000 rpm)
h) Pour small amount of isopropanol on polyimide every ~12 seconds of the “1 min at 3000
rpm” spin cycle (total of 5 times), pour line of isopropanol from middle to outside each
time
i) Allow sample to dry during spin cycle of “1 min at 5000 rpm,” make sure nothing drips
onto the surface of the polyimide substrate from the lid of the resist spinner
j) Remove the polyimide from the Si wafer, handle by short edges of the polyimide
k) Store in individual membrane boxes (in between nylon lint-free wipes) or taped face up in
a box (with the surface not touching anything)
4) Sputtering
a) Bake polyimide substrates on a hot plate for at least 15 min at 200 °C to reduce the water
content (cover the substrate with a glass Petri dish during baking to minimize dust
accumulation on the surface)
b) Load the sample into the sample holder
c) Blow substrate with compressed air at a distance of 2-4 cm to remove dust (blow with
Argon or bitterant-free compressed air to avoid bitterant contamination, if using
compressed air, make sure not to tilt the can as that will cause some residue to form on the
substrate)
d) When choosing sputtering conditions, a sample with no curvature (no residual stresses) is
optimal for tensile testing, and small compressive stresses are preferable to small tensile
stresses
5) Removing from chamber
a) Film may curl up when removed from sample holder
b) May have to cut the film free from the top plate of the sample holder by dragging an Exacto
knife along the corner, in order to avoid the film lifting off of the polyimide when the
sample holder cover is removed
6) General storage: secure the sample by taping down an edge in a clear plastic box, this way the
surface is not touched by anything, use as little tape as possible so that the sample is not
deformed when removing the tape or cutting the tape in order to remove the sample from the
box

Appendix A.3 Additional Samples
Descriptions of additional NMM samples which have been sputtered onto polyimide substrates are
given in Table 8, Table 9, and Table 10. The sample architecture was directed by the following
goals: 1) 60 nm total thickness for the material whose layer thicknesses were being varied (not
106

achieved for samples with 2 nm layers, due to film stresses and machine limitations), 2) The same
initial layer for each system, in order to have similar stress transfer from the substrate (initial layer
is Nb for Table 8, Zr on a Ti seed layer for Table 9, and CuZr for Table 10), and 3) The capping
layer was the material whose thickness was not being varied in that sample set.
Table 8 describes a set of Cu/Nb NMMs (an FCC/BCC system), with varying Nb layer thicknesses
(2-20 nm) and constant Cu layer thicknesses (10 nm). These samples were sputtered under
conditions similar to those described in Section 7 and should complement the data for Cu/Nb
NMMs with varying Cu layer thickness [107].
Table 8: Sputtered Cu/Nb NMMs with varying Nb layer thicknesses
Cu/Nb sample
name
Cu layer
thickness
(nm)
Number
of Cu
layers
Total Cu
thickness
(nm)
Nb layer
thickness
(nm)
Number
of Nb
layers
Total Nb
thickness
(nm)
Total
thickness
(nm)
Cu/Nb-Poly 24 10 12 120 5 12 60 180
Cu/Nb-Poly 25 10 6 60 10 6 60 120
Cu/Nb-Poly 26 10 3 30 20 3 60 90
Cu/Nb-Poly 27 10 22 220 2 22 44 264

Table 9 describes a set of Cu/Zr NMMs (an FCC/HCP system), with varying Cu layer thicknesses
(Cu/Zr Poly 8 through Cu/Zr Poly 13) and varying Zr layer thicknesses (Cu/Zr Poly 14 through
Cu/Zr Poly 18). The testing should elucidate the tensile deformation behaviors in this system,
which has previously been studied by nanoindentation, nanopillar compression, tensile testing
(without in-situ XRD), and MD simulations [79, 170-174]. A Ti seed layer was used to improve
adhesion, as the high stresses in the samples would cause the films to peel off after sputtering. The
necessity of a seed layer was avoided in the Cu/Nb samples by varying the sputtering conditions
to find a set of Cu and Nb conditions which gave low overall stresses. However, the desire to
match the Cu sputtering conditions for the Cu/Zr samples with those used for the Cu/Nb samples
107

limited the ability to reduce the Cu/Zr stresses during sputtering, causing the Ti seed layer to be
necessary.
Table 9: Sputtered Cu/Zr NMMs with varying Cu and Zr layer thicknesses
Cu/Zr sample
name
Ti seed layer
thickness
(nm)
Cu layer
thickness
(nm)
Number of
Cu layers
Zr layer
thickness
(nm)
Number of
Zr layers
Total
thickness
(nm)
Cu/Zr Poly 8 5 5 12 10 13 190
Cu/Zr Poly 9 5 10 6 10 7 130
Cu/Zr Poly 10 5 20 3 10 4 100
Cu/Zr Poly 11 5 2 20 10 21 250
Cu/Zr Poly 12 5 60 1 10 2 80
Cu/Zr Poly 13 5 2 12 10 13 154
Cu/Zr Poly 14 5 10 12 5 12 180
Cu/Zr Poly 15 5 10 6 10 6 120
Cu/Zr Poly 16 5 10 3 20 3 90
Cu/Zr Poly 17 5 10 25 2 25 300
Cu/Zr Poly 18 5 10 1 60 1 70

Table 10 describes a set of Cu/CuZr NMMs (an FCC/amorphous system), with varying Cu layer
thicknesses (Cu/CuZr Poly 6 through Cu/CuZr Poly 11), varying CuZr layer thicknesses (Cu/CuZr
Poly 12 through Cu/CuZr Poly 15), and once again varying Cu layer thicknesses (Cu/CuZr Poly 2
through Cu/CuZr Poly 5). The difference for the last set (Cu/CuZr Poly 2 through Cu/CuZr Poly
5) is that its Cu layers where sputtered at a higher Ar pressure than the other samples, while the Ar
sputtering pressure for Cu layers in Cu/CuZr Poly 6 through Cu/CuZr Poly 15 matched those used
for the Cu/Nb and Cu/Zr samples. The Cu/CuZr system has been studied by tensile testing and
nanopillar compression and has shown notable ductility [76, 174-179]; in-situ XRD tensile testing
should elucidate some of the reasons for the ductility.

108

Table 10: Sputtered Cu/CuZr NMM samples on polyimide substrates with varying Cu and CuZr
layer thicknesses
Sample Name
Cu layer
thickness
(nm)
Number
of Cu
layers
Total Cu
thickness
(nm)
Cu/Zr layer
thickness
(nm)
Number
of CuZr
layers
Total CuZr
thickness
(nm)
Total
thickness
(nm)
Cu/CuZr Poly 6 5 12 60 10 13 130 190
Cu/CuZr Poly 7 10 6 60 10 7 70 130
Cu/CuZr Poly 9 2 25 50 10 26 260 310
Cu/CuZr Poly 10 60 1 60 10 2 20 80
Cu/CuZr Poly 11 20 3 60 10 4 40 100
Cu/CuZr Poly 12 10 6 60 10 6 60 120
Cu/CuZr Poly 13 10 12 120 5 12 60 180
Cu/CuZr Poly 14 10 6 60 20 6 120 180
Cu/CuZr Poly 15 10 20 200 2 20 40 240
Cu/CuZr Poly 2 5 12 60 10 13 130 190
Cu/CuZr Poly 3 2 30 60 10 31 310 370
Cu/CuZr Poly 4 10 6 60 10 7 70 130
Cu/CuZr Poly 5 20 3 60 10 4 40 100

Appendix A.4 Other NMM Systems Sputtered onto Polyimide
In addition to the Cu/Nb, Cu/Zr, and Cu/CuZr systems mentioned in Section 7 and Appendix A.3,
several other samples were sputtered onto polyimide. Al/Nb samples were sputtered, but these
were never tested, because the roughness of the layers that was seen in the TEM study in Section
6 would make interpretation of the stress values more complicated. Future studies of how to reduce
layer roughness would allow for the deformation of the miscible Al/Nb system to be compared to
the immiscible Cu/Nb system.
Several Cu/CuAl samples were also sputtered onto polyimide, with the goal of producing an
FCC/twinned FCC structure, as CuAl has been shown to readily twin when sputtered in bulk [99].
However, when examined by cross-sectional TEM, though some twins were evident, the twins
were not numerous, the layered structure was not apparent, and the sample surface was rough.
More twinning may be induced by increasing the Al sputtering rate/Al concentration, and it may
109

be possible to reduce sample roughness and improve the layered structure by mounting the sample
onto a spinning substrate holder (these samples were sputtered with both the Cu and Al guns
inclined 12° to the sample normal in a co-sputtering configuration). Altering these sputtering
parameters may produce Cu/CuAl samples that are suitable for in-situ XRD testing.
A Cu/CuTa sample was also sputtered with the goal of sputtering an FCC/amorphous sample.
However, the as-sputtered sample presented a cloudy surface. This was attributed to surface
roughness, so the Cu/CuTa system was abandoned in favor of the FCC/amorphous Cu/CuZr
system, which presented a smooth, shiny surface. The Cu/CuTa system may be pursued in the
future by using a spinning substrate holder, which should reduce sample roughness.
Appendix A.5 In-situ XRD Tensile Testing Notes
When capturing in-situ XRD data during tensile testing, the overall thickness of your materials
will greatly affect the testing times. This is because the scattering volume will determine the length
of time necessary to capture a sufficiently defined diffraction pattern. For our scattering volumes
(60 nm total Cu thickness, Section 7), the capture times were up to 75 s per diffraction pattern,
meaning individual data points on the “stress-strain” curves required over a minute to capture in
some cases. Increasing the diffracting volume by sputtering a thicker sample will decrease the
total time to conduct a test, but that may cause the sample to curl up from the residual stresses, as
explained in Appendix A.1.
If the recording times for the diffraction patterns are sufficiently long, then one can analyze not
only the radial distances of the different diffraction rings, but also the relative intensities at
different angles, which show the texture that is developed in the films and indicate the deformation
mechanisms present [139-143]. An example of a diffraction pattern is given in Figure 58, further
110

analysis can show the developed texture, as shown in Figure 49. In order to optimize the use of
beamtime, it may be best to record diffraction patterns with short exposure times during the tensile
tests so that the test can be completed in a reasonable amount of time, and also record diffraction
patterns before and after the tests with longer exposure times, so that the texture can be clearly
seen before and after tensile deformation.

Figure 58: Synchrotron XRD pattern from a multilayered 10 nm Cu/10 nm Nb sample prior to
tensile testing. The Cu (111) and Nb (110) rings are identified. The inner rings are from the
polyimide substrate.

Cu
(111)
Nb
(110)
111

Appendix B. Various Sputtering Notes
Appendix B.1 Variation in Thickness across Substrate Positions
When sputtering, the deposition rate will vary with the inverse square of the distance from the
sputtering gun. Since the substrates onto which we sputter are straight and not curved, the
thickness of the deposited material will vary at different locations on the substrates. This variation
in thickness will be exaggerated at close sputtering distances. As an example of this variation,
Figure 59 is a plot of the thickness of a Cu film which was sputtered at a distance of 152 mm (Cu-
23). The measurements were made over the entire length of the sputtered area, which is
approximately 20 mm long, resulting in about 2 mm between measurement positions, with the
middle positions being from the part of the substrate closest to the sputtering gun.

Figure 59: Variation in sample thickness measured at multiple points along a sputtered 20 mm
diameter Cu film deposited at a sputtering distance of 152 mm.

The reduction in thickness is most evident at the ends of the substrate, with a difference in thickness
of approximately 5% between the middle and the edges. Another substrate was placed in the
sample holder next the substrate measured in Figure 59 and simultaneously sputtered. The average
thickness of that sample, which was located approximately 32 mm off of the axis of the sputtering
gun, was 975 nm. Therefore, the deposition variation that would be seen from the middle of a 75-
112

mm substrate to its outside would be approximately 10% (this will of course depend upon the
deposition conditions). These variations must be considered when one is testing different parts of
a sputtered sample, especially if the sputtered area is large.
Appendix B.2 Substrate Holder for Polyimide Substrates
A new substrate holder was designed for holding the polyimide substrates. The goal was to be
able to sputter samples which could be cut into test strips with dimensions of at least 10 mm X 40
mm. The substrate holder can be seen in Figure 60. As mentioned in Appendix B.1, there will be
a variation in thickness across the substrate, which will be between 5% and 10% for the samples
produced using this substrate holder. Therefore, it is best to perform XRD tensile testing using
strips from as near to the middle of the sample as possible.

Figure 60: Polyimide substrate holder for sputtering - (a) assembled view, (b) exploded view.
The recess cut into the base accommodates a 50 mm x 60 mm polyimide substrate.

Appendix B.3 Checking of Phases by XRD
One of the techniques used to characterize the samples was XRD. One of the initial goals of this
project was to fabricate and mechanically test Cu with a BCC crystal structure (the native crystal
structure of Cu is FCC). Therefore, XRD was used to determine which crystal structures were
present in Cu/Nb NMMs with varying Cu layer thicknesses. XRD curves from Cu/Nb NMM
samples with varying Cu layer thicknesses and constant 100 nm Nb layer thicknesses are shown
113

in Figure 61 and Figure 62 (these XRD curves are from the samples Cu/Nb 29, 31, 33, 35, 39, 41,
43, 45, 47, 49, and 51; Si peaks from the substrate has been removed by flattening the 32°-35°,
60°-80°, and sometimes the 115°-118° regions of the XRD curves – see Figure 74). These two
figures are for samples with slow (0.16 nm/s) and fast (0.70 nm/s) Cu sputtering rates, while the
Nb sputtering rate was held constant (0.12 nm/s). With decreasing Cu layer thickness, the Cu
peaks decrease in intensity, until they are no longer seen for the samples with 2 and 1 nm Cu layers.
While this may indicate a change in Cu crystal structure to match the BCC Nb crystal as seen in
other studies [111-113, 180, 181], it is also possible that the Cu scattering volume is simply too
small to result in strong Cu XRD peaks (total Cu thickness < 25 nm). Therefore, further
characterization, such as HRTEM, would be necessary to determine whether BCC Cu is truly
present.

Figure 61: XRD plots for as-sputtered Cu/Nb samples on Si (100) substrates with Nb and Cu
sputtering rates of approximately 0.12 nm/s and 0.16 nm/s, respectively. The Nb layer
thicknesses are 100 nm and the Cu layer thicknesses are given on the right for the different
samples. Cu and Nb peaks are indicated. Si peaks were removed from the curves.
114

Figure 62: XRD plots for as-sputtered Cu/Nb samples on Si (100) substrates with Nb and Cu
sputtering rates of approximately 0.12 nm/s and 0.70 nm/s, respectively. The Nb layer
thicknesses are 100 nm and the Cu layer thicknesses are given on the right for the different
samples. Cu and Nb peaks are indicated. Si peaks were removed from most of the curves.

NMM samples consisting of Ta/Cu were also sputtered, in which the Cu was not expected to
change crystal structure. The goal of these samples was to determine whether a small scattering
volume of 1 nm Cu layers would result in discernible XRD peaks (40 nm total Cu thickness). If it
did, then the lack of Cu peaks for the samples with 2 nm and 1 nm Cu layers in Figure 61 and
Figure 62 could perhaps be attributed to a change in Cu crystal structure, rather than being due
simply to the small scattering volume.
The expectation was that the Ta would deposit as beta-Ta, which, when sputtered, generally does
not have XRD peaks which are near to the Cu (111) peak at 43.3°, thus allowing for the Cu peak
to be discerned even if it is very small [88]. However, the sample (Ta-1) sputtered as alpha-Ta
instead, Figure 63, which has a strong peak at 38.5° that can overlap with the Cu (111) peak and
make it indiscernible if the Cu peak is small. To eliminate the alpha phase, Ta was sputtered with
a poor starting base pressure (3x10
-2
Pa), which has been shown to favor beta-Ta growth, and this
did in fact produce beta-Ta when interrupted sputtering was used on a glass substrate, Figure 63
115

[88]. However, when similar conditions were used to sputter Cu/Ta multilayers, alpha-Ta was still
formed, making the presence of small Cu (111) peaks indiscernible, Figure 64. Therefore, it was
not clear from the Cu/Ta samples whether a scattering volume of 40 nm total Cu thickness will
produce a discernible Cu XRD peak. Peak deconvolution software may aid in the identification
of the Cu peak, but there is the added complication of multiple possible beta-Ta peaks in the 35-
43° range, though, as pointed out above, these peaks are not generally seen in sputtered Ta.

Figure 63: XRD curves for sputtered pure Ta samples. The sample names, starting base
pressures, substrate types, and interruption during sputtering are indicated on the right for the
different curves. The relevant alpha and beta Ta peaks are identified. Si peaks have been
removed for Ta-2.
116

Figure 64: XRD curves for sputtered 1 nm Cu/10 nm Ta NMM samples. The sample names,
starting base pressures, and substrate types are indicated on the right for the different curves.
The relevant alpha and beta Ta peaks are identified. Si peaks have been removed for Cu-Ta 1
and Cu-Ta 3.

Appendix B.4 Sputtering Amorphous Alloys
Sputtering has been shown to be a route for producing amorphous materials, especially when
sputtering immiscible elements. Two alloys were sputtered which have been reported to deposit
in an amorphous state - Ta-20at.%Cu and Cu-40at.%Zr [108, 174, 182-185]. Their amorphous
structure was confirmed by XRD, Figure 65, where the CuZr sample (a) is fully amorphous, and
the Ta-Cu sample (b) may have some crystallites and/or short range order. These amorphous
materials were used in the sputtering of multilayer samples (Appendix A.3 and Appendix A.4).
117

Figure 65: XRD patterns showing the amorphous structures of co-sputtered samples (a) CuZr-1
and (b) TaCu-15.

118

Appendix C. Residual Stresses
Residual stresses in the sputtered samples were determined by measuring the profiles of the Si
substrates before and after sputtering using an XP-2 profilometer (AMBiOS). The deflections of
the substrates due to sputtering were then input into a modified Stoney’s equation to determine the
residual stresses in the films [186-188]. Several sample stresses from Cu/Nb and Al/Nb
multilayers films are shown in Table 11 and Table 12, respectively. These films have constant Nb
layer thicknesses of 100 nm and varying Cu and Al layer thicknesses with fast and slow sputtering
rates (slow deposition rates – Cu: 0.15 nm/s, Al: 0.07 nm/s, fast deposition rates – Cu: 0.69 nm/s,
Al: 0.29 nm/s, Nb deposition rate - 0.12 nm/s).
Positive stress values are defined as being tensile. For the Cu/Nb films, Table 11, the stresses for
the faster Cu sputtering rate are higher than those for slower Cu sputtering rate, except for the
samples with 20 nm Cu layer thicknesses. These values are for the odd-numbered Cu/Nb samples
between Cu/Nb 39 and Cu/Nb 59.
Table 11: Residual stresses in sputtered Cu/Nb NMMs*
Fast Cu sputtering rate
Slow Cu sputtering
rate
Cu layer
thickness
(nm)
Cu
sputtering
rate (nm/s)
Residual
stress
(MPa)
Cu
sputtering
rate
(nm/s)
Residual
stress
(MPa)
1 0.15 360 0.69 560
2 0.15 300 0.69 530
5 0.15 390 0.69 610
10 0.15 330 0.69 630
20 0.15 540 0.69 420
*all Nb layer thicknesses are 100 nm
The Al/Nb samples shown in Table 12 are for the odd-numbered samples between Al/Nb 3 and
Al/Nb 23. The residual stress values for the samples with fast and slow Al sputtering rates are
119

relatively similar, with the biggest difference seen for the samples with 2 nm thick Al layers. While
all of these multilayer samples showed substantial tensile stresses (>300 MPa), this will change
with sputtering parameters, as explained in Appendix A.1.
Table 12: Residual stresses in sputtered Al/Nb NMMs*
Fast Al sputtering rate Slow Al sputtering rate
Al layer
thickness
(nm)
Al
sputtering
rate (nm/s)
Residual
stress
(MPa)
Al
sputtering
rate
(nm/s)
Residual
stress
(MPa)
1 0.07 580 0.29 550
2 0.07 520 0.29 420
5 0.07 510 0.29 460
10 0.07 380 0.29 340
20 0.07 390 0.29 410
*all Nb layer thicknesses are 100 nm
The stresses of monolithic films change with overall thickness, as shown below for Cu, Nb, and
Ti films, Table 13. These monolithic films were sputtered onto Si substrates at 100 W, 0.67 Pa Ar
pressure, and a sputtering distance of 152 mm. There is a wide range of stresses, from a tensile
stress of 517 MPa to a compressive stress of 124 MPa. Thus, sputtering parameters, layer
thicknesses, overall film thicknesses, and the type of material were all observed to significantly
affect the residual stresses in this study.
Table 13: Residual stresses in monolithic films
Cu Nb Ti
Coating
Thickness
(nm)
Average
Stress
(MPa)
Coating
Thickness
(nm)
Average
Stress
(MPa)
Coating
Thickness
(nm)
Average
Stress
(MPa)
47 298 60 517 35 58
102 116 98 161 90 70
191 90 198 161 188 0
474 100 488 172 453 -107
1013 393 886 -124

120

Appendix D. TEM Imaging of NMMs
TEM imaging of NMMs can yield much valuable information: grain orientation, texture, grain
size, layer epitaxy, elemental distribution, and layer thickness, among other things. This section
presents several specific uses of TEM for NMM characterization to complement those presented
in Sections 3.2.1, 5.2, 0, 0, and 0.
Figure 66 is a HRTEM image of a 100 nm Nb/2 nm Cu NMM and the Cu layer is outlined by
white lines. Based on the directions of the Nb lattice fringes above and below the Cu layer, it is
evident that the Nb grain orientation has changed. Thus, the Cu layers interrupted the transmission
of grain orientation information.

Figure 66: TEM image of 100 nm Nb/2 nm Cu NMM. The growth direction is vertical. The Cu
layer location is shown by the white lines. The Nb grain orientations are changed by the Cu
layer.

On the other hand, in Figure 67 (100 nm Nb/5 nm Cu), the direction of the Nb lattice fringes above
and below the Cu layer remain the same (the Cu layer is delineated by the white lines). When
performing lattice spacing measurements to determine which materials are present, multiple lattice
121

fringes (e.g. 10 planes) were measured together and their total spacing divided among them, see
arrows in Figure 67, rather than attempting to measure individual lattice fringe spacings. The
uninterrupted direction of the Nb lattice fringes indicates that the Nb has maintained its orientation
through the Cu layer, and further TEM imaging by HRTEM can be utilized to ascertain whether
there is an epitaxial relationship between these layers.

Figure 67: TEM image of 100 nm Nb/5 nm Cu NMM. The growth direction is vertical. The Cu
layer is roughly delineated by the white lines. The spacing of ten lattice planes is shown for Nb
at the top and Cu in the middle.

In some cases, the lattice fringes are not as clear and aligned as they are in Figure 67. For example,
in Figure 68a (100 nm Nb/5 nm Cu) the Cu layer is present but not easily distinguishable.
However, since the lattice fringes are at least somewhat visible, image processing can be used to
separate the Cu regions. This is done by performing an FFT of the image, selecting the FFT rings
that correspond to Cu lattice spacings, and performing an inverse FFT to highlight the Cu regions.
This is shown in Figure 68b, where the Cu layer has a visible periodicity, while the Nb regions
above and below have a somewhat random mosaic pattern.
122

Figure 68: TEM images of a 100 nm Nb/5 nm Cu NMM, with a Cu layer in the middle. The
growth direction is vertical. Image (a) is the raw TEM image. That image was processed by FFT,
the Cu rings were isolated, and an inverse FFT was performed, and this is shown in (b). The Cu
layer shows periodicity, while the Nb regions have a random mosaic pattern.

One major consideration when imaging NMMs, especially by HRTEM, is that layer
waviness/roughness should be minimized during the fabrication. This is because TEM imaging is
always in projection and roughness results in overlap of the different layers during imaging.
Especially for thin layers, even a small amount of roughness can result in the different layers being
indistinguishable.
Overall, TEM can yield much valuable information for NMM investigations, as demonstrated in
both this section and the other studies presented in this dissertation.

123

Appendix E. SEM and FIB Imaging of NMMs
The resolution of SEM (several nm) is on the order of the layer thicknesses used for NMMs.
Though this may limit the information which can be obtained by SEM, there is still some data
which can be gathered.
Appendix E.1 Cross-sectional
NMM samples can be fractured in order to image their cross-sections. If the layer thicknesses are
large (> 20 nm), then they may be evident in the SEM. In addition, the fracture surface may
provide an indication of ductility of the sample. For example, the two Cu/Nb NMMs in Figure
69a and b have identical nominal layer thicknesses, but the fracture surface of Figure 69b shows
much more dimpling and indicates some plasticity.

Figure 69: SEM images of fractured cross-sections of as-sputtered 20 nm Cu/100 nm Nb NMM
samples, (a) Cu/Nb 29 and (b) Cu/Nb 43. The fracture surface of (a) appears to be more brittle
that that of (b).

The cross-sections of NMMs with relatively thick (> 20 nm) layers can also be imaged by FIB,
such as the Cu/Nb NMM shown in Figure 70. However, the resolution of the FIB is on the order
of tens of nm, so the information which can be gained is likely to be rather limited. In addition,
124

especially in the case of light elements, imaging will quickly damage the sample, as is evident in
Figure 70b, where the damage from capturing multiple higher magnification images is clear.

Figure 70: (a) FIB cross-sectional images of an as-sputtered 20 nm Cu/100 nm Nb NMM sample
(Cu/Nb 1). The SEM image (b) shows damage incurred during FIB imaging (that cross-sectional
surface was flat prior to FIB imaging).

Appendix E.2 Topographical
The surfaces of sputtered samples can also be imaged by SEM, though the use of such images is
mostly qualitative. That is, a large change in the surface morphology may indicate that the sample
has experienced a substantial change in structure that could be further investigated by TEM or
some other method. This is shown in Figure 71, where the surfaces of Hf-Ti samples co-sputtered
under different conditions have different morphologies. In the case of NMMs on polyimide, SEM
can provide images of the crack propagation through the samples, which will provide an indication
of ductility, see Figure 48.
125

Figure 71: SEM images of the surfaces of two co-sputtered Hf-Ti samples, (a) Hf-Ti 5 and (b) Hf-
Ti 9.

In the case of Hf-Ti imaging, the e-beam often quickly damaged the sample surface, as evidenced
by the multiple dark boxes in Figure 72 that are due to higher magnification imaging. It is not
clear why this happens, but it may be due to a buildup of oxides or organics.

Figure 72: SEM images of the surface of a co-sputtered Hf-Ti sample (Hf-Ti 5) showing damage
produced during imaging.

Some samples showed small dots on their surfaces under SEM imaging, as shown in Figure 73 for
a heat-treated Hf-Ti sample (a) and an as-sputtered Ta-Cu sample (b). These were not identified
in the present study, because these samples were not used for further testing. However, the use of
SEM imaging to identify similar NMM surface features in future investigations can give an
indication of oxide or precipitate growth.
126

Figure 73: Two surface SEM images of (a) a heat-treated Hf-Ti sample (Hf-Ti 9 HT2) and (b) an
as-sputtered Ta-Cu sample (Ta-Cu 5). The small dots appear to be some form of oxide or
precipitate.

127

Appendix F. Additional XRD Investigations
In this section, XRD curves for several additional sample sets are presented (see Appendix B.3 for
other XRD data). The XRD scans in Figure 74 are for single crystal Si (100) substrates. The
presence of the double peak around 117° was found to depend upon the orientation of the Si wafer
during the XRD scan. This is because the substrates are single crystals, and the presence of certain
reflections will depend upon orientation. For XRD scans of thin films on Si (100) substrates, the
peaks that appear in Figure 74 can be attributed to the Si substrate.

Figure 74: XRD scans of Si (100) substrates at different orientations.

XRD plots for monolithic Ti films sputtered onto Si (100) substrates are given in Figure 75. The
goal of these samples was to check whether Ti deposited in its native hexagonal close-packed
(HCP) crystal structure at a variety of thicknesses, before we attempted to induce crystal structure
changes by use of Ti/Al multilayers. These samples were sputtered using the following
parameters: 100 W, 0.67 Pa Ar pressure, and 152 mm sputtering distance, while the nominal
thicknesses are given on the right of Figure 75. Since the focus is on the Ti, the peaks
corresponding to the Si substrate, see Figure 74, will be ignored. The 38.5° and 82° peaks correlate
to Ti peaks. The peaks at 53°, 56.5°, and 94° appear to be shifted from the standard Ti peaks. The
128

peaks at 40° may be from TiO2. The peaks at 114° do not match Ti or Ti oxide peaks. Overall,
the XRD peak positions do not match well with titanium or titanium oxide.
A study by Chakraborty et al. found that sputtered Ti can form a combination of FCC and HCP
phases in films up to a few hundred nanometers in thickness [189]. Since the goals of our study
were to investigate the effect of layering Ti and Al as an NMM and determine possible crystal
structure changes, the fact that Ti can change crystal structure with just a change in thickness
would complicate the interpretation of the effect using multilayers. Therefore, the Ti/Al NMM
system was not further studied.

Figure 75: XRD curves for sputtered monolithic Ti films. The sample names and nominal
thicknesses are given on the right.

Another system of interest with regard to crystal structure or phase changes with decreasing NMM
layer thickness is Cu/Nb [111-113]. Therefore, the presence of phase changes in pure Nb and Cu
samples with varying layer thicknesses was investigated. Figure 76 is a collection of XRD curves
for Nb samples sputtered using the following parameters: 100 W, 0.67 Pa Ar pressure, and 152
mm sputtering distance, with the nominal thicknesses given on the right. The 38.5° and 55° peaks
match Nb. The 82° peaks match Nb well, except for Nb-3, where that peak is slightly shifted to
the left. The Nb-3 107° peak is correct, while the Nb-5 109° peak is shifted from 107°, which may
129

be due to residual stresses. Both the Nb-3 and Nb-5 plots indicate the presence of a 121° Nb peak
just past the scan range. There is a low-angle 14° peak for Nb-3 and Nb-5 which does not match
Nb or its oxides. Overall, the XRD peaks of the sputtered Nb samples match the expected BCC
Nb peak positions, indicating that there were no phase changes in the Nb with decreasing layer
thickness, down to approximately 60 nm.

Figure 76: XRD scans of sputtered monolithic Nb films. The samples names and nominal
thicknesses are given on the right.

Figure 77 contains XRD plots of Cu samples sputtered using the following parameters: 100 W,
0.67 Pa Ar pressure, and 152 mm sputtering distance, with the nominal thicknesses given on the
right. The peaks in the plots all match standard Cu peaks (ignoring the Si peaks from the substrate).
Since neither Cu nor Nb showed phase changes with decreasing monolithic film thickness, the
Cu/Nb system is a good system for studying the effects of varying layer thickness on the phases in
the materials.
130

Figure 77: XRD scans of sputtered monolithic Cu films. The sample names and nominal
thicknesses are given on the right.

XRD plots from Al/Nb samples sputtered with slow (0.07 nm/s) and fast (0.29 nm/s) Al sputtering
rates are given in Figure 78 (samples Al/Nb 3, 5, 7, 9, 11, 13, 15, 17, 19, 21). Si peaks from the
substrate have been removed by flattening the 32°-35°, 60°-80°, and sometimes the 115°-118°
regions of the XRD curves – see Figure 74. All of the remaining peaks match Nb (the samples are
greater than 80 vol.% Nb), and since the main Al peak overlaps with the main Nb peak at
approximately 38.5°, it is not clear whether any of the peaks are due to the Al layers.

Figure 78: XRD scans of multilayered Al/Nb samples. The Nb layer thicknesses are 100 nm, and
the Al layer thicknesses and relative Al sputtering rates (fast = 0.29 nm/s, slow = 0.07 nm/s) are
indicated on the right of each plot. The curves have been flattened from 60°-80° to remove the
large Si (4 0 0) peak.

131

Appendix G. Co-sputtering Investigation
Two main methods were used to co-sputter different materials. The first used the previously
existing flanges in the sputtering chamber (Figure 23) and the second used a specially designed
co-sputtering flange (Figure 24).
Appendix G.1 Existing Flanges
The possible gun positions for our existing chamber are shown schematically in Figure 79a.
Different configurations were tested, with the goal of being able to co-sputter multiple elements,
with a significant amount of the minor element (up to 30 at.%).

Figure 79: Schematic of the existing sputtering configuration. All four possible sputtering gun
positions are shown in (a), with the substrate holder in the middle. The substrate holder can be
made to face any of the four guns. Two guns are shown in (b) and they are each inclined 45° to
the substrate holder.

One set of sputtering investigations were conducted with the substrate not pointed directly at the
sputtering gun. For example, Cu was sputtered from gun position 1 while the substrate was facing
position 3, so the substrate face was facing away from the sputtering direction, Figure 79a. In this
configuration, the deposition rate relative to direct sputtering varied from 2% to 10% for different
sputtering conditions (samples Cu-47 through Cu-52). The deposition rate was also tested using
gun position 1 while the substrate was facing position 2, with the substrate surface parallel to the
132

sputtering direction. The deposition rates relative to direct sputtering in this case varied from
approximately 10% to 25% (samples Cu-37 through Cu-46, Cu-53, Cu-54). Both when the
substrates faced away from the sputtering direction and when they were parallel to the sputtering
direction, higher Ar pressures resulted in significantly higher deposition rates. However, even at
high Ar pressures, the deposition rates were relatively small, and much of the sputtered material
was wasted by being deposited onto the sputtering chamber rather than the substrate. Therefore,
it was decided that co-sputtering would be conducted by sputtering from positions 1 and 2, with
the substrate inclined 45° to both of them, see Figure 79b.
Before co-sputtering in this configuration, the difference in film thickness across the substrate was
measured for sputtering with one material. A significant thickness difference was expected,
because, when sputtering from position 1, the far side of the substrate is ~15%-30% farther away
from the sputtering gun than the near side of the substrate (this depends upon the sputtering
distance and the substrate size). Since the deposition rate scales with the inverse of the square of
distance, the expected differences in thickness are on the order of 25% for a sputtering distance of
127 mm and a 25 mm substrate. In fact, the measured variation in thickness across a 25 mm
substrate for a pure Hf sample sputtered at a distance of 127 mm was approximately 35% (sample
Hf-3).
Individual sputtering rates were determined for several materials with the substrate at a 45° incline.
These sputtering rates were then used to calculate the sputtering conditions which should give the
desired compositions. However, the actual compositions that were achieved using this method
were significantly different from those expected. For example, the sputtering conditions for Ta
and Cu that were expected to result in a composition of Ta-20at.%Cu instead resulted in a film
where no Cu signal was detected by EDX (sample TaCu-1). This is likely due to the higher mass
133

and higher energy of the Ta atoms than the Cu atoms (higher atomic number and higher melting
temperature), which may cause the Cu atoms to evaporate or be sputtered off of the substrates
during the Ta deposition [190].
Sputtering conditions were eventually determined that resulted in the desired average
compositions, but there were significant composition variations across the samples, due to the fact
that different ends of the substrates were closer to different sputtering guns. Among the material
systems deposited in this manner, the largest composition variation was seen for Ta-Cu samples,
e.g. the atomic percentage of Cu varied from 4 to 40 across a 25 mm substrate (sample Ta-Cu 9).
In summary, using this method of inclining the substrate 45° to both guns allowed for the
fabrication of co-sputtered samples with the desired compositions. However, there was a large
variation in composition across the co-sputtered samples.
Appendix G.2 Co-sputtering Flange
The large variation in composition was addressed by designing a specialized flange which allowed
for co-sputtering of multiple materials at much smaller angles (~12°), Figure 24. The smaller
angles resulted in the different sides of the substrates being at similar distances from the guns,
giving a more uniform film thickness for single-element films and more uniform compositions for
multi-element films. As an example of the decreased variability in composition across the sample,
a Ta-Cu sample (Ta-Cu 15) was sputtered with a variation in composition of 19-23at.%Cu, which
is much lower than the 4-40at.%Cu mentioned in Appendix G.1. An additional benefit of the co-
sputtering flange is that the actual sputtering conditions necessary to achieve the desired
compositions were closer to those calculated based on the relative sputtering rates of the individual
134

materials. The next step to achieve better film uniformity would be installation of a rotating
substrate holder.

135

Appendix H. Additional Thermal Stability Research
In addition to the thermal stability investigation that was described in Section 8, several other
points of note with regard to thermal stability study will be explained in this section.
Appendix H.1 Oxidation Considerations
When heat-treating samples, oxidation must be considered, since adsorbed oxygen can diffuse into
the samples at elevated temperatures. This is especially relevant for thin films, where the entire
thickness of the sample is relatively close to the surface. Other reactions are also possible, such
as nitriding and carburizing, but this section will focus on oxidation.
Several different atmospheric conditions can be used to reduce oxidation in the furnace. One is a
flowing ultra-pure argon atmosphere, where the purity of the argon will limit the degree of
oxidation prevention. A second is a reducing atmosphere, which reacts with oxygen and reduces
the oxygen concentration. However, the elements in the atmosphere should not be reactive with
the heat-treated material. For example, a reducing hydrogen atmosphere will react with Ti, Zr, Hf,
Ta, and Nb, among other elements, at high temperatures [191]. A third possible atmospheric
condition is a vacuum, where the concentrations of all the gases are reduced.
The necessary vacuum levels and temperatures to prevent oxidation will vary for different
materials, as shown in Figure 80 [191]. This chart makes clear that prevention of oxidation in, for
example, Ti will require much higher temperatures and/or better vacuum levels than prevention of
oxidation in Mo. Such charts and Ellingham diagrams make clear the thermodynamically
necessary conditions to prevent oxidation.
136

Figure 80: Chart showing necessary temperatures and partial pressure of water vapor
necessary to prevent oxidation. [191]

In addition to the thermodynamic prevention of oxidation, oxidation can be minimized kinetically.
If the vacuum level is sufficient, the time for oxides to form on the surface and diffuse into the
material may be sufficiently long that, during the course of the heat treatment, only a small portion
of the sample will oxidize. Surface monolayer formation time equations must be used, and they
generally state that the monolayer formation time will be on the order of seconds for a vacuum of
1E-3 Pa and on the order of thousands of seconds for a vacuum of 1E-6 Pa [192].
Appendix H.2 Removing Samples from Substrates
When samples are sputtered as thin films, subsequent heat treatments may result in diffusion of
the substrate material into the thin film, though significant diffusion may not occur if the materials
are immiscible or the heat treatment time is short and/or the temperature is low. To prevent
substrate diffusion, it is generally necessary to remove the samples from the substrates. Some
137

films will peel off immediately upon removal from the sputtering chamber, especially if there is
poor bonding between the film and substrate or the film has high residual stresses. If the film does
not peel off by itself, a brief high temperature heat treatment (e.g. reach 600 °C and immediately
cool) may result in the film peeling off of the substrate. However, if a brief heat treatment under
vacuum is used, it must be kept in mind that the films will be unsecured after the heat treatment.
Therefore, venting of the vacuum furnace must be performed slowly, or else the samples may be
blown away.
After removal of the film from the substrate, it may be necessary to polish or etch any substrate
residue from the bottom of the film, since even 50 nm of substrate material may significantly affect
the composition of a film that is only a few μm thick. In the study in Section 8, Si substrate etching
was performed by submerging the samples into an actively stirred solution of 10.1 mol KOH
solution (using purified de-ionized water) at 85 °C for 120 seconds.
Appendix H.3 Quench Vacuum Furnace Specifics
In order to conduct heat treatments and quenching under vacuum, a vertical tube furnace
arrangement was used, as shown in Figure 81. The film samples were placed into quartz boats,
the boats were wrapped with NiCr wire in order to secure the films, and the boats were suspended
in the furnace by a thicker NiCr wire. The quartz boats have a loop on either end, which aids with
both wrapping with the thinner NiCr wire and suspending from the thicker NiCr wire. NiCr wire
was used because it has relatively high melting temperatures and does not significantly evaporate
or weaken at the temperatures used in these studies. Several pieces of iron were attached to the
top of the thick NiCr wire and these were held up by an external magnet.
138

Figure 81: Schematic of quench vacuum furnace. (not to scale)

The sample was placed in a long quartz tube that could be pumped down to achieve vacuum, and
a glass jar of low vapor pressure oil (Invoil 705, Inland Vacuum Industries) was placed in the
bottom of the tube for quenching, far enough away from the furnace to avoid heating (which would
result in oil evaporation and degradation of the vacuum level). After the completion of the heat
treatment, the external magnet was removed, which caused the quartz boat with the films to fall
into the quenching oil. A long tube is necessary to avoid heating either the quenching oil (which
increases its evaporation rate) or the magnet (which would degrade its magnetism).
Appendix H.4 Additional Heat-Treated Systems
In addition to Hf-Ti (Section 8), the thermal stability of several other systems was also investigated
- Hf, Ti, Ta-Hf, Ti-Hf, Ta-Cu, and Mo-Au. The alloy systems were selected due to their expected
139

nanostructure stability (Ta-Hf, Ta-Cu, Mo-Au) or bulk stability (Ti-Hf) [32]. However, these
systems presented difficulties, as explained below.
Hf
Multiple Hf samples were sputtered and heat-treated in order to measure their grain growth and
compare it to that of Hf-Ti samples. However, they all showed nanoscale cracking when imaged
by TEM, Figure 82 (Hf-10 HT6 sample - 800 °C, 4 days). This is likely due to both the residual
stresses from sputtering and some oxidation of the sample, since oxidation will generally make a
material more brittle.

Figure 82: Cross-sectional TEM image of a heat-treated sputtered Hf sample. The bright areas
are cracks.

Ti
The thermal stability of sputtered Ti samples was investigated. However, the Ti samples remained
attached to the substrates and did not peel off after a brief heat treatment (see Appendix H.2). This
was also observed for Ti-Hf samples, and some possible solutions to this problem are explained in
that section below.
140

Ta-Hf
Ta-Hf samples were co-sputtered as monolithic films and several of these samples were
subsequently heat-treated. However, the heat-treated Ta-Hf samples would crack and some would
turn to powder during heat-treatment. This may be an exaggerated version of the nanoscale
cracking seen for heat-treated Hf, and may occur for similar reasons.
Ti-Hf
Ti-Hf samples (> 50 at.% Ti) were co-sputtered as monolithic films and several of these samples
were subsequently heat-treated. Some of these films had the issue that, after the brief heat
treatment mentioned in Appendix H.2 for the purpose of removing the films from the substrates,
the Ti-Hf samples often remained attached to the substrates, meaning that the Ti bonded well to
the Si substrate. This would result in a large amount of diffusion of the Si substrate into the sample
during any subsequent heat treatment. One solution may be to sputter a thin (< 50 nm) layer of Hf
onto the Si prior to depositing the Ti-Hf film. In this manner, the samples should peel off of the
substrates after a brief heat treatment, and the thin layer of Hf should not greatly affect the global
composition.
Another solution for the strong adhesion of Ti-Hf to the Si substrates is to etch the entire Si
substrate (see Appendix H.2). However, this takes a lot of time and etchant and may still not result
in a complete etch of the Si substrates. Selecting sputtering conditions which produce high residual
stresses may also aid in removal of the films from the substrates. Finally, using higher
temperatures during the brief heat treatment aided in film removal for certain samples.
Ta-Cu
The Ta-Cu sample that was subjected to a heat treatment of 800 °C for 4 days under vacuum (Ta-
Cu 15, ~21 at% Cu) showed a sharp decrease in Cu composition, such that Cu was not detected at
141

some points on the sample when scanned by EDX in the SEM. EDX in the TEM, on the other
hand, detected about 18 at.% Cu. However, the detected Cu may not be from the sample itself,
because, even though a Mo TEM grid was used in order to not generate Cu signal from the grid,
there are other Cu pieces in the TEM, such as the TEM grid clip and the pole piece, which can
give spurious Cu signal. For example, the Hf-Ti 17 HT11 sample was attached to the same Mo
grid as Ta-Cu15 HT11 and EDX scans detected ~5 at.% Cu in the Hf-Ti sample, even though
there is no Cu in either the sample or the Mo grid. Therefore, some of the Cu that was detected in
the Ta-Cu sample may in fact be from the TEM grid clip, and that would explain the discrepancy
in Cu signal between the SEM and TEM EDX scans.
The reason for the reduced Cu composition after heat treatment may be due to the vapor pressure
of Cu. At a temperature of 800 °C, the vapor pressure of Cu is approximately 2.3E-5 Pa, which is
only about an order of magnitude better than the lowest pressure during the heat treatment (~3E-4
Pa) [193]. Therefore, it is likely that the Cu was evaporating during the heat treatment and
reducing the Cu composition of the Ta-Cu samples. This may be addressed by conducting heat
treatments in poorer vacuum or in Ar. However, the oxidation of the samples may be increased if
they are not under good vacuum. As an added consideration, a balance must be struck between
selecting heat treatment times and temperatures which are long and hot enough to result in
significant diffusion of the Ta (melting temperature 3,020 °C), while minimizing the evaporation
of the Cu. The evaporation of Cu may be sidestepped by sputtering samples which have a higher
concentration of Cu (30-35 at.%), which would decrease during heat treatment to the desired
compositions.
A co-sputtered and subsequently heat-treated monolithic (not multilayered) Ta-Cu sample was
imaged by TEM (Ta-Cu15 HT11), and a STEM image is shown in Figure 83. In this figure,
142

nanoscale features (below 100 nm) are evident. However, the structure and composition of these
features is not yet clear. An EDX map of the sample did not clearly differentiate between Ta-rich
and Cu-rich areas. This may be due to the thickness of the area that was scanned, which can result
in projection of Ta-rich and Cu-rich regions. Therefore, an EDX scan of a thinner region should
more clearly delineate regions with different compositions (keep in mind that Ta and Cu are
immiscible and are expected to segregate). In addition, line scan EDX composition profiles using
long dwell times may reveal a variation in composition. DF-TEM imaging should reveal whether
the preferred orientation of the as-sputtered state has been retained. Finally, the EDX map detected
a significant amount of Si (~ 10 at.%, 2 wt.%). This is from the substrate, which was not etched
in this case, and any future testing of this system should include etching of residual Si after removal
of the film from the substrate (see Appendix H.2).

Figure 83: Cross-sectional STEM image of a heat-treated co-sputtered Ta-Cu sample (800 °C, 4
days). There are sub-100 nm features, but the composition of those features is not yet known.

Overall, the Ta-Cu system is promising for nanostability, as predicted by the model in Ref. [32].
However, the considerations explained in the preceding paragraphs must be addressed, and a
multilayered geometry may be preferable, as was found for Hf-Ti in Section 8.
143

Mo-Au
A Mo-Au sample was sputtered in a multilayer configuration consisting nominally of layers of 2
nm Au and 20 nm Mo-10 at.% Au (Mo-Au 9). This sample was subsequently heat-treated, and
two TEM images of the heat-treated sample are shown in Figure 84, with (a) taken from the top of
the imaged area and (b) from the bottom. The growth direction is vertical, but it is not known
whether the top of images corresponds to the top or bottom of the original film (the samples are
thin films, and keeping track of orientation during heat treatment and quenching is not
straightforward, though it can be done if needed). In Figure 84a, the large grains are evident. On
the other hand, Figure 84b shows smaller grains near the bottom, with some retention of the
multilayered structure at the very bottom of the image.

Figure 84: Cross-sectional TEM image of a heat-treated multilayered Mo-Au sample (800 °C, 4
days). The different views, (a) and (b), are from different sides of the sample. Note that (a)
shows much larger grains, while (b) shows a partial retention of the multilayer structure at the
bottom of the image. The growth direction is vertical.

It is likely that this region with a retained multilayer structure corresponds to the bottom of the
original sputtered sample (next to the Si substrate), because it has previously been reported that
NMMs will generally increase in roughness with increasing thickness, and this roughness
corresponds to less thermal stability [84, 194-197]. Therefore, it is possible that the first part of
the sample was sputtered with a smaller roughness, and was better able to preserve the multilayered
144

structure, while the part of the sample that was sputtered later had more roughness and was more
prone to grain growth. However, due to time constraints, the as-sputtered Mo-Au sample was not
imaged by TEM. This would be worth investigating in the future. In addition, sputtering a
multilayer Mo-Au sample with the co-sputtering flange and a rotating substrate holder (which was
not available for the sample shown in Figure 84) may decrease the roughness and improve the
thermal stability. Overall, multilayered Mo-Au shows some promise with regard to thermal
stability and should be studied further. On a final note, it is best to sputter Mo-Au samples with
Au as the initial layer, because Au does not form compounds with Si, thus allowing for easy
removal of the Si residue from the Mo-Au film by etching.

145

Appendix I. List of Samples
This appendix contains tables with sample descriptions of sputtered films produced in the course
of this study. The layer thicknesses are assumed to vary by ±10% as is typical for sputtered
samples. The unit mTorr is used for Ar pressure and is equivalent to 0.133 Pa.
Appendix I.1 Monolithic Single-Element Films on Si and Glass Substrates
Table 14: Monolithic Al samples
Al
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Al-1 thick glass 250 152 100 5 62 p
Al-2 thick glass 251 152 250 5 14 p
Al-3 thick glass 394 152 50 5 90 p
Al-4 thick glass 401 152 250 5 22 p
Al-5 thick glass 2804 81 400 3 15 p
Al-6 thick glass 3063 81 400 3 15 p
Al-7 thick glass 3312 117 400 3 30 p
Al-8 thick glass 763 51 50 5 25 p
Al-9 thick glass 1341 51 50 5 50 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 15: Monolithic Au samples
Au
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed
Au-1 Si 904 51 50 5 15 profilometry

Table 16: Monolithic Al-5456 samples
Al-5456
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed
Al-5456 1 Si 76 65 2 52 profilometry
Al-5456 2 glass 1425 76 65 2 52 profilometry
Al-5456 3 Si 76 65 2 52 profilometry
Al-5456 4 glass 1404 76 65 2 52 profilometry
146

Table 17: Monolithic Cu samples
Cu
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Cu-1 Si 182 152 50 5 21 p
Cu-2 Si 278 152 100 5 16 p
Cu-5 152 100 5 29 x, p
Cu-6 Si 474 152 100 5 29 p
Cu-7 Si 152 100 5 12 x, p
Cu-8 Si 191 152 100 5 12 p
Cu-9 Si 152 100 5 6.5 x, p
Cu-10 Si 102 152 100 5 6.5 p
Cu-11 Si 152 100 5 3 x, p
Cu-12 Si 47 152 100 5 3 p
Cu-13 Si 90 241 50 5 47 p
Cu-14 Si 113 241 30 1 50 p
Cu-15 Si 70 241 30 5 60 p
Cu-16 Si 219 241 50 1 60 p
Cu-17 Si 17 241 100 36 30 p
Cu-18 Si 20 152 50 5 2.38 p
Cu-19 glass 20 152 50 5 2.38 p
Cu-20 glass 213 152 100 5 11.5 p
Cu-21 glass 1220 152 50 5 130 p
Cu-22 glass 58 p
Cu-23 Si 1087 152 100 5 60 p
Cu-24 thick glass 975 152 100 5 60 p
Cu-24 b thick glass 644 152 150 5 25 p
Cu-25 thick glass 621 152 250 5 15 p
Cu-26 thick glass 674 152 250 5 15 p
Cu-27 thick glass 427 152 50 5 45 p
Cu-28 thick glass 422 152 50 5 45 p
Cu-29 thick glass 589 152 50 5 60 p
Cu-30 thick glass 626 152 250 5 15 p
Cu-32 thick glass 127 65 2 80
Cu-33 Si 127 64 2 80
Cu-34 Si 127 64 2 210
Cu-35 thick glass 1605 152 150 3 55 x, p
Cu-36 thick glass < 100 nm 152 150 3 55 p
Cu-37 thick glass 141 152 150 3 55 p
Cu-38 thick glass 168 152 150 3 55 p
147

Cu-39 thick glass 105 118 150 3 35 p
Cu-40 thick glass 124 118 150 3 35 p
Cu-41 thick glass 118 150 10 2
Cu-42 thick glass 118 150 10 2
Cu-43 thick glass 489 118 150 10 35 p
Cu-44 thick glass 500 118 150 10 35 p
Cu-45 thick glass 188 118 150 5 35 p
Cu-46 thick glass 302 118 150 5 35 p
Cu-47 thick glass 1934 118 150 5 35 p
Cu-48 thick glass 45 118 150 5 35 p
Cu-49 thick glass 1931 118 150 10 35 p
Cu-50 thick glass 137 118 150 10 35 p
Cu-51 thick glass 1555 152 150 10 55 p
Cu-52 thick glass 148 152 150 10 55 p
Cu-53 thick glass 501 152 150 10 55 p
Cu-54 thick glass 518 152 150 10 55 p
Cu-55 Si 1010 127 200 4 30.5 x, p
Cu-56 thick glass 982 127 200 4 30.5 p
Cu-57 thick glass 430 152 100 3 30 p
Cu-58 thick glass 421 152 100 20 30 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 18: Monolithic Hf samples
Hf
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Hf-1 thick glass 655 127 200 3 20 p
Hf-2 Si 499 127 200 3 30 x, s, p
Hf-3 thick glass 494 127 200 3 30 p
Hf-4 Si 1993 127 200 4 120 x, p
Hf-5 thick glass 127 200 4 120
Hf-6 Si 2522 127 200 4 80 x, p
Hf-7 thick glass 127 200 4 80
Hf-8 Si 127 200 10 120 p
Hf-9 thick glass 2662 127 200 10 120 p
Hf-10 Si 127 200 20 120 p
Hf-11 thick glass 3373 127 200 20 120 p
Hf-12 Si 127 200 15 110 p
Hf-13 thick glass 2785 127 200 15 110 p
Hf-14 Si 2823 127 200 15 80 p
Hf-15 Si 2745 127 200 20 80 p
148

Hf-16 Si 2706 127 200 30 90 p
Hf-17 Si 2374 127 100 20 160 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 19: Monolithic Nb samples
Nb
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Nb-1 glass 215 152 50 2 60 p
Nb-2 glass 220 152 100 5 30 p
Nb-3 Si 152 100 5 140 x, n, p
Nb-4 Si 1013 152 100 5 140 p
Nb-5 Si 152 100 5 69 x, p
Nb-6 Si 488 152 100 5 69 p
Nb-7 Si 152 100 5 28 x, p
Nb-8 Si 198 152 100 5 28 p
Nb-9 Si 152 100 5 14 x, p
Nb-10 Si 98 152 100 5 14 p
Nb-11 152 100 5 8 x, p
Nb-12 Si 60 152 100 5 8 p
Nb-13 Si 91 152 100 36 20 p
Nb-14 Si 225 152 100 5 30 x, p
Nb-15 Si 154 152 50 5 40 p
Nb-16 glass 572 152 100 5 73 p
Nb-17 glass 152 100 5 70
Nb-18 glass 481 152 100 5 70 p
Nb-18
b
glass 43 p
Nb-19 glass 581 152 100 5 70 p
Nb-20 glass 506 152 100 5 70 p
Nb-21 glass 560 152 100 5 70 p
Nb-22 Si 152 100 5 150 x, n, p
Nb-24 glass 423 152 100 5 55 p
Nb-25 thick glass 444 152 100 5 55 p
Nb-26 thick glass 437 152 100 5 50 p
Nb-27 thick glass 503 152 100 5 60 p
Nb-28 thick glass 428 152 100 5 50 p
Nb-29 thick glass 461 152 100 5 60 p
Nb-30 thick glass 919 152 200 5 66 p
Nb-31 Si 152 100 5 120 x, s, a, p
Nb-32 thick glass 960 152 100 5 120 p
149

Nb-33 Si 152 100 3 128 x, s, a, p
Nb-34 thick glass 1013 152 100 3 128 p
Nb-35 Si 152 50 3 244 x, s, a, p
Nb-36 thick glass 874 152 50 3 244 p
Nb-37 Si 152 200 3 70 x, s, a, p
Nb-38 thick glass 1055 152 200 3 70 p
Nb-39 Si 127 200 5 50 x, s, a, p
Nb-40 thick glass 1146 127 200 5 50 p
Nb-41 Si 127 200 3 14.66 x, s, p
Nb-42 thick glass 329 127 200 3 14.66 p
Nb-43 Si 127 200 3 46 x, s, a, p
Nb-44 thick glass 954 127 200 3 46 p
Nb-45 Si 127 250 5 41 x, s, a, p
Nb-46 thick glass 1083 127 250 5 41 p
Nb-47 Si 118 250 3 33 x, s, a, p
Nb-48 thick glass 938 118 250 3 33 p
Nb-49 Si 118 250 2 33 x, s, a, p
Nb-50 thick glass 850 118 250 2 33 p
Nb-51 thick glass 564 127 200 3 25 p
Nb-52 thick glass 482 127 200 3 30 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 20: Monolithic Ta samples
Ta
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed
Ta-1 glass 250 152 100 5 26 profilometry
Ta-2 Si 206 152 100 5 21 profilometry
Ta-3 glass 211 152 100 5 21 profilometry
Ta-4 Si 923 127 200 4 60 profilometry
Ta-5 glass 866 127 200 4 60 profilometry
Ta-6 Si 2326 127 200 5 120 profilometry
Ta-7 glass 1147 127 200 5, 10 2 profilometry

150

Table 21: Monolithic Ti samples
Ti
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Ti-1 glass 159 152 50 2 68.25 p
Ti-2 glass 218 152 100 5 40 p
Ti-3 Si 152 100 5 180 x, p
Ti-4 Si 886 152 100 5 180 p
Ti-5 Si 152 100 5 91 x, p
Ti-6 Si 453 152 100 5 91 p
Ti-7 Si 152 100 5 37 x, p
Ti-8 Si 188 152 100 5 37 p
Ti-9 Si 152 100 5 18 x, p
Ti-10 Si 90 152 100 5 18 p
Ti-11 Si 152 100 5 9 x, p
Ti-12 Si 35 152 100 5 9 p
Ti-13 Si 152 100 5 10 x, p
Ti-14 Si 44 152 100 5 10 p
Ti-15 thick glass 502 127 200 3 36 p
Ti-18 Si 382 127 100 5 59 x, p
Ti-19 thick glass 373 127 100 5 59 p
Ti-20 Si 675 127 200 5 60 x, p
Ti-21 thick glass 798 127 200 5 60 p
Ti-22 Si 127 200 4 150 x, p
Ti-23 thick glass 1658 127 200 4 150 p
Ti-24 Si 51 50 5 480 s, p
Ti-25 thick glass 5816 51 50 5 480 p
Ti-26 Si 289 178 100 5 80 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 22: Monolithic Zr samples
Zr
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed
Zr-1 thick glass 335 127 100 10 30 profilometry
Zr-2 thick glass 602 127 200 5 30 profilometry
Zr-3 Si 219 127 100 10 20 profilometry

151

Appendix I.2 Monolithic Multi-Element Films on Si and Glass Substrates
Table 23: Monolithic CuAl samples
Cu Al
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
CuAl 1 thick glass 1333 102 200 5 20 152 25 5 20 e, p
CuAl 2 thick glass 1411 102 200 5 20 152 50 5 20 e, p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 24: Monolithic CuZr samples
Cu Zr
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
CuZr1 thick glass 592 152 100 20 21 127 163 20 21 x, e, p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 25: Monolithic HfTi samples
Hf Ti
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
HfTi 3 Si 127 200 5 14 127 45 5 14
HfTi 4 thick glass 127 200 5 14 127 45 5 14
HfTi 5 Si 1942 127 200 3 90 127 45 3 90 x, s, e, p
HfTi 6 thick glass 127 200 3 90 127 45 3 90
HfTi 7 Si 1954 127 200 4 90 127 100 4 90 x, s, e, p
HfTi 8 thick glass 127 200 4 90 127 100 4 90
HfTi 9 Si 2431 127 200 4 90 127 200 4 90 x, s, e, t, p
HfTi 10 thick glass 127 200 4 90 127 200 4 90
HfTi 13 Si 102 200 5 120 102 150 5 120 x, e
HfTi 14 thick glass 102 200 5 120 102 150 5 120
HfTi 15 Si 4553 102 200 4 120 102 100 4 120 x, e, t, p
HfTi 16 thick glass 102 200 4 120 102 100 4 120
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

152

Table 26: Monolithic MoAu samples
Mo Au
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Mo-Au 1 Si 102 200 5 30 102 35 5 30 s, e, p
Mo-Au 2 glass 1176 102 200 5 30 102 35 5 30 p
Mo-Au 3 Si 102 200 5 30 102 25 5 30 e, p
Mo-Au 4 glass 926 102 200 5 30 102 25 5 30 p
Mo-Au 5 Si 3443 102 200 5 90 102 25 5 90 x, e, p
Mo-Au 6 glass 3280 102 200 5 90 102 25 5 90 p
Mo-Au 7 Si 3503 102 200 5 90 127 25 5 90 x, e, p
Mo-Au 8 glass 3315 102 200 5 90 127 25 5 90 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 27: Monolithic TaCu samples
Ta Cu
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Ta-Cu 1 Si 1940 127 200 4 120 127 20 4 120 x, s, e, p
Ta-Cu 2 thick glass 127 200 4 120 127 20 4 120 p
Ta-Cu 3 Si 2243 127 200 4 120 127 30 4 120 x, s, e, p
Ta-Cu 4 thick glass 127 200 4 120 127 30 4 120 p
Ta-Cu 5 Si 4342 127 200 4 110 127 150 4 110 x, s, e, p
Ta-Cu 6 thick glass 127 200 4 110 127 150 4 110 p
Ta-Cu 7 Si 3566 127 200 4 120 127 90 4 120 x, p
Ta-Cu 8 thick glass 127 200 4 120 127 90 4 120 p
Ta-Cu 9 Si 3435 127 200 5 120 127 90 5 120 x, s, e, p
Ta-Cu 10 thick glass 127 200 5 120 127 90 5 120 p
Ta-Cu 11 Si 127 200 5 90 127 90 5 90 s, e, p
Ta-Cu 12 thick glass 2597 127 200 5 90 127 90 5 90 p
Ta-Cu 13 Si 2441 127 200 5 120 127 50 5 120 x, e, p
Ta-Cu 14 thick glass 127 200 5 120 127 50 5 120 p
Ta-Cu 15 Si 2996 127 200 5 120 127 65 5 120 x, e, p
Ta-Cu 16 thick glass 2712 127 200 5 120 127 65 5 120 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 28: Monolithic TaHf samples
Ta Hf
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed*
Ta-Hf 1 Si 127 200 4 120 127 50 4 120 x, s, e, p
Ta-Hf 2 thick glass 2157 127 200 4 120 127 50 4 120 p
Ta-Hf 3 Si 3934 102 200 4 120 102 50 4 120 e, p
Ta-Hf 4 thick glass 102 200 4 120 102 50 4 120 p
Ta-Hf 5 Si 102 200 5 120 102 57 5 120 p
Ta-Hf 6 thick glass 4295 102 200 5 120 102 57 5 120 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB
153

Table 29: Monolithic TiHf samples*
Ti Hf
Sample
Name
Substrate
Material
Total
thickness
(nm)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Sputtering
distance
(mm)
Power
(W)
Argon
Pressure
(mTorr)
Time
(min)
Characterization
performed**
Ti-Hf 1 Si 1581 152 200 5 100 127 55 5 100 x, s, e, p
Ti-Hf 2 thick glass 1504 152 200 5 100 127 55 5 100 p
Ti-Hf 3 Si 152 200 4 120 127 30 4 120 x, s, e, p
Ti-Hf 4 thick glass 1600 152 200 4 120 127 30 4 120 p
Ti-Hf 5 Si 127 200 5 150 127 27 5 150 x, s, e, p
Ti-Hf 6 thick glass 2155 127 200 5 150 127 27 5 150 p
*Ti-Hf 5 and Ti-Hf 6 have 50 nm-thick Hf seed layers
** x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Appendix I.3 Multilayered Samples on Si and Glass Substrates
Table 30: Multilayered Al/Nb samples
Al Nb
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed*
Al/Nb 3 Si 5 11 100 12 1255 x, n, p
Al/Nb 4 glass 1189 p
Al/Nb 5 Si 5 11 100 12 1255 x, n, p
Al/Nb 6 glass 1153 p
Al/Nb 7 Si 1 11 100 12 1211 x, n, p
Al/Nb 8 glass 1091 p
Al/Nb 9 Si 1 11 100 12 1211 x, n, p
Al/Nb 10 glass
Al/Nb 11 Si 20 10 100 11 1300 x, t, n, p
Al/Nb 12 glass 1150 p
Al/Nb 13 Si 20 10 100 11 1300 x, n, p
Al/Nb 14 glass 1121 p
Al/Nb 15 Si 2 11 100 12 1222 x, n, p
Al/Nb 16 glass 1080 p
Al/Nb 17 Si 2 11 100 12 1222 x, t, n, p
Al/Nb 18 glass 1090 p
Al/Nb 19 Si 10 11 100 12 1310 x, t, n, p
Al/Nb 20 glass 1109 p
Al/Nb 21 Si 10 11 100 12 1310 x, n, p
Al/Nb 22 glass 1090 p
Al/Nb 23 Si 5 11 100 12 1255 x, t, n, p
Al/Nb 24 glass 1218 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB
154

Table 31: Multilayered Cu/Nb samples*
Cu Nb
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed**
Cu/Nb-1 Si 20 5 100 6 700 x, s, n, p, f
Cu/Nb-2 Si 10 10 100 11 1200 x, n, p
Cu/Nb-3 Si 5 3 100 4 415 x, p
Cu/Nb-4 Si 5,2 1,3 100 5 x, p
Cu/Nb-5 Si 2 4 100 5 508 x, n, p
Cu/Nb-6 Si 1 4 100 5 504 x, n, p
Cu/Nb-7 Si 2 20 20 21 460 x, p
Cu/Nb-9 Si 1 20 20 21 440 x, p
Cu/Nb-10 glass 430 x, p
Cu/Nb-11 Si 2 20 20 20 540 x, p
Cu/Nb-12 glass 479 x, p
Cu/Nb-13 Si 2 20 20 20 540 x, p
Cu/Nb-14 glass 480 x, p
Cu/Nb-15 Si 5 8 100 9 940 x, p
Cu/Nb-16 glass 805 x, p
Cu/Nb-17 Si 5 8 100 9 940 x
Cu/Nb-18 glass 877 x, p
Cu/Nb-19 Si 5 8 100 9 940 x, p
Cu/Nb-20 glass 801 x, p
Cu/Nb-21 Si 1 50 5 50 300 x, p
Cu/Nb-22 glass 313 x, p
Cu/Nb-23 Si 1 40 10 40 490 x, s, p
Cu/Nb-24 glass 463 p
Cu/Nb-25 Si 1 50 5 50 350 x, p
Cu/Nb-26 glass 325 p
Cu/Nb-27 Si 1 40 10 40 476 x, p
Cu/Nb-28 glass 446 x, p
Cu/Nb-29 Si 20 10 100 11 1300 s, e, t, n, p
Cu/Nb-30 glass 1123 p
Cu/Nb-31 Si 2 11 100 12 1222 x, s, n, p
Cu/Nb-32 glass 1039 p
Cu/Nb-33 Si 1 11 100 12 1211 x, n, p
Cu/Nb-34 glass 1012 p
Cu/Nb-35 Si 10 11 100 12 1310 x, s, n, p
Cu/Nb-36 glass 1100 p
Cu/Nb-37 Si 10 1 100 1.5 160 s
155

Cu/Nb-38 glass
Cu/Nb-39 Si 10 11 100 12 1310 x, t, n, p
Cu/Nb-40 glass 1161 p
Cu/Nb-41 Si 1 11 100 12 1211 x, n, p
Cu/Nb-42 glass 1078 p
Cu/Nb-43 Si 20 10 100 11 1300 x, s, e, t, n, p
Cu/Nb-44 glass 1205 p
Cu/Nb-45 Si 2 11 100 12 1222 x, t, n, p
Cu/Nb-46 glass p
Cu/Nb-47 Si 5 11 100 12 1255 x, t, n, p
Cu/Nb-48 glass 1049 p
Cu/Nb-49 Si 5 11 100 12 1255 x, n, p
Cu/Nb-50 glass 1060 p
Cu/Nb-51 Si 20 10 100 11 1300 s, e, n, p
Cu/Nb-52 glass 1183 p
Cu/Nb-53 Si 20 10 100 11 1300 s, x, n, p
Cu/Nb-54 glass 1210 p
Cu/Nb-55 Si 2 11 100 12 1222 x, n, p
Cu/Nb-56 glass 1058 p
Cu/Nb-57 Si 10 11 100 12 1310 x, n, p
Cu/Nb-58 glass 1099 p
Cu/Nb-59 Si 1 11 100 12 1211 x, n, p
Cu/Nb-60 glass 1021 p
*these Cu/Nb samples had the following seed layers: Cu/Nb-11, 12, 13, 14 - 100 nm Nb, Cu/Nb-
23, 24, 25, 26 - 50 nm Nb, Cu/Nb 27, 28 - 36 nm Nb
** x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 32: Multilayered Cu/Ta samples*
Cu Ta
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed**
Cu/Ta 1 Si 1 40 10 40 488 x, p
Cu/Ta 2 glass 1 40 10 40 488 x, p
Cu/Ta 3 Si 1 30 10 30 378 x, p
Cu/Ta 4 glass 1 30 10 30 376 x, p
*All Cu/Ta samples had 50 nm-thick Ta seed layers
** x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

156

Table 33: Multilayered Hf/Ti samples
Hf Ti
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed*
Hf-Ti 1 Si 50 10 2-50 x, t, p
Hf-Ti 2 thick glass 51 10 2-50 1012 x, p
Hf-Ti 11 Si 50 5 x, t, p
Hf-Ti 12 thick glass 1748 p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 34: Multilayered HfTi/Ti samples
Hf-Ti Ti
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed*
Hf-Ti 17 Si 20 2 1261 x, e, t, p
Hf-Ti 18 thick glass 20 2
Hf-Ti 19 Si 21 3 x, e
Hf-Ti 20 thick glass 21 3
Hf-Ti 21 Si 5 0.5 1451 e, t, p
Hf-Ti 22 Si 50 5 1902 e, t, p
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 35: Multilayered MoAu/Au samples
Mo-Au Au
Sample
Name
Substrate
Material
Thickness of
layer (nm)
Number of
layers
Thickness of
layer (nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed*
Mo-Au 9 Si 20 2 x, e
Mo-Au 10 glass 20 2
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

157

Appendix I.4 Samples on Polyimide Substrates
Table 36: Multilayered Al/Nb samples on polyimide substrates
Al Nb
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Al/Nb-Poly 1 polyimide 20 2 100 3 340
Al/Nb-Poly 2 polyimide 20 2 100 3 340
Al/Nb-Poly 3 polyimide 20 2 100 3 340

Table 37: Cu samples on polyimide substrates
Sample
Name
Substrate
Material
Thickness
of layer
(nm)
Sputtering
distance
(mm)
Power (W)
Argon
Pressure
(mTorr)
Time
(min)
Poly-1 polyimide 540 152 100 5 30
Poly-2 polyimide 540 152 100 5 30
Cu-Poly 1 polyimide 1040 152 100 5 56
Cu-Poly 2 polyimide 1040 152 100 5 56
Cu-Poly 3 polyimide 500 152 100 3 26.88
Cu-Poly 4 polyimide 500 152 250 3 11.86
Cu-Poly 5 polyimide 500 152 100 3.5 26.5
Cu-Poly 6 polyimide 500 152 250 3.5 11.66

Table 38: Multilayered Cu/CuAl samples on polyimide substrates
Cu CuAl
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed
Cu-CuAl Poly 2 polyimide 20 26 20 25 1020
Cu-CuAl Poly 3 polyimide 15 26 15 25 765 TEM

Table 39: Multilayered Cu/CuTa sample on a polyimide substrate
Cu CuTa
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed
Cu-CuTa Poly 1 polyimide ~10 37 ~10 36 730 FIB

158

Table 40: Multilayered Cu/CuZr samples on polyimide substrates*
Cu CuZr
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Cu/CuZr Poly 1 polyimide 10 6 10 7 130
Cu/CuZr Poly 2 polyimide 5 12 10 13 190
Cu/CuZr Poly 3 polyimide 2 30 10 31 370
Cu/CuZr Poly 4 polyimide 10 6 10 7 130
Cu/CuZr Poly 5 polyimide 20 3 10 4 100
Cu/CuZr Poly 6 polyimide 5 12 10 13 190
Cu/CuZr Poly 7 polyimide 10 6 10 7 130
Cu/CuZr Poly 8 polyimide 20 3 10 4 100
Cu/CuZr Poly 9 polyimide 2 25 10 26 310
Cu/CuZr Poly 10 polyimide 60 1 10 2 80
Cu/CuZr Poly 11 polyimide 20 3 10 4 100
Cu/CuZr Poly 12 polyimide 10 6 10 6 120
Cu/CuZr Poly 13 polyimide 10 12 5 12 180
Cu/CuZr Poly 14 polyimide 10 6 20 6 180
Cu/CuZr Poly 15 polyimide 10 20 2 20 240
*XRD was performed on Cu/CuZr Poly 1, 6, 14

Table 41: Multilayered Cu/Nb samples on polyimide substrates
Cu Nb
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed*
Cu/Nb-Poly 1 polyimide 20 8 100 9 1060
Cu/Nb-Poly 2 polyimide 20 8 100 9 1060
Cu/Nb-Poly 3 polyimide 5 9 100 10 1045
Cu/Nb-Poly 4 polyimide 20 1 100 2 220
Cu/Nb-Poly 5 polyimide 20 3 100 4 460
Cu/Nb-Poly 6 polyimide 20 3 100 4 460
Cu/Nb-Poly 8 polyimide 20 3 100 4 460
Cu/Nb-Poly 9 polyimide 20 2 100 3 340
Cu/Nb-Poly 10 polyimide 2 2 100 3 304
Cu/Nb-Poly 11 polyimide 2 2 100 3 304
Cu/Nb-Poly 12 polyimide 5 8 10 9 130
Cu/Nb-Poly 13 polyimide 2 30 10 31 370 x, t
Cu/Nb-Poly 14 polyimide 10 6 10 7 130
Cu/Nb-Poly 15 polyimide 5 12 10 13 190 x, t
Cu/Nb-Poly 16 polyimide 20 3 10 4 100 x, s, t
159

Cu/Nb-Poly 17 polyimide 10 6 10 7 130 x, t
Cu/Nb-Poly 18 polyimide 10 13 5 12 190
Cu/Nb-Poly 19 polyimide 10 12 5 12 180
Cu/Nb-Poly 20 polyimide 10 12 5 12 180
Cu/Nb-Poly 21 polyimide 10 6 10 6 120 x
Cu/Nb-Poly 22 polyimide 10 3 20 3 90
Cu/Nb-Poly 23 polyimide 10 30 2 30 360
Cu/Nb-Poly 24 polyimide 10 12 5 12 180
Cu/Nb-Poly 25 polyimide 10 6 10 6 120
Cu/Nb-Poly 26 polyimide 10 3 20 3 90
Cu/Nb-Poly 27 polyimide 10 22 2 22 264
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Table 42: Multilayered Cu/Zr samples on polyimide substrates*
Cu Zr
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total Cu/Zr
thickness
(nm)
Cu/Zr Poly 1 polyimide 5 12 10 13 190
Cu/Zr Poly 2 polyimide 10 6 10 7 130
Cu/Zr Poly 3 polyimide 20 3 10 4 100
Cu/Zr Poly 4 polyimide 5 12 10 13 190
Cu/Zr Poly 5 polyimide 2 20 10 21 250
Cu/Zr Poly 6 polyimide 5 12 10 13 190
Cu/Zr Poly 7 polyimide 10 6 10 7 130
Cu/Zr Poly 8 polyimide 5 12 10 13 190
Cu/Zr Poly 9 polyimide 10 6 10 7 130
Cu/Zr Poly 10 polyimide 20 3 10 4 100
Cu/Zr Poly 11 polyimide 2 20 10 21 250
Cu/Zr Poly 12 polyimide 60 1 10 2 80
Cu/Zr Poly 13 polyimide 2 12 10 13 154
Cu/Zr Poly 14 polyimide 10 12 5 12 180
Cu/Zr Poly 15 polyimide 10 6 10 6 120
Cu/Zr Poly 16 polyimide 10 3 20 3 90
Cu/Zr Poly 17 polyimide 10 25 2 25 300
Cu/Zr Poly 18 polyimide 10 1 60 1 70
*XRD was performed on Cu/Zr Poly 1; Cu/Zr Poly 8 through 18 had 5 nm-thick Ti seed layers

160

Table 43: Multilayered Cu/Nb samples on polyimide substrates tested by in-situ XRD tensile
testing
Cu Nb
Sample Name
Substrate
Material
Thickness
of layer
(nm)
Number of
layers
Thickness
of layer
(nm)
Number of
layers
Total
thickness
(nm)
Characterization
performed
Cu/Nb-Poly 12-1 polyimide 5 8 10 9 130 SEM
Cu/Nb-Poly 13-1 polyimide 2 30 10 31 370 SEM
Cu/Nb-Poly 15-1 polyimide 5 12 10 13 190 SEM
Cu/Nb-Poly 16-1 polyimide 20 3 10 4 100 SEM
Cu/Nb-Poly 17-1 polyimide 10 6 10 7 130 SEM

Appendix I.5 Heat-treated Samples
Table 44: Conducted heat treatments
Heat
treatment
name
Temperature
(°C)
Time
(h)
Quenched Samples
HT1 500 2 no Ti-20, Hf-2, Hf-Ti 5, Ti-Hf 1
HT2 800 10 no Hf-Ti 9
HT3 800 96 no Hf-Ti 9, Hf-4
HT4 800 96 no Hf-Ti 11
HT5 800 96 no Hf-10, Ta-Hf 1, Ti-Hf 3
HT6 800 96 no Ta-Hf 1, Ti-Hf 3, Hf-10, Hf-12
HT7 800 96 no Hf-14, Hf-15, Hf-16, Hf-17
HT8 1000 93 no Hf-Ti 15, Hf-Ti 9, Ti-Hf 5, Hf-Ti 9 HT3
HT9 800 1 no
Hf-Ti 7, Ta-Cu 15, Hf-Ti 17, Ti-Hf 5, Ta-
Hf 5, Mo-Au 7, Hf-Ti 15
HT11 800 96 yes
Ta-Cu 15 HT9, Ti-Hf 5 HT9, Mo-Au 7
HT9,Hf-Ti 17 HT9b, Hf-Ti 15 HT9
HT12 500 166 no
Ta-Cu 15 HT9, Hf-Ti 15 HT9, Ti-Hf 5
HT9, Hf-Ti 17 HT9b, Mo-Au 7 HT9
HT13 800 1 no Mo-Au 5, Hf-Ti 13, Ta-Cu 11, Ta-Hf 5
HT16 800 1 no Hf-Ti 19, Mo-Au 9
HT17 800 96 yes
Hf-Ti 19 HT16, Mo-Au 9 HT16, Hf-Ti
13 HT13
HT18 600 0 no Hf-Ti 22, Mo-Au 9, Hf-Ti 17, Hf-Ti 21
HT19 800 96 yes Hf-Ti 21 HT18, Hf-Ti 22 HT18
HT20 800 96 yes
Mo-Au 9 HT18 Etch, Hf-Ti 21 HT 18
Etch, Hf-Ti 22 HT18 Etch, Hf-Ti 17
HT9b Etch
161

Table 45: Heat-treated samples
Sample name
Characterization
performed*
Sample name
Characterization
performed*
Hf-2 HT1 x, s

Mo-Au 7 HT9
Hf-2 HT2 Mo-Au 7 HT10
Hf-4 HT3 t Mo-Au 7 HT11
Hf-10 HT5 Mo-Au 9 HT16
Hf-10 HT6 t Mo-Au 9 HT18
Hf-12 HT6 Mo-Au 9 HT18 Etch HT20 t
Hf-14 HT7 s, e
Hf-15 HT7 Ta-Cu 15 HT9
Hf-16 HT7 s, e, t, f Ta-Cu 15 HT11 e, t
Hf-17 HT7
Ta-Hf 1 HT5
Hf-Ti 5 HT1 x, s, e Ta-Hf 1 HT6
Hf-Ti 9 HT2 s, e Ta-Hf 5 HT9
Hf-Ti 9 HT3 t, f
Hf-Ti 9 HT3 HT8 Ti-20 HT1
Hf-Ti 9 HT8 t
Hf-Ti 11 HT4 t Ti-Hf 1 HT1 x, s, e
Hf-Ti 13 HT15 t Ti-Hf 3 HT5
Hf-Ti 15 HT8 Ti-Hf 3 HT6
Hf-Ti 15 HT9 Ti-Hf 5 HT8
Hf-Ti 15 HT10 Ti-Hf 5 HT9
Hf-Ti 15 HT11 e, t Ti-Hf 5 HT10
Hf-Ti 17 HT9a Ti-Hf 5 HT11
Hf-Ti 17 HT9b
Hf-Ti 17 HT10
Hf-Ti 17b HT11 e, t
Hf-Ti 17 HT15 t
Hf-Ti 17 HT18
Hf-Ti 17 H9b Etch HT20 t
Hf-Ti 19 HT16
Hf-Ti 21 HT18 t
Hf-Ti 21 HT18 HT19 t
Hf-Ti 21 HT18 Etch HT20
Hf-Ti 22 HT18
Hf-Ti 22 HT18 HT19 t
Hf-Ti 22 HT18 Etch HT20 t
* x-XRD, s-SEM, e-EDX, t-TEM, p-profilometry, n-nanoindentation, a-AFM, f-FIB

Abstract (if available)

Abstract Nano metallic multilayers (NMMs) are nanostructured composite materials which consist of thin (< 100 nm) alternating layers of metals. They have many attractive mechanical, electromagnetic, and optical properties. These properties can be tailored by adjusting the layer and interface parameters of the NMMs. In addition, the ability to tailor these sample characteristics allows for NMMs to be used as model systems in the study of a variety of nanostructure phenomena. In the studies described in this dissertation, the structure and behavior of NMMs under mechanical and thermal loading were investigated. Several general results are as follow: 1) A novel transmission electron microscopy technique revealed how 20 nm thick layers of Cu and Al more greatly refined the overall structure of Cu/Nb and Al/Nb NMMs than 2 nm layers, 2) The load-sharing behavior among sub-20 nm thick layers of Cu and Nb within Cu/Nb NMMs was deconvoluted, and, by tracking the changes in deformation behavior for varying layer thicknesses, a competing interplay of strengthening and weakening was demonstrated for the different materials, 3) The high-temperature behavior of nanostructured Hf-Ti was investigated, with an observed increased thermal stability for Hf-Ti NMMs as compared to monolithic Hf-Ti films, showing that the starting configuration may strongly influence the stability of a nanostructured system. Overall, these studies demonstrated the ability to more fully characterize NMMs and configure them for optimized behavior under mechanical and thermal loading.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Synthesis and mechanical behavior of highly nanotwinned metals

PDF

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

PDF

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

PDF

Mechanical behavior and deformation of nanotwinned metallic alloys

PDF

Synthesis, characterization, and mechanical properties of nanoporous foams

PDF

Development and characterization of hierarchical cellular structures

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

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

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

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

Thermal properties of silicon carbide and combustion mechanisms of aluminum nanoparticle

PDF

3D printing and compression testing of biomimetic structures

PDF

Ageing and mechanical failure of fiber reinforced polymers

$On the dynamic fracture behavior of polymeric materials subjected to extreme conditions$

PDF

On the dynamic fracture behavior of polymeric materials subjected to extreme conditions

PDF

Development and characterization of transparent metal/ceramic and ceramic/ceramic nanomultilayers

PDF

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

PDF

Tailoring compositional and microstructural complexity in nanostructured alloys

PDF

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

PDF

The effect of lattice structure and porosity on thermal conductivity of additively-manufactured porous materials

Asset Metadata

Creator Polyakov, Mikhail N. (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Mechanical Engineering

Publication Date 09/11/2015

Defense Date 06/24/2015

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

Tag deformation,in-situ,microstructure,multilayers,Nano,OAI-PMH Harvest,thermal stability

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 (committee member), Ravichandran, Jayakanth (committee member)

Creator Email mnpolyakov@gmail.com,mpolyako@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-178750

Unique identifier UC11275932

Identifier etd-PolyakovMi-3895.pdf (filename),usctheses-c40-178750 (legacy record id)

Legacy Identifier etd-PolyakovMi-3895.pdf

Dmrecord 178750

Document Type Dissertation

Format application/pdf (imt)

Rights Polyakov, Mikhail N.

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